This textbook presents the classical treatment of the problems of heat transfer in an exhaustive manner with due emphasi
549 29 26MB
English Pages 1162 Year 2020
Table of contents :
Preface
Contents
About the Author
List of Symbols
Dimensionless Numbers
Greek Symbols
Superscript and Subscript
Space Coordinates
1 Introduction
1.1 Introduction
1.2 Heat Transfer by Conduction
1.3 Heat Transfer by Convection
1.4 Heat Transfer by Radiation
1.5 Simultaneous Heat Transfer
1.6 Summary
Review Questions
2 OneDimensional SteadyState Heat Conduction
2.1 Introduction
2.2 Temperature Field and Temperature Gradient
2.3 Thermal Conductivity
2.3.1 Thermal Conductivity of Solids
2.3.2 Thermal Conductivity of Metals and Alloys
2.3.3 Thermal Conductivity of Construction and HeatInsulating Materials
2.3.3.1 RValues of Insulating Materials
2.3.4 Thermal Conductivity of Gases
2.3.5 Thermal Conductivity of Liquids
2.4 General Heat Conduction Equations
2.4.1 General Heat Conduction Equation in Cartesian Coordinates
2.4.1.1 Thermal Diffusivity
2.4.2 General Heat Conduction Equation in Cylindrical Coordinates
2.4.3 General Heat Conduction Equation in Spherical Coordinates
2.5 OneDimensional SteadyState Heat Conduction
2.5.1 Composite Plane Wall
2.5.2 OneDimensional SteadyState Heat Conduction Through a Plane Homogeneous Wall Considering Film Coefficients
2.5.2.1 Conduction Heat Transfer Through a Composite Plane Wall Considering Film Coefficients
2.5.3 OneDimensional SteadyState Conduction Heat Transfer Through a Plane Homogeneous Wall Considering Heat Transfer by Convection and Radiation from the Wall Surface
2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
2.6.1 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell Considering Film Coefficients
2.6.1.1 Overall Heat Transfer Coefficient
2.6.2 Composite Cylindrical Wall
2.7 OneDimensional SteadyState Heat Conduction Through a Spherical Shell
2.7.1 OneDimensional SteadyState Heat Conduction Through a Spherical Shell Considering the Film Coefficients
2.7.2 Composite Spherical Shell
2.8 Measurement of Thermal Conductivity
2.8.1 Thermal Conductivity Measurement of Solids
2.8.2 Thermal Conductivity Measurements of Liquids and Gases
2.9 Effect of Variable Thermal Conductivity
2.9.1 Plane Wall
2.9.2 Cylindrical Shell
2.9.3 Spherical Shell
2.10 Critical Thickness of Insulation
2.10.1 Critical Thickness of Insulation for Cylinders
2.10.2 Critical Thickness of Insulation for Spherical Vessel
2.11 Thermal Contact Resistance
2.12 Summary
Review Questions
Problems
References
3 Extended Surfaces (Fins)
3.1 Introduction
3.2 Heat Transfer from a Fin of Uniform CrossSection
3.2.1 A Very Long Fin
3.2.1.1 Comments on Parameter M
3.2.2 Negligible Heat Transfer from the Fin End as Compared to the Heat Transferred from the Fin Surface (Ac ≪ PL)
3.2.3 Short Fins (Fin with Heat Loss from the Fin End)
3.3 Hollow Fins
3.4 Composite Fins
3.5 Effectiveness and Efficiency of Fins
3.6 Heat Transfer from a Finned Wall
3.7 Intensification of Heat Transfer by Finning
3.8 Error in Temperature Measurement with Thermometer Well
3.9 When Fins Are to Be Used?
3.10 Heat Transfer from a Bar Connected to Two Heat Sources at Different Temperatures
3.11 Generalized Equation of Fin
3.12 Fin of Minimum Weight (Isachenko et al.1977)
3.13 Straight Fin of Triangular Section
3.14 Straight Fin of Trapezoidal Section
3.15 Annular Fin
3.16 Fin Efficiency Plots
3.17 Summary
Review Questions
Problems
References
4 Conduction with Heat Generation
4.1 Plane Wall with Uniform Heat Generation
4.1.1 Case (A) Surfaces at Different Temperatures
4.1.1.1 The Maximum Temperature and Its Location Within the Wall
4.1.2 Case (B) Surfaces at the Same Temperature
4.2 Cylinder with Uniform Heat Generation
4.3 Solid Sphere with Uniform Heat Generation
4.4 Heat Transfer Through Piston Crown
4.5 Summary
Review Questions
Problems
5 SteadyState TwoDimensional Heat Conduction
5.1 Introduction
5.2 Analytical Solution of TwoDimensional Heat Conduction Problems
5.3 Conduction Through a Flat Semiinfinite Homogeneous Plate
5.4 Mean Value Theorem
5.5 Graphical Analysis of Two Dimensional SteadyState Conduction: Thermal Flux Plotting and Shape Factor
5.6 Experimental Investigation of Conduction Process by Method of Analogy: ElectroThermal Analogy
5.7 Numerical Solution Methods
5.7.1 FiniteDifference Method
5.7.2 Solution of Nodal Equations
5.7.2.1 Relaxation Method
5.7.2.2 Gaussian Elimination
5.7.2.3 The Gauss–Seidel Iteration Method
5.8 TwoDimensional SteadyState Heat Conduction with Heat Generation
5.9 Summary
Review Questions
Problems
6 Unsteady or Transient Heat Conduction
6.1 Introduction
6.2 Lumped Heat Capacity Analysis
6.2.1 Instantaneous and Total Heat Flow
6.2.2 Applicability of the Lumped Heat Capacity Analysis
6.3 Lumped Capacitance, Varying Fluid Temperature
6.4 MultipleLumped Capacity Systems
6.5 Transient Heat Flow in Semiinfinite Solids
6.6 Transient Heat Conduction in Infinite Plate
6.7 Heisler and Grober Charts
6.8 Two and ThreeDimensional Transient Heat Conduction Systems
6.8.1 TwoDimensional Systems
6.8.2 ThreeDimensional Systems
6.9 Numerical Method of Solving Transient Conduction Problems
6.9.1 The Explicit and Implicit Formulations
6.10 The Schmidt Graphical Method for OneDimensional Problems
6.11 Summary
Review Questions
Problems
References
7 Convective Heat Transfer
7.1 Introduction
7.1.1 Natural Convective Heat Transfer
7.1.2 Forced Convection Heat Transfer
7.2 Flow of Fluid Past a Flat Plate
7.3 Flow in Tubes
7.3.1 Laminar Flow Through a Tube
7.3.2 Turbulent Flow Through a Tube
7.4 Equation of Continuity
7.4.1 The Displacement and Momentum Thickness
7.4.2 The Enthalpy and Conduction Thickness
7.5 Momentum Equation of Laminar Boundary Layer Over a Flat Plate
7.5.1 Solution of Momentum Equation (Blasius Solution)
7.6 Integral Momentum Equation of Laminar Boundary Layer Over a Flat Plate: von Karman Solution
7.7 Energy Equation of Laminar Boundary Layer Over a Flat Plate
7.7.1 Pohlhausen’s Solution
7.7.2 von Karman Integral Technique (Integral Analysis of Energy Equation for the Laminar Boundary Layer)
7.8 Turbulent Boundary Layer Over a Flat Surface
7.9 Laminar Flow in Tubes
7.10 Turbulent Flow in Tubes
7.11 Momentum and Heat Exchange in Turbulent Flow (Eddy Viscosity and Eddy Thermal Diffusivity)
7.12 Reynolds Analogy for Flow Past a Flat Surface
7.12.1 Reynolds–Colburn Analogy
7.12.2 Application of Colburn Analogy to Turbulent Heat Transfer from a Flat Plate
7.13 Prandtl–Taylor Modification of Reynolds Analogy for Turbulent Flow Over Flat Plates
7.13.1 von Karman Analogy for Flat Plates
7.14 Reynolds Analogy for Turbulent Flow in Tubes
7.14.1 Prandtl–Taylor Modification of Reynolds Analogy for Turbulent Flow in Tubes
7.14.2 Friction Drag: Flow Over a Flat Plate Parallel to the Flow
7.14.2.1 Laminar Flow
7.14.2.2 Turbulent Flow
7.15 Natural or Free Convection
7.16 Integral Momentum and Energy Equation of Free Convection on a Vertical Plate
7.17 Liquid Metal Heat Transfer for Laminar Flow Over a Flat Plate
7.18 Summary
Review Questions
Problems
References
8 Empirical Relations for Forced Convection Heat Transfer
8.1 Introduction
8.2 Dimensional Analysis
8.3 Dimensional Analysis Applied to Forced Convection
8.3.1 Rayleigh’s Method
8.3.2 Buckingham’s PiMethod
8.3.3 Physical Significance of Dimensionless Numbers
8.4 Experimental Determination of Forced Convection Heat Transfer Coefficient
8.4.1 Uniform Temperature Condition
8.4.2 Uniform Heat Flux Condition
8.5 Friction Factor and Heat Transfer Coefficient Correlations for Circular Ducts
8.5.1 Laminar Flow in Circular Tubes
8.5.1.1 Friction Factor Correlations
8.5.1.2 Heat Transfer Coefficient Correlations
8.5.2 Turbulent Flow in Circular Tubes
8.5.2.1 Friction Factor Correlations
8.5.2.2 Heat Transfer Coefficient Correlations
8.6 Effects of Temperature Varying Properties
8.7 Heat Transfer and Friction in Concentric Circular Tube Annuli and Parallel Plate Duct
8.7.1 Laminar Flow
8.7.2 Turbulent Flow
8.8 Heat Transfer and Friction in Rectangular Duct
8.8.1 Laminar Flow
8.8.2 Turbulent Flow
8.9 Correlations for External Forced Flow Over a Flat Plate
8.9.1 Laminar Flow
8.9.2 Turbulent Flow
8.10 Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
8.10.1 Cylinder in Cross Flow
8.10.2 Flow Around a Sphere
8.10.3 Flow Across Tube Banks
8.11 Heat Transfer in Liquid Metals
8.12 Influence of Duct Wall Roughness in Turbulent Flow
8.13 Summary
Review Questions
Problems
References
9 Empirical Relations for Natural or Free Convection
9.1 Introduction
9.2 Buoyancy Force in Natural Convection
9.3 Dimensional Analysis Applied to Natural Convection
9.3.1 Rayleigh’s Method
9.3.2 Buckingham’s Pi Method
9.3.3 Physical Interpretation of Grashof Number
9.4 Experimental Determination of Natural Convection Heat Transfer Coefficient
9.5 Empirical Relations for Free or Natural Convection
9.5.1 Vertical Plate and Cylinders
9.5.2 Inclined Plate
9.5.3 Horizontal Plate
9.5.4 Horizontal Cylinder of Diameter D and Length L greaterthan greaterthan d
9.5.5 Sphere of Diameter d
9.6 Free Convection in Parallel Plate Channels
9.6.1 Vertical Channels
9.6.2 Inclined Channels
9.7 Empirical Correlations for Enclosed Spaces
9.8 Combined Free and Forced Convection (Kays and Crawford 1980; Gebhart 1961; Holman 1992; Cengel 2007)
9.9 Summary
Review Questions
Problems
References
10 Laws of Thermal Radiation
10.1 Introduction
10.2 Reflection, Absorption and Transmission of Radiation
10.3 Emissivity and a Perfect Blackbody
10.4 Planck’s Spectral Distribution of Emissive Power
10.5 Wein’s Displacement Law
10.6 Total Emissive Power: Stefan–Boltzmann Law
10.7 Blackbody Radiation in a Wave Length Interval
10.8 Real and Gray Bodies
10.9 Kirchhoff’s Law
10.10 Intensity of Radiation and Lambert’s Cosine Law
10.11 Summary
Review Questions
Problems
References
11 Exchange of Thermal Radiation Between Surfaces Separated by Transparent Medium
11.1 Introduction
11.2 Radiation Heat Exchange Between Two Black Surfaces and the Shape Factor
11.3 Evaluation of the Shape Factor
11.3.1 Salient Features of the Radiation Shape Factor
11.4 Reciprocity Relation
11.5 Radiation Exchange Between Infinite Parallel Planes
11.6 Radiation Exchange Between Infinite Long Concentric Cylinders
11.7 Radiation from a Gray Cavity
11.8 Small Gray Bodies
11.9 Electric Network Method for Solving Radiation Problems
11.9.1 Electric Network for a System Consisting of Two Gray Surfaces
11.9.2 System Consisting of Two Black Surfaces
11.9.3 Closed System of NBlack Surfaces
11.9.4 Systems Consisting of Two Black Surfaces Connected by a Single Refractory Surface
11.9.5 System Consisting of Two Gray Surfaces Connected by a Single Refractory Surface
11.9.6 System Consisting of Four Gray Surfaces Which See Each Other and Nothing Else
11.10 Radiation Shields
11.11 Radiation from a Gray Cavity (Alternative Method)
11.12 Newton’s Law of Cooling and Overall Heat Transfer Coefficient
11.12.1 Determination of Specific Heat Using Newton’s Law of Cooling
11.13 Radiation Heat Transfer Coefficient
11.14 Summary
Review Questions
Problems
References
12 Heat Transfer in Absorbing and Emitting Media (Gaseous Radiation)
12.1 Introduction
12.2 Specific Features of Gaseous Radiation
12.2.1 Selective Emitters
12.2.2 Beer’s Law
12.2.3 Transmissivity, Emissivity and Absorptivity
12.2.4 Total Emissive Power
12.3 Heat Exchange
12.3.1 Radiation Emitted by a Gas
12.3.2 Radiation Heat from Surface (Wall)
12.3.3 Net Rate of Heat Transfer
12.3.4 Mixture of CO2 and H2O Vapour
12.3.5 Gray Enclosure
12.4 Gray Gas Surrounded by Diffuse Gray Surfaces at Different Temperatures
12.5 Flames
12.5.1 Luminous Flames
12.5.2 Nonluminous Flames
12.6 Summary
Review Questions
Problems
References
13 Heat Transfer in Condensing Vapours and Boiling Liquids
13.1 Part A: Heat Transfer in Condensing Vapours
13.1.1 Introduction
13.1.1.1 Different Types of Condensation
Dropwise Condensation
Filmwise Condensation
13.1.2 Nusselt’s Film Condensation Theory
13.1.2.1 Laminar Film Condensation on a Vertical Surface
Comparison of Horizontal and Vertical Orientation of Tubes
13.1.2.2 Turbulent Film Flow
13.1.3 Factors Affecting Film Condensation
13.2 Part B: Heat Transfer in Boiling Liquids
13.2.1 Introduction
13.2.2 Boiling Heat Transfer
13.2.2.1 Pool Boiling
13.2.2.2 ForcedFlow Boiling
Flow Pattern in a Vertical Heated Tube with Upward Flow
Flow Pattern in Horizontal Evaporator Tube
13.3 Relations for Boiling Heat Transfer in Pool Boiling
13.3.1 Nucleate Boiling
13.3.1.1 The Peak Heat Flux
13.3.2 Simplified Relations for Boiling Heat Transfer with Water
Review Questions
Problems
References
14 Heat Exchangers
14.1 Part A: Heat Exchangers Fundamentals
14.1.1 Introduction
14.1.2 Heat Transfer Equation for Double Pipe (Concentric Tube) Heat Exchanger
14.1.3 Log Mean Temperature Difference (LMTD)
14.1.3.1 Parallel Flow Arrangement
14.1.3.2 Counterflow Arrangement
14.1.4 LMTD for Other Flow Arrangements
14.1.5 EffectivenessNTU Method
14.1.5.1 EffectivenessNTU Method for Counterflow Heat Exchanger
14.1.5.2 EffectivenessNTU Method for Parallel Flow Heat Exchanger
14.1.6 EffectivenessNTU Relations for Other Flow Arrangements
14.2 Part B: Design of Heat Exchangers
14.2.1 Introduction
14.2.2 Double Pipe Exchangers
14.2.3 Clean and Design Overall Heat Transfer Coefficients
14.3 Summary
Review Questions
Problems
References
15 Mass Transfer
15.1 Introduction
15.2 Fick’s Law of Diffusion
15.2.1 Fick’s Law for Gases in Terms of Partial Pressures
15.2.2 Fick’s Law on Mass Basis and Mole Basis
15.3 Diffusion Coefficient
15.4 Diffusion of Vapour Through a Stationary Gas: Stefan Law
15.5 Convective Mass Transfer
15.5.1 Convective Mass Transfer Equation in Terms of Partial Pressure Difference
15.6 Dimensional Analysis Applied to Convective Mass Transfer
15.6.1 Forced
15.6.2 Free
15.7 Mass Transfer Correlations
15.8 Reynolds and Colburn (or ChiltonColburn) Analogies
15.9 Summary
Review Questions
Problems
16 Special Topic: Performance of Solar Air Heater
16.1 Introduction
16.2 Mathematical Model for Thermohydraulic Performance Prediction (Karwa et al. 2007; Karwa and Chauhan 2010)
16.2.1 Top Loss
16.2.2 Wind Heat Transfer Coefficient
16.2.3 Sky Temperature
16.2.4 Convective Heat Transfer Coefficient Between the Absorber Plate and Glass Cover
16.2.5 Back and Edge Losses
16.2.6 Heat Transfer and Friction Factor Correlations
16.3 Enhanced Performance Solar Air Heaters
16.3.1 Introduction
16.3.2 Artificial Roughness for Heat Transfer Enhancement
16.3.2.1 Effect of Rib Shape and Pitch
16.3.2.2 Effect of Rib Arrangement
16.4 Heat Transfer and Friction Factor Correlations for Roughened Rectangular Ducts
16.5 Summary
References
Appendix_1
Appendix_2
Bibliography for Further Reading
Index
Rajendra Karwa
Heat and Mass Transfer Second Edition
Heat and Mass Transfer
Rajendra Karwa
Heat and Mass Transfer Second Edition
123
Rajendra Karwa Jodhpur Institute of Engineering and Technology Jodhpur, Rajasthan, India
ISBN 9789811539879 ISBN 9789811539886 https://doi.org/10.1007/9789811539886
(eBook)
1st edition: © Springer Science+Business Media Singapore 2017 2nd edition: © Springer Nature Singapore Pte Ltd. 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, speciﬁcally the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microﬁlms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a speciﬁc statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional afﬁliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #2101/04 Gateway East, Singapore 189721, Singapore
Preface
This textbook of heat and mass transfer has been written to meet the need that exists in the opinion of the author among the undergraduate students of Mechanical, Automobile and Production Engineering, graduate students of Thermal Engineering, and students appearing in various competitive examinations. The heat transfer part of the book will also be useful for students of Chemical Engineering. A classical treatment of the fundamentals of heat transfer has been presented in this book. The basic approach of separate discussions of conduction, convection and radiation has been used. The book contains 16 chapters and 3 appendices. Each chapter contains sufﬁcient number of solved problems. Stepwise answers are given to all unsolved problems. In Chap. 1, an overview of different modes of heat transfer has been presented. Chapters 2–4 present the conventional treatment of onedimensional heat conduction through plane wall, cylindrical and spherical systems, ﬁns, and simple systems with volumetric heat generation. In Chap. 5, analytical treatment of some cases of twodimensional steadystate heat conduction has been presented followed by a discussion of ﬁnitedifference numerical methods which are often used in practice for solving complex problems. Chapter 6 is devoted to the transient heat conduction where lumped heat capacity analysis and solution of problems based on Heisler charts have been given. Numerical method of solving transient conduction problems has been presented with a number of illustrative examples. Analytical solutions of some simple convection heat transfer problems, especially the convection with laminar flow, have been presented in Chap. 7. Empirical relations for forced convection and natural or free convection heat transfer have been presented in Chaps. 8 and 9, respectively. Chapters 10 and 11 deal with the fundamentals of radiation heat transfer and the exchange of thermal radiation between surfaces separated by transparent medium, respectively. The method of radiation network has been used extensively in the analysis of v
vi
Preface
radiation problems. Gaseous radiation problems have been dealt with using the conventional Hottel charts in Chap. 12. Chapter 13 has been divided into two parts. In the ﬁrst part, the basic modes of condensation have been presented followed by the presentation of the analytical solution due to Nusselt for laminar ﬁlm condensation on a vertical surface. The second part discusses the phenomenon of pool boiling followed by discussion on forced boiling in vertical and horizontal pipes. The conventional thermal analysis of heat exchangers (the logmeantemperature difference and effectiveness approaches) is presented in Chap. 14 followed by introduction to design methodology of heat exchangers considering the design of doublepipe heat exchanger. Chapter 15 presents a brief introduction to mass diffusion in a quiescent medium and convective mass transfer. Analogies between heat, mass and momentum transfer have been presented. In Chap. 16, thermal analysis and discussion of conventional and enhanced performance solar air heaters are presented along with the mathematical model of the solar air heater. The ﬁrst edition of the book received an excellent response. In this edition, some minor modiﬁcations and corrections have been made, and at the end of each chapter, summary of the chapter has been incorporated. Many new solved and unsolved problems have been added to provide problems of varying complexity. The students are advised to refer to the reference books, handbooks and journals, some of which are also listed at the end of book, for details beyond the coverage of this textbook and also for the new developments in the ﬁeld of heat transfer. Computers have made possible the numerical solution of quite complex problems. Readers are advised to refer to advanced references for the computeraided solution of heat transfer problems. The author sincerely expresses deep sense of gratitude and indebtedness to the authors and publishers of various advanced books, handbooks, journals and other references which have been consulted and whose material has been used in the preparation of this book. In spite of the care taken in preparing the manuscript of this book and reading the proofs, there is always a scope for improvement and some errors might have crept in. I will be grateful to the readers if they can suggest ways to improve the contents and bring the errors to my attention, if any, noticed by them. Jodhpur, India
Dr. Rajendra Karwa
Contents
1
Introduction . . . . . . . . . . . . . . . . . . . . 1.1 Introduction . . . . . . . . . . . . . . . 1.2 Heat Transfer by Conduction . . 1.3 Heat Transfer by Convection . . 1.4 Heat Transfer by Radiation . . . 1.5 Simultaneous Heat Transfer . . . 1.6 Summary . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . .
. . . . . . . .
1 1 1 2 4 5 5 6
2
OneDimensional SteadyState Heat Conduction . . . . 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Temperature Field and Temperature Gradient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Thermal Conductivity. . . . . . . . . . . . . . . . . . . . . 2.3.1 Thermal Conductivity of Solids . . . . . 2.3.2 Thermal Conductivity of Metals and Alloys . . . . . . . . . . . . . . . . . . . . . 2.3.3 Thermal Conductivity of Construction and HeatInsulating Materials . . . . . . . . . . . . . . . . . . . . . . . 2.3.4 Thermal Conductivity of Gases . . . . . . 2.3.5 Thermal Conductivity of Liquids . . . . 2.4 General Heat Conduction Equations . . . . . . . . . . 2.4.1 General Heat Conduction Equation in Cartesian Coordinates . . . . . . . . . . . 2.4.2 General Heat Conduction Equation in Cylindrical Coordinates . . . . . . . . . 2.4.3 General Heat Conduction Equation in Spherical Coordinates . . . . . . . . . . . 2.5 OneDimensional SteadyState Heat Conduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.1 Composite Plane Wall. . . . . . . . . . . . . 2.5.2 OneDimensional SteadyState Heat Conduction Through a Plane Homogeneous Wall Considering Film Coefﬁcients . . . . . . . . . . . . . . . . .
7 7
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
. . . . . . . .
7 9 9 10
10 11 11 12 12 17 22 25 27
28
vii
viii
Contents
2.5.3
OneDimensional SteadyState Conduction Heat Transfer Through a Plane Homogeneous Wall Considering Heat Transfer by Convection and Radiation from the Wall Surface . . . . . . . . . . . . . . . . . . . . 2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell . . . . . . 2.6.1 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell Considering Film Coefﬁcients . . . . . . . . . . . . . . . . . . . . . 2.6.2 Composite Cylindrical Wall . . . . . . . . 2.7 OneDimensional SteadyState Heat Conduction Through a Spherical Shell . . . . . . . . 2.7.1 OneDimensional SteadyState Heat Conduction Through a Spherical Shell Considering the Film Coefﬁcients . . . . . . . . . . . . . . . . . . . . . 2.7.2 Composite Spherical Shell . . . . . . . . . 2.8 Measurement of Thermal Conductivity . . . . . . . 2.8.1 Thermal Conductivity Measurement of Solids . . . . . . . . . . . . . . . . . . . . . . . 2.8.2 Thermal Conductivity Measurements of Liquids and Gases . . . . . . . . . . . . . . . . . . . . . . 2.9 Effect of Variable Thermal Conductivity . . . . . . 2.9.1 Plane Wall . . . . . . . . . . . . . . . . . . . . . 2.9.2 Cylindrical Shell . . . . . . . . . . . . . . . . . 2.9.3 Spherical Shell . . . . . . . . . . . . . . . . . . 2.10 Critical Thickness of Insulation . . . . . . . . . . . . . 2.10.1 Critical Thickness of Insulation for Cylinders . . . . . . . . . . . . . . . . . . . . 2.10.2 Critical Thickness of Insulation for Spherical Vessel . . . . . . . . . . . . . . 2.11 Thermal Contact Resistance . . . . . . . . . . . . . . . . 2.12 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Extended Surfaces (Fins) . . 3.1 Introduction . . . . . . . 3.2 Heat Transfer from a CrossSection . . . . . .
..................... .....................
30 43
47 50 58
61 62 66 66
69 70 71 73 74 89 89 91 101 105 106 108 112 113 113
Fin of Uniform .....................
114
Contents
ix
3.2.1 3.2.2
4
5
A Very Long Fin . . . . . . . . . . . . . . . . Negligible Heat Transfer from the Fin End as Compared to the Heat Transferred from the Fin Surface (Ac PL) . . . . . . . . . . . . . . . . . . . . . 3.2.3 Short Fins (Fin with Heat Loss from the Fin End) . . . . . . . . . . . . . . . . 3.3 Hollow Fins . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Composite Fins . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 Effectiveness and Efﬁciency of Fins . . . . . . . . . . 3.6 Heat Transfer from a Finned Wall . . . . . . . . . . . 3.7 Intensiﬁcation of Heat Transfer by Finning . . . . 3.8 Error in Temperature Measurement with Thermometer Well . . . . . . . . . . . . . . . . . . . . . . . 3.9 When Fins Are to Be Used? . . . . . . . . . . . . . . . 3.10 Heat Transfer from a Bar Connected to Two Heat Sources at Different Temperatures . . . . . . . 3.11 Generalized Equation of Fin . . . . . . . . . . . . . . . 3.12 Fin of Minimum Weight (Isachenko et al. 1977) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.13 Straight Fin of Triangular Section . . . . . . . . . . . 3.14 Straight Fin of Trapezoidal Section . . . . . . . . . . 3.15 Annular Fin . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.16 Fin Efﬁciency Plots . . . . . . . . . . . . . . . . . . . . . . 3.17 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
116
Conduction with Heat Generation . . . . . . . . . . . . . . . . 4.1 Plane Wall with Uniform Heat Generation . . . . . 4.1.1 Case (A) Surfaces at Different Temperatures . . . . . . . . . . . . . . . . . . . 4.1.2 Case (B) Surfaces at the Same Temperature . . . . . . . . . . . . . . . . . . . . 4.2 Cylinder with Uniform Heat Generation . . . . . . 4.3 Solid Sphere with Uniform Heat Generation . . . 4.4 Heat Transfer Through Piston Crown. . . . . . . . . 4.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
195 195
118 121 124 124 135 139 139 149 151 158 172 173 175 179 180 183 188 190 192 194
196 202 206 219 222 248 249 250
SteadyState TwoDimensional Heat Conduction . . . . 253 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 5.2 Analytical Solution of TwoDimensional Heat Conduction Problems . . . . . . . . . . . . . . . . . . . . . 254
x
Contents
5.3
Conduction Through a Flat Semiinﬁnite Homogeneous Plate . . . . . . . . . . . . . . . . . . 5.4 Mean Value Theorem. . . . . . . . . . . . . . . . . 5.5 Graphical Analysis of Two Dimensional SteadyState Conduction: Thermal Flux Plotting and Shape Factor . . . . . . . . . . . . . 5.6 Experimental Investigation of Conduction Process by Method of Analogy: ElectroThermal Analogy . . . . . . . . . . . . . . 5.7 Numerical Solution Methods . . . . . . . . . . . 5.7.1 FiniteDifference Method . . . . . . 5.7.2 Solution of Nodal Equations . . . . 5.8 TwoDimensional SteadyState Heat Conduction with Heat Generation . . . . . . . 5.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6
.... ....
258 263
....
269
. . . .
. . . .
. . . .
. . . .
283 285 285 298
. . . .
. . . .
. . . .
. . . .
319 323 324 324
Unsteady or Transient Heat Conduction . . . . . . . . . . 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Lumped Heat Capacity Analysis . . . . . . . . . . . . 6.2.1 Instantaneous and Total Heat Flow . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2.2 Applicability of the Lumped Heat Capacity Analysis . . . . . . . . . . . . . . . . 6.3 Lumped Capacitance, Varying Fluid Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 MultipleLumped Capacity Systems . . . . . . . . . . 6.5 Transient Heat Flow in Semiinﬁnite Solids . . . . 6.6 Transient Heat Conduction in Inﬁnite Plate . . . . 6.7 Heisler and Grober Charts . . . . . . . . . . . . . . . . . 6.8 Two and ThreeDimensional Transient Heat Conduction Systems . . . . . . . . . . . . . . . . . . . . . . 6.8.1 TwoDimensional Systems . . . . . . . . . 6.8.2 ThreeDimensional Systems . . . . . . . . 6.9 Numerical Method of Solving Transient Conduction Problems . . . . . . . . . . . . . . . . . . . . . 6.9.1 The Explicit and Implicit Formulations . . . . . . . . . . . . . . . . . . . . 6.10 The Schmidt Graphical Method for OneDimensional Problems . . . . . . . . . . . . . 6.11 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
329 329 329 334 334 347 348 353 360 363 379 379 380 384 389 399 407 407 408 412
Contents
xi
7
Convective Heat Transfer . . . . . . . . . . . . . . . . . . . . . . 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.1.1 Natural Convective Heat Transfer . . . . 7.1.2 Forced Convection Heat Transfer . . . . 7.2 Flow of Fluid Past a Flat Plate. . . . . . . . . . . . . . 7.3 Flow in Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.1 Laminar Flow Through a Tube . . . . . . 7.3.2 Turbulent Flow Through a Tube . . . . . 7.4 Equation of Continuity . . . . . . . . . . . . . . . . . . . . 7.4.1 The Displacement and Momentum Thickness . . . . . . . . . . . . . . . . . . . . . . 7.4.2 The Enthalpy and Conduction Thickness . . . . . . . . . . . . . . . . . . . . . . 7.5 Momentum Equation of Laminar Boundary Layer Over a Flat Plate . . . . . . . . . . . . . . . . . . . 7.5.1 Solution of Momentum Equation (Blasius Solution) . . . . . . . . . . . . . . . . 7.6 Integral Momentum Equation of Laminar Boundary Layer Over a Flat Plate: von Karman Solution . . . . . . . . . . . . . . . . . . . . . . . . 7.7 Energy Equation of Laminar Boundary Layer Over a Flat Plate . . . . . . . . . . . . . . . . . . . . . . . . 7.7.1 Pohlhausen’s Solution . . . . . . . . . . . . . 7.7.2 von Karman Integral Technique (Integral Analysis of Energy Equation for the Laminar Boundary Layer) . . . . . . . . . . . . . . . . . . . . . . . . . 7.8 Turbulent Boundary Layer Over a Flat Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.9 Laminar Flow in Tubes . . . . . . . . . . . . . . . . . . . 7.10 Turbulent Flow in Tubes . . . . . . . . . . . . . . . . . . 7.11 Momentum and Heat Exchange in Turbulent Flow (Eddy Viscosity and Eddy Thermal Diffusivity) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.12 Reynolds Analogy for Flow Past a Flat Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.12.1 Reynolds–Colburn Analogy . . . . . . . . 7.12.2 Application of Colburn Analogy to Turbulent Heat Transfer from a Flat Plate . . . . . . . . . . . . . . . . . 7.13 Prandtl–Taylor Modiﬁcation of Reynolds Analogy for Turbulent Flow Over Flat Plates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.13.1 von Karman Analogy for Flat Plates . . . . . . . . . . . . . . . . . . . . . . . . . 7.14 Reynolds Analogy for Turbulent Flow in Tubes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
413 413 413 413 415 416 418 419 420 421 425 426 429
439 459 463
468 484 491 507
509 514 517
519
526 529 530
xii
Contents
7.14.1
Prandtl–Taylor Modiﬁcation of Reynolds Analogy for Turbulent Flow in Tubes . . . . . . . . . . . . . . . 7.14.2 Friction Drag: Flow Over a Flat Plate Parallel to the Flow . . . . . . 7.15 Natural or Free Convection . . . . . . . . . . . . 7.16 Integral Momentum and Energy Equation of Free Convection on a Vertical Plate . . . 7.17 Liquid Metal Heat Transfer for Laminar Flow Over a Flat Plate . . . . . . . . . . . . . . . . 7.18 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
....
533
.... ....
535 539
....
540
. . . . .
552 554 558 560 563
. . . . .
. . . . .
. . . . .
Empirical Relations for Forced Convection Heat Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Dimensional Analysis. . . . . . . . . . . . . . . . . . . . . 8.3 Dimensional Analysis Applied to Forced Convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.1 Rayleigh’s Method . . . . . . . . . . . . . . . 8.3.2 Buckingham’s PiMethod . . . . . . . . . . 8.3.3 Physical Signiﬁcance of Dimensionless Numbers . . . . . . . . . . . 8.4 Experimental Determination of Forced Convection Heat Transfer Coefﬁcient . . . . . . . . . 8.4.1 Uniform Temperature Condition . . . . . 8.4.2 Uniform Heat Flux Condition . . . . . . . 8.5 Friction Factor and Heat Transfer Coefﬁcient Correlations for Circular Ducts . . . . . . . . . . . . . 8.5.1 Laminar Flow in Circular Tubes . . . . . 8.5.2 Turbulent Flow in Circular Tubes . . . . 8.6 Effects of Temperature Varying Properties . . . . . 8.7 Heat Transfer and Friction in Concentric Circular Tube Annuli and Parallel Plate Duct . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.7.1 Laminar Flow . . . . . . . . . . . . . . . . . . . 8.7.2 Turbulent Flow . . . . . . . . . . . . . . . . . . 8.8 Heat Transfer and Friction in Rectangular Duct . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.8.1 Laminar Flow . . . . . . . . . . . . . . . . . . . 8.8.2 Turbulent Flow . . . . . . . . . . . . . . . . . . 8.9 Correlations for External Forced Flow Over a Flat Plate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.9.1 Laminar Flow . . . . . . . . . . . . . . . . . . . 8.9.2 Turbulent Flow . . . . . . . . . . . . . . . . . .
565 565 565 566 567 568 576 578 578 578 580 580 589 598
601 601 608 611 611 612 614 614 615
Contents
xiii
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies . . . . . . . 8.10.1 Cylinder in Cross Flow . . . . . . . . 8.10.2 Flow Around a Sphere . . . . . . . . 8.10.3 Flow Across Tube Banks . . . . . . 8.11 Heat Transfer in Liquid Metals . . . . . . . . . 8.12 Inﬂuence of Duct Wall Roughness in Turbulent Flow . . . . . . . . . . . . . . . . . . . . . 8.13 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9
. . . . .
. . . . .
. . . . .
. . . . .
620 621 624 626 643
. . . . .
. . . . .
. . . . .
. . . . .
645 672 677 678 681
Empirical Relations for Natural or Free Convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Buoyancy Force in Natural Convection . . . . . . . 9.3 Dimensional Analysis Applied to Natural Convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3.1 Rayleigh’s Method . . . . . . . . . . . . . . . 9.3.2 Buckingham’s Pi Method . . . . . . . . . . 9.3.3 Physical Interpretation of Grashof Number . . . . . . . . . . . . . . . . . . . . . . . . 9.4 Experimental Determination of Natural Convection Heat Transfer Coefﬁcient . . . . . . . . . 9.5 Empirical Relations for Free or Natural Convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.5.1 Vertical Plate and Cylinders . . . . . . . . 9.5.2 Inclined Plate . . . . . . . . . . . . . . . . . . . 9.5.3 Horizontal Plate . . . . . . . . . . . . . . . . . 9.5.4 Horizontal Cylinder of Diameter D and Length L >> d . . . . . . . . . . . . . . . 9.5.5 Sphere of Diameter d . . . . . . . . . . . . . 9.6 Free Convection in Parallel Plate Channels . . . . 9.6.1 Vertical Channels . . . . . . . . . . . . . . . . 9.6.2 Inclined Channels . . . . . . . . . . . . . . . . 9.7 Empirical Correlations for Enclosed Spaces . . . . 9.8 Combined Free and Forced Convection (Kays and Crawford 1980; Gebhart 1961; Holman 1992; Cengel 2007) . . . . . . . . . . . . . . . 9.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
683 683 683 684 685 686 690 691 693 693 695 695 696 697 710 711 711 711
719 726 728 728 730
10 Laws of Thermal Radiation . . . . . . . . . . . . . . . . . . . . . 733 10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 733 10.2 Reﬂection, Absorption and Transmission of Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 734
xiv
Contents
10.3 10.4
Emissivity and a Perfect Blackbody . . . . . . . . . . Planck’s Spectral Distribution of Emissive Power . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5 Wein’s Displacement Law . . . . . . . . . . . . . . . . . 10.6 Total Emissive Power: Stefan–Boltzmann Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.7 Blackbody Radiation in a Wave Length Interval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.8 Real and Gray Bodies . . . . . . . . . . . . . . . . . . . . 10.9 Kirchhoff’s Law . . . . . . . . . . . . . . . . . . . . . . . . . 10.10 Intensity of Radiation and Lambert’s Cosine Law . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.11 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11 Exchange of Thermal Radiation Between Surfaces Separated by Transparent Medium . . . . . . . . . . . . . . 11.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Radiation Heat Exchange Between Two Black Surfaces and the Shape Factor . . . . . . . . . . . . . . 11.3 Evaluation of the Shape Factor . . . . . . . . . . . . . 11.3.1 Salient Features of the Radiation Shape Factor . . . . . . . . . . . . . . . . . . . . 11.4 Reciprocity Relation . . . . . . . . . . . . . . . . . . . . . . 11.5 Radiation Exchange Between Inﬁnite Parallel Planes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.6 Radiation Exchange Between Inﬁnite Long Concentric Cylinders . . . . . . . . . . . . . . . . . . . . . 11.7 Radiation from a Gray Cavity . . . . . . . . . . . . . . 11.8 Small Gray Bodies . . . . . . . . . . . . . . . . . . . . . . . 11.9 Electric Network Method for Solving Radiation Problems . . . . . . . . . . . . . . . . . . . . . . 11.9.1 Electric Network for a System Consisting of Two Gray Surfaces . . . . 11.9.2 System Consisting of Two Black Surfaces . . . . . . . . . . . . . . . . . . . . . . . 11.9.3 Closed System of NBlack Surfaces . . . . . . . . . . . . . . . . . . . . . . . 11.9.4 Systems Consisting of Two Black Surfaces Connected by a Single Refractory Surface . . . . . . . . . . . . . . . 11.9.5 System Consisting of Two Gray Surfaces Connected by a Single Refractory Surface . . . . . . . . . . . . . . .
736 738 740 741 743 748 750 755 762 763 764 765 767 767 767 770 772 779 806 807 809 813 819 821 822 823
823
828
Contents
xv
11.9.6
System Consisting of Four Gray Surfaces Which See Each Other and Nothing Else . . . . . . . . . . . . . . . . 11.10 Radiation Shields . . . . . . . . . . . . . . . . . . . . . . . . 11.11 Radiation from a Gray Cavity (Alternative Method) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.12 Newton’s Law of Cooling and Overall Heat Transfer Coefﬁcient . . . . . . . . . . . . . . . . . . . . . . 11.12.1 Determination of Speciﬁc Heat Using Newton’s Law of Cooling . . . . 11.13 Radiation Heat Transfer Coefﬁcient . . . . . . . . . . 11.14 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Heat Transfer in Absorbing and Emitting Media (Gaseous Radiation) . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2 Speciﬁc Features of Gaseous Radiation . . . . . . . 12.2.1 Selective Emitters . . . . . . . . . . . . . . . . 12.2.2 Beer’s Law . . . . . . . . . . . . . . . . . . . . . 12.2.3 Transmissivity, Emissivity and Absorptivity . . . . . . . . . . . . . . . . . . . . 12.2.4 Total Emissive Power . . . . . . . . . . . . . 12.3 Heat Exchange . . . . . . . . . . . . . . . . . . . . . . . . . . 12.3.1 Radiation Emitted by a Gas . . . . . . . . 12.3.2 Radiation Heat from Surface (Wall) . . . . . . . . . . . . . . . . . . . . . . . . . 12.3.3 Net Rate of Heat Transfer . . . . . . . . . . 12.3.4 Mixture of CO2 and H2O Vapour . . . . 12.3.5 Gray Enclosure . . . . . . . . . . . . . . . . . . 12.4 Gray Gas Surrounded by Diffuse Gray Surfaces at Different Temperatures . . . . . . . . . . . 12.5 Flames . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.5.1 Luminous Flames . . . . . . . . . . . . . . . . 12.5.2 Nonluminous Flames . . . . . . . . . . . . . 12.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Heat Transfer in Condensing Vapours and Boiling Liquids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.1 Part A: Heat Transfer in Condensing Vapours . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.1.1 Introduction. . . . . . . . . . . . . . . . . . . . . 13.1.2 Nusselt’s Film Condensation Theory . . . . . . . . . . . . . . . . . . . . . . . .
829 831 855 856 858 859 893 894 895 902 903 903 903 903 904 905 906 906 906 908 910 911 911 921 925 925 926 927 928 928 929 931 931 931 932
xvi
Contents
13.1.3
Factors Affecting Film Condensation . . . . . . . . . . . . . . . . . . . 13.2 Part B: Heat Transfer in Boiling Liquids . . . . . . 13.2.1 Introduction. . . . . . . . . . . . . . . . . . . . . 13.2.2 Boiling Heat Transfer . . . . . . . . . . . . . 13.3 Relations for Boiling Heat Transfer in Pool Boiling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.3.1 Nucleate Boiling . . . . . . . . . . . . . . . . . 13.3.2 Simpliﬁed Relations for Boiling Heat Transfer with Water . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Heat Exchangers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.1 Part A: Heat Exchangers Fundamentals . . . . . . . 14.1.1 Introduction. . . . . . . . . . . . . . . . . . . . . 14.1.2 Heat Transfer Equation for Double Pipe (Concentric Tube) Heat Exchanger . . . . . . . . . . . . . . . . . . . . . . 14.1.3 Log Mean Temperature Difference (LMTD) . . . . . . . . . . . . . . . . . . . . . . . . 14.1.4 LMTD for Other Flow Arrangements . . . . . . . . . . . . . . . . . . . 14.1.5 EffectivenessNTU Method . . . . . . . . . 14.1.6 EffectivenessNTU Relations for Other Flow Arrangements . . . . . . . . . . 14.2 Part B: Design of Heat Exchangers . . . . . . . . . . 14.2.1 Introduction. . . . . . . . . . . . . . . . . . . . . 14.2.2 Double Pipe Exchangers . . . . . . . . . . . 14.2.3 Clean and Design Overall Heat Transfer Coefﬁcients . . . . . . . . . . . . . . 14.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Mass Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.2 Fick’s Law of Diffusion . . . . . . . . . . . . . . . . . . . 15.2.1 Fick’s Law for Gases in Terms of Partial Pressures . . . . . . . . . . . . . . . 15.2.2 Fick’s Law on Mass Basis and Mole Basis . . . . . . . . . . . . . . . . . . . . . . . . . . 15.3 Diffusion Coefﬁcient . . . . . . . . . . . . . . . . . . . . . 15.4 Diffusion of Vapour Through a Stationary Gas: Stefan Law. . . . . . . . . . . . . . . . . . . . . . . . . 15.5 Convective Mass Transfer . . . . . . . . . . . . . . . . .
951 952 952 953 959 959 959 964 964 965 967 967 967
969 970 982 988 995 1011 1011 1011 1012 1033 1036 1037 1039 1041 1041 1042 1043 1045 1046 1047 1050
Contents
xvii
15.5.1
Convective Mass Transfer Equation in Terms of Partial Pressure Difference . . . . . . . . . . . . . . . . . . . . . . 15.6 Dimensional Analysis Applied to Convective Mass Transfer . . . . . . . . . . . . . . . . . . . . . . . . . . 15.6.1 Forced . . . . . . . . . . . . . . . . . . . . . . . . . 15.6.2 Free. . . . . . . . . . . . . . . . . . . . . . . . . . . 15.7 Mass Transfer Correlations . . . . . . . . . . . . . . . . 15.8 Reynolds and Colburn (or ChiltonColburn) Analogies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15.9 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Review Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Special Topic: Performance of Solar Air Heater . . . . 16.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16.2 Mathematical Model for Thermohydraulic Performance Prediction (Karwa et al. 2007; Karwa and Chauhan 2010) . . . . . . . . . . . . . . . . . 16.2.1 Top Loss . . . . . . . . . . . . . . . . . . . . . . . 16.2.2 Wind Heat Transfer Coefﬁcient. . . . . . 16.2.3 Sky Temperature . . . . . . . . . . . . . . . . . 16.2.4 Convective Heat Transfer Coefﬁcient Between the Absorber Plate and Glass Cover . . . . . . . . . . . . . 16.2.5 Back and Edge Losses . . . . . . . . . . . . 16.2.6 Heat Transfer and Friction Factor Correlations. . . . . . . . . . . . . . . . . . . . . 16.3 Enhanced Performance Solar Air Heaters . . . . . . 16.3.1 Introduction. . . . . . . . . . . . . . . . . . . . . 16.3.2 Artiﬁcial Roughness for Heat Transfer Enhancement . . . . . . . . . . . . . 16.4 Heat Transfer and Friction Factor Correlations for Roughened Rectangular Ducts . . . . . . . . . . . 16.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1050 1051 1051 1053 1055 1056 1064 1065 1065 1067 1068
1070 1071 1072 1073
1074 1074 1076 1081 1081 1082 1085 1089 1090
Appendix A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1093 Appendix B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1103 Bibliography for Further Reading . . . . . . . . . . . . . . . . . . . 1137 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1139
About the Author
Dr. Rajendra Karwa pursued his Bachelor’s and Master’s degree in Mechanical Engineering from the University of Jodhpur, Jodhpur, India. He obtained his Ph.D. in 1998 from IIT Roorkee, India. He retired from M.B.M. Engineering College, Faculty of Engineering & Architecture, Jai Narain Vyas University, Jodhpur in 2012 as Professor and Head, Department of Mechanical Engineering after serving more than 30 years. He has served as visiting faculty in Addis Ababa University, Addis Ababa (Ethiopia) from 20032005. He is actively involved in design of industrial systems and products, and has successfully completed a large number of industrial projects. Currently he is working as Director (Academics) at Jodhpur Institute of Engineering & Technology, Jodhpur. He has supervised 15 research scholars leading to the award of Ph.D. and M.Tech. He has published about 70 research papers and has authored two books. His areas of interest are heat transfer, machine design, solar thermal applications, and exergy analysis.
xix
List of Symbols1
A A Ac b c c cp cv cfx CD C C C d, D d, D D D h, d h e e/Dh, e/D E E Ebk f fapp FD Fij g G G h
Frontal area, m2; Heat transfer surface area, m2; Crosssectional area, m2; Width, m; Speed of light, m/s; Speciﬁc heat of solids and liquids, J/(kg K); Speciﬁc heat at constant pressure, J/(kg K); Speciﬁc heat at constant volume, J/(kg K); Local skin friction coefﬁcient = swx/(1/2qU2∞); Drag coefﬁcient = (FD/ A)/(1/2qU2∞); Heat capacity = mc, J/K; Mass concentration, kg/m3; Molar concentration, kmol/m3; Characteristic dimension, m; Diameter, m; Diffusion coefﬁcienta, m2/s; Hydraulic diameter = 4A/P, m; Roughness height, m; Relative roughness height; Electric potential; Energy, J; Monochromatic hemispherical emissive power of black body, W/(m.m2); Fanning friction factor = sw/(1/2qU2); Apparent Fanning friction factor in the hydrodynamic entrance region; Drag force, N; View factor; Gravitational acceleration, m/s2; Mass velocity, kg/(s m2); Irradiation, W/m2; Heat transfer coefﬁcient, W/(m2 K);
1
(at most of the places, small t has been used when temperature is in °C) xxi
xxii
hfg hm hr hx hw H I Ik J k l L L Lc Lhy Lth m m M M n N NTU p pi P P q q” , qw Q qg r R R R Rk S S Ta Tb,Tfm Ti To Ts Tsky Tsat, Ts Tw
List of Symbols
Enthalpy of evaporation (latent heat), J/kg; Mass transfer coefﬁcient, m/s; Radiation heat transfer coefﬁcient, W/(m2 K), W/(m2 °C); Local heat transfer coefﬁcient at position x, W/(m2 K), W/ (m2 °C); Wind heat transfer coefﬁcient, W/(m2 K), W/(m2 °C); Height, m; Solar radiation intensity (insolation), W/m2; Monochromatic intensity of radiation; Radiosity, W/m2; Thermal conductivity, W/(m K), W/(m °C); Mean free path, m; Length, m; Fundamental dimension of length; Corrected ﬁn length, m; Hydrodynamic entrance length, m; Thermal entrance length, m; Mass flow rate, kg/s; A ﬁn parameter, 1/m; Molecular weight; Fundamental dimension of mass; Number of radiation shields; Molal diffusion rate; Number of transfer units; Pressure, Pa; Partial pressure, Pa; Power, W; Perimeter, m; Heat transfer rate, W; Heat flux, W/m2; Quantity of heat, J; Volumetric heat generation rate, W/m3; Radius (usually variable), m; Radius, m; Gas constant, J/(kg K); Temperature group (T1–T2)/(t2–t1); Thermal resistance to heat conduction, K/W; Temperature group (t2–t1)/(T1–t1); Conduction shape factor; Ambient temperature, °C, K; Bulk mean air temperature = (To + Ti)/2, °C; Inlet temperature, °C; Outlet temperature, °C; Surface temperature, °C, K; Sky temperature, K; Saturation temperature, °C; Wall temperature, °C, K;
List of Symbols
xxiii
T∞ u, v, w U U U∞ v V W W W/H x, y, z xi
Freestream temperature, °C; Velocity, m/s; Overall heat transfer coefﬁcient, W/(m2 K), W/(m2 °C); Velocity, m/s; Freestream velocity, m/s; Speciﬁc volume, m3/kg; Volume, m3; Weight, N; Width of the duct, m; Duct aspect ratio; Variable distances in space; Mole fraction;
Dimensionless Numbers Bi e+ Ec Fo g Gr Gz Le p/e Nu Pe Pr Prt R Ra Ra* Re Rex Recr Sc Sh Stm St Stx
Biot number = hL/ksolid; Roughness Reynolds number, Eq. (8.63); Eckert number = u2/cpDt; Fourier number = as/L2; Heat transfer function, Eq. (8.69); Grashof number = gbL3Dt/m2; Graetz number = Re Pr (D/L); Lewis numbera = Sc/Pr = a/D; Relative roughness pitch; Nusselt number = hL/kfluid; Peclet number = Re Pr; Prandtl number = lcp/k; Turbulent Prandtl number = eM/eH; Roughness function, Eq. (8.67); Rayleigh number = Gr Pr = gbH3Dt/am; Rayleigh number (based on heat flux) = gbq”H4/amk; Reynolds number = qUd/l = GL/l; Reynolds number based on longitudinal length = U∞x/m; Critical Reynolds number; Schmidt numbera = m/D =l/qD = Le Pr; Sherwood numbera = hmL/D; Mass transfer Stanton number = hm/U; Stanton number = h/(Gcp); Local Stanton number = hx/(Gcp);
xxiv
List of Symbols
Greek Symbols a, b, c, /, w a a b b b d d dmd dvd dp, Dp dt Dt, DT e e e eH eM / η ηf k k kmax l m p h h q q r r s s (sa) x x w
Angle (degree or rad); Thermal diffusivity = k/qc, m2/s; Absorptivity (radiation); Coefﬁcient of volumetric expansion, 1/K; Temperature coefﬁcient of thermal conductivity, 1/K; Collector slope (degree); Velocity boundary layer thickness, m; Thickness, m; Momentum displacement thickness, m; Velocity displacement thickness, m; Pressure drop in the duct, Pa; Thermal boundary layer thickness; Temperature difference, °C, K; Fin effectiveness; Heat exchanger effectiveness; Emissivity; Thermal eddy diffusivity, m2/s; Momentum eddy diffusivity or viscosity, m2/s; Relative humidity; Thermal efﬁciency; Efﬁciency of ﬁn; Darcy friction factor (= 4f) Wavelength, m; Wavelength at maximum value of Ebk; Dynamic viscosity, Pa s, N s/m2, kg/(m s); Kinematic viscosity = l/q, m2/s; Dimensionless group; Temperature excess, K; Time, s; Density of fluid, kg/m3; Reflectivity; Stefan–Boltzmann constant; Surface tension, N/m; Time, s; Shear stress between fluid layers, Pa; Transmittance–absorptance product; Solid angle, sr; Speciﬁc humidity; Stream function;
List of Symbols
xxv
Superscript and Subscript a b cr f f fd g hy i, 1 i l m m max min o, 2 o o s s th v w x ∞ ¯ (overbar)
Ambient Bulk, blackbody Critical state Fluid Film Fully developed Gas Hydrodynamic Inlet or initial Based on the inside surface of a pipe Laminar or liquid Mass transfer quantity Mean Maximum Minimum Outlet Based on the outside surface of a pipe Stagnation Smooth surface Surface Thermal Vapour Wall Based on variable length Freestream condition Mean or molar
Space Coordinates r, h, z r, h, / x, y, z a
Cylindrical, m, rad, m; Spherical, m, rad, rad; Cartesian, m, m, m.
Diffusion D is mass diffusivity, m2/s. Note: The symbol L in the dimensionless groups stands for a generic length and is deﬁned according to the particular geometry under consideration; it may be diameter, hydraulic diameter, plate length, etc.
1
Introduction
1.1
Introduction
Thermodynamically, heat is a form of energy that is transferred between two systems (or a system and its surroundings) by virtue of a temperature difference. Classical thermodynamics treats the processes as though only the end states exist, presuming that the system exists in a state of thermodynamic equilibrium. It provides no information about the rates of irreversible flows. The heat flow with temperature difference is one example of irreversible flow. In studies of processes that involve flow of heat or mass, the rate of flow is an important parameter. Heat can be transferred in three different ways: by conduction, convection and radiation. In most of the engineering applications, it is a combination of the two or three modes. Pure conduction is found only in solids, and convection is possible only in fluids. A detailed study of these modes will be given in this book. Here, a brief description of each mode is presented to familiarize the reader with them.
1.2
Heat Transfer by Conduction
Thermal conduction is a process by which heat is transmitted by the direct contact between particles of a body without any motion of the material as a whole. The phenomenon of conduction heat transfer can be experienced by a simple experiment. Heat one end of a metal rod. The other end of the rod will become hotter and hotter with the passage of time. Heat reaches from the heated end of the rod to the other end by conduction through the material of the rod. Conduction occurs in all media—solids, liquids, and gases when a temperature gradient exists. In opaque solids, it is the only mechanism by which the heat can flow. In the fluids, the molecules have freedom of motion and energy is also transferred by movement of the fluid. However, if the fluid is at rest, the heat is transferred by conduction. Heat transfer by conduction requires that the temperature distribution in a medium is nonuniform, that is, a temperature gradient exists in the body. Fourier noted in 1811 that the heat flow in a homogenous solid is directly proportional to the temperature gradient. Consider a plane wall of thickness dx, whose area perpendicular to the direction x is A as shown in Fig. 1.1. If the thickness of the wall is very small as compared to its height and width, it is a © Springer Nature Singapore Pte Ltd. 2020 R. Karwa, Heat and Mass Transfer, https://doi.org/10.1007/9789811539886_1
1
2
1
Introduction
dx t1 dt
t2
q x
Fig. 1.1 Conduction heat flow through a solid
case of onedimensional (in direction x only) heat flow. Let one of the faces of the wall is at a temperature t1 and the other at temperature t2. For the elemental thickness dx, the temperature difference is dt. Then, the conduction heat flux q/A (the heat flux is the heat flow rate per unit area of the surface), according to Fourier, is q dt / A dx
ð1:1Þ
When the constant of the proportionality is inserted in Eq. (1.1), we get q dt ¼ k A dx
ð1:2Þ
The constant of proportionality k is known as thermal conductivity of the material. It is a physical property of a substance and characterizes the ability of a material to conduct heat. The negative sign in the equation indicates that the heat flow is in the direction of falling temperature. From Eq. (1.2), we get k¼
q dx A dt
W=ðm KÞ
ð1:3Þ
Thus thermal conductivity determines the quantity of heat flowing per unit time through the unit area with a temperature drop of 1°C (K) per unit length.
1.3
Heat Transfer by Convection
The term is applied to transport of heat as a volume of liquid or gas moves from a region of one temperature to that of another temperature. Thus, the transport of heat is linked with the movement of the medium itself. The convection can be observed in liquids if we carry out a simple experiment. Consider a pot with water which is placed over a burner. The water at the bottom of the pot is heated and becomes less dense than before due to its thermal expansion. Thus, the water at the bottom, which is less dense than the cold water in the upper portion, rises upwards. It transfers its heat by mixing as it rises. The movement of the water, referred
1.3 Heat Transfer by Convection
3
to as the convection currents, can be observed by putting a few crystals of potassium permanganate in the bottom of the pot. Density differences and the gravitational force of the earth act to produce a force known as buoyancy force, which drives the flow. When the flow is due to the density differences only, it is called natural or free convection. Density differences may also be caused by the composition gradients. For example, water vapour rises mainly due to the lower density of water vapour present in the moist air. Convection is termed as forced if the fluid is forced to flow over a surface or in a duct by external means such as a fan, pump or blower, that is, the forced convection implies mechanically induced flow. Heat transfer processes involving change of phase of a fluid (boiling of liquids or condensation of vapours) are also considered to be convection because of the motion of fluid that is set up due to the rising vapour bubbles during boiling or the falling liquid droplets during the condensation. The heat transfer by convection is always accompanied by conduction. The combined process of heat transfer by convection and conduction is referred to as convective heat transfer. The heat transfer, between a solid surface and a fluid, is expressed by Newton’s law of cooling as q ¼ hADt
ð1:4Þ
where Δt is the temperature difference between the fluid and the surface, A is the area of the surface transferring heat and h is known as heat transfer coefﬁcient or ﬁlm coefﬁcient. In general, the value of the heat transfer coefﬁcient h depends on the fluidflow conditions, the thermophysical properties of the fluid and the type of flow passage. Order of magnitude of convective heat transfer coefﬁcients is given in Table A6, Appendix. Consider a heated plate at temperature tw as shown in Fig. 1.2, which is exposed to a fluid flowing parallel to the plate surface at velocity U∞ and is at a temperature t∞. The velocity of the fluid at the plate surface will be zero because of the viscous effect. Since the velocity of the fluid at the wall is zero, we can imagine that there exists a thin ﬁlm of the fluid quite close to the wall, which is practically stationary. In most of the cases, the temperature gradient is conﬁned to this thin layer where at a greater distance from the wall, only a small temperature difference exists because of the mixing of the fluid. The heat is, thus, transferred across the ﬁlm by conduction. This heat is then carried away by fluid motion. Assuming that the temperature variation in the fluid ﬁlm is linear, the flow of heat across the ﬁlm can be expressed by Fourier’s law as q ¼ kf A
ðtw t1 Þ df
ð1:5Þ
where df is the thickness of the fluid ﬁlm and kf is the thermal conductivity of the fluid.
Fluid film U∞, t∞
tw Plate
Fig. 1.2 Cooling of a hot plate by flowing fluid
4
1
Introduction
Comparison of Eq. (1.5) with Newton’s equation, which is widely used, gives h = kf/df. This equation relates the convection heat transfer to the conduction. It shows that the value of the heat transfer coefﬁcient is directly proportional to the thermal conductivity of the fluid. It is to be kept in mind that there are other factors also which govern the magnitude of the heat transfer coefﬁcient as mentioned earlier. From the order of magnitude of the heat transfer coefﬁcient in Table A6, it can be seen that the gases provide a lower value of heat transfer coefﬁcient in comparison to the liquids. This can be attributed to their lower thermal conductivity as compared to the liquids. In most of the heat exchangers, heat is transferred between hot and cold fluid streams across a solid wall. In such cases, it is convenient to combine the two ﬁlm coefﬁcients (i.e. of the hot and cold fluid streams) to give a single coefﬁcient known as overall heat transfer coefﬁcient U, which is deﬁned as U¼
q ADt
ð1:6Þ
Here Δt is the temperature difference between the two fluids.
1.4
Heat Transfer by Radiation
The thermal radiation is the process of heat propagation by means of electromagnetic waves produced by virtue of the temperature of the body. It depends on both the temperature and an optical property known as emissivity e of the body. In contrast to the conduction and convection heat transfers, radiation can take place through a perfect vacuum. Solids, liquids and gases may radiate energy. For example, water vapour and carbon dioxide are the principal sources of the gaseous radiation in furnaces. Boltzmann established that the rate at which a body gives out the heat by radiation is proportional to the fourth power of the absolute temperature of the body, that is, q / AT 4
ð1:7Þ
When the constant of the proportionality is inserted, q ¼ rAT 4
ð1:8Þ
where the constant of proportionality r is known as Stefan–Boltzmann constant, and A is the surface area of the body. A body may absorb, transmit or reflect radiant energy. A blackbody absorbs the entire radiation incident on it. Thus, it is a perfect absorber. Technically, the blackbody is a hypothetical body. It does not necessarily refer to the colour of the body, though bodies black in colour usually absorb most. A blackbody is also a perfect or ideal radiator, for which the emissivity e = 1. For real bodies, e is less than 1 and they do not emit as much energy as a blackbody. The net heat radiated between two black bodies 1 and 2 at temperatures T1 and T2 that see each other completely (i.e. they exchange heat by radiation between themselves only); the net energy exchange is proportional to the difference in T14 and T24 . Thus
1.4 Heat Transfer by Radiation
5 t1 Electrical heater
qcond
qconv ●t∞ qrad
k Insulation δ
t2 Plane wall
Fig. 1.3 Combination of modes of heat transfer
q12 ¼ rAðT14 T24 Þ
ð1:9Þ
The radiation exchange between bodies which are not black is quite complex and will be dealt with in detail later.
1.5
Simultaneous Heat Transfer
Heat can be transferred in three different modes: by conduction, convection and radiation. In most of the engineering applications, it is a combination of the two or three modes. Pure conduction is found only in opaque solids but in the case of semitransparent solids radiation heat transfer is also involved. Convection is possible only in fluids. In vacuum, heat transfer is possible only by radiation. In case of a solid surface exposed to a fluid (gas or liquid), convection heat transfer is always involved. If the gas to which the solid surface is exposed is transparent to radiation, heat transfer also takes place by radiation. It is to note that liquids and some gases (presented in Chap. 12) are absorbers of radiation. Figure 1.3 presents a case where all the three modes of heat transfer are present. Heat transfers by conduction through the wall from left face to the right face and then is rejected from the right face by convection to the surrounding air and by radiation to the surroundings. The energy balance would give qcond ¼ qconv þ qrad
1.6
ð1:10Þ
Summary
In this chapter, the basic modes of heat transfer have been presented, which will be discussed in greater detail in coming chapters to apply them to various problems of heat transfer. Newton’slaw of cooling presented for convection heat transfer looks quite simple but it is to be noted that the heat transfer coefﬁcient h is a function which depends on the fluidflow conditions, the thermophysical properties of the fluid and the geometric conﬁguration of the heated surface or duct, and for its estimate mathematical or experimental approach is used. A simple case of radiation exchange between two black bodies that see each other completely (i.e. they exchange heat by radiation between themselves only) has been considered in Eq. (1.9). In general, the radiation heat exchange between bodies, which are not black, involves complex interaction because of surface emissive properties and when a system of radiation heat exchange between a number of bodies is involved.
6
1
Introduction
Review Questions 1:1. 1:2. 1:3. 1:4. 1:5.
State and explain in brief the different modes of heat transfer. Write Fourier’s equation for onedimensional steadystate conduction heat transfer. Deﬁne thermal conductivity and heat transfer coefﬁcient. Discuss the mechanism of convective heat transfer. State Boltzmann’s law for the radiation heat transfer. What is the net heat exchange between two blackbodies, which exchange heat by radiation between themselves only?
2
OneDimensional SteadyState Heat Conduction
2.1
Introduction
We presented Fourier’s law of heat conduction in Chap. 1. In this chapter, we shall use the law to calculate the heat flow in systems where onedimensional heat flow occurs. A plane wall made of isotropic material, whose thickness is much smaller compared to its length and width, is the simplest case of such onedimensional conduction heat flow. The heat flow will also be onedimensional in cylindrical and spherical solid when the temperature gradient is only in the radial direction. Some two or threedimensional systems can also be approximated as onedimensional if the temperature variation, represented by the temperature gradient, in one direction is signiﬁcantly greater than the other directions. Before attempting the analysis of onedimensional heat conduction problems, we shall discuss some important terms relating to the conductors and establish the general heat conduction equations in the rectangular, cylindrical and spherical coordinates.
2.2
Temperature Field and Temperature Gradient
Heat flows by conduction when different points in a body are at different temperatures. In general, the temperature may vary both in space (x, y and z) and time (s). The analytical investigation of the heat conduction is basically a study of the spacetime variation of the temperature and determination of the equation: t ¼ tðx; y; z; sÞ
ð2:1Þ
which is the mathematical expression of the temperature ﬁeld. Equation (2.1) represents a set of temperatures at all points of the space at any given time. The temperature ﬁeld expressed by Eq. (2.1) is referred to as transient or nonsteady temperature ﬁeld. If the temperature in the space does not change with time, the temperature ﬁeld is a function of the space coordinates only and it is referred to as a steady state. Mathematically, it is expressed as t ¼ tðx; y; z; Þ
© Springer Nature Singapore Pte Ltd. 2020 R. Karwa, Heat and Mass Transfer, https://doi.org/10.1007/9789811539886_2
7
8
2
OneDimensional SteadyState Heat Conduction
and @t ¼0 @s
ð2:2Þ
The temperature ﬁelds represented by Eqs. (2.1) and (2.2) are threedimensional ﬁelds since they are function of three coordinates. A temperature ﬁeld, which is a function of two coordinates say x and y, is termed as twodimensional and is described by the following equation: t ¼ tðx; yÞ and @t ¼0 @z
ð2:3Þ
Similarly for a onedimensional case, the equation of temperature ﬁeld is t ¼ tðxÞ and @t @t ¼ ¼0 @y @z
ð2:4Þ
If different points of the body having the same temperature are joined, we obtain an isothermal surface. Intersection of such isothermal surfaces by a plane gives a family of isotherms on this plane as shown in Fig. 2.1. The ﬁgure shows isotherms which differ by temperature ±dt. The temperature gradient is a vector normal to the isothermal surface. Mathematically, it is the derivative in this direction: *
grad ðtÞ ¼ n
@t @n
ð2:5Þ
@t where n is the unit vector normal to the isothermal surface and @n is the temperature derivative along the normal. Projections of the vector grad t can be made on the coordinate *
y n Isotherms x t z
Fig. 2.1 Temperature vectors
t  δt
t + δt
2.2 Temperature Field and Temperature Gradient
9
y qy q
qx
qz
x
z
Fig. 2.2 Vector representation of Fourier’s law
axes ox, oy and oz. Accordingly, the components of the normal heat flow vector ~ q are (Fig. 2.2) @t @x @t qy ¼ k @y @t qz ¼ k @z
qx ¼ k
ð2:6Þ
where the negative sign indicates that the heat flow is in the direction of negative temperature gradient. The vector representing the rate of heat flow is ~ q ¼ iqx þ jqy þ kqz
ð2:7Þ
The vector q at a point is along a path normal to the isothermal surface at that point.
2.3
Thermal Conductivity
As already mentioned, the thermal conductivity is a physical property of a substance. In general, it depends on temperature, pressure and nature of the substance. Mostly, it is determined experimentally. Typical values of the thermal conductivity of several common engineering materials are given in a tabulated form in Appendix A. The thermal conductivity of the materials of interest differs by many thousand times. The highest values are for the metals, followed by dense ceramics, organic solids and liquids, while the lowest values are for the gases. Superinsulations have been developed for cryogenic applications with thermal conductivity as low as 0.3 mW/(m °C).
2.3.1 Thermal Conductivity of Solids Heat is transmitted through the solids by the elastic vibrations of the atoms and molecules (crystal lattice vibrations) and by free electrons (electronic thermal conduction). The transfer of heat by free electrons is very effective. The mechanism of electronic thermal conduction is similar to the electric conduction. When the temperature at one end of a rod is raised, the
10
2
OneDimensional SteadyState Heat Conduction
electrons in that region receive a somewhat higher energy and acquire an increased velocity. As these electrons move through the solid, they collide with other particles and transfer their energy to these particles. Thus the heat is transferred from one end to the other. As the electrons can be accelerated to a very high velocity, the electronic thermal conduction is more effective than the molecular conduction.
2.3.2 Thermal Conductivity of Metals and Alloys Metals have large number of free electrons (referred to as electron gas) and they contribute most to the heat transfer. Hence, the metals have high thermal conductivity. Since the electrons are the carriers of the heat and electricity in metals, the thermal conductivity of a metal is proportional to the electric conductivity. In general, good electrical conductors are almost always good heat conductors. Examples are silver, copper and aluminium. Since the scattering of the electrons intensiﬁes with rising temperature, the thermal and electrical conductivities decline with the increase in the temperature. The thermal conductivity of the metals drops sharply in the presence of impurities. The phenomenon can also be explained by an increase in structure heterogeneity, which causes electron scattering. In contrast to the pure metals, the thermal conductivity of the alloys increases with rise in the temperature. Electrical insulators are usually bad heat conductors.
2.3.3 Thermal Conductivity of Construction and HeatInsulating Materials Insulators and nonmetals have lower conductivities because the heat is transferred by the vibrations of the atoms only. Thermal conductivity of the nonmetals (solid dielectrics) depends on their density, structure, porosity and moisture content. Their thermal conductivity usually increases with temperature. Many construction and heatinsulating materials (bricks, concrete, asbestos, etc.) have a porous structure, and it is not possible to consider them continuous. Their thermal conductivity is taken as the thermal conductivity of a solid of the same shape and size, through which the same amount of heat is transmitted under the given temperature and boundary conditions. The thermal conductivity of porous solids and powders depends on their volumetric density. For example, the thermal conductivity of asbestos powder increases from about 0.1 to 0.25 W/(m K) with the increase in the density from 400 to 800 kg/m3 because the thermal conductivity of the air ﬁlling the pores is much less than that of the solid constituents of the porous material. The thermal conductivity of the damp porous material is higher than the dry material. The thermal conductivity of a damp material is signiﬁcantly greater than the thermal conductivities of the dry material and water taken separately. This can be attributed to the capillary movement of water within the material. The thermal conductivity of the construction and building materials range from about 0.03 to 3.0 W/(m K). Materials having thermal conductivity below 0.25 W/(m K) are used for heat insulation.
2.3 Thermal Conductivity
11
2.3.3.1 RValues of Insulating Materials An insulating material’s resistance to conductive heat flow is measured in terms of Rvalue, which is calculated from R ¼ d=k where d is the thickness of the material in m and k is the thermal conductivity in W/(m K). Thus, the Rvalue is measured in m2 K/W in SI units. In FPS system of units (used in the United States), Rvalues are measured in ft2.°F. hr/Btu, which are 5.67 times larger than Rvalues in SI units. The Rvalue should not be confused with thermal resistivity. It can be interpreted as the thermal resistance of a 1 m2 crosssection of the material. The higher the Rvalue of an insulating material, the greater is the thermal resistance the material offers to conduction heat flow. For calculating the Rvalue of a multilayered insulation system, the Rvalues of the individual layers are added. It is to note that the Rvalues take only conduction into account. It does not include convection and radiation.
2.3.4 Thermal Conductivity of Gases The heat conduction in gases at ordinary pressure and temperature occurs through the transport of the kinetic energy of the molecules because of their random motion and collision. The molecules in the hightemperature region have higher kinetic energy than those in the lower temperature region. When the molecules from the hightemperature region move to the lowtemperature region, they give up a part of their kinetic energy through collision with the lowtemperature molecules. From the kinetic theory of gases, the thermal conductivity of a gas is given by k ¼ ð1=3Þvm lqcv
ð2:8Þ
where l is the mean free path and vm is the mean velocity of the molecules. With the increase in the pressure, the gas density increases but the product (lq) remains constant. Therefore, the thermal conductivity does not depend on pressure with the exception of very high (of the order of critical pressure or more) or very low pressures. Since the mean travel velocity of the gas molecules depends on the temperature, and the speciﬁc heat of diatomic and polyatomic gases increase with the temperature, their thermal conductivity increase with the temperature. The thermal conductivity of helium and hydrogen is ﬁve to ten times greater than that of other gases. Due to the small mass of their molecules, the mean velocity is high. This explains the reason for their high thermal conductivity. The thermal conductivity of steam and other imperfect gases depend on pressure. The thermal conductivity of a mixture of the gases cannot be determined from the additive law and is found experimentally.
2.3.5 Thermal Conductivity of Liquids The mechanism of heat conduction in liquids is qualitatively the same as in the gases. However, due to the closely spaced molecules, the molecular force exerts a strong influence on the energy transfer by the collision. Since the density of liquids decreases with the increase in the temperature, their thermal conductivity decreases with the increase in the temperature. However, water and glycerin are
12
2
OneDimensional SteadyState Heat Conduction
important exceptions. In case of water, it increases with temperature in one temperature range and decreases in another. With rising pressure, the thermal conductivity of the liquids increases. In brief, the gases transfer heat by direct collisions between molecules and their thermal conductivity is low compared to most solids because they are dilute media. Nonmetallic solids transfer heat by lattice vibrations so that there is no net motion of the media as the energy propagates through them. Such heat transfer is often described in terms of ‘phonons’, quanta of lattice vibrations. Metals are much better thermal conductors than nonmetals because the mobile electrons which participate in electrical conduction also cause the transfer of heat. The metals which are good electrical conductors are also good heat conductors. At a given temperature, the thermal and electrical conductivities of metals are proportional. This behaviour is quantiﬁed by the Wiedemann–Franz Law (Fraas Arthur 1989). The law (Holman 1992) states that the ratio of the thermal conductivity k to the electrical conductivity r of a metal is proportional to the temperature T, i.e. k ¼ LT r
ð2:9Þ
where L is a constant of proportionality known as the Lorenz number. Its value is 2.44 10−8 WXK−2. However, the experiments have shown that the value of Lorenz number is not exactly the same for all materials. Its value, as reported in the literature, ranges from 2.23 10−8 WXK−2 for copper to 3.08 10−8 WXK−2 for tungsten at 0°C. Qualitatively, the relationship of the thermal and electrical conductivities is based on the fact that the both heat and electrical transports involve the transport of free electrons in the metals. Interested readers may refer to https://www.doitpoms.ac.uk/tlplib/thermal_electrical/ printall.php for greater details.
2.4
General Heat Conduction Equations
In general, the temperature gradient may exist in all three directions of a solid. There may be internal heat generation. The temperature can also vary with the time (unsteady state). Hence, it is necessary to develop a general heat conduction equation, which can be used to evaluate the heat transfer in any direction under steady or unsteady state and with or without heat generation. First, we consider a rectangular parallelepiped and then a cylinder and spherical system of coordinates.
2.4.1 General Heat Conduction Equation in Cartesian Coordinates Consider an elementary parallelepiped of a solid as shown in Fig. 2.3. The volume of the element is dV = dx dy dz. Let dQx is the quantity of the heat entering the solid through the face area (dy dz) in time ds as shown in the ﬁgure. Using Fourier’s heat conduction equation for the unidirectional conduction heat flow, we have
2.4 General Heat Conduction Equations
13 y
δQy+δy δQz δQx+δx
δQx δy
x δQz+δz
δz δx
δQy
z
Fig. 2.3 Heat conduction in rectangular parallelepiped system
dQx ¼ kðdydzÞ
@t ds @x
ðiÞ
where the temperature gradient in the xdirection has been taken as partial derivative because the temperature is function of x, y, z and time s. The quantity of heat leaving the element can be obtained by Taylor’s expansion of dQx: dQx þ dx ¼ dQx þ
@ ðdQx Þdx þ higher order terms @x
ðiiÞ
The net heat inflow into the element due to the difference of heat entering and leaving is (leaving the higher order terms) @ @t dQx dQx þ dx ¼ kx ðdydzÞ ds dx @x @x @ @t kx ¼ dxdydz ds @x @x @ @t kx dVds ¼ @x @x
ðiiiÞ
Similarly in the y and zdirections,
@ @t ky dVds @y @y @ @t kz ¼ dVds @z @z
dQy dQy þ dy ¼
ðivÞ
dQz dQz þ dz
ðvÞ
The net amount of heat stored in the element due to difference in conduction heat flow in and out of the element is the sum of Eqs. (iii), (iv) and (v), i.e. dQ ¼
@ @t @ @t @ @t kx ky kz þ þ dVds @x @x @y @y @z @z
ðviÞ
14
2
OneDimensional SteadyState Heat Conduction
If there is heat generation in the element at a rate of qg per unit volume per unit time, then the heat generated in the element in time ds is dQg ¼ qg dVds
ðviiÞ
The heat stored in the element due to the difference of conduction heat flow and heat generated within the element increases the internal energy of the element, i.e. ðviiiÞ
dE ¼ dQ þ dQg The change in the internal energy is also given by dE ¼ mcdT ¼ qdVc
@t ds @s
ðixÞ
where m is the mass of the element and c is the speciﬁc heat of the material. Hence, dQ þ dQg ¼ qdVc
@t ds @s
ðxÞ
Substitution of the values of dQ and dQg gives
@ @t @ @t @ @t @t kx ky kz þ þ dVds þ qg dVds ¼ qdVc ds @x @x @y @y @z @z @s
or @ @t @ @t @ @t @t kx ky kz þ þ þ qg ¼ qc @x @x @y @y @z @z @s
ð2:10Þ
which is the general heat conduction equation in Cartesian coordinates. Case (A) Homogeneous materials (kx = ky = kz = k) If the thermal conductivity is constant, i.e. kx = ky = kz = k, which is true for homogeneous materials, then the general heat conduction equation can be written as @2t @2t @2t qg qc @t ¼ þ þ þ 2 2 2 @x @y @z k @s k
ð2:11Þ
The term (k/qc) is called the thermal diffusivity and is denoted by a, which has been discussed later on. Hence, the above equation can be written as @2t @2t @2t qg 1 @t ¼ þ þ þ 2 2 2 @x @y @z a @s k
ð2:12Þ
2.4 General Heat Conduction Equations
15
Case (B) No Heat generation When there is no heat generation qg = 0 and we have @2t @2t @ 2 t 1 @t þ þ ¼ @x2 @y2 @z2 a @s
ð2:13aÞ
or D2 t ¼
1 @t a @s
ð2:13bÞ
The above equation is known as Fourier’s equation. Case (C) Steadystate heat conduction For steadystate heat conduction, the derivative of the temperature with time is zero and the general heat conduction equation yields @2t @2t @2t qg ¼0 þ 2þ 2þ 2 @x @y @z k
ð2:13cÞ
This equation is known as Poisson’s equation. Case (D) Steadystate heat conduction with no heat generation For this condition, general heat conduction equation yields @2t @2t @2t þ þ ¼0 @x2 @y2 @z2
ð2:13dÞ
D2 t ¼ 0
ð2:13eÞ
or
This equation is known as the Laplace equation. Case (E) Two and onedimensional steadystate heat conduction without heat generation For two and onedimensional steadystate heat conduction without heat generation, the Laplace equation yields the following equations. For twodimensional case, @2t @2t þ 2¼0 2 @x @y
ð2:14Þ
@2t ¼0 @x2
ð2:15Þ
For onedimensional case,
16
2
OneDimensional SteadyState Heat Conduction
2.4.1.1 Thermal Diffusivity In steady state, the thermal conductivity is the only property of the substance that determines the temperature distribution across the substance. But in the unsteady state, the temperature distribution is influenced by the thermal conduction as well as the thermal capacity (qc) of the material. Thus, the thermal diffusivity, which is a combination of these thermophysical properties of the material, controls the temperature distribution in the unsteady state. Thermal diffusivity is a property of a material (being a combination of three physical properties q, c and k). The higher the thermal diffusivity of a material, the higher the rate of temperature propagation, i.e. the equalization of the temperature at all points of the space will proceed at a higher rate. It can also be termed as a measure of the thermal inertia of a substance. In general, the metals have high thermal conductivity, which impart them high thermal diffusivities. The nonmetals and insulators have lower heat capacities than the metals but their thermal conductivities are much lower, and this leads to low values of thermal diffusivity of these solids compared to the metals. Typical values of q, c, k and a of some substances are given in Table 2.1. More details can be seen in Appendix A. Temperature gradients by local heating cause localized stress in a material, which may be of great signiﬁcance in some applications. Local heating is observed in welded joins, parts of internal combustion engines, at furnace inner surfaces, and in massive concrete structures during hydration of the cement. When heat is applied locally to a small area on the surface of a solid, the solid absorbs heat locally at a rate depending on its heat capacity. The heat is conducted away from the heated location at a rate depending on its conductivity. Thus, a high heat capacity means higher temperature gradients, while a high thermal conductivity causes
Table 2.1
Thermophysical properties of some materials
Material
Density q, kg/m3
Speciﬁc heat c, kJ/(kg K) at 20°C
Thermal conductivity k, W/(m K)
Thermal diffusivity a, (m2/s) 106
Metals (at 20°C) Aluminium
2710
0.895
204
84.1
Brass (70% Cu, 30% Zn)
8520
0.380
110
34
Copper
8950
0.380
386
113.5
Silver (99.9%)
10,520
0.234
419
170.2
Steel (C = 0.5%)
7830
0.465
55
15.1
18–8 Stainless steel
7820
0.460
16.3
4.53
Mercury(at 0°C)
13,630
0.14
8.2
4.3
Brick
1600
0.84
0.69
0.51
Concrete
1900
0.88
1.37
0.82
Water (at 25°C)
997.4
4.179
0.604
0.145
Air (at 300 K)
1.177
1.006
0.0262
22.1
Nonmetals
Liquids and gases
2.4 General Heat Conduction Equations
17
high rate of carrying away of heat to other parts of the solid that tends to reduce the temperature gradients. These two opposing effects are combined in the thermal diffusivity. A high thermal diffusivity reduces the temperature gradients caused by the local heating and, thus, reduces the local stresses.
2.4.2 General Heat Conduction Equation in Cylindrical Coordinates The equation in cylindrical coordinates is useful in dealing with the conduction heat transfer in systems with cylindrical geometry such as pipes, wires, rods, etc. Figure 2.4 shows the primary element with dimensions dr in direction r, rdh in direction h and dz in direction z (along the axis of the cylinder). The volume of the element is (dr rdh dz). Let dQr is the quantity of heat entering the element through the face whose area is rdh dz in time ds. Using Fourier’s equation for unidirectional heat flow, we have dQr ¼ kðrdhdzÞ
@t ds @r
ðiÞ
where the temperature gradient in direction r is the partial derivative in radial direction. The quantity of heat leaving the element can be obtained by Taylor’s expansion of dQr: dQr þ dr ¼ dQr þ
@ ðdQr Þdr þ higher order terms @r
ðiiÞ
The net flow into the element due to the difference of the heat entering and leaving the element in rdirection (neglecting higher order terms) is
δQr+δr
δQθ+δθ
δQz+δz δz
δr δQz Element
r
δQr
δθ
δQθ Element enlarged
δr rδθ
δr δz
Projections
Fig. 2.4 Heat conduction in a cylindrical system
18
2
dQr dQr þ dr
OneDimensional SteadyState Heat Conduction
@ @t kðrdh dzÞ ds dr ¼ @r @r @ @t ðdr rdh dzÞ r ds ¼k @r @r r [email protected] @t r ¼k dV ds r @r @r 1 @t @2t þ r 2 dV ds ¼k r @r @r 1 @t @2t þ dVds ¼k r @r @r 2
ðiiiÞ
The area of the face perpendicular to the hdirection of the heat flow is (dr dz), and the net heat flow into the element is dQh dQh þ dh
@ @t kðdr dzÞ ds rdh ¼ [email protected] [email protected] 1 @ @t ¼k ðdr rdh dzÞds r @h [email protected] 1 @2t ¼ k 2 2 dVds r @h
ðivÞ
Similarly, the area of the face perpendicular to the zdirection of the heat flow is (dr rdh) and the net heat flow into the element is dQz dQz þ dz
@ @t kðdr rdhÞ ds dz ¼ @z @z 2 @ t ¼ k 2 dr rdh dz ds @z @2t ¼ k 2 dVds @z
ðvÞ
The net amount of heat stored in the element of Fig. 2.4 is the sum of the net heat flow into the element in r, h and zdirections, i.e. the sum of the righthand side terms of Eqs. (iii), (iv) and (v), and is dQ ¼ k
1 @t @2t 1 @2t @2t þ 2 þ 2 2 þ 2 dVds r @r @r r @h @z
ðviÞ
If there is internal heat generation at the rate of qg per unit volume per unit time, then the heat generated in the element in time ds is dQg ¼ qg dVds
ðviiÞ
2.4 General Heat Conduction Equations
19
The net heat dQ stored in the element due to the difference of in and outflows, and the heat generation dQg within the element increase the internal energy of the element hence dE ¼ dQ þ dQg
ðviiiÞ
Substitution of the values of dQ and dQg gives dE ¼ k
1 @t @2t 1 @2t @2t þ 2 þ 2 2 þ 2 dVds þ qg dVds r @r @r r @h @z
ðixÞ
The change in the internal energy is also given by dE ¼ qdVc
@t ds @s
ðxÞ
Substitution gives @t 1 @t @2t 1 @2t @2t qdVc ds ¼ k þ þ þ dV ds þ qg dV ds @s r @r @r 2 r 2 @h2 @z2 or 1 @t @2t 1 @2t @2t qg qc @t þ 2þ 2 2þ 2 þ ¼ r @r @r r @h @z k @s k or 1 @t @2t 1 @2t @2t qg 1 @t þ 2þ 2 2þ 2 þ ¼ r @r @r r @h @z a @s k
ð2:16Þ
which is the general heat conduction equation in cylindrical coordinates, where the thermal conductivity has been assumed to be constant. The equation can be arrived at by the transformation of Eq. (2.12) to cylindrical coordinates, see Example 2.1. For variable thermal conductivity, the equation in general form can be written as 1 [email protected] @t 1 @ @t @ @t qg @t kr r kh kz ¼ þ 2 þ þ qc r @r @r r @h @h @z @z qc @s
ð2:17aÞ
If there is no internal heat generation and the thermal conductivity is constant, then qg = 0 and kr = kh = kz = k. Equation (2.16) reduces to 1 @t @2t 1 @2t @ 2 t 1 @t þ 2þ 2 2þ 2¼ r @r @r r @h @z a @s
ð2:17bÞ
For the steadystate heat conduction with no heat generation and constant value of the thermal conductivity k, Eq. (2.16) reduces to
20
2
OneDimensional SteadyState Heat Conduction
1 @t @2t 1 @2t @2t þ 2þ 2 2þ 2 r @r @r r @h @z
¼0
ð2:17cÞ
In the case of steadystate and onedimensional heat conduction (in the radial direction only) without heat generation, the above equation reduces to 1 dt d2 t þ 2¼0 r dr dr
ð2:17dÞ
The temperature distribution in this case is a function of radius r only; hence, the partial derivative has been changed to the full derivative. Example 2.1 Beginning with the threedimensional heat conduction equation in Cartesian coordinates, obtain the equation in cylindrical coordinates. Solution The rectangular and cylindrical coordinates are interrelated by the following relations, refer to Fig. 2.5: x ¼ r cos h y ¼ r sin h and z¼z
ðaÞ
The general heat conduction equation in rectangular coordinates is @2t @2t @2t qg 1 @t ¼ þ 2þ 2þ 2 @x @y @z a @s k
ð2:12Þ
which is transformed into cylindrical coordinates using Eq. (a), knowing that x and y are functions of r and h while z is independent of r and h.
z P(r, θ, z)
z θ y
Fig. 2.5 Example 2.1
x = r cosθ
r
x y = r sinθ
2.4 General Heat Conduction Equations
21
@t @t @x @t @y @t @t þ ¼ cos h þ sin h ¼ @r @x @r @y @r @x @y
ðiÞ
and
@t @h
¼
@t @x @t @y @t @t þ ¼ r sin h þ r cos h @x @h @y @h @x @y
Solution of Eqs. (i) and (ii) gives
@t sin h @t ¼ cos h @r r @h
@t cos h @t ¼ sin h þ @r r @h
@t @x
and @t @y
Second derivatives of t with x and y are @ @t @ @t sin h @ @t ¼ cos h @x @x @r @x r @h @x or @ @t @ @t sin h @t sin h @ @t sin h @t cos h cos h ¼ cos h @x @x @r @r r @h r @h @r r @h or 2 2 2 2 @2t sin h cos h @t sin h @t sin h @ t sin h cos h @t 2 @ t þ þ ¼ cos h þ þ @x2 @r 2 r2 @h r @r r2 r2 @h @h2 Similarly, @ @t @ @t cos h @ @t ¼ sin h þ @y @y @r @y r @h @y or @ @t @t @t cos h @t cos h @ @t cos h @t sin h þ sin h þ ¼ sin h þ @y @y @r @r r @h r @h @r r @h or 2 @2t sin h cos h @t cos2 h @t cos2 h @ 2 t sin h cos h @t 2 @ t þ ¼ sin h þ @y2 @r 2 r2 @h r @r r 2 @h2 r2 @h
ðiiÞ
22
2
OneDimensional SteadyState Heat Conduction
Hence, @2t @2t @2t 2 1 @t 1 @2t 2 þ ¼ cos h þ sin h þ þ 2 2 @x2 @y2 @r 2 r @r r @h or @2t @2t @2t 1 @t 1 @2t þ ¼ þ þ @x2 @y2 @r 2 r @r r 2 @h2 Substitution in Eq. (2.12) gives the desired equation as 1 @t @2t 1 @2t @2t qg 1 @t þ 2þ 2 2þ 2 þ ¼ r @r @r r @h @z a @s k
ð2:16Þ
2.4.3 General Heat Conduction Equation in Spherical Coordinates Figure 2.6a shows the elemental area on the surface of a sphere of radius r. For the strip ABCD in the ﬁgure, the radius is rsinh. Also refer to the section of the sphere in projections in Fig. 2.6c. Thus, the sides of the elemental area are rdh and rsinh d/. The elemental volume can be made by moving a distance dr in the direction r as shown in Fig. 2.6b. Thus, the sides of the element in Fig. 2.6b are dr, rdh and rsinh d/ in the directions r, h and /, respectively. The volume of the element is dV ¼ dr rdh r sin h d/
ðiÞ
The heat dQr in rdirection enters the element through face of area (rdh rsinh d/), which is perpendicular to rdirection. The heat inflow in time ds is dQr ¼ kðrdh r sin h d/Þ
@t ds @r
ðiiÞ
The outflow is dQr þ dr ¼ dQr þ
@ ðdQr Þdr @r
ðiiiÞ
The net heat inflow into the element is @ @ dQr dQr þ dr ¼ dQr dQr þ ðdQr Þdr ¼ ðdQr Þdr @r @r @ 2 @t r ¼ kðdr dh sin h d/Þ ds @r @r 1 @ 2 @t r ds ¼ kdV 2 r @r @r
ðivÞ
2.4 General Heat Conduction Equations
23
Fig. 2.6 Heat conduction in a spherical system
Flow area perpendicular to the h direction is (dr rsinh d/), and the elemental distance in h direction is rdh. The net heat inflow into the element in this direction is
dQh dQh þ dh
@ @ ¼ dQh dQh þ ðdQh Þrdh ¼ ðdQh Þrdh [email protected] [email protected] @ @t kðdr r sin h d/Þ ds rdh ¼ [email protected] [email protected] @ @t sin h ¼ kðdr d/Þdh ds @h @h 1 @ @t sin h ds ¼ kðdr rdh r sin h d/Þ 2 r sin h @h @h 1 @ @t sin h ds ¼ kdV 2 r sin h @h @h
ðvÞ
24
2
OneDimensional SteadyState Heat Conduction
In the similar way, the flow area perpendicular to the / direction is (dr rdh) and the elemental distance in / direction is rsinh d/. The net heat inflow into the element in this direction is @ @ dQ/ dQ/ þ d/ ¼ dQ/ dQ/ þ dQ/ r sin d/ ¼ dQ/ r sin hd/ r sin [email protected]/ r sin [email protected]/ @ @t kðdr rdhÞ ds r sin hd/ ¼ r sin [email protected]/ r sin [email protected]/ 2 1 @ t ¼ kðdr rdh r sin h d/Þ 2 2 ds r sin h @/2 2 1 @ t ds ðviÞ ¼ kdV 2 2 r sin h @/2
The total of the net inflows into the element is the summation of the net of the conduction heat inflow and outflow in all three directions of the heat flow, and is 1 @ 1 @ @t 2 @t r sin h ds þ kdV ds r 2 @r @r r 2 sin h @h @h 2 1 @ t ds þ kdV 2 2 r sin h @/2 2 1 @ 1 @ @t 1 @ t 2 @t r sin h þ 2 þ 2 2 ¼ kdV 2 ds r @r @r r sin h @h @h r sin h @/2
dQ ¼ kdV
ðviiÞ
Heat generated in the element in time ds is dQg ¼ qg dVds
ðviiiÞ
Change in the internal energy of the element in time ds is dE ¼ qdVc
@t ds @s
ðixÞ
Energy balance for the element gives dE ¼ dQ þ dQg Substitution of the values of dE, dQ and dQg gives
@t qdV c ds @s 2 1 @ 1 @ @t 1 @ t 2 @t r sin h þ 2 þ 2 2 ds ¼ kdV 2 r @r @r r sin h @h @h r sin h @/2 þ qg dVds or
ðxÞ
2.4 General Heat Conduction Equations
25
2 qc @t 1 @ 1 @ @t 1 @ t qg 2 @t r sin h ¼ 2 þ 2 þ 2 2 þ 2 k @s r @r @r r sin h @h @h k r sin h @/ or 2 1 @ 1 @ @t 1 @ t qg 1 @t 2 @t r sin h ¼ þ þ þ r 2 @r @r r 2 sin h @h @h a @s k r 2 sin2 h @/2
ð2:18Þ
which is the desired equation assuming thermal conductivity as constant.
2.5
OneDimensional SteadyState Heat Conduction
For the onedimensional steadystate heat conduction without heat generation, @2t ¼0 @x2
ð2:15Þ
Integrating the above equation, we obtain @t ¼ C1 @x
ðiÞ
t ¼ C1 x þ C2
ðiiÞ
and
The boundary conditions in Fig. 2.7 are at x ¼ x1 ;
t ¼ t1
at x ¼ x2 ;
t ¼ t2
and
t1 q
t2 Thickness of the wall = x2 — x1 = δ
x1 x2 x=0
Fig. 2.7 Onedimensional steadystate heat conduction through a plane wall
26
2
OneDimensional SteadyState Heat Conduction
Substitution in Eq. (ii) gives t1 ¼ C1 x1 þ C2 t2 ¼ C1 x2 þ C2 Solution of these equations gives t1 t2 t1 t2 ¼ x1 x2 x2 x1 t1 t2 C2 ¼ t1 þ x1 x2 x1
C1 ¼
Substitution of values of constants C1 and C2 in Eq. (ii) gives t¼
t1 t2 t1 t2 xþ x1 þ t 1 x2 x1 x2 x1
or t t1 x1 x ¼ t 1 t 2 x2 x1 which indicates a linear distribution of the temperature, refer to Fig. 2.7. From Eq. (i), @t t1 t2 ¼ C1 ¼ @x x2 x1 Hence, the rate of heat transfer is @t t1 t2 t1 t2 ¼ kA ¼ kA @x x2 x1 d t1 t2 Dt ¼ ¼ Rk d=kA
q ¼ kA
ð2:19Þ
where A is wall area perpendicular to the direction of heat flow and Rk = d/kA. We compare the above equation with Ohm’s law for an electric conductor, which is I¼
E V 1 V2 ¼ R R
The electric current I corresponds to the heat flow q, the electrical potential E corresponds to the thermal potential and the electrical resistance corresponds to resistance Rk to the heat conduction. Thus, Fourier’s equation of heat conduction is exactly analogous to Ohm’s law for an electrical conductor. We shall use this electrical analogy frequently as it is quite useful in solving the complex heat conduction problems. We represent Eq. (2.19) as in Fig. 2.8, where the temperature difference Δt is the driving force for the flow of heat and Rk = R (= d/kA) is the thermal resistance, which the wall offers to the flow of heat by conduction. The reciprocal of the thermal resistance is known as thermal conductance of the wall.
2.5 OneDimensional SteadyState Heat Conduction
27
R = δ/kA t2
t1
q Δt
Fig. 2.8 Thermal network
2.5.1 Composite Plane Wall Walls made of several layers of different materials are called composite walls, refer to Fig. 2.9. Walls of houses, furnaces, boilers, pipes with layer of insulating materials, etc. are some of the examples of composite walls. The composite wall in Fig. 2.9 consists of three layers of thicknesses d1, d2 and d3. The thermal conductivities of these layers are k1, k2 and k3, respectively. The temperature of the outer layers of the wall is t1 and t4 as shown in the ﬁgure, with interface temperatures as t2 and t3. It is being assumed that different layers are having perfect contact between them, and hence the adjacent surfaces are at the same temperature. In the steadystate condition, the heat flow q is the same for all the layers and is constant. The equations of heat transfer through these layers are t1 t2 d1
q ¼ k1 A q ¼ k2 A
t2 t3 d2
q ¼ k3 A
for the first layer
ðiÞ
for the second layer
ðiiÞ
for the third layer
ðiiiÞ
t3 t4 d3
The temperature differences across the layers, from above equations, are t1 t2 ¼ q
k1
q R1 = δ1/k1A 1
t2 t3
R2 = δ2/k2A 1 T4
δ1 t1
δ2 t2
R1
Fig. 2.9 A composite wall
ðivÞ
k3
k2
t1
d1 k1 A
q R2
δ3 t3
t4 R3
R3 = δ3/k3A 1
28
2
OneDimensional SteadyState Heat Conduction
d2 t2 t3 ¼ q k2 A d3 t3 t4 ¼ q k3 A
ðvÞ ðviÞ
Adding the above equations, we get
d1 d2 d3 þ þ t1 t4 ¼ q k1 A k2 A k3 A
ðviiÞ
or q¼
d1 k1 A
t1 t4 þ kd22A þ
d3 k3 A
or q¼
t1 t4 R 1 þ R 2 þ R3
ðviiiÞ
where R1, R2 and R3 are resistances of the layers to the conduction heat flow. For n layers, the equation can be written as t1 tn þ 1 t1 tn þ 1 q ¼ Pn ¼ 1 Pn d i i¼1 Ri i¼1 k A
ð2:20Þ
i
2.5.2 OneDimensional SteadyState Heat Conduction Through a Plane Homogeneous Wall Considering Film Coefficients In Fig. 2.10, the two faces of the wall are exposed to fluids at temperatures ti and to. The heat is transferred to and from the wall by convection. In the convection heat transfer process, a thin fluid ﬁlm is formed on the surface, which offers resistance to the flow of heat (refer to Chap. 7 for details). Fluid film ti• t1 •
q
• ho •
hi •
• to
δ
ti
1/(hi A)
(δ/kA) t1
1/(ho A)
to
t2
Fig. 2.10 Onedimensional steadystate heat conduction through a plane wall considering ﬁlm coefﬁcients
2.5 OneDimensional SteadyState Heat Conduction
29
Newton’s equation for convection heat transfer gives q ¼ hADt
ð1:4Þ
where h A Δt
is known as convective heat transfer or ﬁlm coefﬁcient, is the area of the surface transferring heat by convection and is temperature difference between the wall (surface) and the fluid in contact. Hence, in the present case, at the left face, q ¼ hi Aðti t1 Þ ti t1 ti t1 ¼ ¼ 1=hi A Ri
ðiÞ
where Ri = 1/hiA is known as ﬁlm resistance. Similarly at the right face q ¼ ho Aðt2 to Þ t2 to t2 to ¼ ¼ 1=ho A Ro
ðiiÞ
From Eqs. (i) and (ii), ti t1 ¼ qRi t2 to ¼ qRo For the wall,
d t1 t2 ¼ q kA
ðiiiÞ
Combing the temperature equations, we get
1 d 1 þ þ ðti t1 Þ þ ðt1 t2 Þ þ ðt2 to Þ ¼ q hi A kA ho A ti to q¼ 1 d 1 hi A þ kA þ ho A
ð2:21Þ
which is the desired equation for the present case.
2.5.2.1 Conduction Heat Transfer Through a Composite Plane Wall Considering Film Coefficients Introducing ﬁlm coefﬁcients in Eq. (2.20), the heat transfer equation takes the form as q¼
1 hi A
þ
1 A
ti to Pn d i i¼1 ki
þ
1 ho A
ð2:22Þ
30
2
OneDimensional SteadyState Heat Conduction
For example, for a wall with three layers (n = 3), the equation yields q¼
1 hi A
þ
d1 k1 A
ti to þ kd22A þ
d3 k3 A
þ
1 ho A
2.5.3 OneDimensional SteadyState Conduction Heat Transfer Through a Plane Homogeneous Wall Considering Heat Transfer by Convection and Radiation from the Wall Surface In general, the wall surface will transfer heat both by convection and radiation to its surroundings. The expression for the heat flow to the wall can be written in the following form: q ¼ ðhri þ hci Þ:A:ðti t1 Þ ¼
ðti t1 Þ 1=½ðhri þ hci ÞA
ðiÞ
where hr is termed as the radiation heat transfer coefﬁcient. Similarly, heat flow from the wall is q ¼ ðhro þ hco Þ:A:ðt2 to Þ ¼
ðt2 to Þ 1=½ðhro þ hco ÞA
ðiiÞ
Considering the electrical analogy, the overall resistance to heat flow is 1 d 1 þ þ ðhri þ hci ÞA kA ðhro þ hco ÞA
ðiiiÞ
Here 1/hr and 1/hc are two resistances Rr and Rc, respectively, in parallel, refer to Fig. 2.11. The heat transfer equation takes the form q¼
1 ðhri þ hci ÞA
ti to d þ kA þ
1 ðhro þ hco ÞA
ð2:23Þ
Example 2.2 The heating surface of a boiler is in contact with flue gas (at 1200°C) on one side and boiling water (at 200°C) on the other side. The metal wall thickness is 10 mm, and its thermal conductivity is 45 W/(m K). The local or surface coefﬁcient of heat transfer from the flue gas to the wall is 50 W/(m2 K) and from wall to the boiling water is 4000 W/(m2 K). Calculate the rate of heat flow from the flue gas to the water and the surface temperatures. What will be the effect on heat flow rate if heat transfer coefﬁcients for one or both sides are doubled?
2.5 OneDimensional SteadyState Heat Conduction
31
ti • q t1•
• (hr + hc)o • t2
(hr + hc)i •
• to
δ1
a 1/hri A
1/hro A
ti
t1
δ/kA
t2
to
q 1/hci A
1/hco A
b
Fig. 2.11 Onedimensional steadystate heat conduction through a plane wall considering heat transfer by convection and radiation from wall surface
ti• t1• h i•
q
•ho • t2 • to
Fig. 2.12 Example 2.2
Solution (i) The resistance to the heat flow is (refer to Fig. 2.12) Rk ¼ Ri þ Rwall þ Ro 1 d 1 þ ¼ þ Ahi kA Aho 1 10 1 þ þ ¼ 0:02047: ¼ 1 50 1000 1 45 1 4000 The rate of heat transfer per unit area, q¼
ti to 1000 ¼ 48;852 W: ¼ 0:02047 Rk
Considering heat flow through the flue gas ﬁlm at the wall,
32
2
OneDimensional SteadyState Heat Conduction
q ¼ hi Að t i t 1 Þ or 48;852 ¼ 50 1 ð1200 t1 Þ or t1 ¼ 223 C Similarly at the other face, t2 ¼
q 48;852 þ 200 ¼ 212:2 C þ to ¼ ho A 4000
It is to note that the resistance of the wall material (Rwall) and that of the ﬁlm on the waterside (Ro) are negligible compared to the resistance of the ﬁlm (Ri) on the gas side. Thus, the resistance Ri is the controlling resistance and its change has a greater effect on the heat transfer. By following the above procedure, it can be shown that heat transfer increases to 95,493 W if hi is doubled as compared to q = 49,147 W when ho is doubled. Example 2.3 A furnace wall consists of 200 mm of refractory ﬁreclay brick, 100 mm of kaolin brick and 6 mm of steel plate. The ﬁreside of the refractory is at 1150°C, and the outside of the steel is at 30°C. An accurate heat balance over the furnace shows the heat loss from the wall to be 300 W/m2. It is known that there may be thin layers of air between the layers of brick and steel. To how many millimetres of kaolin are these air layers equivalent? The thermal conductivities are as follows: Refractory ﬁreclay bricks, kaolin brick, steel,
Solution
kf = 1.7 W/(m K) ki = 0.17 W/(m K) ks = 17 W/(m K)
Aðt1 t4 Þ q ¼ df ds di k f þ ki þ ks
Substitution gives 300 ¼
1 ð1150 30Þ di 0:2 0:006 1:7 þ 0:17 þ 17
which gives effective thickness of kaolin, di = 0.6146 m = 614.6 mm. Thus, the air layers are equivalent to (614.6 – 100) = 514.6 mm of kaolin. Example 2.4 The thermal conductivity of a cylindrical specimen 150 mm in diameter was determined using an instrument consisting of two flat plates between which the cylindrical specimen sides were placed. When heat transfer rate was 100 W, the hot and cold end plate
2.5 OneDimensional SteadyState Heat Conduction
33
temperatures were recorded to be 170°C and 20°C, respectively. The cylindrical specimen is 15 mm in length. It was found that due to improper matching air clearances of thickness of about 0.15 mm were formed between the cold and hot surfaces of the instrument and the specimen ends. Determine the thermal conductivity of the specimen. Conductivity of the air can be taken as 2.7 10−2 and 3.8 10−2 W/(m °C) at the cold and hot surfaces, respectively. Solution The resistance to the heat transfer, considering the air resistances, is d d d R¼ þ þ k air;cold k specimen k air;hot 0:15 15 0:15 ¼ þ þ 1000 2:7 102 1000 k 1000 3:8 102 0:015 ¼ 9:503 103 þ k The heat transfer from Fourier’s equation is q¼
Aðt1 t4 Þ Pd k
Substitution gives p 0:015 1 100 ¼ ð0:15Þ2 ð170 20Þ 9:503 103 þ 4 k Solution of above equation gives k = 0.882 W/(m K), which is the thermal conductivity of the specimen. Example 2.5 The walls of a paint drying chamber are builtup of a layer of brick [thickness d = 250 mm and k = 0.7 W/(m K)]. The temperature in the chamber is estimated to be 115° C. The heat flow from 1 m2 of the chamber wall is not to exceed 100 W when ambient temperature is 25°C for which a layer of felt [k = 0.045 W/(m K)] is to be applied outside the brick layer. Calculate the thickness of the felt if the surface heat transfer coefﬁcients at inner and outer walls are 30 W/(m2 °C) and 20 W/(m2 °C), respectively. Solution Total resistance to heat flow R¼
1 1 db df 1 þ þ þ A hi ho kb kf
34
2
OneDimensional SteadyState Heat Conduction
where subscripts b and f refer to brick and felt, respectively. R¼
1 1 0:25 df 1 þ þ þ 1:0 30 0:7 0:045 20
or ¼ 0:4405 þ 22:22df
ðiÞ
From heat transfer equation, q¼
t t i o R
or 100 ¼
115 25 R
or Rk ¼ 0:9 Substitution of this value of Rk in Eq. (i) gives df ¼ 0:0207 m ¼ 20:7 mm; say 21 mm
Example 2.6 The walls of a house have a composite construction of bricks [thickness = 330 mm, k = 0.8 W/(m K)] and 20mmthick plaster [k = 1.2 W/(m K)] on both sides. The inner and outer surface heat transfer coefﬁcients are 15 and 30 W/(m2 K). Determine the heat loss per m2 of the wall surface if the temperature inside the room is 22°C and outside temperature is –5°C. When the wind blows fast, the outside ﬁlm coefﬁcient increases to 200 W/(m2 °C). Determine the increase in the heat loss. Identify the controlling resistance that determines the heat flow rate. Solution The values of various resistances, for the unit area of the wall, are 1 1 ¼ 0:067 m2 K=W ¼ hi 15 d 20 ¼ 0:0167 m2 K=W Inner plaster; Rpi ¼ ¼ k 1000 1:2 d 330 ¼ 0:4125 m2 K=W Bricks; Rb ¼ ¼ k 1000 0:8 d 20 ¼ 0:0167 m2 K=W Plaster ðouter); Rpo ¼ ¼ k 1000 1:2 1 1 ¼ 0:033 m2 K=W Convective ðouterÞ; Ro ¼ ¼ ho 30 Convective ðinner); Ri ¼
2.5 OneDimensional SteadyState Heat Conduction
35
The total resistance, X
R ¼ Ri þ Rpi þ Rb þ Rpo þ Ro ¼ 0:5459
The heat flow rate, Ti To 22 ð5Þ ¼ 49:46 W=m2 q¼ P ¼ 0:5459 R When the wind blows, the outer convective resistance Ro changes to 1/200 = 0.005 and the total resistance, X
R ¼ Ri þ Rpi þ Rb þ Rpo þ Ro ¼ 0:5179
The new heat flow rate, q0 ¼
22 ð5Þ ¼ 52:13 W=m2 0:5179
Percentage change in heat flow ¼
52:13 49:46 100 ¼ 5:4% 49:46
which is small because the controlling resistance is the resistance of the bricks and is almost 75–80% of the total resistance to the heat flow. Example 2.7 A house has a multilayer composite wall constructed as shown in Fig. 2.13. The temperature of the air inside the room is 22°C, and the surface heat transfer coefﬁcient between the room air and the wall is 6 W/(m2 K). The outer surface heat transfer coefﬁcient is 20 W/(m2 K). The different wall thicknesses (d) in mm and the thermal conductivities
a ka = 0.05 kb = 0.04 kc = 0.25 kd = 0.3 Thickness in mm, height in m, and thermal conductivity in W/(m K).
1.2 c
d
1.8
b
200 50
10 (δ/k A)a
(δ/k A)c
ti 1/hi A (δ/k A)b
Fig. 2.13 Example 2.7
q
(δ/k A)d
1/ho A
to
36
2
OneDimensional SteadyState Heat Conduction
(k) in W/(m K) are indicted in the ﬁgure. Calculate the heat transfer rate across the wall section per m length of the wall. Outside air temperature is 5°C. Solution The resistance to heat transfer, refer to the thermal network in Fig. 2.13, P
¼
1 þ 63
dc dd 1 þ þ 1 1 þ ho A kc A kd A da þ kbdAb b ka Aa 1 0:2 0:05 1 ¼ 0:47 þ þ 1 0:01 1 þ 0:25 3 0:3 3 20 3 þ
1 R¼ þ hi A
0:01 0:051:2
1
0:041:8
Hence, the heat transfer rate is Dt 22 5 q¼P ¼ ¼ 36:17 W 0:47 R Example 2.8 Find the heat flow through a wall, per m2 of wall face area, shown in Fig. 2.14. The thermal conductivities of brick material and steel are 1.0 W/(m K) and 40 W/ (m K), respectively. Calculate the percentage reduction in the heat flow if the bolts are not used. What will be the temperature at the outer surface of the wall neglecting the effect of bolts? Solution Part (i) Figure 2.15 shows the analogous electrical circuit for various thermal resistances. It is to be noted that the bolts provide a parallel path to the flow of heat. The effect of bolt heads and nuts has been neglected.
Steel plate
Steel plate Steel bolts 10 nos. per m2 of the wall for fastening steel plates.
Bolt 20 mm dia. Furnace side 900 K
Air side T∞ = 300K
b
ho = 60 W/(m2 K) Thickness in mm.
100 5
Fig. 2.14 Wall crosssection
5
2.5 OneDimensional SteadyState Heat Conduction (δ/k A)steel
t1
(δ/k A)bricks
37 (δ/k A)steel
1/ho A
t∞
(δ/k A)bolts
Fig. 2.15 Thermal network with bolts
The heat flow areas and length of heat paths for different layers and bolts are 2
2
Abolts ¼ 10 p4 ðdbolt Þ ¼ 10 p4 ð0:02Þ ¼ 3:1416 103 m2 Asteel ¼ Abricks ¼ A Abolts ¼ 1:0 3:1416 103 ¼ 0:99686 m2 dsteel ¼ 0:005 m; dbrick ¼ 0:1 m; dbolt ¼ 0:110 m The values of various resistances are
d 0:005 ¼ 1:254 104 Rsteel ¼ ¼ kA steel 40 0:99686 d 0:1 ¼ 0:1003 Rbrick ¼ ¼ kA brick 1 0:99686 d 0:110 Rbolt ¼ ¼ ¼ 0:87535 kA bolt 40 3:1416 103 1 1 ¼ 0:0167: ¼ Ro ¼ ho A 60 1 The total resistance to heat transfer is X
R¼
1
Rsteel þ Rbrick þ Rsteel
þ
1 Rbolt
1
þ Ro ¼ 0:1069
The heat transfer, t1 t1 900 300 ¼ ¼ 5612:7 W q¼ P R 0:1069
Part (ii) Heat transfer without bolts: The values of various resistances are
d 0:005 Rsteel ¼ ¼ 1:25 104 ¼ kA steel 40 1:0 d 0:1 ¼ 0:1 Rbrick ¼ ¼ kA brick 1 1:0 1 1 ¼ 0:0167 Ro ¼ ¼ ho A 60 1
38
2
t1
(δ/k A)steel
(δ/k A)bricks
OneDimensional SteadyState Heat Conduction (δ/k A)steel
1/ho A
t∞
Fig. 2.16 Thermal network without bolts
The total resistance to heat transfer is (refer to network in Fig. 2.16) X
R ¼ Rsteel þ Rbrick þ Rsteel þ Ro ¼ 0:11695
The heat transfer t1 t1 900 300 ¼ 5130:4 W ¼ q¼ P 0:11695 R Percentage reduction in heat loss is Dq ¼
5612:7 5130:4 100 ¼ 8:6%: 5612:7
Part (iii) Temperature at the outer surface of the wall: The temperature T2 can be found from the heat transfer equation (neglecting the effects of the bolts): q¼
t2 t1 Ro
or 5130:3 ¼
T2 300 0:0167
or t2 ¼ 385:7 K
Example 2.9 A masonry wall consists of 100 mm brick outer face with 10 mm mortar joint, a 200 mm concrete wall and a 10 mm insulating board on the inside as shown in Fig. 2.17a. The outside and inside bulk air temperatures are 40 and 25°C, respectively. Determine the heat flux if the outside heat transfer coefﬁcient is 15 W/(m2 K) and the inside ﬁlm coefﬁcient is 5 W/(m2 K). Thermal conductivities given are bricks, kb = 1.2; mortar, km = 0.6; concrete, kc = 0.8; board, kb = 0.15 W/(m K).
2.5 OneDimensional SteadyState Heat Conduction
39
Concrete Bricks Insulating board 75 10 Mortar
Thickness in mm. 100
200 10
a Crosssection of wall. Rm Rc
Rib
ti
to Ro
q
Rb
b Thermal network. Fig. 2.17 Example 2.9
Solution Resistances for unit height and width of the wall: (i) Convective (outer) Ro ¼
1 1 ¼ 0:0666 ¼ ho Ao 15 1
(ii) Brick Rb ¼
db 0:1 ¼ 0:0944 ¼ kb Ab 1:2 75=85
Rm ¼
dm 0:1 ¼ 1:4166 ¼ km Am 0:6 10=85
(iii) Mortar
(iv) Concrete Rc ¼
dc 0:2 ¼ 0:25 ¼ kc Ac 0:8 1
(v) Insulating board Rib ¼
dib 0:01 ¼ 0:0666 ¼ kib Aib 0:15 1
Ri
40
2
OneDimensional SteadyState Heat Conduction
(vi) Convective (inner) Ri ¼
1 1 ¼ 0:2 ¼ h i Ai 5 1
All resistances are in K/W. The total resistance to heat flow is (refer to the network in Fig. 2.17b) X
R ¼ R0 þ
1 1 þ Rb R m
1
þ Rc þ Rib þ Ri ¼ 0:6717 K=W
Heat flux, To Ti 40 25 q00 ¼ P ¼ 22:33 W=m2 ¼ 0:6717 R which is from the outside to inside. Example 2.10 Rods of copper, brass and steel are welded together to form a Yshaped ﬁgure as shown in Fig. 2.18. The crosssectional area of each rod is 400 mm2. The end of the copper rod is maintained at 120°C and the ends of the brass and steel rods at 20°C. The lengths of the rods are copper 768 mm, brass 172 mm and steel 45 mm. (a) What is the temperature of the junction point? (b) What is the heat flowing in the copper rod? The thermal conductivities of copper, brass and steel are 384, 86 and 45 W/(m K), respectively. The rods are insulated from outside. Solution Heat balance at the junction gives Heat inflow at the junction through copper rod; qc ¼ heat out flow at the junction through brass and steel rods; ðqb þ qs Þ
Brass Steel qb
qs qc
Fig. 2.18 Example 2.10
Copper
2.5 OneDimensional SteadyState Heat Conduction
41
where, from Fourier’s equation, qc ¼ kA
Dt 400 120 t ¼ 384 ¼ 0:2 ð120 tÞ L 0:768 ð1000Þ2
qb ¼ kA
Dt 400 t 20 ¼ 86 ¼ 0:2 ðt 20Þ 2 L 0:172 ð1000Þ
qs ¼ kA
Dt 400 t 20 ¼ 45 ¼ 0:4 ðt 20Þ 2 L 0:045 ð1000Þ
Substitution in heat balance equation gives 0:2 ð120 tÞ ¼ 0:2 ðt 20Þ þ 0:4 ðt 20Þ or t ¼ 45 C This gives qc ¼ 0:2ð120 45Þ ¼ 15 W
Example 2.11 The walls of a cold storage are proposed to be made of three layers as shown in Fig. 2.19a. The temperature of the inside wall of the storage is likely to be –10°C. The outside surface of the wall is likely to be exposed to surrounding air at 20°C (dew point = 15°C) with ﬁlm coefﬁcient ho = 20 W/(m2 K). Determine the position in the wall
Wood [k = 0.2 W/(m K)] 10oC •
Cork [k = 0.04 W/(m K)]
Expanded Polystyrene [k = 0.035 W/(m K)]
• t∞ = 20oC ho = 20 W/(m2 K) 150
Thickness in mm.
25
25
a 10oC
R1
R3
R2 T1
T2
1/ho T3
b Fig. 2.19 a Crosssection of cold storage wall, b thermal network
20oC
42
2
OneDimensional SteadyState Heat Conduction
where (i) condensation of the moisture of the air, and (ii) freezing of the condensed moisture may take place if the air diffuses into the wood and cork. Also determine the heat flow rate across the wall. Solution Figure 2.19b shows the thermal network for A = 1 m2. The magnitudes of the resistances are (i) Polystyrene R1 ¼ dk ¼ 0:025 0:035 ¼ 0:7143 d 0:15 (ii) Cork R2 ¼ k ¼ 0:04 ¼ 3:75 (iii) Wood R3 ¼ dk ¼ 0:025 0:2 ¼ 0:125 1 (iv) Convective Ro ¼ h1o ¼ 20 ¼ 0:05 The heat flow equation for various layers can be written as t1 ð10Þ 0:7143 t2 t1 q¼ 3:75 t3 t2 q¼ 0:125
q¼
q¼
20 t3 0:05
ðiÞ ðiiÞ ðiiiÞ ðivÞ
and heat flow across the wall is q¼
20 ð10Þ ¼ 6:47 W/m2 0:7143 þ 3:75 þ 0:125 þ 0:05
Substituting the value of q in Eqs. (i)–(iv) gives t1 = –5.4°C, t2 = 18.86°C and t3 = 19.67°C. The analysis shows that the temperatures of the two sides of the cork layer are –5.4°C (inside) and 18.86°C (outside). Hence, the condensation of the moisture of the diffusing air and freezing of the condensate will take place in the cork layer. Let the planes with 0°C and 15°C lie at distances d1 and d2 from the inner side of the cork as shown in Fig. 2.20. We can write q¼
0 ð5:4Þ d1 =0:04
and q¼
15 0 d2 =0:04
Putting q = 6.47 W/m2, we get d1 = 33.4 mm (for 0°C) and d2 = 92.7 mm (for 15°C).
2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
43
Cork 5.4oC •
•15oC •18.86 oC q = 6.47 W/m2 o
•0 C
δ1 δ2
Fig. 2.20 Heat flow across cork layer
2.6
OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
@t For steadystate conduction, @s ¼ 0: In case of onedimensional radial heat flow, the derivatives with respect to h and z are zero. If there is no heat generation qg = 0 and Eq. (2.17b) reduces to
d2 t 1 dt ¼0 þ 2 dr r dr or r
d2 t dt ¼0 þ dr 2 dr
or d dt r ¼0 dr dr or r
dt ¼ C1 dr
or dt C1 ¼ dr r
ðiÞ
t ¼ C1 ln r þ C2
ðiiÞ
or
44
2
OneDimensional SteadyState Heat Conduction
The boundary conditions are at r ¼ r1 ;
t ¼ t1
ðiiiÞ
at r ¼ r2 ;
t ¼ t2
ðivÞ
t1 ¼ C1 ln r1 þ C2
ðvÞ
t2 ¼ C1 ln r2 þ C2
ðviÞ
and
This gives
Solution of Eqs. (v) and (vi) gives t1 t2 lnðr2 =r1 Þ t1 t2 ln r1 C2 ¼ t1 þ lnðr2 =r1 Þ C1 ¼
Substitution of values of the constants in Eq. (ii) gives the equation of temperature distribution as t¼
t1 t2 t1 t2 ln r þ t1 þ ln r1 lnðr2 =r1 Þ lnðr2 =r1 Þ
Rearrangement of the terms gives t t1 ln r ln r1 lnðr=r1 Þ ¼ ¼ lnðr2 =r1 Þ t2 t1 lnðr2 =r1 Þ
ð2:24Þ
This is equation of variation of the temperature in nondimensional form, which is an equation of logarithmic curve. It is to be kept in mind that the wall material has been assumed to be homogeneous and the thermal conductivity is constant. The conduction heat transfer from Fourier’s equation is q ¼ kA
dt dr
or q ¼ kð2pr2 LÞ
dt dr r¼r2
where 2pr2L is the heat transfer area at r = r2 and of the value of C1 gives derivative as
dt
C1 dr r¼r2 ¼ r2
; refer to Eq. (i). Substitution
2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
45
dt 1 t1 t2 ¼ dr r¼r2 r2 lnðr2 =r1 Þ Hence, 1 t1 t2 q ¼ kð2pr2 LÞ r2 lnðr2 =r1 Þ or q ¼ ð2pkLÞ
t1 t2 lnðr2 =r1 Þ
or q¼
t1 t2 1 2pkL lnðr2 =r1 Þ
¼
t1 t2 R
ð2:25Þ
where R¼
1 lnðr2 =r1 Þ 2pkL
ð2:26Þ
is the resistance to the heat conduction through a cylindrical wall. Alternative Method Alternatively, Eq. (2.25) can be derived by considering an elemental cylindrical shell of thickness dr at radius r, refer to Fig. 2.21. The heat flow, from Fourier’s equation, is q ¼ kA
dt dr
or dt q ¼ kð2prLÞ dr
r1
dr r
r2
Fig. 2.21 Heat flow across wall of a cylindrical shell
Cylinder length = L
46
2
OneDimensional SteadyState Heat Conduction
R = (1/2πkL) ln(r2/r1) t1
t2 q Δt
Fig. 2.22 Thermal network for a cylindrical shell
where dt/dr is the temperature gradient. The surface area at radius r is 2prL. Rearranging q
dr ¼ ð2pkLÞdt r
Integration between the limits r = r1, t = t1 and r = r2, t = t2 gives Zr2 q r1
dr ¼ ð2pkLÞ r
Zt2 dt t1
or q lnðr2 =r1 Þ ¼ 2pkLðt1 t2 Þ or q¼
t1 t2 1 2pkL lnðr2 =r1 Þ
¼
t1 t2 R
ð2:25Þ
The thermal network in this case can be presented as in Fig. 2.22, where the temperature 1 difference Δt is the driving force for the heat flow and R ¼ lnðr2 =r1 Þ is the thermal 2pkL resistance, which the wall offers to the flow of the heat by conduction. Logarithmic mean radius Equation (2.25) can be rewritten as q¼k
2pLðr2 r1 Þ ðt1 t2 Þ lnðr2 =r1 Þ ðr2 r1 Þ
ðiÞ
If the heat transfer equation is written in a simple plane wall format, then q ¼ kA0m
ðt1 t2 Þ ðr2 r1 Þ
Comparing Eqs. (i) and (ii), we obtain A0m ¼
2pLðr2 r1 Þ ð2pr2 L 2pr1 LÞ ðA2 A1 Þ ¼ ¼ lnðr2 =r1 Þ lnð2pr2 L=2pr1 LÞ lnðA2 =A1 Þ
ðiiÞ
2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
47
where A1 and A2 are the inner and outer surface areas of the cylindrical shell. Area A0m can be termed as logarithmic mean area. Putting A0m ¼ 2prm0 L; we have 2prm0 L ¼
2pLðr2 r1 Þ lnðr2 =r1 Þ
or rm0 ¼
r2 r1 lnðr2 =r1 Þ
ð2:26Þ
where rm0 can be termed as logarithmic mean radius. The arithmetic mean radius is rm ¼
r1 þ r2 2
and rm rm0 r2 r1 2 ¼1 r1 þ r2 rm lnðr2 =r1 Þ ðr2 =r1 1Þ 2 ¼1 lnðr2 =r1 Þ 1 þ r2 =r1 r r0
For r2/r1 = 2, mrm m ¼ 0:0382; i.e. for r2/r1 < 2, the arithmetic mean radius rm deviates from the logarithmic mean radius by less than 3.8% and can be used without signiﬁcant error for the calculation of the heat transfer through a cylindrical shell.
2.6.1 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell Considering Film Coefficients When the process of heat conduction is accompanied with convection heat transfer at the inner and outer surfaces of the cylinder (refer to Fig. 2.23), Eq. (2.25) is modiﬁed to include the ﬁlm resistances and is written in terms of inner and outer fluid temperatures ti and to, respectively. The convection heat transfer at the inner surface gives q ¼ hi Ai ðti t1 Þ ¼ hi ð2pr1 LÞðti t1 Þ or q¼
ti t1 ti t1 ti t1 ¼ ¼ 1=ðhi Ai Þ 1=½hi ð2pr1 LÞ Ri
where Ri = 1/hiAi is the ﬁlm resistance at the inner surface of the cylinder.
ðiÞ
48
2
dr
r1
OneDimensional SteadyState Heat Conduction
Cylinder length = L
r Fluid films
r2 Rk
Ri ti
t1
q
Ro t2
to
Fig. 2.23 Onedimensional steadystate heat conduction through a cylindrical shell considering ﬁlm coefﬁcients
Similarly, the convection heat transfer at the outer surface gives q ¼ ho Ao ðt2 to Þ ¼ ho ð2pr2 LÞðt2 to Þ or q¼
t2 to t2 to t2 to ¼ ¼ 1=ðho Ao Þ 1=½ho ð2pr2 LÞ Ro
ðiiÞ
where Ro = 1/hoAo is the ﬁlm resistance at the outer surface of the cylinder. The conduction heat transfer through the wall is q¼
t1 t2 t1 t2 ¼ ½1=ð2pkLÞ lnðr2 =r1 Þ Rk
ðiiiÞ
From Eqs. (i), (ii) and (iii), we get ti t1 ¼ qRi t1 t2 ¼ qRk t2 to ¼ qRo Combining the above equations, we obtain ti to ¼ qðRi þ Rk þ Ro Þ or q¼
ti to ti to ¼ P Ri þ Rk þ Ro R
ðivÞ
2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
49
or q¼
1 hi Ai
ti to 1 þ 2pkL ln rr21 þ
ð2:27Þ
1 ho Ao
i.e. the three resistances are in series. The thermal network is shown in Fig. 2.23.
2.6.1.1 Overall Heat Transfer Coefficient In heat exchangers, the equation of heat exchange between the fluids flowing in and outside the tube is written as q ¼ UAðti to Þ
ð2:28Þ
where U is termed as overall heat transfer coefﬁcient. Since the heat flow area (surface area) for a cylinder is different at the inner and outer surfaces, the equation of heat exchange must be written as q ¼ Ui Ai ðti to Þ
ðiÞ
q ¼ Uo Ao ðti to Þ
ðiiÞ
or
We derived the heat exchange equation as q¼
1 hi Ai
ti to 1 þ 2pkL ln rr21 þ
ð2:27Þ
1 ho Ao
Comparison with Eq. (i) gives 1 1 1 r2 1 1 ln þ þ Ai hi Ai 2pkL r1 ho Ao 1 Ai r2 Ai 1 ln þ ¼ þ hi 2pkL r1 ho Ao
Ui ¼
ð2:29aÞ
Since Ai = 2pr1L and Ao = 2pr2L, the alternative form of the equation of the overall heat transfer coefﬁcient is Ui ¼
1 Ai Ao Ai ln þ þ hi 2pkL Ai h o Ao
1
ð2:29bÞ
Similarly, we can write Uo ¼
Ao Ao Ao 1 ln þ þ ho hi Ai 2pkL Ai
1
ð2:30aÞ
where the overall heat transfer coefﬁcient has been referred to the outer surface area.
50
2
OneDimensional SteadyState Heat Conduction
Deﬁning hio = hiAi/Ao, which is the ﬁlm coefﬁcient at the inner surface referred to the outer surface, we obtain Uo ¼
1 Ao Ao 1 ln þ þ hio 2pkL Ai ho
1
ð2:30bÞ
Most of the heat exchangers employ metallic tubes and the tube wall resistance is negligible compared to the two ﬁlm resistances and is neglected. Then the equation of the overall heat transfer coefﬁcient takes the following form: Uo ¼
1 1 þ hio ho
1
¼
hio ho hio þ ho
ð2:31Þ
2.6.2 Composite Cylindrical Wall Pipes and tubes carrying high or lowtemperature fluids are generally insulated with one or more layers of insulating materials (termed as lagging of the pipes) to reduce heat loss or gain by the fluid while passing through the tube. The heat transfer in such cases can be determined as follows. Consider a composite cylinder wall comprising three layers as shown in Fig. 2.24. The radii and thermal conductivities of these layers are mentioned in the ﬁgure. In the steadystate condition, the quantity of heat passing through each layer is the same. The adjacent surfaces are assumed to be in perfect contact and hence are at the same temperature. The heat flow equation for the layers can be written as 2pk1 Lðt1 t2 Þ lnðr2 =r1 Þ 2pk2 Lðt2 t3 Þ q¼ lnðr3 =r2 Þ 2pk3 Lðt3 t4 Þ q¼ lnðr4 =r3 Þ
q¼
r4
r3 r2 r1 t1
t2 t3 t4
k1
Fig. 2.24 Composite cylindrical wall
k2
k3
ðiÞ
2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
51
From the above equations, lnðr2 =r1 Þ 2pk1 L lnðr3 =r2 Þ ðt2 t3 Þ ¼ q 2pk2 L lnðr4 =r3 Þ ðt3 t4 Þ ¼ q 2pk3 L
ðt1 t2 Þ ¼ q
ðiiÞ
Adding the equations, we obtain ðt1 t4 Þ ¼ q
lnðr2 =r1 Þ lnðr3 =r2 Þ lnðr4 =r3 Þ þ þ 2pk1 L 2pk2 L 2pk3 L
or q¼h
ðt1 t4 Þ lnðr2 =r1 Þ 2pk1 L
þ
lnðr3 =r2 Þ 2pk2 L
þ
lnðr4 =r3 Þ 2pk3 L
ðt1 t4 Þ R1 þ R2 þ R3
i¼
ð2:32Þ
The denominator is the sum of the resistances of the different layers. For n layers, the equation can be written as q¼
1 2pL
ðt1 tn þ 1 Þ ðt1 tn þ 1 Þ Pn lnðrj þ 1 =rj Þ ¼ Pn j¼1 Rj j¼1
ð2:33Þ
kj
Introducing the ﬁlm resistances, above equations give q¼h ¼
ðti to Þ 1 hi Ai
þ
lnðr2 =r1 Þ 2pk1 L
þ
lnðr3 =r2 Þ 2pk2 L
þ
lnðr4 =r3 Þ 2pk3 L
þ
1 ho Ao
i
ðti to Þ R i þ R 1 þ R2 þ R3 þ Ro
and q¼
1 hi Ai
þ
1 2pL
ðti to Þ Pn lnðrj þ 1 =rj Þ j¼1
kj
þ
1 ho Ao
ð2:34Þ
Example 2.12 Determine steadystate heat transfer rate through the wall of a 5mlong cylinder having inside and outside radius of 2.7 and 3 m, respectively. The inside and outside wall temperatures are 100°C and 30°C, respectively. The thermal conductivity of the wall material is 45 W/(m K). Also determine the error in estimate when the cylindrical wall is treated as a flat plate.
52
2
OneDimensional SteadyState Heat Conduction
Solution Heat flow through a cylindrical wall is given by q¼
2pkLðt1 t2 Þ lnðr2 =r1 Þ
or q¼
2p 45 5 ð100 30Þ ¼ 939;253 W lnð3=2:7Þ
Treating the cylindrical shell as a flat plate, r þ r 1 2 A ¼ 2prm L ¼ 2p L 2 2:7 þ 3 5 ¼ 89:53 m2 ¼ 2p 2 and q0 ¼ kA
t1 to 100 30 ¼ 940;065 W ¼ 45 89:53 3 2:7 r2 r1
The error in heat transfer estimate is q0 q ¼ 0:086% q which is very small. Comments: Large radii thin cylinders can be treated as flat plates. If the above problem refers to a cylinder with r1 = 10 mm and r2 = 20 mm, then q = 142,769 W and q′ = 148,440 W. The error, in this case, is 4%. The flat plate equation gives a higher value of the heat transfer rate. Example 2.13 A 15mmdiameter copper rod loses heat to the surrounding air with a surface heat transfer coefﬁcient of 10 W/(m2 K). If the rod is to be uniformly coated with a material of thermal conductivity 0.2 W/(m K) and the coated surface has a heat transfer coefﬁcient of 15 W/(m2 K), ﬁnd the thickness of the coating which will keep the state of heat loss unchanged. Solution For the rod without insulation, the heat is transferred by convection from its surface and is given by tw ta per m length of the rod q1 ¼ 1 2pr1 h1
where tw is the temperature of the surface of the rod and ta is the surrounding air temperature.
2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
53
For the rod with insulation, the heat flows from the wire surface through the insulation layer and is then transferred by convection from its surface. It is given by q2 ¼
tw ta 1 2pki
ln rr21 þ
1 2pr2 h2
per m length of the rod
where ki is the thermal conductivity of the insulating material. Equating the above equations for an equal rate of heat transfer in both conditions, we have t ta w ¼ 1 2pr1 h1
1 2pki
tw ta ln rr21 þ
1 2pr2 h2
or 1 1 r2 1 ¼ ln þ r1 h1 ki r1 r2 h2 Substituting the values of various terms, we get 1 1 r2 1 ¼ ln þ 0:0075 10 0:2 0:0075 r2 15 or 2:2262 ¼ ln r2 þ
0:0133 r2
Solution by trial and error gives r2 = 0.094 m = 94 mm. Example 2.14 Calculate the overall heat transfer coefﬁcient based on the inner diameter for a steel pipe covered with ﬁbre glass insulation. The following data are given: ID of pipe, di = 20 mm Thickness of pipe, ds = 2 mm Thickness of insulation, di = 20 mm Heat transfer coefﬁcient (inside), hi = 10 W/(m2 K) Heat transfer coefﬁcient (outside), ho = 5 W/(m2 K) Conductivity of insulation, ki = 0.05 W/(m K) Conductivity of steel, ks = 46 W/(m K) Inside fluid temperature, ti = 200°C Ambient temperature, to = 30°C Also, ﬁnd the heat loss from the pipe per m of its length.
54
2
OneDimensional SteadyState Heat Conduction
Solution Extending Eq. (2.29b) for the present case, the overall heat transfer coefﬁcient in this case is
1 Ai r2 Ai r3 Ai 1 ln þ ln þ þ hi 2pks L r1 2pki L r2 ho Ao 1 1 r1 r2 r1 r3 r1 1 ¼ þ ln þ ln þ h i ks r1 ki r2 r 3 ho 1 10 12 10 32 10 1 1 þ ln ln ¼ þ þ 10 1000 46 10 1000 0:05 12 32 5
Ui ¼
¼ 2:79 W=ðm2 C) The heat transfer rate is q ¼ Ui Ai ðti to Þ ¼ Ui ð2pr1 LÞ ðti to Þ ¼ 2:79 ð2pÞ ð10=1000Þ 1 ð200 30Þ ¼ 29:80 W per m length of the pipe Alternatively, Eq. (2.34) may be used to determine the heat transfer rate: q¼h
1 hi Ai
ðti to Þ þ
lnðr2 =r1 Þ 2pks L
þ
lnðr3 =r2 Þ 2pki L
þ
1 ho Ao
i
where Ai = 2pr1L and Ao = 2pr3L. Example 2.15 Air at 100°C flows through a thinwalled steel tube with inside diameter of 25 mm. The wall thickness is 0.5 mm. Thermal conductivity of the steel is 40 W/(m K). The tube is exposed to an environment at 20°C with heat transfer coefﬁcient of 10 W/(m2 K). The ﬁlm coefﬁcient at the inner surface is 50 W/(m2 K). Calculate (i) Overall heat transfer coefﬁcient, Ui, (ii) Heat loss per m length, (iii) What thickness of insulation having thermal conductivity of 0.035 W/(m K) must be applied to reduce the heat loss by 50%? Solution (i) From Eqs. (2.29a), (2.29b), the overall heat transfer coefﬁcient based on the inner surface of pipe can be written as Ui ¼
1 1 r1 r2 r1 1 þ ln þ hi ks r1 r 2 ho
2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
55
Substituting r1 = 12.5 10−3 m, r2 = 13 10−3 m, ks = 40 W/(m K), hi = 50 W/(m2 K), ho = 10 W/(m2 K), we get Ui ¼
1 12:5 13 12:5 1 1 þ ln ¼ 8:608 W= m2 K þ 50 1000 40 12:5 13 10
(ii) Heat loss per m length, q ¼ Ui Ai ðti to Þ ¼ Ui ð2pr1 LÞ ðti to Þ ¼ 8:608 ð2pÞ ð12:5=1000Þ 1 ð100 20Þ ¼ 54:1 W per m length of the pipe (iii) The heat loss after insulation is reduced to the half, i.e. q′ = 54.1/2 = 27.05 W. The required insulation thickness can be determined from Eq. (2.34): ðti to Þ lnðr2 =r1 Þ lnðr3 =r2 Þ 1 þ þ þ 2pks L 2pki L h o Ao
q0 ¼
1 hi Ai
or 27:05 ¼
ð100 20Þ 1 lnð0:013=0:0125Þ lnðr3 =0:013Þ 1 þ þ þ 50 2p 12:5 103 2p 40 2p 0:035 10 2p r3
By trial and error, r3 18 mm. Example 2.16 A steam pipe having an outside diameter of 20 mm is to be covered with two layers of insulation, each of thickness 10 mm. The average thermal conductivity of one material is 5 times that of the other. Assuming that the inner and outer surface temperatures of the composite insulation are ﬁxed, by how much will be the heat transfer be reduced when the better insulating material is placed next to the pipe surface than when it is placed away from the pipe? Solution From the given data, R1 = 10 mm, R2 = 20 mm, R3 = 30 mm and let k2 = 5k1. Heat transfer q1 with better insulating material next to the pipe from Eq. (2.33) is q1 ¼ h
2pLðt1 t3 Þ lnðr2 =r1 Þ k1
þ
i¼h
lnðr3 =r2 Þ k2
2pLðt1 t3 Þ
lnð20=10Þ k1
þ
lnð30=20Þ 5k1
i ¼ 1:2916 ½2pk1 Lðt1 t3 Þ
56
2
OneDimensional SteadyState Heat Conduction
Heat transfer q2 with better insulating material placed as outer layer is q2 ¼ h
2pLðt1 t3 Þ lnðr2 =r1 Þ k1
þ
2pLðt1 t3 Þ
i¼h
lnðr3 =r2 Þ k2
lnð20=10Þ 5k1
þ
lnð30=20Þ k1
i ¼ 1:8379 ½2pk1 Lðt1 t3 Þ
Hence, q2 1:8379 ½2pk1 Lðt1 t3 Þ ¼ 1:423 ¼ q1 1:2916 ½2pk1 Lðt1 t3 Þ The analysis shows that for the given surface temperatures, the better insulating material must be placed next to the pipe. Example 2.17 A conical cylinder is insulated along its tapered surface. The cylinder is of length L with one end of radius R1 and the other of radius R2. The surface with radius R1 is maintained at temperature T1 and the other at temperature T2. Find the expression of the heat flow along the length of the cylinder. Consider onedimensional heat flow along the axis of the cylinder only. Solution The heat flow through section xx at distance x is, refer to Fig. 2.25, q ¼ kAx
dT dx
ðiÞ
The radius at distance x is ðR2 R1 Þx L ¼ R1 þ Cx
R ¼ R1 þ
1Þ : where C ¼ ðR2 R L The area of the crosssection xx is
Ax ¼ pðR1 þ CxÞ2
Insulation R1
x
R
T1 •
q dx
A x
R2 B • T2
x L
Fig. 2.25 Heat conduction in axial direction of a conical cylinder
2.6 OneDimensional SteadyState Heat Conduction Through a Cylindrical Shell
57
Substitution of the value of Ax in Eq. (i) gives q ¼ kpðR1 þ CxÞ2
dT dx
ðiiÞ
Integration of Eq. (ii) gives ZL q 0
dx ðR1 þ CxÞ2
ZT2 ¼ kp
dT T1
Let (R1 + Cx) = a, then Cdx = da. The new boundary conditions are x ¼ 0;
a ¼ R1
x ¼ L;
a ¼ R1 þ CL ¼ R1 þ ½ðR2 R1 Þ=LL ¼ R2
Equation (iii) transforms to ZR2 R1
qda ¼ kp Ca2
ZT2 dT T1
or q 1 R2 ¼ kpðT1 T2 Þ C a R1 or q 1 1 ¼ kpðT1 T2 Þ C R1 R2 or
L q R2 R1
R2 R1 R 2 R1
¼ kpðT1 T2 Þ
or q ¼ kp
ðT1 T2 ÞR1 R2 : L
ðiiiÞ
58
2
2.7
OneDimensional SteadyState Heat Conduction
OneDimensional SteadyState Heat Conduction Through a Spherical Shell
Threedimensional general heat conduction equation in spherical coordinates is 2 1 @ 1 @ @t 1 @ t qg 1 @t 2 @t r sin h ¼ þ þ þ r 2 @r @r r 2 sin h @h @h a @s k r 2 sin2 h @/2 For steadystate condition, derivatives
@t @s
ð2:18Þ
¼ 0 and, for onedimensional radial heat flow, the
@ @2 and are also zero. Further if heat generation is not present, the equation @h @/2
reduces to
1 d 2 dt r ¼0 r 2 dr dr
or r2
dt ¼ C1 dr
or dt C1 ¼ dr r 2
ðiÞ
or t¼
C1 þ C2 r
ðiiÞ
The boundary conditions are t ¼ t1
at r ¼ r1
t ¼ t2
at r ¼ r2
and
Applying these boundary conditions, we get t1 ¼
C1 þ C2 r1
ðiiiÞ
t2 ¼
C1 þ C2 r2
ðivÞ
and
2.7 OneDimensional SteadyState Heat Conduction Through a Spherical Shell
59
From Eqs. (iii) and (iv), we have C1 ¼
t1 t2 1=r2 1=r1
and C2 ¼ t1 þ
t1 t2 1 r1 1=r2 1=r1
Substitution of the values of constants C1 and C2 in Eq. (ii) gives t t1 r1 r2 1 1 ¼ t2 t1 r2 r1 r1 r which is the temperature distribution equation in nondimensional form. It is an equation of hyperbola. Therefore, the temperature distribution across the wall of a spherical shell is hyperbolic. The heat transfer by conduction through the wall of the shell can be determined by substituting value of dt/dr from Eq. (i) in Fourier’s equation. Hence, q ¼ kA
dt C1 ¼ kA 2 dr r
Substituting value of C1 and A = 4pr2 for the spherical surface, we obtain
t1 t2 1 q ¼ kð4pr Þ 2 1=r2 1=r1 r 2
or
t1 t2 q ¼ 4pk 1=r1 1=r2
or q ¼ 4pk
r1 r2 ðt1 t2 Þ r2 r1
ð2:35aÞ
or q¼
ðt1 t2 Þ ðr2 r1 Þ 4pkr1 r2
¼
ðt1 t2 Þ R
ð2:35bÞ
where R¼
ðr2 r1 Þ 4pkr1 r2
ð2:35cÞ
is the resistance of the wall to the conduction heat transfer. The thermal network is shown in Fig. 2.26.
60
2
OneDimensional SteadyState Heat Conduction
Fig. 2.26 Onedimensional steadystate heat conduction through a spherical shell
Alternative method Equations (2.35a)–(2.35c), (2.35a) and (2.35b) can be obtained by considering an elementary shell of thickness dr at radius r, refer to Fig. 2.26. The surface area of the elemental shell is 4pr2. Hence, from Fourier’s equation, q ¼ kA
dt dt ¼ kð4pr 2 Þ dr dr
or q
dr ¼ 4pkdt r2
Integration between the limits, r = r1, t = t1, and r = r2, t = t2, gives Zr2 q r1
dr ¼ 4pk r2
Zt2 dt t1
or
1 1 q r1 r2
¼ 4pkðt2 t1 Þ
or ðt2 t1 Þ q ¼ 4pk 1 1 r1 r2
2.7 OneDimensional SteadyState Heat Conduction Through a Spherical Shell
61
or q ¼ 4pkðt1 t2 Þ
r1 r2 ðr2 r1 Þ
ð2:35aÞ
ðt1 t2 Þ R
ð2:35bÞ
or q¼
ðt1 t2 Þ ðr2 r1 Þ 4pkr1 r2
¼
which is the desired equation. Effective mean radius Equation (2.35a) can be rewritten as q ¼ kð4pr1 r2 Þ
ðt1 t2 Þ ðr2 r1 Þ
If the heat transfer equation is written in a simple plane wall format, then q ¼ kA0m
ðt1 t2 Þ ðr2 r1 Þ
Comparing the equations, we obtain A0m ¼ 4pr1 r2 ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ A 1 A2
ð2:36aÞ
where A1 and A2 are the inner and outer surface areas of the shell. Putting A0m ¼ 4prm2 , we have A0m ¼ 4prm2 ¼ 4pr1 r2 or rm ¼
pﬃﬃﬃﬃﬃﬃﬃﬃ r1 r2
Thus the mean effective radius is the geometric mean of the radii r1 and r2.
2.7.1 OneDimensional SteadyState Heat Conduction Through a Spherical Shell Considering the Film Coefficients Following the procedure adopted for the cylindrical shell, we have q ¼ hi Ai ðti t1 Þ ¼ hi ð4pr12 Þðti t1 Þ
ð2:36bÞ
62
2
OneDimensional SteadyState Heat Conduction
or q¼
ti t1 ti t1 ti t1 ¼ ¼ 2 1=ðhi Ai Þ 1=½hi ð4pr1 Þ Ri
ðiÞ
where Ri = 1/hiAi is the ﬁlm resistance at the inner surface of the shell. Similarly, the convection heat transfer at the outer surface gives q ¼ ho Ao ðt2 to Þ ¼ ho ð4pr22 Þðt2 to Þ or q¼
t2 to t2 to t2 to ¼ ¼ 1=ðho Ao Þ 1=½ho ð4pr22 Þ Ro
ðiiÞ
where Ro = 1/hoAo is the ﬁlm resistance at the inner surface of the shell. The conduction heat transfer through the wall is q¼
ðt1 t2 Þ ðr2 r1 Þ 4pkr1 r2
¼
ðt1 t2 Þ Rk
ðiiiÞ
From Eqs. (i), (ii) and (iii), we get ti t1 ¼ qRi t1 t2 ¼ qRk
ðivÞ
t2 to ¼ qRo Combining the above equations, we obtain ti to ¼ qðRi þ Rk þ Ro Þ or q¼
ti to ti to ¼ P Ri þ Rk þ Ro R
or q¼
1 hi Ai
þ
ti to ðr2 r1 Þ 1 4pkr1 r2 þ ho Ao
ð2:37Þ
i.e. the three resistances are in series. The thermal network is shown in Fig. 2.27.
2.7.2 Composite Spherical Shell Spherical shells containing high or lowtemperature fluids are generally insulated with one or more layers of insulating materials to reduce heat loss or gain by the fluid. The heat transfer in such cases can be determined as follows.
2.7 OneDimensional SteadyState Heat Conduction Through a Spherical Shell
r1
dr r Fluid films
r2
Ri
Rk
ti
63
t1
Ro to
t2
q
Fig. 2.27 Onedimensional steadystate heat conduction through a spherical shell considering the ﬁlm coefﬁcients
r3 r2
r4
r1 k1 k2 k3
t1
t2
Temperature distribution
t3 t4
a ti
Ri
R2
R1 t1
t2
Ro
R3 t3
to
t4
b Fig. 2.28 a A composite shell, b thermal network
Consider a spherical shell with the wall comprising three layers as shown in Fig. 2.28a. The radii and thermal conductivities of these layers are mentioned in the ﬁgure. In the steadystate condition, the quantity of heat passing through each layer is the same. The adjacent surfaces are assumed to be in perfect contact and hence are at the same temperature. The heat flow equation for the layers can be written as 4pk1 r1 r2 ðt1 t2 Þ r2 r1 4pk2 r2 r3 ðt2 t3 Þ q¼ r3 r2 4pk3 r3 r4 ðt3 t4 Þ q¼ r4 r3
q¼
64
2
OneDimensional SteadyState Heat Conduction
From the above equations, r2 r1 4pk1 r1 r2 r3 r2 ðt2 t3 Þ ¼ q 4pk2 r2 r3 r4 r3 ðt3 t4 Þ ¼ q 4pk3 r3 r4
ðt1 t2 Þ ¼ q
Adding the equations, we obtain ðt1 t4 Þ ¼ q
r2 r1 r3 r2 r4 r3 þ þ 4pk1 r1 r2 4pk2 r2 r3 4pk3 r3 r4
or q¼
r2 r1 4pk1 r1 r2
ðt1 t4 Þ r3 r2 þ 4pk þ 2 r2 r3
ð2:38aÞ
r4 r3 4pk3 r3 r4
or q¼
ðt1 t4 Þ R 1 þ R 2 þ R3
The denominator is the sum of the resistances of the different layers. For n layers, the equation can be written as q¼
ðt1 tn þ 1 Þ ðt1 tn þ 1 Þ Pn rj þ 1 rj ¼ Pn j¼1 Rj j¼1 kj rj rj þ 1
ð2:39Þ
1 4p
Introducing the ﬁlm resistances, q¼h
ðti to Þ 1 hi Ai
r2 r1 4pk1 r1 r2
þ ¼
þ
r3 r2 4pk2 r2 r3
þ
r4 r3 4pk3 r3 r4
þ
1 ho Ao
i
ð2:40aÞ
ðti to Þ Ri þ R1 þ R 2 þ R 3 þ Ro
The thermal network is shown in Fig. 2.28b. The equation can be written in a general form as q¼
1 hi Ai
þ
1 4p
ðti to Þ Pn rj þ 1 rj j¼1 kj rj rj þ 1
þ
1 ho Ao
ð2:40bÞ
Example 2.18 A hemispherical electric heated oven of 1.0 m internal diameter is made of 225 mm thick layer of ﬁre bricks. On the outer surface, a magnesia 85 layer of 100 mm thickness is applied for insulation. The temperature at the inner surface of the oven is 800°C. If the ambient temperature is 20°C, and outside convection ﬁlm coefﬁcient is 10 W/(m2 K),
2.7 OneDimensional SteadyState Heat Conduction Through a Spherical Shell
65
calculate the heat loss from the oven through the hemispherical shell with and without insulation. Also ﬁnd out the temperature of the outer surface of the ﬁre bricks. The thermal conductivities of ﬁre bricks and magnesia 85 are 0.3 W/(m K) and 0.05 W/(m K), respectively. Solution (i) Heat loss Heat loss through a spherical shell with composite wall is given by q¼
r2 r1 4pk1 r1 r2
ðt1 to Þ r3 r2 þ 4pk þ 2 r2 r3
1 4pr32 ho
¼
ðt1 to Þ R1 þ R2 þ Ro
Here r1 = 500 mm, r2 = 500 + 225 = 725 mm, r3 = 725 + 100 = 825 mm, t1 = 800°C, to = 20°C, k1 = 0.3 W/(m K), k2 = 0.05 W/(m K) and ho = 10 W/(m2 K). Substitution of the values of various parameters gives r2 r1 ð725 500Þ 1000 ¼ 0:1646 ¼ 4pk1 r1 r2 4p 0:3 500 725 r3 r2 ð825 725Þ 1000 ¼ 0:266 R2 ¼ ¼ 4pk2 r2 r3 4p 0:05 725 825
R1 ¼
Ro ¼
1 ð1000Þ2 ¼ ¼ 0:0117 3 4pr3 ho 4p ð825Þ2 10
The heat loss from the hemispherical shell will be half of the spherical shell, i.e. 1 t1 to 1 800 20 q¼ ¼ 881:75 W ¼ 2 R1 þ R2 þ Ro 2 0:1646 þ 0:266 þ 0:0117 Heat loss without magnesia85 layer will be q¼
1 t1 to 1 800 20 ¼ 2212:1 W ¼ 2 R1 þ Ro 2 0:1646 þ 0:0117
(ii) Temperature of the outer surface of the ﬁrebrick layer Heat flow through the brick layer equals the heat loss. Hence, (i) with magenia85 layer: 881:75 2 ¼
t1 tb R1
¼
or tb ¼ 509:73 C
800 tb 0:1646
66
2
OneDimensional SteadyState Heat Conduction
(ii) without magenia85 layer: 2212:1 2 ¼
t1 tb R1
¼
800 tb 0:1646
or tb ¼ 71:78 C:
2.8
Measurement of Thermal Conductivity
2.8.1 Thermal Conductivity Measurement of Solids Experimental determination of thermal conductivity of a material, based on Eq. (1.3), involves measurement of heat flow rate through the material and the temperature difference across the material specimen under steadystate condition. Various methods used are known as the plate, tube and sphere methods. The plate method is based on onedimensional heat conduction through a plane wall. Figure 2.29 shows the schematic layout of a guarded hot plate apparatus for measurement of thermal conductivity of solids. The material, whose thermal conductivity is to be determined, is in the form of a circular disc of area A and thickness d. Heat is supplied to the lower side of the disc by a flat electric heater M. Heat is removed from the top of the disc by arranging a cooler above the disc. Compensation or guard heater G surrounds the main heater and is installed to prevent escape of heat of the main heater from the edge. Additional compensation heater C1 is provided below the back insulation, under the main heater M, to compensate the loss of heat in the downward direction. The installation of guard heater G and compensation heater C1 ensures onedimensional heat flow through the disc, i.e. from the lower heated face to upper cooled face when there is no temperature difference across the side and back insulations (e.g. t3 = t4). The whole assembly is further adequately insulated to prevent loss of heat to the surrounding. When steady state is reached, the electric power input (q) to the main heater and temperatures of the hot and cold faces of the disc t1 and t2, respectively, are measured with the help of thermocouples installed there. These measurements must be made only when there is no temperature difference across the side and back insulations (e.g. t3 = t4), as explained earlier. In this situation, all the heat input to the main heater passes through the specimen
Cooled surface M
Specimen disc
t2
Side insulation G
t1
G t4
C1
t3
Insulation t3 = t4
Fig. 2.29 Schematic layout of experimental setup to determine thermal conductivity by plate method
2.8 Measurement of Thermal Conductivity
67
disc. Knowing q, t1 and t2, the thermal conductivity is determined from the following relation: k¼
qd qd ¼ Aðt1 t2 Þ pr 2 ðt1 t2 Þ
ð2:41Þ
The equipment, described above, is widely used to determine the thermal conductivity of nonmetals, i.e. solid of low thermal conductivity. In the case of high thermal conductivity such as metals, the temperature difference (t1 – t2) is small and their precise measurement will be required. Figure 2.30 shows a simple instrument which has been used to measure the thermal conductivity of metals. Here, the rod 1 whose thermal conductivity is to be measured is joined axially with another metal rod 2 whose thermal conductivity is known. Thermocouples are afﬁxed axially to both the rods. The rods are properly insulated from outside to minimize heat loss and ensure onedimensional heat flow in the axial direction only. Knowing the temperature gradients through the two rods, the thermal conductivity of the rod 1 can be determined using the following relation:
Dt q ¼ k1 A Dx
Dt ¼ k2 A Dx 1
2
or k2 ¼ k1
Dt Dx
Dt = Dx 1 2
ð2:42Þ
The tube method of determining thermal conductivity of insulating materials is based on law of heat conduction through a cylindrical wall, i.e. k¼
q lnðr2 =r1 Þ 2pLðt1 t2 Þ
ð2:43Þ
The instrument, shown schematically in Fig. 2.31a, consists of a central electric heater which is surrounded by a copper tube of external diameter d1. The material whose thermal
Heat source q
Rod 1 Insulation Thermocouple
q Rod 2
Heat sink
Fig. 2.30 Schematic layout of experimental setup to determine thermal conductivity by rod method
68
2
OneDimensional SteadyState Heat Conduction Side insulation
Heater
a
Heater
b
Copper tubes
Thermocouple
Fig. 2.31 Schematic layout of experimental setup to determine thermal conductivity by tube method
conductivity is to be determined envelops the copper tube. The temperatures t1 and t2 of the outer surface of the copper tube and that of the outer surface of the material, respectively, are measured by afﬁxing thermocouples as shown in the ﬁgure. Temperatures t1 and t2 used in Eq. (2.43) are the average temperatures of these surfaces. Knowing the heat input (i.e. the power supplied to the electric heater), dimensions r1, r2, and length L of the tube, the thermal conductivity is calculated from the relation given above. In the case of materials in powder form, another copper tube of inside radius r2 is used. The powder is ﬁlled in the space between these tubes, see Fig. 2.31b. Temperature t2, in this case, is the temperature of the inner surface of the outer copper tube. However, the thermal conductivity of the powder insulating material is strong function of the density of the material packed into the space between tubes. Loosely ﬁlled material shows a lower thermal conductivity. In the tubetype instruments described above, the length of the tube must be large (say 2–3 m) so that the end loss is negligible and heat flows only in radial direction as desired. In the sphere method, the instrument consists of two spheres of diameter d1 and d2, where d2 > d1, refer to Fig. 2.32. The test material ﬁlls the space between the spheres. The inner sphere houses an electric heater. Measuring the heat input (power to the heater), temperatures t1 and t2 in the steadystate condition and knowing the sphere diameters, the thermal conductivity of the ﬁlled material is determined from the following relation: k¼
q 1 1 4pðt1 t2 Þ r1 r2
ð2:44Þ
Sphere dia. d2
Sphere dia. d1
Test material
t2 t1 Electric heater
Fig. 2.32 Schematic layout of experimental setup to determine thermal conductivity by sphere method
2.8 Measurement of Thermal Conductivity
69
2.8.2 Thermal Conductivity Measurements of Liquids and Gases To determine thermal conductivity of liquids and gases, it is an essential requirement that the convection currents are suppressed. Figure 2.33a depicts one of the methods, wherein the heat from the heated plate flows downwards through a very thin layer of liquid d (of the order of 0.5 mm) to the cold plate at the bottom. The liquid layer is very thin and is heated from above; hence, the convection currents are not setup. An alternative method of concentric cylinders (Fig. 2.33b) may be used wherein the heat flows in radial direction through a very thin layer of liquid between two concentric cylinders. This arrangement has also been used for gases. The thermal conductivity is calculated from Eq. (2.43). Example 2.19 In an instrument used to determine the thermal conductivity, the specimen to be tested is placed between the hot and cold surfaces. The specimen is a circular disc of 150 mm diameter and 15 mm thickness. The temperatures of the hot and cold surfaces of the instrument are recorded to be 170°C and 20°C, respectively. Under steadystate condition, the heat flow through the specimen is 105 W. Calculate the thermal conductivity of the material of the specimen if due to the presence of the guard heaters there is no heat flow in the radial direction. Solution k¼ ¼
qd qd ¼ 2 Aðt1 t2 Þ pr ðt1 t2 Þ 105 0:015 p ð0:075Þ2 ð170 20Þ
¼ 0:594 W=ðm KÞ:
Example 2.20 A hollow aluminium sphere, with an electric heater at its centre, is used to determine the thermal conductivity of an insulating material. The inner and outer radii of the sphere are 0.1 and 0.125 m, respectively. In a particular test, a spherical shell of an insulating material of 100 mm thickness was cast on the outer surface of the sphere. In the steady state,
Insulation Hot plate
Fluid space
Heater Fluid layer
Cylinders
Cold plate
a
b Thermocouples
Fig. 2.33 Schematic layout of experimental setups to determine thermal conductivity of liquid or gases
70
2
OneDimensional SteadyState Heat Conduction
the temperature at the outer surface of the aluminium wall was recorded as 200°C. The system is exposed to room air at 30°C, and the convective heat transfer coefﬁcient is estimated to be 20 W/(m2 K). If the energy supplied to the electric heater is at the rate of 100 W, what is the thermal conductivity of the insulation? Neglect heat loss by radiation. Solution From the convective heat transfer equation, q ¼ hAðt2 ta Þ where t2 is the temperature at the outer surface of the shell and ta is the temperature of the surrounding air. The above equation gives 100 ¼ 20 4p ð0:225Þ2 ðt2 30Þ The solution gives t2 ¼ 37:86 C The heat flow rate by conduction through the wall of the spherical shell is q¼
4pkr1 r2 ðt1 t2 Þ r2 r1
Substitution of values of various terms gives 100 ¼
4p k 0:125 0:225 ð200 37:86Þ 0:225 0:125
Solution gives k = 0.1745 W/(m K).
2.9
Effect of Variable Thermal Conductivity
Thermal conductivity of a dry solid substance depends on temperature and when the temperature difference across the wall of such substance is large, proper account of the variation of thermal conductivity with the temperature must be taken. For most of the materials, this dependence is linear if the difference between boundary temperatures of the substance is not large. Mathematically, it can be expressed as k ¼ ko ð1 þ btÞ
ð2:45Þ
where ko is the thermal conductivity at 0°C, and b is a constant which is negative for metals and positive for insulating materials. The constant b is usually determined experimentally. When variation of thermal conductivity with temperature is considered linear, the heat flow and temperature distribution through the plane, cylindrical and spherical walls can be determined as under.
2.9 Effect of Variable Thermal Conductivity
71
2.9.1 Plane Wall Fourier’s equation for conductive heat transfer is q ¼ kA
dt dx
Introducing value of k from Eq. (2.45), q ¼ ko ð1 þ btÞA
dt dx
or qdx ¼ ko Að1 þ btÞdt
ðiÞ
Integrating between the limits of boundary conditions of the plane wall, refer to Fig. 2.34, we get Zd
Zt2 dx ¼ Ako
q 0
ð1 þ btÞ dt t1
or t t2 2 qd ¼ Ako t þ b 2 t1 or q¼
Ako b 2 t2 t12 ðt2 t1 Þ þ 2 d
q¼
Ako b ðt2 t1 Þ 1 þ ðt2 þ t1 Þ 2 d
ðiiÞ
or
t
δ q
β (+ve)
t1 •
β=0 • t2
β (ve)
x x=0
Fig. 2.34 Temperature distribution across a plane wall with temperaturedependent thermal conductivity
72
2
OneDimensional SteadyState Heat Conduction
Hence, heat flow rate is ðt2 t1 Þ ðt2 þ t1 Þ 1þb d 2 ðt2 t1 Þ ¼ Ako ð1 þ btm Þ d ðt1 t2 Þ ¼ Akm d
q ¼ Ako
ð2:46Þ
where km is the thermal conductivity at mean temperature tm ¼ ðt1 þ2 t2 Þ of the wall. The equation establishes that in the case of linear variation of thermal conductivity with temperature, a constant value of the thermal conductivity km corresponding to the mean temperature may be used. However, the variation of the thermal conductivity with temperature is not linear in all cases and in such cases the above simpliﬁcation is not applicable (see Example 2.25). For the expression of the temperature distribution, integrate Eq. (i) between the limits 0 to x and t1 to t, which gives Zx
Zt dx ¼ ko A
q 0
ð1 þ btÞdt t1
or t t2 qx ¼ Ako t þ b 2 t1 or Ako b 2 2 t t1 q¼ ðt t1 Þ þ 2 x Substitution of value of q from Eq. (ii) gives Ako b 2 Ako b 2 2 2 t t1 ¼ t t1 ðt2 t1 Þ þ ðt t1 Þ þ 2 2 2 d x Simplifying and rearranging the terms, we get t¼
h i x o1=2 1 1n ð1 þ bt1 Þ2 ð1 þ bt1 Þ2 ð1 þ bt2 Þ2 b d b
ð2:47Þ
The equation gives the temperature distribution through the wall. In general, for the positive values of b, the curve is convex and the curve is concave for the negative values of b, see Fig. 2.34. For b = 0, i.e. when k is independent of the temperature, the distribution is linear.
2.9 Effect of Variable Thermal Conductivity
73
2.9.2 Cylindrical Shell Fourier’s equation for conductive heat transfer, in this case, is q ¼ kA
dt dt ¼ kð2prLÞ dr dr
Introducing value of k from Eq. (2.45), q ¼ ko ð1 þ btÞð2prLÞ
dt dr
or q
dr ¼ 2pLko ð1 þ btÞdt r
ðiÞ
Integrating between the limits of boundary conditions of the cylindrical wall, we get Zr2 q r1
dr ¼ 2pLko r
Zt2 ð1 þ btÞdt t1
or t t2 2 q lnðr2 =r1 Þ ¼ 2pLko t þ b 2 t1 or b 2 t2 t12 q lnðr2 =r1 Þ ¼ 2pLko ðt2 t1 Þ þ 2
ðiiÞ
or b ðt2 t1 Þ q ¼ 2pLko 1 þ ðt2 þ t1 Þ 2 lnðr2 =r1 Þ or q ¼ 2pLko ½1 þ btm
ðt2 t1 Þ lnðr2 =r1 Þ
or q ¼ 2pLkm
ðt2 t1 Þ lnðr2 =r1 Þ
ð2:48Þ
74
2
OneDimensional SteadyState Heat Conduction
ðt 1 þ t 2 Þ of the wall, i.e. for 2 the linear variation of thermal conductivity with temperature, a constant value of the thermal conductivity km corresponding to the mean temperature may be used.
where km is the thermal conductivity at mean temperature tm ¼
2.9.3 Spherical Shell Fourier’s equation for conductive heat transfer, in this case, is q ¼ kA
dt dt ¼ kð4pr 2 Þ dr dr
Introducing value of k from Eq. (2.45), q ¼ ko ð1 þ btÞð4pr 2 Þ
dt dr
or q
dr ¼ 4pko ð1 þ btÞdt r2
ðiÞ
Integrating between the limits of boundary conditions of the spherical wall, we get Zr2 q r1
dr ¼ 4pko r2
Zt2 ð1 þ btÞdt t1
or r2 t 1 t2 2 q ¼ 4pko t þ b r r1 2 t1 or q
1 1 r1 r2
b 2 t2 t12 ¼ 4pko ðt2 t1 Þ þ 2
or q
r2 r1 r1 r2
b ¼ 4pko 1 þ ðt2 þ t1 Þ ðt2 t1 Þ 2
or q ¼ 4pko ½1 þ btm
r1 r2 ðt2 t1 Þ r2 r1
ðiiÞ
2.9 Effect of Variable Thermal Conductivity
75
or q ¼ 4pkm
r1 r2 ðt1 t2 Þ ; r2 r1
ð2:49Þ
ðt 1 þ t 2 Þ of the wall. In this 2 case also, for the linear variation of thermal conductivity with temperature, a constant value of the thermal conductivity km corresponding to the mean temperature may be used. where km is the thermal conductivity at mean temperature tm ¼
Example 2.21 Derive the equation of temperature distribution through the wall of a cylindrical shell if the thermal conductivity of the wall material is a linear function of temperature. Solution Fourier’s equation for conductive heat transfer, in this case, is q ¼ kA
dt dt ¼ ko ð1 þ btÞð2prLÞ dr dr
or q
dr ¼ 2pLko ð1 þ btÞdt r
Integrating between the limits of boundary conditions of the cylindrical wall, we get
b 2 t t12 q lnðr2 =r1 Þ ¼ 2pLko ðt2 t1 Þ þ 2 2
ðiÞ
Integrating between the limits of t1 at r1 and t at r
b 2 t t12 q lnðr=r1 Þ ¼ 2pLko ðt t1 Þ þ 2
From Eqs. (i) and (ii), b lnðr2 =r1 Þ ðt2 t1 Þ þ 2 t22 t12 ¼ lnðr=r1 Þ ðt t1 Þ þ b2 t2 t12 or ðt t1 Þ þ n Putting
lnðr=r1 Þ lnðr2 =r1 Þ
lnðr=r1 Þ b 2 b 2 t t12 t2 t12 ¼ 0 ðt 2 t 1 Þ þ 2 lnðr2 =r1 Þ 2
h i o ðt2 t1 Þ þ b2 t22 t12 ¼ C in the above equation, we get
ðiiÞ
76
2
ðt t1 Þ þ
OneDimensional SteadyState Heat Conduction
b 2 t t12 C ¼ 0 2
or t2 þ
2 2 2 t t12 t1 C ¼ 0 b b b
or t2 þ
2 t C1 ¼ 0 b
where C1 ¼ t12 þ b2 t1 þ b2 C: Solution of Eq. (iii) gives 8 " #1=2 9 = 1< 2 2 2 þ 4C1 t¼ ; 2: b b or 1 t¼ b
" #1=2 1 2 þ C1 b
Substituting value of C1, we have 1 t¼ b
" #1=2 1 2 2 2 2 þ t1 þ t1 þ C b b b
The equation of C may be rearranged as follows: C¼
lnðr=r1 Þ b 2 2 t t1 ðt2 t1 Þ þ lnðr2 =r1 Þ 2 2
or " # lnðr=r1 Þ ð1 þ bt1 Þ2 ð1 þ bt2 Þ2 C¼ lnðr2 =r1 Þ 2b Substituting value of C, we get 1 t¼ b
" " ##1=2 1 2 2 2 2 lnðr=r1 Þ ð1 þ bt1 Þ2 ð1 þ bt2 Þ2 þ t1 þ t1 b b b lnðr2 =r1 Þ 2b
ðiiiÞ
2.9 Effect of Variable Thermal Conductivity
77
or i1=2 1 1 lnðr=r1 Þ h ð1 þ bt1 Þ2 ð1 þ bt2 Þ2 t ¼ 1 þ b2 t12 þ 2bt1 b b lnðr2 =r1 Þ or i1=2 1 1 lnðr=r1 Þ h 2 2 2 ð1 þ bt1 Þ ð1 þ bt2 Þ t ¼ ð1 þ bt1 Þ b b lnðr2 =r1 Þ At r = r2, t = t2. The condition is satisﬁed when positive sign of square root is used. This gives the required temperature distribution equation as i1=2 1 1 lnðr=r1 Þ h 2 2 2 t¼ þ ð1 þ bt1 Þ ð1 þ bt1 Þ ð1 þ bt2 Þ b b lnðr2 =r1 Þ
Example 2.22 Following the procedure of the preceding example, derive the equation of temperature distribution through the wall of a spherical shell if the thermal conductivity of the wall material is a linear function of temperature. Solution Fourier’s equation for conductive heat transfer, in this case, is q ¼ kA
dt dt ¼ kð4pr 2 Þ dr dr
Introducing value of k from Eq. (2.45), q ¼ ko ð1 þ btÞð4pr 2 Þ
dt dr
or q
dr ¼ 4pko ð1 þ btÞdt r2
Integrating between the limits of boundary of t1 at r1 and t2 at r2, we get Zr2 q r1
dr ¼ 4pko r2
Zt2 ð1 þ btÞdt t1
or r2 t 1 t2 2 q ¼ 4pko t þ b r r1 2 t1
78
2
OneDimensional SteadyState Heat Conduction
or
b 2 t2 t12 ¼ 4pko ðt2 t1 Þ þ 2
b ¼ 4pko 1 þ ðt2 þ t1 Þ ðt2 t1 Þ 2
1 1 r1 r2
q or q
r2 r1 r1 r2
ðiÞ
Similarly integrating between the limits of boundary of t1 at r1 and t at r, we get r r1 b q ¼ 4pko 1 þ ðt þ t1 Þ ðt t1 Þ 2 r1 r
ðiiÞ
From Eqs. (i) and (ii), h i b 4pk 1 þ ð t þ t Þ ðt t1 Þ o 1 2 r r1 r1 r2 h i ¼ r1 r r2 r1 4pko 1 þ b2 ðt2 þ t1 Þ ðt2 t1 Þ h i ðt t1 Þ þ b2 t2 t12 r r1 r2 ¼h i r2 r1 r ðt2 t1 Þ þ b t2 t2 2
2
1
or
b 2 r2 r r1 b 2 2 2 t t1 t t1 ¼ 0 ðt t1 Þ þ ðt2 t1 Þ þ 2 2 2 r r2 r1
or 2 2 2 r2 r r1 b 2 2 2 t t1 ¼ 0 t þ t t1 t1 ðt2 t1 Þ þ b b b r 2 2 r2 r1 2
or t2 þ
r r r ð1 þ bt Þ2 ð1 þ bt Þ2 2 2 2 1 1 2 t t12 t1 þ ¼0 b b r r2 r1 b2
or t2 þ where t12 b2 t1 þ
2 tþC ¼ 0 b
r rr ð1 þ bt1 Þ2 ð1 þ bt2 Þ2 2
r
1
r2 r1
b2
¼ C in the above equation.
ðiiiÞ
2.9 Effect of Variable Thermal Conductivity
79
Solution of Eq. (iii) gives 8 " #1=2 9 = 1< 2 2 2 t¼ 4C ; 2: b b or 1 t¼ b
" #1=2 1 2 C b
Substituting value of C, we have 1 t¼ b
" # r r r ð1 þ bt Þ2 ð1 þ bt Þ2 1=2 1 2 2 2 2 1 1 2 þ t1 þ t1 b b r r2 r1 b2
or i 1=2 r r r h 1 1 2 1 1 þ b2 t12 þ 2bt1 ð1 þ bt1 Þ2 ð1 þ bt2 Þ2 t¼ b b r r2 r1 or i1=2 r r r h 1 1 2 1 2 2 2 ð1 þ bt1 Þ ð1 þ bt2 Þ t ¼ ð1 þ bt1 Þ b b r r2 r1 At r = r2, t = t2. The condition is satisﬁed when positive sign of square root is used. This gives the required temperature distribution equation as i1=2 r r r h 1 1 2 1 2 2 2 t¼ þ ð1 þ bt1 Þ ð1 þ bt1 Þ ð1 þ bt2 Þ b b r r2 r1
Example 2.23 The thermal conductivity of the material of a plane wall varies linearly with temperature as k = ko(1 + bt). Verify that for b > 0, the temperature proﬁle is concave downwards as shown in Fig. 2.34. Solution Fourier’s equation for a plane wall is q ¼ ko ð1 þ btÞA
dt dx
80
2
OneDimensional SteadyState Heat Conduction
or q dt ¼ ð1 þ btÞ ko A dx or
q dt dt þ bt ¼ ko A dx dx
Differentiating with respect to x, d2 t 0 ¼ 2 þb dx
" # dt 2 d2 t þt 2 dx dx
or 2 d2 t b dt ¼ \0 dx2 1 þ bt dx The second derivative is negative; hence, the temperature proﬁle is concave downwards. Example 2.24 The surface temperatures of a 200mmthick ﬁreclay wall are 1000°C and 200°C. The thermal conductivity of the ﬁreclay is a linear function of the temperature given by the following equation: k ¼ 0:8ð1 þ 0:0007tÞ W=ðm CÞ where t is in °C. Calculate the rate of heat flow, in W/m2, through the wall and determine the temperature distribution in the wall. Solution (i) For linear variation of the thermal conductivity, the mean value of the thermal conductivity may be used, which is h t þ t i 1 2 km ¼ 0:8ð1 þ 0:0007tm Þ ¼ 0:8 1 þ 0:0007 ¼ 1:136 W=ðm CÞ 2 Heat transfer q ¼ km A
t1 t2 1000 200 ¼ 4544 W ¼ 1:136 1 0:2 d
2.9 Effect of Variable Thermal Conductivity
81
1200
Temperature, oC
1000 800 600 400 200 0 0
50
100
150
200
x, mm Fig. 2.35 Temperature distribution across the wall of Example 2.24
(ii) Temperature distribution t¼
h i xo1=2 1 1n ð1 þ bt1 Þ2 ð1 þ bt1 Þ2 ð1 þ bt2 Þ2 b d b
Here b = 0.0007. Hence, ð1 þ bt1 Þ2 ¼ ð1 þ 0:0007 1000Þ2 ¼ 2:89 ð1 þ bt2 Þ2 ¼ ð1 þ 0:0007 200Þ2 ¼ 1:2996 Hence, t ¼ 1428:6 ð2:89 7:952xÞ1=2 1428:6 Figure 2.35 shows the temperature distribution. Example 2.25 The thermal conductivity of a plane wall varies as k ¼ ko ð1 þ Bt þ Ct2 Þ where ko is the thermal conductivity at temperature t = 0 and B and C are constants. If the wall thickness is d and the surface temperatures are t1 and t2, show that the steady heat flux q″ through the wall is given by q00 ¼ ko
ðt2 t1 Þ B C 1 þ ðt2 þ t1 Þ þ ðt22 þ t1 t2 þ t12 Þ d 2 3
82
2
OneDimensional SteadyState Heat Conduction
Solution From Fourier’s equation, the heat flow is given by Zd
Zt2 dx ¼ A
q 0
kdt t1
or Zt2 qd ¼ A
ko ð1 þ Bt þ Ct2 Þ dt
t1
or t t2 t3 2 ¼ Ako t þ B þ C 2 3 t1 or C 3 B 2 t2 t12 þ t2 t13 ¼ Ako ðt2 t1 Þ þ 2 3 or B C qd ¼ Ako ðt2 t1 Þ 1 þ ðt2 þ t1 Þ þ ðt22 þ t1 t2 þ t12 Þ 2 3 Hence, heat flux is q00 ¼
q ðt2 t1 Þ B C ¼ ko 1 þ ðt2 þ t1 Þ þ ðt22 þ t1 t2 þ t12 Þ A d 2 3
Example 2.26 For ko = 50, B = –0.0005 and C = –1 10−6, show that for small temperatures t1 = 200°C and t2 = 100°C, the mean value of thermal conductivity can be used. Solution (i) The effective thermal conductivity, from the previous example, is B C 2 2 k ¼ ko 1 þ ðt2 þ t1 Þ þ ðt2 þ t1 t2 þ t1 Þ 2 3 Substitution gives k = 45.08 W/(m K).
2.9 Effect of Variable Thermal Conductivity
83
(ii) From equation, k = ko(1 + Bt + Ct2) = 47.0 W/(m K) at t1 = 100°C ¼ 43:0 W=ðm KÞ at t1 ¼ 200 C Hence, mean value km = (k1 + k2)/2 = 45.0 W/(m K). (iii) The mean temperature, tm = (t1 + t2)/2 = 150°C At 150 C; k ¼ 50 ð1 0:0005 150 1 106 1502 Þ ¼ 45:12 The analysis shows that if the effect of the temperature is not signiﬁcant and temperature difference is not large, value of k corresponding to the mean temperature can be used. Example 2.27 Show that in case of a spherical shell, the heat conducted with variable thermal conductivity k = ko(1 + bT) is higher by 50b(T1 + T2) per cent over the shell with constant value of thermal conductivity ko, and T1 and T2 are wall temperatures. Solution With the constant value of the thermal conductivity, q¼
4pko r1 r2 ðT1 T2 Þ r2 r1
The heat conducted with the variable thermal conductivity is q0 ¼
4pko 1 þ b T1 þ2 T2 r1 r2 ðT1 T2 Þ r2 r1
Hence, the percentage increase in the heat transfer is Dq ¼
q0 q 100 q
Substitution of the values of q and q′ gives Dq ¼ 50bðT1 þ T2 Þ percent It can be shown, using appropriate equations, that the above result is also valid for a plane wall and tube. Since the value of b is positive for insulating materials and negative for metals, the effect of the rise in temperature is to increase the heat loss through the insulation and decrease the conduction heat transmission through the metal walls. Example 2.28 A spherical shell is made of a material with thermal conductivity k = koT2. Derive the expression for the conduction heat transfer rate, if the inner and outer walls are held at temperatures T1 and T2, respectively.
84
2
OneDimensional SteadyState Heat Conduction
Solution Substituting k = koT2 and A = 4pr2 in Fourier’s equation, we get dT q ¼ ko T 2 4pr 2 dr Separating the variables and integrating, ZRo q Ri
dr ¼ 4pko r2
ZT2
T 2 dT
T1
or q 1 1 1 ¼ T23 T13 4pko Ri Ro 3 or 4 Ro Ri 3 q ¼ pko T1 T23 3 Ro Ri Example 2.29 The wall of a steam boiler furnace is made of layer of ﬁrebricks of thickness 125 mm (kb = 0.28 + 0.002T) and insulation bricks layer of 250 mm thickness (ki = 0.7 W/ (m K)). The inside surface temperature of the ﬁrebrick wall is 1000°C, and outside insulation brick temperature is 60°C. (i) Calculate the amount of heat lost per m2 of the furnace wall and the temperature at the interface of the two layers. (ii) If the layer of the insulation brick wall is reduced to half of its original thickness and an insulating material with km = 0.113(1 + 0.002T) is to be ﬁlled between the ﬁre brick and insulation brick layers, what must be the thickness of this insulating material so that the heat loss remains the same. Assume that the inside and outside wall temperatures remain as before. Temperature T in thermal conductivity equations is the mean temperature in K. Solution (i) Given data are kb = 0.28 + 0.002T, ki = 0.7 W/(m K), Ab = Ai = A = 1 m2. Let temperature at the interface of ﬁre and insulation bricks is T2. The heat transfer, refer to Fig. 2.36a, q ¼ k b Ab
1273 T2 T2 333 ¼ k i Ai di db
2.9 Effect of Variable Thermal Conductivity Fire bricks (kb)
85
Insulation bricks [ki = 0.7 W/(m K)]
•T2
Ti = 1273 K•
Insulation bricks [ki = 0.7 W/(m K)]
Fire bricks (kb)
•T2
Ti = 1273 K • • T3 = 333 K
δb
• T3 = 333 K
δi
δm δb
a
b
δ i’
Fig. 2.36 Example 2.29
or 1273 þ T2 1273 T2 T2 333 ¼ 0:7 1 0:28 þ 0:002 1 250 2 125 or T2 ¼ 1165:6 K And the heat loss is q ¼ ki A
T2 333 1165:6 333 ¼ 2331:3 W ¼ 0:7 1 di 0:25
(ii) With three layers, the heat flow equation for the reduced thickness di′ of the insulation bricks and the same heat loss rate, q ¼ ki A
T3 333 d0i
or 2331:3 ¼ 0:7 1
T3 333 0:125
or T3 ¼ 749:3 K Heat flow equation for the ﬁrebrick layer is q ¼ kb A b
1273 T2 ¼ 2331:3 W db
As there is no change as far as ﬁrst layer is concerned, hence, T2 is the same as in Part (i), i.e. T2 = 1165.6 K.
86
2
OneDimensional SteadyState Heat Conduction
For the second layer of the insulating material ﬁlled between layers of ﬁre bricks and insulation bricks, 2331:3 ¼ km Am
T2 T3 dm
where Am ¼ Ab ¼ Ai ¼ 1 m2 ; km ¼ 0:113 ½1 þ 0:002 ðT2 þ T3 Þ=2 ¼ 0:113 ½1 þ 0:002 ð1165:6 þ 749:3Þ=2 ¼ 0:3294 W=ðm KÞ Substitution gives 2331:3 ¼ 0:3294 1
1165:6 749:3 dm
Solution gives dm = 0.0588 m = 59 mm. Example 2.30 The wall of a furnace is builtup of a 250 mm thick layer of ﬁreclay bricks whose thermal conductivity is given by k ¼ 0:83 ð1 þ 0:0007tm Þ W=ðm CÞ Calculate the rate of heat loss per m2 of the wall if the temperature of the gas in the furnace is 1200°C and the ambient temperature is 20°C. The ﬁlm coefﬁcients from the inner and outer wall surfaces are 30 W/(m2 K) and 20 W/(m2 K), respectively. Solution The total heat resistance to the heat transfer per m2 of the wall is 1 d 1 þ þ hi k h0 1 0:25 1 þ þ ¼ 30 k 20
R¼
As the thermal conductivity of the wall material is function of its temperature, the problem will be solved by following an iterative method. We assume mean wall temperature tm as 600°C for the ﬁrst trial. This gives k ¼ 0:83ð1 þ 0:0007 600Þ ¼ 1:1786 W=ðm CÞ and R ¼ 0:296
2.9 Effect of Variable Thermal Conductivity
87 Fluid film
ti• t1 •
q
• ho • t2
hi •
• to δ
ti
1/(hi A)
1/(ho A)
(δ/kA) t1
to
t2
Fig. 2.37 System of Example 2.30 and thermal network
Knowing the total wall resistance, the heat transfer rate is q ti to 1200 20 ¼ ¼ 3986:5 W=m2 ¼ A 0:296 R Let the wall surface temperatures are t1 and t2, refer to Fig. 2.37, then q ti t1 t1 t2 t2 to ¼ ¼ ¼ ¼ 3986:5 W=m2 A 1=hi d=k 1=ho This gives q 1 1 ¼ 1200 3986:5 ¼ 1067:1 C A hi 30 q 1 1 ¼ 219:3 C t2 ¼ t0 þ ¼ 20 þ 3986:5 A ho 20
t1 ¼ ti
This gives ﬁrst estimate of the mean temperature as tm ¼
t1 þ t2 ¼ 643:2 C 2
The second trial with the estimated values of tm gives k ¼ 0:83ð1 þ 0:0007 643:2Þ ¼ 1:2037 W=ðm CÞ For k = 1.2037 W/(m °C), R ¼ 0:291 and q ¼ 4055 W=m2 A Third trial is not needed as the value of q/A from the second trial is not in much difference from that obtained from the ﬁrst trial.
88
2
OneDimensional SteadyState Heat Conduction
Example 2.31 A 50mmthick magnesia insulation has its faces at 25°C and 45°C, and the heat loss is found to be 25 W/m2. When the face temperatures are changed to 240°C and 300°C, respectively, the heat loss becomes 100 W/m2. Find the thermal conductivity–temperature relation of the material assuming it to be linear. Solution The heat loss for a plane wall is given by q ¼ kA
t1 t2 d
ðiÞ
Let k t a ko d A
ko(1 + at), i.e. a linear relation as speciﬁed mean temperature = (t1 + t2)/2 = (25 + 45)/2 = 35°C in the ﬁrst case= (240 + 300)/2 = 270°C in the second case a constant thermal conductivity at 0°C thickness of the wall = 0.05 m heat flow area = 1 m2
Case (i) q = 25 W/m2 and from Eq. (i), 25 ¼ ko ð1 þ 35aÞ 1
45 25 0:05
or ko ð1 þ 35aÞ ¼ 0:0625
ðiiÞ
Case (ii) q = 100 W/m2 and from Eq. (i), 100 ¼ ko ð1 þ 270aÞ 1
300 240 0:05
or ko ð1 þ 270aÞ ¼ 0:0833 Solution of Eqs. (ii) and (iii) gives ko = 0.0594 and a = 1.491 10−3. Hence, the thermal conductivity–temperature relation is k ¼ 0:0594ð1 þ 1:491 103 tÞ where t is the mean temperature of the material in °C.
ðiiiÞ
2.10
2.10
Critical Thickness of Insulation
89
Critical Thickness of Insulation
2.10.1 Critical Thickness of Insulation for Cylinders The thermal resistance to heat transfer for a hollow cylinder with a layer of insulating material applied to the cylindrical surface is given by, refer to Fig. 2.38a, ln rr1i ln rr21 1 1 þ Rt ¼ þ þ 2pri lhi 2pr2 lho 2pkm l 2pkl ¼ Ri þ Rw þ Rk þ Ro The ﬁrst term Ri in the equation of Rt is the resistance due to the inner ﬁlm, second term Rw is the metal wall resistance, the third term Rk represents the resistance of the insulation material and the last term Ro is the outer ﬁlm resistance. The ﬁrst two terms are constant, while Rk and Ro vary with the change in the thickness of the insulation. Neglecting the ﬁrst two terms, we have ln rr21 1 R0t ¼ þ ðiÞ 2pr2 lho 2pkl As we increase the thickness of the insulation, i.e. the radius r2, keeping inside radius of insulation r1 constant, the resistance Rk to conduction heat flow through the layer of insulation increases while the resistance Ro to the heat transfer by convection from the surface of insulation decreases, see Fig. 2.38. The sum of the resistances, Rt′ = Rk + Rc, ﬁrst decreases to a certain minimum at r = rc and then again increases. It means that as the radius r2 increases, the heat flow q ﬁrst increases to a certain maximum value at r2 = rc and then decreases. The value of r2 at the maximum of heat flow q (or the minimum of the resistance R) is called the critical radius of insulation. The critical radius of the insulation can be determined as under. The heat transfer rate through the wall and insulation of thickness (r2 – r1) with ﬁlm coefﬁcients hi and ho at the inner and outer surfaces, respectively, is given by q¼
ðt i t o Þ Rt
q
Tube wall ho
Insulation R, q
• • hi r2
R Rk
ri
Ro
r1 0
a Fig. 2.38 Critical radius of insulation
r1
rc
b
r
90
2
OneDimensional SteadyState Heat Conduction
where Rt ¼
ln
r1 ri
ln
r2
r1 1 1 þ þ þ 2pri lhi 2pr2 lho 2pkm l 2pkl
The condition for the maximum or minimum heat transfer is obtained by differentiating the equation of the total resistance Rt with respect to r2 and equating to zero, i.e. 2 3 ln rr1i ln rr21 @Rt @ 4 1 1 5 þ þ ¼0¼ þ @r2 2pri lhi 2pr2 lho 2pkm l 2pkl @r2 or @Rt 1 1 r1 1 1 ¼ ¼0 @r2 2pl k r2 r1 ho r22 or r2 ¼ rc ¼
k ho
ð2:50Þ
This is the equation of the critical radius of insulation for a cylinder, which is independent of the tube radius r1. It can be shown that the value of the second derivative of the resistance Rt at r2 = rc refers to the minimum value of the resistance, i.e. @ 2 Rt 1 @ 1 1 1 1 ¼ 2pl @r2 k r2 r22 ho @r22 or ¼
1 1 1 2 1 2 þ 3 2pl k r2 r2 ho
Putting r2 = rc = k/ho, 2 @ 2 Rt 1 ho 2h2o 3 þ 3 ¼ 2pl k k @r22 or @ 2 Rt h2o ¼ 2plk3 @r22 which is a positive quantity. Thus, the condition r2 = rc refers to the minimum resistance to the heat transfer or to the maximum heat transfer.
2.10
Critical Thickness of Insulation
91
We can conclude that, for the heat insulation of a pipe of radius r1, the insulating material must be selected such that the critical radius rc is less than r1 under the given conditions while in the case of the electric current carrying cables, the insulation radius r2 must be equal to rc to dissipate maximum heat. Hence, the insulating material for the cables must be selected such that the critical radius rc is greater than the bare cable radius r1.
2.10.2 Critical Thickness of Insulation for Spherical Vessel The critical radius of the insulation for this can be determined as under. The heat transfer rate through the wall of the sphere and insulation of thickness (r2 – r1) with ﬁlm coefﬁcients hi and ho at the inner and outer surfaces, respectively, is given by q¼
ðti to Þ : Rt
where Rt ¼
1 1 1 1 1 1 1 1 þ þ þ 4pk r1 r2 4pri2 hi 4pkm ri r1 4pr22 ho
Again the condition for the maximum or minimum heat transfer is obtained by differentiating the equation of the total resistance Rt with respect to r2 and equating to zero, i.e. @Rt 1 @ 1 1 1 1 1 1 1 1 ¼0¼ þ þ þ 4p @r2 ri2 hi km ri r1 k r1 r2 @r2 r22 ho or 1 1 1 2 2 þ 3 ¼0 4p k r2 r 2 ho or r2 ¼ rc ¼
2k ho
ð2:51Þ
This is the equation of the critical radius of insulation for a sphere, which is independent of the sphere radius r1. Again it can be shown that the value of the second derivative of the resistance Rt at r2 = rc refers to the minimum value of the resistance. 1 @ 1 1 2 2 þ 3 4p @r2 k r2 r ho 2 1 1 ð2Þ 2 ð3Þ 3 þ ¼ 4p k r24 ho r2
92
2
OneDimensional SteadyState Heat Conduction
Putting r2 = rc = 2 k/ho, @ 2 Rt 1 h3o ¼ 4p 8k4 @r22 which is a positive quantity. Thus, the condition r2 = rc refers to the minimum resistance to the heat transfer or to the maximum heat transfer. The discussion presented for cylindrical cell regarding the selection of the insulation material also applies here. Example 2.32 An electric current carrying cable of radius r1 is covered with a layer of insulation of thickness r2 – r1. If the resistance heating rate is q per m length of the cable, show that in the steadystate condition, the cable surface temperature is given by q 1 r2 1 ln þ þ t1 2p k r1 r2 h where k is the thermal conductivity of the insulation material, h is the outside ﬁlm coefﬁcient and t∞ is the surrounding temperature. Hence, show that r2 = k/h gives the minimum cable surface temperature. Solution The rate of heat transfer per unit length of the insulated cable at surface temperature ts is given by q¼h
ts t1 r2 1 2pk ln r1
þ
1 2pr2 h
i
or q 1 r2 1 ln þ ts ¼ þ t1 2p k r1 r2 h For the cable surface temperature to be minimum or maximum, the condition is @ts ¼0 @r2 or @ q 1 r2 1 ln þ þ t1 ¼ 0 @r2 2p k r1 r2 h
2.10
Critical Thickness of Insulation
93
or
q 1 r1 1 1 þ 2 ¼0 2p k r2 r1 r2 h
or
q 1 1 1 ¼0 2p k r2 r22 h
or r2 ¼
k h
The second derivative is @ q 1 1 1 @r2 2p k r2 r22 h q 1 1 2 ¼ 2p k r22 r23 h @ 2 ts q h2 Substituting r2 = k/h, the second derivate 2 ¼ which is a positive quantity. @r2 2p k3 Hence, the condition r2 = k/h refers to the minimum temperature of the cable surface. Example 2.33 For the cable of the previous example, show that the condition for keeping the insulated cable at the same temperature as if it were not insulated is ln
r2 k 1 1 ¼ r1 h r1 r2
Solution The cable surface temperature with insulation of thickness (r2 – r1) from the previous example is q 1 r2 1 ln þ ts ¼ þ t1 2p k r1 r2 h When insulation is not present, the cable surface temperature ts0 can be obtained from the above equation by excluding the insulation resistance and putting r2 = r1. This gives ts0
q 1 ¼ þ t1 2p r1 h
94
2
OneDimensional SteadyState Heat Conduction
Equating the two temperature equations for the given condition, we have q 1 r2 1 q 1 ln þ þ t1 ¼ þ t1 2p k r1 r2 h 2p r1 h Simpliﬁcation gives r2 k 1 1 ln ¼ r1 h r1 r2
Example 2.34 An electric current cable is 5 mm in radius. It is to be covered by a uniform sheathing of plastic (k = 0.2 W/(m K)). The convective ﬁlm coefﬁcient between the surface of the plastic sheathing and surrounding air is 15 W/(m2 K). Calculate (i) thickness of insulation which will not alter the temperature of the wire surface, and (ii) the insulation thickness for minimum wire temperature. Comment on the result. Solution (i) The condition for keeping the insulated cable at the same temperature as if it were not insulated has been obtained in previous example, which is ln
r2 k 1 1 ¼ r1 h r1 r2
From the given data, r1 = 5 mm, k = 0.2 W/(m K) and h = 15 W/(m2 K). Substitution gives, by trial and error, r2 = 0.057 m = 57 mm. (ii) For the minimum wire temperature, the condition is r2 ¼ rc ¼
k 0:2 1000 ¼ 13:33 mm ¼ ho 15
Since r2 calculated in Part (i) of the example is greater than rc, any insulation thickness which is less than r2 calculated in Part (i) will cause decrease in the wire surface temperature for the same heat generation rate in the wire due to Joule heating. The wire surface temperature is minimum when r2 = rc. The wire surface temperature for r2 = r1 = 5 mm (i.e. the bare wire) and r2 = 57 mm is the same. Example 2.35 Show that heat dissipation from an electric cable provided with critical insulation thickness is given by q¼
2pkLðts t1 Þ 1 þ ln hokr1
where the notations have the usual meaning.
2.10
Critical Thickness of Insulation
95
Solution The heat dissipation from an electric cable of radius r1 with insulation thickness (ro – r1) is given by q¼ ¼
ts t1 ro 1 1 2pkL ln r1 þ 2pro Lho 2pkLðts t1 Þ ln rro1 þ rokho
For the critical insulation thickness, substitute ro by k/ho, which gives the desired result. Example 2.36 An electric wire of 3 mm radius is to be provided with plastic sheathing [k = 0.15 W/(m K)]. The convective ﬁlm coefﬁcient on the surface of the bare cable as well as insulated cable can be assumed to be 10 W/(m2 K). For a particular value of current flow I1, the wire surface temperature is found to be 60°C when the surrounding temperature was 20°C. Find (i) The insulation thickness so that the wire surface temperature will be minimum when current flow is unchanged. What will be wire surface temperature in this case? Assume that the wire resistance does not change with change in its temperature. (ii) The maximum possible increase in the current carrying capacity if the wire surface temperature is to be limited to the same value after the insulation of critical thickness. Solution (i) To keep the wire as cool as possible, the thickness of the insulation must be equal to the critical thickness, which is rc ¼
k 0:15 ¼ 0:015 m ¼ ho 10
i.e. the thickness of the insulation is 15 – 3 = 12 mm. For the bare wire, the heat dissipation rate is q ¼ ho ð2pr1 LÞ ðt1 t1 Þ ¼ 10 ð2p 0:003 1Þ ð60 20Þ ¼ 7:54 W per m length of the cable Heat dissipation rate from the insulated cable is given by, refer to Example 2.35, q0 ¼ ¼
2pkLðt1 t1 Þ 1 þ ln hokr1
2p 0:15 1 T10 20 0:15 1 þ ln 100:003
¼ 0:3612ðT10 20Þ
96
2
OneDimensional SteadyState Heat Conduction
When the current flow is unchanged, the heat generation rate will remain the same, i.e. q′ = q. Hence, 7:54 ¼ 0:3612ðt10 20Þ or t10 ¼ 40:87 C (ii) The heat dissipation rate and hence the current carrying capacity of the cable will be maximum when the critical thickness of the insulation is provided. For t1 ¼ t10 ¼ 60 C; the heat dissipation rate for ro = rc will be, from Part (i), q ¼ 0:3612ð60 20Þ ¼ 14:45 W For the given wire resistance, q ðwithout insulationÞ ¼ I12 R ¼ 7:54 W q0 ðwith insulationÞ ¼ I22 R ¼ 14:45 W Thus, I2 ¼ I1
rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 14:45 ¼ 1:384 7:54
i.e. the current can be increased by 38.4%. Example 2.37 A vertical noninsulated steam pipe of 4 m length and 80 mm diameter has surface temperature tw = 160°C. The surrounding air is at 20°C. Find the loss of the heat from the pipe by convection if the surface heat transfer coefﬁcient is 9 W/(m2 K). What will be the heat loss from the pipe length if an asbestosbased insulation [k = 0.16 W/(m K)] of 30 mm thickness is applied? Give your comments. Solution Heat loss without insulation q ¼ hðpDLÞðtw t1 Þ ¼ 9 p 0:08 4 ð160 20Þ ¼ 1266:7 W Heat loss when insulation is provided (r1 = 40 mm, r2 = 70 mm) q0 ¼ ¼
tw t1 r2 1 1 2pkL ln r1 þ 2pr2 Lho
160 20 ¼ 691:97 W 1 þ 2p0:0749
1 70 2p0:164 ln 40
2.10
Critical Thickness of Insulation
97
Comments: The critical radius of insulation, from the given data, is rc ¼
k 0:16 ¼ 0:0178 m ¼ 17:8 mm ¼ ho 9
As the critical radius is less than the tube radius, the given insulation will be effective whatever may be thickness of the insulation applied. The loss will decrease with the increase in the thickness of insulation. Note: In the above analysis, it has been assumed that the surface temperature of the tube and the convective heat transfer coefﬁcient remains the same even after the layer of the insulation is applied. Example 2.38 If the pipe in Example 2.37 is covered with insulation having thermal conductivity of 0.5 W/(m K), discuss the effect of variation in insulation thickness. Solution Heat loss without insulation q ¼ hðpDLÞðtw t1 Þ ¼ 9 p 0:08 4 ð160 20Þ ¼ 1266:7 W Heat loss with insulation provided for r1 = 40 mm q0 ¼ ¼
tw t1 r2 1 1 ln 2pkL r1 þ 2pr2 Lho
160 20 þ 2pr2149
r2 1 2p0:54 ln 40
The critical thickness of insulation rc ¼
k 0:5 ¼ 0:0555 m ¼ ho 9
The variation of the heat loss with r2 is shown in table below and depicted in Fig. 2.39. r2 (mm)
Heat loss, Q (W)
r2 = r1 = 40 45 50 rc = 55.5 60 70 80 90
1266.70 1300.96 1318.68 1324.36 1321.44 1300.0 1267.92 1231.8
Example 2.39 An electric wire is 6 mm in diameter. The wire is to be covered by at least 4 mm rubber insulation [k = 0.1 W/(m K)] from electric consideration. How the cooling conditions will be changed? The heat transfer coefﬁcient from the surface is 10 W/(m2 K).
98
2
OneDimensional SteadyState Heat Conduction
Fig. 2.39 Heat loss versus insulation radius
Solution The critical radius of insulation is rc ¼
k 0:1 ¼ 0:01 m ¼ 10 mm ¼ ho 10
As rc > rwire and rwire < r2 < rc, the insulated wire will be cooled better than the bare wire. At insulation thickness of 7 mm, the cooling effect will be at maximum. The following calculations can be carried out for clear understanding: Bare wire: q ¼ ho ðpDLÞðtwire t1 Þ or twire t1 ¼
q q 1000 ¼ 5:305q for 1 m length ¼ ho pDL 10 p 6 1
Insulated wire: q¼
twire r2 1 2pkL ln r1
t1 þ 2pr12 Lho
2.10
Critical Thickness of Insulation
99
or twire t1 ¼ q
1 7 1 ln þ 2p 0:1 1 3 2p 0:007 1 10
¼ 3:622q
Similarly, for the insulation thickness of 7 mm, twire t1 ¼ 3:508q and for the insulation thickness of 10 mm, twire t1 ¼ 3:557q: From the above calculations, it can be seen that the wire temperature, over the surrounding temperature, (twire – t∞) decreases with the increase in the insulation thickness up to r2 = rc and then increases with the further increase in the thickness. Again the wire temperature is minimum at r2 = rc. Example 2.40 A cylindrical pipe of 20 mm outer diameter is to be insulated so that the heat loss from the pipe surface is not more than 65 W per m length. The pipe surface is 280°C, and it can be assumed that the surface temperature remains the same after application of the insulation layer. The surrounding temperature is 30°C. The heat transfer coefﬁcient from the surface is 10 W/(m2 K). Available for the service is asbestos [k = 0.14 W/(m K)] or slag wool [k = 0.08 W/(m K)]. Determine the thicknesses of these insulating materials and comment on the result. Solution The critical radius of insulation, rc ¼
k 0:14 ¼ 0:014 m ¼ 14 mm for asbestos ¼ ho 10
rc ¼
k 0:08 ¼ 0:008 m ¼ 8 mm for slag wool ¼ ho 10
and
Thus the critical radius of insulation rc is less than the pipe radius r1 for the slag wool. Therefore, the use of the slag wool as insulation is recommended. If asbestos is used as insulation, the thickness of it should be such that r2 is well beyond rc, refer to Fig. 2.40. For slag wool, the curve below r = r1 is shown dashed because it is hypothetical only. Heat loss from the bare pipe: q ¼ hðpDLÞðtw t1 Þ ¼ 10 p 0:02 1 ð280 30Þ ¼ 157:1 W Heat loss with insulation:
100
2
OneDimensional SteadyState Heat Conduction 2 Asbestos 1 Slag wool
q
2 1
rc1 r1 rc2
r
Fig. 2.40 Heat loss versus radius for alternative insulations of Example 2.40
q¼
tw t1 r2 1 1 ln 2pkL r1 þ 2pr2 Lho
For asbestos r2 = rc = 14 mm (r1 = 10 mm). Heat loss from the above equation is qa ¼
280 30 ¼ 164:54 W: 1 þ 2p0:014110
1 14 2p0:141 ln 10
Asbestos insulation thickness for heat loss of 65 W per m length: 65 ¼
280 30 þ 2pr21110
r2 1 2p0:141 ln 10
Solution gives r2 = 280 mm, by trial and error. Similarly for the slag wool (rc = 8 mm), 65 ¼
280 30 þ 2pr21110
r2 1 2p0:081 ln 10
which gives r2 = 61 mm by trial and error. Example 2.41 Show that for a small diameter tube, when critical radius is greater than the tube radius r1, the heat loss can be reduced by applying an insulating layer of thickness (r2 – r1) such that 1 r2 1 1 ln þ [ k r 1 ho r 2 ho r 1 Solution Resistance to heat flow when insulation layer is applied, 1 r2 1 ln R ¼ þ 2plk 2pr2 lho r1 0
2.10
Critical Thickness of Insulation
101
Resistance to heat flow from a bare tube, R¼
1 2pr1 lho
For reduction in the heat loss R′ must be greater than R, i.e. 1 r2 1 1 ln [ þ 2plk 2pr2 lho 2pr1 lho r1 Simpliﬁcation of the above equation gives the desired result. For the cooling of the electric cables, the condition will be R′ < R.
2.11
Thermal Contact Resistance
In many engineering applications, solid surfaces are not having metallurgical bond. At the interface of such surfaces, a resistance to the heat flow has been observed. This resistance is called thermal contact resistance. It can be the dominant resistance when high conductivity metals are involved. A drop in temperature at the interface of the two materials due to the resistance of the interface can be seen in Fig. 2.41. Real surfaces are not perfectly smooth. Hence, when such surfaces are pressed against each other, contact between them occurs only at a limited number of spots as shown in Fig. 2.41b, which has been exaggerated for a better understanding. The voids are ﬁlled with the surrounding fluid. The heat transfer across the interface takes place through two parallel paths: 1. The conduction at the high spots, which are in direct contact 2. The conduction through the fluid entrapped in the voids. Since the fluid layer is very thin, there is no convection. At high temperatures, radiation across the gap may also contribute to the heat transfer.
T1
1
2 Drop in the temperature
Ti1 Ti2
δg
T2
b a
Interface
Rs1
Rs2
Rv
c Fig. 2.41 a Effect of the thermal contact resistance on temperature proﬁle, b contact surfaces (enlarged), c the thermal network at the interface
102
2
OneDimensional SteadyState Heat Conduction
The heat flow equation can be written as Ti1 Ti2 1 1 1 þ Rs Rv
q¼
ðiÞ
where Rs Rv (Ti1 – Ti2)
resistance of the spots in direct contact, resistance of the fluid layer, temperature drop across the interface.
The resistance Rs consists of two resistances Rs1 and Rs2 in series; the resistances Rs and Rv are in parallel as shown in Fig. 2.41c. If it is assumed that the void thickness dg, refer to Fig. 2.41b, is equally divided between the solid surfaces 1 and 2, then the resistance Rs can be expressed as Rs ¼
dg =2 dg =2 þ k 1 As k 2 As
ðiiÞ
The fluid resistance can be expressed as Rv ¼
dg kf Av
ðiiiÞ
where k1 k2 As Av kf
thermal conductivity of solid 1, thermal conductivity of solid 2, area of the spots of solidtosolid contact, void area and thermal conductivity of the entrapped fluid. Substitution in Eq. (i) gives " q ¼ ðTi1 Ti2 Þ
dg dg þ 2k1 As 2k2 As
1
dg þ kf A v
1 #
ðivÞ
The interface or the contact conductance is deﬁned as q ¼ hAðTi1 Ti2 Þ
ðvÞ
where A is the total contact area of the surfaces. Comparison of Eqs. (iv) and (v) gives the expression of the contact conductance as 1 h¼ dg
As 2k1 k2 Av þ kf A k1 þ k2 A
ð2:52Þ
2.11
Thermal Contact Resistance
103
Ti = 150oC • T1
T2 q
hi = 20 W/(m2 K)
T3
δ1
(δ1/k1A)
1/(hiA) Ti
T1
ho = 10 W/(m2 K)
T2
a
•
• To = 30oC
δ2
Rc /A q
T4
(δ2/k2A) T3
1/(hoA)
T4
To
b
Fig. 2.42 Example 2.42
It is not easy to use this equation because of the difﬁculty in determination of the values of As, Av and dg. However, the equation shows that h is a function of void thickness dg, thermal conductivity kf of the entrapped fluid, and area As. The void thickness dg depends on the roughness of the surfaces in contact. Thus, the roughness of the surfaces plays an important role in determining the contact resistance. The resistance increases with the increase in the roughness. Increase in the joint pressure for the surfaces in contact decreases the resistance. With the increase in the joint pressure, the high spots are deformed (crushed). This increases the contact area As and reduces dg. Experimental studies have shown that the conductance varies almost directly with the joint pressure. At very low gas pressures, when the free path of the molecules is large compared to the dimension of the void space, the effective thermal conductivity of the entrapped gas decreases. This leads to an increase in the contact resistance. Studies have shown that the contact conductance ranges from 3120 to 71,000 W/(m2 K) for steel, brass and aluminium surfaces ground to various degrees of roughness for pressures from 1.31 to 55.1 bar with air, oil or glycol between the surfaces (Frass 1989). Thin foils made of materials of good thermal conductivity when inserted between the surfaces reduce the resistance if the foil is softer than the materials of the surfaces in contact. The resistance is reported to reduce by as much as 75% when thermal grease like Dow 3401 is used (Holman 1992). The values of the contact resistance of some surfaces are also tabulated in Holman (1992) and Mills (1995). However, the data is scarce and not reliable (Mills 1995). Example 2.42 The composite wall (A = 2 m2) shown in Fig. 2.42 consists of layers of bricks (k1 = 0.6 W/(m K), d1 = 0.1 m) and mineral wool (k2 = 0.045 W/(m K), d2 = 0.05 m). The contact resistance is estimated to be 0.25 m2 K/W. Determine (a) the rate of heat transfer through the wall and (b) the temperature distribution.
A grease like silicone fluid thickened with metal oxide ﬁller [k = 0.4 W/(m K)].
1
104
2
OneDimensional SteadyState Heat Conduction
Solution (a) The electric network is shown in Fig. 2.42b. The total resistance to heat transfer is 1 d1 Rc d2 1 þ þ þ þ hi A k 1 A A k2 A ho A 1 0:1 0:25 0:05 1 þ þ þ þ ¼ 0:839 K=W ¼ 20 2 0:6 2 2 0:045 2 10 2
Rt ¼
Heat transfer rate, q¼
Ti To 150 30 ¼ 143:03 W ¼ 0:839 Rt
(b) We can write q¼
Ti T1 1=hi A
which gives q hi A 143:03 ¼ 146:42 C ¼ 150 20 2
T1 ¼ Ti
Similarly, T2 ¼ T1
qd1 k1 A
¼ 146:42
143:03 0:1 ¼ 134:5 o C, 0:6 2
qRc A 143:03 0:25 ¼ 116:62 o C, ¼ 134:5 2 qd2 T4 ¼ T3 k2 A 143:03 0:05 ¼ 37:15 o C ¼ 116:62 0:045 2 T3 ¼ T2
and q ho A 143:03 ¼ 30 C ¼ 37:15 10 2
T0 ¼ T4
The temperature distribution is shown in the ﬁgure.
2.12
2.12
Summary
105
Summary
In this chapter, basically Fourier’s law of heat conduction has been applied to systems where onedimensional heat flow occurs. Mechanism of heat conduction in metals and alloys, construction and heatinsulating materials, and liquid and gases has been discussed. In general, the thermal conductivity depends on temperature, pressure and nature of the substance. It differs by many thousand times for the materials of interest. The highest values are for the metals, followed by dense ceramics, organic solids and liquids, while the lowest values are for the gases. Superinsulation materials have been developed for cryogenic applications. Typical values of the thermal conductivity of several common engineering materials are given in a tabulated form in Appendix A. In general, the temperature gradient may exist in all three directions of a solid. There may be internal heat generation. The temperature can also vary with the time (unsteady state). Hence, a general heat conduction equation must be developed, which can be used to evaluate the heat transfer in any direction under steady or unsteady state and with or without heat generation. The general heat conduction equations in the rectangular, cylindrical and spherical coordinates have been developed. In the unsteady state, a material property termed as thermal diffusivity, which is ratio of thermal conductivity (k) and thermal capacity (qc) of the material, controls the temperature distribution. The higher the thermal diffusivity of a material, the higher the rate of temperature propagation, i.e. the equalization of the temperature at all points of the space will proceed at a higher rate. Fourier’s equation of heat conduction is exactly analogous to Ohm’s law for an electrical conductor. We use this electrical analogy frequently as it is quite useful in solving the complex heat conduction problems. The temperature difference Δt is the driving force for the flow of heat and d/kA is the thermal resistance in case of a plane wall. Where the wall face is exposed to a fluid, the heat is transferred to and from the wall by convection. In such cases, ﬁlm resistance 1/hA is considered along with conduction resistance of the wall. If heat also transfers by radiation from the surface of the solid, the resistance is 1/[(hr + hc)A], where hr is termed as the radiation heat transfer coefﬁcient, which has been discussed in Chap. 11. In case of onedimensional steadystate heat conduction through a cylindrical shell, the wall resistance is lnðr2 =r1 Þ=ð2pkLÞ and ﬁlm resistances at the inner and outer surface of the shell are 1/hiAi and 1/hoAo, respectively. In heat exchangers, the equation of heat exchange between the fluids flowing in and outside the tube is written as q ¼ UAðti to Þ where U is termed as overall heat transfer coefﬁcient, refer to Eqs. (2.29a), (2.29b), (2.30a) and (2.30a). In case of onedimensional steadystate heat conduction through a spherical shell, ðr2 r1 Þ=ð4pkr1 r2 Þ is the resistance of the wall to the conduction heat transfer and ﬁlm resistances are 1/hiAi and 1/hoAo. In Sect. 2.8, experimental schemes of measurement of thermal conductivity of solids, liquids and gases have been presented. Thermal conductivity of a dry solid substance depends on temperature and when the temperature difference across the wall of such substance is large, the effect of the variation of thermal conductivity with the temperature must be considered. This has been considered in Sect. 2.9.
106
2
OneDimensional SteadyState Heat Conduction
In case of a hollow cylinder or sphere, with the increase in the thickness of the insulating material, the conduction resistance Rk of the insulation increases while the ﬁlm resistance Ro decreases. Mathematical analysis shows that the sum of the resistances (Rk + Ro) ﬁrst decreases to a certain minimum value at radius r = rc and then again increases. It means that as the insulation radius r2 increases, the heat flow q ﬁrst increases to a certain maximum value at r2 = rc and then decreases. The value of r2 at the maximum of heat flow q (or the minimum of the resistance) is called the critical radius of insulation. The knowledge of the concept of critical radius of insulation helps in the selection of type of insulation material. For the heat insulation of a pipe of radius r1, the insulating material must be selected such that the critical radius rc is less than r1 while in the case of the electric current carrying cables, the insulation radius r2 must be equal to rc to dissipate maximum heat. Hence, the insulating material for the cables must be selected such that the critical radius rc is greater than the bare cable radius r1. Mathematical expressions for the critical radius of insulation have been deduced. In many engineering applications, solid surfaces are not having metallurgical bond. At the interface of such surfaces, a resistance to the heat flow has been observed, which is termed as thermal contact resistance. It can be a dominant resistance when high conductivity metals are involved. A basic mathematical treatment for estimate of the thermal contact resistance has been given. A signiﬁcant number of illustrative examples have been included in this chapter on onedimensional steadystate heat conduction.
Review Questions 2:1. Deﬁne one and twodimensional heat transfer when applied to conduction problems. 2:2. Discuss the mechanism of heat conduction in metals and nonmetals. What is the effect of alloying on thermal conductivity of metals? 2:3. Write a short note on insulating materials. 2:4. Discuss the factors that affect thermal conductivity of metals, insulating materials, liquids and gases. 2:5. (i) Derive threedimensional general heat conduction equation in (a) Cartesian coordinates (b) Cylindrical coordinates (c) Spherical coordinates for transient condition considering uniform volumetric heat generation. (ii) Deduce onedimensional steadystate conduction equations from the equations derived in Part (i) if there is no heat generation. The thermal conductivity of the material can be assumed to be constant. 2:6. Deﬁne thermal diffusivity and explain its physical signiﬁcance giving suitable example. 2:7. What do you mean by critical radius of insulation? Show that it is given by (i) (k/ho) for a cylindrical conﬁguration and (ii) (2 k/ ho) for a spherical vessel, where k is the thermal conductivity of the insulating material and ho is the convection coefﬁcient of the heat loss to the surrounding.
2.12
Summary
107
2:8. What is the signiﬁcance of critical radius of insulation with reference to (a) insulation of pipes to reduce the heat flow, and (b) insulation of the electric cables? 2:9. Deﬁne overall heat transfer coefﬁcient and develop its expression for heat transfer in the tube flow. 2:10. How would you measure the thermal conductivity of an insulating material in the form of powder? Draw a neatlabelled sketch of the equipment suggested. 2:11. Present a scheme for experimental measurement of thermal conductivity of material of a flat plate. 2:12. What is thermal contact resistance? Discuss its mechanism and the parameters on which it depends. What are the engineering applications where it is relevant? 2:13. Show that the geometric mean radius Rm of a spherical shell of outside radius R2 and inside radius R1 is given by Rm ¼ ðR1 R2 Þ0:5 2:14. Show that the temperature distribution across the wall of a hollow sphere operating with steady and uniform surface temperature T1 on the inner surface radius R1 and T2 on the outer surface radius R2 is given by T T1 ¼ T 2 T1
1=R1 1=r 1=R1 1=R2
where R1 < r < R2. 2:15. The thermal conductivity of a certain material varies with temperature as k ¼ ko 1 þ aT þ bT 2 where a and b are constants. Derive from the ﬁrst principles an expression for the heat flow rate for a hollow sphere of the inner and outer radii R1 and R2, respectively. The temperatures of the inner and outer surfaces are T1 and T2, respectively. dt ; Integra[Ans. Following the method of Sect. 2.9.3, q ¼ ko ð1 þ aT þ bT 2 Þ4pr 2 dr tion gives
q
R2 R1 R1 R2
q
RR2 dr R1
r2
¼ 4pko
RT2
ð1 þ aT þ bT 2 Þdt;
T1
¼ 4pko ðT2 T1 Þ þ a2 T22 T12 þ b3 T23 T13
or q ¼ 4pko
R 1 R2 a b 2 2 ðT2 T1 Þ 1 þ ðT2 þ T1 Þ þ T2 þ T1 T2 þ T1 2 3 R 2 R1
108
2
OneDimensional SteadyState Heat Conduction
Problems 2:1 Draw the thermal network (electrical analog) and determine the heat flow rate through the wall shown in Fig. 2.43a. Assume onedimensional heat flow and uniform temperature at the interface of different layers. [Ans. The thermal network is shown in Fig. 2.43b; Ra = 0.0375; Rb = 0.25; Rc = 0.25; Rd = 0.2; Re = 2.86; Rf = 2.5; Rg = 2.143; Resistances Rb, Rc and Rd are in parallel, and Re and Rf are also in parallel; The total resistance is 3.59 and q = 125.35 W/m2.] 2:2 The wall of a house may be approximated by 225 mm layer of ordinary bricks [k = 0.5 W/(m K)] with both sides of it plastered with 20mmthick layer of cement plaster [k = 1.2 W/(m K)]. What thickness of plywood sheet [k = 0.15 W/(m K)] be added to the inside of the wall to reduce the heat loss (or gain) through the wall by 20 percent? Neglect the contact resistance. If there is an air space of average 0.25 mm thickness [k = 0.0257 W/(m K)] between the wall and the plywood, determine the effect on the heat flow rate. Assume only conduction heat flow across the air layer. P P [Ans. R = 0.483 without plywood sheet; Since q / 1= R , after application of P plywood desired R′ = 0.483/0.8 = (0.483 + dply/0.15); This gives dply = 18.1 mm. P With the air layer Ra′ = 0.483/0.8 + 0.25/(1000 0.0257) = 0.6135. Reduction in P P P heat flow = (1/ R – 1/ R′)/(1/ R) = 21.3%.] 2:3 A vessel with flat bottom, refer to Fig. 2.44, is made of 2mmthick nickel sheet [k = 60 W/(m K)]. The outside of its bottom has a plating of 0.5mmthick copper [k = 350 W/(m K)]. Calculate the reduction in the heat transfer rate due to the deposition of scale on the lower side of the bottom. Rscale = 0.001 K/W. The ﬁlm
b ka = 2.0 W/(m K), δa = 75 mm kb = 0.8 W/(m K), δb = 50 mm kc = 0.4 W/(m K), δc = 50 mm kd = 1.0 W/(m K), δd = 50 mm ke = 0.07 W/(m K), δe = 100 mm kf = 0.08 W/(m K), δf = 100 mm kg = 0.035 W/(m K), δg = 75 mm
0.25 m a 1m
e g
c d
f
0.5
Depth = 1 m. Wall side temperatures are 0.25 m 500oC and 50oC. δ = 50 mm
a Wall crosssection. Rb
Re 50oC
500oC Ra Rc
Rd
Rf
b Thermal network. Fig. 2.43 Problem 2.1
Rg
2.12
Summary
109
hi = 1000
ho = 500
Fig. 2.44 Problem 2.3
coefﬁcients on the two sides of the bottom are 1000 W/(m2 K) and 500 W/(m2 K), respectively. P P [Ans. Without scale, R = 3.035 10−3; q / 1= R ¼ 329:5; With the scale, P P R = 4.035 10−3; q / 1= R ¼ 247:8; Reduction in the heat transfer rate due to the deposition of scale = (329.5 – 247.8)/329.5 = 24.8%.] 2:4 Two layers of glass (each 4 mm thick) in a window panel are 3 mm apart. The air [k = 0.025 W/(m K)] in between these layers can be assumed to be dry and stagnant. Find the heat loss through the panel for a temperature difference across the panel of 20°C. If the two layers of glass are replaced by a single glass of 10 mm thickness, how the loss will change? [kglass = 0.8 W/(m K)]. [Ans. q1/A = 153.85 W/m2; q2/A = 1600 W/m2.] 2:5 Derive an expression for the heat transfer rate from a cylindrical tank of length L with hemispherical covers. The wall thickness throughout is (R2  R1). The temperatures at the inner and outer surfaces are T1 and T2, respectively (T1> T2). Assume onedimensional heat flow. Þ 1 T 2Þ þ 2 2pk ðT1 T2 [Ans. q ¼ 2pkL ðT ln
R2 R1
1 1 R1 R 2
2:6 An aluminium pipe carries steam at 110°C. The pipe [k = 185 W/(m K)] has an inside diameter of 100 mm and outside diameter of 120 mm. The pipe is located in a room where the ambient air temperature is 30°C and convective heat transfer coefﬁcient between the pipe and air is 15 W/(m2 K). Determine the heat transfer rate per unit length of the pipe. To reduce the heat loss from the pipe, it is covered with a 50mmthick layer of insulation [k = 0.20 W/(m K)]. Determine the heat transfer rate per unit length from the insulated pipe. Assume that the convective resistance of the steam is negligible. [Ans. Here R1 = 50 mm; R2 = 60 mm; R3 = 110 mm; kal = 185 W/(m K); ki = 0.2 W/(m K); ho = 15 W/(m2 K); This gives (i) q ¼ 2p 1 ð11030Þ ¼ 451:98 W 1 kal
(ii) q ¼ 2p 1
ln
R2 R1
þ ho1R
2
ð11030Þ 1 kal
ln
R2 R1
þ k1 ln i
R3 R2
þ ho1R
¼ 138:18 W 3
2:7 A steel pipe carrying a hot fluid has inside diameter of 120 mm and an outside diameter of 160 mm. It is insulated at the outside with asbestos. The fluid temperature is 150°C, and the surrounding air temperature is 20°C, h (fluid side) = 100 W/(m2 K), h (air side) = 8 W/(m2 K), kasbestos = 0.18 W/(m K) and ksteel = 42 W/(m K). How thick should the asbestos layer be provided in order to limit the heat loss to 150 W per metre length of the pipe?
110
2
[Ans. 150 ¼ 2p 1
OneDimensional SteadyState Heat Conduction
ð15020Þ 1 h i R1
þk
1 steel
ln
R2 R1
1 asbestos
þk
ln
R3 R2
þ ho1R
This gives R3 = 182.7 mm 3
and thickness of insulation = R3 – 80 = 102.7 mm.] 2:8 Determine overall heat transfer coefﬁcient (based on the inner diameter) for the data of Problem 2.7 and calculate the heat transfer rate. [Ans. Ui = 3.0606 W/(m2 K); q = UAi(Ti – To) = 150 W.] 2:9 Determine the heat flow rate for the data of Problems 2.6 and 2.7 neglecting the pipe wall resistance and comment on the result. [Ans. Neglect (1/k) ln(R2/R1) term. The heat transfer rates are 138.22 W and 150.18 W as against 138.18 W and 150 W, respectively. The effect of the wall resistance in both cases is negligible because the resistances offered by the insulation are much larger in magnitude.] 2:10 A steel pipe d1/d2 = 100/110 mm is covered with 80 mm thick layer of an insulating material. The thermal conductivity of the insulation material depends on temperature and is given by k ¼ 0:06 ð1 þ 0:4 102 TÞ W=ðm CÞ: Determine the heat loss per m length of the pipe, if the temperature of the outer surface of the pipe is 200°C and the temperature of the outer surface of the insulation is 40°C. [Ans. For insulation, R1 = 110/2 = 55 mm, R2 = 55 + 80 = 135 mm; k ¼ 0:06 h i 1 T 2Þ 1 þ 0:4 102 ðT1 þ2 T2 Þ ¼ 0:0888; Lq ¼ 2pk ðT ¼ 99:4 W ln
R2 R1
2:11 A steam pipeline of d1/d2 = 100/110 mm is covered with two layers of insulation of equal thickness d1 = d2 = 50 mm. The average thermal conductivity of the ﬁrst layer of insulation material is 0.05 W/(m K) and that of the second layer is 0.1 W/(m K). If the layer of greater thermal conductivity is placed next to the pipe, what will be the effect on the heat loss? Neglect thermal resistance of the pipe material. R3 R2 1 1 [Ans. Rt ¼ 2pk ln ln þ 2pk2 R1 R2 ; R1 = 55 mm, R2 = 105 mm, R3 = 155 mm. For 1 the ﬁrst part of the problem, k1 = 0.05, k2 = 0.1 W/(m K) and L = 1 m, which gives Rt = 2.678 K/W. In the second part, k1 = 0.1 and k2 = 0.05. W/(m K). This gives Rt = 2.2688 K/W. Thus the resistance decreases by 18%. Hence, the heat loss will increase by 18%. The results show that the better insulation must be put next to the pipe.] 2:12 An electronic instrument for a probe is contained in a spherical shell made of 30 mm thick layer of mild steel (inside radius of 150 mm) and a 10 mm thick outside layer of stainless steel. The two layers can be assumed to be in perfect thermal contact. The instrument inside the shell generates heat at a rate of 6 kW. Estimate the inside surface temperature, if the outer surface is estimated to be at 5°C [kms = 40 W/(m K) and kss = 15 W/(m K)]. [Ans. T1 = 27.57°C.] 2:13 The energy loss from a 16 mm outer diameter pipe, which carries a hot fluid, is to be reduced. Available for the service are magnesia 85 [k = 0.071 W/(m K)] and glass wool [k = 0.038 W/(m K)]. Would you help in making the correct selection if heat transfer coefﬁcient h is likely to be 5 W/(m2 K)? [Ans. The critical radius, (rc)mag = 14.2 mm, (rc)glass = 7.6 mm; (rc)mag is greater than outer radius of the pipe; hence, it should not be used. For the glass wool, any thickness will reduce the heat loss.]
2.12
Summary
111
2:14 A thinwalledlong cylindrical vessel of 0.8 m outer diameter contains a chemical undergoing exothermic reaction. The outer surface temperature of the vessel is 180°C when the surrounding air is at 30°C. Determine the thickness of polyurethane foam (rigid) insulation required so that the outer surface temperature of the insulation is not more than 40°C. The convection heat transfer coefﬁcient on outer surface is estimated to be 20 W/(m2 K). Neglect radiation heat loss. [Ans. Heat flow by conduction through the insulation of thickness (r2  r1) equals the ðT1 T 2 Þ ¼ h 2pr2 l convective heat transfer from the outer surface, i.e. q ¼ 2pki l ln
r2 r1
ðT2 T1 Þ: Substituting T1 = 180°C, T2 = 40°C, T∞ = 30°C, r1 = 0.4 m, h = 20 W/ (m2 K) and ki = 0.025 W/(m K) from Table A3.2 of Appendix A, we get r2 = 0.417 m; Insulation thickness (r2  r1) = 0.017 m = 17 mm.] 2:15 A hollow aluminum sphere [ka = 204 W/(m K)] with a 100 W electric heater at its centre is used to determine the thermal conductivity of an insulating material. The inner and outer radii of the sphere are 0.1 and 0.125 m, respectively. In a particular test, a spherical shell of an insulating material of 75 mm thickness was applied on the outer surface of the sphere. In the steady state, the temperature at the inner surface of the aluminum wall was recorded as 200°C. The system is exposed to room air at 30°C and the convective heat transfer coefﬁcient is estimated to be 20 W/(m2 K). What is the thermal conductivity of the insulation? Neglect heat loss by radiation and contact resistance. [Ans. r1 = 0.1 m, r2 = 0.125 m, r3 = 0.2 m, T1 = 200°C, T∞ = 30°C, ka = 204 W/ ðT1 T1 Þ ; (m K), h = 20 W/(m2 K), q = 100 W; Heat flow rate, q ¼ 1 4p
1 ka
1 1 r1 r2
þ k1
i
1 1 r2 r3
þ
1 hr 2 3
solution after substitution of various known values gives ki = 0.15 W/(m K).] 2:16 A fluid is stored in a thin metallic spherical shell (outside diameter of 2 m) with 200 mm thick insulation [ki = 0.05 W/(m K)]. The shell outer surface temperature is –20°C. Determine the thickness of the insulation which experiences temperature 0°C when the surrounding air temperature is 30°C and convective heat transfer coefﬁcient is 10 W/(m2 K). [Ans. r1 = 1 m, r2 = 1.2 m, T1 = −20°C, T∞ = 30°C, ki = 0.05 W/(m K), h = 10 W/ (m2 K); Heat flow rate, q ¼ h ðT1 T 1 Þ i :Substitution gives q = 184.65 W; Using 1 1 4p ki
1 1 r1 r2
þ
1 hr 2 2
heat flow equation between shell radius r1 and radius r0 where the temperature is 0°C, 184:65 ¼ ð0T1 Þ ; which gives r0 = 1.073 m, i.e. the thickness of the insulation 1 4pki
1 1 r1 r0
which experiences temperature 0°C is 73 mm.] 2:17 A cylinder is covered with three different insulation materials A, B and C as shown in Fig. 2.45. Determine the heat flow rate per unit length of the cylinder. Assume radial heat flow only.
112
2
OneDimensional SteadyState Heat Conduction
kC = 0.075 W/(m K) t1 = 100oC
r1 = 50 mm r2 = 75 mm kA = 0.025 W/(m K) r1 h = 20 W/(m2 K)
kB = 0.05 W/(m K)
(Rk)A
t1
(Rk)B
qA
qB
o
● t∞ = 30 C
(Rk)C
qC
Ro
Ro
t∞
Ro
Fig. 2.45 Problem 2.16
[Ans. Heat flows through three paths comprising the insulation materials A, B and C (see the thermal network). 8 q qA qB qC 1 < ðt t Þ h 1 1 þ þ ¼ ¼ L L L L 3 : 1 1 ln r2 þ r1 2p kA
ðt t Þ i þ h 1 1 r2 1 1 1 hr2 2p kB ln r1 þ
ðt t Þ i þ h 1 1 r2 1 1 1 hr2 2p kC ln r1 þ
1 hr2
i
9 = ;
¼ 49:53 W=m
Note: The insulations will have different temperatures but circumferential heat flow has not been considered.] 2:18 Crosssectional area of a rod varies with x as Ax = A0ebx where A0 and b are constants. Determine temperature distribution equation for onedimensional steadystate axial conduction heat flow in the rod. The thermal conductivity of the material is constant, and lateral surface of the rod is insulated. C C bx [Ans. qx = constant gives kAx dT dx: Integration dx ¼ C: Hence, dT ¼ kAx dx ¼ kA0 e
gives T ¼ kAC0 b ebx þ C2 ¼ C1 ebx þ C2 where C1 ¼ kAC0 b ; Constants C1 and C2 can be determined from temperature boundary conditions.]
References Fraas Arthur P (1989) Heat exchanger design, 2nd edn. Wiley, New York Holman JP (1992) Adapted for SI units by White PRS. Heat transfer. McGrawHill Book Co., New York Mills AF (1995) Heat and mass transfer. Irwin, USA
3
Extended Surfaces (Fins)
3.1
Introduction
The heat flow from or to a surface in contact with a fluid is given by q ¼ hADt
ð3:1Þ
The heat transfer rate q in this case can be increased by increasing the surface area. This can be achieved by attaching metal pieces to the heattransferring surface. These extended surfaces are known as ﬁns. Fins are widely used in various engineering equipments. They are provided on the cylinders of aircooled internal combustion engines. Small diameter metal rods attached vertically to the condenser tubes at the back of a domestic refrigerator also work as ﬁns. Electrical appliances such as transformers and motors are provided with the ﬁns for efﬁcient dissipation of heat generated in these electrical equipments. Fins of various shapes and in different arrangements have been employed. Some of them are shown in Fig. 3.1. The longitudinal ﬁns are long metal strips attached to the surface as shown in Fig. 3.1a. These ﬁns are commonly employed with tubes where the fluid moves along the axis of the tube. Transverse ﬁns, Fig. 3.1b, are employed in crossflow (flow perpendicular to the axis of the tube). Disctype transverse ﬁns are either welded to the tube surface or shrunk on to the tube. These ﬁns may be attached to the tube in the form of helix (such as welding a metal strip to the tube continuously in the form of a helix). Discontinuous and star ﬁns are other types of transverse ﬁns. A majority of ﬁns are of rectangular crosssection, Fig. 3.1d. Thick ﬁns and ﬁns made by casting are generally of the trapezoidal crosssection with rounded edges. Spine or studtype ﬁns are basically metal cones, cylinders or pipes which extend from the pipe or tube surface. They are employed both in case of longitudinal and crossflows. These ﬁns are mechanically rugged and hence have longer life in corrosive atmosphere than thin plate ﬁns. Let us consider ﬁns attached to the outer surface of a cylinder, as shown in Fig. 3.1, to increase the heat transfer rate from the cylinder surface to the surrounding. For the heat to be conducted to the ﬁn, the ﬁn temperature tf must be lower than the cylinder surface temperature ts. Hence, the difference between the ﬁn and surrounding temperatures (tf – t∞) is less than the difference between the cylinder surface and surrounding temperatures (ts – t∞). Further, the temperature of the ﬁn continuously decreases from the ﬁn base (i.e. the cylinder wall) to the ﬁn tip. Thus, each unit area of the ﬁn surface is less effective in heat transfer (heat © Springer Nature Singapore Pte Ltd. 2020 R. Karwa, Heat and Mass Transfer, https://doi.org/10.1007/9789811539886_3
113
114
3
Extended Surfaces (Fins)
A cylinder with fins. Fin
a Longitudinal fins
Tube
b Transverse fins
c Helical fins
d Rectangular fins e Trapezoidal fins
f Spine or stud fins
g Star fins
Fig. 3.1 Fin shapes and arrangements
transfer is proportional to the area temperature difference) than the area of the cylinder surface to which the ﬁns have been attached. However, the effectiveness with which ﬁns transfer heat also depends on the ﬁn proﬁle (shape), its length and the number of ﬁns provided per m2 of the surface. In the end, it may be noted that the ﬁnal choice of the type of ﬁn and their arrangement depends not only on the heat transfer performance but also on the resistance offered to the flow of the surrounding fluid, cost and the ease of fabrication.
3.2
Heat Transfer from a Fin of Uniform CrossSection
Figure 3.2 shows a ﬁn of uniform crosssection projecting from a wall, which is to be cooled. Let the ﬁn is of uniform crosssection Ac throughout its length and the wall is at temperature ts. The heat, which is conducted to the ﬁn, is rejected by convection from the ﬁn surface to the surrounding fluid at temperature t∞.
3.2 Heat Transfer from a Fin of Uniform CrossSection Wall
115 Crosssectional area Ac, perimeter P
• t∞ qc
ts •
qx
qx+dx
x L
t ts
Crosssection
dx
t
Fin temperature x
Fig. 3.2 Heat balance on a ﬁn of uniform crosssection
An exact analytical solution of the heat propagation through a ﬁn involves considerable difﬁculties. However, an approximate solution can be made using the following assumptions: (i) Temperature at any crosssection of the ﬁn is uniform. The assumption is valid for ﬁns of small crosssection and made of high thermal conductivity material. This assumption reduces the problem to that of onedimensional heat conduction along the axis of the ﬁn only. (ii) Steadystate condition. (iii) Thermal conductivity of the ﬁn material is constant, and (iv) The heat transfer coefﬁcient h from the ﬁn surface to the surrounding is constant for the entire ﬁn surface. Most of the ﬁns used in practice are plate type, Fig. 3.1a, with small thickness of the plate and the error due to the ﬁrst approximation is less than 1%. The greatest uncertainty is in the value of the heat transfer coefﬁcient, which is seldom uniform over the entire ﬁn surface. In the case of severe nonuniformity of the heat transfer coefﬁcient and two or threedimensional conduction heat flow, numerical techniques are used to solve the problem. Consider a very small elemental length dx of the ﬁn at a distance x from the wall surface, Fig. 3.2. The rate of heat flow into the element by conduction is qx ¼ kAc
dt dx
ðiÞ
The rate of heat flow out of the element by conduction is qx þ dx ¼ qx þ
d ðqx Þdx dx
Heat transfer to the surrounding fluid by convection from the surface area (perimeter dx) of the element from Newton’s law is qc ¼ hAðt t1 Þ ¼ hðPdxÞðt t1 Þ where t is the temperature of the element and P is the perimeter of the ﬁn. In the steady state, heat inflow by conduction must equal the heat outflow by conduction and convection, i.e.
116
3
Extended Surfaces (Fins)
qx ¼ qx þ dx þ qc or qx ¼ qx þ
d ðqx Þdx þ hðPdxÞðt t1 Þ dx
or d ðqx Þdx þ hðPdxÞðt t1 Þ ¼ 0 dx Substitution of the value of qx from Eq. (i) gives d dt kAc dx þ hðPdxÞðt t1 Þ ¼ 0 dx dx or d2 t hP ¼ ðt t 1 Þ dx2 kAc
ðiiÞ
Let (t − t∞) = h, then d2 t d2 h ¼ dx2 dx2 Substitution in Eq. (ii) gives d 2 h hP ¼ h dx2 kAc Putting
hP kAc
¼ m2 , the equation is transformed into d2 h ¼ m2 h dx2
which is a secondorder differential equation. Its general solution is h ¼ C1 emx þ C2 emx
ð3:2Þ
The constants C1 and C2 are to be determined from the boundary conditions of the problem. For this purpose, we shall discuss three different cases.
3.2.1 A Very Long Fin If the ﬁn is very long compared to its crosssection dimensions, then the ﬁn end approaches the surrounding fluid temperature. The boundary conditions for this case are
3.2 Heat Transfer from a Fin of Uniform CrossSection
117
(i) at x = 0, t = ts and (ii) as x ! ∞, t! t∞ The ﬁrst condition gives ts t1 ¼ C1 þ C2
ðiÞ
and the second condition gives t1 t1 ¼ 0 ¼ C1 em1 þ C2 em1 or 0 ¼ C1 em1 This gives C1 ¼ 0. Hence, from Eq. (i), C2 ¼ ts t1 Substitution in Eq. (3.2) gives h ¼ ðt t1 Þ ¼ ðts t1 Þemx or ðt t1 Þ ¼ emx ðts t1 Þ
ð3:3Þ
The equation reveals that the temperature along the ﬁn length varies exponentially as depicted in Fig. 3.2. Heat rejection rate from the entire surface area of the ﬁn qﬁn equals the heat flow by conduction at the ﬁn base, i.e. qfin
dt ¼ kAc dx x¼o
Using Eq. (3.3), we have
qfin qfin
d mx ¼ kAc ½ðts t1 Þe þ t1 dx x¼o rﬃﬃﬃﬃﬃﬃﬃ hP kAc ðts t1 Þ ¼ mkAc ðts t1 Þ ¼ kAc
or qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ
ð3:4Þ
118
3
Extended Surfaces (Fins)
3.2.1.1 Comments on Parameter M From the equation of m = √(hP/kAc), it follows that m is proportional to √h, inversely proportional to √k and is a function of ﬁn geometry presented by √(P/Ac). Figure 3.3 shows the dependence of the temperature distribution in the ﬁn along its length for different values of parameter m. For a ﬁn made of high thermal conductivity material, which will give a low value of m for a given ﬁn and heat transfer coefﬁcient h, a large excess temperature (t – t∞) is obtained along the length of the ﬁn. Similar effect will be seen for low values of the heat transfer coefﬁcient and ratio of ﬁn perimeter to its area of crosssection.
3.2.2 Negligible Heat Transfer from the Fin End as Compared to the Heat Transferred from the Fin Surface (Ac PL) This situation can occur when the area of the ﬁn end is very small in comparison to the surface area of the entire ﬁn (Ac PL) as in the case of a thin long ﬁn. Mathematically, this can be stated by equating the heat reaching the ﬁn end by conduction to zero, which is equivalent to an insulated tip ﬁn. Hence, qend ¼ kAc
dt ¼0 dx x¼L
or dt ¼0 dx x¼L The differentiation of Eq. (3.2) with respect to x gives dt ¼ C1 memx þ C2 ðmÞemx dx Hence, dt ¼ 0 ¼ C1 memL mC2 emL dx x¼L
t = ts
m1 > m2 > m3 m3 m2
t∞
m1 x
Fig. 3.3 Effect of ﬁn parameter m on temperature distribution along the ﬁn
3.2 Heat Transfer from a Fin of Uniform CrossSection
119
or C1 emL C2 emL ¼ 0
ðiÞ
At the ﬁn base (x = 0), the temperature is ts, i.e. ts t1 ¼ C1 þ C2
ðiiÞ
From Eqs. (i) and (ii), ts t1 1 þ e2mL ts t1 C2 ¼ 1 þ e2mL C1 ¼
Substitution of the values of constant C1 and C2 in Eq. (3.2) gives h¼
ts t1 mx ts t1 mx e þ e 2mL 1þe 1 þ e2mL
or t t1 emx emx ¼ þ ts t1 1 þ e2mL 1 þ e2mL Using the relation cosh mL ¼ e
mL
þ emL , 2
the above equation can be transformed into
t t1 cosh mðL xÞ ¼ cosh mL ts t1
ð3:5Þ
The rate of heat transfer from the ﬁn equals the heat conducted into the ﬁn base, i.e. qfin
dt ¼ kAc dx x¼o
Using Eq. (3.5), we have qfin ¼ kAc
d cosh mðL xÞ þ t1 ðts t1 Þ dx cosh mL x¼o
or qfin
1 d ½cosh mðL xÞ ¼ kAc ðts t1 Þ cosh mL dx x¼o
or qfin ¼ kAc ðts t1 Þ
1 ½m sinh mðL xÞx¼o cosh mL
120
3
Extended Surfaces (Fins)
or qfin ¼ kAc mðts t1 Þ
sinh mL cosh mL
Substituting the value of m, we obtain the heat transfer equation as qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ tanh mL
ð3:6Þ
The values of exponential and hyperbolic functions are tabulated in Table 3.1 and Fig. 3.4 shows variation of tanh(mL) with mL. It can be seen that for mL = 5.0, tanh(mL) = 0.9999 1 Table 3.1 Exponential and hyperbolic functions mL
emL
e−mL
sinh(mL)
cosh(mL)
tanh(mL)
0.0
1.00
1.00
0.00
1.00
0.00
0.1
1.105
0.905
0.10
1.005
0.0997
0.2
1.221
0.819
0.20
1.020
0.1973
0.3
1.350
0.741
0.305
1.045
0.2913
0.4
1.492
0.670
0.411
1.0311
0.380
0.5
1.649
0.607
0.521
1.128
0.462
0.6
1.822
0.549
0.637
1.186
0.537
0.7
2.014
0.497
0.759
1.255
0.6044
0.8
2.226
0.449
0.888
1.337
0.664
0.9
2.460
0.407
1.027
1.433
0.7163
1.0
2.718
0.368
1.175
1.543
0.7616
1.1
3.004
0.333
1.336
1.669
0.8005
1.2
3.320
0.301
1.509
1.811
0.8337
1.3
3.670
0.272
1.698
1.971
0.862
1.4
4.055
0.247
1.904
2.151
0.8854
1.5
4.482
0.223
2.129
2.352
0.905
1.6
4.953
0.202
2.376
2.577
0.922
1.7
5.474
0.1827
2.646
2.828
0.9354
1.8
6.050
0.1653
2.942
3.107
0.947
1.9
6.686
0.150
3.268
3.418
0.956
2.0
7.389
0.1353
3.627
3.762
0.964
2.1
8.166
0.1224
4.022
4.144
0.9705
2.2
9.025
0.111
4.457
4.568
0.976
2.3
9.974
0.100
4.937
5.037
0.980
2.4
11.02
0.0907
5.466
5.557
0.984
2.5
12.18
0.0821
6.05
6.132
0.987
2.6
13.46
0.074
6.695
6.770
0.989
2.7
14.88
0.067
7.406
7.473
0.991
2.8
16.445
0.061
8.192
8.253
0.9926
2.9
18.174
0.055
9.060
9.115
0.994
3.0
20.09
0.050
10.018
10.068
0.995
4.0
54.60
0.0183
27.29
27.31
0.9993
74.20
74.21
0.9999
5.0
148.4
6.74 10
−3
3.2 Heat Transfer from a Fin of Uniform CrossSection
121
1.0
tanhmL
0.8 0.6 0.4 0.2 0.0 0.0
1.0
2.0
3.0
4.0
5.0
6.0
mL
Fig. 3.4 Variation of tanhmL with mL
and Eqs. (3.4) and (3.6) give the same values of rate of heat transfer. Since the advantage gained by using a ﬁn length greater than given by mL = 2 is negligible (tanhmL = 0.964 for mL = 2), this must be considered the maximum useful limit of the ﬁn length.
3.2.3 Short Fins (Fin with Heat Loss from the Fin End) In the case of short ﬁns, the heat loss from the ﬁn end cannot be neglected. The heat loss from the ﬁn end equals the heat reaching the end by conduction, i.e. dt hL Ac ðtL t1 Þ ¼ kAc dx x¼L
ðiÞ
where hL is the heat transfer coefﬁcient at the ﬁn end. Equation (3.2) can be written as h ¼ C1 ðsinh mx þ cosh mxÞ þ C2 ðcosh mx sinh mxÞ or ðt t1 Þ ¼ ðC1 C2 Þ sinh mx þ ðC1 þ C2 Þ cosh mx or ðt t1 Þ ¼ C3 sinh mx þ C4 cosh mx At x = 0, t = ts. This gives ðts t1 Þ ¼ C4 Hence, ðt t1 Þ ¼ C3 sinh mx þ ðts t1 Þ cosh mx
ðiiÞ
122
3
Extended Surfaces (Fins)
Using this temperature relation in Eq. (i), we get kAc
d ½C3 sinh mx þ ðts t1 Þ cosh mx þ t1 dx
¼ hL Ac ½C3 sinh mL þ ðts t1 Þ cosh mL x¼L
or k½C3 m cosh mL þ mðts t1 Þ sinh mL ¼ hL ½C3 sinh mL þ ðts t1 Þ cosh mL Simpliﬁcation gives C3 ¼ ðts t1 Þ
hL cosh mL þ mk sinh mL hL sinh mL þ mk cosh mL
Substitution of the value of C3 in Eq. (ii) gives ðt t1 Þ ¼ ðts t1 Þ cosh mx ðts t1 Þ
hL cosh mL þ mk sinh mL sinh mx hL sinh mL þ mk cosh mL
Simpliﬁcation gives t t1 cosh mðL xÞ þ ðhL =mkÞ sinh mðL xÞ ¼ ðhL =mkÞ sinh mL þ cosh mL ts t1
ð3:7Þ
which is the temperature distribution equation. The rate of heat transfer from the ﬁn equals the heat conducted into the ﬁn base, i.e. dt qfin ¼ kAc dx x¼o Using Eq. (3.7), we have qfin ¼ kAc
d cosh mðL xÞ þ ðhL =mkÞ sinh mðL xÞ þ t1 ðts t1 Þ dx ðhL =mkÞ sinh mL þ cosh mL x¼o
or sinh mðL xÞ ðhL =mkÞ cosh mðL xÞ qfin ¼ kAc ðts t1 Þm ðhL =mkÞ sinh mL þ cosh mL x¼o or qfin ¼ mkAc ðts t1 Þ
sinh mðL xÞ þ ðhL =mkÞ cosh mðL xÞ ðhL =mkÞ sinh mL þ cosh mL x¼o
or qfin ¼ mkAc ðts t1 Þ
sinh mL þ ðhL =mkÞ cosh mL ðhL =mkÞ sinh mL þ cosh mL
3.2 Heat Transfer from a Fin of Uniform CrossSection
123
or qfin ¼ kmAc ðts t1 Þ
ðhL =mkÞ þ tanh mL ðhL =mkÞ tanh mL þ 1
Substituting the value of m, we get the heat transfer equation as qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ðhL =mkÞ þ tanh mL hPkAc ðts t1 Þ ðhL =mkÞ tanh mL þ 1
ð3:8Þ
Equations (3.7) and (3.8) reduce to Eqs. (3.5) and (3.6) when hL = 0, i.e. when there is no heat rejection from the ﬁn end. Further value of factor mL > 5 reduces Eq. (3.8) to Eq. (3.4) of the very long ﬁn. The heat transfer equation of the short ﬁn can be expressed in the same form as that of ﬁn of Sect. 3.2.2 if a corrected length Lc, as deﬁned below, is used in all equations of this case. The heat lost, in case of a short ﬁn, from the ﬁn end is hL Ac ðtL t1 Þ
ðiiiÞ
We extend the ﬁn length L by ΔL, see Fig. 3.5, such that the heat lost by the lateral surface of the extended ﬁn length equals the heat loss from the ﬁn end given by Eq. (iii). The heat transfer from the lateral surface of the extended length is hL ðPDLÞðtL t1 Þ
ðivÞ
assuming that the ﬁn end temperature does not change when extended by length ΔL. Equating Eqs. (iii) and (iv), we get hL ðPDLÞðtL t1 Þ ¼ hL Ac ðtL t1 Þ or DL ¼
Ac P
and the corrected length is Lc ¼ L þ DL For a circular crosssection ﬁn, DL ¼
Ac ðp=4Þd2 d ¼ ¼ 4 P pd
L ΔL
Fig. 3.5 Corrected ﬁn length
124
3
Extended Surfaces (Fins)
For a rectangular crosssection ﬁn, DL ¼
Ac Wd d ¼ 2W 2 P
Replacing L by Lc in Eqs. (3.5) and (3.6), we get the equations for the short ﬁn as t t1 cosh mðLc xÞ ¼ cosh mLc ts t1 and qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ tanh mLc
ð3:9Þ
ð3:10Þ
The error which results from this approximation is very small if (hd/2 k)1/2 0.5.
3.3
Hollow Fins
It is important to note that the equations developed here can be applied to hollow ﬁns. For such ﬁns, the perimeter will be the sum of the inside and outside if the heat transfer takes place from the inside surface also. The crosssectional area is the solid area. The equations will not apply to these conditions unless the heat transfer coefﬁcient at the inside and outside surfaces is equal. This is evident from the examination of the derivation.
3.4
Composite Fins
Such ﬁns usually have a core of high thermal conductivity metal such as copper and a sheathing of steel. They are used in hightemperature applications where the core metal may get oxidized. The sheathing protects the core metal. Figure 3.6 shows the crosssection of such a ﬁn. The resistance to heat flow from the core caused by the sheath is usually very small compared with the external convection coefﬁcient, and hence the ﬁn crosssection at any axial position can be assumed to be at a uniform temperature. With this approximation, the conduction heat transfer equation at any section is dt q ¼ ðk1 A1 þ k2 A2 þ . . .Þ ð3:11Þ dx x and
sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP m¼ ðk1 A1 þ k2 A2 þ . . .Þ
where P is the perimeter of the composite ﬁn, i.e. the ﬁn taken as a whole. Core
Sheathing
Fig. 3.6 A composite ﬁn
ð3:12Þ
3.4 Composite Fins
125
Example 3.1 Heat generated in a bearing by friction causes the temperature at the shaft end to increase to 60°C above the ambient temperature. How is the temperature distributed along the shaft as we move away from the bearing? Calculate the amount of heat transferred through the shaft if the convective heat transfer coefﬁcient for the shaft surface is 7 W/(m2 K) and the thermal conductivity of the shaft material is 60 W/(m K). The shaft diameter is 60 mm and may be assumed a rod of inﬁnite length. Solution The temperature distribution is given by ðt t1 Þ ¼ emx ðts t1 Þ
ð3:3Þ
Here, rﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP hpd 4h 47 m¼ ¼ ¼ ¼ ¼ 2:79 2 kAc kðp=4Þd kd 60 0:06 and ðts t1 Þ ¼ 60o C Hence, ðt t1 Þ ¼ 60e2:79x Heat rejection rate from the entire surface area of the shaft is qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ
ð3:4Þ
or qfin ¼
qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 7 p 0:06 60 ðp=4Þ ð0:06Þ2 60 ¼ 28:39 W
Example 3.2 A 600mmlong pump shaft of stainless steel [k = 20 W/(m K)] is 30 mm in diameter. The heat transfer coefﬁcient for the outside surface of the shaft to the cooling air at 30°C is 30 W/(m2 K). The impeller end of the shaft is immersed in a hot fluid which may raise the temperature of the shaft end to 250°C. At what distance the bearing must be placed so that the maximum temperature at the bearing is not greater than 80°C? If the shaft is hollow with di = 20 mm, what will be the temperature at the bearing location determined above? The shaft is not cooled from the inner surface.
126
3
Extended Surfaces (Fins)
Solution (i) Solid Shaft Here rﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP hpd 4h 4 30 m¼ ¼ ¼ ¼ ¼ 14:14 m1 2 kAc kðp=4Þd kd 20 0:03 and mL ¼ 8:485 [ 5 The temperature distribution for this case is ðt t1 Þ ¼ emx ðts t1 Þ
ð3:3Þ
Substitution gives 80 30 ¼ e14:14x 250 30 or x ¼ 0:1048 m (ii) Hollow Shaft Ac ¼ ðp=4Þðdo2 di2 Þ ¼ ðp=4Þð0:032 0:022 Þ; P ¼q pd ﬃﬃﬃﬃﬃo ¼ p 0:03; hP ¼ 18:97 m1 m ¼ kA c and the product mL is again greater than 5. The temperature distribution for this case is also given by ðt t1 Þ ¼ emx ðts t1 Þ
ð3:3Þ
or t ¼ ðts t1 Þemx þ t1 At x = 0.1048 m it is t ¼ ð250 30Þe18:970:1048 þ 30 ¼ 60:12o C The temperature is lower due to the reduced area for conduction heat transfer along the shaft.
3.4 Composite Fins
127
Example 3.3 Three solid rods, made of silver [ks = 420 W/(m K)], aluminium [ka = 205 W/(m K)] and iron [ki = 70 W/(m K)], are coated with a uniform layer of wax all around. The rods are placed vertically in boiling water bath with 250 mm length of each rod projecting outside. If all rods are 5 mm in diameter, 300 mm in length and have the same surface heat transfer coefﬁcient of 40 W/(m2 K), work out the ratio of lengths up to which the coated material will melt on each rod if the melting temperature of the wax is 40°C. The surrounding temperature is 20°C. Solution The rods work as ﬁns of length 250 mm each. The corrected length for each rod is Lc ¼ L þ d=4 ¼ 250 þ 5=4 ¼ 251:25 mm: The value of parameter m is rﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP hpd 4h 4 40 178:9 ¼ pﬃﬃﬃ m¼ ¼ ¼ ¼ kAc kðp=4Þd2 kd k 0:005 k Hence, parameters m and mLc for the three rods are 178:9 178:9 ms ¼ pﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 8:73 and ms Lc ¼ 8:73 0:25125 ¼ 2:19 ks 420 178:9 178:9 ma ¼ pﬃﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 12:5 and ma Lc ¼ 12:5 0:25125 ¼ 3:14 ka 205 178:9 178:9 mi ¼ pﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃ ¼ 21:38 and mi Lc ¼ 21:38 0:25125 ¼ 5:37 ki 70 Let the lengths up to which wax will melt are L1, L2 and L3 on the ﬁrst, second and third rods, respectively. It means that at lengths L1, L2 and L3, the three rods are having the same temperature, i.e. t = 40°C. The equation of temperature distribution for a short ﬁn is t t1 cosh mðLc xÞ ¼ cosh mLc ts t1
ð3:9Þ
or 0:25 ¼
cosh mðLc xÞ cosh mLc
or x ¼ Lc
cosh1 ð0:25 cosh mLc Þ m
ðiÞ
128
3
Extended Surfaces (Fins)
For silver rod, let the length x = L1 at which t = 40°C, then from Eq. (i), L1 ¼ 0:25125
cosh1 ½0:25 coshð2:19Þ ¼ 0:193 m 8:73
Similarly for the aluminium and steel rods, the lengths L2 and L3, respectively, are L2 ¼ 0:25125
cosh1 ½0:25 coshð3:14Þ ¼ 0:113 m 12:5
L3 ¼ 0:25125
cosh1 ½0:25 coshð5:37Þ ¼ 0:065 m 21:38
and
Hence, L1 : L2 : L3 :: 2:97 : 1:74 : 1 If we assume the ﬁns to be very long, then the temperature distribution is given by ðt t1 Þ ¼ emx ðts t1 Þ
ð3:3Þ
and the calculated lengths using this equation are L1 ¼ 0:159 m; L2 ¼ 0:111 m; L3 ¼ 0:065 m Recalling that the product mL for these cases is 2.18, 3.13 and 5.36, respectively, based on the length L. The assumption of a very long ﬁn (mL 5) for the abovecalculated values leads to the greatest error for the ﬁrst case (mL = 2.18), a small error for the second case (mL = 3.13) and no error for the third case (mL = 5.36). Thus, it can be concluded that the assumption of a very long ﬁn can lead to a signiﬁcant error in the results if product mL is not 3. Example 3.4 An aluminium ﬁn [k = 200 W/(m K)] is 3.0 mm in thickness and 100 mm in width. It protrudes 75 mm from a wall at temperature of 200°C. The surrounding temperature is 20°C. Calculate heat loss from the ﬁn if the surface heat transfer coefﬁcient is 10 W/(m2 K). Solution qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ þ 2dÞ ¼ hð2W ¼ 102061000 2001003 ¼ 5:86 kWd mL ¼ 5:86 0:075 ¼ 0:4395 h=mk ¼ 10=ð5:86 200Þ ¼ 8:532 103 :
m¼
qﬃﬃﬃﬃﬃ hP kAc
3.4 Composite Fins
129
(a) From Eq. (3.8) of the short ﬁn,
qfin
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ =mkÞ þ tanh mL qfin ¼ hPkAc ðts t1 Þ ðhðhLL=mkÞ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ tanh mL þ 1 8:532103 þ tanhð0:4395Þ ¼ 10 206 103 200 100 3 106 ð200 20Þ 8:53210 3 tanhð0:4395Þ þ 1 ¼ 26:6 W:
(b) Using corrected length Lc in Eq. (3.6), we have qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ tanh mLc
where Lc = L + d/2 = 0.075 + 0.0015 = 0.0765 m. This gives mLc ¼ 5:86 0:0765 ¼ 0:44829 and qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ
10 206 103 200 100 3 106 ð200 20Þ tanhð0:44829Þ
¼ 26:61 W: The analysis clearly shows that even for mL = 0.44, the equation with corrected length can be used. Example 3.5 A 20mmdiameter and 500mmlong steel rod [k = 50 W/(m K)] is projecting through a furnace wall as shown in Fig. 3.7. The rod portion in wall may be assumed insulated. Determine the temperature of the rod just outside the furnace wall. Comment on the result.
t2
o
t1 = 250 C
t∞ = 20oC h = 10 W/(m2 K)
L = 275 mm
L1 = 225 mm
Fig. 3.7 Example 3.5
130
3
Extended Surfaces (Fins)
Solution Rod projecting outside the wall acts as a ﬁn of length L = 275 mm with base temperature t2. Fin parameter, rﬃﬃﬃﬃﬃﬃﬃ hP m¼ kAc sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 10 p 0:02 ¼ ¼ 6:324 50 ðp=4Þ 0:022 mL ¼ 6:324 0:275 ¼ 1:739\3: Since product mL is less than 3, the heat transfer from ﬁn will be calculated from the following equation: pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ qfin ¼ hPkAc ðt2 t1 Þ tanhðmLc Þ where Lc is the corrected length = L + d/4 for circular crosssection ﬁn. Hence, pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ h p d k ðp=4Þ d2 ðt2 t1 Þ tanh½mðL þ d=4Þ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 10 p 0:02 50 ðp=4Þ 0:022 ðt2 20Þ tanh½6:324 ð0:275 þ 0:02=4Þ ¼ 0:09375 ðt2 20Þ
qfin ¼
The heat flow by conduction through the rod length L1 equals the heat rejected by the ﬁn. Hence, t1 t2 kAc ¼ 0:09375 ðt2 20Þ L1 or 50 ðp=4Þ 0:022
250 t2 ¼ 0:09375 ðt2 20Þ 0:225
or t2 ¼ 118:17 o C: Example 3.6 Obtain an expression for the optimum thickness of a straight rectangular ﬁn. Use the equation of heat transfer for the ﬁn of Sect. 3.2.2. Solution The weight of the ﬁn shown in Fig. 3.8 is W ¼ qbLd where q is the density of the ﬁn material and b is the ﬁn width. Optimization here is to ﬁnd a combination of d and L for a given ﬁn width b and ﬁn area A1 = Ld.
3.4 Composite Fins
131 L δ
b
Fig. 3.8 Example 3.6
The heat flow through the ﬁn in consideration is given by qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ tanh mL
ð3:6Þ
Putting L = A1/d, parameter P 2b (for d b) and m = √(hP/kAc) = √(2 hb/kbd) = √(2 h/kd), the heat flow is qfin
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ b 2hkd tanh
! rﬃﬃﬃﬃﬃ 2h A1 ðts t1 Þ kd d
The heat flow will be maximum for dqﬁn/dd = 0. Differentiating the equation of qﬁn with respect to d and equating to zero, we get dqfin dd ¼ 1=2
0
! 3 rﬃﬃﬃﬃﬃ 2h A1 tanh þ7 6 kd d 2 7 6 7 6 rﬃﬃﬃﬃﬃ 7ðts t1 Þ ¼ 6 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 6 1 2h 3 5=2 7 5 4 b 2hkd A1 d qﬃﬃﬃﬃ
k 2 A1 2 2h cosh 2
pﬃﬃﬃﬃﬃﬃﬃﬃ d b 2hk
kd
Simpliﬁcation gives
d
2 qﬃﬃﬃﬃ
3 ! rﬃﬃﬃﬃﬃ A1 2h 3 kd d 2h A1 6 7 tanh ¼4 qﬃﬃﬃﬃ
5 kd d A 2 2h 1 cosh kd
qﬃﬃﬃﬃ Let
2h kd
d
Ad1 ¼ a, then tanhðaÞ ¼
3a cosh2 ðaÞ
Solution gives a = 1.4192. This gives L ¼ 0:71 d
rﬃﬃﬃﬃﬃ 2k hd
ð3:13Þ
132
3
Extended Surfaces (Fins)
Example 3.7 Show that for a long ﬁn of ﬁnite length L the heat transfer from the ﬁn surface to the surrounding is given by qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ coth mL
if at the ﬁn end t t∞. Solution From Eq. (3.2), h ¼ C1 emx þ C2 emx
ð3:2Þ
The constants C1 and C2 are to be determined from the boundary conditions of the problem. The boundary conditions for this case are (a) at x = 0, t = ts, and (b) at x = L, t = t∞. The ﬁrst condition gives ts t1 ¼ C1 þ C2
ðiÞ
and the second condition gives t1 t1 ¼ 0 ¼ C1 emL þ C2 emL or C1 emL þ C2 emL ¼ 0
ðiiÞ
Solution of Eqs. (i) and (ii) gives emL ðts t1 Þ emL emL emL C2 ¼ mL ðts t1 Þ e emL C1 ¼
Substitution of values of C1 and C2 in Eq. (3.2) gives ðt t1 Þ emðLxÞ emðLxÞ sinh½mðL xÞ ¼ ¼ ðts t1 Þ sinhðmLÞ emL emL
ð3:14Þ
Heat rejection rate from the entire surface area of the ﬁn qﬁn equals the heat flow by conduction at the ﬁn base, i.e. dt qfin ¼ kAc dx x¼o
3.4 Composite Fins
133
Using Eq. (3.3), we have qfin ¼ kAc
d sinh½mðL xÞ þ t1 ðts t1 Þ dx sinhðmLÞ x¼o
or qfin ¼ kAc
1 fcosh½mðL xÞgx¼o ðmÞðts t1 Þ sinhðmLÞ
or qfin ¼ kAc m
coshðmLÞ ðts t1 Þ sinhðmLÞ
or qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ coth mL
ð3:15Þ
Equation (3.4) can be obtained by putting L = ∞. Example 3.8 Show that for a ﬁn of ﬁnite length L the heat transfer from the ﬁn surface to the surrounding is given by t1 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ coshðmLÞ ttLs t 1 qfin ¼ hPkAc ðts t1 Þ sinhðmLÞ if at the ﬁn end temperature is tL. Other terms have usual meaning. Solution From Eq. (3.2), ts t1 ¼ h ¼ C1 emx þ C2 emx The constants C1 and C2 are to be determined from the boundary conditions of the problem. The boundary conditions for this case are (c) at x = 0, t = ts, and (d) at x = L, t = tL. The ﬁrst condition gives ts t1 ¼ C1 þ C2 or hs ¼ C1 þ C2 and the second condition gives tL t1 ¼ C1 emL þ C2 emL
ðiÞ
134
3
Extended Surfaces (Fins)
or hL ¼ C1 emL þ C2 emL
ðiiÞ
Solution of Eqs. (i) and (ii) gives hL =hs emL mL mL e e hL =hs emL C2 ¼ hs 1 mL e emL C1 ¼ hs
Substitution of values of C1 and C2 in Eq. (3.2) gives h ¼ hs ¼
hL =hs emL mx hL =hs emL mx e þ 1 mL e emL emL e emL ðhL =hs Þðemx emx Þ þ ½emðLxÞ emðLxÞ emL emL
or t t1 ¼ ts t1
tL t1 ts t1
sinh mx þ sinh mðL xÞ
ð3:1Þ
sinh mL
Heat rejection rate from the entire surface area of the ﬁn qﬁn equals the heat flow by conduction at the ﬁn base, i.e. dt qfin ¼ kAc dx x¼o Using Eq. (3.1), we have ( qfin ¼ kAc
" #) tL t1 d ts t1 sinh mx þ sinh mðL xÞ þ t1 ðts t1 Þ dx sinh mL
x¼o
or qfin
1 tL t1 ¼ kAc cosh mx cosh½mðL xÞ ðmÞðts t1 Þ sinhðmLÞ ts t1 x¼o
or qfin ¼ kAc m
t1 coshðmLÞ ttLs t 1
sinhðmLÞ
ðts t1 Þ
or qfin ¼
t1 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ coshðmLÞ ttLs t 1 hPkAc ðts t1 Þ sinhðmLÞ
ð3:2Þ
3.4 Composite Fins
135
Check: If tL = t∞, Eq. (3.2) gives qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ coth mL
which is Eq. (3.15).
3.5
Effectiveness and Efficiency of Fins
To study the performance of a ﬁn in transferring the heat, a term called ﬁn effectiveness eﬁn is used. It is a ratio of heat transfer with the ﬁn to the heat that would be transferred without ﬁn. Effectiveness of a ﬁn must be always greater than 1. Mathematically, it can be expressed as qfin ð3:16Þ efin ¼ hAc ðts t1 Þ Thus for a very long ﬁn (Sect. 3.2.1), it is rﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ Pk efin ¼ ¼ hAc ðts t1 Þ hAc
ð3:17Þ
From the above result, it can be concluded that the ﬁn effectiveness is high when (i) the ﬁn is made of high thermal conductivity material, (ii) the ﬁn has high ratio of parameter to the base area (P/Ac) and (iii) the convective heat transfer coefﬁcient is low. The third conclusion indicates that the ﬁns are very effective in transferring heat from surfaces with low heat transfer coefﬁcient. Another term that is used to assess the performance of a ﬁn in transferring heat is ﬁn efﬁciency ηﬁn, which is deﬁned as gfin ¼
qfin qideal
ðiÞ
where qideal is the heat that would be transferred if the entire ﬁn surface were at the ﬁn base temperature. That is, qfin ðiiÞ gfin ¼ hAf ðts t1 Þ where Af (=PL) is the ﬁn surface area. Considering the ﬁn of Sect. 3.2.2 (a ﬁn with negligible heat transfer from its end as compared to its surface), the ﬁn efﬁciency is
136
3
gfin
Extended Surfaces (Fins)
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ tanh mL ¼ hPLðts t1 Þ rﬃﬃﬃﬃﬃﬃﬃ kAc tanh mL ¼ L hP tanh mL ¼ mL
Mathematically, for small values of mL, tanh mL ðmLÞ2 ¼1 mL 3
ðiiiÞ
tanh mL 1 mL mL
ðivÞ
while for large values of mL,
tanh mL tends to the limiting value of mL unity. From Eq. (iv), and also from the variation of tanh(mL) with mL presented in Table 3.1, it tanh mL can be seen that the ﬁn efﬁciency gfin ¼ decreases rapidly as mL increases. TypimL cally at mL = 1.0, ηﬁn = 71.63%; for mL = 2.0, it is 48.2% and it reduces to 25% at mL = 4.0. For the trivial case of mL = 0, when L = 0 (i.e. no ﬁn at all), the ﬁn efﬁciency reaches its maximum value of unity. This situation is meaningless. Thus, it may be concluded that the ﬁn efﬁciency cannot be maximized with respect to its length. Generally, the ﬁn performance is maximized with respect to its mass, volume or cost. This maximization process has economic meaning also. As regards the parameter m, a small value of this parameter leads to a high ﬁn efﬁciency, which is achieved when From Eq. (iii), it can be seen that as mL tends to zero,
(i) ﬁn is made of high thermal conductivity material, (ii) ﬁn has low P/Ac ratio, i.e. a thick rectangular, square or circular crosssection ﬁn, and (iii) the convection heat transfer coefﬁcient is low. The ﬁrst and third conclusions are the same for both ﬁn effectiveness and efﬁciency. It means that for high effectiveness and efﬁciency, the ﬁns must be made of high thermal conductivity materials and must be provided on surfaces with low value of heat transfer coefﬁcient. It will be shown by taking an illustrative example that the ﬁns may reduce the heat transfer rate in applications where very high values of heat transfer coefﬁcient are encountered. The second conclusion for high ﬁn efﬁciency, regarding the ratio of ﬁn parameter to crosssectional area, is opposite to that for high effectiveness. Since the basic purpose of the installation of ﬁns on a surface is to enhance the heat transfer rate, the ﬁn effectiveness is a more meaningful parameter for assessing the usefulness of the ﬁns. The ﬁn effectiveness can be related to the ﬁn efﬁciency. From Eq. (ii), qfin ¼ gfin hAf ðts t1 Þ Substitution in Eq. (3.16) gives
3.5 Effectiveness and Efficiency of Fins
137 Af
Fin surface
Bare surface Ao
Fig. 3.9 Surface with multiple ﬁns
efin ¼
gfin hAf ðts t1 Þ hAc ðts t1 Þ
or efin
Af ¼ gfin Ac
ð3:18Þ
It means that the ﬁn effectiveness equals the ﬁn efﬁciency multiplied by the ratio of ﬁn surface and crosssectional areas. The analysis given above applies only to a single ﬁn. Usually, a surface with a number of ﬁns consists of bare portion between the ﬁns, refer to Fig. 3.9. In such cases, the total heat transfer from the surface is calculated by combining the heat flow from the bare portion of the surface with the heat flow from the ﬁns. Accordingly, the efﬁciency of the ﬁnned surface can be deﬁned as the ratio of the total heat transfer rate of the combined area of the bare surface Ao and that of the ﬁn Af to the heat which would be transferred if this total area were maintained at the ﬁn base temperature ts. The total heat flow can be expressed for a surface with multiple ﬁns as qtotal ¼ qfin þ qo ¼ hAf ðts t1 Þgfin þ hAo ðts t1 Þ
ð3:19aÞ
¼ hðgfin Af þ Ao Þðts t1 Þ The total heat flow can also be found in terms of the effectiveness. In this case qtotal ¼ qfin þ qo ¼ efin hAb ðts t1 Þ þ hAo ðts t1 Þ ¼ hðefin Ab þ Ao Þðts t1 Þ where Ab is the sum of the base area of ﬁns = number of ﬁns Ac.
ð3:19bÞ
138
3
Extended Surfaces (Fins)
The overall efﬁciency of the ﬁnned surface can be expressed as gtotal ¼
hAf ðts t1 Þgfin þ hAo ðts t1 Þ hðAf þ Ao Þðts t1 Þ
or gtotal ¼
Af gfin þ Ao Af gfin þ Ao ¼ ðAf þ Ao Þ Afw
where Afw = total surface area of the ﬁnned wall = Af + Ao. Putting Ao = Afw − Af, we get gtotal ¼
Af gfin þ Afw Af Afw
gtotal ¼
Afw ð1 gfin ÞAf Afw
or
or gtotal ¼ 1
Af ð1 gfin Þ Afw
ð3:19cÞ
In Eq. (3.19a, 3.19b, 3.19c), the values of the heat transfer coefﬁcient from the ﬁns and bare portion of the wall have been assumed to be equal. In fact they differ. Let the values of these coefﬁcients are hw and hf for the base portion and the ﬁns, respectively. Then the heat flow rate from the wall is qtotal ¼ qfin þ qo ¼ hf Af ðts t1 Þgfin þ hw Ao ðts t1 Þ
ðiÞ
Let qtotal ¼ hred AfW ðts t1 Þ
ðiiÞ
where Afw = Af +Ao is the total area of the surface transferring the heat. Comparing Eqs. (i) and (ii), we get hred ¼ hf gfin
Af Ao þ hw Afw Afw
ð3:20Þ
The coefﬁcient of heat transfer hred is called the reduced heat transfer coefﬁcient and is a mean or effective heat transfer coefﬁcient for a ﬁnned wall, which accounts for the heat removal from the bare portion of the wall between the ﬁns, ﬁn surface and ﬁn efﬁciency.
3.6 Heat Transfer from a Finned Wall
139
q
tf1
tf2
h1, tw1 h2, tw2
Fig. 3.10 A ﬁnned wall
3.6
Heat Transfer from a Finned Wall
The heat flow through a ﬁnned wall, with plane side at temperature tw1 and the ﬁnned side at temperature tw2, can be described by the following set of equations (Fig. 3.10): q ¼ h1 A1 ðtf 1 tw1 Þ k q ¼ A1 ðtw1 tw2 Þ d q ¼ hred Afw ðtw2 tf 2 Þ Reducing the equation to q ¼ A1
1 h1
ðtf 1 tf 2 Þ
A1 1 þ dk þ hred Afw
ð3:21Þ
¼ Uðtf 1 tf 2 Þ where U is the overall heat transfer coefﬁcient for heat flow through a ﬁnned wall and is deﬁned as U¼
1 1 h1
þ
d k
þ
1 hred
A1 Afw
ð3:22Þ
The ratio of the area of the ﬁns’ surface Afw to the base surface area A1 is called as ﬁnning factor.
3.7
Intensification of Heat Transfer by Finning
Use of ﬁn on a heat transfer surface is an effective method of heat transfer enhancement when the heat transfer coefﬁcient for this wall is much smaller compared to the other side. This is illustrated by a numerical problem given below. For heat transfer through any wall, the ﬁlm resistance is 1/hA. Thus, the ﬁlm resistance depends not only on the heat transfer coefﬁcients but also on the size of the surfaces represented by area A. Hence, the ﬁlm resistance can be decreased by increasing the surface area by ﬁnning.
140
3
Extended Surfaces (Fins)
Example 3.9 For a wall, h1 = 1000 W/(m2 K) and h2 = 10 W/(m2 K). Determine the enhancement of heat transfer by providing ﬁns on the wall with the lower heat transfer coefﬁcient. The ﬁnning factor is 3 and hred= h2. Solution The overall heat transfer coefﬁcient for the surface without ﬁns is (neglecting wall resistance) U¼
1 1000
When surface is ﬁnned, U¼
1 1000
1 þ
1 10
¼ 9:90 W= m2 K
1 ¼ 29:13 W= m2 K 1 1 þ 10 3
Example 3.10 The following data refer to a heat exchanger, which transfers heat from a heated liquid to air: Liquidside surface area, A1 = 2 m2 Airside ﬁnned surface area, Af = 7 m2 Airside total area, Afw = 10 m2 Fins: Length = 15 mm Thickness = 0.5 mm Material: steel [k = 45 W/(m K)] Heat transfer coefﬁcients: Liquid side, hl = 250 W/(m2 K) Air side, ha = 50 W/(m2 K). Determine the overall heat transfer coefﬁcient U based on the liquidside area. Neglect the wall resistance to conduction heat flow. Compare the value with surface without ﬁns. Solution The overall heat transfer coefﬁcient for a ﬁnned wall is given by (Eq. 3.22) U¼
1 1 h1
þ
þ
d k
1 hred
A1 Afw
Neglecting wall resistance, we have U¼
1 1 h1
ðiÞ
Af Ao þ hw Afw Afw
ðiiÞ
þ
1 hred
A1 Afw
where hred ¼ hf gfin
3.7 Intensification of Heat Transfer by Finning
141
Here hf = hw = 50, Af = 7 m2, Afw = 10 m2, Ao = Afw − Af = 10 – 7 = 3 m2 and let gfin ¼
tanh mL mL
ð3:18Þ
Putting m = √(2h/kd) = √(2 50 1000/45 0.5) = 66.67, we have gfin ¼
tanhð66:67 0:015Þ ¼ 0:7616 66:67 0:015
Substituting the values of various terms in Eq. (ii) gives hred ¼ 50 0:7616
7 3 þ 50 ¼ 41:65 W= m2 K 10 10
And this gives U¼
¼ 113:61 W= m2 K
1 1 250
þ
1 2 41:65 10
For the wall without ﬁns, the overall heat transfer coefﬁcient is U¼
1 1 þ hl ha
1
1 1 þ 250 50
¼
1
¼ 41:67 W= m2 K
Example 3.11 A ﬁnned tube of heat exchanger, shown in Fig. 3.11, carries hot water which is cooled by blowing air over the ﬁns. Determine the overall heat transfer coefﬁcient. Solution From Eqs. (3.22) and (2.29a), overall heat transfer coefﬁcient for the present case is Ui ¼
1 hi
þ
1
R0 Ai 2pkL ln Ri
þ
1 hred
Ai Afw
where hred is reduced heat transfer coefﬁcient given by
δ
ri ro
Fig. 3.11 A tube with longitudinal ﬁns
Lf
Di = 20 mm Do = 25 mm δ = 2 mm Lf = 20 mm k = 40 W/(m K) hi = 5000 W/(m2 K) ho = 100 W/(m2 K) N = 12
142
3
hred ¼ hf gfin
Extended Surfaces (Fins)
Af Ao þ hw Afw Afw
ð3:20Þ
For the given thin rectangular section ﬁns (P 2b), rﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP h 2b 2h 2 100 m¼ ¼ ¼ ¼ ¼ 50 kAc kbd kd 40 2=1000 mLf ¼ 50
20 ¼ 1:0: 1000
Fin efﬁciency, neglecting heat transfer from ﬁn end, gfin ¼
tanh mLf tanh 1:0 ¼ 0:762 ¼ 1:0 mLf
Fin surface area for unit length of tube, Af ¼ Nð2Lf þ dÞ ¼ 12 ð2 20 þ 2Þ=1000 ¼ 0:504 m2 Total outside heat transfer area for unit length of tube, Afw ¼ Af þ ðpDo NdÞ ¼ 0:504 þ ðp 25 12 2Þ=1000 ¼ 0:5585 m2 Hence, for hf = hw = ho = 100 W/(m2 K), the reduced heat transfer coefﬁcient hred ¼ 100 0:762
0:504 0:5585 0:504 þ 100 ¼ 78:52 0:5585 0:5585
Substitution gives the overall heat transfer coefﬁcient as Ui ¼
1
1 5000
þ
pð20=1000Þ1 25 1 2p401 ln 20 þ 78:52
¼ 592:2 W= m2 K
p20=10001 0:5585
If ﬁns are not provided, Ui ¼ ¼
1 hi
þ
1
R0 Ai 2pkL ln Ri
þ
1 ho
Ai Ao
1
1 5000
þ
pð20=1000Þ1 25 1 20 2p401 ln 20 þ 100 25 2
¼ 121:13 W= m K
The effect of ﬁns on heat transfer intensiﬁcation can be seen. Example 3.12 One end of a very long aluminium rod [k = 205 W/(m2 K)] is in contact with a heated wall. Its surface is in contact with a cold fluid. Neglect the contact resistance between the rod and the wall.
3.7 Intensification of Heat Transfer by Finning
143
(i) By what percentage the rate of heat removal will increase if the rod diameter is doubled? (ii) If the rod is made of stainless steel [k = 15 W/(m2 K)], by what percentage the heat transfer rate will change? Solution Using the condition of inﬁnite length ﬁn, pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ qfin ¼ hPkAc ðts t1 Þ / d1=2 k1=2 ðd 2 Þ1=2 ¼ d3=2 k1=2 / P1=2 k1=2 A1=2 c (i) When the diameter of the rod is doubled, q0fin / ð2dÞ3=2 k1=2 Increase in the heat transfer rate is Dq0fin ¼
q0fin qfin ð2dÞ3=2 d 3=2 ¼ 100 ¼ 182:8% qfin d3=2
(ii) When the material of the rod is changed to stainless steel, 1=2
q0fin / d3=2 k1 Change in the heat transfer rate is Dq0fin ¼
q0fin qfin ðk1 Þ1=2 k1=2 ð15Þ1=2 ð205Þ1=2 ¼ 100 ¼ 100 ¼ 72:95% qfin k1=2 ð205Þ1=2
Example 3.13 Two long 10mmdiameter copper rods [k = 360 W/(m K)] are to be soldered together end to end. The melting point of the solder is 650°C. If the heat transfer coefﬁcient between the copper rod surface and air is 5 W/(m2 K), and the surrounding air temperature is 25°C, determine the minimum heat input rate to keep the soldered surfaces at 650°C. Solution Treating the junction of the two rods as x = 0 plane (refer to Fig. 3.12), the problem is that of a very long ﬁn with Ts = 650°C. The required heat input rate will be twice the heat transfer rate by conduction into the ﬁn, i.e. pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ q ¼ 2qfin ¼ 2 hPkAc ðts t1 Þ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 2 5 p 0:01 360 ðp=4Þ 0:012 ð650 25Þ ¼ 83:3 W
144
3
Extended Surfaces (Fins)
650oC qfin
qfin
θ=0
x x=0
Fig. 3.12 Example 3.13
Example 3.14 Two long rods of the same diameter, one made of brass [k = 85 W/(m K)] and the other of copper [k = 375 W/(m K)], have one of their ends inserted into a furnace, refer to Fig. 3.13. Both rods are exposed to the same environment. At a plane 105 mm away from the furnace wall, the temperature of the brass rod is 120°C. At what distance from the furnace end, the same temperature would be reached in the copper rod? Solution For a long ﬁn, the equation of temperature distribution is t ¼ ðts t1 Þemx þ t1
ðiÞ
For the brass rod, t = 120°C at x = 0.105 m, i.e. 120 ¼ ðts t1 Þe0:105m1 þ t1
ðiiÞ
For the copper rod, Eq. (i) yields 120 ¼ ðts t1 Þem2 a þ t1
ðiiiÞ
Equating Eqs. (ii) and (iii), we get ðts t1 Þe0:105m1 þ t1 ¼ ðts t1 Þem2 a þ t1 or
0:105m1 ¼ am2
or m1 a ¼ 0:105 ¼ 0:105 m2
Furnace
ts
rﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃ! hP kAc kAc 1 hP
t =120oC
Brass rod
• 0.105 m
a
Fig. 3.13 Example 3.14
2
• t∞
Copper rod
3.7 Intensification of Heat Transfer by Finning
or
rﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃ! hP kAc a ¼ 0:105 kAc 1 hP
145
2
Since (hP/Ac)1 = (hP/Ac)2, rﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ k2 375 a ¼ 0:105 ¼ 0:105 ¼ 0:2205 m 85 k1 The desired answer is 220.5 mm. Example 3.15 Heat dissipation from a surface is to be enhanced by providing wide rectangular section ﬁns. Total of 200 ﬁns are to be provided in 1 m height. Each ﬁn is 1 mm thick and 50 mm long. The convection heat transfer coefﬁcient for the surface without ﬁns is h = 20 W/(m2 K), which drops to hf = 15 W/(m2 K) when ﬁns are installed on the surface. Calculate the enhancement in heat transfer if the thermal conductivity of the ﬁn material is 210 W/(m K). Solution Heat dissipation from the wall without ﬁns, qo ¼ ho Ac ðts t1 Þ ¼ 20 1 b ðts t1 Þ ¼ 20ðbhs Þ W where hs = ts − t∞ and b = width of the ﬁn. Heat dissipation from ﬁnned wall: Surface area without ﬁns (bare surface area), Ab = 1 b – b 1/1000 200 = 0.8b m2. Contribution of the bare surface, qb ¼ hb Ab hs ¼ 15 0:8b hs ¼ 12ðbhs Þ W From the given data for the ﬁn, rﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP hð2bÞ 2h 2 15 1000 ¼ ¼ m¼ ¼ ¼ 11:95 kAc kðbdÞ kd 210 1 Assuming parameter P 2b for rectangular section ﬁns, mL ¼ 11:95 0:05 ¼ 0:5975 Using corrected length Lc in Eq. (3.6), we have qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ tanh mLc
146
3 L
L 1
Wall
Extended Surfaces (Fins)
8
2
Wall 16
Fig. 3.14 Example 3.16
where Lc = L + d/2 = 0.05 + 0.001/2 = 0.0505 m. This gives qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ
15 2b 210 b 1 103 hs tanhð11:95 0:0505Þ ¼ 1:354ðbhs Þ W
Total heat transfer qtotal from the ﬁned wall is qtotal ¼ N qfin þ qb ¼ 200 1:354ðbhs Þ þ 12ðbhs Þ ¼ 282:8ðbhs Þ W Heat transfer enhancement, qtotal qo 282:8ðbhs Þ 20ðbhs Þ 100 ¼ 1314% 100 ¼ 20ðbhs Þ qo
Example 3.16 In order to enhance heat transfer from a heat exchanger surface at 90°C, 40 mm long ﬁns are to be installed. There are two alternative schemes from equal weight consideration (Fig. 3.14): (i) 1mmthick plate ﬁns at 8 mm pitch, or (ii) 2mmthick plate ﬁns at 16 mm pitch. The ﬁns are to be made of aluminium [k = 205 W/(m K)]. Assuming surface heat transfer coefﬁcient h = 10 W/(m2 K) and t∞ = 20°C, select the better arrangement based on the heat transfer consideration. Solution Heat transfer from a ﬁn of ﬁnite length is given by qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ tanh mLc
For both the arrangements, h ¼ 10 W= m2 K ; P 2b; k ¼ 205 W= m2 K ; Ac ¼ b d; and ðts t1 Þ ¼ 70 C
3.7 Intensification of Heat Transfer by Finning
147
Substituting the above values, heat transfer equation for unit width becomes qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 10 2 1 205 1 d ð70Þ tanh mLc
or qfin ¼ 4482 Case (i) m¼
qﬃﬃﬃﬃﬃ hP kAc
¼
qﬃﬃﬃﬃﬃﬃﬃﬃ hð2bÞ kðdbÞ
¼
pﬃﬃﬃ d tanh mLc
qﬃﬃﬃﬃ 2h kd
¼
qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2101000 2051
ðiÞ
¼ 9:877
Lc ¼ L þ d=2 ¼ 40 þ 0:5 ¼ 40:5 mm mLc ¼ 9:877 40:5=1000 ¼ 0:4 Substitution in Eq. (i) gives qfin ¼ 4482 Case (ii) m¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1=1000 tanhð0:4Þ ¼ 53:86 W per m width:
qﬃﬃﬃﬃﬃ hP kAc
¼
qﬃﬃﬃﬃﬃﬃﬃﬃ hð2bÞ kðdbÞ
¼
qﬃﬃﬃﬃ 2h kd
¼
qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2101000 2052
¼ 6:984
Lc ¼ L þ d=2 ¼ 40 þ 1 ¼ 41 mm mLc ¼ 6:984 41=1000 ¼ 0:2864:
Substitution in Eq. (i) gives qfin ¼ 4482
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2=1000 tanhð0:2864Þ ¼ 55:9 W per m width:
Since two 1 mm thick ﬁns can be installed in one pitch space for 2 mm thick ﬁns, the ﬁrst arrangement of 1 mm ﬁns will be better. Example 3.17 In the previous example, what is the overall effectiveness of the ﬁrst arrangement? Solution In 1.0 m height of the surface, there will be 1000/8 = 125 ﬁns. The bare surface in between the ﬁns will be 125 7/1000 = 0.875 m in height. The total heat transfer from the surface would be qtotal ¼ qfin þ qo ¼ 125 53:86 þ hAo ðts t1 Þ ¼ 6732:5 þ 10 0:875 70 ¼ 7345 W/m2 The overall effectiveness of the ﬁnned surface is the ratio of the heat transfer from the surface with ﬁns to the heat transfer from the surface without ﬁns, i.e. e¼
qtotal 7345 ¼ 10:5 ¼ hAðts t1 Þ 10 1 70
148
3
Extended Surfaces (Fins)
Example 3.18 A 12.5mmdiameter aluminium rod [k = 200 W/(m K)] is exposed to an environment at 30°C, while the base temperature is 100°C. The heat transfer coefﬁcient h = 10 W/(m2 K) for the surface as well as end of the ﬁn. Determine the total heat transfer rate from the ﬁn if its length is 25, 50, 100, 150 or 200 mm. Also determine the effectiveness of the ﬁn for all these lengths and comment on the result. Solution The value of parameter m is rﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP hpd 4h 4 10 m¼ ¼ ¼ ¼ 4:0 m1 ¼ kAc kðp=4Þd 2 kd 200 0:0125 The maximum value of product mL = 0.8, which is less than 3.0. Hence, the heat transfer rate will be calculated considering it to be a short ﬁn, i.e. qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hL =mk þ tanh mL hPkAc ðts t1 Þ hL =mk tanh mL þ 1
Here, hL 10 ¼ 0:0125 ¼ mk 4 200 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ ¼ 10 p 0:0125 200 ðp=4Þ 0:01252 70 ¼ 6:871
Substitution gives qfin ¼ 6:871
0:0125 þ tanhð4LÞ 0:0125 tanhð4LÞ þ 1
ðiÞ
and qfin qfin ¼ hAc ðts t1 Þ h p=4 d2 ðts t1 Þ qfin ¼ 11:64qfin ¼ 10 p=4 0:01252 70
e¼
ðiiÞ
Length of ﬁn, L (m)
mL
tanhmL
qﬁn (W)
e
0.025
0.1
0.0997
0.77
8.96
0.05
0.2
0.1974
1.438
16.74
0.1
0.4
0.38
2.684
31.24
0.15
0.6
0.537
3.75
43.65
0.2
0.8
0.664
4.61
53.66
The values of heat transfer rate and effectiveness of the ﬁn as calculated from Eqs. (i) and (ii) for different ﬁn lengths are tabulated.
3.7 Intensification of Heat Transfer by Finning
149
The effectiveness increases with the length of the ﬁn. It is to be noted that while the length of the ﬁn and hence its weight for the given crosssection of the ﬁn increases in the present example by 8 times, the effectiveness increases by 6 times only. Thus the shorter ﬁns are more economical. Note: For an inﬁnitely long ﬁn, qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ
Substitution of values of various terms gives qfin ¼ 6:87 W For mL = 3, we get qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ tanh mL ¼ 6:84 W
The heat transfer rate for a ﬁn with mL = 3 is only 0.5% lower than for a ﬁn of inﬁnite length and 8.9 times of the ﬁn with 25 mm length. For mL = 3, L is 3/m = ¾ = 0.75 m, i.e. the length is 30 times of the minimum length of 25 mm in this example while the effectiveness is only 8.9 times. Thus, it can be seen that the short ﬁns are economical.
3.8
Error in Temperature Measurement with Thermometer Well
The arrangement shown in Fig. 3.15 is used to measure the temperature of fluid flowing through a pipe or duct. The thermometer well is a thin hollow metallic or ceramic cylinder. The metallic cylinder may be welded to the pipe wall. Thermometer to measure the temperature is dipped into the oil ﬁlled in the well and thus reads the temperature at the end of the well. It will be shown by the ﬁn analysis that the thermometer does not indicate the actual temperature of the fluid flowing through the duct. The thermometer well can be regarded as a hollow ﬁn with the end at temperature tL and the base at temperature ts. The actual temperature of the fluid in the duct is t∞. Assuming negligible heat transfer from the ﬁn end as compared to the heat transferred from the ﬁn surface, we have (refer to Sect. 3.2.2) t t1 cosh mðL xÞ ¼ cosh mL ts t1
ð3:5Þ
Thermometer well ts Fluid at t∞
L
δ do
Fig. 3.15 A thermometer well
x
tL
150
3
Extended Surfaces (Fins)
Putting t = tL at x = L, we get tL t1 ¼ ðts t1 Þ
1 cosh mL
ð3:23Þ
For a hollow ﬁn, rﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ hP hpdo h m¼ ¼ ¼ kAc kd kpdo d
ð3:24Þ
For the error (t∞ − tL) to be minimum, cosh(mL), i.e. m and L must be large. This can be achieved by the following measures: (i) The well must be made of low thermal conductivity material, such as ceramic; (ii) The well should be as thin as possible. However, it must have sufﬁcient structural strength and must be able to withstand the fluid pressure and corrosion during the service. (iii) The well should be as large as possible in length. In a given pipe diameter, the length can be increased by installing the well in an inclined position. (iv) The heat transfer rate from the fluid to the well can be effectively increased by providing ﬁns on the outer surface of the well, refer to Fig. 3.16. The thermometer indicates a temperature lower than the gas temperature ﬁrstly because of the temperature drop in the ﬁlm formed at the well surface and secondly due to the conduction of the heat through the wall of the well towards the pipe surface which is at a lower temperature. The resistance to the heat flow from the fluid to the well can be reduced by providing ﬁns or increasing the heat transfer coefﬁcient. The heat flow by conduction to the pipe wall can be reduced by increasing the conduction resistance, i.e. by increasing the length of the heat flow path, reducing the heat flow area and using a low thermal conductivity material for the well. The conduction heat flow can also be reduced by reducing the temperature difference between the well end and the pipe surface. This can be achieved by insulating the pipe from the outside, which will increase the pipe surface temperature. Note: In the abovepresented analysis, we have not considered radiation heat exchange between the surface of the well and the pipe wall.
Well with fins
Fluid
Fig. 3.16 A thermometer well with ﬁns
3.8 Error in Temperature Measurement with Thermometer Well
151
Example 3.19 Temperature of air flowing through a pipe is measured with a mercuryinglass thermometer placed in a steel well ﬁlled with oil. How great is the error in the measurement if the thermometer indicated a temperature of 100°C? The temperature of the base of the well is 60°C. The well is 120 mm long, and the thickness of the wall of the well is 1.0 mm. The thermal conductivity of the well material is 45 W/(m K), and air to the well surface heat transfer coefﬁcient is 30 W/(m2 K). Solution From Eq. (3.23), tL t1 ¼ ðts t1 Þ Here,
L tL ts
1 cosh mL
rﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ h 30 m¼ ¼ ¼ 25:82 kd 45 1=1000
= 0.12 m = 100°C = 60°C. Substitution in the equation gives 100 t1 ¼ ð60 t1 Þ
1 coshð25:82 0:12Þ
or t1 ¼ 103:96 o C Thus the error in measurement of the temperature is (t∞ − tL) = 3.96°C. Note: Radiation heat exchange between the well surface and the pipe inner surface may introduce additional error.
3.9
When Fins Are to Be Used?
The installation of ﬁns on a heattransferring surface does not necessarily increase the heat transfer rate. In fact, there are certain conditions when the ﬁns on a surface may cause a reduction in the heat transfer. Here, we shall discuss the conditions when the ﬁns are useful. In Sect. 3.5, ﬁn effectiveness has been deﬁned. The equation of effectiveness can be written for a very long ﬁn (Sect. 3.2.1) as efin
rﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ Pk hP k mk 1 ¼ ¼ ¼ ¼ ¼ hAc ðts t1 Þ hAc kAc h h ðh=mkÞ
ðiÞ
When the effectiveness is unity, the heat transfer rate with or without ﬁns is the same. From the above equation, we can see that the combined term (h/mk) is unity in this case. It is
152
3
Extended Surfaces (Fins)
to note that this conclusion is also applicable to the short ﬁn of Sect. 3.2.3. The combined term (h/mk) has special signiﬁcance for the ﬁns. We can write h h ¼ mk k
rﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃ0 kAc h Ac hd ¼ ¼ k P hP k
ðiiÞ
where d′ = Ac/P = is a linear dimension. It equals D/4 for a circular crosssection ﬁn and d/2 for a thin rectangular crosssection ﬁn. The combined term (hd′/k) is a nondimensional term and is known as Biot number (Bi). That is, Bi ¼
hd0 k
ð3:25Þ
The Biot number compares the relative magnitudes of internal resistance to conduction heat flow and the external (surface) resistance to convection heat transfer. A low value of the Biot number means that the internal resistance to conduction heat transfer is small in comparison to the external one to the convection heat transfer. In terms of the Biot number, the heat transfer rate from Eq. (3.8) for the short ﬁn is pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ Bi þ tanh mL qfin ¼ hPkAc ðts t1 Þ pﬃﬃﬃﬃﬃ ð3:26Þ Bi tanh mL þ 1 and, from Eq. (3.4) of Sect. 3.2.1, qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ
¼ hAc ðts t1 Þ=ðh=mkÞ pﬃﬃﬃﬃﬃ ¼ hAc ðts t1 Þ= Bi
ð3:27Þ
Now we discuss the effect of the Biot number on the heat transfer rate. (a) When Bi = 1, qfin ¼ hAc ðts t1 Þ which means the heat transfer with ﬁns qﬁn is the same as heat transfer without ﬁns. It means that when Bi = 1 the ﬁns do not contribute to heat transfer from the surface, and hence there is no advantage of ﬁns. (b) When Bi > 1, qfin \hAc ðts t1 Þ which means that ﬁnned surface heat transfer qﬁn is less than the heat transfer without ﬁns, i.e. the ﬁns will act as insulator and heat transfer rate is reduced when the ﬁns are installed.
3.9 When Fins Are to Be Used?
153
(c) When Bi < 1, qfin [ hAc ðts t1 Þ In this case, heat transfer rate increases when the ﬁns are provided on the surface. From the above discussion of desirability of Bi < 1, the following can be concluded: (i) For a given value of the heat transfer coefﬁcient h, the ﬁn should be made of high conductivity material. (ii) For the given value of the heat transfer coefﬁcient and thermal conductivity of ﬁn material k, the ﬁn geometrical dimensions must be selected to give a high value of (P/Ac). One example of such a ﬁn is a thin and wide ﬁn. (iii) For the given value of thermal conductivity of ﬁn material k and ﬁn shape, as the value of heat transfer coefﬁcient h increases the ﬁn effectiveness or utility decreases and when hAc/P equals k, the ﬁn does not enhance the heat transfer rate. If h is so high that the Biot number is greater than unity, the ﬁns act as heat insulator and reduce the heat transfer rate. For surfaces in contact with condensing vapours or boiling liquids and liquid metals, the magnitude of heat transfer coefﬁcient h is very high and the ﬁns are not useful. In general, the heat transfer coefﬁcient is quite low for surfaces in contact with a gas at low velocities or in the condition of natural convection heat transfer. The ﬁns are very effective for such surfaces to enhance the heat transfer rate. The magnitude of the heat transfer coefﬁcient for the surfaces in contact with liquids is moderate to high and due attention must be given to the selection of material and shape of the ﬁn so that the Biot number is sufﬁciently less than unity. In the case of short and thick ﬁns, the assumption of onedimensional heat flow made for the analysis presented in the previous sections is not valid. The heat flow in such ﬁns becomes twodimensional as shown in Fig. 3.17 (note that the heat flow lines are not parallel to the axis of the ﬁn). This will have effect on the conclusion regarding the minimum value of the Biot number. To be sure, that the use of ﬁn in such cases is advantageous, it is advised to fulﬁl the following condition (Eckert and Drake 1972): Bi 0:2
ð3:28Þ
The effectiveness and efﬁciency deﬁned in Sect. 3.5 can be related to the Biot number as shown below.
Heat flow lines
Fig. 3.17 Temperature distribution and heat flow paths in a short rectangular ﬁn
154
3
Extended Surfaces (Fins)
For a very long ﬁn, the effectiveness of the ﬁn, from Eq. (3.17), can be expressed as rﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ kP k 1 e¼ ¼ 0 ¼ hAc Bi hd
ð3:29Þ
Thus the effectiveness increases as the Biot number decreases. Similarly, Eq. (3.18) of efﬁciency for the ﬁn of Sect. 3.2.2 (a ﬁn with negligible heat transfer from its end as compared to its surface) in the terms of the Biot number is gfin ¼
pﬃﬃﬃﬃﬃ tanh mL tanh L Bi=d0 ¼ pﬃﬃﬃﬃﬃ 0 mL L Bi=d
ð3:30Þ
The efﬁciency of the ﬁn in this case tends to its maximum value of unity as the Biot number tends to zero. Example 3.20 Three rods, one made of glass, one of pure aluminium and the third of steel, each having diameter of 10 mm and a length of 250 mm, are used as ﬁns. When the base temperature is 180°C for each ﬁn and ambient temperature 30°C, ﬁnd the distributions of temperature in the rods and their heat dissipations. The convective heat transfer coefﬁcient is 25 W/(m2 K). Determine the effectiveness and efﬁciency of these ﬁns. The thermal conductivities of the three materials are 0.8, 200 and 50 W/(m K), respectively. Solution The three rods projecting from the surface work as ﬁns. The value of the parameter m for the circular section rod is rﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP 4h 4 25 100 m¼ ¼ ¼ ¼ pﬃﬃﬃ kAc kd k 0:01 k Therefore, the value of parameter m for the glass ﬁn mg, aluminium ﬁn ma and steel ﬁn ms is 111.8, 7.07 and 14.14, respectively. The product mL for these ﬁns is 27.95, 1.768 and 3.535, respectively. For the glass and steel ﬁns, mL 3; hence, equation of Case (A) may be used. For the aluminium ﬁn, equations of Case (B) with end correction may be used. The equations of temperature distribution are Glass ﬁn: t t1 ¼ emx ts t1 h ¼ e111:8x ho Aluminium ﬁn:
h cosh mðLc xÞ ¼ ho cosh mðLc Þ
3.9 When Fins Are to Be Used?
155
Here Lc = L + d/4 = 0.2525 m, cosh(mLc) = 3.064. This gives h ¼ 0:3263 coshð1:785 7:07xÞ ho Steel ﬁn: h ¼ e14:14x ho Heat dissipation: Glass ﬁn: pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 25 ðp 0:01Þ 0:8 ðp=4Þ ð0:01Þ2 ð180 30Þ
q¼
¼ 1:054 W: Aluminium ﬁn: pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ tanhðmLc Þ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 25 ðp 0:01Þ 200 ðp=4Þ ð0:01Þ2 ð180 30Þ tanhð7:07 0:2525Þ
q¼
¼ 15:75 W Steel ﬁn: pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðts t1 Þ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 25 ðp 0:01Þ 50 ðp=4Þ ð0:01Þ2 ð180 30Þ
q¼
¼ 8:33 W Heat dissipation without ﬁn, q ¼ hAðts t1 Þ ¼ 25 ðp=4Þ ð0:01Þ2 ð180 30Þ ¼ 0:2945 W Effectiveness Glass ﬁn: e¼
1:054 ¼ 3:58 0:2945
e¼
15:75 ¼ 53:48 0:2945
Aluminium ﬁn:
156
3
Extended Surfaces (Fins)
Steel ﬁn: e¼
8:33 ¼ 28:28 0:2945
Efﬁciency The heat dissipation, if whole ﬁn surface is at the base temperature, q ¼ hAf ðts t1 Þ ¼ 25 ½p 0:01 0:25 þ ðp=4Þ ð0:01Þ2 ð180 30Þ ¼ 29:75 W This gives efﬁciency of the three ﬁns as 1:054 100 ¼ 3:54% 29:75 15:75 gAl ¼ 100 ¼ 52:94% 29:75 8:33 gs ¼ ¼ 28% 29:75 gg ¼
Despite the same length, crosssection and the base temperature, it can be seen that the effectiveness and efﬁciency of the aluminum ﬁn (a high thermal conductivity material) is very high. Example 3.21 Typical heat transfer coefﬁcient data is given in column 2 of Table 3.2. We have to investigate the utility of ﬁns for surfaces in contact with fluids given in column 1 of the table. Fins are to be made either of copper [k = 385 W/(m K)] or steel [k = 70 W/(m K)]. Fin geometric data is as follows: thickness d = 2 mm, width W = 100 mm and length L = 60 mm.
Table 3.2 Example 3.21 h, W/(m2 K)
Copper ﬁns, k = 385 W/(m K)
Steel ﬁns, k = 70 W/(m K)
e
Bi = hd’/k
e
Bi = hd’/k
60.1–57.4 52.7– 27.1
(1.3–5.2) 10−5 (1.3–13) 10−4
55.2–45.8 34.68–11.8
(7.14–28.5) 10−5 (7.14–71.4) 10−4
60.147.14 52.7–6.2
(1.3 − 26) 10−5 1.3 10−4  0.026
55.2–26.2 34.68–2.64
7.14 10−51.4 10−3 7.14 10−4 0.143
1. Free Convection Gases Liquids
5–20 50–500
2. Forced Convection Gases Liquids
5–100 50–10000
3. Condensation Gases Liquids
5000–15000 5 104 – 15 104
8.77–5.06 2.77–1.6
0.013–0.039 0.13–0.39
3.75–2.16 1.19–0.68
0.071–0.214 0.71–2.14
4. Boiling
500–50000
27.73–2.77
1.3 10−3  0.13
11.9–1.19
7.1 10−3  0.715
3.9 When Fins Are to Be Used?
157
Solution The ﬁn effectiveness e is given by Case (A): A very long ﬁn:
1 e ¼ pﬃﬃﬃﬃﬃ Bi
Case (B): A ﬁn with negligible heat rejection from its end: tanh mL 1 e ¼ pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ Bi Bi Case (C): A short ﬁn: e¼
if mL [ 3
pﬃﬃﬃﬃﬃ 1 þ ðtanh mLÞ= Bi pﬃﬃﬃﬃﬃ 1 þ Biðtanh mLÞ
From the given data, ﬁn crosssectional area A = Wd = 200 mm2 and ﬁn surface area Af = PL = 2(W + d)L = 12240 mm2. Thus A < < PL, i.e. the conditions of Case (A) and (B) may be applied. Further, the product mL for a rectangular section duct and given data is Copper ﬁn: rﬃﬃﬃﬃﬃ pﬃﬃﬃ 2h mL ¼ L ¼ 0:0967 h kd Steel ﬁn: pﬃﬃﬃ mL ¼ 0:2268 h From the above values of product mL, we can see that when h 960 for the copper ﬁn and h 175 for the steel ﬁn, the value of product mL 3 and the effectiveness equation of Case (A) can be used. Calculated values of effectiveness e and the Biot number for the different conditions are given along with the corresponding values of h in Table 3.2. It can be readily seen that the effectiveness decreased with the increase in h. For the given geometric dimensions of the ﬁn in the present example, the effectiveness e = 82.4 for h = 5 W/(m2 K) and is only 1.6 for the copper ﬁn when dropwise condensation occurs (h = 15 104) and even less than unity for the steel ﬁn in contact with the condensing vapour. The Biot number is a quite useful parameter to decide the utility of a ﬁn. It must be kept in mind that the Biot number depends on the thermal conductivity of the ﬁn material, heat transfer coefﬁcient and the ﬁn geometry represented by d’.
158
3.10
3
Extended Surfaces (Fins)
Heat Transfer from a Bar Connected to Two Heat Sources at Different Temperatures
In Fig. 3.18, thermal reservoirs 1 and 2 are connected with a solid bar of uniform crosssection A and parameter P. The length of the bar is L. For the small element of length dx, shown in the ﬁgure as shaded, heat qx entering the element by conduction is qx ¼ kAc
dt dx
Heat leaving the element by conduction is qx þ dx ¼ qx þ
d ðqx Þdx dx
Heat transferred by convection is qc ¼ hPdxðt t1 Þ Heat balance for the element gives qx qx þ dx qc ¼ 0 or
d ðqx Þdx qc ¼ 0 dx
or d dt kAc dx þ hPdxðt t1 Þ ¼ 0 dx dx or d2 t hP ðt t1 Þ ¼ 0 dx2 kAc The above equation is the same as Eq. (ii) of Sect. 3.2. Its solution is ðt t1 Þ ¼ C1 emx þ C2 emx
qc t1
qx x
qx+dx
t2
dx
Fig. 3.18 Heat transfer from a bar connected to two heat sources
3.10
Heat Transfer from a Bar Connected …
where m¼
159
rﬃﬃﬃﬃﬃﬃﬃ hP kAc
The boundary conditions for the present problem are At x ¼ 0; t ¼ t1 At x ¼ L; t ¼ t2 Applying the boundary conditions, t1 t1 ¼ C1 þ C2 ðt2 t1 Þ ¼ C1 emL þ C2 emL Solving the above equations, we get h2 h1 emL emL emL h1 emL h2 C2 ¼ mL e emL
C1 ¼
where h1 = t1 − t∞ and h2 = t2 − t∞ Substitution of the values of C1 and C2 in Eq. (i) gives t t1 ¼
h2 h1 emL mx h1 emL h2 mx e þ mL e emL emL e emL
which can be transformed into the following form: h1 sinh mðL xÞ h2 sinh mx þ sinh mL sinh mL h1 sinh mðL xÞ þ h2 sinh mx ¼ sinh mL
t t1 ¼
The rate of heat loss from the bar can be found from the relation ZL ðqc Þtotal ¼
hPdxðt t1 Þ 0
ZL ¼ hP
ðt t1 Þdx 0
ZL ¼ hP 0
h1 sinh mðL xÞ þ h2 sinh mx dx sinh mL
hP ¼ sinh mL ¼
ZL ½h1 sinh mðL xÞ þ h2 sinh mxdx 0
hP h1 ð1 cosh mLÞ þ h2 ðcosh mL 1Þ sinh mL m
ð3:31Þ
160
3
Extended Surfaces (Fins)
or qc ¼
hP ðh1 þ h2 Þðcosh mL 1Þ m sinh mL
ð3:32Þ
The position of maximum temperature in the bar can be determined by differentiating Eq. (3.31) with respect to x and equating to zero. Thus d 1 d ðt t1 Þ ¼ ½h1 sinh mðL xÞ þ h2 sinh mx ¼ 0 dx sinh mL dx This gives h1 m cosh mðL xÞ þ mh2 cosh mx ¼ 0 Thus h1 cosh mx ¼ h2 cosh mðL xÞ
ð3:33Þ
When both the reservoirs are at the same temperature, i.e. h1 = h2, then the above equation gives cosh mx ¼ cosh mðL xÞ or x ¼ L=2 i.e. the maximum temperature is at midplane of the rod, which is a logical result. Example 3.22 Can we obtain the temperature distribution equation of the bar connected to two heat sources at different temperatures from the equation of the ﬁn? Solution From Example 3.7, the temperature distribution equation for a bar with one end at temperature ts and the other at temperature t∞ is t t1 sinh mðL xÞ ¼ sinh mL ts t1
ð3:14Þ
Consider the subproblem (a) and (b) shown in Fig. 3.19. Their temperature distribution equations will be t t1 ¼ ðt1 t1 Þ
sinh mðL xÞ sinh mL
and t t1 ¼ ðt2 t1 Þ
sinh mx sinh mL
3.10
Heat Transfer from a Bar Connected …
161 θ=0
θ = θ1 a θ=0
θ = θ2 b
Fig. 3.19 Examples 3.22
Superposition of subproblems (a) and (b) gives the original problem. So the temperature distribution for the bar with ends at temperature excess h1 and h2 is t t1 ¼ h1
sinh mðL xÞ sinh mx þ h2 sinh mL sinh mL
which is the desired result. Example 3.23 One end of a long rod is inserted into a furnace while the other end projects into outside air. Under steady state, the temperature of the rod is measured at two points 75 mm apart and found to be 125°C and 88.5°C when the ambient temperature is 20°C. The rod diameter is 250 mm, and the convective heat transfer coefﬁcient is 6 W/(m2 K). Find the thermal conductivity of the rod material. Solution Given data is h = 6 W/(m2 K), P = pD = p 0.25 m and Ac = (p/4) D2 = (p/4) (0.25)2 m2, and k is unknown. For a very long ﬁn, t t1 ¼ emx ts t1 rﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP 6 p 0:25 4 9:8 m¼ ¼ ¼ pﬃﬃﬃ kAc k p ð0:25Þ2 k Let given temperatures are at x = L1 (t1 = 125°C) and x = L1 + 0.075 (t2 = 88.5°C), then 125 20 ¼e ts 20
ﬃ
9:8L1 p k
and 88:5 20 ¼e ts 20
h
ﬃ
9:8ðL1 þ 0:075Þ p k
Dividing the ﬁrst equation by the second, we get
i
162
3
125 20 ¼e 88:5 20
Extended Surfaces (Fins)
ﬃ
9:8 p 0:075 k
Simpliﬁcation gives k = 2.96 W/(m K). Alternatively, x1 may be considered to be zero, then t1 = ts = 125°C and the substitution gives 88:5 20 ¼e 125 20
ﬃ
9:8 p 0:075 k
which gives the same result. Example 3.24 An iron rod 10 mm in diameter and 500 mm long is giving up heat to its surrounding at an average rate of 10 W/(m2 K) by combined radiation and conduction. The ends of the rod are ﬁrmly connected to two heat sources at 120°C. The thermal conductivity of the iron is 40 W/(m K). The temperature of the surrounding is 20°C. (i) What is the temperature at the middle of the rod? (ii) How much heat is flowing out of each source? (iii) How much heat is flowing at the middle of the rod? Solution The parameter m is, from the given data, rﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP 10 p 0:01 4 m¼ ¼ ¼ 10 kAc p ð0:01Þ2 40 Product mL = 10 0.5 = 5, emL = 148.41, e−mL = 6.74 10−3 and emL/2 = 12.182. The temperature excess at each end, h1 = h2 = 120–20 = 100°C. (i) From Eq. (3.31), we have t t1 ¼
h1 sinh mðL xÞ þ h2 sinh mx sinh mL
Temperature at the middle of the bar tL=2 t1 ¼ 2h1 or tL=2 t1 ¼ 2h1
sinhðmL=2Þ sinh mL
mL mL=2 emL=2 emL=2 e 1 mL=2 e ¼ 2h ¼ 2h e 1 1 e2mL 1 emL emL emL þ 1
3.10
Heat Transfer from a Bar Connected …
163
or emL=2 þ t1 emL þ 1 12:182 þ 20 ¼ 36:3o C ¼ 2 100 148:41 þ 1
tL=2 ¼2h1
(ii) The heat transfer rate q¼
hP ðh1 þ h2 Þðcosh mL 1Þ m sinh mL
ð3:32Þ
where emL emL ¼ 74:2 2 mL mL e þe cosh mL ¼ ¼ 74:208 2 sinh mL ¼
Hence, q¼
10 ðp 0:01Þ ð100 þ 100Þ ð74:208 1Þ ¼ 6:2 W 10 74:2
(iii) The maximum of the temperature occurs at the middle of the rod when h1 = h2, and hence at this plane dh/dx = 0. Thus, the heat does not flow at the middle of the rod. Example 3.25 The ends of a 30mmdiameter and 1mlong rod are maintained at a constant temperature of 100°C. The rod transfers heat to the surrounding air at 25°C. The heat transfer coefﬁcient is estimated to be 32 kW/(m2 °C). If the centre of the rod is maintained at 50°C, ﬁnd the thermal conductivity of the rod material. Solution Temperature at the middle of the bar (refer to Example 3.24) tL=2 ¼ 2h1
emL=2 þ t1 emL þ 1
Substituting h1 = 100–25 = 75°C, and tL/2 − t∞ = 50–25 = 25°C, the equation transforms to emL þ 1 ¼ 6emL=2 Let emL/2 = x, then x2 6x þ 1 ¼ 0
164
3
Extended Surfaces (Fins)
The equation gives x = emL/2 = 5.8284 and 0.1716. The ﬁrst value gives m = 3.5254 for L = 1 m. Using the equation of parameter m, we get rﬃﬃﬃﬃﬃﬃﬃ hP m¼ kAc or 3:5284 ¼
sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 32 p 0:03 4 k p ð0:03Þ2
or k ¼ 343:3 W/(m K) The second value of x gives negative value of k, and hence is not applicable. Example 3.26 Derive an expression for the temperature distribution along a uniform crosssection ﬁn with internal heat generation at the rate of qg W/m3 along its length. The ﬁn end is insulated. The base temperature of the ﬁn is ts while the surrounding air is at temperature t∞. Can the heat flow from the base towards the ﬁn be equal to zero? Find out the condition. Solution We make an energy balance on an element of the ﬁn of thickness dx as shown in Fig. 3.20. In the steady state, heat generated in the element and heat inflow by conduction must equal to the heat outflow by conduction and convection, i.e. Energy inflow into left face þ heat generated in the element ¼ energy outflow from the right face þ energy lost by Convection
or qx þ qg ¼ qx þ dx þ qc or kAc
dt dt d dt þ qg ðAc dxÞ ¼ kAc þ kAc dx þ hðPdxÞðt t1 Þ dx dx dx dx
qc ts
qx
qx+dx
x dx
Fig. 3.20 Example 3.26
3.10
Heat Transfer from a Bar Connected …
165
where t is the temperature of the element and P is the perimeter. Simpliﬁcation gives d2 t hP qg ¼0 ðt t1 Þ þ dx2 kAc k Let (t − t∞) = h, then
d2 t dx2
2
¼ ddxh2 . Putting
hP kAc
¼ m2 , the equation is transformed to
d2 h qg ¼0 m2 h þ dx2 k The above equation can be further transformed by using h0 ¼ h kmg2 to give q
d 2 h0 m2 h0 ¼ 0 dx2 which is a secondorder differential equation. Its general solution is h0 ¼ C1 emx þ C2 emx
ðiÞ
The constants C1 and C2 are to be determined from the boundary conditions of the problem. (i) at x = 0, h0 ¼ h01 ¼ ts t1 kmg2 and q
(ii) at x = L,
dh dx
0
¼ dh dx ¼ 0
The ﬁrst condition gives C1 þ C2 ¼ ts t1
qg km2
ðiiÞ
and the second condition gives dh0 ¼ 0 ¼ mC1 emL mC2 emL dx or C1 emL C2 emL ¼ 0
ðiiiÞ
Solving Eqs. (ii) and (iii) for the constants C1 and C2, we get q
ts t1 kmg2 1 þ e2mL q ts t1 kmg2 C2 ¼ 1 þ e2mL C1 ¼
Substitution in Eq. (i) gives h0 ¼
q
q
ts t1 kmg2 mx ts t1 kmg2 mx e þ e 1 þ e2mL 1 þ e2mL
166
3
Extended Surfaces (Fins)
or
t t1
mx qg qg
e emx ¼ t t þ s 1 km2 km2 1 þ e2mL 1 þ e2mL
or q
t t1 kmg2 emx emx þ qg ¼ ts t1 km2 1 þ e2mL 1 þ e2mL or q
t t1 kmg2 cosh mðL xÞ q ¼ cosh mL ts t1 kmg2
ð3:34Þ
The condition of zero heat flow at the base is qfin
dt ¼ kAc ¼0 dx x¼o
or dt ¼0 dx x¼o From Eq. (3.34), we have dt qg
1 d ½cosh mðL xÞ ¼ ts t1 2 dx x¼0 km cosh mL dx x¼0 qg
1 ðm sinh mLÞ ¼ ts t1 2 km cosh mL From the above equation, it can be seen that the heat flow at the base of the ﬁn is zero q when ts t1 ¼ kmg2 . Example 3.27 A composite ﬁn consists of a cylindrical rod 3 mm diameter and 100 mm long of one material. It is uniformly covered with another material forming outside diameter of 10 mm and length of 100 mm. Thermal conductivity of inner material is 15 W/(m K) and that of outer is 45 W/(m K). Convective heat transfer coefﬁcient is 12 W/(m2 K). Determine (i) the effectiveness of the composite ﬁn neglecting heat rejection from the ﬁn end, and (ii) the efﬁciency of the ﬁn. Solution We make the energy balance on an element of the composite ﬁn of thickness dx as shown in Fig. 3.21.
3.10
Heat Transfer from a Bar Connected …
167
di x
do dx
Fig. 3.21 A composite ﬁn
Energy inflow into left face ¼ Energy outflow from the right face þ energy lost by convection or
qx ¼ qx þ dx þ qc
or qx ¼ qx þ
d ðqx Þdx þ hðPdxÞðt t1 Þ dx
or d ðqx Þdx þ hðPdxÞðt t1 Þ ¼ 0 ðiÞ dx dt dt where qx ¼ ki Ai þ ko Ao for the composite ﬁn. dx dx The above equation is based on the assumption that there is no temperature gradient in radial direction of the composite ﬁn, i.e. the temperature at any plane parallel to the wall is uniform throughout the crosssection. Substituting the value of qx, we get d dx
dt dt ki Ai þ ko Ao dx þ hðPdxÞðt t1 Þ ¼ 0 dx dx
or ðki Ai þ ko Ao Þ
d2 t hPðt t1 Þ ¼ 0 dx2
or d2 t hP ðt t 1 Þ ¼ 0 dx2 ðki Ai þ ko Ao Þ Let (t − t∞) = h, then
d2 t d2 h hP ¼ m2 , the equation is transformed ¼ 2 . Putting 2 dx dx ðki Ai þ ko Ao Þ
to d2 h m2 h ¼ 0 dx2
168
3
Extended Surfaces (Fins)
Heat transfer from the ﬁn is dt dt qfin ¼ ki Ai þ ko Ao dx x¼0 dx x¼0 Following the procedure used earlier qfin ¼ ðki Ai þ ko Ao Þmðts t1 Þ tanh mL
ð3:35Þ
sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP m¼ ðki Ai þ ko Ao Þ
ð3:36Þ
where
Heat transfer rate from the surface when there is no ﬁn, q ¼ hðAi þ Ao Þðts t1 Þ Thus the effectiveness of the ﬁn is e¼
qfin ðki Ai þ ko Ao Þ ¼ m tanh mL hðAi þ Ao Þ q
where ki ko Ai Ao h P m¼ L
qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP ðki Ai þ ko Ao Þ
= = = = = = =
15 W/(m K) 45 W/(m K) (p/4) (3/1000)2 m2 (p/4) [(10/1000)2−(3/1000)2] m2 12 W/(m2 K) pdo = p (10/1000) m 10.652
= 0.1 m. Substitution gives e = 29.85. Fin efﬁciency for case B gives gfin ¼
tanh mL ¼ 0:74 mL
Example 3.28 Heat dissipation from a surface (1 m 1 m) is to be increased by 50% by providing rectangular section ﬁns. The ﬁns are to be 5 mm thick, and 20 ﬁns are to be provided. The surface temperature is 100°C. It is expected that the surface temperature will drop to 90°C when the ﬁns will be installed but the convection heat transfer coefﬁcient can be assumed to remain unchanged at 20 W/(m2 K). Calculate the length of ﬁns if the thermal conductivity of the ﬁn material is 200 W/(m K) and the surrounding temperature is 20°C.
3.10
Heat Transfer from a Bar Connected …
169
Solution Heat dissipation without ﬁns, q ¼ hAc ðts t1 Þ ¼ 20 1 ð100 20Þ ¼ 1600 W Desired heat dissipation rate, q′ = 1.5 1600 = 2400 W. The surface area without ﬁns (bare surface area) = 1 1 – 1 5/1000 20 = 0.9 m2. Contribution of the bare surface, qb = 20 0.9 (90 − 20) = 1260 W. Heat to be dissipated by the ﬁns, qf′ = 2400 – 1260 = 1140 W. Heat to be dissipated by one ﬁn, qf = qf′ /(number of ﬁns) = 1140/20 = 57 W. The ﬁn is quite thin. Hence, we take account of the heat rejection by the ﬁn end by using corrected ﬁn length. Thus, pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ qfin ¼ PhkAc ðts t1 Þ tanh mLc ðiÞ where P = 2 (1000 + 5)/1000 = 2.01 m, Ac = 1000 5/10002 = 5 10−3. Value of the parameter m from the given data is
m¼
rﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP 20 2:01 ¼ ¼ 6:34 kAc 200 5 103
Substitution of the values of various terms in Eq. (i) gives 57 ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2:01 20 200 5 103 ð90 20Þ tanh mLc
Simpliﬁcation gives tanh mLc ¼ 0:12843 or mLc ¼ 0:12915 or 0:12915 0:12915 ¼ ¼ 0:02037 m m 6:34 ¼ 20:4 mm:
Lc ¼
Hence, the ﬁn length, d L ¼ Lc ¼ 20:4 2:5 ¼ 17:9 mm 2 Check: Considering the heat loss from ﬁn end, the heat dissipation equation is qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ h=mk þ tanh mL PhkAc ðts t1 Þ ðh=mkÞ tanh mL þ 1
170
3
Extended Surfaces (Fins)
Fig. 3.22 A rectangular section ﬁn
Substituting mL = (6.34 0.0179) = 0.11349 and values of other parameters, we obtain qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2:01 20 200 0:005 ð90 20Þ
¼ 57:13 W
20=ð6:34 200Þ þ tanhð0:11349Þ ½20=ð6:34 200Þ tanhð0:11349Þ þ 1
Example 3.29 Rectangular section (800 3 mm2) aluminium ﬁns of 30 mm length, as shown in Fig. 3.22, are provided on a flat wall. The heat transfer coefﬁcient from the ﬁn surface is 10 W/(m2 K). The thermal conductivity of the ﬁn material is 200 W/(m K). The temperature at the base of the ﬁn is 300°C, and the temperature of the surrounding air is 20° C. Calculate the temperature at the end of the ﬁn. Solution From Eq. (3.7), we have t t1 cosh mðL xÞ þ ðhL =mkÞ sinh mðL xÞ ¼ ðhL =mkÞ sinh mL þ cosh mL ts t1 Let the temperature at the ﬁn end (x = L) is tL, then tL t1 1 ¼ ts t1 ðhL =mkÞ sinh mL þ cosh mL or tL ¼ ðts t1 Þ
1 þ t1 ðhL =mkÞ sinh mL þ cosh mL
From the given data, rﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP 10 2 ð800 þ 3Þ 1000 ¼ 5:784 m¼ ¼ kAc 200 800 3 This gives mL ¼ 5:784 ð30=1000Þ ¼ 0:17353
3.10
Heat Transfer from a Bar Connected …
171
Hence, tL ¼ ð300 20Þ
1 þ 20 ¼ 295:35o C ½10=ð5:784 200Þ sinhð0:17353Þ þ coshð0:17353Þ
Note: The area of the ﬁn end is 800 3 = 2400 mm2, while the total surface area is 50580 mm2. Thus the ﬁn end area is only 4.75%. If we neglect the heat rejection from the ﬁn end, then we have (Eq. 3.23) tL t1 1 ¼ ts t1 cosh mL which gives tL = 295.84°C, which is only marginally different from the earlier calculated value. Example 3.30 A tube consists of integral longitudinal ﬁns as shown in Fig. 3.23. The outside diameter of the tube is 100 mm. Length of the tube is 1 m. Each ﬁn is 5 mm in thickness and 50 mm in height. There are 20 such ﬁns. The temperature at the base of the ﬁns is 100°C, and the surrounding temperature is 20°C. The heat transfer coefﬁcient from the ﬁn surface and also from the surface between the ﬁns can be assumed to be 10 W/(m2 K). The thermal conductivity of the ﬁn and tube material is 50 W/(m K). Calculate the amount of heat transferred from the ﬁnned wall to the surrounding. Also calculate the amount of heat that would be transferred from the tube wall without ﬁns under the same conditions. Solution From Eq. (3.8), the heat transfer from a ﬁn is qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ðhL =mkÞ þ tanh mL hPkAc ðts t1 Þ ðhL =mkÞ tanh mL þ 1
Here h = 10 W/(m2 K), P = 2(b + d) = 2(1000 + 5) 1/1000 m, Ac = 1 5/1000 m2, k = 50 W/(m K), ts = 100°C, t∞ = 20°C and L = 50 mm. The values of various parameters give rﬃﬃﬃﬃﬃﬃﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hP 10 2 ð1000 þ 5Þ 1=1000 ¼ 8:97 m¼ ¼ kAc 50 1 5=1000
δ
ro
Fig. 3.23 A tube with longitudinal ﬁns
L
172
3
Extended Surfaces (Fins)
This gives mL ¼ 8:97 50=1000 ¼ 0:4483 h 10 ¼ ¼ 0:0223 mk 8:97 50 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ¼ 10 ð2 1005=1000Þ 50 1 5=1000 ¼ 2:2417 Hence, qfin ¼ 2:2417 ð100 20Þ
0:0223 þ tanh 0:4483 ¼ 78:67 W 0:0223 tanh 0:4483 þ 1
Total heat transfer from 20 ﬁns = 20 78.67 = 1573.4 W. The bare surface of the tube = (pdo − 20 ﬁn thickness d) length of the tube ¼
ðp 100 20 5Þ 1 ¼ 0:2142 m2 1000
Heat transfer from the tube surface ¼ hAðts t1 Þ ¼ 10 0:2142 ð100 20Þ ¼ 171:3 W Total heat transfer from the tube = 1573.4 + 171.3 = 1744.7 W. Heat transfer from the tube without ﬁns, ¼ hðpdo LÞ ðts t1 Þ ¼ 10 ðp 100 1Þ=1000 ð120 20Þ ¼ 251:33 W Hence, effectiveness = 1744.7/251.33 = 6.94.
3.11
Generalized Equation of Fin
We have studied the performance of uniform crosssection ﬁns protruding from a flat wall. In practical applications, the proﬁle or crosssection of the ﬁn may vary along its length as shown in Fig. 3.24 and may be attached to circular surfaces. The calculation for ﬁns with variable crosssection is more complicated than that for the straight ﬁns of uniform crosssection. Let the shape of the ﬁn is such that AðxÞ ¼ f ðxÞ and PðxÞ ¼ uðxÞ
3.11
Generalized Equation of Fin
173 y
x
Fig. 3.24 A ﬁn of variable crosssection
Taking inﬁnitesimal element of thickness dx of the ﬁn at a distance x from the wall, refer to Fig. 3.24, the heat balance equation can be written as kAx
dt dt d dt ¼ kAx þ kAx dx þ hPx dxðt t1 Þ dx dx dx dx
or d dt h Ax ¼ Px ðt t1 Þ dx dx k or Ax
d2 t dAx dt h Px ðt t1 Þ ¼ 0 þ dx2 dx dx k
or d2 t dAx =dx dt h Px þ ðt t1 Þ ¼ 0 dx2 Ax dx k Ax
ð3:37Þ
This is the generalized equation applicable to ﬁns of any proﬁle or crosssection where the crosssection is some function of x. Equation (3.37) is a modiﬁed Bessel’s equation, and its solution is obtained with the help of Bessel’s function. Note: For a ﬁn of uniform crosssection, Ax and Px are not functions of x but are constants. Hence, their derivatives are zero and the above equation reduces to d2 t hP ðt t1 Þ ¼ 0 dx2 kA which is the equation for the uniform crosssection ﬁn.
3.12
Fin of Minimum Weight (Isachenko et al.1977)
The design of ﬁn for maximum heat transfer at minimum weight is of special signiﬁcance in some applications. This can be achieved if the speciﬁc rate of heat flow (heat flow through the unit area of crosssection of the ﬁn) remains constant over the entire length of the ﬁn.
174
3
Extended Surfaces (Fins)
ts t t∞ L x ts
r
Fig. 3.25 Fin of minimum weight
Under such condition, the lines of heat flow will be parallel to the axis of the ﬁn and the temperature distribution along the ﬁn length will be linear, Fig. 3.25. Therefore, we may write for any crosssection of the ﬁn: x t t1 ¼ ðts t1 Þ L
ðiÞ
where x is the distance measured from the tip and L is the total length of the ﬁn. Consider an elementary surface of the ﬁn at distance x. If the rate of heat flow along the ﬁn axis is q, the heat flow through the elemental surface will be q sin/, where / is angle of the elemental surface with the axis of the ﬁn. The heat balance at the elemental surface gives q sin / ¼ hðt t1 Þ or q sin / ¼
hx ðts t1 Þ L
This gives sin / ¼
hx x ðts t1 Þ ¼ Lq Lq=½hðts t1 Þ
ð3:38Þ
The contour of this ﬁn is a circular arc of radius r, since sin/ = x/r. From Eq. (3.38), r = Lq/[h(ts − t∞). This ﬁn possesses the least weight. However, it is difﬁcult to manufacture such ﬁns. A triangular ﬁn differs only slightly from the circular arc ﬁn and can be manufactured with ease. Such ﬁns are commonly used in applications where the minimum weight is the design consideration.
3.13
3.13
Straight Fin of Triangular Section
175
Straight Fin of Triangular Section
The area normal to the heat flow is a function of the distance along the ﬁn. With origin at the vertex of the triangle, the area of crosssection at distance x is x
Ax ¼ d b L
ðiÞ
Following the procedure outlined for uniform crosssection ﬁns, the heat balance for the inﬁnitesimal element of thickness dx gives, refer to Fig. 3.26, d dt kAx þ hPx ðt t1 Þ ¼ 0 dx dx When b > > dx, Px = 2(b + dx) 2b and is constant. Assuming constant thermal conductivity and substituting the value of Ax from Eq. (i), the above equation transforms to db d dt x k þ hð2bÞðt t1 Þ ¼ 0 L dx dx or
d dt 2hL x ðt t 1 Þ ¼ 0 dx dx kd
or x where b ¼
2hL . kd
Let h = t − t∞, then
d2 t dt bð t t 1 Þ ¼ 0 þ dx2 dx
dt dh d2 t d2 h ¼ , and 2 ¼ 2 . Hence, we obtain dx dx dx dx d 2 h 1 dh b h¼0 þ dx2 x dx x
ð3:39Þ
Equation (3.39) is a modiﬁed Bessel equation and its solution is h ¼ C1 I0 ð2
x
pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ bxÞ þ C2 K0 ð2 bxÞ
t∞
L
b δ θs
δx x dx
Fig. 3.26 Atriangular ﬁn
ð3:40Þ
176
3
Extended Surfaces (Fins)
Table 3.3 Modiﬁed Bessel functions of ﬁrst and second kinds a
I0(a)
I1(a)
K0(a)
K1(a)
0.0 0.2 0.4 0.6 0.8 1.0
1.0000 1.0100 1.0404 1.0921 1.1665 1.2661
0.0000 0.1005 0.2040 0.3137 0.4329 0.5652
∞ 1.753 1.114 0.777 0.565 0.421
∞ 4.775 2.185 1.303 0.862 0.602
1.2 1.4 1.6 1.8 2.0
1.394 1.553 1.750 1.990 2.280
0.715 0.886 1.085 1.371 1.591
0.319 0.244 0.188 0.146 0.114
0.435 0.321 0.241 0.183 0.140
2.2 2.4 2.6 2.8 3.0
2.629 3.049 3.553 4.157 4.881
1.914 2.298 2.755 3.301 3.953
0.0893 0.0702 0.0554 0.0438 0.0347
0.1079 0.0837 0.0653 0.0511 0.0402
3.2 3.4 3.6 3.8 4.0
5.747 6.785 8.028 9.517 11.302
4.734 5.670 6.793 8.140 9.760
0.02760 0.02196 0.01750 0.01397 0.0112
0.0316 0.0250 0.0198 0.0157 0.0125
4.2 4.4 4.6 4.8 5.0
13.442 16.010 19.023 22.794 27.240
11.706 14.046 16.863 20.253 24.336
0.00893 0.00715 0.00573 0.00460 0.00369
0.00994 0.00792 0.00633 0.00505 0.00404
5.2 5.4 5.6 5.8 6.0
32.584 39.009 46.738 56.038 67.234
29.254 35.182 42.328 50.946 61.342
0.00297 0.00238 0.00192 0.00154 0.00125
0.00324 0.00260 0.00208 0.00167 0.00134
6.2 6.4 6.6 6.8 7.0
80.72 96.98 116.54 140.14 168.6
73.89 89.03 107.30 129.38 156.04
0.001005 0.00081 0.00065 0.000526 0.000425
0.00108 0.00087 0.00067 0.000563 0.000454
where I0 is modiﬁed Bessel function of I kind, and K0 is modiﬁed Bessel function of II kind. The values of functions I0(a) and K0(a) are tabulated in Table 3.3 as function of a. In the present case, a = 2√(bx). At x = 0, a = 2√(bx) = 0. For a = 0, Ko(a) = ∞ from the table. It means that at the ﬁn tip (x = 0), the temperature excess h is inﬁnity from Eq. (3.40), which is physically not possible. Hence, the constant C2 must be zero. This gives h ¼ C1 I0 ð2
pﬃﬃﬃﬃﬃ bxÞ
The other boundary condition is h = hs (the temperature at the base of the ﬁn) at x = L.
3.13
Straight Fin of Triangular Section
177
This gives hs ¼ C1 I0 ð2
pﬃﬃﬃﬃﬃﬃ bLÞ
or C1 ¼
hs pﬃﬃﬃﬃﬃﬃ I0 ð2 bLÞ
and hence the temperature distribution equation applicable to the present case is pﬃﬃﬃﬃﬃ I0 ð2 bxÞ pﬃﬃﬃﬃﬃﬃ h ¼ hs I0 ð2 bLÞ
ð3:41Þ
The temperature ho of the ﬁn end is obtained by putting x = 0 in the above equation ho ¼ hs
1 pﬃﬃﬃﬃﬃﬃ I0 ð2 bLÞ
ð3:42Þ
The heat flow rate qﬁn equals the heat flow by conduction at the ﬁn base, i.e.
dt dx x¼L pﬃﬃﬃﬃﬃ d I0 ð2 bxÞ pﬃﬃﬃﬃﬃﬃ hs ¼ kðbdÞ dx I0 ð2 bLÞ x¼L h pﬃﬃﬃﬃﬃ i 1 d pﬃﬃﬃﬃﬃﬃ I0 ð2 bxÞ ¼ kðbdÞhs I0 ð2 bLÞ dx x¼L
qfin ¼
Since
d ½In ðaÞ dx
kA
¼ In þ 1 ðaÞ x¼L
d ðaÞ, we get dx
pﬃﬃﬃﬃﬃ pﬃﬃﬃ 1 1 pﬃﬃﬃﬃﬃﬃ I1 ð2 bxÞ 2 b pﬃﬃﬃ 2 x x¼L I0 ð2 bLÞ pﬃﬃﬃ pﬃﬃﬃﬃﬃﬃ b I1 ð2 bLÞ pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ ¼ kðbdÞhs I0 ð2 bLÞ L
qfin ¼ kðbdÞhs
Substituting b ¼
2hL and simplifying, we get kd qfin
pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ I1 ð2 bLÞ pﬃﬃﬃﬃﬃﬃ ¼ 2hkd bhs I0 ð2 bLÞ
ð3:43Þ
Example 3.31 The heat dissipation from an aircooled flat surface is to be enhanced by installing either triangular or rectangular ﬁns. Either ﬁn is to be 25 mm thick, 100 mm long and made from aluminium [k = 230 W/(m K)]. The wall temperature is 630°C, and the heat transfer coefﬁcient between the surface and the ambient air at 30°C is 72 W/(m2 K). Compare the effectiveness of the two ﬁns based on the heat flow per unit weight.
178
3
Extended Surfaces (Fins)
Solution (i) Triangular ﬁn: From Eq. (3.43), qfin ¼
pﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ I1 ð2 bLÞ pﬃﬃﬃﬃﬃﬃ 2hkd bhs I0 ð2 bLÞ
ðiÞ
Here h = 72 W/(m2 K), k = 230 W/(m K), d = 0.025 m, b = 1 m (assumed), hs = ts − t∞ = 630 – 30 = 600°C and L = 0.1 m. The factor b = 2 hL/kd = 2 72 0.1/(230 0.025) = 2.504 and 2√(bL) = 2√(2.504 0.1) = 1. From Table 3.3, I1(1) = 0.565 and I0(1) = 1.266. Substitution of values of various parameters in Eq. (i) gives pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 0:565 1 600 2 72 230 0:025 1:266 ¼ 7705 W per m width of the fin
qfin ¼
(ii) Rectangular ﬁn: From the given data, P ¼ 2ð1 þ 0:025Þ ¼ 2:05 m2 2 Ac ¼ bd ﬃﬃﬃﬃﬃ0:025 m q¼ hP m ¼ kAc ¼ 5:07 h mk
¼ 0:062
Hence, from Eq. (3.8), pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ h=mk þ tanh mL PhkAc ðts t1 Þ ðh=mkÞ tanh mL þ 1 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 0:062 þ tanhð0:507Þ ¼ 2:05 72 230 0:025 600 0:062 tanhð0:507Þ þ 1
qfin ¼
¼ 8990 W per m width The weight of the rectangular section ﬁn per unit width is twice that of the triangular ﬁn, while the heat transfer capacity is only 16.7 percent higher. Thus, the triangular ﬁn is more effective per unit weight.
3.14
Straight Fin of Trapezoidal Section
3.14
179
Straight Fin of Trapezoidal Section
Consider a ﬁn of trapezoidal crosssection as shown in Fig. 3.27. The ﬁn operates in conditions speciﬁed for the forgoing case. The dimensions and notations are given in the ﬁgure. The general solution given in Eq. (3.40) is h ¼ C1 I0 ð2
pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ bxÞ þ C2 K0 ð2 bxÞ
ðiÞ
where I0 is modiﬁed Bessel function of I kind, and K0 is modiﬁed Bessel function of II 1 kind. Here b ¼ 2hL kd1 . One of the boundary conditions is h = hs (the temperature at the base of the ﬁn) at x = L1. This gives hs ¼ C1 I0 ð2
pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ bL1 Þ þ C2 K0 ð2 bL1 Þ
ðiiÞ
For simpliﬁcation, we assume that heat transfer from the ﬁn end is negligible, i.e. dh ¼ 0 at x ¼ L2 dx x¼L2 This gives pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ ﬃ C2 K1 ð2 bL2 Þ2 b p1ﬃﬃﬃﬃ 0 ¼ C1 I1 ð2 bL2 Þ2 b 2p1ﬃﬃﬃ 2 L2 pﬃﬃﬃﬃﬃﬃﬃLﬃ 2 pﬃﬃﬃﬃﬃﬃﬃﬃ 0 ¼ C1 I1 ð2 bL2 Þ C2 K1 ð2 bL2 Þ
ðiiiÞ
where
d ½Kn ðaÞ dx
¼ Kn þ 1 ðaÞ x¼L
d ðaÞ dx
Solving Eqs. (ii) and (iii), we get values of constants C1 and C2. Finally substitution of their values in Eq. (i) gives pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ h Io ð2 bxÞK1 ð2 bL2 Þ þ I1 ð2 bL2 ÞKo ð2 bxÞ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ ¼ hs Io ð2 bL1 ÞK1 ð2 bL2 Þ þ I1 ð2 bL2 ÞKo ð2 bL1 Þ
ð3:44Þ
The temperature of the ﬁn end h2 can be found from the above equation by putting x = L2.
L δ1 θ s
t∞
L1
L2
δx x dx
Fig. 3.27 A trapezoidal ﬁn
(δ2, θ2)
180
3
Extended Surfaces (Fins)
The heat flow rate qﬁn equals the heat flow by conduction at the ﬁn base, i.e.
dt dx x¼L1 pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ I1 ð2 bL1 ÞK1 ð2 bL2 Þ I1 ð2 bL2 ÞK1 ð2 bL1 Þ kd1 b b pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ hs ¼ L1 I0 ð2 bL1 ÞK1 ð2 bL2 Þ þ I1 ð2 bL2 ÞK0 ð2 bL1 Þ
qfin ¼
kAbase
1 Substituting b ¼ 2hL kd1 and simplifying, we get
pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ I1 ð2 bL1 ÞK1 ð2 bL2 Þ I1 ð2 bL2 ÞK1 ð2 bL1 Þ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 2hkd1 bhs I0 ð2 bL1 ÞK1 ð2 bL2 Þ þ I1 ð2 bL2 ÞK0 ð2 bL1 Þ
qfin
ð3:45Þ
The heat transfer from the tip of the ﬁn can be approximately accounted for by an increase in the ﬁn length L by half of its thickness.
3.15
Annular Fin
Figure 3.28 shows an annular ﬁn. Consider an elemental ring of width dr at radius r. The surface area of the ring rejecting heat to the surrounding is 2(2prdr) and the crosssectional area is (2prd). Heat entering the elemental ring by conduction is dt dt qr ¼ kAc ¼ kð2prdÞ dr dr Heat leaving the element is
qr þ dr
dqr ¼ qr þ dr dr
Heat rejected by convection to the surrounding from the elemental surface area is qc ¼ hAðt ts Þ ¼ h½2ð2prdrÞðt t1 Þ
Pipe dr r2
r1 r
dr
qc δ
Fig. 3.28 An annular ﬁn
r qr r1
r2
qr+dr
3.15
Annular Fin
181
In equilibrium, the heat entering the element equals the heat leaving the element, i.e. qr ¼ qr þ dr þ qc dqr ¼ qr þ dr þ qc dr or
dqr dr ¼ qc dr
or
d dt kð2prdÞ dr ¼ 4prdr h ðt t1 Þ dr dr
or kð2pdÞ
d dt r ¼ 4pr h ðt t1 Þ dr dr
or d dt 2h r ¼ r ðt t1 Þ dr dr kd Putting h = t – t∞, d dh 2h r ¼ rh dr dr kd or dh d2 h 2h þr 2 rh ¼ 0 dr dr kd or d2 h 1 dh 2hh ¼0 þ dr 2 r dr kd
ðiÞ
This is a form of Bessel’s equation of zero order, and its solution is h ¼ C1 I0 ðbrÞ þ C2 K0 ðbrÞ
ðiiÞ
where b = √(2 h/kd) and I0 and K0 are as deﬁned earlier. Constants C1 and C2 are to be determined from the boundary conditions. One of the boundary conditions is h = hs, the temperature at the base of the ﬁn, i.e. at r = r1.
182
3
Extended Surfaces (Fins)
This gives hs ¼ C1 I0 ðbr1 Þ þ C2 K0 ðbr1 Þ
ðiiiÞ
For simpliﬁcation, we assume that heat transfer from the ﬁn end is negligible. This is true when d < < (r2 – r1). This gives dh ¼0 dx r¼r2 Hence, 0 ¼ C1 I1 ðbr2 Þb C2 K1 ðbr2 Þb
ðivÞ
Solving Eqs. (iii) and (iv), we get values of constants C1 and C2 as K1 ðbr2 Þ I0 ðbr1 ÞK1 ðbr2 Þ þ I1 ðbr2 ÞK0 ðbr1 Þ I1 ðbr2 Þ C 2 ¼ hs I0 ðbr1 ÞK1 ðbr2 Þ þ I1 ðbr2 ÞK0 ðbr1 Þ C 1 ¼ hs
Substitution of their values in Eq. (ii) gives h I0 ðbrÞK1 ðbr2 Þ þ I1 ðbr2 ÞK0 ðbrÞ ¼ hs I0 ðbr1 ÞK1 ðbr2 Þ þ I1 ðbr2 ÞK0 ðbr1 Þ
ð3:46Þ
The temperature of the ﬁn end h2 can be found from the above equation by putting r = r2. The heat flow rate qﬁn equals the heat flow by conduction at the ﬁn base, i.e. qfin
dt ¼ kAbase dr r¼r1 I1 ðbr2 ÞK1 ðbr1 Þ I1 ðbr1 ÞK1 ðbr2 Þ ¼ 2pkdbr1 hs I0 ðbr1 ÞK1 ðbr2 Þ þ I1 ðbr2 ÞK0 ðbr1 Þ
ð3:47Þ
The heat transfer from the tip of the ﬁn can be approximately accounted for by increasing r2 by half of the ﬁn thickness d. Example 3.32 An annular aluminium alloy ﬁn [k = 160 W/(m K)] is installed on a 25 mm OD tube. The ﬁn has uniform thickness of 0.4 mm and has outer radius of 37.5 mm. If the tube wall temperature is 150°C, the surrounding air temperature is 20°C and the average convective heat transfer coefﬁcient is 30 W/(m2 K), calculate the heat loss from the ﬁn. Solution Neglecting the heat loss from the ﬁn end,
qfin
I1 ðbr2 ÞK1 ðbr1 Þ I1 ðbr1 ÞK1 ðbr2 Þ ¼ 2pkdbr1 hs I0 ðbr1 ÞK1 ðbr2 Þ þ I1 ðbr2 ÞK0 ðbr1 Þ
ð3:47Þ
3.15
Annular Fin
183
Here k = 160 W/(m K), d = 0.4/1000 m, hs = (150–20) = 130°C, h = 30 W/(m2 K). Hence, br1 ¼
qﬃﬃﬃﬃ
qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 12:5 ¼ 2301000 1600:4 1000 ¼ 0:3827; br2 ¼ br1 rr21 ¼ 1:1482
2h kdr1
This gives qfin ¼ 2p 160 0:0004 0:3827 130
I1 ð1:1482ÞK1 ð0:3827Þ I1 ð0:3827ÞK1 ð1:1482Þ I0 ð0:3827ÞK1 ð1:1482Þ þ I1 ð1:1482ÞK0 ð0:3827Þ
Substitution of the values of Bessel’s function from Table 3.3 gives qfin ¼ 23:1 W
3.16
Fin Efficiency Plots
The equations derived in previous sections for rectangular, triangular and circumferential ﬁns are very inconvenient for calculations. The heat flow rate can be determined approximately for most of the engineering applications using the following equation: qfin ¼ gf hAs hs
ð3:48Þ
where ηf is the ﬁn efﬁciency, which can be read from Fig. 3.29 for rectangular, triangular and circumferential ﬁns of rectangular crosssection, and As is surface area of the ﬁn. The 1/2 abscissa parameter in Fig. 3.29 is L3/2 where Lc (the corrected length) and Am are c (h/kAm) deﬁned in the ﬁgure. For other ﬁn shapes, the readers can refer to Gardner (1945) and advanced texts on the subject. Example 3.33 Repeat Example 3.31 using the efﬁciency approach. Solution (i) Triangular ﬁn: Lc ¼ L ¼ 0:1; Am ¼ ðt=2ÞL ¼ ð0:025=2Þ 0:1 ¼ 0:00125 rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃ h 72 3=2 3=2 Parameter Lc ¼ 0:1 ¼ 0:5 kAm 230 0:00125
184
3
Extended Surfaces (Fins)
100 90
Fin efficiency, %
t
Rectangular fin
80
L
Triangular fin
Lc = L + t/2 Am = tLc
70 60
Rectangular fin.
50 t
40 30
L
20
Lc = L Am = (t/2)L
10 0
Triangular fin. 0
0.5
1.0 1.5 Lc3/2(h/kAm)1/2
2.0
2.5
a 100 90
Fin efficiency, %
80 t
70
r1
L r2
60 50 40
3
30
4
20
r2c/r1 1 2
Lc = L + t/2 r2c = r1 + Lc Am = tLc
5
10 0
0
0.5
1.5 1.0 2.0 Lc3/2(h/kAm)1/2
2.5
3.0
b Fig. 3.29 a Efﬁciencies of rectangular and triangular ﬁns, b Efﬁciency of circumferential or annular ﬁns of rectangular section. Holman JP, adapted for SI units by White PRS, Heat Transfer, McGrawHill Book Co, New York, Copyright 1992. The material is reproduced with permission of McGrawHill Education (Asia)
From Fig. 3.29a, the ﬁn efﬁciency, ηf = 0.88 for the calculated value of the above parameter. Hence,
qfin
qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ hAs hs gf ¼ h ð2bÞ L2 þ ðt=2Þ2 hs gf qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 72 ð2 1Þ 0:12 þ ð0:025=2Þ2 ð630 30Þ 0:88 ¼ 7662 W=m width of the fin
3.16
Fin Efficiency Plots
185
(ii) Rectangular ﬁn: Lc ¼ L þ t=2 ¼ 0:1125; t ¼ 0:025; Am ¼ tLc ¼ 0:025 0:1125 ¼ 0:0028125 rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃ h 72 3=2 3=2 Parameter Lc ¼ 0:1125 0:4 kAm 230 0:0028125 From Fig. 3.29a, the ﬁn efﬁciency ηf 0.91 for the calculated value of above parameter. Hence, qfin ¼ hAs hs gf ¼ h ð2bLc Þ hs gf ¼ 72 ð2 1 0:1125Þ ð630 30Þ 0:91 ¼ 8845 W/m width of the fin The results are quite close to the solution in Example 3.31. Example 3.34 Repeat above example for the triangular ﬁn with L = 0.2 m and 0.05 m keeping ﬁn base thickness ﬁxed at 0.025 m. Solution (i) L = 0.2 m: Lc ¼ L ¼ 0:2; Am ¼ ðt=2ÞL ¼ ð0:025=2Þ 0:2 ¼ 0:0025 rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃ h 72 3=2 3=2 Parameter Lc ¼ ð0:2Þ ¼ 1:0 kAm 230 0:0025 From Fig. 3.29, the ﬁn efﬁciency, ηf = 0.7 and qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ qfin ¼ hAs hs gf ¼ h ð2bÞ L2 þ ðt=2Þ2 hs gf qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 72 ð2 1Þ 0:22 þ ð0:025=2Þ2 ð630 30Þ 0:7 ¼ 12120 W/m width of the fin i.e. 1.55 times of that for ﬁn of length 0.1 m while the weight of the ﬁn increased to 2 times. (ii) L = 0.05 m: Lc ¼ L ¼ 0:05; Am ¼ ðt=2ÞL ¼ ð0:025=2Þ 0:05 ¼ 0:000625 rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃ h 72 3=2 3=2 Parameter Lc ¼ ð0:05Þ ¼ 0:25 kAm 230 0:000625 From Fig. 3.29, the ﬁn efﬁciency, ηf = 0.95 and
186
3
Extended Surfaces (Fins)
qfin
qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ hAs hs gf ¼ h ð2bÞ L2 þ ðt=2Þ2 hs gf qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 72 ð2 1Þ 0:052 þ ð0:025=2Þ2 ð630 30Þ 0:95 ¼ 4229 W/m width of the fin
i.e. 0.55 times of that for ﬁn of length 0.1 m while the weight of the ﬁn is reduced to half. We can conclude that the heat transfer per unit weight is greater for ﬁns of smaller heights. Example 3.35 Repeat Example 3.32 using the ﬁn efﬁciency approach with no correction for the ﬁn end loss (as ﬁn thickness is very small). Solution For Lc = L = 0.025, k = 160 W/(m K),
Am = tL = 0.0004 0.025 = 0.00001,
Parameter L3=2 c
h = 30 W/(m2
K)
and
rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃ h 30 3=2 ¼ ð0:025Þ ¼ 0:54 kAm 160 0:00001
and r2c ¼3 r1 From Fig. 3.29b, ﬁn efﬁciency ηf 0.73. Hence, qfin ¼ gf hAs hs ¼ 0:73 30 ½2pð0:03752 0:01252 Þ ð150 20Þ ¼ 22:36 W which is quite close to the solution given in Example 3.32. Example 3.36 A steel pipe carries hot gas at 800°C. In order to increase heat transfer rate from the pipe, a ﬁnned aluminum cylinder is slipped over the steel pipe as shown in Fig. 3.30a. The contact resistance to heat transfer at the pipe and cylinder interface is estimated to be 10−4 m2 K/W. For the data given in the ﬁgure, determine heat transfer rate per unit length of pipe. Solution The thermal network for the system is shown in Fig. 3.30b. The heat transfer rate is q¼
tgas t1 Rtotal
where Rtotal = Ri + Rwall + Rc + R2 + Req. The various resistances for unit length of pipe are calculated below.
3.16
Fin Efficiency Plots Steel pipe, ks = 40 W/(m K)
187 ro r2 r1 ri
Aluminium pipe, ka = 210 W/(m K)
hi = 100 W/(m2 K), tgas = 650oC
ho = 50 W/(m2 K), t∞ = 30oC 3 mm 2 mm
a ri = 80 mm, r1 = 90 mm, r2 = 105 mm and ro = 130 mm. Rbare Rwall
tgas
Rc
R2
t∞
q
Ri
b Thermal network
Rfin
Fig. 3.30 Example 3.36
Inner surface ﬁlm resistance: Ri ¼
1 1 1 ¼ 0:0199 ¼ ¼ hi Ai hi ð2pri LÞ 100 2 p 0:08 1
Pipe wall resistance: Rwall
1 r1 1 90 ln ln ¼ ¼ 0:468 103 ¼ 2pks L 2 p 40 1 80 ri
Contact resistance: Rc ¼
104 104 ¼ ¼ 0:176 103 2pr1 L 2 p 0:09 1
Aluminium wall resistance: 1 r2 1 105 ln ln R2 ¼ ¼ 0:117 103 ¼ 2pka L 2 p 210 1 90 r1 Film resistance of surface of aluminium cylinder without ﬁns: Rbare ¼
1 1 1 ¼ 0:0505 ¼ ¼ ho Abare ho ð2pr2 Lbare Þ 50 2 p 0:105 0:6
Fin resistance using Eq. (3.48): Rfin ¼
ts t1 1 ¼ gfin ho Afin qfin
188
3
Extended Surfaces (Fins)
Efﬁciency of the ﬁn (refer to Fig. 3.29b): Lc ¼ L þ t=2 ¼ 25 þ 2=2 ¼ 26 mm r2c ¼ r1 þ Lc ¼ 105 þ 26 ¼ 131 mm Am ¼ tLc ¼ 2 26 ¼ 52 mm2 r2c / r1 ¼ 131=105 ¼ 1:248
1=2 3=2 Lc ðh=kAm Þ1=2 ¼ 0:0263=2 50= 210 52 106 ¼ 0:284 From Fig. 3.29b, ηﬁn = 0.91. Fin surface area per unit length of pipe: 2 Afin ¼ no: of fins pðr2c r22 Þ 2 ¼ 200 p ð0:1312 0:1052 Þ 2 ¼ 7:71 m2
Hence,
Req ¼
h
1 ¼ 0:91507:71 ¼ 2:85 103 i1 1 1 1 þ R1fin ¼ 0:0505 þ 2:8510 ¼ 2:69 103 3
Rfin ¼ g
1 Rbare
1
fin ho Afin
Total resistance, Rtotal ¼ Ri þ Rwall þ Rc þ R2 þ Req ¼ 0:0199 þ 0:468 103 þ 0:176 103 þ 0:117 103 þ 2:69 103 ¼ 0:02335 Heat transfer rate, q¼
3.17
650 30 ¼ 26552 W=m length 0:02335
Summary
The heat transfer rate from or to a surface in contact with a fluid can be enhanced by providing ﬁns, which basically increase the surface area for heat transfer. Fins are widely used in various engineering equipments. Fins of various shapes and in different arrangements have been employed. A majority of ﬁns are of rectangular cross section. Thick ﬁns and ﬁns made by casting are generally of the trapezoidal crosssection with rounded edges. Spine or studtype ﬁns are mechanically rugged, and hence have longer life in corrosive atmosphere than thin plate ﬁns. It may be noted that the ﬁnal choice of the type of ﬁn and their arrangement depends not only on the heat transfer performance but also on the resistance offered to the flow of the surrounding fluid, cost and the ease of fabrication. In this chapter, mathematical treatments of various types of uniform crosssection, triangular, trapezoidal and annular ﬁns have been presented to determine the temperature variation along the ﬁn length and heat transfer rate from the ﬁn surface assuming onedimensional steadystate heat flow condition. It is to note that an exact analytical solution of the heat propagation through a ﬁn involves considerable difﬁculties. Mathematical solutions presented here are approximate because of the following major simplifying assumptions.
3.17
Summary
189
(i) Temperature at any crosssection of the ﬁn is uniform. The assumption is valid for ﬁns of small crosssection and made of high thermal conductivity material. This assumption reduces the problem to that of onedimensional heat conduction along the axis of the ﬁn only. (ii) The heat transfer coefﬁcient h from the ﬁn surface to the surrounding is constant for the entire ﬁn surface. The greatest uncertainty is in the value of the heat transfer coefﬁcient, which is seldom uniform over the entire ﬁn surface. In the case of severe nonuniformity of the heat transfer coefﬁcient and two or threedimensional conduction heat flow, numerical techniques are used to solve the problem. A discussion on the same has been presented in Chap. 5. In order to do mathematical analysis of the uniform crosssection ﬁns, the ﬁns have been divided into three classes: (i) a very long ﬁn (mathematically a ﬁn of inﬁnite length), (ii) a ﬁn with negligible heat transfer from the ﬁn end as compared to the heat transferred from the ﬁn surface (treated as a ﬁn with insulated end) and (iii) a short ﬁn (ﬁn with heat loss from the ﬁn end). The heat transfer equation of the short ﬁn can be expressed in the same form as that of ﬁn with insulated end if a corrected length Lc, deﬁned as Lc = L + ΔL where ΔL = Ac/P, is used. pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ In fact, we can use heat transfer equation as qfin ¼ hPkAc ðts t1 Þ tanh mLc for all cases because in case of a very long ﬁn product mL will be greater than 5 and tanh(mL) can be taken as unity, refer to Table 3.1. To study the performance of a ﬁn in transferring the heat, a term called ﬁn effectiveness has been used, which is a ratio of heat transfer with the ﬁn to the heat that would be transferred without ﬁn. Effectiveness of a ﬁn must be always greater than 1 and as high as possible. In general, the ﬁn effectiveness is high when (i) the ﬁn is made of high thermal conductivity material, (ii) it has high ratio of parameter to the base area (P/Ac) and (iii) the convective heat transfer coefﬁcient is low. Condition for enhancement of heat transfer from a ﬁn has been deduced in terms of a nondimensional number termed as Biot number Bi. Since the basic purpose of the installation of ﬁns on a surface is to enhance the heat transfer rate, the ﬁn effectiveness is a meaningful parameter for assessing the usefulness of the ﬁns. However, another term that has been used to assess the performance of a ﬁn in transferring heat is ﬁn efﬁciency ηﬁn, which is deﬁned as ratio of the heat transfer with the ﬁn to the heat that would be transferred if the entire ﬁn surface were at the ﬁn base temperature. Since most of the heattransferring surfaces consist of a number of ﬁns with bare portion of the surface between the ﬁns, the total heat transfer from the surface is obtained by combining the heat flow from the ﬁns qﬁn with the heat flow from the bare portion of the surface qo. Using the ﬁn efﬁciency, the heat flow rate from the wall can be readily expressed as qtotal ¼ qfin þ qo ¼ hf Af ðts t1 Þgfin þ hw Ao ðts t1 Þ, where hw and hf are heat transfer coefﬁcients for the bare portion and the ﬁns, respectively. In Sect. 3.12, design of ﬁn for maximum heat transfer at minimum weight has been developed. The contour of this ﬁn is a circular arc. However, it is difﬁcult to manufacture such ﬁns. A triangular ﬁn differs only slightly from the circular arc ﬁn and can be manufactured with ease. Such ﬁns are commonly used in applications where minimum weight is design consideration. A modiﬁcation of the triangular ﬁn is the ﬁn of trapezoidal crosssection. Mathematical equations for temperature distribution and heat transfer from triangular and trapezoidal ﬁns have been derived in Sects. 3.13 and 3.14, respectively. Analysis of annular or circumferential ﬁns for cylindrical surfaces is given in Sect. 3.15. The equations derived for triangular, trapezoidal and circumferential ﬁns are very inconvenient for
190
3
Extended Surfaces (Fins)
calculations. The heat flow rate can be determined approximately for most of the engineering applications from Eq. (3.48) knowing the ﬁn efﬁciency from efﬁciency plots in Fig. 3.29 for rectangular, triangular and circumferential ﬁns of rectangular crosssection, surface area of the ﬁn and the temperature difference between the wall and the ambient air. A signiﬁcant number of illustrative examples have been included in the chapter.
Review Questions 3:1 Write a short note on heat transfer enhancement from a surface. 3:2 Why ﬁns are used? Discuss various types of ﬁns used in engineering applications. 3:3 Derive the expression for temperature distribution and total heat flow rate under steadystate conditions for a ﬁn of uniform crosssection which is so long that the temperature of the end of the ﬁn can be assumed to be equal to the surrounding temperature. 3:4 Derive the expression for temperature distribution and total heat flow rate under steadystate conditions for a uniform crosssection ﬁn of length L. The heat transfer from the free end of the ﬁn may be neglected. Explain the different parameters that form a basis for assessing the utility of a ﬁn. 3:5 Derive the equation for heat dissipation by a ﬁn of uniform crosssection if the area of the ﬁn end is a very small proportion of the total ﬁn surface area and its contribution to the heat dissipation can be neglected. 3:6 Show that the temperature distribution for a ﬁn of ﬁnite length, when the heat loss from the ﬁn end by convection is taken into account, is given by t t1 cosh mðL xÞ þ ðhL =mkÞ sinh mðL xÞ ¼ ðhL =mkÞ sinh mL þ cosh mL ts t1 Hence, show that the heat transfer from this ﬁn is given by qfin ¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ðhL =mkÞ þ tanh mL hPkAc ðts t1 Þ ðhL =mkÞ tanh mL þ 1
where hL is the heat transfer coefﬁcient from the ﬁn end. 3:7 How would you introduce correction in the ﬁn length to take account of the heat transfer from the ﬁn end? What will be the form of equation of the heat transfer from the ﬁn? 3:8 Deﬁne the efﬁciency and effectiveness of a ﬁn. Discuss the parameters that affect the ﬁn efﬁciency and effectiveness. 3:9 Show that the total efﬁciency of a ﬁnned wall is given by gtotal ¼ 1
Af ð1 gfin Þ Afw
Review Questions
191
2
1
L
Fig. 3.31 Q. 3.16
where Af Afw ηﬁn
= surface area of all ﬁns attached to the wall = total surface area of the ﬁnned wall rejecting heat = ﬁn efﬁciency.
3:10 Discuss the design of thermometer well to minimize the error in the measurement of the temperature of fluids flowing through a duct. 3:11 ‘Fins are more effective for surfaces with low value of surface heat transfer coefﬁcient’. Justify the statement. 3:12 Show that ﬁns provided on a surface will increase the heat transfer rate from the surface only if the Biot number is less than 1. 3:13 A thin rod of length L has its ends connected to two walls, which are maintained at temperatures T1 and T2, respectively. The rod loses heat to the surrounding fluid at T∞ by convection. Derive the equations of temperature distribution in the rod and total heat rejection rate from the entire surface of the rod. 3:14 Present a scheme of experimental setup to measure heat transfer coefﬁcient from a ﬁn surface. 3:15 Show that for the triangular proﬁle ﬁn shown in Fig. 3.26, the temperature variation equation is pﬃﬃﬃﬃﬃ I0 2 bx pﬃﬃﬃﬃﬃﬃ h ¼ hs I0 2 bL where Io(a) is Bessel’s function, b = (2hL/kd), hs = temperature excess at the ﬁn base and d is thickness of the ﬁn at the base. Determine the total heat rejected by the ﬁn. 3:16 Figure 3.31 shows a ﬁn of trapezoidal proﬁle. Derive the equation for the temperature distribution. Neglect the heat transfer from the tip of the ﬁn. h1 and h2 are the temperature excesses above the ambient. The width of the ﬁn (perpendicular to the plane of the paper) is very large compared to the height L of the ﬁn. Fin thicknesses at the base and tip are d1 and d2, respectively.
192
3
Extended Surfaces (Fins)
Problems 3:1 A part of a very long 25mmdiameter copper rod [k = 370 W/(m K)] is inserted into a furnace. Temperatures at two points 100 mm apart along its length are measured to be 150°C and 120°C when the steady state was achieved. If the ambient air is at 30° C, estimate the effective heat transfer coefﬁcient from the rod surface. [Hint: Apply the ﬁn equation for temperature distribution, i.e. t1 t1 ¼ exp½mðx2 x1 Þ t2 t1 [Ans. m = 2.88 m−1; h = 19.2 W/(m2 K)] 3:2 One end of a long rod is inserted into a furnace, while the other end projects into outside air. Under steady state, the temperature of the rod is measured at two points 75 mm apart and found to be 125°C and 88.5°C when the ambient temperature is 20° C. The rod diameter is 250 mm, and the thermal conductivity of the rod material is 3.0 W/(m K). Find the convective heat transfer coefﬁcient. [Ans. Given k = 3.0 W/(m K), P = pD = p 0.25 m, Ac = (p/4) D2 = (p/ t t1 4) (0.25)2 m2 and h is unknown. For a very long ﬁn, ¼ emx , where t s t1 qﬃﬃﬃﬃﬃ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ hP ¼ 2:31 h; Distance x1 may be considered to be zero m ¼ kA ¼ hp0:254 c pð0:25Þ2 3 (refer to Example 3.23), then ts = 125°C and t = 88.5°C at x = 0.075 m. Substitution pﬃﬃ ð2:31 h0:075Þ gives 88:520 or h = 6.08 W/(m2 K).] 12520 ¼ e 3:3 One end of a long rod is inserted into a furnace, while the other end projects into outside air. Under steadystate condition, the temperatures of the rod at 50 mm and 125 mm from the furnace wall measured with the help of thermocouples indicate 125°C and 88.5°C when the ambient temperature is 20°C. Find the temperature of the rod at the furnace wall. 1 [Ans. For a very long ﬁn, we have ttt ¼ emx ; for the given temperatures of t1 = 125° s t1 125 20 ¼ e0:05m and C at x1 = 0.05 m and t2 = 88.5°C at x2 = 0.125 m, we have ts 20 88:5 20 125 20 ¼ e0:125m , solution gives m =5.695 m−1. With this value of m, ¼ ts 20 ts 20 e0:05m gives ts = 159.6°C.] 3:4 Heat generated in a bearing causes the temperature of the end of a shaft to rise to 50° C above the ambient. The shaft is 50 mm in diameter and 800 mm long. Determine the equation of temperature distribution along the shaft if the heat transfer coefﬁcient for the shaft surface is 10 W/(m2 K) and the thermal conductivity of the shaft material is 45 W/(m K). What is the shaft temperature 100 mm away from the heated end? [Ans. m = 4.216 m−1, mL = 3.373 > 3, hence, the equation of a very long ﬁn can be used, which gives h =50e−4.216x. Temperature at 100 mm from the heated end, that is, at x = 100 mm is 32.8°C above the ambient.] 3:5 The head of a solid steel valve is at 600°C. The stem of the valve has a diameter of 10 mm and is cooled by water at 60°C. The heat transfer coefﬁcient is 60 W/(m2 K). The length of the stem is 200 mm. Determine temperature of the stem 50 mm from the heated end. k = 40 W/(m K).
Problems
3:6
3:7
3:8
3:9
193
[Ans. m = 24.495, mL = 4.9 > 3. The temperature distribution equation of a very long ﬁn gives T = 218.67°C. Equation (3.9) with end correction gives 218.75°C.] Heat dissipation by convection from a metal tank containing cooling oil is to be increased by 50 per cent by addition of ﬁns to the wall. The ﬁns will be 10 mm thick and spaced 100 mm apart between the centres. The surface temperature of the tank is 100°C, and the surrounding air temperature is 20°C. Determine the height of each ﬁn on the assumption that the convective heat transfer coefﬁcient remains unchanged and the surface temperature of the tank is expected to drop to 95°C when ﬁns are ﬁtted. The thermal conductivity of the ﬁns and tank material is 200 W/(m K). Take for simplicity 1 m 1 m surface area of the tank. h = 20 W/(m2 K). [Ans. Heat dissipation without ﬁns, q = hAc(ts − t∞) = 1600 W/m2. Desired heat dissipation rate = 1.5 1600 = 2400 W. Fin width = width of the tank = 1 m; number of ﬁns, n = (height of tank/ﬁn pitch) = 1000/100 = 10; ﬁn thickness, d = 10 mm. Surface area not containing ﬁns = 1 1 – 10 10/1000 = 0.9 m2; heat dissipation by this surface area = 20 0.9 (95 – 20) = 1350 W; Heat to be dissipated by the ﬁns = 2400–1350 = 1050 W; Heat to be dissipated by one ﬁn = 1050/10 = 105 W; Fin perimeter, P = 2(W + d) 2.0 m; Ac = 1 10/1000 = 1 10−2; m = 4.47. Considering ﬁn end correction, qﬁn = √(hPkAc) (ts − t∞) tanh (mLc) gives tanh(mLc) = 0.1565 and mLc= 0.1578; Lc = 0.1578/4.47 = 35.3 mm; Required length of ﬁn = Lc – DL = 35.3 − d/2 = 30.3, say 31 mm.] A steel well of 12 mm ID is placed in a duct of 75 mm ID for a thermometer. The fluid temperature in the duct is 300°C. Determine the length of the well so that the error in the temperature measurement is less than 1.5°C. The heat transfer coefﬁcient is 100 W/(m2 K). Wall thickness of the well is 1 mm. The temperature at the well base = 270°C. k = 50 W/(m K). [Ans. m = 44.72; (tL − t∞)/(ts − t∞) = 1/coshmL gives L = 82.5 mm for tL − t∞ = 1.5°C and ts − t∞ = 30°C. As the length of the well is greater than the diameter of the duct, the well is to be located inclined.] For a rectangular section ﬁn, the thickness and length are 5 mm and 100 mm, respectively. The width of the ﬁn is 300 mm. The thermal conductivity of the ﬁn material is 20 W/(m K). How will the effectiveness of the ﬁn change as the heat transfer coefﬁcient changes from 10 to 100 W/(m2 K)? [Ans. First case, h = 10 W/(m2 K): mL = 1.426, Bi = 1.25 10−3, and e = tanhmLc/ √Bi = 25.4; Second case, h = 100 W/(m2 K): mL = 4.509 > 3, Bi = 0.0125 and e = 1/√Bi = 8.94; Fins are more effective when the heat transfer coefﬁcient is low.] 20mmdiameter steel rods [k = 50 W/(m K)] are to be used as ﬁns on one side of a wall. The heat transfer coefﬁcients for the two sides of the wall are 900 W/(m2 K) and 25 W/(m2 K), respectively. The length of the ﬁns is to be restricted to 100 mm. On which side of the wall, would you install the ﬁns? Also determine the heat transfer enhancement or the overall effectiveness if one ﬁn is installed per 2000 mm2 of the wall area. [Ans. For the high h side, Biot number Bi1 = h1D/4k = 900 (0.02/4)/50 = 0.09; for the other side, Bi2 = 0.0025, as Bi2 < Bi1, the ﬁns must be installed on the low heat transfer coefﬁcient side. m = √(hP/kAc) = 10 and Lc = L + D/4 = 105 mm. mLc = 1.05, qﬁn = √(hPkAc) (ts − t∞) tanhmLc. Heat transfer from ﬁnned surface = h (2000 10−6 – p/4 D2) (ts − t∞) + √(hPkAc) (ts − t∞) tanhmLc = 0.16495(ts − t∞). Heat transfer from 2000 mm2 area without ﬁns is
194
3
Extended Surfaces (Fins)
h (2000 10−6) (ts − t∞) = 0.05(ts − t∞). The ratio of these heat transfer rates = 3.3, which is the enhancement.] 3:10 The following data refer to a uniform section ﬁn. Thickness = 6 mm; length = 60 mm; width = 1 m; temperature excess at the base = 100°C; h = 10 W/(m2 K); thermal conductivity of the ﬁn material = 60 W/(m K). Determine the temperature at the ﬁn end and the efﬁciency of the ﬁn. [Ans. m = 7.45 m−1; mLc = 0.47; qﬁn = √(hPkAc) (ts − t∞) tanhmLc = 117.93 W; h2/hs = 1/coshmLc gives h2 = 89.89°C; Fin efﬁciency = 117.93/[hPLc(ts − t∞)] = 93.04%.] 3:11 Determine the end temperature and heat flow from a straight trapezoidal ﬁn of length L = 75 mm, width = 1 m, d1 = 2 mm, d2 = 0.5 mm, h = 250 W/(m2 K), k = 400 W/(m K) and h1= 100°C. [Ans. L1 = 0.1 m, L2 = 0.025 m, b = 2hL1/(kd1) = 62.5, 2√(bL2) = 2.5, 2√(bL1) = 5, I0[2√(bL1)] = 27.24, I0[2√(bL2)] = 3.301, I1[2√(bL1)] = 24.336, I1[2√(bL2)] = 2.5265, K0[2√(bL1)] = 0.00369, K0[2√(bL2)] = 0.0628, K1[2√(bL1)] = 0.00404, K1[2√(bL2)] = 0.0745; From Eq. (3.44) for L2 = 0.025 m, h2= 19.85°C; From Eq. (3.45), q = 1768.6 W.] 3:12 In order to increase the heat transfer rate from a cylindrical pipe (OD = 100 mm), circumferential ﬁns of 4 mm thickness and 50 mm height are added at a pitch of 10 mm. If the heat transfer coefﬁcient h = 40 W/(m2 K) and the thermal conductivity of the ﬁn and pipe material is 50 W/(m K), determine the heat transfer enhancement. [Ans. Without ﬁns, q = 12.566 (ts − t∞); For the circumferential ﬁns, efﬁ1/2 ciency 69% from Fig. 3.29b for L3/2 0.73 and r2c/r1 2; Fin surface c (h/kAm) 2 2 area = 2p(r2 − r1), qﬁn = ﬁn efﬁciency ﬁn surface area h (ts − t∞) = 1.3, 4.0 (ts − t∞); From Eq. (3.47), qﬁn = 1.3(ts − t∞) for b = √(2 h/kd) = 20 (where d = t), br1 = 1, br2 = 2, Io(br1) = 1.2661, I1(br1) = 0.5652, I1(br2) = 1.591, Ko(br1) = 0.421, K1(br1) = 0.602, K1(br2) = 0.14. Number of ﬁns/m = 100; Heat transfer from bare surface = 0.6 12.566(ts − t∞); Heat rejection from the ﬁnned tube = number of ﬁns 1.3 (ts − t∞) + 0.6 12.566(ts − t∞) = 137.54(ts − t∞); Enhancement = 137.54/12.566 = 10.95.]
References Eckert ERG, Drake RM Jr (1972) Analysis of heat and mass transfer. International student edition, McGrawHill Kogakusha, Tokya Gardner KA (1945) Efﬁciency of extended surfaces. Trans ASME. J Heat Transfer 67:621–631 Holman JP (1992) Adapted for SI units by White PRS. Heat transfer, McGrawHill Book Co, New York Isachenko VP, Osipova VA, Sukomel AS (1977) Heat transfer, 2nd edn. MIR Publishers, Moscow
4
Conduction with Heat Generation
4.1
Plane Wall with Uniform Heat Generation
Figure 4.1 shows a plane wall of thickness d. Heat is generated within the wall at uniform rate qg per unit volume and is liberated over the entire volume. The generated heat is transferred to the wall surfaces by conduction and is rejected to the surrounding from one or both faces of the wall. In equilibrium, when the heat generated equals the heat rejected, the wall surface temperatures are T1 and T2. For the analysis, the following assumptions are being made: 1. steadystate conditions, 2. onedimensional heat flow, 3. constant thermal conductivity of the wall material (homogeneous material). Consider an elemental strip of thickness dx at distance x from the left face of the wall. From the Fourier’s conduction equation, the heat entering the left face of the elemental strip is qx ¼ kA
dt dx
ð1:2Þ
The heat leaving the face at distance x + dx is qx þ dx ¼ qx þ
d ðqx Þdx dx
Net heat leaving the elemental strip is d ðqx Þdx dx d dt kA ¼ dx dx dx 2 d t dx ¼ kA dx2
qx þ dx qx ¼
© Springer Nature Singapore Pte Ltd. 2020 R. Karwa, Heat and Mass Transfer, https://doi.org/10.1007/9789811539886_4
195
196
4 t
Conduction with Heat Generation
δ Temperature profile
t1 t2 qx
qx+dx
x
dx
x
Fig. 4.1 Conduction with heat generation in a plane wall
The heat generated in the elemental strip is q0g ¼ ðAdxÞqg where Adx is the volume of the elemental strip and qg is the rate of heat generation per unit volume. In the equilibrium, the total heat generated in the elemental strip must equal the net heat leaving the strip, i.e. kA
d2 t dx ¼ ðAdxÞqg dx2
or d2 t qg ¼0 þ dx2 k
ð4:1Þ
Equation (4.1) can also be obtained from the general heat conduction equation by assuming onedimensional steadystate heat conduction with constant thermal conductivity. Integrating Eq. (4.1), we get dt qg ¼ x þ C1 dx k and t¼
qg 2 x þ C1 x þ C2 2k
ð4:2Þ
The constants of the integration in Eq. (4.2) are determined from the boundary conditions. We consider two different cases.
4.1.1 Case (A) Surfaces at Different Temperatures Heat is rejected from both the surfaces and the surfaces are at different temperatures t1 and t2. This may happen when the heat rejection rates from the two faces are not the same due to the different surface heat transfer coefﬁcients at the two faces. This gives the following boundary conditions:
4.1 Plane Wall with Uniform Heat Generation
197
at x ¼ 0;
t ¼ t1
and; at x ¼ d;
t ¼ t2
Applying the boundary conditions to Eq. (4.2), we get at x ¼ 0; and; at x ¼ d; t2 ¼
t1 ¼ C2
qg 2 qg d þ C1 d þ C2 ¼ d2 þ C1 d þ t1 2k 2k
This gives C1 ¼
t2 t1 qg þ d d 2k
Inserting the value of C1 and C2 in Eq. (4.2), we get t¼
t t qg 2 t2 t1 qg qg 2 1 x þ þ d x þ t1 ¼ ðd xÞx þ x þ t1 2k d 2k 2k d
ð4:3Þ
Equation (4.3) is the equation of temperature distribution through the wall. For a given value of heat generation rate qg, the temperature distribution is a function of the distance x only and the equation can be written as tðxÞ ¼
t t qg 2 1 ðd xÞx þ x þ t1 2k d
ð4:4Þ
Equation (4.4) can be transformed into nondimensional form as follows: tðxÞ t2 ¼
t t qg 2 1 ðd xÞx þ x þ t1 t2 2k d
or tðxÞ t2 qg ðd xÞx x ¼ þ1 d t1 t2 2k t1 t2 2 qg d x x x ¼ þ 1 1 d d d 2kðt1 t2 Þ x x x þ 1 ¼B 1 d d d x Bx þ1 ¼ 1 d d or tðxÞ t2 x Bx þ1 ¼ 1 d d t1 t2 where
ð4:5Þ
198
4
B¼
Conduction with Heat Generation
q g d2 : 2kðt1 t2 Þ
ð4:6Þ
Parameter B is a dimensionless term.
4.1.1.1 The Maximum Temperature and Its Location Within the Wall It can be determined by differentiating Eq. (4.5) with respect to the dimensionless distance (x/d) and setting the derivative to zero. d tðxÞ t2 Bx x þ1 þB 1 ¼0¼ dðx=dÞ t1 t2 d d Bx x ¼ 1þB B d d 2x ¼B 1 1 d This gives the condition as x B1 ¼ d 2B
ð4:7Þ
Equation (4.7) gives the value of dimensionless distance x/d, where the maximum value of the temperature occurs. By substituting this value in Eq. (4.5), we get the value of maximum temperature as
tmax t2 ð B þ 1Þ 2 ¼ t1 t2 4B
ð4:8Þ
Effect of factor B on the temperature distribution: The factor B has been deﬁned as
qg d2 B¼ 2kðt1 t2 Þ
ð4:6Þ
(i) When there is no heat generation, i.e. qg = 0, factor B is zero and Eq. (4.5) reduces to
tðxÞ t2 x ¼ 1 ; d t1 t2
ð4:9aÞ
which gives the temperature distribution as x tðxÞ ¼ t1 þ ðt2 t1 Þ d The temperature distribution is linear in this case as shown in Fig. 4.2.
ð4:9bÞ
4.1 Plane Wall with Uniform Heat Generation
199
t ( x) − t2 t 1 − t2 1 1
1 0
3 2
0
B
1
x
δ Fig. 4.2 Dimensionless temperature distribution as a function of factor B
For the case of heat generation, the positive value of qg, factor B is positive. For B = 1, Eq. (4.7) gives (x/d) = 0, i.e. the maximum value of the temperature occurs at x = 0. With the increase in the value of B, the value of (x/d) given by (4.7) increases and in the limit approaches 0.5 when B has a very high value, refer Fig. 4.2. In general with heat generation, the maximum temperature is located within the wall (x/d = 0–0.5). When the wall absorbs heat, qg is negative and factor B will have a negative value. The rate of heat transfer from any face of the wall is q ¼ kA
dt dx
The value of the derivative (dt/dx) from Eq. (4.4) is t t dt qg qg 2 1 ¼ d xþ dx 2k k d t t qg 2 1 ¼ ðd 2xÞ þ 2k d Substituting the value of dt/dx, we get q ¼ kA
ht t q i 1 2 g ðd 2xÞ d 2k
ð4:10Þ
When there is no heat generation, q ¼ kA
t t 1 2 d
With the heat generation, the heat transfer from the left face, qL, can be obtained by putting x = 0 in Eq. (4.10) to give ht t q i 1 2 g d d 2k t t 1 2 ¼ kA ð1 B Þ d
qL ¼ kA
Similarly, the heat transfer from the right face, qR, can be obtained by putting x = d in Eq. (4.10) to give
200
4
Conduction with Heat Generation
ht t q i 1 2 g þ d d 2k t t 1 2 ¼ kA ð 1 þ BÞ d
qR ¼ kA
The total heat loss from the wall is qt ¼ qL þ qR
ð4:11Þ
where the negative sign with qL indicates heat flow from the left face is in the negative xdirection. Substitution of values of qL and qR in Eq. (4.10) gives qt ¼ Adqg i.e. the heat from the wall equals the heat generated in the wall. It can be seen that when B = 1, i.e. tmax at x = 0, qL = 0, which means that there is no heat loss from the lefthand face and the total heat is rejected from the righthand face. Example 4.1 Two steel plates are separated by a steel rod of 25 mm diameter and 300 mm long. The rod is welded to each plate. The space between the plates is ﬁlled with insulation material as shown in Fig. 4.3. Voltage difference between the two plates causes the current to flow through the steel rod. Due to the flow of current, heat is generated in the rod at a rate of 15 W. Under steadystate condition, the temperatures at the ends are 100°C and 70°C. Find the maximum temperature in the rod. Also, calculate the heat flow from each end. The thermal conductivity of the steel rod is 40 W/(m K). Solution Heat generated per unit volume 15 ¼ 101:86 103 ðp=4Þ ð0:025Þ2 0:3 qg d2 101:86 103 0:32 B¼ ¼ ¼ 3:82 2kðt1 t2 Þ 2 40 ð100 70Þ qg ¼
tmax t2 ð3:82 þ 1Þ2 ¼ 1:52 ¼ t1 t2 4 3:82
Plates Insulation
Fig. 4.3 Example 4.1
... ............ Rod ... ............ ... ............ ... ............ ............ t1 ... t2 ... ............ ... ............ ... ............ ... ............ ... ............ 300 ... ............ ... ............ ... ............ ... ............
4.1 Plane Wall with Uniform Heat Generation
201
or tmax ¼ 1:52ðt1 t2 Þ þ t2 ¼ 1:52ð100 70Þ þ 70 ¼ 115:6 Heat transfer from the left face is qL ¼ kA
ht t q i 1 2 g d d 2k
¼ 40 ðp=4Þ ð0:025Þ2
100 70 101:86 103 0:3 0:3 2 40
¼ 5:535 W: Heat transfer from the right face is ht t q i 1 2 g þ d d 2k 100 70 101:86 103 ¼ 40 ðp=4Þ ð0:025Þ2 0:3 þ 0:3 2 40
qR ¼ kA
¼ 9:461 W: Total heat flow from the rod ends qL and qR equals the heat generation. Example 4.2 Heat is generated at an interface between two slabs. One is steel [k = 35 W/(m K)] of 50 mm thickness and the other is brass [k = 70 W/(m K)] of 50 mm thickness. The temperatures of the outer surfaces of the steel and brass slabs are maintained at 100°C and 60°C, respectively. The heat generation rate at the contact surface of the slabs is 1.5 105 W/m2. Determine the heat flow through the steel slab. Solution Let the temperature at the interface of the slabs is ti. From Fourier’s heat conduction equation, the heat flow through the left slab is (refer Fig. 4.4) ti t1 qL ¼ k1 A d1 ti 100 ¼ 35 1 ¼ 700 ðti 100Þ: 0:05
Steel
Brass
t1 ● ● ti
Fig. 4.4 Example 4.2
k1
k2
δ1
δ2
● t2
202
4
Conduction with Heat Generation
and through the right slab is ti t2 d2 ti 60 ¼ 70 1 ¼ 1400 ðti 60Þ: 0:05
qR ¼ k 2 A
The total heat flow must equal the heat generated. Thus qL þ qR ¼ qg or 700 ðti 100Þ þ 1400 ðti 60Þ ¼ 1:5 105 or ti ¼ 144:76 C: Knowing the interface temperature ti, the values of qL and qR through the steel and brass, respectively, are qL ¼ 3:13 104 W qR ¼ 1:18 105 W:
4.1.2 Case (B) Surfaces at the Same Temperature This gives the following boundary conditions: at x ¼ 0;
t ¼ t1
and; at x ¼ d;
t ¼ t1
Applying the boundary conditions to Eq. (4.2), we get at x ¼ 0; and at x ¼ d; t1 ¼
t1 ¼ C2
qg 2 qg d þ C1 d þ C2 ¼ d2 þ C1 d þ t1 2k 2k
This gives C1 ¼
qg d 2k
Inserting the value of C1 and C2 in Eq. (4.2) we get t¼
qg 2 qg qg x þ d x þ t1 ¼ ðd xÞx þ t1 2k 2k 2k
ð4:12Þ
4.1 Plane Wall with Uniform Heat Generation
203
Equation (4.12) is the equation of temperature distribution through the wall. For a given value of heat generation rate qg, the temperature distribution is a function of the distance x only and the equation can be written as tðxÞ ¼
qg ðd xÞx þ t1 2k
ð4:13Þ
The above equation can also be obtained by putting t2 = t1 in Eq. (4.4). It is obvious that in the present case, the maximum temperature must be at x = d/2 and qg d d þ t1 d 2 2 2k qg ¼ d2 þ t1 8k
tmax ¼
ð4:14Þ
The heat flow from the right face is dt qR ¼ kA dx x¼d or qR ¼ kA
q g
2k
1 d ¼ ðAdÞ qg 2
From the symmetry, the total heat transfer rate qt is qt ¼ 2qR ¼ ðAdÞ qg The heat conducted to the face qR is rejected to the surrounding by convection. If the surrounding fluid is at temperature t∞ and h is the heat transfer coefﬁcient at the wall, then ðAdÞ qg ¼ hAðt1 t1 Þ 2 or t1 ¼
qg d þ t1 2h
Substituting the value of t1 in Eq. (4.13), we get t¼
qg qg ðd xÞx þ d þ t1 : 2k 2h
ð4:15Þ
The maximum temperature is at the midpláne of the wall, and its value can be obtained by putting x = d/2. Example 4.3 An electric current I = 10A flows through a metal strip of rectangular crosssections (8 mm 1 mm) and 1 m long. The voltage difference between the strip ends is 230 V. Find the temperature in the mid of strip thickness if the temperature of the surface of the strip is 200°C. The thermal conductivity of the strip material is 25 W/(m K).
204
4
Conduction with Heat Generation
Solution Heat generated per unit volume, qg ¼
V I 230 10 ¼ ¼ 2:875 108 W/m3 : ðcrosssectionÞ length 8 1 106
Temperature at the midplane of the strip is qg 2 d þ ts 8k 2:875 108 0:0012 þ 200 ¼ 201:44 C: ¼ 8 25
tmax ¼
Example 4.4 A 100 mm thick plane wall generates heat at the rate of 1 105 W/m3. One side of the wall is exposed to the ambient air at 30°C while the other side of the wall is insulated. The convective heat transfer coefﬁcient on the airside is 400 W/(m2 K). Proceeding from the basic principles, determine the location and value of the maximum temperature to which the wall is subjected. The thermal conductivity of the wall material is 2 W/(m K). Solution The onedimensional steadystate conduction equation with heat generation is d2 t qg ¼0 þ dx2 k
ð4:1Þ
Integrating the above equation, we get dt qg ¼ x þ C1 dx k
ðiÞ
and t¼
qg 2 x þ C1 x þ C2 2k
ðiiÞ
The constants C1 and C2 are to be found from the boundary conditions of the problem. At x = 0, the wall is insulated and hence dt ¼ 0; dx which gives C1 = 0. At x = d, the heat reaching the wall by conduction is rejected to the surrounding by convection (refer Fig. 4.5). Thus dt kA ¼ hA½tðdÞ t1 dx x¼d
4.1 Plane Wall with Uniform Heat Generation
100
205
● t∞ = 30oC h = 400 W/(m2 K)
x=0
x=δ
Fig. 4.5 Example 4.4
or
dt h ¼ ½tðdÞ t1 dx x¼d k
ðiiiÞ
From Eq. (i), with C1 = 0, dt/dx at x = d is dt qg ¼ d dx x¼d k and from Eq. (ii), t(d) is tðdÞ ¼
qg 2 d þ C2 2k
ðivÞ
Substitution of the values of dt/dx and t(d) in Eq. (iii), we have i qg h h qg d ¼ d2 þ C2 t1 k k 2k which gives C2 ¼
qg qg d þ d2 þ t1 h 2k
Putting the value of C2 in Eq. (ii), we get the equation of the temperature distribution through the wall as t¼
qg 2 qg qg x þ d þ d2 þ t 1 2k h 2k
At the insulated wall side, dt/dx = 0, hence the temperature is maximum. Putting x = 0, we get the maximum temperature qg qg d þ d2 þ t 1 h 2k 1 105 1 105 0:1 þ 0:12 þ 30 ¼ 305 C ¼ 400 22
tmax ¼
206
4.2
4
Conduction with Heat Generation
Cylinder with Uniform Heat Generation
Consider an elemental cylinder of thickness dr at distance r from the central axis in Fig. 4.6. From the Fourier’s conduction equation, the heat entering the shell surface at radius r is dt dr
ðiÞ
d ðqr Þdr dr
ðiiÞ
qr ¼ k2prL The heat leaving the face at distance r + dr is qr þ dr ¼ qr þ Net heat leaving the elemental cylinder is
d ðqr Þdr dr d dt k2prL ¼ dr dr dr d dt r dr ¼ k2pL dr dr
qr þ dr qr ¼
ðiiiÞ
The heat generated in the elemental cylindrical cell is q0g ¼ ð2prLdrÞqg
ðivÞ
where 2rLdr is the volume of the elemental cylinder and qg is the rate of heat generation per unit volume. In the equilibrium, the total heat generated in the elemental cylinder must equal the net heat leaving the elemental cylinder, i.e. d dt k2pL r dr ¼ ð2prLdrÞqg dr dr or
d dt qg r þ r¼0 dr dr k
qr
qr+dr
dr r
Fig. 4.6 Conduction with heat generation in a cylinder
ð4:16Þ
4.2 Cylinder with Uniform Heat Generation
207
Equation (4.16) can also be obtained from the general heat conduction equation by assuming onedimensional steady state heat conduction with constant thermal conductivity. Integrating Eq. (4.16), we get dt qg r 2 r þ þ C1 ¼ 0 dr k 2 or dt qg C1 þ rþ ¼0 dr 2k r
ðvÞ
qg 2 r þ C1 ln r þ C2 ¼ 0 4k
ðviÞ
Further integration gives tþ
The constants of the integration in Eq. (vi) are determined from the boundary conditions, which are at r ¼ 0;
dt=dr ¼ 0
and at r ¼ R; i:e:at the surface
t ¼ t1
Applying the ﬁrst boundary condition to Eq. (v), we get C1 ¼ 0 Inserting C1 = 0 and applying the second boundary condition to Eq. (vi) gives t1 þ
qg 2 R þ 0 þ C2 ¼ 0 4k
or C2 ¼
q
g
4k
R2 þ t1
Inserting the value of C2 in Eq. (vi) we get tþ
qg 2 qg 2 r R þ t1 ¼ 0 4k 4k
or t¼
qg 2 ðR r 2 Þ þ t1 4k
ð4:17Þ
which is the equation of temperature distribution through the cylindrical wall. The heat conducted to the surface of the cylinder is transferred by convection to the surrounding fluid, hence q ¼ hAðt1 t1 Þ dt ¼ hAðt1 t1 Þ or kA dr q r¼R g or k R ¼ hðt1 t1 Þ 2k
208
4
Conduction with Heat Generation
or t1 ¼ t1 þ
qg R 2h
Substituting the value of t1 in Eq. (4.17), we get t¼
qg 2 qg ðR r 2 Þ þ R þ t1 4k 2h
ð4:18Þ
The maximum temperature is at the centre r = 0 where dt/dr = 0 and is tmax ¼
qg 2 qg R þ R þ t1 4k 2h
ð4:19Þ
Example 4.5 A current of 100 A flows through a stainless steel wire of 5 mm diameter and 1 m long. The wire is submerged in a liquid bath maintained at 10°C. The heat transfer coefﬁcient for the wire surface is 500 W/(m2 K). If the thermal conductivity of the wire material is 20 W/(m K) and resistivity is 70 10−6 Xcm, calculate the centreline temperature of the wire. Solution Resistance of the wire Re ¼ ¼
qL A 70 106 100 p ð0:25Þ2
¼ 0:03565 X::
Heat generated due to ohmic heating of wire I 2 Re ¼ 1002 0:03565 ¼ 356:5 W: Heat generation rate per unit volume qg ¼
356:5 p ð2:5=1000Þ2 1
¼ 1:8156 107 W/m3 :
Maximum temperature from Eq. (4.19) is qg 2 qg R þ R þ t1 4k 2h 1:8156 107 2:5 2 1:8156 107 2:5 ¼ þ þ 10 1000 1000 4 20 2 500
tmax ¼
¼ 56:81 C:
4.2 Cylinder with Uniform Heat Generation
209
Example 4.6 A plate with uniformly distributed inner heat source with a volumetric heat generation rate of qg is exposed on both sides to a flowing fluid. The local coefﬁcients of heat transfer from the surfaces of the plate to the surrounding fluid and the bulk fluid temperature on the two sides of the plate are h1 and h2 and tf1 and tf2, respectively. The thickness of the plate is d and thermal conductivity is k. (a) Show that when the maximum temperature is at the middle of the plate 1 1 1 tf 2 tf 1 ¼ qg d h1 h2 2 (b) Show that when tf1 = tf2 and h2 = 2h1, location of the maximum temperature is given by xo h1 d þ k ¼ 2h1 d þ 3k d Solution (a) For the maximum temperature, with uniform volumetric heat generation rate, to occur at the middle t1 ¼ t2 The heat lost from the left face is given by, refer Fig. 4.7, t t qg 1 2 þ d qL ¼ kA d 2k
ðiÞ
Hence, for t1= t2, qL ¼ Ad
qg 2
ðiiÞ
The heat lost from the left face is also given by qL ¼ h1 Aðt1 tf 1 Þ
ðiiiÞ
Equating Eqs. (ii) and (iii), ðt1 tf 1 Þ ¼
● t1 tf1 ●
Fig. 4.7 Example 4.6
qg d 2h1
ðivÞ
t2● δ/2
δ/2
● tf2
210
4
Similarly for the righthand face, ðt2 tf 2 Þ ¼
Conduction with Heat Generation
qg d 2h2
ðvÞ
From Eqs. (iv) and (v) for t1 = t2, qg d 1 1 ðtf 2 tf 1 Þ ¼ : 2 h1 h2 (b) From Eqs. (i) and (iii), t t qg 1 2 þ d ¼ h1 Aðt1 tf 1 Þ kA d 2k or t t qg h1 1 2 þ d ¼ ðt1 tf 1 Þ d 2k k
ðviÞ
t t qg h2 1 2 þ d ¼ ðt2 tf 2 Þ d 2k k
ðviiÞ
Similarly at the right face,
Putting h2 = 2h1 and tf2 = tf1 in the above equation, we obtain t t qg 2h1 1 2 þ d ¼ ðt2 tf 1 Þ d 2k k
ðviiiÞ
Multiplying both sides of Eq. (vi) by 2, we get 2
t1 t2 qg 2h1 þ d¼ ðt1 tf 1 Þ d k k
Subtracting Eq. (viii) from Eq. (ix), we get 3 ¼
t1 t2 qg 2h1 þ d¼ ðt1 tf 1 t2 þ tf 1 Þ d 2k k
2h1 ðt1 t2 Þ k
or 3 þ Putting
qg d2 2kðt1 t2 Þ
qg 2h1 d2 ¼ d 2kðt1 t2 Þ k
¼ B we get 3 þ B ¼
2h1 d k
ðixÞ
4.2 Cylinder with Uniform Heat Generation
211
The simpliﬁcation gives the value of parameter B as B¼
2h1 dþ3 k
The position of maximum temperature from Eq. (4.7), x B1 ¼ d 2B Substituting the value of B, we get x 2hk 1 d þ 3 1 ¼ d 2 2hk 1 d þ 3 or xo h1 d þ k ¼ : 2h1 d þ 3k d Example 4.7 A 5 mm thick plate with uniformly distributed internal heat source having a volumetric rate of heat libration qg = 3.0 108 W/m3, is exposed on both sides to cooling fluids. The local coefﬁcients of heat transfer from the two sides of the plate to the fluid to which it is exposed are 3000 W/(m2 K) and 1800 W/(m2 K) and the temperatures of the fluid are 140°C and 150°C, respectively. If the thermal conductivity of the plate material is 30 W/ (m K), determine position and magnitude of the maximum temperature in the plate and temperature of both sides of the plate. Solution From Example 4.6,
t1 t2 qg h1 þ d ¼ ðt1 tf 1 Þ d 2k k
or
t1 t2 3:0 108 3000 ðt1 140Þ þ 0:005 ¼ 30 0:005 2 30
or 1:5t1 þ t2 ¼ 195 Similarly for the righthand face
ðiÞ
212
4
Conduction with Heat Generation
t1 t2 qg h2 þ d ¼ ðt2 tf 1 Þ d 2k k or t1 t2 3:0 108 1800 ðt2 150Þ þ 0:005 ¼ 30 0:005 2 30 or t1 1:3t2 ¼ 170
ðiiÞ
Solving Eqs. (i) and (ii) for t1 and t2, we get t1 ¼ 445:8 C t2 ¼ 473:7 C:
Value of parameter B is B¼
qg d2 3:0 108 0:0052 ¼ ¼ 4:48 2kðt1 t2 Þ 2 30 ð445:8 473:7Þ
The position of maximum temperature from Eq. (4.7), x B1 4:48 1 ¼ ¼ ¼ 0:6116 d 2B 2 ð4:48Þ or x ¼ 3:06 mm: Maximum temperature from Eq. (4.8) is
tmax t2 ð B þ 1Þ 2 ¼ t1 t2 4B
or tmax ¼
ð B þ 1Þ 2 ð4:48 þ 1Þ2 ð445:8 473:7Þ þ 473:7 ¼ 492:55 C: ðt1 t2 Þ þ t2 ¼ 4B 4 ð4:48Þ
Example 4.8 A plate, with a uniform rate of volumetric heat generation qg, is exposed on both sides to flowing fluids. The thickness of the plate is d and the thermal conductivity of the material is k. The temperature of the fluid on one side of the plate is tf1 with heat transfer coefﬁcient h1. Determine the temperature of the fluid on the opposite side of the plate so that the rate of heat flow from this surface is equal to zero.
4.2 Cylinder with Uniform Heat Generation
213
Solution The temperature of the opposite surface of the plate must be equal to the fluid temperature as there is no heat transfer from this surface, i.e. t2 = tf2 since q = h2 A (t2 − tf2) = 0. The total heat generated in the plate, q ¼ qg Ad The heat is rejected from only one face, i.e. h1 A t1 tf 1 ¼ qg Ad or t1 ¼
qg Ad þ tf 1 h1 A
The conduction heat flow to this face is t t qg 1 2 þ d ¼ qg Ad kA d 2k or ðt1 t2 Þ ¼
qg 2 d 2k
Substitution of the values of t2 and t1, we get
qg d q g d2 tf 1 þ tf 2 ¼ h1 2k
or
d 1 þ t f 2 ¼ t f 1 þ qg d : 2k h1
Example 4.9 A hollow cylindrical copper bar, having inner and outer diameters of 13 mm and 50 mm, respectively, carries a current density of 5000 A/cm2. When the outer surface temperature is maintained at 40°C and no heat is removed through the inner surface, ﬁnd the position and value of the maximum temperature. For copper, assume electrical resistivity q = 2 10−6 Ωcm and thermal conductivity k = 381 W/(m K). Solution Heat generation rate per unit volume, 2
qg ¼
I 2 R ð5000 104 AÞ qL ¼ ¼ ð5000 104 Þ2 2 108 ¼ 50 106 W/m3 : AL A AL
214
4
Conduction with Heat Generation
The differential equation for onedimensional heat flow with heat generation in cylindrical coordinates is d dt qg r þ r¼0 dr dr k
ð4:16Þ
Integration of Eq. (4.16) gives r
dt qg r 2 þ þ C1 ¼ 0 dr k 2
or dt qg C1 þ rþ ¼0 dr 2k r Further integration gives the temperature distribution equation as tþ
qg 2 r þ C1 ln r þ C2 ¼ 0 4k
ðiÞ
The constants of the integration in Eq. (i) are determined from the boundary conditions, which are
and
at r ¼ r1 ;
dt=dr ¼ 0 as no heat is removed from the inner surface;
at r ¼ r2 ;
t ¼ t2
The ﬁrst boundary condition gives 0þ
qg C1 r1 þ ¼0 2k r1
or C1 ¼
qg 2 r 2k 1
Inserting value of constant C1 in the temperature distribution equation, we get tþ
qg 2 qg 2 r r1 ln r þ C2 ¼ 0 4k 2k
The second boundary condition gives t2 þ
qg 2 qg 2 r r ln r2 þ C2 ¼ 0 4k 2 2k 1
or C2 ¼
qg 2 qg r ln r2 r22 t2 2k 1 4k
4.2 Cylinder with Uniform Heat Generation
215
Using the numerical values given in the problem, qg 2 50 106 13 2 ¼ 2:772 C 1 ¼ r1 ¼ 2000 2k 2 381 50 106 13 2 50 50 106 50 2 C2 ¼ ln 40 ¼ 70:82 2000 2000 2000 2 381 4 381 Substitution of the values of the constants and other parameters in Eq. (i) gives tþ
qg 2 r þ C1 ln r þ C2 ¼ 0 4k
or tþ
50 106 2 r 2:772 ln r 70:82 ¼ 0 4 381
The maximum temperature is at the inner surface (r = 13/2000 m) and is t¼
50 106 13 2 13 þ 2:772 ln þ 70:82 ¼ 55:47 C: 2000 4 381 2000
The heat transfer at the outer surface of the bar is q ¼ kA
dt dr r¼r2
where dt qg C1 ¼ r2 dr r¼r2 2k r2 ¼
50 106 50 2:772 2000 ¼ 1529:6 50 2 381 2000
Hence, q ¼ 381 2p
50 1 ð1529:6Þ ¼ 91542 W: 2000
Check: The heat transfer must equal the heat generation rate, which is p q ¼ qg ðr2 r1 Þ2 L ¼ 50 106 p 4
"
50 2000
2 # 13 2 1:0 ¼ 91539 W: 2000
216
4
Conduction with Heat Generation
Example 4.10 A fuel element is made in the form of a long hollow cylinder having an inner diameter (2r1) of 20 mm and outer diameter (2r2) = 30 mm. The element is made of uranium with a thermal conductivity k1 = 30 W/(m K). The inner and outer surfaces of the fuel element carry tightly ﬁtting stainless steel cladding of 1 mm thickness. The thermal conductivity of steel k2 is 20 W/(m K). The volumetric rate of heat generation is 6 107 W/m3 and is assumed to be uniform over the crosssection. The fuel element is being cooled with carbon dioxide flowing along an outer channel. The bulk temperature of the coolant in the outer channel is 200°C. The local coefﬁcient of heat transfer from the outer cladding to the coolant is 500 W/(m2 K). Determine the maximum temperature for the fuel element. Solution The heat generated in unit length of the fuel element is q ¼ qg V ¼ qg p ðr22 r12 Þ 1 ¼ 6 107 p ð152 102 Þ 106 1 ¼ 23561:94 W/m length:
If the temperature of the outer surface of the fuel element is t2, then q¼
t2 r3 1 2pk2 L ln r2
tf 2 þ 2pr13 L h12
Substituting L = 1 m, k2 = 20 W/(m K), r3 = (r2 + 1) = 0.016 m, r3/r2 = 16/15, h2 = 500 W/(m2 K) and tf2 = 200°C, we get 23561:94 ¼
t2 200 1 16 1 ln 2p 20 1 15 þ 2p 0:016 1
1 500
or t2 ¼ 680:85 C: From Eq. (4.16), by integration, we have dt qg C1 þ rþ ¼0 dr 2k r
ðiÞ
qg 2 r þ C1 ln r þ C2 ¼ 0 4k
ðiiÞ
Further integration gives tþ The boundary conditions are
and
at r ¼ r1 ¼ 10 mm; dt=dr ¼ 0 as no heat is removed from the inner surface; at r ¼ r2 ¼ 15mm; t ¼ t2 ¼ 680:85 C:
4.2 Cylinder with Uniform Heat Generation
217
Applying the boundary conditions to Eqs. (i) and (ii), we have, for the given data of the problem, qg 6 107 C1 ¼ r12 ¼ ð0:01Þ2 ¼ 100 2k 2 30 and 680:85 þ
6 107 ð0:015Þ2 100 ln ð0:015Þ þ C2 ¼ 0 4 30
or
C2 ¼ 1213:32
Substitution of the values of C1 and C2 in Eq. (ii) we get the temperature distribution equation as tþ
qg 2 r 100 ln r 1213:32 ¼ 0 4k
The maximum temperature t = tmax will occur at the inner surface where dt/dr = 0. Hence, from the above equation, qg 2 r þ 100 ln r1 þ 1213:32 4k 1 6 107 ð0:01Þ2 þ 100 lnð0:01Þ þ 1213:32 ¼ 702:8 C: ¼ 4 30
tmax ¼
Example 4.11 A stainless steel tube with an inner diameter of 9.5 mm and an outer diameter of 10 mm is heated by passing an electric current. The tube carries a current of 300 A. The speciﬁc resistance and the thermal conductivity of the steel are 0.8 X.mm2/m and 18.0 W/(m K), respectively. Calculate the volumetric rate of heat liberation from the tube and the temperature drop across the wall of the tube if all the heat generated in the wall of the tube is transferred from its inner surface. Consider onemeter length of the tube. Solution Electric resistance of the tube, R¼
qL 0:8 1 ¼ ¼ 0:1044 X=m: A ðp=4Þ ð102 9:52 Þ
The rate of heat generation per m length of the tube, q ¼ I 2 R ¼ 3002 0:1044 ¼ 9396 W=m: Volumetric rate of heat generation, qg ¼
I2R 9396 106 ¼ ¼ 1:227 109 W/m3 : AL ðp=4Þ ð102 9:52 Þ 1
218
4
Conduction with Heat Generation
Following the procedure of the previous example, we have dt qg C1 þ rþ ¼0 dr 2k r and tþ
qg 2 r þ C1 ln r þ C2 ¼ 0 4k
The boundary conditions are
and
at r ¼ r1 ;
t ¼ t1 ;
at r ¼ r2 ;
dt=dr ¼ 0 as no heat is removed from the outer surface
Applying the second boundary condition, we have C1 ¼
qg 2 r 2k 2
This gives the temperature distribution equation as tþ
qg 2 qg 2 r r2 ln r þ C2 ¼ 0 4k 2k
The condition t = t1 at r = r1 gives C2 ¼ t1
qg 2 qg 2 r þ r ln r1 4k 1 2k 2
Substituting the value of constant C2 gives the temperature distribution equation as tþ
qg 2 qg 2 qg qg r r2 ln r t1 r12 þ r22 ln r1 ¼ 0 4k 2k 4k 2k
The temperature drop across the wall (t2 − t1), where t2 is the temperature at radius r2, 2 qg r r2 2 þ r22 ln r2 r22 ln r1 þ 1 2k 2 2 " 2 # 1:227 109 52 9:5 9:5 1 2 2 þ 5 ln 5 5 ln ¼ 106 þ 2 2 2 2 18 2
t2 t1 ¼
¼ 2:25 C: Example 4.12 Show that the temperature distribution relation for a long cylinder with uniform heat generation can be expressed in nondimensional form as r 2 t tw ¼1 R tmax tw where tw is the temperature at the outer surface of the cylinder.
4.2 Cylinder with Uniform Heat Generation
219
Solution From Eq. (4.17), we have t¼
qg 2 ðR r 2 Þ þ tw 4k
or t tw ¼
qg 2 ðR r 2 Þ 4k
ðiÞ
The maximum temperature is at r = 0, thus tmax tw ¼ From Eqs. (i) and (ii), we get
qg 2 R 4k
r 2 t tw ¼1 R tmax tw
ðiiÞ
ð4:20Þ
Example 4.13 Show that for an electric current carrying wire, the maximum temperature at the axis of the wire is given by i2 2 tmax ¼ tw þ R 4kke where k and ke are the thermal and electric conductivities of the wire material, respectively, and i is current density (electrical conductivity ke is reciprocal of the electrical resistivity q). Solution From Example 4.12, we have tmax ¼ tw þ
qg 2 R 4k
For an electric current carrying wire, qg = I2Re/AL) = I2 (qL/A)/(AL)= (I/A)2/ke = i2/ke, where Re is the electrical resistance (the electrical conductivity ke is reciprocal of the resistivity). Hence, tmax ¼ tw þ
4.3
i2 2 R 4kke
Solid Sphere with Uniform Heat Generation
Consider an elemental spherical shell of thickness dr at distance r from the centre, see Fig. 4.8. From the Fourier’s conduction equation, the heat entering the shell surface at radius r is qr ¼ kA
dt dt ¼ kð4pr 2 Þ dr dr
220
4
qr
Conduction with Heat Generation
qr+dr
dr r
Fig. 4.8 Conduction with heat generation in a sphere
The heat leaving the face at distance (r + dr) is qr þ dr ¼ qr þ
d ðqr Þdr dr
Net heat leaving the elemental shell is d ðqr Þdr dr d 2 dt 4kpr ¼ dr dr dr d dt r2 dr ¼ 4kp dr dr
qr þ dr qr ¼
The heat generated in the elemental shell is q0g ¼ ð4pr 2 drÞqg where 4pr2dr is the volume of the elemental cell and qg is the rate of heat generation per unit volume. In the equilibrium, the total heat generated in the elemental spherical cell must equal the net heat leaving the elemental sphere, i.e. 4kp
d dt r2 dr ¼ ð4pr 2 drÞqg dr dr
or d dt qg r2 þ r2 ¼ 0 dr dr k
ð4:21Þ
Equation (4.21) can also be obtained from the general heat conduction equation by assuming onedimensional steadystate heat conduction with constant thermal conductivity. Integrating Eq. (4.21), we get r2
dt qg r 3 ¼ þ C1 dr k 3
ðiÞ
4.3 Solid Sphere with Uniform Heat Generation
221
At r = 0, dt/dr = 0 for a solid sphere. Applying this boundary condition, we get C1 ¼ 0 Hence, dt r qg ¼ dr 3k Its integration gives t¼
qg r 2 þ C2 3k 2
ðiiÞ
The constant of the integration C2 can be determined from the following boundary condition. At r ¼ R; t ¼ t2 . Applying the second boundary condition, we get t2 ¼
qg 2 R þ C2 6k
C2 ¼
qg 2 R þ t2 6k
This gives
Inserting the value of C2 in Eq. (ii), we get t¼
qg 2 qg 2 r þ R þ t2 6k 6k
or t¼
qg 2 ðR r 2 Þ þ t2 6k
ð4:22Þ
which is the equation of temperature distribution through the spherical wall. The maximum temperature is at the centre (r = 0), where dt/dr = 0. Hence, tmax ¼
qg 2 R þ t2 6k
ð4:23Þ
The heat conducted to the surface of the sphere is transferred by convection to the surrounding fluid, hence q ¼ hAðt2 t1 Þ or kA
dt ¼ hAðt2 t1 Þ dr r¼R
222
4
Conduction with Heat Generation
or q g k R ¼ hðt2 t1 Þ 3k or t2 ¼ t1 þ
qg R 3h
Substituting the value of t2 in Eq. (4.22), we get the temperature distribution equation in terms of the surrounding temperature t∞ as t¼
qg 2 qg ðR r 2 Þ þ R þ t1 6k 3h
ð4:24Þ
Example 4.14 A 90 mm diameter orange generates 5.25 103 W/m3 heat while it undergoes the ripening process. If the external surface of the orange is at 5°C, determine the temperature at the centre of the orange. Also, determine the heat flow from the outer surface of the orange. The orange may be treated as a spherical body. The average thermal conductivity may be taken as 0.25 W/(m K) for the orange material. Solution The temperature at the centre of the orange from Eq. (4.23) is qg 2 R þ t2 6k 5:25 103 0:09 2 ¼ þ 5 ¼ 12:1 C: 2 6 0:25
tmax ¼
The heat flow from the outer surface of the orange equals the heat generated in the orange and is 4 4 0:09 3 ¼ pR3 qg ¼ p 5:25 103 ¼ 2:0 W: 3 3 2
4.4
Heat Transfer Through Piston Crown
In internal combustion engines, the piston head (crown) is subjected to a very high gas temperature during the combustion. The heat flowing to the piston crown from the hot gases is transferred by conduction through the crown to its edge and then to the cylinder wall. The centre of the crown is at the farthest point from the cylinder wall and is at the highest temperature. A very high temperature at the crown centre may cause piston failure and may also lead to combustion problems in the case of petrol engines. An approximate analysis is presented below to estimate this temperature. Piston head is a thin disc of thickness b which is subjected to uniform heat flux qgas due to convection and radiation from the gases, refer Fig. 4.9. It is assumed in the presented
4.4 Heat Transfer Through Piston Crown
223
qr
qr+dr
dr r Heat of gas to Piston
b Cylinder
Fig. 4.9 Heat transfer through piston crown
analysis that the heat transfer from the lower part of the head to the surrounding air is negligible. Thus, there is only radial heat transfer through the piston crown by conduction. Consider an elemental ring of radial thickness dr at distance r from the centre. From the Fourier’s conduction equation, the heat entering the ring at radius r is qr ¼ kA
dt dt ¼ kð2prbÞ dr dr
The heat leaving the face at distance (r + dr) is qr þ dr ¼ qr þ
d ðqr Þdr dr
Net heat leaving the elemental ring is d ðqr Þdr dr d dt k2prb ¼ dr dr dr d dt r dr ¼ 2pbk dr dr
qr þ dr qr ¼
ðiÞ
The heat incident at the surface of the elemental ring is q0g ¼ ð2prdrÞqgas
ðiiÞ
where 2prdr is the surface area of the elemental ring and qgas is the heat flux due to the gas. In the equilibrium, the heat incident must equal the net heat leaving the elemental ring, i.e. d dt r 2pbk dr ¼ ð2prdrÞqgas dr dr
224
or
4
Conduction with Heat Generation
qgas d dt r r¼0 þ dr dr bk
ð4:25Þ
Integrating the above equation, we get qgas r 2 dt r þ ¼ C1 dr bk 2 At r = 0, dt/dr = 0 for a solid circular plate. Applying this boundary condition, we get C1 ¼ 0 Hence, dt r qgas þ ¼0 dr 2 bk Further integration gives tþ
qgas r 2 ¼ C2 2bk 2
ðiiiÞ
The constant of the integration C2 can be determined from the following boundary condition. At r ¼ R; t ¼ to . Applying the second boundary condition, we get to þ
qgas 2 R ¼ C2 4bk
Inserting the value of C2 in Eq. (iii), we get the equation of temperature distribution as t¼
qgas 2 ðR r 2 Þ þ t0 4bk
ð4:26Þ
The maximum temperature will occur at the centre (r = 0) and is tmax ¼
qgas 2 R þ t0 4bk
ð4:27Þ
The total heat transferred to the piston crown is qtotal ¼ pR2 qgas Example 4.15 The base of an electric iron is made of aluminium. The weight of the base is 1.5 kg. The surface area of the base is 0.04 m2. An electric heater of 1000 W capacity is placed over the base. When heating is turned on, the base is placed in a vertical position. How long the iron will take to reach a temperature of 120°C if at the start of the heating the base is at the room temperature? The room temperature is 20°C and the heat transfer coefﬁcient from the surface of the base to the surrounding is 20 W/(m2 K). Take following data for aluminium q = 2700 kg/m3, c = 900 J/(kg °C), and k = 200 W/(m K).
4.4 Heat Transfer Through Piston Crown
225
Solution The energy balance gives qheater ¼ increase in heat content of the base þ heat lost to the surroundings or qheater ¼ mc
dh þ hAh ds
ðiÞ
where h dh ds
m c A
tbase − tambient rate of increase of the temperature of the base, mass of the base, speciﬁc heat of the base material, area of base transferring heat by convection. Rearranging Eq. (i), we have mc dh qheater ¼ h hA ds hA Let mc/hA = a and qheater/hA = b, then a
dh ¼bh ds
or dh 1 ¼ ds bh a Integration between h = 0 at s = 0 to h = h1 at s = s1 gives 1 ½ lnðb hÞh01 ¼ s1 a or
or
ln
b b h1
s1 ¼ a ln
1 ¼ s1 a
b b h1
ðiiÞ
226
4
Conduction with Heat Generation
From the given data mc 1:5 900 ¼ ¼ 1687:5 hA 20 0:04 qheater 1000 b¼ ¼ 1250 ¼ 20 0:04 hA a¼
Substitution gives
s1 ¼ 1687:5 ln
1250 1250 100
¼ 140:7 s:
Example 4.16 Prove that the temperature distribution in a plane wall shown in Fig. 4.10 with uniform heat generation qg is given by 1 t¼ þ b
sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 2 qg 2 tc þ x b bko
where the thermal conductivity is given by k = ko(1 + bt). Solution Heat balance for the elemental strip gives qx þ dx qx ¼ ðAdxÞqg or qx þ
d ðqx Þdx qx ¼ Adxqg dx
or d ðqx Þ ¼ Aqg dx or
d dt kA ¼ Aqg dx dx
tc t1 qx
t1 qx+dx
δ dx
Fig. 4.10 Example 4.16
x
4.4 Heat Transfer Through Piston Crown
227
or d dt k ¼ qg dx dx Integration gives k
dt þ qg x ¼ C 1 dx
ðiÞ
At x = 0, dt/dx = 0, hence C1 = 0. Putting k = ko(1 + bt), we have ko ð1 þ btÞ
dt þ qg x ¼ 0 dx
or dt dt qg þ bt þ x ¼ 0 dx dx ko Further integration gives tþ
bt2 qg x2 þ ¼ C2 2 ko 2
At x = 0, t = tc, hence C2 ¼ tc þ
btc2 2
Substitution of the value of C2 in Eq. (ii) gives tþ
bt2 qg x2 bt2 þ tc c ¼ 0 2 ko 2 2
Solution of this quadratic equation is
t¼
b2
þ
r ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ﬃ 2 b
2
4
qg 2 bko x
2 tbc tc2
2 sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 1 2 qg 2 ¼ þ tc þ x ; b b bko
which is the desired result.
ðiiÞ
228
4
Conduction with Heat Generation
Example 4.17 Derive the expression of the distribution of temperature in a cylindrical object with uniform heat generation. The material thermal conductivity is given by ko(1 + bt). Solution The differential equation for a cylinder is d dt kr ¼ qg r dr dr Integration gives kr
dt r2 þ qg ¼ C1 dr 2
ðiÞ
At the centre (r = 0), dt/dr = 0, hence C1 = 0. Putting k = ko(1 + bt), we have ko ð1 þ btÞr
dt r2 þ qg ¼ 0 dr 2
or dt dt qg þ bt þ r¼0 dr dr 2ko Further integration gives tþ
bt2 qg r 2 þ ¼ C2 2 2ko 2
At r = 0, t = tc hence 1 C2 ¼ tc þ btc2 2 Substitution of the value of C2 in Eq. (ii) gives tþ
bt2 qg 2 1 þ r tc btc2 ¼ 0 2 2 4ko
Solution of this quadratic equation is
t¼
b2
þ
r ﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2 b
2
4
qg 2 2bko r
2 tbc tc2
2 ﬃ sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2 1 1 qg 2 ¼ þ tc þ r ; b b 2bko
which is the desired result.
ðiiÞ
4.4 Heat Transfer Through Piston Crown
229
Example 4.18 The left face of a plane wall of thickness L and thermal conductivity k is exposed to a microwave radiation which causes volumetric heat generation in the wall, refer Fig. 4.11. The heat generation varies as x qg ¼ qo 1 L where qo (W/m3) is a constant. The other side of the wall is insulated. If the exposed surface of the wall is at a constant temperature Ts, obtain an expression of temperature distribution. Solution The differential equation for the elemental strip in the wall is d2 t k 2 þ qg ¼ 0 dx Substituting the given equation of heat generation, we get d2 t x k 2 þ q0 1 ¼0 dx L Integrating once, dt q0 q0 x2 þ x ¼ C1 dx k kL 2 At x = L, dt/dx = 0 since there is no heat transfer at this face. This gives qo L2 qo L C1 ¼ L ¼ 2k k 2L Equation (i) takes the form dt qo qo x 2 qo L þ x ¼0 dx 2k k k 2L Integrating again, we get
qo x2 qo x3 qo L tþ x ¼ C2 2k k 2 k 6L
Insulated side
Radiation
L x=0
Fig. 4.11 Example 4.18
ðiÞ
230
4
Conduction with Heat Generation
Applying the second boundary condition of t = ts at x = 0, we get the value of constant C2 = ts. Substitution gives the expression of the temperature distribution as t ¼ ts þ
qo 3 x 3x2 L þ 3L2 x 6kL
At x = L, the temperature is maximum. Its value is obtained by putting x = L in the above equation as qo L2 tmax ¼ ts þ 6k Example 4.19 The outer surfaces of the composite wall shown in Fig. 4.12 are exposed to a fluid at 20°C. The convective heat transfer coefﬁcients are 500 W/(m2 K) and 1000 W/(m2 K) on the left and right faces, respectively. The middle wall B experiences uniform volumetric heat generation rate qg, while there is no heat generation in walls A and C. The temperature at the interfaces are T1 = 220°C and T2 = 200°C. Assuming onedimensional heat flow and negligible contact resistance at the two interfaces, determine (i) The heat generation rate, qg (W/m3), (ii) The thermal conductivity kb, and (iii) If the coolant flow on the left side of the wall fails causing practically no loss of heat from this face (h1 = 0), determine the temperatures T1, T2 and Tmax of both wall sides. Solution (i) Heat generation rate, qg Heat loss from the left face (A = 1 m2), qL ¼
t1 t1 220 20 ¼ 60 ¼ 5 104 W: La 1 1 100030 þ 500 ka þ h1
Similarly, heat loss from the right face, qR ¼
qL
h2 = 1000
T2
T1
h1 = 500
Fig. 4.12 Example 4.19
t2 t1 200 20 ¼ 30 ¼ 9 104 W: Lc 1 1 þ þ 100030 1000 h2 kc
qR
qg A
B
La
Lb
C
Lc
La = 60 mm, Lb = 50 mm, Lc = 30 mm, ka = 30 W/(m K), kb = ?, kc = 30 W/(m K), t1 = 220oC, t2 = 200oC, t∞ = 20oC.
4.4 Heat Transfer Through Piston Crown
231
Heat generated in the walls equals the sum of qL and qR. Hence, the heat generation rate is qg ¼
qL þ qR 5 104 þ 9 104 ¼ 2:8 106 W/m3 : ¼ ALb 1 0:05
(ii) For the middle wall with uniform heat generation rate and surface temperatures T1 and T2, the heat flow qR from Eq. (4.10) by putting x = Lb is t1 t2 qg Lb qR ¼ k b þ Lb 2 or 220 200 2:8 106 0:05 þ 9 10 ¼ kb 0:05 2 4
This gives kb = 50 W/(m K). (iii) Temperature distribution through the middle wall is given by Eq. (4.4): qg t2 t1 ðLb xÞx þ x þ t1 2kb Lb 2:8 106 200 220 ð0:05 xÞx þ ¼ x þ 220 0:05 2 50
t¼
or t ¼ 1000x 2:8 104 x2 þ 220 Position of Tmax from Eq. (4.7) is x B1 ¼ Lb 2B where B¼
qg L2b 2:8 106 ð0:05Þ2 ¼ ¼ 3:5 2kb ðt1 t2 Þ 2 50 ð220 200Þ
Substitution gives x¼
3:5 1 0:05 ¼ 0:0179 m from left side of the middle wall; 2 3:5
and tmax ¼ 1000 0:0179 2:8 104 0:01792 þ 220 ¼ 228:9 C: Alternatively, Eq. (4.8) may be used.
232
4
Conduction with Heat Generation
The temperatures of sides of the wall can be found from the convection heat transfer equations as follows. qL ¼ h1 A t10 t1 t10 ¼ Similarly, t20 ¼
qL 5 104 þ t1 ¼ þ 20 ¼ 120 C h1 A 500 1
qR 9 104 þ t1 ¼ þ 20 ¼ 110 C: h2 A 1000 1
The temperature distribution through the composite wall is shown in Fig. 4.13a. (iv) When h1 = 0,
t10 ¼ t1
and
qR ¼ ð9 þ 5Þ 104 W:
Therefore, t20 ¼
qR 14 104 þ t1 ¼ þ 20 ¼ 160 C: h2 A 1000 1
From conduction heat transfer equation, qR ¼ kc
t2 t20 Lc
or t2 ¼
qR Lc 14 104 0:03 þ 160 ¼ 300 C: þ t20 ¼ 30 kc
17.9 mm
o
370 C
A
B
a
Fig. 4.13 Example 4.19
o
200 C o
qg
o
120 C
o
300 C
o
•220 C
160 C
o
110 C C
A
B
b
C
4.4 Heat Transfer Through Piston Crown
233
The maximum temperature (at the left interface), tmax ¼ t2 þ
qg 2 2:8 106 ð0:05Þ2 ¼ 370 C: Lb ¼ 300 þ 2kb 2 50
The temperature distribution is shown in Fig. 4.13b. Example 4.20 Heat is uniformly generated in the wall of a very long hollow cylindrical shell of thickness (R2 − R1). The heat is removed through the inner surface. The thermal conductivity of the wall material varies with temperature as k = ko(1 + bt). Show that the temperature distribution in the wall is given by vﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ " #ﬃ u 2 2 2 1 u 1 q R r r g 2 t ¼ þ t t2 þ 2 ln 1 b b R2 2ko b R2 where t2 is the temperature of the outer surface of the shell. Solution The differential equation for onedimensional heat flow in cylindrical coordinates is d dt kr þ qg r ¼ 0 dr dr Integration of the above equation gives kr
dt r2 þ qg ¼ C1 dr 2
ðiÞ
At r = R2, dt/dr = 0 since no heat is removed from the outer surface. This boundary condition gives C1 ¼
qg R22 2
Equation (i) takes the form kr
dt r 2 qg R22 þ qg ¼0 dr 2 2
or ko ð1 þ btÞ
dt r qg R22 þ qg ¼0 dr 2 2r
or dt dt qg qg R22 1 þ bt þ ¼0 r dr dr 2ko 2ko r
234
4
Conduction with Heat Generation
Integration of the above equation gives tþ
bt2 qg r 2 qg R22 þ : ln r ¼ C2 2 2ko 2 2ko
ðiiÞ
The second boundary condition is t = t2 at r = R2. This gives bT22 qg R22 qg R22 þ : ln R2 2 2ko 2 2ko
C2 ¼ t2 þ
Substitution of the value of C2 in Eq. (ii) gives bt2 qg R22 þ tþ 2 4ko or 2t qg R22 t2 þ þ b 2bko
"
r R2
"
r R2
2
2
# 1
qg R22 r bt2 ln t2 2 ¼ 0 R2 2ko 2
ðiiiÞ
#
qg R22 r 2t2 t22 ¼ 0 1 ln R2 bko b
This is a quadratic equation whose solution is vﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ " #ﬃ u 2 2 2 1 u 1 q R r r g 2 t ¼ þ t t2 þ 2 ln 1 b b R2 2ko b R2
ðivÞ
If thermal conductivity is constant b = 0 and putting ko by k in Eq. (iii), we get qg R22 tþ 4k
"
r R2
2
# 1
qg R22 r ln t2 ¼ 0 R2 2k
or qg R22 t ¼ t2 4k
"
r R2
2
r 2 ln R2
#
1
Example 4.21 How the abovederived expression will change if the shell is cooled from the outer surface instead of the inner one? Solution We can get the desired result by replacing R2 by R1 and t2 by t1. This gives vﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ " #ﬃ u 2 2 2 1 u 1 q R r r g 1 t ¼ þ t t1 þ 2 ln 1 b b R1 2ko b R1
4.4 Heat Transfer Through Piston Crown
235
and qg R21 t ¼ t1 4k
"
r R1
2
r 2 ln R1
#
1
when k is constant. The expression can be derived by following the procedure of the above example. Example 4.22 Inner and outer surfaces of a cylindrical shell are maintained at uniform temperatures T1 and T2, respectively. If there is a uniform rate of heat generation qg within the shell, obtain the following expressions for the steady state, onedimensional radial distribution of temperature, heat flux and heat rate. ) " " 2 # ( 2 # qg R22 r qg R22 R1 ln r ln R2 TðrÞ ¼ T2 þ : 1 1 þ ðT2 T1 Þ R2 4k 4k R2 ln R1 ln R2 Heat rate, 2pLk qg R22 R21 1 2 þ ðT2 T1 Þ qðrÞ ¼ pLqg r lnðR2 =R1 Þ 4k R2 2
Heat flux, qg r k qg R22 R21 1 q ðrÞ ¼ 1 2 þ ð T2 T 1 Þ 2 r 4k lnðR2 =R1 Þ R2 00
Solution (i) Temperature distribution From Eq. (vi), Sect. 4.2, tþ
qg 2 :r þ C1 ln r þ C2 ¼ 0 4k
The boundary conditions are t ¼ T1 at r ¼ R1 and
t ¼ T2 at r ¼ R2 :
Applying the boundary conditions, T1 þ
qg 2 R þ C1 ln R1 þ C2 ¼ 0 4k 1
T2 þ
qg 2 R þ C1 ln R2 þ C2 ¼ 0 4k 2
and
ðiÞ
236
4
Conduction with Heat Generation
Solving for C1 and C2, we get hq i 1 g R22 R21 þ ðT2 T1 Þ ln R1 ln R2 4k h i 2 qg R 2 qg 2 ln R2 2 C2 ¼ T2 R R 1 þ ð T2 T 1 Þ 4k 4k 2 ln R1 ln R2
C1 ¼
Substitution of values of C1 and C2 in Eq. (i) gives " 2 # qg R22 r : 1 TðrÞ ¼ T2 þ R2 4k ( ) " 2 # qg R22 R1 ln r ln R2 1 þ ðT2 T1 Þ 4k R2 ln R1 ln R2
ðiiÞ
(ii) Heat rate qðrÞ ¼ kð2prLÞ
dT dr
ðiiiÞ
Differentiating Eq. (ii), dT qg r ¼ dr 2k
(
) " 2 # qg R22 R1 1=r 1 þ ðT2 T1 Þ ln R1 ln R2 4k R2
Substitution in Eq. (iii) gives 2pLk qg R22 R21 1 2 þ ðT2 T1 Þ qðrÞ ¼ pLqg r lnðR2 =R1 Þ 4k R2 2
(iii) Heat flux qðrÞ qðrÞ ¼ AðrÞ 2prL qg r k qg R22 R2 1 ¼ 1 12 þ ðT2 T1 Þ 2 r 4k lnðR2 =R1 Þ R2
q00 ðrÞ ¼
Example 4.23 Show that the heat generation rate in the cylinder of Fig. 4.14 for a linear radial temperature distribution varies inversely with radius. Solution For steadystate radial heat conduction with heat generation for a cylinder, Eq. (2.16) gives k
1 @t þ qg ¼ 0: r @r
ðiÞ
4.4 Heat Transfer Through Piston Crown
237
t2 t1
r2
r1
t1 t2 r1
r2
Fig. 4.14 Hollow cylinder with heat generation
@t @r
@t For linear temperature distribution with radius, @r = constant, say C. Integration of relation t2 t1 ¼ C gives C ¼ r2 r1 . Substitution in Eq. (i) gives
qg ¼
t1 t2 k 1 / ; r2 r1 r r
i.e. the heat generation rate qg varies inversely with radius r. Example 4.24 Heat flows axially through a rod of variable crosssectional area given by Ax(x) = Aoebx where Ao and b are constants. Considering volumetric heat generation rate qg = qoe−bx, obtain the expression for qx when the left face and lateral surface are insulated. Thermal conductivity of the rod material is constant. Solution For the control volume, refer Fig. 4.15, qx þ ðAx dxÞqg ¼ qx þ dx Substituting qx þ dx ¼ qx þ
d dx ðqx Þdx,
we have, from Eq. (i), d Ax dxqg ¼ ðqx Þdx dx
or d ðqx Þ ¼ Ax qg dx ¼ A0 ebx q0 ebx ¼ A0 q 0
ðiÞ
238
4
Conduction with Heat Generation
dx Ax
qx
qx+dx x
Fig. 4.15 Example 4.24
or qx ¼ A0 q0 x þ C1
ðiiÞ
Left face is insulated, i.e. at x = 0, qx = 0. This condition gives C1 = 0. Hence, Eq. (ii) reduces to qx ¼ A0 q0 x:
Example 4.25 Heat generation in a cylindrical container of radioactive waste varies as "
2 # r qo 1 ro If the heat is rejected from the surface by convection to fluid at temperature t∞, determine the surface temperature of the container. Solution The rate of heat generation in the cylinder is Zro Qg ¼ qg ð2prLdrÞ 0
Zro ¼ 0
"
2 # r qo 1 ð2prLdrÞ ro Zro "
2 # r 1 ¼ 2pLqo rdr ro 0 2 r r r 4 o pLqo ro2 : ¼ ¼ 2pLqo 2 ro2 0 2 The rate of heat rejection by convection is Qc ¼ hð2pro LÞðts t1 Þ
ðiÞ
ðiiÞ
4.4 Heat Transfer Through Piston Crown
239
In equilibrium, Qg = Qc, which gives pLqo ro2 ¼ hð2pro LÞðts t1 Þ 2 or ts ¼ t1 þ
qo r o : 4h
Example 4.26 For the heat generation rate given in Example 4.25, determine the equation of radial temperature distribution. Solution From the general heat conduction equation in cylindrical coordinates, we have for onedimensional heat conduction 1d dt qg r ¼ r dr dr k
" 2 # qo r 1 ¼ ro k
or d dt qo r3 r r 2 ¼ dr dr k ro Integration of the above equation gives
dt qo r 2 r4 r ¼ þ C1 dr k 2 4ro2
At r = 0, dt/dr = 0, which gives C1 = 0. Hence, dt qo r 2 r4 r ¼ dr k 2 4ro2 or
dt qo r r3 ¼ 2 dr k 2 4ro Further integration gives t¼
qo r 2 r4 þ C2 k 4 16ro2
ðiÞ
ðiiÞ
240
4
Conduction with Heat Generation
Hence, the surface temperature is qo ro2 ro4 tðro Þ ¼ þ C2 k 4 16ro2 or tðro Þ ¼
3qo ro2 þ C2 16k
ðiiiÞ
At the surface of the cylinder, heat reaching by conduction equals the heat rejected by convection hence k
dt ¼ h½tðro Þ t1 dr
Substitution from Eqs. (i) to (iii) in the above equation gives qo r r3 3qo ro2 þ C2 t1 ¼h k k 2 4ro2 r¼ro 16k or qo ro 3qo ro2 ¼ þ C2 t1 4h 16k C2 ¼
qo ro 3qo ro2 þ þ t1 : 4h 16k
Substitution of the value of C2 in Eq. (ii) gives qo r 2 r4 qo ro 3qo ro2 þ þ t1 t¼ þ k 4 16ro2 4h 16k or " # qo ro qo ro2 3 1 r 2 1 r 4 þ t¼ þ þ t1 ; 16 ro 4h k 16 4 ro which is the desired radial temperature distribution equation. Check: At r = ro, the temperature from the above equation is tðro Þ ¼
qo r o þ t1 ; 4h
which is the same as determined in the previous example.
4.4 Heat Transfer Through Piston Crown
241
Example 4.27 The composite wall shown in Fig. 4.16a consists of two layers (k1 = 50 W/ (m K), d1 = 0.05 m and k2 = 210 W/(m K), d2 = 0.025 m). There is a uniform generation of heat at the rate of 1 106 W/m3 in layer 1. The contact resistance between the layers is estimated to be 0.001 m2 K/W. Left face of layer 1 is insulated. The open face of layer 2 is subjected to convection environment. Determine the temperature distribution through the wall. Solution Starting from the fundamental equation of conduction heat transfer through a plane wall, we have d2 t qg ¼0 þ dx2 k Integration gives dt qg þ x C1 ¼ 0 dx k
ðiÞ
At x = 0, dt/dx = 0 as the wall is insulated. This give C1 = 0. Equation (i) reduces to dt qg þ x¼0 dx k Further integration gives t¼
qg x2 þ C2 2k
t2 ¼
qg d21 þ C2 2k
ðiiÞ
At x = d1, t = t2 hence
t1
ho = 500 2 W/(m K)
t2 q t3
•
δ1 x
t4 o
• to = 30 C
δ2 a δ2/k2
Rc t2 q
t4
t3 b
Fig. 4.16 Example 4.27
1/ho to
242
4
Conduction with Heat Generation
or C2 ¼
qg d21 þ t2 2k
Substitution of the value of C2 in Eq. (ii) gives t¼
qg 2 ðd x2 Þ þ t2 2k 1
ðiiiÞ
Equation (iii) gives temperature distribution in layer 1 with heat generation, which is parabolic. Temperature distribution through layer 2 is a straight line because there is no heat generation in layer 2 and the thermal conductivity is constant. The temperature t1–t4 can be determined as under. Heat flow rate through layer 2 equals the heat generated in the wall, i.e. qgd1 for unit face area of the wall. Heat flow rate equation gives t2 to q ¼ qg d1 ¼ ; ðivÞ Rt where refer Fig. 4.16b, Rt ¼ Rc þ
d2 1 0:025 1 þ ¼ 0:00312: þ ¼ 0:001 þ ho 210 500 k2
Equation (iv) gives t2 ¼ qg d1 Rt þ to ¼ 1 106 0:05 0:00312 þ 30 ¼ 186 C: From Eq. (iii), at x = 0 qg 2 d þ t2 2k 1 1 106 ð0:05Þ2 þ 186 ¼ 211 C: ¼ 2 50
t1 ¼
We can write q¼
t2 t3 Rc
which gives t3 ¼ t2 qRc ¼ 186 ð1 106 0:05Þ 0:001 ¼ 136 C: Similarly, qd2 k2 ð1 106 0:05Þ 0:025 ¼ 130:05 C: ¼ 136 210
t4 ¼ t3
4.4 Heat Transfer Through Piston Crown
243
and t0 ¼ t4
q ho
¼ 130:05
ð1 106 0:05Þ ¼ 30:05 C; 500
which is approximately equal to the given value. Example 4.28 Heat generation in a spherical container of radioactive waste varies as "
2 # r qo 1 : ro If the heat is rejected from the surface by convection to fluid at temperature t∞, determine the equation of radial temperature distribution. Solution From the general heat conduction equation in spherical coordinates, we have for onedimensional heat conduction 1 @ qg 2 @t r ¼ r 2 @r @r k
" 2 # qo r 1 ¼ ro k
or @ qo 2 r 4 2 @t r r 2 ¼ @r @r k ro Integration of the above equation gives r2
dt qo r 3 r5 ¼ 2 þ C1 dr k 3 5ro
At r = 0, dt/dr = 0, which gives C1 = 0. Hence, dt qo r 3 r5 ¼ r dr k 3 5ro2 2
or dt qo r r3 ¼ dr k 3 5ro2
ðiÞ
244
4
Conduction with Heat Generation
qo r 2 r4 t¼ þ C2 k 6 20ro2
ðiiÞ
Further integration gives
Hence, the surface temperature is qo ro2 ro4 tðro Þ ¼ þ C2 k 6 20ro2 or tðro Þ ¼
7qo ro2 þ C2 60k
ðiiiÞ
At the surface of the sphere, heat reaching by conduction equals the heat rejected by convection hence k
dt ¼ h½tðro Þ t1 dr
Substitution from Eqs. (i) and (iii) in the above equation gives qo r r3 7qo ro2 2 k þ C2 t1 ¼h k 3 5ro r¼ro 60k or 2qo ro 7qo ro2 ¼ þ C2 t1 15h 60k C2 ¼
2qo ro 7qo ro2 þ þ t1 : 15h 60k
Substitution of the value of C2 in Eq. (ii) gives qo r 2 r4 2qo ro 7qo ro2 þ þ t1 t¼ þ k 6 20ro2 15h 60k or " # 2qo ro qo ro2 7 1 r 2 1 r 4 þ t¼ þ þ t1 ; 20 ro 15h k 60 6 ro which is the desired radial temperature distribution equation.
4.4 Heat Transfer Through Piston Crown
245
Example 4.29 A cylindrical reactor fuel element with uniform heat generation rate qg has steel cladding as shown in Fig. 4.17. Heat is rejected from the outer surface by convection to a coolant at temperature t∞. Determine the expressions for temperature distributions in fuel element and cladding. Solution For the cylindrical fuel element, 1d dtf qg r ¼ ; r dr dr kf or d dtf qg r r ¼ dr kf dr Integration gives r
dtf qg r 2 ¼ þ C1 dr 2kf
At r = 0, dtf/dr = 0, therefore C1 = 0. Hence, the equation is r
dtf qg r 2 ¼ dr 2kf
or dtf qg r ¼ 2kf dr Further integration gives tf ¼
qg r 2 þ C2 4kf
Fuel element
ðiÞ
Steel cladding
h, t∞
r1
Fig. 4.17 Example 4.29
r2
246
4
Conduction with Heat Generation
For the cylindrical cladding, 1d dtc r ¼ 0; r dr dr or d dtc r ¼0 dr dr Integration gives r
dtc ¼ C3 dr
or dtc C3 ¼ dr r Further integration gives tc ¼ C3 ln r þ C4
ðiiÞ
The constants C2, C3 and C4 can be found from the following boundary conditions. At radius r1, tf ðr1 Þ ¼ tc ðr1 Þ
ðiiiÞ
Conduction heat flow at r1 gives kf
dtf dtc ¼ kc dr r¼r1 dr r¼r1
ðivÞ
At the cladding surface, heat is rejected by convection, hence dtc kc ¼ h½tc ðr2 Þ t1 dr r¼r2
ðvÞ
From Eq. (iii),
qg r12 þ C2 ¼ C3 ln r1 þ C4 4kf
ðviÞ
4.4 Heat Transfer Through Piston Crown
From Eq. (iv), kf
247
dtf dtc ¼ kc dr r¼r1 dr r¼r1
or
kf
qg r 1 2kf
¼ kc
C3 r1
or C3 ¼
qg r12 2kc
From Eq. (v), kc
C3 ¼ h½C3 ln r2 þ C4 t1 r2
or qg r12 qg r 2 ¼ 1 ln r2 þ C4 t1 2r2 h 2kc or C4 ¼
qg r12 qg r12 þ ln r2 þ t1 2r2 h 2kc
Substituting values of C3 and C4 in Eq. (vi), we obtain C2 ¼
qg r12 qg r12 qg r12 qg r12 þ ln r1 þ þ ln r2 þ t1 2kc 4kf 2r2 h 2kc
or qg r12 r2 qg r12 qg r12 þ t1 C2 ¼ ln þ þ 2kc r1 4kf 2r2 h Substituting values of constants C2–C4 in Eqs. (i) and (ii), we obtain the expressions for temperature distributions in fuel element and cladding as tf ¼
qg r 2 qg r12 r2 qg r12 qg r12 þ t1 þ ln þ þ 4kf 2kc r1 4kf 2r2 h
or tf ¼
qg 2 qg r12 r2 qg r12 þ t1 ðr1 r 2 Þ þ ln þ 4kf 2kc r1 2r2 h
ð1Þ
248
4
Conduction with Heat Generation
and tc ¼
qg r12 qg r12 qg r12 þ ln r þ ln r2 þ t1 2kc 2r2 h 2kc
or tc ¼
4.5
qg r12 r2 qg r12 þ t1 ln þ 2kc r 2r2 h
ð2Þ
Summary
This chapter is devoted to the heat conduction processes when there is heat generation in the solid itself. Firstly the treatment has been presented for plane wall, cylindrical and spherical solids with a uniform rate of heat generation per unit volume with constant thermal conductivity of the solid material, and for the steadystate onedimensional heat conduction. The following basic differential equations for these cases have been developed in Sects. 4.1–4.3. Plane wall with uniform heat generation, d2 t qg ¼0 þ dx2 k
ð4:1Þ
Cylinder with uniform heat generation, d dt qg r þ r¼0 dr dr k
ð4:16Þ
Solid sphere with uniform heat generation, d qg 2 dt r þ r2 ¼ 0 dr dr k
ð4:21Þ
The students are advised to start from these basic differential equations for solution to problems. Integration of the above equations and the determination of the constants of the integration from the given boundary conditions will provide the applicable equation of temperature distribution for different cases of interest, which can be utilized for calculation of the rate of heat transfer using the Fourier’s equation or to locate the position and magnitude of maximum temperature in the solid. For cases with variable thermal conductivity, the students may start from equations obtained from the general heat conduction equations in Chap. 1; for example, from equation dt dt d d dx k dx ¼ qg for a plane wall (refer Example 4.16), or from equation dr kr dr þ qg r ¼ 0 for a cylindrical element (refer Examples 4.17 and 4.20). Alternatively, one may start from fundamentals. For the problems involving the heat generation, which is variable as in Example 4.18, it is advised to start from fundamentals only. An approximate treatment of the heat transfer through piston crown has been presented in Sect. 4.4 to determine temperature distribution and maximum temperature in the crown.
4.5 Summary
249
A number of solved examples have been included to illustrate the application of the abovediscussed approaches for a variety of problems.
Review Questions 4:1 Show that for a plane wall of thickness 2 l with a uniformly distributed heat generation qg per unit volume, the temperature to at the midplane is given by to ¼
qg l2 þ tw 2k
where tw is the temperature on either side of the wall and k is the thermal conductivity of the wall material. 4:2 Develop an expression of temperature distribution in a sphere of radius Ro made of a homogeneous material in which energy is released at a uniform rate per unit volume. The surface temperature of the sphere is To and the thermal conductivity of the sphere material may be assumed to be constant. Also, give the expression of maximum temperature in the sphere. 4:3 Inner and outer surfaces of a spherical shell are maintained at uniform temperatures t1 and t2, respectively. If there is a uniform rate of heat generation qg within the shell, obtain the following expressions for the steadystate, onedimensional radial distribution of temperature, heat flux and heat rate. ) " " 2 # ( 2 # qg R22 r qg R22 R1 1=r 1=R2 tðrÞ ¼ t2 þ : 1 1 þ ðt 2 t 1 Þ R2 6k 6k R2 1=R1 1=R2 Heat rate, 4pqg r 3 qg R22 R21 1 qðrÞ ¼ 4pk 1 2 þ ðt2 t1 Þ 1=R1 1=R2 3 6k R2 Heat flux, q00 ðrÞ ¼
qg r k qg R22 R2 1 2 1 12 þ ðt2 t1 Þ 3 r 1=R1 1=R2 6k R2
[Hint: Follow the procedure of Example 4.22]
250
4
Conduction with Heat Generation
Problems 4:1 Heat is generated at the interface between two slabs each 50 mm thick. One of the slabs is of steel [k = 35 W/(m K)] and other is of brass [k = 75 W/(m K)]. The temperature at the outside of the steel slab is 95°C, while at the outside of the brass is 47°C. Calculate the heat flow rate at the outer surfaces and the interface temperature when the rate of heat generation is 175 kW/m2 of the contact area. 1 2 qR ¼ k2 A ti t [Ans. qL ¼ k1 A ti t ¼ 700 ðti 95Þ; ¼ 1500 ðti 47Þ; d1 d2 qL þ qR ¼ qg ; ti ¼ 141:82 C; qL ¼ 32:77 kW/m2 ; qR ¼ 142:23 kW/m2 : 4:2 A long cylindrical rod of 100 mm radius experiences uniform heat generation rate of 2 104 W/m3. It carries a tightly ﬁtting sleeve having an outer radius of 150 mm and a thermal conductivity of 5 W/(m K). The surface of the sleeve is exposed to crossflow of air at 30°C. If the thermal conductivity of the rod material is 0.5 W/(m K) and the convective heat transfer coefﬁcient is 40 W/(m2 K), determine (i) Temperature at the interface of the rod and the sleeve, and (ii) The maximum temperature and its location. [Ans. Lq ¼
t1 t1 1 2pk2
ln
R2 R1
þ 2pR1
, where q/L equals the heat generated per unit length of the rod, 2h
i.e. q/L = qg pR21 1. This gives interface temperature, t1 = 54.77°C. q tmax ¼ 4kg1 R21 þ t1 ¼ 154:77 C:] 4:3 A large concrete platform 1.0 m high is to be constructed by pouring concrete on a brick base. The top surface of the concrete is exposed to the atmosphere at 300 K and the unit surface conductance is estimated to be 10 W/(m2 K). The concrete sets by an exothermal chemical reaction, which produces heat at the rate of 60 W/m3. Assuming onedimensional steadystate conduction with a uniform rate of heat generation per unit volume, ﬁnd the maximum temperature within the concrete. Take thermal conductivity of the concrete as 1.0 W/(m K) and neglect the heat loss from the base. (Hint: The bottom surface is insulated, hence at the bottom dt/dx is zero and the temperature is the q d
q d2
g maximum. Use the following equation: Tmax ¼ hg þ 2k þ T1 ) [Ans. tmax = 63°C] 4:4 The outer freshly plastered wall surface is held at 35°C during curing, when the rate of generation of heat due to the setting is 5 104 W/m3. The plaster is 15 mm thick. Determine the temperature at inner surface of the plaster. The inner surface can be assumed to be insulated. The thermal conductivity of the plaster material is 1 W/(m K). [Ans. The temperature at the wall is maximum, hence the boundary conditions are: x = 0, dt/dx = 0; at x = d = 0.015 m, t = 35°C, refer Fig. 4.18. Integrate Eq. (4.1) and applying q boundary condition to obtain temperature distribution equation t ¼ 2kg ðd2 x2 Þ þ t1 ; Putting x = 0 in the temperature distribution equation, we get tmax = 40.6°C.] 4:5 A proposed nuclear fuel element 60 mm in diameter is to be clad in a 5 mm thick aluminium sheathing. The fuel element generates heat at a rate of 4 105 kW/m3. The outside surface of the aluminium cover is likely to be at 80°C because of the contact with
Problems
251
Insulated side
o
35 C Plaster 15 mm thick
Wall x=0
Fig. 4.18 Problem 4.4
Insulated side
Radiation
L x=0
Fig. 4.19 Problem 4.7
a cooling fluid. Estimate the maximum temperature of the aluminium cladding if its thermal conductivity can be expressed as k = 210(1  t/1000), where t is in °C. ðti to Þ [Ans. qg ðpR2 LÞ ¼ 2pLko ½1 bðto þ ti Þ=2 lnðR , where b = 1/1000; ti = 237°C.] o =Ri Þ
4:6 A current of 80 A is passed through a 2 mm diameter 500 mm long stainless steel wire [k = 20 W/(m K) and electrical resistivity q = 76 lΩcm]. If the wire is dipped in a liquid bath at 80°C and h = 1500 W/(m2 K), calculate the maximum temperature of the wire. [Ans. Resistance of wire R = qL/A = 0.121Ω, heat generated, H = I2R = 774.4 W, heat generation rate, qg = H/Volume of wire = 493 106 W/m3. Convective heat transfer equation hA(ts − t∞) = H gives ts = 244.33°C. Maximum temperature at the centre, q R2
tmax ¼ g4k o þ ts ¼ 250:5 C:] 4:7 The plane wall shown in Fig. 4.19 is subjected to a radiation which causes volumetric heat generation in the wall. The heat generation varies as qg ¼ q0
x L
Derive the equation of distribution and determine the maximum temperature in the wall if the temperature at the wall surface (at x = L) is ts. qo qo 2 ðL3 x3 Þ; tmax ¼ ðtÞx¼0 ¼ ts þ 6k L :] [Ans. t ¼ ts þ 6kL
5
SteadyState TwoDimensional Heat Conduction
5.1
Introduction
In the preceding chapters, the cases of onedimensional steadystate conduction heat flow were analysed. We consider now twodimensional steadystate conduction heat flow through solids without heat sources. The Laplace equation that governs the temperature distribution for two dimensional heat conduction system is @2t @2t þ 2¼0 2 @x @y
ð5:1Þ
Equation (5.1) is based on the assumption of constant and equal thermal conductivities in both x and y space coordinates. Solution of this equation gives the temperature in the body as a function of coordinates x and y. Knowing the temperature distribution in the body, the heat flow in the x and ydirections at a point can be determined from qx ¼ kAx
@t @x
ð5:2aÞ
qy ¼ kAy
@t @y
ð5:2bÞ
The rate of heat flow at the point is the resultant of the components qx and qy, i.e. q ¼ iqx þ jqy
ð5:3Þ
It follows that in order to determine the heat flow, the temperature ﬁeld must be known. Thus, the problem reduces to the solution of Eq. (5.1). The common techniques available for the solution of Eq. (5.1) are (i) (ii) (iii) (iv)
Mathematical analysis (analytical solution), Graphical analysis, Method of analogy, Numerical solutions using either a ﬁnitedifference or ﬁniteelement method.
© Springer Nature Singapore Pte Ltd. 2020 R. Karwa, Heat and Mass Transfer, https://doi.org/10.1007/9789811539886_5
253
254
5.2
5 SteadyState TwoDimensional Heat Conduction
Analytical Solution of TwoDimensional Heat Conduction Problems
Consider a rectangular section bar, as shown in Fig. 5.1, which is very long in zdirection. Three lateral sides of the bar are maintained at a constant temperature To. For the fourth side we consider different conditions, case (i)–(iii), as outlined below. Case (i) The fourth side (y = H) has a sinusoidal temperature distribution, T = Tm sin(px/W), imposed on it. Using h = T – To, the Laplace equation, Eq. (5.1), is transformed to @2h @2h þ 2¼0 @x2 @y
ð5:4Þ
The boundary conditions are ðiÞ h ¼ 0 ðiiiÞ h ¼ 0
at y ¼ 0 ðiiÞh ¼ 0 at x ¼ 0; at x ¼ W; ðivÞ h ¼ hm sinðpx=WÞ at y ¼ H
ð5:5Þ
where hm = Tm – To is amplitude of the sine function. T = Tm sin(px/W) is transformed to h = hm sin(px/W). The solution of Eq. (5.4), using the separation of variable technique (based on the assumption that the solution to the differential equation takes a product form), is hðx; yÞ ¼ XY
ð5:6Þ
2 2 1 d X 1 dY ¼ 2 X dx Y dy2
ð5:7Þ
where X = X (x) and Y = Y (y). Substitution in Eq. (5.4) gives
Each side of Eq. (5.7) is independent of the other, since x and y are independent variables. Hence, the left side of Eq. (5.7) can equal the right side only if both the sides have a constant value, greater than zero, say k2. That is,
y T H To
To
x To
W
Fig. 5.1 A rectangular section bar with given thermal boundary conditions
5.2 Analytical Solution of TwoDimensional Heat Conduction Problems
2 2 1 d X 1 dY ¼ ¼ k2 2 X dx Y dy2
255
ð5:8Þ
Thus, we get two ordinary differential equations as d2 X þ k2 X ¼ 0 dx2
ð5:9Þ
d2 Y k2 Y ¼ 0 dy2
ð5:10Þ
and
The value of the constant k2 is to be determined from the given boundary conditions. The general solution of Eq. (5.9) is X ¼ C1 cos kx þ C2 sin kx and that of Eq. (5.10) is Y ¼ C3 expðkyÞ þ C4 expðkyÞ Substitution in Eq. (5.6) gives hðx; yÞ ¼ XY ¼ ðC1 cos kx þ C2 sin kxÞ½C3 expðkyÞ þ C4 expðkyÞ
ð5:11Þ
Applying the boundary conditions, we have 0 ¼ ðC1 cos kx þ C2 sin kxÞðC3 þ C4 Þ
ðiÞ
0 ¼ C1 ½C3 expðkyÞ þ C4 expðkyÞ
ðiiÞ
0 ¼ ðC1 cos kW þ C2 sin kWÞ½C3 expðkyÞ þ C4 expðkyÞ
ðiiiÞ
hm sinðpx=WÞ ¼ ðC1 cos kx þ C2 sin kxÞ½C3 expðkHÞ þ C4 expðkHÞ
ðivÞ
Equations (i) and (ii) give C3 ¼ C4 and C1 ¼ 0: Substitution in Eq. (iii) gives 0 ¼ C4 C2 sinðkWÞ½expðkyÞ expðkyÞ Equation (5.12) requires that
ð5:12Þ
256
5 SteadyState TwoDimensional Heat Conduction
sinðkWÞ ¼ 0 or k ¼ np=W
ð5:13Þ
where n is a positive integer. By substituting values of constants C1 to C4 and k in Eq. (5.11), we get hðx; yÞ ¼ T To ¼
1 X
Cn sin
n¼1
npx npy sinh W W
ð5:14Þ
where the term sinh(nky/W) replaces the exponential term and constants C2 and C4 have been combined. The solution of the differential equation is a sum of the solutions for each value of n, up to inﬁnity. Using the fourth boundary condition of h = hm sin(px/W) at y = H, we get hm sin
1 px X npx npH Cn sin ¼ sinh W W W n¼1
ð5:15Þ
This holds only for Cn = 0 when n > 1. For n = 1, we get pH hm sin ¼ C1 sin sinh ; W W W px
px
which gives C1 ¼
hm : sinh pH W
The ﬁnal solution of the differential equation is " hðx; yÞ ¼ T To ¼ hm
sin
# sinh py W sinh pH W
px W
ð5:16Þ
Case (ii) If the temperature along y = H side is given by an arbitrary function f(x), the boundary condition (iv) in Eq. (5.5) changes. The remaining conditions (i) to (iii) remain valid and the solution is as determined above, refer Eq. (5.14), i.e. hðx; yÞ ¼ T To ¼
1 X n¼1
The fourth boundary condition gives
Cn sin
npx npy sinh W W
ð5:14Þ
5.2 Analytical Solution of TwoDimensional Heat Conduction Problems
257
npx npH sinh W W
ð5:17Þ
f ð xÞ ¼
1 X
Cn sin
n¼1
The quantities Cn sinh(npH/W) must be the coefﬁcients of the Fourier sine series for f(x) in the interval (0 < x < L), i.e. ZW npx npH 2 Cn sinh f ð xÞ sin ¼ dx W W W
ð5:18Þ
0
and thus, 2 3 ZW 1 npx npx npy 2X 1 4 npH f ð xÞ sin hðx; yÞ ¼ T To ¼ dx5 sin sinh ð5:19Þ W n¼1 sinh W W W W 0
The temperature ﬁeld is shown in Fig. 5.2. The heat flow lines (dashed lines) are drawn perpendicular to the isotherms (ﬁrm lines). Case (iii) The three sides of the bar are held at temperatures To, while the fourth side (y = H) is at a constant temperature Tb. The fourth boundary condition (h = Tb – To = hb at y = H) gives hb ¼
1 X n¼1
Cn sin
npx W
sinh
npH W
ð5:20Þ
The value of Cn is determined by expanding the constant temperature difference hb in a Fourier series over the interval (0 < x < W), which is hb ¼ hb
1 2X ð1Þn þ 1 þ 1 npx sin p n¼1 n W
Comparison of Eqs. (5.20) and (5.21) gives
1
f(x)
0
Fig. 5.2 Isotherms and heat flow lines in a rectangular plate
1
ð5:21Þ
258
5 SteadyState TwoDimensional Heat Conduction
Cn ¼ hb
2 ð1Þn þ 1 þ 1 p n sinh npH W
ð5:22Þ
The ﬁnal solution is 1 npx npy h T To 2X ð1Þn þ 1 þ 1 npH sin ¼ ¼ sinh hb Tb To p n¼1 n sinh W W W 1 4X 1 ð2j 1Þpx ð2j 1Þpy h i sin sinh ¼ p j¼1 ð2j 1Þ sinh ð2j1ÞpH W W
ð5:23Þ ð5:23bÞ
W
In Eq. (5.23b), only the oddorder terms appear. The evenorder terms vanish because of the bracketed term [(−1)n+1 + 1].
5.3
Conduction Through a Flat Semiinfinite Homogeneous Plate
Consider the flat plate shown in Fig. 5.3, whose length in ydirection is inﬁnite. It is assumed that the plate is relatively thin in zdirection and xoy surfaces are insulated so that there is no temperature gradient in zdirection and the temperature ﬁeld is two dimensional. The general solution, Eq. (5.11), is hðx; yÞ ¼ XY ¼ ðC1 cos kx þ C2 sin kxÞ½C3 expðkyÞ þ C4 expðkyÞ
ð5:11Þ
The boundary conditions are ði) h ¼ 0 at x ¼ 0; ðii) h ¼ 0 at y ! 1 ; ðiii) h ¼ 0 at x ¼ W; ðiv) h ¼ FðxÞ at y ¼ 0 where h ¼ T To and f ðxÞ To ¼ FðxÞ:
y
∞
To
To
To
o T = f(x)
W
x
Fig. 5.3 A semiinﬁnite flat plate with given thermal boundary conditions
5.3 Conduction Through a Flat Semiinfinite Homogeneous Plate
259
(i) X ¼ /ðxÞ ¼ ðC1 cos kx þ C2 sin kxÞ will give /(x) = 0 at x = 0 when C1 = 0 (ii) Y ¼ wðyÞ ¼ C3 expðkyÞ þ C4 expðkyÞ satisﬁes the condition h = 0 as y ! ∞. Function w(y) = 0 at y ! ∞ and this is possible when C4 = 0. Using C1 = 0 and C4 = 0, we obtain from Eq. (5.11) hðx; yÞ ¼ C2 sinðkxÞ½C3 expðkyÞ ¼ C sinðkxÞ½expðkyÞ
ð5:24Þ
(iii) To satisfy the third boundary condition of h = 0 at x = W, sinðkxÞ ¼ 0 or k ¼ np=W where n = 1, 2, 3, ……. The general solution is, therefore, a sum of solutions for the each value of n up to inﬁnity, i.e. hðx; yÞ ¼ T To ¼
1 X
Cn sin
npx W
n¼1
exp
npy W
ð5:25Þ
The integral constant Cn is determined from the fourth boundary condition, h = F(x) at y = 0. Thus F ð xÞ ¼
1 X
Cn sin
n¼1
npx
ð5:26Þ
W
The expression is sine expansion of function F(x) into Fourier series. Hence, 2 Cn ¼ W
ZW F ð xÞ sin 0
npx dx W
ð5:27Þ
and the ﬁnal expression of temperature distribution is hðx; yÞ ¼
1 npx npy 2X sin exp W n¼1 W W
ZW F ð xÞ sin 0
npx dx W
ð5:28Þ
For the special case of T = T1 = constant at y = 0, function f(x) = T1 and F(x) = T1 – To = h1. The integral of Eq. (5.27) is
260
5 SteadyState TwoDimensional Heat Conduction
ZW F ð xÞ sin 0
h npx npxiW 2W W ¼ dx ¼ h1 cos h1 W np W np 0
Substitution in Eq. (5.28) gives 1 npy npx h 4X 1 exp ¼ sin h1 p n¼1 n W W 1 4X 1 ð2j 1Þpy ð2j 1Þpx exp sin ¼ p j¼1 ð2j 1Þ W W
ð5:29aÞ ð5:29bÞ
In Eq. (5.29b), only oddorder terms appear. Example 5.1 Consider the bar shown in Fig. 5.1 with W = H. The boundary conditions are: h = 100°C for the y = H surface while for the remaining three sides of the bar, h = 0°C. Determine the temperature along the centreline of the bar. Solution Substituting H = W, x = W/2 and y = H/2 = W/2 in the righthand side of Eq. (5.23), we obtain 1 np np hc 2 X ð1Þn þ 1 þ 1 sin ¼ sinh 2 2 hb p n¼1 n sinhðnpÞ Retaining the ﬁrst three oddorder terms (evenorder terms are zero), we obtain p p hc 2 2 2 3p 3p sin sin ¼ sinh þ sinh 2 2 3 sinhð3pÞ 2 2 hb p sinhðpÞ 2 5p 5p sin þ sinh 5 sinhð5pÞ 2 2 ¼ ð2=pÞ½0:3985 0:00599 þ 0:000155 ¼ 0:24998 or
hc ¼ 24:998 C: *The series will converge to 0.25.
Example 5.2 If in the above example H/W = 2, determine the centreline temperature. Solution Substituting H = 2 W, x = W/2 and y = H/2 = W in the righthand side of Eq. (5.23), we obtain 1 np hc 2 X ð1Þn þ 1 þ 1 sin ¼ sinhðnpÞ 2 hb p n¼1 n sinhð2npÞ Retaining the ﬁrst two oddorder terms (even order terms are zero), we obtain
5.3 Conduction Through a Flat Semiinfinite Homogeneous Plate
261
p hc 2 2 2 3p sin sin ¼ sinhðpÞ þ sinhð3pÞ 2 3 sinhð6pÞ 2 hb p sinhð2pÞ ¼ 0:054919 0:000034 ¼ 0:054885 or
hc ¼ 5:49 C: *The series will converge to 0.0549.
Note: From the results of the above two examples, it can be seen that (i) hc/hb is a strong function of ratio H/W, and (ii) the convergence of the series improves as H/W increases. When H/W 2, the ﬁrst term practically equals the sum of the series. Example 5.3 The material of the long square crosssection bar, shown in Fig. 5.4a, has homogeneous composition. The top side of the bar is held at t1 = 200°C and the temperature of the remaining three sides is 100°C. Determine the temperature along the centreline of the bar. Solution The problem can be solved by the method of superposition explained below. Consider the problem on the left in Fig. 5.5, which has uniform temperature on all sides. Its centreline temperature will be 100°C. The centreline temperature of all the subproblems on the right will be the same by symmetry. The subproblems on the right can be superimposed to yield the problem on the left. Let hc is the temperature at the centreline of the subproblems on the right. Then superposition will give 4hc ¼ 100 C or
hc ¼ 25 C
t1 = 200oC θ1 = 100oC t2 = 100oC
θ2 = 0oC
θ4 = 0oC θ3 = 0oC
o
t3 = 100 C θ=T a
Fig. 5.4 Example 5.3
1000C b
262
5 SteadyState TwoDimensional Heat Conduction 100o 100o
100o =
100o 100o 0o
100o 0o
θc
0o + 100o
0o 0o + 0o
θc
θc
0o
0o
0o 0o + 0o
100o
θc
100o
0o
Fig. 5.5 Problem of Fig. 5.4 and its subproblems
which is the desired centreline temperature of the original problem and hence tc = hc + 100 = 125°C. Example 5.4 Using the method of superposition, determine the temperature at point (1, 1) in Fig. 5.6. Solution This problem has more than one nonhomogeneous boundary condition. It can be reduced into a set of simpler problems with the geometry of the original problem and each having a nonhomogeneous boundary condition. Then the solution of these simpler problems can be superimposed at the point (1, 1) to give the solution of the original problem as explained through Fig. 5.7. Refer Example (5.3), hc1 = 75°C and hc2 = 25°C. The superposition gives hc = 75°C + 25°C = 100°C.
θ1 = 300oC 2
.
(1, 1) θ2 = 0oC
θ4 = 100oC
θ3 = 0oC 2
Fig. 5.6 Example 5.4 300oC
300oC 0oC
θc
100oC = 0oC
0oC
Fig. 5.7 Problem of Fig. 5.6 and its subproblems
θc1 0o C
0oC 0oC + 0oC
θc2 0oC
100oC
5.4 Mean Value Theorem
5.4
263
Mean Value Theorem
The results of the previous two examples can be extended to a homogeneous bar having a regular nsided polygon, where side i is at temperature Ti (i = 1, 2, 3, …n). Its centreline temperature will be the arithmetic mean of the temperatures around its perimeter, i.e. Tc ¼ ð1=nÞðT1 þ T2 þ Tn Þ
ð5:30aÞ
When n tends to inﬁnity, the polygon becomes a circle and 1 Tc ¼ 2p
Z2p ð5:30bÞ
Tð/Þd/ 0
where T(/) is the equation of angular distribution of temperature around the circumference of the circle. This result is termed the mean value theorem. Example 5.5 Using the method of superposition, obtain the equation of temperature distribution for the long bar in Fig. 5.8. Solution As explained earlier, this problem has more than one nonhomogeneous boundary condition. It can be reduced into a set of simpler problems with the geometry of the original problem and each having nonhomogeneous boundary condition as shown in Fig. 5.9. Then the equations of the temperature distribution of these simpler problems can be superimposed to give the solution of the original problem as explained below.
y θ1 H θ2
x, y θ4 x
W
θ3
Fig. 5.8 A long bar with nonhomogeneous thermal boundary conditions
0o
θ1 0o
0o + θ 2 0o (i)
y
(W 0o
x)
0o
+ 0o
0o + 0o
0o
θ3
(ii)
(iii) Origin
Fig. 5.9 A set of subproblems for the problem in Fig. 5.8
0o H  y
Hy
(W
x)
θ4
x 0o (iv)
264
5 SteadyState TwoDimensional Heat Conduction
Part (i) Referring Fig. 5.1, the temperature distribution equation is given by Eq. (5.23b) as h
1 4X
hðx; yÞ ¼ h1 p
j¼1
sinh
ð2j1Þpy W
ð2j 1Þ sinh
h
i
ð2j1ÞpH W
i sin
ð2j 1Þpx W
ðiÞ
We can denote this function as t(x, y/h1, W, H). Then the functions for the parts (ii) to (iv) of Fig. 5.9 and the relevant temperature distribution equations can be written as below. Part (ii) The function of this case is t(y, W  x/h2, H, W), see the new position of the origin in the Fig. 5.9b for the measurement of distances instead of distances x and y of Case (i), and h
1 4X
hðx; yÞ ¼ h2 p
j¼1
sinh
ð2j1ÞpðWxÞ H
ð2j 1Þ sinh
h
i
ð2j1ÞpW H
i sin
ð2j 1Þpy H
ðiiÞ
Part (iii) The function of this case is t(W–x, H–y/h3, W, H) and 1 4X
hðx; yÞ ¼ h3 p
j¼1
h sinh
ð2j1ÞpðHyÞ W
ð2j 1Þ sinh
h
i
ð2j1ÞpH W
i sin
ð2j 1ÞpðW xÞ W
ðiiiÞ
Part (iv) The function of this case is t(H–y, x/h4, H, W) and h i Þpx 1 sinh ð2j1 H hðx; yÞ 4 X ð2j 1ÞpðH yÞ h i sin ¼ h4 p j¼1 ð2j 1Þ sinh ð2j1ÞpW H
ðivÞ
H
Superposition, i.e., the summation of Eqs. (i)–(iv), gives the desired result. Example 5.6 Does the result of Example 5.5 conﬁrm the mean value theorem for the square? Solution Substitution of W = H and x = y = H/2 in the result gives h i Þp 1 sinh ð2j1 2 H H 4X ð2j 1Þp sin ; ¼ ðh1 þ h2 þ h3 þ h4 Þ h 2 2 p J¼1 ð2j 1Þ sinhð2j 1Þp 2 p 3p sinh 2 sinh 5p 4 sinh 2 p 3p 5p 2 sin sin sin þ þ ¼ ðh1 þ h2 þ h3 þ h4 Þ þ 3 sinhð3pÞ 5 sinhð5pÞ p sinhðpÞ 2 2 2 ¼ ðh1 þ h2 þ h3 þ h4 Þ ð4=pÞ ð0:19926 0:002994 þ 0:00007764Þ ¼ 0:24999 ðh1 þ h2 þ h3 þ h4 Þ; for first three terms ¼ ðh1 þ h2 þ h3 þ h4 Þ=4:001;
which conﬁrms the mean value theorem for a square.
5.4 Mean Value Theorem
265
Example 5.7 Solve Example 5.4 using the equations developed in Example 5.5. Solution The applicable equation is, refer Example 5.5, h i Þpy 1 sinh ð2j1 W 4 X ð2j 1Þpx h i sin hðx; yÞ ¼ h1 p j¼1 ð2j 1Þ sinh ð2j1ÞpH W W h i Þpx 1 sinh ð2j1 H 4X ð2j 1ÞpðH yÞ h i þ h4 sin p J¼1 ð2j 1Þ sinh ð2j1ÞpW H
ððiÞÞ
H
Here W = H = 2 and x = y = 1. This gives 4 hð1; 1Þ ¼ ðh1 þ h4 Þ p
h
1 X j¼1
sinh
ð2j1Þp 2
i
ð2j 1Þ sinh½ð2j 1Þp
sin
ð2j 1Þp 2
For the ﬁrst three terms, p sinh p2 sinh 3p sinh 5p 4 3p 5p 2 2 þ þ hð1; 1Þ ¼ ð300 þ 100Þ sin sin sin p 2 2 2 sinhðpÞ 3 sinhð3pÞ 5 sinhð5pÞ ¼ 1600 0:19927 0:002994 þ 7:764 105 =p ¼ 100 C:
and hence tc ¼ 100 þ 100 ¼ 200 C: Example 5.8 Determine temperature at point (3/2, ½) for the conﬁguration of Example 5.4. Solution Eq. (i) of Example 5.7 applies. We get (refer Fig. 5.10)
300oC
0oC
(3/2, ½) 0oC
Fig. 5.10 Example 5.8
θ=T
100oC
100oC
266
5 SteadyState TwoDimensional Heat Conduction
h i ð2j1Þp 1 sinh X 4 3 1 4 ð2j 1Þ3p sin h ; ¼ h1 2 2 p j¼1 ð2j 1Þ sinh½ð2j 1Þp 4 h i Þ3p 1 sinh ð2j1 4 4X ð2j 1Þ3p sin þ h4 p j¼1 ð2j 1Þ sinh½ð2j 1Þp 4 For the ﬁrst three terms, we have 3p 5p 3p sin 4 sinh p4 sin 9p sin 15p 3 1 4 4 sinh 4 4 sinh 4 h ; þ þ ¼ 300 2 2 p sinh p 3 sinh 3p 5 sinh 5p 3p 9p 9p 15p 3p sin 4 sinh 4 sin 4 sinh 4 sin 4 sinh 15p 4 4 þ þ þ 100 ¼ 63:64 C: p sinh p 3 sinh 3p 5 sinh 5p and
3 1 t ; ¼ 63:64 þ 100 ¼ 163:64 C: 2 2
Alternative Method The problem can be reduced into a set of simple subproblems as shown in Fig. 5.11 and then equations of Example 5.5 can be applied. Example 5.9 Determine temperature at point (3/2, ½) for the long rectangular section bar in Fig. 5.12. 0o
400o 0o
0o
0oC + 100o
0o
0o + 0o C
0o (i)
0oC + 0oC
0o
100o
0o
(ii)
(iii)
(iv)
Origin
Fig. 5.11 A set of subproblems of Fig. 5.10
y
2 400oC
100oC 100oC
(3/2, ½)
100oC
Fig. 5.12 Example 5.9
x
2
x
200o
5.4 Mean Value Theorem
267
Solution The temperature from Eq. (5.23b) is
1 4 X 1 ð2j 1Þpx ð2j 1Þpy h i sin sinh ðiÞ hðx; yÞ ¼ h p j¼1 ð2j 1Þ sinh ð2j1ÞpH W W W
The given problem can be transformed into the problem shown in Fig. 5.13 for which h = 300°C, x = 3/2, y = ½, W = H = 2. The calculated values of various terms in the above equation for j = 1–3 have been listed in the Table 5.1. Substitution in Eq. (i) gives hðx; yÞ ¼ ð4=pÞ 300 ½0:8687 0:707=11:5488 þ 5:227 0:707=ð6195:8 3Þ þ 25:367 ð0:707Þ=ð5 3317812Þ ¼ 20:4 C: Hence,
3 1 t ; ¼ 20:4 þ 100 ¼ 120:4 C: 2 2
Example 5.10 A rectangular section bar (W H) is very long in the zdirection. Its three sides are held at temperatures To while the fourth side (y = H) is held at a constant temperature Tb as shown in Fig. 5.14 Determine the equation of heat flow rate from the face y = 0.
2 300oC
0oC 0oC
(3/2, ½)
0oC
2
Fig. 5.13 Equivalent system of Fig. 5.12
Table 5.1 Example 5.9
j=1
sinh[(2j – 1) py/W]
sin[(2j – 1) px/W]
0.8687
0.707
11.5488
sinh[(2j–1) pH/W]
j=2
5.227
0.707
6195.82
j=3
25.367
–0.707
3317812
268
5 SteadyState TwoDimensional Heat Conduction y
H
Tb
To
0, 0
To
To (q)y=0
x W
Fig. 5.14 Example 5.10
Solution For the given bar, the heat flow from the face y = 0 for unit length in zdirection is x¼W x¼W x¼W ! Z Z Z @T @h qy¼0 ¼ dqy ðx; 0Þdx ¼ k k dx ðiÞ dx ¼ @y @y x¼0
x¼0
y¼0
x¼0
y¼0
where h = T–To. From Eq. (5.23), we have h ¼ ðTb To Þ
1 npx npy 2X ð1Þn þ 1 þ 1 npH sin sinh p n¼1 n sinh W W W
Hence, 1 npx @ h npyi @h 2X ð1Þn þ 1 þ 1 sinh ¼ ðT T Þ sin b o @y y¼0 p n¼1 n sinh npH W @y W y¼o W n þ 1 1 n o 2 X ð1Þ þ1 npx np npy npH sin cosh ¼ ðTb To Þ p n¼1 n sinh W W W W y¼o n þ 1 1 2 X ð1Þ þ1 npx np npH sin ¼ ðTb To Þ p n¼1 n sinh W W W Substitution in Eq. (i) gives qy¼0
x¼W Z 1 npx np 2X ð1Þn þ 1 þ 1 npH sin ¼ kðTb To Þ dx p n¼1 n sinh W W W x¼0
x¼W Z 1 npx 2X ð1Þn þ 1 þ 1 np sin dx ¼ kðTb To Þ p n¼1 n sinh npH W W W x¼0
1 npxix¼W 2X ð1Þn þ 1 þ 1 h npH cos ¼ kðTb To Þ p n¼1 n sinh W W x¼0
5.4 Mean Value Theorem
269
or qy¼0 ¼ kðTb To Þ
5.5
1 2X ð1Þn þ 1 þ 1 ½1 cosðnpÞ p n¼1 n sinh npH W
Graphical Analysis of Two Dimensional SteadyState Conduction: Thermal Flux Plotting and Shape Factor
Consider a quarter, Fig. 5.15b, of a chimney crosssection, Fig. 5.15a, where the inner surface is at some uniform temperature T1 and the outer surface is at uniform temperature T2. It is a twodimensional system. Isotherms (constant temperature) and heat flow lines (adiabats) have been sketched therein following the rule that the heat flow lines cut the isotherms at right angles. To determine the heat flow, let us consider an elemental curvilinear square1 abcd in a heat flow channel. The heat flow through this channel is Dq ¼ kðDx 1Þ
DT Dy
where the depth of the element (perpendicular to the plane of the paper) has been taken as unity. If the curvilinear squares are drawn such that Dx = Dy for all such elements, then Dq ¼ kDT: Let there be N number of temperature increments (uniformly spaced isotherms) between inner and outer surfaces, then the drop in temperature across any two adjacent isotherms is DT ¼ ðT1 T2 Þ=N: and Dq ¼ kðT1 T2 Þ=N: If there are M number of heat flow channels, the total heat flow through the quarter section will be the sum of heat flow through all channels, i.e. q ¼ MDq ¼ k ðM=N ÞðT1 T2 Þ ¼ kSðT1 T2 Þ
ð5:31Þ
where S = M/N is termed as conduction shape factor. Shape factors for some common geometrical systems and thermal conditions of practical utility have been worked out by various researchers. They are summarized in Table 5.2. From the above analysis, it can be seen that to determine the total heat flow through a given conﬁguration, it is required to draw curvilinear square elements and count the number 1
Curvilinear squares closely approximate the true squares only in the limit when their number approaches inﬁnity. In ﬁnite form, they must be drawn such that average lengths of the opposite sides of the curvilinear squares are nearly equal and keeping internal angles close to 90o.
270
5 SteadyState TwoDimensional Heat Conduction
a d
T1
b c
Typical heat flow lines
T2
Isotherms
a
b
Fig. 5.15 a Chimney crosssection b Flux plotting in one quarter
Table 5.2 Conduction shape factors S. No.
Physical system
Shape factor
Restrictions
1
Plane wall of thickness d
Schematic
A/d
Onedimensional heat flow
2
Hollow cylinder, length L (radial heat flow)
2pL ln RR21
L >> R
3
Hollow sphere
4pR1 R2 R2 R1

4
Isothermal sphere of radius R buried in semiinﬁnite medium having an isothermal surface
4pR R 1 2Z
z > D/2
(continued)
5.5 Graphical Analysis of Two Dimensional SteadyState Conduction …
271
Table 5.2 (continued) S. No.
Physical system
Shape factor
Restrictions
5
Isothermal sphere of radius R buried in an inﬁnite medium
Schematic
4pR (This value can be obtained by putting R2 = ∞ in equation of Case 3, or by putting z = ∞ in that of Case 4).

6
Isothermal sphere of radius R buried in semiinﬁnite medium with an insulated surface
4pR R 1 þ 2Z

7
Two isothermal spheres buried in inﬁnite medium
R2 R1
4pD
C4 1 1C 2 2
1
D > 5Rmax 2C2
where C1 = R1/D, C2 = R2/D (R2 refers to smaller sphere)
8
9
10
Isothermal horizontal cylinder of radius R buried in semiinﬁnite medium having isothermal surface
Two isothermal cylinders buried in inﬁnite medium
Eccentric cylinders of equal lengths
2pL cosh1 ZR
L >> R
2pL ln 2Z R
L >> R Z > 3R
2pL h lnðL=2DÞ i ln RL 1lnðL=RÞ
Z >> R L >> z
2pL
L >> R1, R2 L >> D
2pL
L >> 2R2 R2 > R1
D2 R2 R2 cosh1 2R11R2 2
R2 þ R2 e2 cosh1 1 2R1 R2 2
(continued)
272
5 SteadyState TwoDimensional Heat Conduction
Table 5.2 (continued) S. No.
Physical system
Shape factor
Restrictions
11
Cylinder of length L centred in a square duct of equal length
Schematic
2pL ln 0:54W R
L >> W W > 2R
12
Isothermal cylinder of radius R placed vertically in semiinﬁnite medium having isothermal surface
2pL ln 2L R
L >> 2R
13
Thin plate buried in semiinﬁnite medium
pW ln 4W L
z=0 W>L
2pW ln 4W L
z >> W W>L
14
An edge formed by the intersection of two plane walls; temperatures of inner and outer walls are T1 and T2, respectively
0.54L
W > d /5 H > d /5
15
Corner at the intersection of three plane walls, each of thickness d; temperatures of inner and outer walls are T1 and T2, respectively
0.15d
Inside dimension > d/5
16
Isothermal rectangular parallelepiped buried in semiinﬁnite medium having isothermal surface
1:685L
0:59 log 1 þ Z Z 0:078 W
L >> W, H, z
H
5.5 Graphical Analysis of Two Dimensional SteadyState Conduction …
273
of heat flow channels and the temperature increments. However, the thermal flux plotting depends on the skill of the person plotting the lines. Electrical analogy, discussed in Sect. 5.6, may be used to sketch the isotherms and heat flow lines. It is to note that this technique is now not much used but a rough sketching can sometimes help in making a fairly good estimate of the temperatures and the basic concepts of the plots can help in a quick check of the result from other methods. Hence, it must not be totally overlooked. Example 5.11 Determine the heat transfer rate per unit length by flux plotting through the 300 mm thick insulation on a 200 mm outer diameter pipe. The temperature of the inner surface of the insulation is 150°C and outer surface is at 50°C. The thermal conductivity of the insulating material is 0.06 W/(m K). Check the result against the analytical solution. Solution Concentric circles of 200 and 800 mm are drawn accurately. The network of the curvilinear squares can be constructed by freehand plotting in a quarter only, refer Fig. 5.16. There are approximately 5.25 squares in each heat flow lane, and there are a total of 24 flow lanes (6 in each quarter), i.e. N 5.25 and M = 24. This gives S
24 ¼ 4:57 5:25
and q ¼ kSðT1 T2 Þ ¼ 4:57 0:06 ð150 50Þ ¼ 27:42 W/m: L From the analytical solution, q T 1 T2 150 50 ¼ 2pk ¼ 2p 0:06 400 ¼ 27:19 W/m: R2 L ln 100 ln R1
The result of the freehand plotting is in a reasonable agreement with the analytical solution (about 0.8% higher than that of the analytical solution).
Fig. 5.16 Flux plotting Example 5.11
274
5 SteadyState TwoDimensional Heat Conduction
Example 5.12 A 25 mm OD pipe carries a hot fluid (pipe inner surface temperature = 80° C). It is placed centrally in a thinwalled square duct (85 85 mm2). The space between the pipe and the duct is completely ﬁlled with glass wool insulation [k = 0.055 W/(m K)]. Calculate the heat loss for one meter length of the pipe if T2 = 20°C. Solution From the flux plotting, refer Fig. 5.17, Nav ð6:3 þ 6:4 þ 6:8 þ 7:3Þ=4 ¼ 6:7: For the 4 flow lanes in oneeighth region of the conﬁguration as shown in the ﬁgure, the shape factor S ð8 4Þ=6:7 ¼ 4:78 L and the heat loss per meter length of the pipe is q S ¼ k ðT1 T2 Þ ¼ 0:055 4:78 60 ¼ 15:77 W/m: L L From Case 11, Table 5.2, S 2p L ¼ lnð0:54W Þ ¼ 4:83, which is very close to that found from flux plotting. R Example 5.13 Determine the shape factor for the conﬁguration shown in Fig. 5.18. If the conﬁguration refers to a squaresection duct (a a) covered with ﬁreclay of thickness 1.5a to give an outer shape of 4a 4a square, determine the heat loss for a = 20 mm, T1 = 100°C and T2 = 10°C. Thermal conductivity of ﬁreclay is 1.05 W/(m K).
Fig. 5.17 Flux plotting Example 5.12
5.5 Graphical Analysis of Two Dimensional SteadyState Conduction …
275
Very long in direction z.
Fig. 5.18 Flux plotting Example 5.13
Solution Flux plot on oneeight region of the conﬁguration has been shown in Fig. 5.18. Freehand plotting for this region gives Nav ð8:1 þ 8:15 þ 8:25 þ 8:4 þ 8:7Þ=5 ¼ 8:32: The shape factor, S M ¼ ¼ ð8 5Þ=8:32 ¼ 4:81: L Nav Heat loss through the insulation, per meter length of the duct, is q S ¼k ðT1 T2 Þ ¼ 1:05 4:81 ð100 10Þ ¼ 454:54 W/m: L L
Example 5.14 Determine the shape factor for the following systems. (i) (ii) (iii) (iv)
a plane wall, a hollow cylinder of length L (L > > r), a hollow sphere, an isothermal sphere of radius R buried in an inﬁnite medium.
Solution (i) A plane wall (thickness d) q ¼ kAðT1 T2 Þ=d ¼ kðA=dÞ:ðT1 T2 Þ Hence, S = (A/d).
276
5 SteadyState TwoDimensional Heat Conduction
(ii) A hollow cylinder q ¼ 2pkL
T1 T2 2pL ¼k ðT1 T2 Þ lnðR2 =R1 Þ lnðR2 =R1 Þ
Hence, S¼
2pL : lnðR2 =R1 Þ
(iii) A hollow sphere q ¼ 4pkR2 R1
T1 T2 4pR2 R1 ¼k ð T 1 T2 Þ ð R2 R1 Þ ð R2 R1 Þ
Hence, S¼
4pR2 R1 1 1 1 ¼ 4p : R 1 R2 ð R2 R1 Þ
(iv) An isothermal sphere of radius r buried in inﬁnite medium For the inﬁnite medium, put R2 = ∞ and R1 = R in shape factor equation of case (iii). This gives S ¼ 4pR:
Example 5.15 A 1 m diameter sphere of radioactive material generates heat at the rate of 5 kW. It is buried 1 m in the earth [k = 1 W/(m K)], which has surface temperature of 30°C. What is the surface temperature of the sphere? Solution The shape factor from Table 5.2 (Case 4) S¼
4pR ¼ 4p0:5 ¼ 8:378 0:5 ð12zR Þ ð121 Þ q ¼ kSðT1 T2 Þ
or 5000 ¼ 1 8:378 ðTs 30Þ or Ts ¼ 626:8 C:
5.5 Graphical Analysis of Two Dimensional SteadyState Conduction …
277
Example 5.16 A 50 mm onedimensional thinwalled pipe carries a cold fluid at 5°C. It is enclosed by a 200 mm pipe, whose surface is at 40°C. The space between the two pipes is ﬁlled with dry sawdust [k = 0.06 W/(m K)]. What is the heat transfer rate per m length of the pipe? If the inner pipe is placed eccentric with 25 mm eccentricity, what will be the heat transfer rate? Solution (i) Concentric cylinders q ¼ kSðT1 T2 Þ L 2p ¼ k ðT1 T2 Þ ln RR12 2p ¼ 0:06 100 ð40 5Þ ¼ 9:52 W/m: ln 25 (ii) Eccentric cylinders S 2p 2p ¼ 2 þ 1002 252 ¼ 4:77 ¼ L cosh1 R21 þ R22 e2 cosh1 25 225100 2R1 R2
q S ¼ k ðT1 T2 Þ ¼ 0:06 4:77 ð40 5Þ ¼ 10:02 W/m: L L Example 5.17 A 500 mm OD pipeline transports crude oil at 100°C. In order to reduce the heat loss from the pipeline, two schemes have been proposed: (i) one of the schemes is to bury it 3.0 m below the earth’s surface [ke = 1.0 W/(m K)], (ii) the second scheme is to cover it with 50 mm thick glass wool insulation [kg = 0.05 W/(m K)]. If the earth’s surface temperature is 0°C and the temperature of the outer surface of the insulation is also estimated to be 0°C, compare the schemes. Solution (i) Heat loss per unit length of the pipeline buried in the earth, Case 8, Table 5.2 S 2p 2p ¼ 2Z ¼ 23 ¼ 1:977 L ln R ln 0:25 q S ¼ ke ðT1 T2 Þ ¼ 1:0 1:977 ð100 0Þ ¼ 197:7 W/m: L L
278
5 SteadyState TwoDimensional Heat Conduction
(ii) Insulation covering: q S ¼ kg ðT1 T2 Þ L L 2p ¼ kg R2 ðT1 T2 Þ ln R1
2p ¼ 0:05 300 ð100 0Þ ¼ 172:3 W/m: ln 250
Scheme 2 is better. However, the ﬁnal choice of the scheme is based on economics. Example 5.18 A small cubical furnace is of 0.75 m sides on the inside. It is made of 150 mm thick ﬁreclay bricks [k = 1.3 W/(m K)]. If the inside and outside surfaces are at 500°C and 100°C, respectively, determine the rate of heat loss from the furnace through its walls. Solution (i) Walls: There are six 0.75 0.75 m2 surface areas, which can be treated as onedimensional conduction cases. Thus, the heat loss through these walls is qw ¼ 6½kAðT1 T2 Þ=d ¼ 6 ½1:3 ð0:75 0:75Þ ð500 100Þ=0:15=1000 ¼ 11:70 kW: (ii) Edges: There are 12 edges, each 0.75 m long. These are twodimensional problems for which the total shape factor from Table 5.2 is Se ¼ 12 ð0:54LÞ ¼ 12 0:54 0:75 ¼ 4:86: (iii) Corners: The 8 threedimensional corners have the total shape factor, refer Table 5.2, Sc ¼ 8 ð0:15dÞ ¼ 8 0:15 0:15 ¼ 0:18: The heat loss through the edges and corners is qe þ qc ¼ kðSe þ Sc ÞðT1 T2 Þ ¼ 1:3 ð4:86 þ 0:18Þ 400 =1000 ¼ 2:62 kW: Total heat loss = 11.70 + 2.62 = 14.32 kW approx. Example 5.19 Combustion products at an average temperature of 800°C flow through a 20 m high chimney whose crosssection is shown in Fig. 5.19. The ambient temperature is 30°C. Determine the heat loss from the chimney if the inside and outside ﬁlm coefﬁcients are 100 and 10 W/(m2 K), respectively.
5.5 Graphical Analysis of Two Dimensional SteadyState Conduction …
279
k = 1.3
250
320 Sq
Fig. 5.19 Example 5.19
Solution From Table 5.2, Case 11, conduction shape factor, S¼
2pL 2p 20 0:54W ¼ 0:540:32 ¼ 388:06: ln R ln 0:125
The thermal resistances are: Inside convective, Ri ¼
1 1 ¼ 6:366 104 : ¼ 2pR1 Lhi 2p 0:125 20 100
Conductive, Rk ¼
1 1 ¼ ¼ 1:982 103 : kS 1:3 388:06
Outside convective, Ri ¼
1 1 ¼ 3:906 103 : ¼ 4WLho 4 0:32 20 10
The heat loss, q¼
ðT1 T 2 Þ ð800 30Þ ¼ ¼ 118kW: 4 Ri þ Rk þ Ro 6:366 10 þ 1:982 103 þ 3:906 103
Example 5.20 Repeat the above example for the 250 250 mm2 flue passage, Fig. 5.20. Comment on the result.
280
5 SteadyState TwoDimensional Heat Conduction
250 Sq
320 Sq
Fig. 5.20 Example 5.20
Solution From Table 5.2, Swalls ¼ 4ðA=dÞ ¼ 4 ð0:25 20Þ=0:035 ¼ 571:43: Shape factor for edges: Se ¼ 4 ð0:54LÞ ¼ 4 0:54 20 ¼ 43:2: Hence, Stotal ¼ 571:43 þ 43:2 ¼ 614:63: q¼
ðT1 T 2 Þ ð800 30Þ ¼ 136:1 kW: ¼ 1 1 1 1 1 1 þ þ þ þ hi Ai kS ho Ao 100 1 20 1:3 614:63 10 1:28 20
The heat flow in this case is more because the mass of the wall resisting conduction heat flow is less. Example 5.21 A radioactive brick (50 mm 100 mm 200 mm) at 800°C is buried 750 mm deep in the earth [k = 0.5 W/(m K)]. Determine the heat transfer rate when the surface temperature of the earth is 30°C. Solution From Table 5.2, Case 16, h z i0:59 z 0:078 S ¼ 1:685L log 1 þ W H 0:75 0:59 0:75 0:078 ¼ 1:685 0:2 log 1 þ ¼ 0:2849 0:1 0:05
5.5 Graphical Analysis of Two Dimensional SteadyState Conduction …
281
Heat transfer rate, q ¼ kSDT ¼ 0:5 0:2849 ð800 30Þ ¼ 109:7 W:
Example 5.22 A furnace is 1.2 m by 0.9 m by 0.8 m. It is made of a brick wall of 150 mm thickness [for bricks, k = 1.3 W/(m K)]. Inside and outside temperatures are 500°C and 50°C, respectively. Determine the heat loss through the walls. Solution Inside dimensions are 0.9 0.6 0.5 m3. Plane walls: 2 nos. of 0.9 0.6 m2, 2 nos. of 0.9 0.5 m2 and 2 nos. of 0.6 0.5 m2. The total area is Awall ¼ 2ð0:9 0:6 þ 0:9 0:5 þ 0:6 0:5Þ ¼ 2:58 m2 : Heat transfer through the walls, qwall ¼ kADT=d ¼ 1:3 2:58 ð500 50Þ=0:15 ¼ 10:06 kW: Shape factors Edges: Four edges of 0.9 m, four of 0.6 m and four of 0.5 m. Total shape factor for the edges, Se ¼ 4 0:54 ð0:9 þ 0:6 þ 0:5Þ ¼ 4:32 : Corners: There are 8 similar corners. Total shape factor for these corners, Sc ¼ 8 0:15 0:15 ¼ 0:18: Heat loss through the edges and corners is qe þ qc ¼ kSDT ¼ 1:3 ð4:32 þ 0:18Þ ð500 50Þ ¼ 2:63 kW: Total heat loss, q = 10.06 + 2.63 = 12.69 kW. Example 5.23 In order to recover oil, steam is supplied through a 250 mm diameter and 250 m long pipe laid down into the earth [k = 1.0 W/(m K)]. Determine the heat loss from the pipe if the steam is at 150°C and the earth is at 20°C. Solution From Case 12 of Table 5.2, the shape factor, S¼
2pL 2p 250 ¼ 2250 ¼ 189:4 ln 2L ln 0:125 R
282
5 SteadyState TwoDimensional Heat Conduction
The heat loss, q ¼ kSDT ¼ 1:0 189:4 ð150 20Þ ¼ 24:62 kW:
Example 5.24 Determine the equation of shape factor for Case 8 of Table 5.2 from that of Case 9. Solution For Case 9, S 2p ¼ L cosh1 D2 R21 R22 2R1 R2
Putting D = z + R1, the denominator becomes cosh
1
z2 þ 2R1 z R22 2R1 R2
¼ cosh
1
z2 z R2 þ 2R1 R2 R2 2R1
When R1 tends to inﬁnity, the cylinder coverts into plane isothermal surface, i.e. to Case 8, and cosh1
z2 z R2 þ 2R1 R2 R2 2R1
¼ cosh1
where R = R2. Thus S 2p ; ¼ L cosh1 Rz which is the desired result. t2 = 5oC
z = 1.2 m
Oil at 100oC
Insulation ki = 0.27 W/(m K)
hi = 200 W/(m2 K) a
Ri
Rins
ti
t2 q b
Fig. 5.21 Example 5.25
Rsoil
z R
5.5 Graphical Analysis of Two Dimensional SteadyState Conduction …
283
Example 5.25 A 500 mm diameter long pipe carrying oil at 100°C is covered with 250 mm thick insulation. It is buried in soil [ksoil = 0.27 W/(m K)] at a depth of 1.2 m as shown in Fig. 5.21a. Determine the rate of heat transfer from the oil if the convection heat transfer coefﬁcient from oil to pipe surface is 500 W/(m2 K). Soil surface temperature is 5°C. Neglect pipe wall resistance. Solution The electric network is shown in Fig, 5.21b. The total resistance to heat transfer is Rt ¼ Ri þ Rins þ Rsoil For unit length of pipe, 1 1 ¼ hi Ai hi pdi L 1 ¼ 0:00318: ¼ 200 p 0:5 1
Ri ¼
Resistance of insulation,
1 r1 Ri ¼ ln 2pki L ri 1 0:5 ln ¼ ¼ 1:84: 2p 0:06 1 0:25
From Table 5.2, Case 8, Rsoil ¼ ¼
1 cosh1 ðz=RÞ 1 ¼ Sksoil 2pL ksoil cosh1 ð1:2=0:5Þ 1 ¼ 0:897: 2p 1 0:27
Hence, Rt ¼ Ri þ Rins þ Rsoil ¼ 0:00318 þ 1:84 þ 0:897 ¼ 2:74: Heat transfer rate, q¼
5.6
ti t2 100 5 ¼ 34:67 W=m length: ¼ 2:74 Rt
Experimental Investigation of Conduction Process by Method of Analogy: ElectroThermal Analogy
The electrothermal analogy has been employed to sketch the curvilinear squares experimentally. It is based on the similarity of the mathematical equations of the steadystate heat conduction and electric conduction presented by the Laplace equations given below.
284
5 SteadyState TwoDimensional Heat Conduction Perspex sheet Water + salt Copper plates
Copper plates
Copper plate
Perspex sheet A
C Perspex sheet
D
G
Copper plate
R Perspex sheets
Perspex sheets
+
B
(b) A
C
Copper plate (a)
G
D
B (c)
Fig. 5.22 Experimental setup for flux plotting
@2t @2t þ ¼0 @x2 @y2
ð5:1Þ
@2E @2E þ 2 ¼0 @x2 @y
ð5:32Þ
where E is the electrical potential. Thus, the thermal phenomenon can be investigated by the study of the electrical one. The experimental investigation of the later is simpler than the thermal process. The method involves the construction of an electric analog and experimentally drawing the isothermal (equipotential) and heat flow (current flow) lines. An experimental setup based on this analogy is discussed below. The setup, shown in Fig. 5.22a, uses two similar hollow boxes ‘1’ and ‘2’ made with two parallel sides of copper plates and the other two of an insulating material (say perspex sheet). The bottom of the boxes is also made of the insulating material. The boxes represent a scaled model of the chimney crosssection discussed earlier. A metallic tracer at ‘A’ in Fig. 5.22b can trace the whole area of the box at a certain depth. The tracer is connected to the moving point ‘B’ of a rheostat ‘R’ through a galvanometer ‘G’. Some electrical potential difference (say 12 V) is applied to the two ends of the rheostat and the same is applied across the two parallel copper plates of the hollow box. The box water, made conductive by adding some salt, is ﬁlled to such a level that the point of the metallic
5.6 Experimental Investigation of Conduction Process by Method of Analogy …
285
tracer touches the water surface. Thus the rheostat and the box form a Wheatstone bridge, see Fig. 5.22c. Tracer ‘A’ is moved, touching the water surface, such that the galvanometer ‘G’ always shows zero deflection. The path of the tracer is thus a uniform potential line and this path can be plotted on a paper with the help of a traversing mechanism. Setting the movable point of the rheostat at different positions, constant potential lines can be plotted for the whole of the section. Repeating the above experiment with the second box, similar lines can be plotted which will be found to be nearly perpendicular to the lines drawn with the ﬁrst box. The method leads to a better plot of the curvilinear squares than by freehand plotting and, thus, gives a better estimate of the value of the shape factor.
5.7
Numerical Solution Methods
The direct integration of the differential equations has been used to solve simple problems of twodimensional, threedimensional and transient heat conduction problems but success in solving complex problems, involving nonlinear boundary conditions and temperature or positiondepending thermal properties, is limited. Such problems have been solved using numerical methods. Commonly used numerical methods are: ﬁnitedifference and ﬁniteelement methods. The ﬁnitedifference method has been used extensively in solving heat conduction problems because of its simplicity in implementation. The ﬁniteelement method is being widely used to solve problems in structural mechanics. It requires much greater mathematical efforts and has been used for solving heat conduction problems involving complicated geometries. Here only the basic principles of the ﬁnitedifference method are presented. The application of the method has been illustrated with some examples.
5.7.1 FiniteDifference Method The ﬁrst step in the ﬁnitedifference method is to discretize the spatial and time coordinates to form a mesh of nodes. Then by applying the energy balance to the volume elements surrounding the nodes, a set of linear algebraic equations is obtained. These equations, consisting of as many unknowns as the number of nodes in the mesh, are solved by metric inversion or by iteration. The accuracy of the ﬁnitedifference approximation increases with number of nodes hence computers are used to obtain ﬁnitedifference solutions. Standard computer programs are available for this purpose. A coarse mesh with few nodes has been used in this section (suitable for calculation by hand) to clarify the basic principles of the method. Figure 5.23 shows section of a body which has been divided into a number of small volumes by using equal divisions in x and ydirections. The nodal points (nodes) of these volumes are designated for a twodimensional system as (m + 1, n), (m, n), (m–1, n), etc., where m and n indicate the x and y increments, respectively. The thermal properties of each volume are assumed to be concentrated at the nodal points. In order to establish temperatures at the nodal points, we use ﬁnite differences to approximate the differentials in the Laplace equation, Eq. (5.1). We shall consider the case of constant thermal conductivity of the material.
286
5 SteadyState TwoDimensional Heat Conduction
y (m, n+2) (m, n+1) (m2, n)
Δy y
(m+1, n)
(m1, n)
x
Δy y
(m+2, n)
(m, n)
(m, n1) (m, n2)
node
x
Dashed lines indicate the element [the node (m, n)] with volume ( . y.1); unit depth.
x
Fig. 5.23 Nomenclature for twodimensional numerical analysis
The temperature gradients can be written as follows with reference to Fig. 5.23.
@t Tm þ 1;n Tm;n @x m þ 1=2;n Dx @t Tm;n Tm1;n @x Dx m1=2;n @t Tm;n þ 1 Tm;n @y Dy m;n þ 1=2 @t Tm;n Tm;n1 @y m;n1=2 Dy @t @t 2 @x m þ 1=2;n @x m1=2;n Tðm þ 1;nÞ þ Tðm1;nÞ 2Tðm;nÞ @ t ¼ @x2 m;n Dx ðDxÞ2 @t @t 2 @y @y Tðm;n þ 1Þ þ Tðm;n1Þ 2Tðm;nÞ @ t m;n þ 1=2 m;n1=2 ¼ @y2 m;n Dy ðDyÞ2 Substitution in Eq. (5.1) yields Tðm þ 1;nÞ þ Tðm1;nÞ 2Tðm;nÞ ðDxÞ
2
þ
Tðm;n þ 1Þ þ Tðm;n1Þ 2Tðm;nÞ ðDyÞ2
¼0
5.7 Numerical Solution Methods
287
If Dx = Dy, then Tðm þ 1;nÞ þ Tðm1;nÞ þ Tðm;n þ 1Þ þ Tðm;n1Þ 4Tðm;nÞ ¼ 0
ð5:33Þ
This is known as nodal or temperature equation. Equation (5.33) can also be obtained making use of the method of heat balances. Consider node (m, n) in Fig. 5.23. The rates of heat conduction from the nodes (m–1, n), (m +1, n), (m, n–1) and (m, n +1) to node (m, n) are given by the following equations. Tm1;n Tm;n Dx Tm þ 1;n Tm;n kðDy:1Þ Dx Tm;n1 Tm;n kðDx:1Þ Dy Tm;n þ 1 Tm;n kðDx:1Þ Dy kðDy:1Þ
In the steady state, the net heat transfer to the node (m, n) is zero. For Dx = Dy, it gives Eq. (5.33). The use of the method of heat balance is illustrated in the examples to follow. Example 5.26 Write down the nodal equation for the node situated at the corner as shown in Fig. 5.24 with one side insulated and the adjacent side subjected to convective heat transfer. Solution The rate of heat conducted between 2 and 1, and 3 and 1, respectively, is k
bDy ðT2 T1 Þ 2 Dx
k
bDx ðT3 T1 Þ 2 Dy
and
where b is the depth perpendicular to the plane of the paper.
h, T 2
1 3
Fig. 5.24 Example 5.26
Insulated
288
5 SteadyState TwoDimensional Heat Conduction
The rate of heat convected into node 1 is hbDx ð T1 T1 Þ ¼ 0 2 In the steady state, the summation of energy transfer rate into node 1 is zero. Thus bDy ðT2 T1 Þ bDx ðT3 T1 Þ hbDx þk þ k ð T 1 T1 Þ ¼ 0 2 Dx 2 Dy 2 Assuming a square grid (Dx = Dy), the equation simpliﬁes to hDx hDx þ 2 T1 ¼ 0 T2 þ T3 þ T1 k k which is the desired result. Example 5.27 Considering onedimensional heat conduction through a thin ﬁn, show that the nodal equation for the node shown in Fig. 5.25 can be expressed as "
# " # hPðDxÞ2 hPðDxÞ2 þ 2 Tm T1 ðTm þ 1 þ Tm1 Þ ¼ 0 kA kA
Solution The rate of heat conducted between nodes (m–1) and (m), and between nodes (m + 1) and (m) are, respectively, kA
ðTm1 Tm Þ Dx
kA
ðTm þ 1 Tm Þ Dx
and
The rate at which the heat is convected from the ﬁn surface is hPDxðT1 Tm Þ
(m
Fig. 5.25 Example 5.27
1)
(m)
(m + 1)
5.7 Numerical Solution Methods
289
Fig. 5.26 Example 5.28
For the steady state, the summation of the heat transfer rates into the node (m) must be zero, and thus kA
ðTm1 Tm Þ ðTm þ 1 Tm Þ þ kA þ hPDxðT1 Tm Þ ¼ 0 Dx Dx
which on simpliﬁcation gives "
# " # hPðDxÞ2 hPðDxÞ2 þ 2 Tm T1 ðTm þ 1 þ Tm1 Þ ¼ 0 kA kA
Example 5.28 Write down nodal (temperature) equations for the cases shown in Fig. 5.26a–h referring to the numerical solution of twodimensional steadystate conduction problems. Consider the dimension perpendicular to the plane of paper as b.
290
5 SteadyState TwoDimensional Heat Conduction
Solution The solid crosssections in Fig. 5.26a–h have been divided into arbitrary lengths and widths of Dx and Dy, respectively. Let Dx = Dy = d as depicted in Fig. 5.26a. Case (a): The nodal point ‘1’ lies on an isothermal surface, i.e. T1 = T2 = T3. Hence, there is no heat flow between nodal points 1 and 2 and between 1 and 3. Hence, the nodal equation is qb ðbdÞ þ kðbdÞ
ðT4 T1 Þ ¼0 d
or qb
d þ ðT4 T1 Þ ¼ 0: k
Case (b): The nodal point ‘1’ lies on an insulated surface (as T1 = T5). Hence q51 = 0. The nodal equation can be written as q21 þ q31 þ q41 ¼ 0 or k
bd ðT2 T1 Þ ðT3 T1 Þ bd ðT4 T1 Þ þ kðbdÞ þk ¼0 2 d d 2 d
or T2 þ T4 þ 2T3 4T1 ¼ 0 It is to be noted that area of the heat flow between 21 and 41 is half of that between 31. Case (c): The surface is in contact with a fluid at bulk temperature T∞. The ﬁlm coefﬁcient is h at the surface. The nodal equation is q51 þ q41 þ q31 þ q21 ¼ 0 or bd ðT4 T1 Þ bd ðT3 T1 Þ ðT2 T1 Þ þk þ kðbdÞ ¼0 hðbdÞðT1 T1 Þ þ k 2 d 2 d d or hd ðT4 T1 Þ ðT3 T1 Þ þ þ ðT2 T1 Þ ¼ 0 ðT1 T1 Þ þ k 2 2
5.7 Numerical Solution Methods
291
or ðT3 þ T4 Þ hd hd þ T2 þ T1 T1 2 þ ¼0 2 k k Case (d): The nodal point ‘1’ lies at the corner of two insulated surfaces. Hence, heat flows only between nodal points 21 and 31. The nodal equation for this case is q21 þ q31 ¼ 0 or k
bd ðT2 T1 Þ bd ðT3 T1 Þ þk ¼0 2 d 2 d
or T2 þ T3 2T1 ¼ 0 Case (e): The nodal point ‘1’ lies at the corner of two isothermal surfaces. Hence, there is no heat flow to the nodal point 1 and there is no nodal equation for this case. Case (f): The nodal point ‘1’ lies in a corner where the surfaces are in contact with a fluid at bulk temperature T∞. Hence, the nodal equation is
bd ðT2 T1 Þ bd ðT3 T1 Þ ðT4 T1 Þ ðT5 T1 Þ þk þ kðbdÞ þ kðbdÞ ¼0 hðbdÞðT1 T1 Þ þ k 2 d 2 d d d
or hd 1 1 ðT1 T1 Þ þ ðT2 T1 Þ þ ðT3 T1 Þ þ ðT4 T1 Þ þ ðT5 T1 Þ ¼ 0 k 2 2 or T2 þ T3 þ 2T4 þ 2T5 þ
2hd 2hd T1 T1 6 þ ¼0 k k
Case (g): Following the procedure applied to the above cases, the nodal equation can be written as q11 þ q21 þ q31 ¼ 0 or hðbdÞðT1 T1 Þ þ k
bd ðT2 T1 Þ bd ðT3 T1 Þ þk ¼0 2 d 2 d
292
5 SteadyState TwoDimensional Heat Conduction
or 2
hd ðT1 T1 Þ þ ðT2 T1 Þ þ ðT3 T1 Þ ¼ 0 k
or T2 þ T3 þ 2
hd hd T1 2T1 1 þ ¼0 k k
Case (h): For this case, q21 þ q31 þ q41 þ q51 ¼ 0 or k
bd ðT2 T1 Þ bd ðT3 T1 Þ ðT4 T1 Þ ðT5 T1 Þ þk þ kðbdÞ þ kðbdÞ ¼0 2 d 2 d d d
or T2 þ T3 þ 2T4 þ 2T5 6T1 ¼ 0
Example 5.29 Write the nodal equation for the interior node near a curved boundary as shown in Fig. 5.27. Solution The rate of heat conduction from 1, 2, (m, n–1) and (m–1, n) to the node (m, n) are aDx þ Dx T1 Tm;n 2 bDy bDy þ Dy T2 Tm;n k 2 aDx aDx þ Dx Tm;n1 Tm;n k 2 Dy k
1 2 (m 1, n)
(m, n) x
Fig. 5.27 Example 5.29
a. x y (m, n 1)
h, T∞
5.7 Numerical Solution Methods
293
and
k
bDy þ Dy Tm1;n Tm;n : 2 Dy
In the steady state, the summation of the conducted heat into the node (m, n) is zero. For Dx = Dy, this gives k
aþ1 bþ1 T1 Tm;n þ k T2 Tm;n þ kða þ 1Þ Tm;n1 Tm;n þ kðb þ 1Þ Tm1;n Tm;n ¼ 0 b a
or ða þ 1ÞT1 ðb þ 1ÞT2 þ þ ða þ 1ÞTm;n1 þ ðb þ 1ÞTm1;n b a ð a þ 1Þ ð b þ 1Þ þ þ ða þ 1Þ þ ðb þ 1Þ Tm;n b a ¼0 Dividing by (a + 1) (b + 1), we have T1 T2 Tm;n1 Tm1;n þ þ þ bðb þ1Þ aða þ 1Þ ðb þ 1Þ ða þ 1Þ 1 1 1 1 þ þ þ Tm;n bð b þ 1Þ að a þ 1Þ ð b þ 1Þ ð a þ 1Þ ¼0 Simpliﬁcation gives the ﬁnal equation as T1 T2 Tm;n1 Tm1;n 1 1 þ þ þ þ Tm;n ¼ 0; a b bð b þ 1Þ að a þ 1Þ ð b þ 1Þ ð a þ 1Þ which is the desired result. Nodal equations for some conﬁgurations of interest are given in Table 5.3. Example 5.30 Write down the nodal equation for a nodal point on the boundary between two materials as shown in Fig. 5.28. Solution The heat balance equation is, refer Fig. 5.28, ðq21 ÞA þ ðq21 ÞB þ ðq31 ÞA þ ðq31 ÞB þ q41 þ q51 ¼ 0 Using Fourier’s law for conduction heat transfer, we obtain for unit depth Dx T2 T1 Dx T2 T1 Dx T 3 T1 Dx T3 T1 :1 :1 :1 :1 þ kB þ kA þ kB 2 2 2 2 Dy Dy Dy Dy T4 T1 T5 T1 þ kB ðDy:1Þ ¼0 þ kA ðDy:1Þ Dx Dx
kA
294
5 SteadyState TwoDimensional Heat Conduction
Table 5.3 Nodal equation for different conﬁgurations Conﬁguration
Nodal equation Tðm1;nÞ þ Tðm þ 1;nÞ þ Tðm;n1Þ þ Tðm;n þ 1Þ 4Tðm;nÞ ¼ 0
Internal node Tðm;n1Þ þ Tðm;n þ 1Þ þ 2Tðm1;nÞ 4Tðm;nÞ ¼ 0
Node on an insulated surface ð1=2Þ½Tðm;n1Þ þ Tðm;n þ 1Þ þ 2Tðm1;nÞ þ hðDx=kÞT1 ½2 þ hðDx=kÞTðm;nÞ ¼ 0
Node on a surface in contact with fluid (continued)
5.7 Numerical Solution Methods
295
Table 5.3 (continued) Conﬁguration
Nodal equation Tðm;n1Þ þ Tðm1;nÞ 2Tðm;nÞ ¼ 0
Insulated external corner The node lies on the corner of two isothermal surfaces hence there is no nodal equation
External corner on isothermal surfaces Tðm;n1Þ þ Tðm1;nÞ þ 2hðDx=kÞT1 ½2 þ 2hðDx=kÞTðm;nÞ ¼ 0
External corner in contact with a fluid Tðm;n þ 1Þ þ Tðm þ 1;nÞ þ 2ðTðm1;nÞ þ Tðm;n1Þ Þ þ 2hðDx=kÞT1 ½6 þ 2hðDx=kÞTðm;nÞ ¼ 0
Interior corner in contact with a fluid (continued)
296
5 SteadyState TwoDimensional Heat Conduction
Table 5.3 (continued) Conﬁguration
Nodal equation Tðm;n þ 1Þ þ Tðm þ 1;nÞ þ 2ðTðm1;nÞ þ Tðm;n1Þ Þ 6Tðm;nÞ ¼ 0
Interior corner with insulated sides
Material A
Material B
(q21)A
2
4
1
(q31)A
3
Δx (q21)B 5
Δy
(q31)B
Fig. 5.28 Example 5.30
For Δx = Δy, we have kA 2
ðT2 T1 Þ þ
kB 2
ðT2 T1 Þ þ k2A ðT3 T1 Þ þ k2B ðT3 T1 Þ þ kA ðT4 T1 Þ þ kB ðT5 T1 Þ ¼ 0: Hence, ðkA þ kB ÞT2 þ ðkA þ kB ÞT3 þ 2kA T4 þ 2kB T5 4ðkA þ kB ÞT1 ¼ 0
If kA = kB, we get T2 þ T3 þ T4 þ T5 4T1 ¼ 0; which is the same as Eq. (5.33).
5.7 Numerical Solution Methods
297
Fig. 5.29 Example 5.31
Example 5.31 For the cylindrical segment shown in Fig. 5.29, write down nodal equations for node 1 to 3. The radial and angular spacings of nodes are shown in the ﬁgure. Solution Node 1 The heat balance equation for the twodimensional conduction system gives q21 þ q31 þ q41 þ q51 ¼ 0 T2 T1 T3 T1 þ kðDr:1Þ þk kðDr:1Þ ðr þ DrÞD/ ðr þ DrÞD/ 3Dr T 5 T1 ¼0 þk rþ D/:1 2 Dr
Dr T4 T1 rþ D/:1 2 Dr
Simpliﬁcation gives the nodal equation as
Dr 2 1 Dr 3Dr ðT2 þ T3 Þ þ r þ T4 þ r þ T5 D/ " ðr þ DrÞ 2 2 # 2 Dr 1 þ ðr þ Dr Þ T1 2 D/ ðr þ DrÞ ¼0
Node 2 The heat balance equation for this twodimensional system gives kðDr:1Þ
T1 T2 1 Dr T6 T2 1 3Dr T7 T2 rþ D/:1 rþ D/:1 þk þk þ hðDr:1ÞðT1 T2 Þ ¼ 0 2 2 2 2 ðr þ DrÞD/ Dr Dr
Simpliﬁcation gives the nodal equation as 2
Dr 2 1 Dr 3Dr ðDrÞ2 T1 þ r þ T1 T6 þ r þ T7 þ 2h D/" ðr þ DrÞ 2 2# kD/ 2 Dr 1 ðDrÞ2 þ ðr þ Dr Þ þ 2h 2 T2 D/ ðr þ DrÞ kD/ ¼0
298
5 SteadyState TwoDimensional Heat Conduction
Node 3 The heat balance equation for this twodimensional system gives T1 T3 1 Dr T 8 T3 1 3Dr T9 T3 rþ rþ kðDr:1Þ þk þk ¼0 D/:1 D/:1 2 2 2 2 ðr þ DrÞD/ Dr Dr Simpliﬁcation gives the nodal equation as
Dr 2 D/
2
" # 1 Dr 3Dr Dr 2 1 T1 þ r þ þ ðr þ Dr Þ T3 T8 þ r þ T9 2 ðr þ DrÞ 2 2 D/ ðr þ DrÞ
¼0 The above equation can also be obtained by putting h = 0 in the nodal equation of node 2.
5.7.2 Solution of Nodal Equations The use of the numerical method involves writing nodal equations for each node within the body. The resulting set of equations is solved to determine the temperature at each node. It must be noted that smaller the elemental volumes, greater are the number of nodes and more closely the true temperature is approximated. If the number of nodes is very large, the simultaneous solution of the resulting equations might not be possible if carried out by hand. For a small number of nodes, the hand calculation can be carried out. The techniques, which have been used, are being explained in the sections to follow.
5.7.2.1 Relaxation Method This method can be used to determine the steadystate temperatures at nodal points. The process in this method of solution is explained below with the help of an example. Let the nodal equations are Node 1 :
300 þ 400 þ T2 þ T4 4T1 ¼ 0
ðaÞ
Node 2 :
400 þ 100 þ T1 þ T3 4T2 ¼ 0
ðbÞ
Node 3 :
100 þ 200 þ T2 þ T4 4T3 ¼ 0
ðcÞ
Node 4 :
200 þ 300 þ T1 þ T3 4T4 ¼ 0:
ðdÞ
The relaxation method proceeds as follows. 1. It starts with an initial guess of unknown temperatures. Since the initially assumed values will usually be in error, the righthand side of the above nodal equations will differ from zero. So replace the zeros in Eqs. (a)–(d) with R1, R2, R3 and R4, respectively, where R1, R2, etc. are known as residuals. 700 þ T2 þ T4 4T1 ¼ R1
ðeÞ
5.7 Numerical Solution Methods
299
500 þ T1 þ T3 4T2 ¼ R2
ðfÞ
300 þ T2 þ T4 4T3 ¼ R3
ðgÞ
500 þ T1 þ T3 4T4 ¼ R4 :
ðhÞ
2. Next set up a table known as unit change table. This table shows the effect of 1o change of temperatures at a node on the residuals. Also, enter the effect on all residuals of block (overall) unit change (see Table 5.4). 3. Set up the relaxation table. Calculate the initial residuals from the initial gauss of the temperatures T1 through T4 using the residual equations (e)–(h) as illustrated in Table 5.5. The initially guessed temperature values are now changed such that the residuals are reduced to zero. This procedure should begin with relaxing the largest initial residual. If all residuals are of the same sign, make a block change. Note that a good initial guess of temperatures helps to minimize the effort required. In the lefthand column of the table are entered the change in temperatures from the initially assumed values. In the present example, the initial residuals are of the same sign, hence a block change has been made. Next relax the largest residuals, which are R1 and R2 taking the help of unit change table. First, choose any one, say R1 and overrelax it but not too much. The change in T1 has increased R2. Now, relax R2. Change in DT2 in the table has relaxed all the residuals. Check the result, so obtained, by substituting the temperature values so obtained in Eqs. (e)–(h). Example 5.32 Derive an expression for the residual of the node at the end of a ﬁn (refer Fig. 5.30). Assume onedimensional heat conduction.
Table 5.4 Unit change DR1
DR2
DR3
DR4
DT1 = + 1
–4
DT2 = + 1
+1
+1
0
+1
–4
+1
0
DT3 = + 1
0
DT4 = + 1
+1
+1
–4
+1
0
+1
–4
Block change = + 1
–2
–2
–2
–2
Table 5.5 Relaxation table
Initial guess
T1
R1
T2
R2
T3
250
125
200
125
175
R3
T4
25
R4
225
25
Block change = + 25
275
75
225
75
200
–25
250
–25
DT1 = + 25
300
–25
225
100
200
–25
250
0
DT2 = + 25
300
0
250
0
200
0
250
Check Solution
0 300
0 250
0 200
0 0
250
300
5 SteadyState TwoDimensional Heat Conduction h, T∞ m
m1 x
Fig. 5.30 Example 5.32
Solution The rate of the heat conducted between nodes (m–1) and m is kAc
ðTm1 Tm Þ Dx
ðiÞ
The rate of heat convected is the sum of the heat convected from the end and the lateral surface of the volume: hAc ðT1 Tm Þ þ h
PDx ðT1 Tm Þ 2
ðiiÞ
In the steady state, summation of Eqs. (i) and (ii) must be zero, i.e. ðTm1 Tm Þ PDx þ hAc ðT1 Tm Þ þ h kAc ðT1 Tm Þ Dx 2 Rearranging the terms, we get the nodal equation as " Tm
# " # Dx hPðDxÞ2 hDx hPðDxÞ2 þ 1þh þ T1 Tm1 ¼ 0 k k 2kAc 2kAc
The residual equation can be written as " Tm
# " # Dx hPðDxÞ2 hDx hPðDxÞ2 þ 1þh þ T1 Tm1 ¼ Rm k k 2kAc 2kAc
Example 5.33 Write down the residual equation for the insulated corner section shown in Fig. 5.31.
(m  1, n)
(m, n) (m, n  1)
Fig. 5.31 Example 5.33
5.7 Numerical Solution Methods
301
Solution The heat flow rates from the nodes (m, n–1) and (m–1, n) to the node (m, n) are Dx:1 Tm;n1 Tm;n k 2 Dy Dy:1 Tm1;n Tm;n k 2 Dx
ð1Þ ð2Þ
The nodal equation is the summation of the above equations and for Dx = Dy, the simpliﬁcation gives Tm;n1 þ Tm1;n 2Tm;n ¼ 0 Hence, the residual equation is Tm;n1 þ Tm1;n 2Tm;n ¼ Rm;n Example 5.34 A squaresection ﬁn (5 5 mm2), see Fig. 5.32, is attached to a wall at 200° C. The surface of the ﬁn is exposed to air at 20°C. The convective heat transfer coefﬁcient is 100 W/(m2 K). The ﬁn is made of stainless steel with thermal conductivity of 20 W/(m K). Set up the residual equations in terms of temperature excess h = T – T∞ and determine the heat transfer rate. Length of the ﬁn is 25 mm. Solution Node 1 (at the base):
ð h1 h2 Þ PDx qb ¼ kAc þh h1 Dx 2
where Ac = 25 10–6 m2, Dx = 5 10–3 m, P = 20 10–3 m, and h1 = 180°C. Hence, ð180 h2 Þ 20 103 5 103 qb ¼ 20 25 10 þ 100 180 5 103 2
6
¼ 18:9 0:1h2 The residual equation is R1 ¼ qb þ 0:1h2 18:9
qb
1
2
3
4
5
6
5 5 θ = T  T∞
Fig. 5.32 Example 5.34
ð1Þ
302
5 SteadyState TwoDimensional Heat Conduction
Node 2: kAc
ð h1 h2 Þ ð h3 h2 Þ þ kAc hðPDxÞh2 ¼ 0 Dx Dx
The substitution of various values gives the residual equation as R2 ¼ 180 þ h3 2:1h2 Similarly, the residual equations for the nodes 3 to 5 can be written as R3 ¼ h2 þ h4 2:1h3 R4 ¼ h3 þ h5 2:1h4 R5 ¼ h4 þ h6 2:1h5 Node 6: kAc
ðh5 h6 Þ PDx h þ A c h6 ¼ 0 Dx 2
Substitution gives R6 ¼ h5 1:075h6 The set of residual equations is R1 ¼ qb þ 0:1h2 18:9 R2 ¼ 180 þ h3 2:1h2 R3 ¼ h2 þ h4 2:1h3 R4 ¼ h3 þ h5 2:1h4 R5 ¼ h4 þ h6 2:1h5 R6 ¼ h5 1:075h6 The unit change table and relaxation tables are given as Tables 5.6 and 5.7, respectively. Since the residuals are reasonably low after step 8, the process can be terminated.
Table 5.6 Unit change DR2
DR3
DR4
DR5
DR6
Dh2 = + 1
–2.1
+1
0
0
0
Dh3 = + 1
+1
–2.1
+1
0
0
Dh4 = + 1
0
+1
–2.1
+1
0
Dh5 = + 1
0
0
+1
–2.1
+1
Dh6 = + 1
0
0
0
+1
–1.075
Block change = + 1
–1.1
–0.1
–0.1
–0.1
–0.075
5.7 Numerical Solution Methods
303
Table 5.7 Relaxation table h2
R2
h
R3
h4
R4
h5
R5
h6
R6
Initial guess
130
7
100
–5
75
7.5
65
– 1.5
60
0.5
Dh2 = + 5
135
– 3.5
100
0
75
7.5
65
– 1.5
60
0.5
Dh3 = + 3.5
135
0
103.5
– 7.35
75
7.5
65
– 1.5
60
0.5
Dh4 = + 7
135
0
103.5
– 0.35
82
3.7
65
5.5
60
0.5
Dh5 = + 3
135
0
103.5
– 0.35
82
0.7
68
– 0.8
60
3.5
Dh6 = + 3
135
0
103.5
– 0.35
82
0.7
68
2.2
63
0.275
Dh5 = + 1
135
0
103.5
– 0.35
82
0.3
69
0.1
63
1.275
Dh6 = + 0.5
135
0
103.5
– 0.35
82
0.3
69
0.6
63.5
0.7375
Approximate solution
135
3
103.5
82
69
63.5
Now from Eq. (1), qb ¼ 18:9 0:1 135 ¼ 5:4 W: Check: From the equation of the ﬁn, qb ¼ where m ¼
qﬃﬃﬃﬃﬃ hP kAc
¼
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðTs T1 Þ tanhðmLc Þ
qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 3 100ð2010 Þ 20ð25106 Þ
¼ 63:25
Ac 25 106 ¼ 25 103 þ ¼ 0:02625 m P 20 103 tanh mLc ¼ tanhð63:25 0:02625Þ ¼ 0:9298
Lc ¼ L þ
Hence, qb ¼
qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 100 ð20 103 Þ 20 ð25 106 Þ 180 0:9298 ¼ 5:29 W:
Thus the solution is in reasonable agreement. Example 5.35 Figure 5.33 shows a rectangular section ﬁn extending from the wall of a furnace. It is exposed to a convection environment (h = 400 W/(m2 K) and T∞ = 25°C). The ﬁn is very wide in the zdirection. Calculate the steadystate temperatures of the nodes shown in the ﬁgure. The thermal conductivity of the material is 4 W/(m K).
304
5 SteadyState TwoDimensional Heat Conduction
20 mm
T∞ = 25 oC 2
1
Ts = 200 oC
4
6
8
3
5
7
10 mm 10 mm
80 mm = 20, = 10
Fig. 5.33 Example 5.35
Solution Due to the symmetry about the horizontal centreline, there are only 8 different nodal conditions. It is to note that Dx 6¼ Dy in this problem. Interior nodes (1, 3 and 5): At node 1, the heat balance gives kAy
ð T s T1 Þ ðT3 T1 Þ ðT2 T1 Þ þ kAy þ 2kAx ¼0 Dx Dx Dy
where Ax ¼ ð20 1Þ 103 m2 for depth ¼ 1m Ay ¼ ð10 1Þ 103 m2 Dx ¼ 20 103 m Dy ¼ 10 103 m Substitution of various values gives k 10
ðTs T1 Þ ðT3 T1 Þ ðT2 T1 Þ þ k 10 þ 2k 20 ¼0 20 20 10
or ðTs T1 Þ ðT3 T1 Þ þ þ 4 ð T 2 T1 Þ ¼ 0 2 2 or 10T1 8T2 T3 Ts ¼ 0 So the nodal equation is 10T1 8T2 T3 200 ¼ 0
ðiÞ
5.7 Numerical Solution Methods
305
Similarly the equation for the nodes 3 and 5 can be written as 10T3 8T4 T5 T1 ¼ 0
ðiiÞ
10T5 8T6 T7 T3 ¼ 0
ðiiiÞ
Nodes 2, 4, and 6: These nodes are on the convective boundary. At node 2, the heat balance equation is kAx
ðT1 T2 Þ Ay ðT4 T2 Þ Ay ðTs T2 Þ þk þk þ hðDx:1ÞðT1 T2 Þ ¼ 0 Dy Dx Dx 2 2
Putting vales of Dx, Dy, Ax, Ay, k and h, we obtain ðT1 T2 Þ ðT4 T2 Þ þ 4 10 103 3 10 10 2 20 103 ð Þ T T s 2 þ 4 10 103 þ 400 20 103 ðT1 T2 Þ ¼ 0 2 20 103
4 20 103
or 8ðT1 T2 Þ þ ðT4 T2 Þ þ ðTs T2 Þ þ 8ðT1 T2 Þ ¼ 0 or 18T2 8T1 T4 Ts 8T1 ¼ 0 Substituting values of Ts and T∞, the nodal equation is 18T2 8T1 T4 400 ¼ 0 Similarly, at node 4, we have 18T4 8T3 T6 T2 200 ¼ 0 And at node 6, 18T6 8T5 T8 T4 200 ¼ 0 Node 8 (the corner node): k
Ay ðT6 T8 Þ Ax ðT7 T8 Þ ðDx þ DyÞ:1 þk þh ð T 1 T8 Þ ¼ 0 Dx Dy 2 2 2
or ðT6 T8 Þ ðT7 T8 Þ þ 4 20 103 þ 400 2 20 103 2 10 103 3 ð20 þ 10Þ 10 ðT1 T8 Þ 2 ¼0
4 10 103
306
5 SteadyState TwoDimensional Heat Conduction
or ð T 6 T 8 Þ þ 4 ð T7 T 8 Þ þ 6 ð T 1 T 8 Þ ¼ 0 or 11T8 4T7 T6 150 ¼ 0 Node 7 at the ﬁn end: kAy
ðT5 T7 Þ Ax ðT8 T7 Þ þ 2k þ hðDy:1ÞðT1 T7 Þ ¼ 0 Dx Dy 2
Substitution of various values gives 7T7 4T8 T5 50 ¼ 0 The complete set of nodal equations is Node 1 :
10T1 8T2 T3 200 ¼ 0
ðiÞ
Node 2 :
18T2 8T1 T4 400 ¼ 0
ðiiÞ
Node 3 :
10T3 8T4 T5 T1 ¼ 0
ðiiiÞ
Node 4 :
18T4 8T3 T6 T2 200 ¼ 0
ðivÞ
Node 5 :
10T5 8T6 T7 T3 ¼ 0
ðvÞ
Node 6 :
18T6 8T5 T8 T4 200 ¼ 0
ðviÞ
Node 7 :
7T7 4T8 T5 50 ¼ 0
ðviiÞ
Node 8 :
11T8 4T7 T6 150 ¼ 0
ðviiiÞ
The above set of equations can be solved by Gaussian elimination or relaxation method. The relaxation method yields T1 = 65.91°C, T2 = T9 = 53.2°C, T3 = 33.69°C, T4 = T10 = 30.5°C, T5 = 26.77°C, T6 = T11 = 26.1°C, T7 = 25.36°C, T8 = T12 = 25.22°C with the residuals as R1 ¼ 0:19; R2 ¼ 0:18; R3 ¼ 0:22; R4 ¼ 0:18; R5 ¼ 0:15; R6 ¼ 0:08; R7 ¼ 0:13; R8 ¼ 0:12: Example 5.36 From the temperatures determined in Example 5.35, determine the heat flow rate from the ﬁn base. Solution From Example 5.35, h = 400 W/(m2 K), k = 4 W/(m K), T∞ = 25°C, Ts = 200°C, T1 = 65.91°C, T2 = T9 = 53.2°C. The heat flow rate from the ﬁn base, refer Fig. 5.34, is
5.7 Numerical Solution Methods
307 h, T∞ qc
2 q2 q1
Ts
Δy
1
Δy
q9 qc’ Δx
9
= 20 mm = 10 mm.
Fig. 5.34 Example 5.36
qfin ¼ qc þ q1 þ q2 þ q9 þ qc0 ; where Dx 20 :1 ðTs T1 Þ ¼ 400 1 ð200 25Þ ¼ 700 W=m q c ¼ q c0 ¼ h 2 2000 ðTs T1 Þ 10 ð200 65:91Þ 1000 q1 ¼ kðDy:1Þ ¼4 1 ¼ 268:18 W=m Dx 1000 20 Dy ðTs T2 Þ 10 ð200 53:2Þ 1000 :1 ¼4 1 ¼ 146:8 W=m q2 ¼ k 2 Dx 2000 20 Dy ðTs T9 Þ 10 ð200 53:2Þ 1000 :1 ¼4 1 ¼ 146:8 W=m: q9 ¼ k 2 Dx 2000 20 Hence, qfin ¼ 700 þ 268:18 þ 146:8 þ 146:8 þ 700 ¼ 1961:78 W=m:
Example 5.37 Display the nodal equations of Example 5.35 in matrix form. Solution The matrix is as given below. 10 8 1
8 18 1
1 10 8 1
1 8 18 1
1 10 8 1
1 8 18 1
1 7 4
T1 200 T2 400 T3 0 T4 200 ¼ T5 0 1 T6 200 4 T7 50 11 T8 150
308
5 SteadyState TwoDimensional Heat Conduction T = 200oC A
T = 300oC
D
B 1
2
4
3
T = 100oC
C T = 400oC
Fig. 5.35 Example 5.38
5.7.2.2 Gaussian Elimination The relaxation method involves trial and error, hence it cannot be applied to the solution of a set of a large number of equations. Such a set of nodal equations can be solved by Gaussian elimination method. This method is systematic, hence easy to understand and apply. It is being explained below with the help of an example. Now the computers have made it possible to solve practical problems, which were earlier believed to be not solvable due to a large number of nodal equations involved in such problems. Example 5.38 Write the nodal equations at the four interior nodal points of the square grid of the section ABCD of a bar shown in Fig. 5.35. The temperature at the boundaries AB, BC, CD and DA are indicated in the ﬁgure. Solution The nodal equations obtained are Node 1 :
200 þ 300 þ T2 þ T4 4T1 ¼ 0
ðaÞ
Node 2 :
200 þ 100 þ T3 þ T1 4T2 ¼ 0
ðbÞ
Node 3 :
100 þ 400 þ T4 þ T2 4T3 ¼ 0
ðcÞ
Node 4 :
300 þ 400 þ T1 þ T3 4T4 ¼ 0
ðdÞ
Example 5.39 Explain and use the Gaussian elimination method to solve the nodal equations of the previous example. Solution The nodal equations are 500 þ T2 þ T4 4T1 ¼ 0
ðaÞ
300 þ T3 þ T1 4T2 ¼ 0
ðbÞ
5.7 Numerical Solution Methods
309
500 þ T4 þ T2 4T3 ¼ 0
ðcÞ
700 þ T1 þ T3 4T4 ¼ 0
ðdÞ
The equations may be rewritten as
(i) The Gaussian elimination is to ﬁrst triangularize the given set of equations. This is accomplished by some basic operations as explained below. Eliminate T1 from Eqs. (f) and (h) by multiplying these equations by 4 and adding Eq. (e) to them. This operation gives
Next, eliminate T2 from Eqs. (k) and (l) by multiplying them by 15 and adding Eq. (j) to them. The result of this operation is
Rewriting the equations in simpliﬁed form, we obtain
310
5 SteadyState TwoDimensional Heat Conduction
Now eliminate T3 from Eq. (p’) by adding to it 2/7 times of Eq. (o’), to obtain
This is the triangularized set of equations. (ii) From Eq. (p”), T4 = 300°C. Now back substitute beginning with the bottom equation and working upwards. The process gives T3 ¼ 250 C; T2 ¼ 200 C; and T1 ¼ 250 C:
Note: When the number of equations in a set is large, the hand calculation may be a timeconsuming process. For this purpose, a standard computer program is available. The given set of equations can be put in a matrix form as given below.
þ4 1 0 1
1 þ4 1 0
0 1 þ4 1
1 T1 500 0 T2 300 ¼ 1 T3 500 þ 4 T4 700
In general, the matrix is A(1,1) A(1,2) A(2,1) A(2,2) ... ... A(N,1) A(N,2)
... ... ... ...
A(1,N) Tð1Þ B(1) A(2,N) Tð2Þ B(2) ¼ ... . . . ... A(N,N) TðNÞ B(N)
2. The addition of the nodal equations of this example yields T1 þ T2 þ T3 þ T4 ¼ 1000: This equals the sum of the four boundary temperatures. This result is exact for any four points that are symmetrically placed about the centre of a square. Example 5.40 Combustion products flow through a chimney whose crosssection is shown in Fig. 5.36. If at any location, the inside surface temperature is 600°C and outside surface temperature is 100°C, determine the heat loss rate from the chimney for its unit length. Solution Refer Fig. 5.36b. Node 1 The heat balance equation for the twodimensional conduction system gives
5.7 Numerical Solution Methods
311
100 100 o o k = 1.3 W/(m K)
Δy
o1
2o
o Δx o 600 600
200 Sq
100 100 o o 3o o 600
4
100 o
o Δx = Δy = 100 mm
b One eight section (enlarged) 400 Sq
a Chimney crosssection Fig. 5.36 Example 5.40
Dx 100 T1 Dx 600 T1 T 2 T1 :1 :1 þk þ kðDy:1Þ ¼ 0: k 2 2 Dy Dy Dx Simpliﬁcation gives the nodal equation for Δx = Δy as 2T2 þ 700 4T1 ¼ 0:
ðiÞ
Node 2 The heat balance equation for the twodimensional conduction system gives kðDx:1Þ
100 T2 600 T2 T1 T2 T3 T2 þ kðDx:1Þ þ kðDy:1Þ þ kðDy:1Þ ¼ 0: Dy Dy Dx Dx
Simpliﬁcation gives the nodal equation as T1 þ T3 þ 700 4T2 ¼ 0:
ðiiÞ
Node 3 The equation is similar to node 2 and is T2 þ T4 þ 700 4T3 ¼ 0:
ðiiiÞ
Node 4 The heat balance equation for the twodimensional conduction system gives kðDx:1Þ
100 T4 T3 T4 þ kðDy:1Þ ¼ 0: Dy Dx
Simpliﬁcation gives the nodal equation as T3 þ 100 2T4 ¼ 0:
ðivÞ
312
5 SteadyState TwoDimensional Heat Conduction
Rewriting the nodal equations as
The equation can be solved by the Gaussian elimination method. We eliminate T1 from Eq. (b) by multiplying this equation by –4 and adding Eq. (a) to it. This operation gives
Next eliminate T2 from Eq. (g) by multiplying this equation by –14 and adding Eq. (f) to it. The result of this operation is
Rewriting the equations in simpliﬁed form, we obtain
Now eliminate T3 from Eq. (p) by multiplying this equation by  26 and adding Eq. (o) to it. This gives
5.7 Numerical Solution Methods
313
This is the triangularized set of equations. From Eq. (t), T4 = 205.55°C. Now back substitute beginning with Eq. (s) and working upwards. The process gives T3 ¼ 311:11 C; T2 ¼ 338:88 C; and T1 ¼ 344:44 C: The heat leaving the outer surface of this chimney section is q1=8 ¼ k
Dx T1 100 T2 100 T3 100 T4 100 :1 þ kðDx:1Þ þ kðDx:1Þ þ kðDx:1Þ 2 Dy Dy Dy Dy
¼ kð0:5T1 þ T2 þ T3 þ T4 350Þ ¼ 1:3 ð0:5 344:44 þ 338:88 þ 311:11 þ 205:55 350Þ ¼ 881:1 W/m length: The heat flow rate can also be calculated from the heat entering the chimney section, which is q1=8
Dx 600 T1 600 T2 600 T3 :1 þ kðDx:1Þ þ kðDx:1Þ ¼k 2 Dy Dy Dy ¼ kð1500 0:5T1 T2 T3 Þ ¼ 1:3 ð1500 0:5 344:44 338:88 311:11Þ ¼ 881:1 W=m length:
Since the considered chimney section is 1/8 of the total crosssection, the heat loss from the chimney is q ¼ 881:1 8 ¼ 7048:8 W=m length: Example 5.41 If in the above problem inner and outer surfaces of the chimney are subjected to convective conditions write the nodal equations. Solution Refer Fig. 5.37b. Node 1 The heat balance equation for the twodimensional conduction system gives k
Dx T5 T1 Dx T8 T1 T2 T1 :1 :1 þk þ kðDy:1Þ ¼ 0: 2 2 Dy Dy Dx
Simpliﬁcation gives the nodal equation for Δx = Δy as 2T2 þ T5 þ T8 4T1 ¼ 0 Node 2 The heat balance equation for the twodimensional conduction system gives
ðiÞ
314
5 SteadyState TwoDimensional Heat Conduction
ho, T∞o k = 1.3 W/(m K) 200 Sq
Δy
8 o
9 o
10 o
11 o
o1
2o
3o
4
o Δx o 6 5
o 7
12 o
o Δx = Δy
hi, T∞i
400 Sq
a Chimney crosssection
b One eight section (enlarged)
Fig. 5.37 Example 5.41
kðDx:1Þ
T6 T2 T9 T2 T1 T2 T3 T2 þ kðDx:1Þ þ kðDy:1Þ þ kðDy:1Þ ¼ 0: Dy Dy Dx Dx
Simpliﬁcation gives the nodal equation as T1 þ T3 þ T6 þ T9 4T2 ¼ 0:
ðiiÞ
Node 3 The equation is similar to node 2 and is T2 þ T4 þ T7 þ T10 4T3 ¼ 0:
ðiiiÞ
Node 4 The heat balance equation for the twodimensional conduction system gives kðDx:1Þ
T11 T4 T 3 T4 þ kðDy:1Þ ¼ 0: Dy Dx
Simpliﬁcation gives the nodal equation as T3 þ T11 2T4 ¼ 0:
ðivÞ
Node 5 The heat balance equation for the twodimensional conduction system gives
Dx T1 T5 Dy T6 T5 Dx :1 :1 :1 ðT1i T5 Þ ¼ 0: k þk þ hi 2 2 2 Dy Dx Simpliﬁcation gives the nodal equation for Δx = Δy as hi Dx hi Dx T1i 2 þ T1 þ T 6 þ T5 ¼ 0: k k
ðvÞ
5.7 Numerical Solution Methods
315
Node 6 The heat balance equation for the twodimensional conduction system gives T 2 T6 Dy T5 T6 Dy T7 T6 :1 :1 kðDx:1Þ þk þk þ hi ðDx:1ÞðT1i T6 Þ ¼ 0: 2 2 Dy Dx Dx Simpliﬁcation gives the nodal equation for Δx = Δy as 2hi Dx hi Dx T1i 2 þ 2 T6 ¼ 0: 2T2 þ T5 þ T7 þ k k
ðviÞ
Node 7 The heat balance equation for the twodimensional conduction system gives kðDx:1Þ
T 3 T7 Dy T6 T7 Dx :1 :1 ðT1i T7 Þ ¼ 0: þk þ hi 2 2 Dy Dx
Simpliﬁcation gives the nodal equation for Δx = Δy as 2T3 þ T6 þ
hi Dx hi Dx T1i 3 þ T7 ¼ 0: k k
ðviiÞ
Node 8 The heat balance equation for the twodimensional conduction system gives
Dx T1 T8 Dy T9 T8 Dx k :1 :1 :1 ðT1o T8 Þ ¼ 0: þk þ h0 2 2 2 Dy Dx Simpliﬁcation gives the nodal equation for Δx = Δy as ho Dx ho Dx T1o 2 þ T1 þ T9 þ T8 ¼ 0: k k
ðviiiÞ
Node 9 The heat balance equation for the twodimensional conduction system gives T2 T9 Dy T8 T9 Dy T10 T9 kðDx:1Þ :1 :1 þk þk þ ho ðDx:1ÞðT1o T9 Þ ¼ 0: 2 2 Dy Dx Dx Simpliﬁcation gives the nodal equation for Δx = Δy as 2T2 þ T8 þ T10 þ
2ho Dx ho Dx T1o 2 þ 2 T9 ¼ 0: k k
ðixÞ
316
5 SteadyState TwoDimensional Heat Conduction
Node 10 The heat balance equation for the twodimensional conduction system gives T3 T10 Dy T9 T10 Dy T11 T10 :1 :1 kðDx:1Þ þk þk þ ho ðDx:1ÞðT1o T10 Þ ¼ 0: 2 2 Dy Dx Dx Simpliﬁcation gives the nodal equation for Δx = Δy as 2T3 þ T9 þ T11 þ
2ho Dx ho Dx T1o 2 þ 2 T10 ¼ 0: k k
ðxÞ
Node 11 The heat balance equation for the twodimensional conduction system gives kðDx:1Þ ¼ 0:
T4 T11 Dy T10 T11 Dy T12 T11 :1 :1 þk þk þ ho ðDx:1ÞðT1o T11 Þ 2 2 Dy Dx Dx
Simpliﬁcation gives the nodal equation for Δx = Δy as 2ho Dx ho Dx T1o 2 þ 2 T11 ¼ 0: 2T4 þ T10 þ T12 þ k k
ðxiÞ
Node 12 The heat balance equation for the twodimensional conduction system gives k
Dy T11 T12 Dx :1 :1 ðT1o T12 Þ ¼ 0: þ ho 2 2 Dx
Simpliﬁcation gives the nodal equation for Δx = Δy as T11 þ
ho Dx ho Dx T1o þ 1 T12 ¼ 0: k k
ðxiiÞ
5.7.2.3 The Gauss–Seidel Iteration Method The iterative method may yield quicker results when the number of nodal equations is large. One of such methods, the Gauss–Seidel method, is being explained below using the nodal equations of Example 5.35. Step 1: Rearrange the nodal equations so that the unknown temperature of the node for which the nodal equation was written is on the left as follows.
5.7 Numerical Solution Methods
317
Table 5.8 The Gauss–Seidel iteration table
a
na
T1
T2
T3
T4
T5
T6
T7
T8
Initial guess
130
80
90
55
50
40
35
30
1
93
65.6
58.3
42.9
41.33
33.53
30.19
27.68
2
78.31
59.41
46.28
35.84
34.47
29.96
27.88
26.50
3
72.16
56.28
39.33
33.38
30.69
28.08
26.67
25.89
4
68.96
54.73
36.67
32.01
28.80
27.13
26.06
25.58
5
67.45
53.98
35.23
31.28
27.83
26.64
25.74
25.42
6
66.71
53.61
34.48
30.89
27.33
26.39
25.57
25.33
7
66.34
53.42
34.08
30.69
27.08
26.26
25.49
25.29
8
66.14
53.32
33.87
30.59
26.94
26.19
25.44
25.27
9
66.04
53.27
33.77
30.53
26.87
26.15
25.42
25.26
10
65.99
53.25
33.71
30.5
26.83
26.13
25.41
25.25
n indicates the number of iterations
T1 ¼ ð1=10Þð8T2 þ T3 þ 200Þ T2 ¼ ð1=18Þð8T1 þ T4 þ 400Þ T 3 ¼ ð1=10Þð8T4 þ T5 þ T1 Þ T 4 ¼ ð1=18Þð8T3 þ T6 þ T2 þ 200Þ T 5 ¼ ð1=10Þð8T6 þ T7 þ T3 Þ T 6 ¼ ð1=18Þð8T5 þ T8 þ T4 þ 200Þ T 7 ¼ ð1=7Þð4T8 þ T5 þ 50Þ T 8 ¼ ð1=11Þð4T7 þ T6 þ 150Þ Step 2: Assume an initial set of values for the temperatures T1 to T8 as shown in Table 5.8. Using each of the equations of T1 to T8, calculate new values of the temperatures. For example, the ﬁrst equation gives a new value of temperature T1 = 93°C. Now using the new value of T1 and still not updated value of T4 (= 55°C), obtain the new value of T2 (= 65.6°C). In this manner, sweep through the set of eight equations. This gives a new set of revised or updated estimates of T1 to T8 (refer row 3 of the table). This completes one cycle of iteration. In essence, it is the calculation of new values of nodal temperatures using the most recent values of the temperatures. Step 3: The process explained in Step 2 is repeated until the calculated nodal temperatures converge. This is assumed to be achieved when successive temperatures differ by very small values, i.e. when
ðTi Þn þ 1 ðTi Þn d
for all Ti
where (Ti)n +1 is the calculated value of the temperature Ti after (n +1) iterations and (Ti)n is the temperature Ti after nth iteration. Value of d depends on the desired accuracy of the result.
318
5 SteadyState TwoDimensional Heat Conduction
It is to note that with the increase in the number of iterations, both the accuracy and the cost (time) of the computation increase. The decision regarding the number of the iterations is the tradeoff between the two. After 10th iteration, ðTi Þ10 ðTi Þ9 ð0:06 CÞ
for all Ti
Comparison of the results of last iteration with those obtained with the relaxation method shows that they differ by a very small amount. Example 5.42 Solve Example 5.9 using the ﬁnitedifference method. Solution A grid with Dx = Dy = 0.5 is being used, see Fig. 5.38. The temperature distribution will be symmetrical about the horizontal axis. The nodal equations are Node 1 :
4T 1 þ 2T2 þ T3 þ 400 ¼ 0
ðiÞ
Node 2 :
4T2 þ T1 þ T4 þ 500 ¼ 0
ðiiÞ
Node 3 :
4T 3 þ 2T4 þ T1 þ T5 ¼ 0
ðiiiÞ
Node 4 :
4T4 þ T3 þ T2 þ T6 þ 100 ¼ 0
ðivÞ
Node 5 :
4T5 þ T3 þ 2T6 þ 100 ¼ 0
ðvÞ
Node 6 :
4T6 þ T5 þ T4 þ 200 ¼ 0:
ðviÞ
The solution of the equations gives T1 = 258.03°C, T2 = 228.57°C, T3 = 175°C, T4 = 156.25°C, T5 = 129.47°C and T6 = 121.43°C. Note: 1. A ﬁner grid will give a better approximation. 2. From the mean value theorem, T3 = [(400 + 100 + 100 + 100)/4] = 175°C.
T = 100oC 2 T = 100oC 1
3
5
y
2
4
6
0
x
T = 400oC
Fig. 5.38 Example 5.42
2 T = 100oC
x = y = 0.5
5.8 Two Dimensional SteadyState Heat Conduction with Heat Generation
5.8
319
TwoDimensional SteadyState Heat Conduction with Heat Generation
(a) Internal node, refer Fig. 5.39a, In the steady state, the summation of heat flow to the node (m, n) from the surrounding nodes is zero.
Tm1;n Tm;n Tm þ 1;n Tm;n Tm;n1 Tm;n þ kðDy:1Þ þ kðDx:1Þ Dx Dx Dy Tm;n þ 1 Tm;n þ qg ðDx:Dy:1Þ ¼ 0 þ kðDx:1Þ Dy
kðDy:1Þ
for Dz = 1. Putting Dx = Dy, and on simpliﬁcation, we obtain Tm;n ¼
1 ðDxÞ2 Tm1;n þ Tm þ 1;n þ Tm;n1 þ Tm;n þ 1 þ qg 4 4k
(m, n 1)
(m 1, n)
(m +1, n)
(m, n)
Δy Δx
(m, n +1)
a (m, n 1) Δy
h, T∞
(m, n)
(m 1, n)
h, T∞
(m 1, n)
(m, n) Δy (m, n +1)
(m, n +1)
Δx
Δx
b
c
Fig. 5.39 Twodimensional steadystate heat conduction with heat generation
320
5 SteadyState TwoDimensional Heat Conduction
(b) Surface node, refer Fig. 5.39(b), Summation of the heat conduction and convection into the node gives
Tm1;n Tm;n Dx:1 Tm;n1 Tm;n Dx:1 Tm;n þ 1 Tm;n þk þk kðDy:1Þ 2 2 Dx Dy Dy ðDx:Dy:1Þ ¼0 þ hðDy:1Þ T1 Tm;n þ qg 2 Simpliﬁcation gives
Dx Dx ðDxÞ2 Tm;n1 þ Tm;n þ 1 þ 2Tm1;n 2 2 þ h ¼0 Tm;n þ 2h T 1 þ qg k k k
for Dx = Dy. (c) External corner node Heat balance equation in this case is ðDy:1Þ Tm1;n Tm;n ðDx:1Þ Tm;n þ 1 Tm;n ðDx þ DyÞ:1 k T1 Tm;n þk þh 2 2 2 Dx Dy Dx Dy : :1 ¼ 0 þ qg 2 2 For Dx = Dy, simpliﬁcation gives Dx Dx ðDxÞ2 Tm1;n þ Tm;n þ 1 2 1 þ h ¼0 Tm;n þ 2h T1 þ q g k k 2k Example 5.43 Write down nodal equations for an interior node and a surface node with convection in a cylindrical system for onedimensional radial heat conduction. Solution In Fig. 5.40, the nodes are spaced at equal radial distance Dr with node m = 0 at the centre. Node m, situated at a radial distance mDr, corresponds to annular ring of radial width Dr and radius mDr. Similarly we can deﬁne the area corresponding to other nodal points. Node n is surface node. The heat balance equation for node m can be written for unit length of cylinder as Dr Tm1 Tm Dr Tm þ 1 Tm k 2p mDr þ 2p mDr þ ¼0 :1 :1 2 2 Dr Dr
5.8 Two Dimensional SteadyState Heat Conduction with Heat Generation
Fig. 5.40 Example 5.43
or
1 Tm1 Tm 1 Tm þ 1 Tm þ 2pDr m þ ¼0 :1 k 2pDr m :1 2 2 Dr Dr Simpliﬁcation gives 1 1 m Tm1 þ m þ Tm þ 1 2mTm ¼ 0 2 2 The heat balance equation for node n for unit length of cylinder is Dr Tn1 Tn k 2p nDr þ hð2pnDr:1ÞðT1 Tn Þ ¼ 0 :1 2 Dr
or 1 k n ðTn1 Tn Þ þ hnDrðT1 Tn Þ ¼ 0 2 or 1 hnDr 1 hnDr T1 n þ n Tn1 þ Tn ¼ 0 2 k 2 k
321
322
5 SteadyState TwoDimensional Heat Conduction
Example 5.44 If the radial heat conduction in Example 5.43 is accompanied with uniform volumetric heat generation, write the nodal equations for node m and n. Solution The heat balance equation is considering the heat generation Dr Tm1 Tm Dr Tm þ 1 Tm k 2p mDr þ 2p mDr þ þ qg DV ¼ 0 :1 :1 2 2 Dr Dr where DV = 2pmDr. Dr .1. Substitution gives
Dr Tm1 Tm Dr Tm þ 1 Tm k 2p mDr þ 2p mDr þ þ qg ð2pmDr:Dr:1Þ :1 :1 2 2 Dr Dr ¼0 or 1 1 qg mðDrÞ2 ¼0 m :ðTm1 Tm Þ þ m þ :ðTm þ 1 Tm Þ þ 2 2 k
or m
1 1 qg mðDrÞ2 ¼0 Tm1 þ m þ Tm þ 1 2mTm þ þ 2 2 k
The heat balance equation for node n for unit length of cylinder considering heat generation is Dr Tn1 Tn k 2p nDr þ hð2pnDr:1ÞðT1 Tn Þ þ qg DV ¼ 0 :1 2 Dr where DV = 2pnDr.(Dr/2) .1 for surface node. Substitution gives Dr Tn1 Tn k 2p nDr þ hð2pnDr:1ÞðT1 Tn Þ þ qg ½2pnDr:ðDr=2Þ:1 ¼ 0 :1 2 Dr or
1 hnDr qg nðDrÞ2 k n ðTn1 Tn Þ þ ¼0 ð T1 T n Þ þ 2 k 2k
or n
1 hnDr 1 hnDr qg nðDrÞ2 T1 n þ ¼0 Tn1 þ Tn þ 2 k 2 k 2k
5.9 Summary
5.9
323
Summary
In this chapter, analytical solution, graphical analysis, method of analogy and numerical solutions have been presented for the study of problems of twodimensional steadystate conduction heat flow through solids without heat sources. The Laplace equation that governs the temperature distribution for twodimensional heat conduction system is ∂2t/∂x2 + ∂2t/∂y2 = 0. Since to determine the conduction heat flow, the temperature ﬁeld must be known, the problem reduces to the solution of the Laplace equation. The common techniques used for the solution are: (i) analytical, (ii) graphical, (iii) method of analogy, and (iv) numerical methods. In Sect. 5.2, the analytical solution for a rectangular section bar (W H), which is very long in the zdirection, has been presented whose three lateral sides (x = 0 and W, and y = 0) are maintained at a constant temperature To and for the fourth side we considered three different boundary conditions: (i) the fourth side (y = H) has a sinusoidal temperature distribution, T = Tm sin(px/W), imposed on it, (ii) temperature along the fourth side is given by an arbitrary function f(x), and (iii) the fourth side is at a constant temperature Tb. In Sect. 5.3, mathematical treatment of the problem of conduction through a flat semiinﬁnite homogeneous plate has been presented, whose length in ydirection is inﬁnite, it is relatively thin in the zdirection and xoy surfaces are insulated so that there is no temperature gradient in the zdirection and the temperature ﬁeld is two dimensional. In Sect. 5.5, the basic concept of the graphical method of thermal flux plotting has been explained and conduction shape factor has been deﬁned. The shape factor presented in Table 5.2 can be readily used for the solution of various physical systems of interest as illustrated by solved problems. However, the graphical method involves drawing of curvilinear square elements and count of the number of heat flow channels and the temperature increments, which depends on the skill of the person plotting the lines. Electrical analogy, discussed in Sect. 5.6, may be used to sketch the isotherms and heat flow lines. It is to note that this technique is now not much used but rough sketching and the concept of the plots can help in a quick check of the result from other methods. Hence, it must not be totally overlooked. The direct integration of the differential equations has been used to solve simple problems of twodimensional, threedimensional and transient heat conduction problems but success in solving complex problems, involving nonlinear boundary conditions and temperature or positiondepending thermal properties, is limited. Such problems have been solved using numerical methods. Commonly used numerical methods are: ﬁnite difference and ﬁnite element methods. The ﬁnitedifference method has been used extensively in solving heat conduction problems because of its simplicity in implementation. The ﬁnite element method is being widely used to solve problems in structural mechanics. It requires much greater mathematical efforts and has been used for solving heat conduction problems involving complicated geometries. Here only the basic principles and application of the ﬁnitedifference method have been presented. The ﬁrst step in the ﬁnitedifference method is to discretize the spatial and time coordinates to form a mesh of nodes. Then by applying the energy balance to the volume elements surrounding the nodes, a set of linear algebraic equations (termed as nodal equations) is obtained. Various techniques, namely the relaxation method, Gaussian elimination, and the Gauss–Seidel iteration methods for the solution of nodal equations developed by the ﬁnitedifference method have been presented.
324
5 SteadyState TwoDimensional Heat Conduction
It is to note that the accuracy of the ﬁnitedifference approximation increases with the number of nodes hence computers are used to obtain ﬁnitedifference solutions. Standard computer programs are available for this purpose.
Review Questions 5:1 What are the various methods of solving twodimensional heat conduction problems? Discuss their basics and limitations. 5:2 What are curvilinear squares? 5:3 Deﬁne conduction shape factor. Support your answer by considering a conﬁguration where twodimensional heat conduction is encountered. 5:4 Explain the basic principle of electrothermal analogy? 5:5 Explain the basic method of ﬁnite difference. What are the various methods available to obtain a solution after writing the ﬁnitedifference equations? 5:6 Compare relaxation and Gaussian elimination methods.
Problems 5:1 Using the equations of temperature distribution obtained in the analytical method for the twodimensional heat conduction, determine the temperature at (3/4, ¼) of the long squaresection rod shown in Fig. 5.41. [Ans. Using the equations presented in Example 5.5, t(x, y) = 127.2°C.] 5:2 A cubical furnace (0.6 m side) is covered with 0.1 m thick layer of insulation [k = 0.035 W/(m K)]. If the temperature difference across the insulation is 80°C, calculate the heat loss through the layer. [Ans. Shape factors: Splane wall = 3.6, Sedge = 0.324, Scorner = 0.015, Total shape factor, S = 25.61 m; q = 71.71 W] 5:3 Combustion gases at an average temperature of 1200°C flow through a 3 m long, 100 mm zerodimensional circular section duct. In order to reduce the heat loss, the duct is covered with insulation [k = 0.05 W/(m2 K)]. The insulated duct measures 250 mm 250 mm (square in shape). Determine the heat loss for the duct if the heat transfer coefﬁcients are hi = 150 W/(m2 K) and h0 = 5 W/(m2 K) for the duct inner and outer surfaces, respectively. T∞ = 30°C. [Ans. From Case (11), Table 5.2, S = 18.98; Resistances are: Ri = 1/(2pRLhi) = P 7.074 10–3, Rk = 1/kS = 1.054; and Ro = 1/(4WLho) = 0.0666, q = Dt/ R = 1037.5 W.]
1 400oC
200oC (3/4, 1/4) 100 oC
100oC
Fig. 5.41 Problem 5.1
1
5.9 Summary
325
h, T∞ Δx
Fig. 5.42 Problem 5.5
5:4 A 75 mm diameter hot fluid pipeline (surface temperature = 80°C) and a cold water pipeline 50 mm in diameter (surface temperature = 20°C) are 200 mm apart on centres in a large duct packed with glass wool insulation [k = 0.038 W/(m K)]. Calculate the heat transfer to the cold fluid for 5 m length of the pipes. [Ans. From Case (9), Table 5.2, S = 8.49, q = k S (80–20) = 19.36 W]. 5:5 For a plate ﬁn of uniform crosssection Ac along its length, show that the temperature for the tip node shown in Fig. 5.42 is given by Dx 1 Dx 1 h þ 1 þ ðmDxÞ2 Tm ¼ h þ ðmDxÞ2 T1 þ Tm1 k 2 k 2 where m2 = hP/kAc. [Hint: hAc ðT1 Tm Þ þ kAc Tm1DxTm þ P Dx 2 hðT1 Tm Þ ¼ 0. On simpliﬁcation and hP 2 putting kAc ¼ m , the result is obtained.] 5:6 If the end of the plate ﬁn of Problem 5.5 is insulated, show that the temperature for the tip node is given by "
# " # 2 2 ðDxÞ ðDxÞ Tm1 þ m2 : T1 1 þ m 2 : Tm ¼ 0 2 2 where m2 = hP/kAc. [Hint: Refer Fig. 5.43. Heat balance equation is kAc Tm1DxTm þ hP Dx 2 hP 2 ðT1 Tm Þ ¼ 0. On simpliﬁcation and putting kA ¼ m , the result is obtained.] c 5:7 The node (m, n) in Fig. 5.44 is situated on a boundary along which uniform heat flux q00 is speciﬁed. Show that in the steadystate, the node temperature is given by Tm;n ¼ ð1=4ÞðTm;n þ 1 þ Tm;n1 þ 2Tm1;n Þ þ ðq00 =2kÞDx Verify that the adiabatic boundary limit result deduced from the equation agrees with the corresponding result in Table 5.3.
h, T∞
m1 Δx Fig. 5.43 Problem 5.6
m
326
5 SteadyState TwoDimensional Heat Conduction
(m, n +1) Δy
q”
(m, n)
(m 1, n)
(m, n 1) Δx
Fig. 5.44 Problem 5.7
∆x
5 ∆y Insulated
Insulated
h, T∞ a
2
1
h, T∞
3
4
b
Fig. 5.45 Problem 5.8
[Ans. The heat balance equation is Tm1;n Tm;n Dx Tm;n þ 1 Tm;n Dx Tm;n1 Tm;n þk þ kðDyÞ þ q00 Dy k 2 2 Dy Dy Dx ¼0
Substitution of Dx = Dy gives the desired result. For adiabatic boundary, q00 ¼ 0] 5:8 Write down the nodal equation for internal corner of a twodimensional system with horizontal boundary insulated and vertical boundary subjected to convection heat transfer as shown in Fig. 5.45a. Verify that the adiabatic boundary limit result deduced from the equation agrees with the corresponding result in Table 5.3. [Ans. The heat balance equation for node 1 gives for unit depth, refer Fig. 5.45b,
1:Dy 1:Dy ðT2 T1 Þ 1:Dx ðT3 T1 Þ ð T 4 T1 Þ þk þ kð1:DyÞ ðT1 T1 Þ þ k 2 2 Dx 2 Dy Dx ðT5 T1 Þ þ kð1:DxÞ ¼0 Dy hDy Putting Dx = Dy, we have T2 þ T3 þ 2T4 þ 2T5 þ hDy 6 þ T 1 k k T1 ¼ 0: h
When vertical boundary is also adiabatic, put h = 0 and the result is T2 þ T3 þ 2T4 þ 2T5 6T1 ¼ 0, which is the same as for case (h) Example 5.28.]
5.9 Summary
327 45o 3 1
Δx 2
Δy
h, T∞
Fig. 5.46 Problem 5.10
5:9 Write down nodal equation for the twodimensional steadystate conduction problem of Case (g) shown in Fig. 5.26 of Example 5.28 when the vertical boundary is insulated. Consider the dimension perpendicular to the plane of paper as b. [Ans. Referring to the heat balance equation of Case (g) of Example 5.28, the convective heat transfer will reduce to half and the heat balance equation for node 1 will be ðT2 T1 Þ bd ðT3 T1 Þ h bd T Þ þ k bd ¼ 0. Simpliﬁcation gives 2 ðT1 d þ k 2 d hd1 2 T ¼ 0. When horizontal boundary is also insulated we get T2 þ T3 þ k T1 2 þ hd 1 k result of Case (d) by putting h = 0, i.e. T2 þ T3 2T1 ¼ 0:] 5:10 Write nodal equation for the nodal point 1 on an inclined surface of a system shown in Fig. 5.46. The inclined surface is subjected to a convective heat transfer. [Ans. The heat balance equation for unit depth is pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ T T1 1 kðDy:1Þ T2Dx For þh Dx2 þ Dy2 :1 ðT1 T1 Þ ¼ 0: þ kðDx:1Þ T3Dy Δx = Δy equation transforms to nodal equation as pﬃﬃﬃ hDx pﬃﬃﬃ hDx T2 þ T3 þ 2 k T1 2 þ 2 k T1 ¼ 0:] 5:11 Write nodal equation for the nodal point 1 on tip of the system shown in Fig. 5.47 The inclined surface is subjected to a convective heat transfer while the horizontal surface is insulated. [Ans. The heat balance equation for unit depth is
45o 2
1
h, T∞
Fig. 5.47 Problem 5.11
Δx
Δy
328
5 SteadyState TwoDimensional Heat Conduction
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 Dx2 þ Dy2 :1 ðT1 T1 Þ ¼ 0: For Δx = Δy equation trans2 pﬃﬃﬃ pﬃﬃﬃ hDx forms to nodal equation as T2 þ 2 hDx k T1 1 þ 2 k T1 ¼ 0:] 5:12 Use the relaxation method to solve the following set of equations.
k
Dy 2 :1
T2 T1 Dx
þh
500 þ T2 þ T4 4T1 ¼ R1
ðaÞ
300 þ T3 þ T1 4T2 ¼ R2
ðbÞ
500 þ T4 þ T2 4T3 ¼ R3
ðcÞ
700 þ T1 þ T3 4T4 ¼ R4
ðdÞ
[Ans. T1 = 250°C, T2 = 200°C, T3 = 250°C, T4 = 300°C.]
6
Unsteady or Transient Heat Conduction
6.1
Introduction
When a solid body is suddenly exposed to an environment at different temperatures, the equilibrium temperature condition in the body is established only after some time. The equilibrium condition is referred to as the steady state. We discussed the temperature distribution and heat transfer in the steady state in Chaps. 2–5. This chapter is devoted to the transient state of conduction, i.e. the heating or cooling of the bodies which takes place in the period before the equilibrium is established. In the transient or unsteadystate heat conduction, the temperature of the solid body varies with time as well as in the space. Such problems are of interest because they are encountered in many industrial processes. The practical problems of the transient heat conduction can be divided into two groups: (i) when the body tends to thermal equilibrium, and (ii) when the temperature of the body is subjected to periodic variation.
6.2
Lumped Heat Capacity Analysis
Consider a body which may be regarded as having uniform temperature throughout at any instant. The assumption of uniform temperature throughout is approximately valid for bodies with a very high thermal conductivity combined with a low value of the convective heat transfer coefﬁcient. Smaller bodies with lower values of the thermal conductivity may also satisfy this condition. Mathematically, we shall assume that the thermal conductivity of the body is inﬁnite. This analysis is called lumped heat capacity method. From the heat balance for the small body shown in Fig. 6.1, which is in interaction with the environment at temperature T∞, the convection heat loss from the surface of the body at any instant will equal the rate of change of the internal energy of the body. Thus
© Springer Nature Singapore Pte Ltd. 2020 R. Karwa, Heat and Mass Transfer, https://doi.org/10.1007/9789811539886_6
329
330
6
Unsteady or Transient Heat Conduction q
Body T(τ )
Fig. 6.1 Cooling of a lump
q ¼ hAs ðt t1 Þ ¼ cðqV Þ
dt ds
ð6:1Þ
where h As t t∞ c q (qV) dt/ds
is is is is is is is is
the the the the the the the the
convection heat transfer coefﬁcient, surface area of the body, temperature of the body at any instant, temperature of the surrounding fluid, speciﬁc heat of the body, density of the material of the body, mass of the body of volume V, rate of change of temperature of the body.
If the body is initially (i.e. at time s = 0) at a uniform temperature ti, the temperature of the body after time s can be determined by integrating Eq. (6.1). Z
Z dt hAs ds ¼ t t1 cqV
or lnðt t1 Þ ¼
hAs s þ C1 cqV
The constant of integration C1 can be found from the condition that t = ti at s = 0. This gives lnðti t1 Þ ¼ C1
Hence, we obtain
hAs lnðt t1 Þ ¼ s þ lnðti t1 Þ cqV
or
t t1 hAs ¼ s ti t1 cqV t t1 hAs ¼ exp s ti t1 cqV ln
ð6:2Þ
6.2 Lumped heat capacity Analysis
331
Cooling of a body
t − t ti − t
Time, τ
Fig. 6.2 Cooling curve of a lumped heat capacity system
From Eq. (6.2), it is evident that the temperature of the body falls or rises exponentially with time s as shown in Fig. 6.2. The exponent of Eq. (6.2) can be transformed into a product of two dimensionless parameters: 2 hAs hV As k s s¼ kAs cqV 2 cqV hL k ð6:3Þ ¼ s k cqL2 hL as ¼ k L2 where L (= V/As) is referred to as a characteristic dimension of the solid, and a (= k/cq) is the thermal diffusivity of the solid. The values of the characteristic dimension L, in Eq. (6.3), for some simple geometric shapes are given in Table 6.1. The ﬁrst dimensionless term (hL/k), in Eq. (6.3), is called the Biot number Bi. The second dimensionless term is dimensionless time and is referred as Fourier number Fo. That is, hL k as Fo ¼ 2 L
Bi ¼
ð6:4Þ ð6:5Þ
Equation (6.2), thus, becomes t t1 ¼ expðBiFoÞ ti t1
ð6:6Þ
The Biot number compares the relative magnitudes of the internal resistance to conduction heat transfer and surface resistance to convection heat transfer. A low value of the Biot number means that the internal resistance is small in comparison to the surface resistance and due to the low internal resistance the temperature throughout the body will be uniform. In
332
6
Unsteady or Transient Heat Conduction
Table 6.1 Characteristic dimension L Type of solid
Volume, V
Surface area, As
Characteristic dimension, L = V/As
Cylinder (diameter D and length L; D 100) From the equation of Bi = hL/k, it follows that when h ! ∞ at the given physical parameters and size of the body, Bi ! ∞, i.e. when heat is removed at a very high rate from the surface of the body. Under these conditions, the cooling of the body is governed by the physical properties and the dimensions of the body. The temperature distribution, for this case, is shown in Fig. 6.5 for a plane wall.
6.2 Lumped heat capacity Analysis
335
t −t ti − t
Fo1 Fo2 Fo3
x
0
x
Fig. 6.5 Temperature distribution in a plane wall cooled at Bi ! ∞; Fo1 < Fo2 < Fo3
t −t ti − t
Fo1 Fo2 Fo3
x
0
x
Fig. 6.6 Temperature distribution in a plane wall cooled at 100 > Bi > 0.1; Fo1 < Fo2 < Fo3
t −t ti − t
Fo1 Fo2 Fo3
x
x 0
Fig. 6.7 Temperature distribution in a plane wall cooled at Bi < 0.1; Fo1 < Fo2 < Fo3
(2) 100 > Bi > 0.1 The rate of cooling in this case depends both on the internal and external resistances. The temperature distribution curve at any moment of time will appear as shown in Fig. 6.6. (3) Very small Biot number (practically Bi < 0.1) The Biot number will be small for a thin plate with a large value of thermal conductivity and small heat transfer coefﬁcient. At the small values of the Biot number, the surface temperature differs little from the temperature at the axis of the plate, that is, the temperature distribution across the plate can be assumed to be uniform at any moment of time, Fig. 6.7. Thus, the lumped heat capacity analysis is applicable to the condition when Bi 0.1. In this case, the rate of cooling depends on the value of the heat transfer coefﬁcient only.
336
6
Unsteady or Transient Heat Conduction
Example 6.1 In a lumped heat capacity system, if there is internal heat generation at a constant volumetric rate qg W/m3, determine the relationship between time and temperature. Solution Following the procedure outlined in Sect. 6.1, the heat balance for the body, which is in interaction with the environment at temperature t∞, the convection heat loss from the surface of the body at any instant will equal the rate of change of the internal energy of the body and the internal heat generation. Thus dt q ¼ hAs ðt t1 Þ ¼ cðqV Þ þ qg V ds or dt hAs qg V þ hAs t1 þ t¼ ds qcV qcV Its solution is the general solution of the homogeneous equation and the particular integral t = t∞ + qgV/hAs. Hence, hAs qg V s þ t1 þ t ¼ C1 exp hAs qcV or hAs qg V s þ t t1 ¼ C1 exp hAs qcV
Applying the initial condition of ti at time s = 0, we obtain qg V ti t1 ¼ C1 þ hAs or C1 ¼ ðti t1 Þ
qg V : hAs
Substitution of the value of C1 in Eq. (i) gives qg V hAs qg V s þ ¼ ðti t1 Þ : exp hAs hAs qcV
t t1
Rearranging the terms and introducing hAss/qcV = Bi Fo, we obtain t t1 ¼ ðti t1 Þ expðBi FoÞ þ which is the desired relation.
qg V ½1 expðBi FoÞ; hAs
ðiÞ
6.2 Lumped heat capacity Analysis
337
or q V
t t1 hAg s
q V
ti t1 hAg s
¼ expðBi FoÞ;
Example 6.2 A 0.8 mm diameter Nichrome wire (q = 8400 kg/m3, c = 420 J/(kg K), k = 12 W/(m K) and electric resistivity qe = 1.2 10–6 Xm) is dipped in oil at bulk temperature of 25°C. When 10 A current is passed through the wire, determine the time for the wire to come within 1°C of its steadystate temperature. The convection heat transfer coefﬁcient is 500 W/(m2 K). Solution The Biot number for a cylindrical conﬁguration, Bi ¼
hðd=4Þ 500 ð0:8=4000Þ ¼ ¼ 0:0083: k 12
Since Biot number is less than 0.1, lumped heat capacity analysis can be applied. For a lumped heat capacity system with internal heat generation at a constant volumetric rate qg W/m3, temperature–time relationship is given by, refer Example 6.1, q V
t t1 hAg s
q V
ti t1 hAg s
¼ expðBi FoÞ
or qg d 4h qg d t1 4h
t t1 ti
¼ expðBi FoÞ
as V ðp=4Þd2 L d ¼ : ¼ As pdL 4 Here qg ¼ I 2 Re =½ðp=4Þd 2 L, where electric resistance, Re ¼
qe L qe L 1:2 106 1 ¼ ¼ ¼ 2:39 X: Ac ðp=4Þd2 ðp=4Þ ð0:8=1000Þ2
Hence, qg ¼
102 2:39 ¼ 4:75 108 W=m3 :
0:8 2 p 4 1000 1:0
ðiÞ
338
6
Unsteady or Transient Heat Conduction
as k s k s Fo ¼ 2 ¼ ¼ 2 L qc L qc ðd=4Þ2 12 s ¼ ¼ 85:03 s: 8400 420 ð0:8=4000Þ2 In steady state, the heat generation rate equals the heat transfer rate from the wire surface, hence qg V ¼ hAs ðtw t1 Þ or tw ¼ t1 þ
qg V qg d 4:75 108 ð0:8=1000Þ ¼ 25 þ ¼ 215 C; ¼ t1 þ hAs 4h 4 500
where tw is the wire temperature in steadystate condition. Hence, t = 215–1 = 214°C. Substitution of the values of various terms in Eq. (i) gives 8
ð0:8=1000Þ 214 25 4:75104500 8
ð0:8=1000Þ 25 25 4:75104500
¼ expð0:0083 85:03sÞ
Solution gives s = 7.4 s. Example 6.3 A 72 mm diameter orange is subjected to a cold environment. The heat transfer coefﬁcient h is estimated to be 10 W/(m2 K). Could a lumped thermal capacity analysis be applied? Thermal conductivity of apple is 0.6 W/(m K) at 20°C. Solution Assuming apple to be spherical, the characteristic dimension, L = D/6 = 72/6 = 12 mm Biot number, Bi ¼
hL 10 0:012 ¼ ¼ 0:2; k 0:6
which is greater than 0.1, hence the lumped heat capacity analysis must not be used. Example 6.4 A right circular cone has a base diameter of 50 mm and a height of 75 mm. The thermal conductivity of the cone material is 15 W/(m K). The heat transfer coefﬁcient is 50 W/ (m2 K). Can we apply the lumped capacity analysis? Neglect the heat transfer from the base. Solution The characteristic dimension is L¼
1 2 V RH 3 pR H ¼ ¼ ; As pRðR2 þ H 2 Þ1=2 3ðR2 þ H 2 Þ1=2
6.2 Lumped heat capacity Analysis
339
where R H
radius of the base = 25 mm, height of the cone = 75 mm. Substitution gives L¼
25 75 3ð252 þ 752 Þ1=2
8 mm:
Biot number, Bi ¼
hL 50 0:008 ¼ ¼ 0:0267\0:1: k 15
Lumped capacity analysis can be applied. Example 6.5 What can be the maximum diameter of a steel ball [k = 40 W/(m K)] subjected to a convective heat transfer coefﬁcient h = 25 W/(m2 K) for the applicability of the lumped capacity analysis? Solution The characteristic dimension is L¼
V D ¼ As 6
For the lump capacity analysis to be applicable, Bi must be less than 0.1. Bi ¼ 0:1 ¼
hLmax 25 Dmax =6 ¼ 40 k
or Dmax ¼ 0:96 m; which is the desired result. Example 6.6 Steel balls 10 mm in diameter are annealed by heating to 880°C and then slowly cooling to 100°C in an air environment at 25°C. The convection heat transfer coefﬁcient is 15 W/(m2 K). Estimate the time required for this cooling process. For steel take k = 40 W/(m K), c = 450 J/(kg K) and q = 7900 kg/m3. Solution For a sphere, characteristic length L = R/3 = 0.005/3 = 1/600 m. Biot number = hL/k = 15/(40 600) = 6.25 10–4 < 0.1, hence the lumped heat capacity solution is applicable. For this condition
340
6
Unsteady or Transient Heat Conduction
t t1 ¼ exp½Bi Fo ti t1 where Fo ¼
This gives
as k s 40 ¼ ¼ s ¼ 4:051s: 2 2 L qc L 7900 450 ð1=600Þ2
100 25 ¼ exp½6:25 104 4:051s 880 25
or s ¼ 962 s: Example 6.7 A 1 mm diameter copper wire initially at a temperature of 140°C is suddenly placed in the atmosphere at 40°C. The convective heat transfer coefﬁcient is 12 W/(m2 K). Calculate the time required for the wire to reach a temperature of 90°C. For copper, q = 8954 kg/m3, c = 0.3831 kJ/(kg K), and k = 386 W/(m K). Solution For a cylindrical body, characteristic length L = V/As = r/2 = 0.25 mm. Biot number = hL/K = 12 0.25/(386 1000) = 7.77 10–6 < 0.1, hence the lumped heat capacity analysis is applicable. For the given data, t t1 90 40 ¼ 0:5 ¼ ti t1 140 40 as k s 386 Fo ¼ 2 ¼ ¼ s ¼ 1800:4s L qc L2 8954 383:1 ð0:25=1000Þ2
From Eq. (6.6),
t t1 Bi Fo ¼ ln ti t1
Substitution of the values of various terms gives 7:77 106 1800:4s ¼ lnð0:5Þ or s ¼ 49:55 s:
6.2 Lumped heat capacity Analysis
341
Example 6.8 How much heat has been removed from the steel balls of Example 6.6 in 962 s? Solution From Eq. (6.13), q ¼ cqVðti t1 Þ½expðBi FoÞ 1 4 5 3 ¼ 450 7900 p ð880 25Þ 3 1000 ð0:08772 1Þ ¼ 1452 W: Alternatively using Eq. (6.14) q ¼ cqVðti t1 Þ
4 5 3 ¼ 450 7900 p 3 1000
ð880 100Þ ¼ 1452 W:
Example 6.9 A cubical piece of aluminium 10 mm on a side is to be heated from 50°C to 300° C directly by flame. How long should the piece remain in the flame, if the flame temperature is 800°C and the heat transfer coefﬁcient between the flame and the aluminium piece is 190 W/ (m2 K)? For aluminium, q = 2719 kg/m3, c = 0.871 kJ/(kg K) and k = 215 W/(m K). Solution The characteristic dimension for a cube with side L is L¼ From Eq. (6.4), Bi ¼
V L3 L ¼ 2¼ A s 6L 6
hL 190 1 102 ¼ ¼ 1:473 103 \0:1: k 6 215
As the Biot number is less than 0.1, the lumped heat capacity analysis can be used, which gives t t1 hAs ¼ exp s ð6:2Þ ti t1 cqV Inserting the given data, we have 300 800 190 6 102 1000 ¼ exp s 50 800 0:871 1000 2719 103
342
6
Unsteady or Transient Heat Conduction
Solution of the above equation gives s ¼ 8:41 s: Example 6.10 A chromel–alumel thermocouple (diameter = 0.7 mm) is used to measure the temperature of a gas stream for which the heat transfer coefﬁcient is 600 W/(m2 K). Estimate the time constant of the thermocouple. Given: q = 8500 kg/m3; c = 400 J/(kg K). Solution The thermocouple bead is very small in diameter, hence the lumped heat capacity analysis can be applied. t t1 hAs ¼ 0:368 ¼ exp s ti t1 cqV or
hAs s¼1 cqV
Substitution of the values of various parameters gives "
600 pð0:7Þ2 1000 400 8500 ðp=6Þ ð0:7Þ3
# s¼1
or s ¼ 0:66 s: Example 6.11 The bead of a copper constant thermocouple is 0.5 mm in diameter. If its initial temperature is 30°C and the surrounding air temperature is 80°C, how long will it take the thermocouple bead to attain (i) 79°C (ii) 79.5°C and (iii) 79.9°C? Given: k = 23 W/ (m K), c = 410 J/(kg K), q = 8900 kg/m3, and h = 50 W/(m2 K). Solution From Eq. (6.2), cqV t t1 s¼ ln hAs ti t1 Substituting values of various terms and V/As = D/6 for a spherical body, we get s ¼ 6:08 ln
t 80 50
For t = 79.0°C, s = 23.8 s, for t = 79.5°C, s = 28.0 s and for t = 79.9°C, s = 37.8 s.
6.2 Lumped heat capacity Analysis
343
Note: If the bead diameter is doubled, the abovecalculated values will be doubled as time s is proportional to L = D/6. It is evident that the measurement of the temperature of an unsteady thermal system, the diameter of the bead of the thermocouple should be as small as possible. Example 6.12 Consider a mercury thermometer (k = 10 W/(m K), c = 140 J/(kg K), q = 13.6 g/cm3) for the measurement of the temperature of the air stream in Example 6.11. The mercury bulb can be assumed to be cylindrical in shape (D = 5 mm). The thermal resistance of the thin glass wall may be neglected. How long will it take the bulb to attain 79°C? Given h = 50 W/(m2 K). Solution Here, s¼
cq V t t1 140 13600 0:005 1 ln ¼ 186:2 s: ln ¼ h As 50 4 50 ti t1
Comparison of the result with that of Example 6.11 clearly shows the disadvantage of the mercury thermometer for recording unsteady temperatures. Example 6.13 If a steel spherical ball of 100 mm diameter is coated with a 1 mm thick layer of dielectric material of thermal conductivity 0.05 W/(m K), estimate the time required to cool it from 500°C to 100°C in an oil bath at 25°C. The convection heat transfer coefﬁcient is 500 W/(m2 K). For steel, k = 40 W/(m K), c = 450 J/(kg K) and q = 7900 kg/m3. Solution The total resistance to heat transfer is r2 r1 1 þ 4pki r1 r2 hð4pr22 Þ 1=1000 1 þ ¼ 4p 0:05 0:05 0:051 500 ð4p 0:0512 Þ
Rt ¼
¼ 0:6853 K=W
The overall heat transfer coefﬁcient, U¼
1 1 1 ¼ ¼ ¼ 44:64 W= m2 K : Rt A2 Rt 4pr22 0:6853 4p 0:0512
Note: Since the difference in r2 and r1 is very small, the overall heat transfer coefﬁcient
0 1 r2 r1 1 1 1=1000 1 1 þ can also be approximated from U ¼ Rt ¼ þ ¼ h 0:05 500 k 2 ¼ 45:45 W=ðm KÞ. For a sphere, characteristic length L = D/6 = 0.102/6 = 0.017 m.
344
6
Unsteady or Transient Heat Conduction
Biot number Bi ¼
UL 44:64 0:017 ¼ ¼ 0:019\0:1 k 40
hence the lumped heat capacity method can be used. From Eq. (6.2),
UAs t t1 s ¼ ln cqV ti t1
or
t t1 s ¼ ln ti t1 where
V As
cqV t t1 cqD ¼ ln UAs 6U ti t1
3
=6 D ¼ pD pD2 ¼ 6 . Hence, time for cooling on substitution of values of various terms
100 25 450 7900 0:102 s ¼ ln ¼ 2499 s ¼ 0:694 h: 500 25 6 44:64 Example 6.14 A rectangular plate of thickness d = 50 mm (width W, length L; d L and d W), which is initially at a uniform temperature of 400°C, is exposed to a fluid at 25°C. The convective heat transfer coefﬁcient is 50 W/(m2 K). Determine the time required to lose 50% of its stored heat and temperature of the plate at that time. Neglect radiation heat exchange. For the plate material, k = 210 W/(m K), c = 900 J/(kg K) and q = 2710 kg/m3. Solution Characteristic dimension L of a plate from Table 6.1 is d/2 = 0.025 m. Biot number, Bi ¼
hL 50 0:025 ¼ ¼ 0:00595\0:1: k 210
Hence, lumped heat capacity method can be applied. From Eq. (6.13), the heat flow qt in time s is
hAs qt ¼ cqVðti t1 Þ exp s 1 cqV
ð6:13Þ
Maximum possible total heat flow is (when body acquires surrounding temperature) qmax ¼ cqVðti t1 Þ Hence, 50% of the maximum possible heat flow in time s is
6.2 Lumped heat capacity Analysis
345
qt ¼ 0:5cqVðti t1 Þ Substitution in Eq. (6.13) gives hAs 0:5cqVðti t1 Þ ¼ cqVðti t1 Þ exp s 1 cqV or
hAs 0:5 ¼ exp s 1 cqV or hAs 0:5 ¼ exp s cqV or
hAs s ¼ ln 0:5 cqV
or
cqV s ¼ ln 0:5 hAs
From Table 6.1, V/As = d/2 for a plate. Hence, cqd 900 2710 0:05 s ¼ ln 0:5 ¼ ln 0:5 ¼ 845 s: 2h 2 50 From Eq. (6.2), temperature after time s is hAs t ¼ t1 þ ðti t1 Þ exp s cqV or 2h ¼ t1 þ ðti t1 Þ exp s cqd or ¼ 25 þ ð400 25Þ exp
Alternatively, simply use
2 50 845 ¼ 212:5 C: 900 2710 0:05
qt cqVðt t1 Þ ¼ 0:5 to calculate t. ¼ qmax cqVðti t1 Þ
346
6
Unsteady or Transient Heat Conduction
Note: When radiation heat exchange is present, radiation heat transfer coefﬁcient is calculated from
2 hr ¼ er T 2 þ Tsur ðT þ Tsur Þ when the body at temperature T is in a large space having surface temperature Tsur, for details about radiation heat transfer coefﬁcient refer Chap. 11. It is to note that hr varies as the body temperature changes. The radiation heat transfer coefﬁcient hr has its maximum value when body temperature T is maximum, which is at the start of cooling when the body is cooled or is the temperature attained after heating if the body is heated. If the Biot number based on the sum of convection heat transfer coefﬁcient and the maximum radiation heat transfer coefﬁcient is less than 0.1, lumped heat capacity analysis may be applied. When both convection and radiation are present, the heat balance for the body gives cðqV Þ
dT 4 ¼ hAs ðT T1 Þ þ eAs r T 4 Tsur ds
Thus the variation of the body temperature is given by ZT Ti
As dT ¼ TðsÞ Ti ¼ qcV
Zs
4 ds; hðT T1 Þ þ er T 4 Tsur
0
which is solved numerically. If only radiation is present (body in an enclosure with vacuum), the heat balance equation is cðqV Þ
dT 4 ¼ eAs r T 4 Tsur ds
Separating variables and integrating from initial condition to time s, we have Zs 0
qcV ds ¼ eAs r
ZT Ti
T4
dT ; 4 Tsur
which gives qcV s¼ 3 4eAs rTsur
Tsur þ T Tsur þ Ti T Ti 1 1 ln þ 2 tan tan : ln Tsur T Tsur Tsur Ti Tsur
6.3 Lumped Capacitance, Varying Fluid Temperature
6.3
347
Lumped Capacitance, Varying Fluid Temperature
Consider the body of Fig. 6.1, which is in contact with a fluid whose temperature increases with time, i.e. t1 ¼ as
ðiÞ
where a is a constant. From Eq. (6.1), we write q ¼ hAs ðt t1 Þ ¼ cðqV Þ
dt ds
dt hAs hAs þ t t1 ¼ 0 ds qcV qcV or dt hAs hAs þ t¼ as ds qcV qcV It is a differential equation whose solution is hAs qcV t ¼ b exp s þa s hAs qcV
ðiiÞ
Initially, the body and the surrounding fluid are in thermal equilibrium, hence from Eq. (i), t = 0 at s = 0. Applying the condition in the above equation, we get b¼a
qcV hAs
and the solution, Eq. (ii), becomes qcV hAs s t ¼ as a 1 exp hAs qcV The variation of the temperature of the body given by this equation is shown in Fig. 6.8. The temperature of the body lags behind the fluid temperature. After the initial transition period, this time lag becomes constant and the temperature of the body varies linearly with time according to the equation qcV t ¼a s hAs
ð6:15Þ
qcV is the time lag by which the linear temperature variation of the body is delayed hAs with respect to the fluid temperature. where
348
6
Unsteady or Transient Heat Conduction
t
t∞ = aτ t(τ)
ρcV
τ
hAs Fig. 6.8 Temperaturetime history: linearly varying fluid temperature
Example 6.15 The thermometer of Example 6.12 is to be used to measure the temperature of an electric oven. Calculate the temperature lag of the thermometer if the oven is heated at the rate of 200°C/h. The convective heat transfer coefﬁcient may be taken as 10 W/(m2 K). Solution From Eq. (6.15), the temperature recorded by the thermometer at any moment after the transition period is qcV t ¼a s hAs Hence, the temperature lag is qcV hAs 200 13600 140 0:005 ¼ 13:2 C: ¼ 3600 10 4
Dt ¼ as t ¼ a
6.4
MultipleLumped Capacity Systems
Such systems can also be analysed by following the procedure given for a singlelump heatcapacity system. Let us consider a twolump system consisting of electrically heated liquid in a container, Fig. 6.9. There is a convective heat transfer from the liquid to the container and from the container to the surrounding fluid at t∞. The energy balance at any moment on the two lumps (liquid 1 and container 2) gives h1 A1 ðt1 t2 Þ ¼ c1 ðq1 V1 Þ
dt1 ds
ðiÞ
6.4 MultipleLumpedCapacity Systems
349 Insulated cover Container
t2
t1 Liquid
1 h1, A1
h2, A2 2
Fig. 6.9 Twolump system
h1 A1 ðt2 t1 Þ þ h2 A2 ðt2 t1 Þ ¼ c2 ðq2 V2 Þ
dt2 ds
ðiiÞ
for the lumps liquid 1 and container 2, respectively These are simultaneous linear differential equations for t1(s) and t2(s). Differentiating Eq. (i) with respect to time s1, we get h1 A1
dt1 dt2 d 2 t1 ¼ c1 ðq1 V1 Þ 2 ds ds ds
Substitution of the value of dt2/ds from Eq. (ii) and elimination of t2 from the resulting equation gives d2 t1 dt1 þ C20 t1 ¼ C20 t1 þ C10 ds2 ds
ðiiiÞ
where h 1 A1 h1 A 1 h 2 A2 þ þ q1 c1 V1 q2 c2 V2 q2 c2 V2 h 1 A1 h2 A2 0 C2 ¼ q1 c1 V1 q2 c2 V2 C10 ¼
The solution of Eq. (iii) is t1 ¼ Aem1 s þ Bem2 s þ t1
ðivÞ
where
m1 ¼ m2 ¼
C10 C10 þ
h
h
2 i1=2 C10 4 C20 2 0 2
i1=2 C1 4 C20 2
The constants A and B can be determined from the initial boundary conditions, which are (i) At the beginning (s = 0), the two lumps are at the same temperature ti, i.e.
350
6
Unsteady or Transient Heat Conduction
t1 ð0Þ ¼ t2 ð0Þ ¼ ti : (ii) This condition of t1(0) = t2(0) = ti gives, from Eq. (i), dt1 ¼ 0: ds These conditions give ti ¼ A þ B þ t1
ðvÞ
dt1 ¼ Am1 þ Bm2 þ t1 ¼ 0 ds s¼0
ðviÞ
and
From Eqs. (v) and (vi), m2 ðti t1 Þ m2 m1 m1 B¼ ðti t1 Þ m2 m1 A¼
Substituting values of A and B in Eq. (iv) gives t1 t1 m2 m1 ¼ em 1 s em 2 s ti t1 m 2 m 1 m2 m1
ð6:16Þ
The temperature of the container t2(s) can be obtained by substituting the value of t1(s) from Eq. (6.16) and its derivative dt1/ds into Eq. (i). It is evident that there will be three simultaneous differential equations for a threelump system, four equations for a fourlump system and so on. To draw the thermal network, the thermal capacitance of each lump of the system can be determined from the known values of their volume, density and speciﬁc heats. The thermal resistance connecting the lump must also be determined. The thermal network of the twolump system shown in Fig. 6.9 is presented in Fig. 6.10. Here the connecting resistance is due to convection between the container and liquid. Temperatures t1 and t2 reach ti when the container is insulated from outside an electric heater is run (this is equivalent to opening of switch K2 and closing of switch K1, and charging the capacitors to potential ti). Thereafter, switch K1 is opened and K2 is closed, which causes the dissipation of stored energy in the capacitors through the resistances. We take another example of a twolumpsystem as shown in Fig. 6.11. The system consists of two solids (lumps) in perfect thermal contact (negligible contact resistance). Block 2 is exposed to a fluid at temperature t∞. Initially, both the lumps are at the same temperature, i.e. t1(0) = t2(0) = ti at s = 0.
6.4 MultipleLumpedCapacity Systems K1
351
● t1
ti
R1
C1
t2
R2 K2
C2
●
t∞ t∞ C1 = ρ1c1V1, R1 = 1/h1A1, C2 = ρ2c2V2, R2 = 1/h2A2
Fig. 6.10 The thermal network of twolump system
● t∞
h2, A2 t2, 2 δ2 A1 = A2 ●tc δ1 t1, 1
Insulated container t1 > t2 > t∞
Fig. 6.11 Example 6.16
The heat transfer from lump 1 to 2 is q t1 tc tc t2 ¼ k1 ¼ k2 A d1 =2 d2 =2 where tc is the temperature at the interface. From the two equations, tc ¼
k1 d2 t1 þ k2 d1 t2 k2 d1 þ k1 d2
Using the ﬁrst equation, we get q q1 2k1 k1 d2 t1 þ k2 d1 t2 ¼ ¼ t1 A A1 d1 k 2 d1 þ k 1 d2 Rearranging the equation, we get q1 ¼ A1 k0 ðt1 t2 Þ where k0 ¼
2k1 k2 k 2 d1 þ k 1 d2
By replacing h1 of the twolump system of Fig. 6.9 by k’, temperature t1(s) can be determined from Eq. (6.16). The application of the results obtained here is illustrated in the example that follows.
352
6
Unsteady or Transient Heat Conduction
Example 6.16 For the system shown in Fig. 6.11, the following data is given: For solid 1: c1 = 385 J/(kg K), k1 = 390 W/(m K) and d1 = 60 mm, q1 = 8950 kg/m3; solid 2: c2 = 900 J/(kg K), k2 = 210 W/(m K) and d2 = 50 mm, q2 = 2700 kg/m3; ti = 200°C, t∞ = 40°C and h2 = 20 W/(m2 K). Determine the expression for the temperature of solid 1, assuming applicability of the lumped analysis. Solution k0 ¼
2k1 k2 2 390 210 ¼ 5103 ¼ k2 d1 þ k1 d2 ð210 60 þ 390 50Þ=1000
k 0 A1 k 0 A1 h 2 A2 þ þ q1 c1 V1 q2 c2 V2 q2 c2 V2 k0 k0 h2 ¼ þ þ q1 c1 d1 q2 c2 d2 q2 c2 d2 5103 1000 5103 1000 20 1000 þ þ ¼ 8950 385 60 2700 900 50 2700 900 50 ¼ 0:06685 0 k A1 h 2 A2 0 C2 ¼ q c1 V1 q2 c2 V2 1 5103 1000 20 1000 ¼ ¼ 4:06 106 8950 385 60 2700 900 50 h
i1=2 2 C10 C10 4 C20 m1 ¼ 2 h
i1=2 0:06685 ð0:06685Þ2 4 4:06 106 ¼ 0:06678 ¼ 2 h
i1=2 2 C10 þ C10 4 C20 m2 ¼ 2 h
i1=2 0:06685 þ ð0:06685Þ2 4 4:06 106 ¼ 0:6 104 ¼ 2 C10 ¼
Then
t1 t1 m2 m1 ¼ em 1 s em 2 s ti t1 m 2 m 1 m2 m1
or 0:6 104 t1 40 e0:06678s ¼ 200 40 0:6 104 ð0:06678Þ ð0:06678Þ 4 e0:610 s 0:6 104 ð0:06678Þ
6.4 MultipleLumpedCapacity Systems
353
or 4
t1 ¼ 40 0:144e0:06678s þ 160:14e0:610
6.5
s
Transient Heat Flow in Semiinfinite Solids
A semiinﬁnite slab is shown in Fig. 6.12, which is bounded by plane x = 0 and extends to inﬁnity in the positive xdirection. The solid also extends to inﬁnity in y and zdirections. Examples of such bodies are a thick or large block of steel, the earth, etc. When such a body is heated or cooled for a relatively short period, the temperature will change only for a short distance from the surface. Case A: Surface Temperature of Semiinﬁnite Body Suddenly Lowered and Maintained at Constant Temperature First, we consider the case when the surface temperature of the semiinﬁnite body is suddenly lowered and maintained at temperature ts. For a onedimensional conduction system without heat generation, temperature distribution equation can be deduced from Eq. (2.13a) as @ 2 t 1 @t ¼ @x2 a @s The boundary conditions are (i) The body is initially at a uniform temperature, tðx; 0Þ ¼ ti
at
s¼0
(ii) At the surface, the temperature is suddenly changed and then maintained at ts, hence tð0; sÞ ¼ ts
at
s[0
(iii) At distance sufﬁciently away from the surface, the temperature in the body does not change, hence tð1; sÞ ¼ ti
at
s[0
ts ● q0 = kA∂t/∂x
● ti
x x=0 Fig. 6.12 A semiinﬁnite solid
354
6
Unsteady or Transient Heat Conduction
The differential equation of this problem is solved by the Laplacetransform technique. The solution for temperature distribution, i.e. the temperature of a parallel plane at distance x at any time s, is tðx; sÞ ts x ¼ erf pﬃﬃﬃﬃﬃ ti ts 2 as where erf
x pﬃﬃﬃﬃﬃ 2 as
ð6:17Þ
is the Gauss error function. It is deﬁned as erf
x pﬃﬃﬃﬃﬃ 2 as
2 ¼ pﬃﬃﬃ p
pxﬃﬃﬃ Z2 as
2
eg dg
ð6:18Þ
0
where η is a dummy variable and the integral is a function of the upper limit. We can express the nondimensional temperature distribution as tðx; sÞ ts 2 ¼ pﬃﬃﬃ ti ts p
pxﬃﬃﬃ Z2 as
2
eg dg
ð6:19Þ
0
The temperature distribution given by Eq. (6.19) is plotted in Fig. 6.13. The values of the error function are tabulated in Table 6.2. The heat flow at any plane at distance x from the surface is qx ¼ kA
@t @x
The differentiation of Eq. (6.19) gives @t 2 @ x x2 4as pﬃﬃﬃﬃﬃ ¼ ðti ts Þ pﬃﬃﬃ e @x @x 2 as p
2 x ðt i t s Þ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ e 4as pas
1.0
0.9953
t ( x,τ ) − t s ti − t s
0, 0
x 2 ατ
Fig. 6.13 Temperature distribution given by Eq. (6.19)
2.0
6.5 Transient Heat Flow in Semiinfinite Solids
355
Table 6.2(a) Error function pxﬃﬃﬃﬃ 2 as
ﬃ erf 2pxﬃﬃﬃ as
pxﬃﬃﬃﬃ 2 as
ﬃ erf 2pxﬃﬃﬃ as
pxﬃﬃﬃﬃ 2 as
ﬃ erf 2pxﬃﬃﬃ as
pxﬃﬃﬃﬃ 2 as
ﬃ erf 2pxﬃﬃﬃ as
0.00
0.0000
0.65
0.6420
1.30
0.9340
1.95
0.9942
0.05
0.05636
0.70
0.6778
1.35
0.9435
2.00
0.9953
0.10
0.1125
0.75
0.7111
1.40
0.9523
2.10
0.9970
0.15
0.1680
0.80
0.7421
1.45
0.9593
2.20
0.9981
0.20
0.2227
0.85
0.7706
1.50
0.9661
2.30
0.9989
0.25
0.2764
0.90
0.7969
1.55
0.9713
2.40
0.9993
0.30
0.3286
0.95
0.8208
1.60
0.9764
2.50
0.9996
0.35
0.3794
1.00
0.8427
1.65
0.9802
2.60
0.9998
0.40
0.4284
1.05
0.8625
1.70
0.9838
2.80
0.99993
0.45
0.4755
1.10
0.8802
1.75
0.9865
3.0
0.99998
0.50
0.5205
1.15
0.8962
1.80
0.9891
3.2
0.999994
0.55
0.5633
1.20
0.9103
1.85
0.9910
3.4
0.999998
0.60
0.6039
1.25
0.9229
1.90
0.9928
3.6
1.000000
Table 6.2(b) Inverted table ﬃ erf 2pxﬃﬃﬃ as
pxﬃﬃﬃﬃ 2 as
0.0
0
0.2
0.1791
0.25
0.2253
0.3
0.2725
0.4
0.3708
0.5
0.4769
0.6
0.5951
0.7
0.7329
0.75
0.8134
0.8
0.9062
0.9
1.1631
1.0
3.6000
Hence, ðti ts Þ qx ¼ kA pﬃﬃﬃﬃﬃﬃﬃﬃ e pas
2 x 4as
ð6:20aÞ
The heat flow at the surface is obtained by putting x = 0, ðti ts Þ qx¼0 ¼ kA pﬃﬃﬃﬃﬃﬃﬃﬃ pas
ð6:20bÞ
356
6
Unsteady or Transient Heat Conduction
The heat flux at the surface is a function of (1/√s), hence it diminishes with an increase in time. The total heat flow in time s can be obtained by integrating Eq. (6.20b) over the time interval 0 to s, i.e. Zs kA 1 Q0 ¼ pﬃﬃﬃﬃﬃﬃ ðti ts Þ pﬃﬃﬃ ds s pa 0
pﬃﬃﬃ kA ¼ pﬃﬃﬃﬃﬃﬃ ðti ts Þ 2 s pa rﬃﬃﬃ s ¼ 1:128kAðti ts Þ a
ð6:21Þ
Case B: Constant Heat Flux on Semiinﬁnite Solid The surface of the body is suddenly exposed to uniform heat flux. The body is initially at a uniform temperature, i.e. tðx; 0Þ ¼ ti
at
s ¼ 0;
and, for s > 0, qo @t ¼ k @x x¼0 A The temperature distribution, in this case, is expressed as t ti ¼
2q0
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ
2 x as=p q0 x x 4as 1 erf pﬃﬃﬃﬃﬃ e kA kA 2 as
ð6:22Þ
Case C: Convective Boundary Condition In most of the cases of interest, the heat is transferred to or from the surface of the solid by convection. Hence, at x = 0,
hAðt1 tÞx¼0
@t ¼ kA @x
x¼0
where t∞ is the temperature of the fluid to which the surface of the solid is exposed. The temperature distribution, in this case, is given by pﬃﬃﬃﬃﬃ t ti x hx h2 as x h as þ 2 1 erf pﬃﬃﬃﬃﬃ þ ¼ 1 erf pﬃﬃﬃﬃﬃ exp k k k t1 ti 2 as 2 as
ð6:23Þ
The above equation can be written as pﬃﬃﬃﬃﬃ t t1 x hx h2 as x h as þ 2 1 erf pﬃﬃﬃﬃﬃ þ ¼ erf pﬃﬃﬃﬃﬃ þ exp k k k ti t1 2 as 2 as
ð6:24Þ
6.5 Transient Heat Flow in Semiinfinite Solids
357
1.0 0.8 0.6 0.4
0.2
t − ti t − ti
0.1 0.08
h√(ατ)/k
0.06 0.05 0.04 0.03
0.3
0.6
3
∞
1.0
0.1 0.05
0.02
0.01 0
0.5
x
1.0
1.5
2 ατ Fig. 6.14 Temperature distribution in semiinﬁnite solid with convection boundary condition. Holman JP, adapted for SI units by White PRS, Heat Transfer, McGrawHill Book Co, New York, Copyright 1992. The material is reproduced with permission of McGrawHill Education (Asia)
If h tends to ∞, the surface temperature of the body approaches the temperature of the fluid to which the surface is exposed. For this case, Eq. (6.24) transforms to t t1 x ¼ erf pﬃﬃﬃﬃﬃ ti t1 2 as which is the result presented by Eq. (6.17). The solution given by Eq. (6.23) or (6.24) is also available in graphical form in Fig. 6.14. Example 6.17 A large block of steel, which is initially at a uniform temperature of 30°C, is suddenly changed to and held at 200°C. Calculate the temperature at a depth of 20 mm after 60 s. For steel k = 45 W/(m K) and a = 1.2 10–5 m2/s. Solution For transient heat conduction in inﬁnite thick solids, Eq. (6.17) applies tðx; sÞ ts x ¼ erf pﬃﬃﬃﬃﬃ ti ts 2 as
358
6
Unsteady or Transient Heat Conduction
The given data is: ts = 200°C, ti = 30°C, x = 0.02, a = 1.2 10–5 m2/s and s = 60 s. Hence,
0:02 tðx; sÞ 200 ¼ ð30 200Þ erf pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2 1:2 105 60
or tð0:02; 60Þ ¼ 170 erf ð0:3727Þ þ 200 or tð0:02; 60Þ ¼ 170 0:406 þ 200 ¼ 130:98 C: Example 6.18 If the surface of the steel block of the previous example is subjected to a uniform heat flux of 200 kW/m2, calculate the temperature at the same depth after 60 s. Solution Equation (6.22) applies, which gives t ti ¼
2q0
pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ
2 x as=p q0 x x 1 erf pﬃﬃﬃﬃﬃ e 4as kA kA 2 as
ð6:22Þ
From given data: x2 0:022 ¼ ¼ 0:1389 4as 4 1:2 105 60 x erf pﬃﬃﬃﬃﬃ ¼ 0:406 2 as q0 ¼ 200 103 W/m2 A ¼ 1 m2 k ¼ 45 W=ðm KÞ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ as=p ¼ 1:2 105 60=p ¼ 0:01514 Inserting the values of various parameters, we obtain t¼
2 200 103 0:01514 200 103 0:02 eð0:1389Þ ½1 0:406 þ 30 ¼ 94:3 C 45 1 45 1
Example 6.19 A water pipe is to be buried underground in wet soil (a = 1.8 10−3 m2/h). The night temperature can fall to –5°C, which can remain at this value for 9 h. Calculate the minimum depth at which the pipe is to be laid down so that the surrounding soil temperature does not reach 5°C. The soil temperature at the onset of the night is estimated to be 10°C.
6.5 Transient Heat Flow in Semiinfinite Solids
359
Solution Let at a depth of L, the temperature t just approaches 5°C in 9 h. Assuming the soil as semiinﬁnite solid, Eq. (6.17) applies. Hence, tðx; sÞ ts x ¼ erf pﬃﬃﬃﬃﬃ ti ts 2 as
or
5 ð5Þ L ¼ erf pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 10 ð5Þ 2 1:8 103 9
or
L 0:667 ¼ erf 0:2546
or L ¼ 0:6494 0:2546 or L ¼ 0:165 m; i.e. the pipe must be laid down at a depth greater than 0.165 m. Example 6.20 A semiinﬁnite aluminium slab is initially at a uniform temperature of 400°C. Its surface is suddenly changed to 50°C. How much heat will be removed per unit area of the slab if the temperature at a depth of 100 mm from the surface drops to 100°C? Given k = 225 W/(m K), a = 8.6 10–5 m2/s. Solution For the given data, tðx; sÞ ts 100 50 ¼ 0:1429 ¼ 400 50 ti ts Thus, from Eq. (6.17),
x pﬃﬃﬃﬃﬃ ¼ 0:1429 2 as
erf From Table 6.2,
x pﬃﬃﬃﬃﬃ 0:126 2 as
360
6
or
s¼
0:1 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2 8:6 105 0:126
Unsteady or Transient Heat Conduction
2
¼ 1831 s:
Heat transfer at the surface, Eq. (6.21), rﬃﬃﬃ q s ¼ 1:128kðts ti Þ A a
rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1831 ¼ 1:128 225 ð50 400Þ ¼ 41 107 J=m2 : 8:6 105
The negative sign indicates heat loss from the slab. Example 6.21 If the ground in Example 6.19 is suddenly exposed to a cold wave at –10°C with h = 50 W/(m2 K), will the water in the pipe remain above 4°C in the period of 9 h? The thermal conductivity of the wet soil is 2 W/(m K). Solution For the given condition, Fig. 6.14 applies. The required parameters are pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ h as 50 1:8 103 9 ¼ 3:18 ¼ 2:0 k x 0:165 pﬃﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0:648 2 as 2 1:8 103 9 For these values of the parameters, the ﬁgure gives tðx; sÞ ti 0:3 t1 ti Substitution of known temperature data gives tðx; sÞ 10 0:3 10 10 or t ¼ 4 C:
6.6
Transient Heat Conduction in Infinite Plate
Figure 6.15 shows an inﬁnite plate of thickness 2L. Physically the plate is so large in length and width that end effects may be ignored. This is a case of onedimensional transient heat conduction. Initially, the plate is at a uniform temperature ti. At the time zero, both sides of the plate are suddenly reduced to temperature t1. The differential equation of onedimensional heat conduction for the unsteady state is
6.6 Transient Heat Conduction in Infinite Plate
361
t1 ● t(x,τ)
x
2L
x=0
Fig. 6.15 Inﬁnite plate
@ 2 t 1 @t ¼ @x2 a @s Introducing h = t–t1, the equation becomes @ 2 h 1 @h ¼ @x2 a @s The initial boundary conditions are h ¼ ti t1 ¼ hi
at
s ¼ 0;
0 x 2L
ðaÞ
h¼0
at
x ¼ 0;
s[0
ðbÞ
h¼0
at
x ¼ 2L;
s[0
ðcÞ
Using the technique of separationofvariables, similar to the one used for the twodimensional steadystate problem, the solution is of the form hðx; sÞ ¼ XðxÞHðsÞ It produces two ordinary differential equations @2X þ k2 X ¼ 0 @x2 1 @H þ k2 H ¼ 0 a @s where k2 is the separation constant. The form of solution is 2
hðx; sÞ ¼ ðC1 cos kx þ C2 sin kxÞek as
ðiÞ
362
6
Unsteady or Transient Heat Conduction
In order to satisfy the boundary conditions, k2 must be > 0. From the boundary condition (b), C1 ¼ 0: Using boundary condition (c), we obtain 2
0 ¼ ðC2 sin 2kLÞek as Since C1 = 0, C2 cannot be zero. Hence, sin(2kL) = 0 gives sinð2kLÞ ¼ sinðnpÞ or k¼
np 2L
where n = 1, 2, 3 ….. Substitution in Eq. (i) gives h¼
1 X
np 2
Cn eð2LÞ
as
sin
npx
n¼1
2L
ð6:25Þ
The initial condition (h = hi at s = 0), when applied to Eq. (6.25), gives hi ¼
1 X
Cn sin
n¼1
npx 2L
ð6:26Þ
Equation (6.26) is simply a Fourier series expansion of the initial temperature, so that 1 Cn ¼ L
Z2L hi sin 0
npx dx 2L
4 hi ¼ np The ﬁnal solution is 1 npx 2 h t t1 4X 1 ðnp e 2LÞ as sin ¼ ¼ hi ti t1 p n¼1 n 2L
ð6:27Þ
where n = 1, 3, 5 ….. Equation (6.27) involves the dimensionless number as/L2, which is the Fourier number based on the halfwidth L of the plate.
6.6 Transient Heat Conduction in Infinite Plate
363
The heat flow, using the Fourier equation, can be obtained as dt dx 1 npx X np 2 4kA ðDtÞmax eð 2 Þ Fo cos ¼ 2L 2L n¼1
q ¼ kA
ð6:28Þ
where Dtmax is the difference between the initial and the new face temperature.
6.7
Heisler and Grober Charts
We considered one geometry of unsteady heat conduction in the above section. Solutions have been presented by the researchers for other geometries. The analytical solutions are not useful for rapid calculation, hence the solutions are presented in the graphical form by Heisler. Temperatures as a function of time and spatial position are given in Figs. 6.17, 6.18, 6.19, 6.20, 6.21 and 6.22 for (i) plates whose thickness is small compared to the other dimensions, (ii) cylinders whose diameter is small compared to its length, (iii) Spheres. All parameters in these ﬁgures are dimensionless. The nomenclature used is shown in Fig. 6.16 Temperature used in charts are deﬁned as h ¼ tðx; sÞt1 hi ¼ ti t1 ho ¼ to t1 where t∞ to
surrounding fluid temperature, centreline or centre temperature. L
L
r0 r
r0 r
x (a) Infinite plate
Fig. 6.16 Nomenclature for the Heisler charts
(b) Infinite cylinder
(c) Sphere
364
6
Unsteady or Transient Heat Conduction
1.0 0.5 0.3 0.2 0.1
10
1
θ 0 0.05 0.03 θi 0.02
8
0.8 4
0.5
0.01 0.005
5
6
3 2.5
0.3
16
1/Bi
0.1
0.003 0.002 0.001
100 80 60 50 40 35 30 25 20
14
0.05
2.0
0 0
2
1
1.2 1.6
3
4
8
12 16
20 24 28 30
60
80 100
Fo = ατ/L2
130 200 300 400 500 600 700 150
Fig. 6.17 Midplane temperature of an inﬁnite plate of thickness 2L. Heisler (1947) Temperature charts for induction and constant temperature heating. Trans ASME 69: 227–236. With permission of ASME 1.0 0.5 0.3 0.2
5 4 3
0.1
θo/θi
0.05
50 45 40 35 30
10
1 1.2
8
1/Bi
6
0.8
0.005
60
16 12
1.6
0.01
100 80
20
2.0
0.03 0.02
0.1 0.3 0.5
0.003 0.002 0.001
25
14
2.5
0.6
0 0
1
2
3
4
6
8
12
16
20
25
Fo = ατ/ro2
30 40
60
80
100
130
150 200
300 350
Fig. 6.18 Axis temperature of an inﬁnite cylinder of radius ro. Heisler (1947) Temperature charts for induction and constant temperature heating. Trans ASME 69: 227–236. With permission of ASME
The centreline temperature h0 is obtained using Figs. 6.17, 6.18 and 6.19 while the offcentre temperatures are obtained using Figs. 6.20, 6.21 and 6.22 as illustrated through worked examples. The heat loss can be determined using Figs. 6.23, 6.24 and 6.25, where Qo Q
initial internal energy of the solid, with reference to the temperature t∞, = qVc (ti–t∞), heat lost in time s.
Note: Inspection of Figs. 6.17, 6.18, 6.19, 6.20, 6.21, 6.22, 6.23, 6.24 and 6.25 shows that the dimensionless temperature distribution and heat flows have been expressed in terms of dimensionless parameters Bi and Fo deﬁned as
6.7 Heisler and Grober Charts
365
1.0 0.5 0.3 0.2 0.1
3
0.05
2.4
θo/θi 0.03
2.0
14 12 10
40
8
30
6 5
0.02
4
1.6 0.75
0.01
25
0.5 0.1
0.005
0.35
0.05
0.003
0
1/Bi
1.2
20 16
0.2 1.0
0
0.002 0.001
100 80 60 50
0.5
1
1.5
2
2.5
3 4
6
8
10
20
30
Fo = ατ/ro2
40
50
100
150
200
250
Fig. 6.19 Centre temperature for a sphere of radius ro. Heisler (1947) Temperature charts for induction and constant temperature heating. Trans ASME 69: 227–236. With permission of ASME
1.0 0.9
0.2
0.8 0.4 0.7 x/L
0.6 0.6
θ t −t = θo to − t
0.5 0.4 0.3
0.8
0.2 0.9 0.1 1.0 0 0.01
0.05 0.1
0.5 1
5 10
50 100
1 k = Bi hL
Fig. 6.20 Temperature as a function of centre temperature in an inﬁnite plate of thickness 2L. Heisler (1947) Temperature charts for induction and constant temperature heating. Trans ASME 69: 227–236. With permission of ASME
366
6
1.0 0.9
0.2
0.8 0.4
r/ro
0.7
θ t −t = θo to − t
Unsteady or Transient Heat Conduction
0.6 0.6 0.5 0.4 0.3 0.8 0.2
0.9
0.1 1.0 0 0.0
5 10 100 20 0.5 1 k 1 = Bi hro Fig. 6.21 Temperature as a function of axis temperature in an inﬁnite cylinder of radius ro. Heisler (1947) Temperature charts for induction and constant temperature heating. Trans ASME 69: 227–236. With permission of ASME 0.05 0.1
1.0 0.9 0.8
0.2
0.4
r/ro
0.7
θ t −t = θo to − t
0.6 0.5
0.6
0.4 0.3 0.2
0.8 0.9
0.1 0 1.0 0.0
0.05 0.1
0.5 1 k 1 = Bi hro
5 10
20
100
Fig. 6.22 Temperature as a function of centre temperature in a sphere of radius ro. Heisler (1947) Temperature charts for induction and constant temperature heating. Trans ASME 69: 227–236. With permission of ASME
6.7 Heisler and Grober Charts
367
1.0 0.9 0.8 0.7 Bi 0.001 0.005 0.02 0.1 0.002 0.01 0.05 0.2
0.6 Q/Qo 0.5
0.5 1
2
5
10 20
10
102
50
0.4 0.3 0.2 0.1 0 5 10
104
103
102
101
2
1
103
104
Bi Fo
Fig. 6.23 Dimensionless heat loss Q/Qo with time for a plane wall of thickness 2L (Grober et al. 1961)
1.0 0.9 0.8 0.7 Bi 0.001 0.005 0.02 0.1 0.002 0.01 0.05 0.2
0.6 Q/Qo 0.5
0.5 1
2
5
10
50
20
0.4 0.3 0.2 0.1 0 5 10
4
10
3
10
2
10
101
Bi2Fo
1
10
2
10
103
4
10
Fig. 6.24 Dimensionless heat loss Q/Qo with time for an inﬁnite cylinder of radius ro (Grober et al. 1961)
hL k as Fo ¼ 2 L
Bi ¼
ð6:4Þ ð6:5Þ
where L is a characteristic dimension of the body. For a plate, it is the halfthickness L, and for the cylinder and sphere it is the radius ro, refer Fig. 6.16. Note that the Heisler charts are applicable for values of Fourier number greater than 0.2.
368
6
Unsteady or Transient Heat Conduction
1.0 0.9 0.8 0.7 Bi 0.001 0.005 0.02 0.1 0.002 0.01 0.05 0.2
0.6 Q/Qo 0.5
0.5 1
2
5
10
20
50
0.4 0.3 0.2 0.1 0 5 10
3
104
10
2
10
1
10
1
10
102
103
4
10
2
Bi Fo
Fig. 6.25 Dimensionless heat loss Q/Qo with time for a sphere of radius ro (Grober et al. 1961)
Example 6.22 A large slab of aluminium 100 mm thick is originally at a temperature of 500°C. It is suddenly immersed in a liquid bath at 100°C resulting in a heat transfer coefﬁcient of 1200 W/(m2 K). Determine the temperature at the centre line and the surface 1 min after the immersion. Also, calculate the total heat removed per unit area of the slab during this period. The properties of the aluminium at the given conditions may be taken as q = 2700 kg/m3, c = 0.9 kJ/(kg K) and k = 215 W/(m K). Solution The characteristic length for a plate with thickness d is L¼
d 100 ¼ ¼ 50 mm: 2 2
The Biot number from Eq. (6.4), Bi ¼
hL 1200 50 103 ¼ ¼ 0:279 [ 0:1: k 215
Since the Biot number is greater than 0.1, the lumped heat capacity analysis is not applicable. The problem may be solved using Heisler charts. The Fourier number, Fo ¼
as k s 215 60 2¼ ¼ ¼ 2:12: 2 L qc L 2700 900 0:052
6.7 Heisler and Grober Charts
369
(i) Centreline temperature, to From the Heisler chart (Fig. 6.17) at Fo = 2.1 and 1/Bi = 3.6, for the temperature variation at the centre of the plate, we have t0 t1 0:6: ti t1 Hence, the temperature at the centre of the plate, t0 ¼ 0:6ðti t1 Þ þ t1 ¼ 0:6ð500 100Þ þ 100 ¼ 340 C: (ii) Plate surface temperature, ts The correction factor from Fig. 6.20 for 1/Bi = 3.6 and x/L = 1.0 (where x is the distance from the midplane) is 0.86. Hence, the plate surface temperature ts is given by t t1 t0 t1 t t1 ¼ ¼ 0:6 0:86 ¼ 0:516: ti t1 ti t1 t0 t1 Simpliﬁcation gives the surface temperature as t ¼ 0:516 ðti t1 Þ þ t1 ¼ 0:516 ð500 100Þ þ 100 ¼ 306:4 C: (iii) Total heat removal The heat transferred from the slab in time s can be calculated using Fig. 6.23. From the given data, FoBi2 ¼ 2:12 ð0:279Þ2 ¼ 0:165; and hL ¼ Bi ¼ 0:279: k From Fig. 6.23, Q 0:4 Q0 where Q0 is the initial internal energy content of the slab in reference to the surrounding temperature. Thus, for the unit surface area of the plate, Qo ¼ ðqV Þcðti t1 Þ ¼ ½2700 ð100=1000Þ 1 0:9 ð500 100Þ ¼ 97200 kJ=m2 :
370
6
Unsteady or Transient Heat Conduction
Hence, the heat lost per unit area of the slab in one minute is Q ¼ 0:4 97200 ¼ 38880 kJ=m2 : Example 6.23 For the slab of the previous example, how long it will take for the midplane temperature to drop to 300°C? What will be the temperature at a depth of 20 mm from one of the faces when the midplane temperature will be 300°C? Solution From the previous example, k 3:6: hL For a midplane temperature of 300°C, t0 t1 300 100 ¼ 0:5: ¼ ti t1 500 100 Corresponding to this value of temperature ratio, Fo = 2.7. This gives Fo ¼
as k s 215 s 2¼ ¼ ¼ 2:7 2 L qc L 2700 900 0:052
or s ¼ 76:3 s for midplane temperature to drop to 300 C: Since distance x is measured from the midplane, x/L for the plane 20 mm from one of the face is x 0:05 0:02 ¼ ¼ 0:6 L 0:05 Correction factor for x/L = 0.6 and 1/Bi = 3.6, t t1 ¼ 0:95: t0 t1 Hence, t t1 t0 t1 t t1 ¼ ¼ 0:5 0:95 ¼ 0:475 ti t1 ti t1 t0 t1
6.7 Heisler and Grober Charts
371
which gives t ¼ 0:475 ðti t1 Þ þ t1 ¼ 0:475 ð500 100Þ þ 100 ¼ 290 C where t is the temperature after 76.3 s at a depth of 20 mm from the surface of the slab when the midplane temperature is 300°C. Example 6.24 A water pipeline is laid along the centreline of a 300 mm thick, very long and wide, brick masonry wall (k = 0.66 W/(m K), a = 0.0046 cm2/s). The water is initially at a temperature of 10°C. The wall is suddenly exposed to a cold wave [average heat transfer coefﬁcient, hav = 20 W/(m2 K)] at –2°C. How long will it take for the water to reach a temperature of 4°C. Solution Given data: Midplane temperature t0 = 4°C, surrounding fluid temperature t∞ = –2°C, ti = 10°C and L = 0.15 m. Hence, t0 t1 4 ð2Þ ¼ 0:5 ¼ ti t1 10 ð2Þ 1 k 0:66 ¼ ¼ ¼ 0:22: Bi hL 20 0:15 From Fig. 6.17, Fo 0.6. This gives L2 0:152 s ¼ Fo ¼ 0:6 ¼ 29347 s ¼ 8 h 9 min: a 0:0046 104 Example 6.25 A long copper cylinder 100 mm in diameter and initially at 20°C is suddenly exposed to a fluid at 80°C. The average value of the heat transfer coefﬁcient is 500 W/(m2 K). Calculate the temperature at a radius of 20 mm and the heat gained per m length of the cylinder in 60 s. For the cylinder material k = 380 W/(m K), q = 8950 kg/m3 and c = 380 J/(kg K). Solution k 380 Thermal diffusivity, a ¼ qc ¼ 8950380 ¼ 11:2 105 m2 =s. 5
60 Fourier number, Fo ¼ as ¼ 11:210 ¼ 2:688 0:052 r2 0
Biot number, Bi ¼ hrk0 ¼ 5000:05 ¼ 0:0658 380 Radius ratio, rr0 ¼ 0:02 0:05 ¼ 0:4. From Figs. 6.18 and 6.21, ho 0:7 hi h 0:98: h0
372
6
Unsteady or Transient Heat Conduction
Hence, h ¼ t t1 ¼
h h0 hi ¼ 0:98 0:7 ð20 80Þ ¼ 41:16 C h0 hi
and t ¼ h þ t1 ¼ 41:16 þ 80 ¼ 38:84 C: Heat gained h2 as ¼ 0:0116 k2 hr0 500 0:05 ¼ 0:0658: ¼ 380 k
FoBi2 ¼
From Fig. 6.24, Q 0:3: Q0 The internal energy at time s = 0 is Q0 ¼ qcðpro2 Þhi L ¼ 8950 380 p 0:052 60 ¼ 1:603 106 J=m: Hence, heat gained, Q ¼ 0:3Q0 ¼ 0:3 1:603 106 ¼ 4:808 105 J=m: Example 6.26 For the orange of Example 6.3, estimate the time required for the centre of the orange to reach 10°C if the initial temperature of the orange is 30°C. The temperature of the cold environment is 5°C. The thermal diffusivity of the orange is nearly equal to that of water since it mainly consists of water. Solution At the average temperature of ½ (30 + 10) = 20°C, a = 1.4 10−7 m2/s and k = 0.6 for water. 1 k 0:6 ¼ ¼ 1:67 ¼ Bi hr0 10 36=1000
and h0 t0 t1 10 5 ¼ 0:2: ¼ ¼ hi ti t1 30 5 From Fig. 6.19, Fo 1.2.
6.7 Heisler and Grober Charts
s ¼ Fo
373
r02 1:2 0:0362 ¼ ¼ 11108 s ¼ 3 h 5 min: a 1:4 107
Example 6.27 A long 50 mm diameter aluminium cylinder initially at 300°C is suddenly exposed to a convective environment at 100°C. The average value of the heat transfer coefﬁcient is 200 W/(m2 K). Determine the heat lost per m length of the cylinder in 30 s. Given: k ¼ 215 W=ðm KÞ; c ¼ 0:9 103 J=ðkg KÞ; q ¼ 2700 kg=m3 ; a ¼ 8:85 105 m2 =s: Solution 5
30 Fourier number, Fo ¼ as ¼ 8:8510 ¼ 4:248 0:0252 r2 0
¼ 0:02326: Biot number, Bi ¼ hrk0 ¼ 2000:025 215 2 FoBi ¼ 0:0023: From Fig. 6.24, Q 0:18: Qi Hence, Q ¼ 0:18Qi ¼ 0:18 qcðpro2 1Þðti t1 Þ ¼ 0:18 2700 0:9 103 p 0:0252 1 ð300 100Þ ¼ 171:8 103 J=m: Lumpedcapacity analysis 5 30 ¼ 8:8510 ¼ 16:99 Fourier number, Fo ¼ ðr as ð0:025=2Þ2 =2Þ2 0
Biot number, Bi ¼ From Eq. (6.13),
hr0 =2 k
¼ 2000:025=2 ¼ 0:01163. 215 Q ¼ 1 expðBi FoÞ Qi ¼ 0:1793;
which is in reasonable agreement with the above result. Example 6.28 It was assumed that when Bi 0.1, the body is having uniform temperature and the lumped heat capacity analysis is applicable. Based on the Heisler charts, determine the error in this approximation. Solution (i) Plate From Fig. 6.20, 1 The temperature ratio hh0 ¼ 0:95, for Bi ¼ 10 and Lx ¼ 1:0. This means that the surface temperature hx = L differs by about 5% from the centreline temperature ho.
374
6
Unsteady or Transient Heat Conduction
(ii) Cylinder For the Heisler chart, the Biot number is deﬁned as Bi ¼
hr0 while, for the lumped k
hr0 =2 . k Hence, Bi = 0.1 refers to Bi = 0.2 for the Heisler chart. From Fig. 6.21, h h0 0:91 for 1/Bi = 5 and r/ro = 1.0. Thus, the error can be 9%. (iii) Sphere Following the explanation given for the cylinder, r ¼ 0:3: ðBiÞHeisler ¼ ðBiÞLumped r=3 heat capacity analysis, it is deﬁned as Bi ¼
From Fig. 6.22,
h 0:85 for 1/Bi = 3.33 and r/ro = 1.0. Thus, the error can be 15%. h0
It can be seen that with the increase in 1/Bi or decrease in the value of the Biot number, the aboveapproximated errors will reduce. When the value of the Biot number will approach 0.01 (or 1/Bi approaches 100), the error will approach zero. Hence, the aboveapproximated error of 5–15%, with respect to the Heisler chart, can be regarded as the maximum for the lumped heat capacity analysis. However, it must be noted that the Heisler charts too are approximate. Example 6.29 A rectangularshaped sandstone block of thickness d = 150 mm (width W d and length L d), which is initially at a uniform temperature of 30°C, is exposed on both sides to hot gases at 500°C. The convective heat transfer coefﬁcient is 80 W/(m2 K). Determine the time required by the block to gain 50% of its maximum storage capacity of heat, and minimum and maximum temperatures of the stone at that time. Neglect radiation heat exchange. For the masonry, k = 2.9 W/(m K), c = 750 J/(kg K) and q = 2200 kg/m3. Solution For the given system, characteristic dimension L from Table 6.1 is d/2 = 0.075 m. Biot number, Bi ¼
hL 80 0:075 ¼ ¼ 2:07 [ 0:1: k 2:9
Hence, the lumped heat capacity method is not applicable. Grober and Heisler charts may be used to solve the problem. Time required by the block to gain 50% of its maximum storage capacity of heat From Grober chart (Fig. 6.23), Bi2Fo 2.9 for Q/Qo = 0.5 and Bi = 2.07. Fourier number, Fo ¼
as ¼ L2
k s 2:9 s ¼ ¼ 3:12 104 s: qc L2 2200 750 0:0752
6.7 Heisler and Grober Charts
375
Hence, as k s 2:9 ¼ Bi2 Fo ¼ Bi2 2 ¼ Bi2 L qc L2 2:9 s ¼ ð2:07Þ2 ¼ 13:37 104 s 2200 750 0:0752 or s ¼ 2169 s. For s = 2169 s, Fo ¼ 3:12 104 2169 ¼ 0:676: Minimum and maximum temperatures The minimum temperature is at x = 0, i.e. at midplane and the maximum is at the surface (x = L). From Heisler chart (Fig. 6.17), with Fo = 0.676 and 1/Bi = 1/2.07 = 0.48, ho/hi 0.47. This gives the minimum temperature, tx¼0 ¼ t1 þ ðho =hi Þ ðti t1 Þ ¼ 500 þ 0:47 ð470Þ ¼ 279:1 C: From Heisler chart (Fig. 6.20), with 1/Bi = 1/2.07 = 0.48 and x/L = 1, h/ho 0.46. This gives the maximum temperature, tx¼L ¼ t1 þ ðh=hi Þ ðto t1 Þ ¼ 500 þ 0:46 ð220:9Þ ¼ 398:4 C: Example 6.30 The wall of a furnace is 150 mm thick and is adequately insulated from outside. It is initially at a uniform temperature of 20°C. The wall is exposed to hot gases at 900°C. The convective heat transfer coefﬁcient is 80 W/(m2 K). Determine the time required for the outer surface to reach a temperature of 725°C. Neglect radiation heat exchange. For the furnace wall, k = 1.2 W/(m K), c = 900 J/(kg K) and q = 2500 kg/m3. Solution The wall is equivalent to onehalf of a wall of thickness d (= 2L) with symmetric convection conditions as shown in Fig. 6.26b. Hence, characteristic dimension L from Table 6.1 is d/2 = 0.3/2 = 0.15 m. For the wall in Fig. 6.26b, Biot number,
h, t∞
h, t∞
0.15 a
Fig. 6.26 Example 6.30
h, t∞
δ = 2L b
376
6
Bi ¼
Unsteady or Transient Heat Conduction
hL 80 0:15 ¼ ¼ 10 [ 0:1: k 1:2
Hence, the lumped heat capacity method is not applicable. Heisler graphs will be used. 1 ¼ 0:1 Bi ho to t1 725 900 ¼ 0:2: ¼ ¼ 20 900 hi ti t1 For 1/Bi = 0.1 and ho/hi = 0.2, Fo 0.85 from Fig. 6.17. Equation Fo ¼ Las2 ¼ gives
s ¼ Fo
qcL2 k
¼ 0:85
k qc
s L2
2500 900 0:152 ¼ 35859 s ¼ 9:96 h: 1:2
Example 6.31 A 2.5 mm thick plastic sheet (k = 0.2 W/(m K), a = 1 10–7 m2/s), initially at 180°C, lies on an insulated surface and is cooled by airflow at 20°C. If the convection heat transfer coefﬁcient is 100 W/(m2 K), determine the time required for the surface of the sheet to reach 40°C. Solution The sheet is equivalent to onehalf of a sheet of thickness d (= 2L) with symmetric convection conditions. Hence, the characteristic length L = 0.0025 m. The Biot number, hL 100 0:0025 Bi ¼ ¼ ¼ 1:25: k 0:2 Figure 6.20 gives hL/ho = 0.6 for x/L = 1 and 1/Bi = 0.8. From given data, hL tL t1 40 20 ¼ 0:125; ¼ ¼ hi ti t1 180 20 ho hL ho 1 ¼ 0:208: ¼ ¼ 0:125 0:6 hi hi hL For ho/hi = 0.208 and 1/Bi = 0.8, Fo 1.9 from Fig. 6.17. Knowing that Fo ¼ Las2 , we get s¼
FoL2 1:9 0:00252 ¼ ¼ 118:75 s: a 1 107
Example 6.32 A long glass rod of 40 mm diameter (k = 0.78 W/(m K), c = 840 J/(kg K) and q = 2700 kg/m3), initially at 80°C, is cooled in air [t∞ = 30°C, h = 20 W/(m2 K)]. Determine the time for which the rod should be cooled so that when removed and isolated from the surroundings it attains an equilibrium temperature of 45°C.
6.7 Heisler and Grober Charts
377
Solution Heat removed, Q, in cooling to the equilibrium temperature of tf = 45°C is the difference of the initial heat content and heat content at 45°C with reference to temperature t∞ of air, i.e. Q ¼ Qo Qf or Q Qf mcðtf t1 Þ tf t1 ¼1 ¼1 ¼1 Qo mcðti t1 Þ Qo ti t1 or Q 45 30 ¼ 0:7: ¼1 Qo 80 30 The Biot number, Bi ¼
hro 20 0:02 ¼ 0:51; ¼ 0:78 k
where L = ro = 0.02 m. For Q/Qo = 0.7 and Bi = 0.51, Bi2Fo = 0.48 from Fig. 6.24, which gives Fo ¼
ðBi2 Fo) 0:48 ¼ ¼ 1:845: 2 0:512 Bi
as k s gives Equation Fo ¼ 2 ¼ ro qc ro2 2 qcro 2700 840 0:022 s ¼ Fo ¼ 1:845 ¼ 2146 s: k 0:78 Example 6.33 A 50 mm diameter long rod (c = 1000 J/(kg K) and q = 1200 kg/m3), initially at 110°C, is cooled in air [t∞ = 30°C, h = 60 W/(m2 K)]. If the axis temperature reaches 50°C after cooling for 10 min, determine the thermal conductivity of the rod material. Solution From given data, ho tr¼0 t1 50 30 ¼ 0:25; ¼ ¼ 110 30 hi ti t1 1 k k ¼ ¼ 0:667k; ¼ Bi hro 60 0:025 as k s k 10 60 Fo ¼ 2 ¼ ¼ ¼ 0:8k: ro qc ro2 1200 1000 0:0252 We have to determine that value of k, which will give ho/hi = 0.25 in Fig. 6.18 at values of Fo and 1/Bi from the above equations. By trialanderror procedure, we ﬁnd k = 1.0 W/(m K).
378
6
Unsteady or Transient Heat Conduction
The Biot number Bi = 1.5 for k = 1.0 W/(m K), hence lumped heat capacity analysis is not applicable. Example 6.34 A 10 mm diameter steel ball, initially at a temperature of 30°C, is to be heated by immersing in a salt bath at 1000°C and then quenched for hardening. Hardening occurs at locations where the temperature is greater than 800°C. The heat transfer coefﬁcient is 4000 W/(m2 K). (a) Determine the time required for heating to harden up to a depth of 1 mm and (b) heat flux when t(4 mm, s) = 800°C,. Given for ball material q = 7860 kg/m3, c = 460 J/(kg K) and k = 40 W/(m K). Solution (a) Time required for heating The temperature up to the depth of 1 mm, i.e. from surface to r = 4 mm must be 800°C. From the given data, r 4 ¼ ¼ 0:8; ro 5 1 k 40 ¼ ¼ 2: ¼ Bi hro 4000 0:005 For applicability of lumped heat capacity analysis, Biot number Bi = hro/ (3 k) = 0.166 > 0.1. Hence, the lumped heat capacity analysis is not applicable. From Fig. 6.22, h/ho 0.84 for 1/Bi = 2 and r/ro = 0.8. Hence, ho ¼ This gives
h 800 1000 ¼ ¼ 238:1 o C: 0:84 0:84
ho ho 238:1 ¼ 0:245: ¼ ¼ hi ti t1 30 1000
From Fig. 6.19, Fo 1.13 for ho/hi = 0.245 and 1/Bi = 2. Equation Fo ¼ as r2 ¼ o
gives s ¼ Fo
2 qcro 7860 460 0:0052 ¼ 1:13 ¼ 2:55 s: k 40
(b) Heat flux From Fig. 6.22, h/ho 0.77 for 1/Bi = 2 and r/ro = 1.0. Hence, h ¼ 0:77ho ¼ 0:77 ð238:1Þ ¼ 183:3 o C or
tðr ¼ ro Þ ¼ 1000 183:3 ¼ 816:7 o C: Heat flux ¼ h½tðro Þ t1 ¼ 4000 183:3 ¼ 733:2 kW=m2 :
k qc
s ro2
6.8 Two and ThreeDimensional Transient Heat Conduction Systems
6.8
379
Two and ThreeDimensional Transient Heat Conduction Systems
The Heisler charts presented in the previous sections are applicable to inﬁnite plates of 2L thickness, a long cylinder or a sphere. When the length and width of the plate are not large compared to its thickness, or the length of the cylinder is not large compared to its diameter, the above charts are not applicable. The temperature distribution in such cases is a function of additional space coordinates.
6.8.1 TwoDimensional Systems The differential equation for the semiinﬁnite bar in Fig. 6.27 is @2t @ 2 t 1 @t þ ¼ @x2 @y2 a @s
ð6:29Þ
The solution, based on the separation of the variables technique, is of the form tðx; y; sÞ ¼ Xðx; sÞYðy; sÞ
ð6:30Þ
It will be shown1 that the bar in Fig. 6.27 can be formed from two inﬁnite plates of thickness 2L1 and 2L2, respectively, and the temperature distribution may be expressed as a product of the solution for these plates. 1
For two inﬁnite plates, the differential equations are @ 2 t1 1 @t1 ¼ a @s @x2
ðiÞ
@ 2 t2 1 @t2 ¼ a @s @y2
ðiiÞ
t1 ¼ t1 ðx; sÞ
ðiiiÞ
t2 ¼ t2 ðy; sÞ
ðivÞ
and their temperature distributions are
Let the solution to Eq. (6.29) is a simple product solution of above functions, i.e. t ¼ t1 ðx; sÞt2 ðy; sÞ Then the derivations are @2t @ 2 t1 ¼ t2 2 2 @x @x
ðvÞ ðviÞ
@2t @ 2 t2 ¼ t1 2 2 @y @y
ðviiÞ
@t @t1 @t2 ¼ t2 þ t1 @s @s @s
ðviiiÞ
Using Eqs. (i) and (ii), Eq. (viii) transforms to @t @ 2 t1 @ 2 t2 ¼ at2 2 þ at1 2 @s @x @y Substitution of the value of ∂t/∂s satisﬁes Eq. (6.29): @ 2 t1 @ 2 t2 1 @ 2 t1 @ 2 t2 at2 2 þ at1 2 t2 2 þ t1 2 ¼ a @x @y @x @y Therefore, the assumed product solution, Eq. (v), is correct and the dimensionless temperature distribution for the rectangular bar can be given by Eq. (6.31).
380
6
Unsteady or Transient Heat Conduction
y x z
2L2 2L1
Fig. 6.27 Twodimensional system
h tðx; y; sÞ t1 ¼ ¼ hi ti t1
tðx; sÞ t1 tðy; sÞ t1 ti t1 ti t1 2L1 Plate 2L2 Plate
ð6:31Þ
where ti and t∞ are deﬁned as earlier. Similarly, the temperature distribution in semiinﬁnite plates and cylinders can be expressed.
6.8.2 ThreeDimensional Systems The results of twodimensional systems can be extended to a threedimensional block (2L1 2L2 2L3) to give h tðx; y; z; sÞ t1 ¼ ¼ hi ti t1
tðx; sÞ t1 tðy; sÞ t1 tðz; sÞ t1 ti t1 ti t1 ti t1 2L1 Plate 2L2 Plate 2L3 Plate ð6:32Þ
The product solutions for the temperature distributions in two or threedimensional systems are presented in Table 6.3 Example 6.35 A long rectangular section steel bar (50 mm 80 mm) is initially at a uniform temperature of 800 °C. It is heat treated by quenching in an oil tank at 80°C. The average heat transfer coefﬁcient is 400 W/(m2 K). What is the temperature at the centreline after 120 s of quenching? a = 1.4 10−5 m2/s and k = 50 W/(m K). Solution The problem is a combination of two inﬁnite plates of thicknesses 2L1 and 2L2, respectively. Hence, equation of Case (2) of Table 6.3 applies 2L1 Plate (2L1 = 50 mm) 1 k 50 ¼ ¼5 ¼ Bi hL1 400 0:025 as 1:4 105 120 Fo ¼ 2 ¼ ¼ 2:69: 0:0252 L1
6.8 Two and ThreeDimensional Transient Heat Conduction Systems
381
Table 6.3 Product solutions for transient heat conduction in two and threedimensional systems System
Solution
1. Semiinﬁnite plate y x
h tðy; sÞ t1 ¼ hi ti t1 tðy; sÞ t1 tðx; sÞ t1 ¼ semiinfinite ti t1 ti t1 2L1 Plate
2L1
●
x
y
slab
h tðx; y; sÞ t1 ¼ hi t t1 i tðx; sÞ t1 tðy; sÞ t1 ¼ ti t1 ti t1 2L1 Plate 2L2 Plate
2. Inﬁnite rectangular bar
y x z 2L2 2L1
h
3. Semiinﬁnite cylinder
ro ●
x
hi tðr; sÞ t1 ¼ ti t1
r
●
x
Cylinder
slab
h tðr; x; sÞ t1 ¼ h ti i t1 tðr; sÞ t1 tðx; sÞ t1 ¼ Infinite ti t1 ti t1 2L1 Plate
4. A short cylinder
ro
tðr; x; sÞ t1 t i t1 tðx; sÞ t1 Infinite Semiinfinite ti t1
¼
r
Cylinder
2L1 5. Semiinﬁnite rectangular bar
y x z
h tðx; y; z; sÞ t1 ¼ ti t1 hi tðx; sÞ t1 tðy; sÞ t1 tðz; sÞ t1 ¼ Semiinfinite ti t1 ti t1 ti t1 2L1 Plate 2L2 Plate solid
2L2 2L1 6. Rectangular parallelepiped
y x z
h tðx; y; z; sÞ t1 ¼ hi ti t1 tðx; sÞ t1 tðy; sÞ t1 tðz; sÞ t1 ¼ ti t1 ti t1 ti t1 2L1 Plate 2L2 Plate 2L3 Plate
2L2 2L3 2L1 1 ¼ 5 and Fo ¼ 2:69, From Fig. 6.17, for Bi h0 0:63: hi 2L1 plate
382
6
Unsteady or Transient Heat Conduction
2L2 Plate (2L2 = 80 mm) 1 k 50 ¼ ¼ 3:125 ¼ Bi hL2 400 0:04 as 1:4 105 120 Fo ¼ 2 ¼ ¼ 1:05: 0:042 L2 From Fig. 6.17, h0 0:8: hi 2L2 plate For the given bar, the solution is h0 h0 h0 0:63 0:8 ¼ 0:504 ¼ hi hi 2L1 plate hi 2L2 plate t0 t1 ¼ 0:504 ðti t1 Þ ¼ 0:504 ð800 80Þ ¼ 362:9 o C: Centreline temperature after 120 s, t0 ¼ ðt0 t1 Þ þ t1 ¼ 362:9 þ 80 ¼ 442:9 o C: Example 6.36 A semiinﬁnite steel cylinder [a = 1.4 10–5 m2/s and k = 45 W/(m K)] is 50 mm in diameter. It is initially at a uniform temperature of 30°C. Calculate the temperature at the axis of the cylinder 40 mm away from its one end after 100 s if it is exposed to an environment of 200°C with heat transfer coefﬁcient of 350 W/(m2 K). Solution The problem is a combination of inﬁnite cylinder and semiinﬁnite slab as given in Table 6.3, Case 3. The solution for the centreline temperature is h0 h0 ¼ hi hi
infinite cylinder
h hi semiinfinite slab
The required temperature ratios in Eq. (i) are calculated below. Inﬁnite Cylinder (ro = 0.025 m) 1 k 45 ¼ ¼ 5:143 ¼ Bi hr0 350 0:025 as 1:4 105 100 Fo ¼ 2 ¼ ¼ 2:24: ro 0:0252 From Fig. 6.18,
ðiÞ
6.8 Two and ThreeDimensional Transient Heat Conduction Systems
383
h0 0:45: hi infinite cylinder Semiinﬁnite Slab (x = 0.04 m) The required parameters for this case are pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ h as 350 1:4 105 100 ¼ ¼ 0:291 k 45 x 0:04 pﬃﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0:534: 2 as 2 1:4 105 100 From Fig. 6.14, h 1 0:085 hi or h 1 0:085 ¼ 0:915: hi The required solution from Eq. (i) is h0 h0 ¼ hi hi
infinite cylinder
h ¼ 0:45 0:915 ¼ 0:412 hi semiinfinite slab
or t0 ¼ 0:412 ðti t1 Þ þ t1 ¼ 0:412 ð30 200Þ þ 200 ¼ 129:96 o C: Example 6.37 A steel cylinder 50 mm in diameter and 80 mm long is initially at a uniform temperature of 650°C. It is dipped in a cooling medium at 100°C. If the heat transfer coefﬁcient is 1000 W/(m2 K), determine the temperature at 20 mm away from one of its flat faces and at radius 10 mm after 150 s. Given: a = 1.2 10–5 m2/s and k = 40 W/(m K). Solution The problem is a combination of an inﬁnite cylinder and a 2L plate, refer Case 4, Table 6.3. Inﬁnite Cylinder (ro = 0.025 m) 1 k 40 ¼ ¼ 1:6 ¼ Bi hr0 1000 0:025 as 1:2 105 150 Fo ¼ 2 ¼ ¼ 2:88: ro 0:0252
384
6
h0 From Fig. 6.18, 0:05. hi
From Fig. 6.21, at r/ro = 10/25 = 0.4, Hence,
h ho
Unsteady or Transient Heat Conduction
0:95.
h h0 h ¼ ¼ 0:05 0:95 ¼ 0:0475: hi infinite h0 hi cylinder
2L Plate (2L = 0.08 m) The required parameters for this case are 1 k 40 ¼ ¼ ¼ 1:0 Bi hL 1000 0:04 as 1:2 105 150 Fo ¼ 2 ¼ ¼ 1:125: L 0:042 h0 From Fig. 6.17, 0:5. hi From Fig. 6.20, at x/L = (40 – 20)/40 = 0.5, Hence,
h ho
0:9.
h h0 h ¼ ¼ 0:5 0:9 ¼ 0:45: hi 2Lplate h0 hi
The solution for the given short cylinder is h t t1 h h ¼ 0:0475 0:45 ¼ 0:02138 ¼ ¼ hi hi infinite hi 2Lplate ti t1 cylinder t ¼ 0:02138 ðti t1 Þ þ t1 ¼ 0:02138 ð650 100Þ þ 100 ¼ 111:76o C:
6.9
Numerical Method of Solving Transient Conduction Problems
In the previous sections, the results of mathematical analysis for regularshaped solids under transient conduction have been presented in the form of charts. However, in many problems of practical interest, the geometrical shapes do not conform to these regular shapes, and in some cases, the boundary conditions may also vary with time. It is not possible to provide a mathematical solution to such problems. The ﬁnitedifference method (a numerical technique, which was presented in Chap. 5 to solve twodimensional steadystate conduction problems), can also be used to solve transient conduction problems. The differential equation which governs the heat flow equation within a solid under twodimensional transient conduction is @2t @ 2 t 1 @t þ ¼ @x2 @y2 a @s
ðiÞ
6.9 Numerical Method of Solving Transient Conduction Problems
385
Following the procedure presented in Chap. 5, the derivatives in Eq. (i) may be expressed as (refer Fig. 6.28 for the nomenclature of the nodes):
p p Tmp þ 1;n þ Tm1;n 2Tm;n @2t @x2 m;n ðDxÞ2 2 p p p Tm;n @ t þ 1 þ Tm;n1 2Tm;n @y2 m;n ðDyÞ2
ðiiÞ ðiiiÞ
pþ1 p Tm;n @t Tm;n @s Ds
ðivÞ
where the superscripts designate the time increments. Using the above ﬁnitedifference relations in Eq. (i), we obtain p p Tmp þ 1;n þ Tm1;n 2Tm;n
ðDxÞ2
þ
p p p Tm;n þ 1 þ Tm;n1 2Tm;n
ðDyÞ2
¼
pþ1 p Tm;n 1 Tm;n a Ds
ðvÞ
pþ1 For Dx = Dy, the equation of Tm;n by simpliﬁcation of Eq. (v) is
pþ1 Tm;n
Introducing
¼
aDs ðDxÞ2
aDs ðDxÞ2
Tmp þ 1;n
p þ Tm1;n
p þ Tm;n þ1
p þ Tm;n1
þ 1
4aDs
!
ðDxÞ2
p Tm;n
= Fo, the Fourier number, we get
p p p pþ1 p Tm;n ¼ Fo Tmp þ 1;n þ Tm1;n þ Tm;n þ 1 þ Tm;n1 þ ð1 4FoÞTm;n
ð6:33Þ
pþ1 Thus, the temperature Tm;n of the node (m, n) after a time increment of Ds may be determined if the temperatures of the surrounding nodes at any particular time are known. Such equations are written for each node leading to a set of equations.
y (m, n+1) x
(m, n) (m1, n)
(m+1, n) (m, n1)
nodes Dashed lines indicate the element [the node (m, n)] with volume ( . .1); unit depth.
Fig. 6.28 Nomenclature for ﬁnitedifference approximation (twodimensional transient conduction)
386
6
Unsteady or Transient Heat Conduction
If the distance and time increments in Eq. (6.33) are chosen such that aDs ðDxÞ
2
¼ Fo ¼
1 4
ð6:34Þ
Then, the temperature of node (m, n) after a time increment Ds is pþ1 Tm;n ¼
p p p Tmp þ 1;n þ Tm1;n þ Tm;n þ 1 þ Tm;n1 4
ð6:35Þ
which is simply the arithmetic mean of the temperature of the four surrounding nodes at the beginning of the time increment. For a onedimensional system, Eq. (6.33) becomes
p Tmp þ 1 ¼ Fo Tmp þ 1 þ Tm1 þ ð1 2FoÞTmp
ð6:36Þ
1 ¼ Fo ¼ ; 2 ðDxÞ
ð6:37Þ
and for aDs
2
Tmp þ 1 ¼
p Tmp þ 1 þ Tm1 2
ð6:38Þ
Equation (6.33) can also be obtained by following the method of heat balances. Consider node (m, n) in Fig. 6.28. The rates of heat conduction from the neighbouring nodes are p Tmp þ 1;n Tm;n kðDy 1Þ Dx p p Tm1;n Tm;n kðDy 1Þ Dx p p Tm;n þ 1 Tm;n kðDx 1Þ Dy p p Tm;n1 Tm;n kðDx 1Þ Dy
The change in the internal energy of the elemental volume due to the conduction heat flow into the node (m, n) is qcðDx:Dy:1Þ
pþ1 p Tm;n Tm;n Ds
Heat balance at the node (m, n) gives p p p p p Tmp þ 1;n Tm;n Tm1;n Tm;n Tm;n þ 1 Tm;n þ kðDy:1Þ þ kðDx:1Þ Dx Dx Dy p p pþ1 p Tm;n1 Tm;n Tm;n Tm;n ¼ qcðDx:Dy:1Þ þ kðDx:1Þ Dy Ds
kðDy:1Þ
6.9 Numerical Method of Solving Transient Conduction Problems
387
For Dx = Dy, we get p p p p Tmp þ 1;n þ Tm1;n þ Tm;n þ 1 þ Tm;n1 4Tm;n ¼
Introducing
kDs qcðDxÞ2
qc pþ1 p ðDxÞ2 Tm;n Tm;n kDs
pþ1 ¼ Fo and solving for Tm;n , we get
p p p pþ1 p Tm;n ¼ Fo Tmp þ 1;n þ Tm1;n þ Tm;n þ 1 þ Tm;n1 þ ð1 4FoÞTm;n
ð6:33Þ
pþ1 The calculation of the temperature Tm;n using the above equations requires the selection of certain values of the Fourier number. This choice is limited by the following conditions, known as stability criteria.
For a twodimensional system; Fo 1=4
ð6:34Þ
for a onedimensional system; Fo 1=2
ð6:37Þ
and
For a threedimensional system, the stability criterion can be shown to be Fo 1/6. By simple analysis, we can show that the rejection of these conditions will lead to unrealistic results. Let us consider a onedimensional system, with the adjoining nodes of node (m, n) at equal temperatures, say 30°C. Let the temperature of node (m, n) be 40°C. If we select Fo = 1, ignoring the condition of Eq. (6.37), then after the time increment Ds, Tmp þ 1 would be 20°C, which is lower than the temperature of the adjoining nodes and, hence it is an unrealistic result. This conclusion can be shown to be valid for all values of Fo > 1/2 for the onedimensional case. Let us select some value of Fo ½, say 1/3. Then from Eq. (6.36), we have Tmp þ 1 ¼
p Tmp þ 1 þ Tm1 þ Tmp 3
ðiÞ
p Tmp þ 1 þ Tm1 þ 2Tmp 4
ðiiÞ
and for Fo = ¼, we get Tmp þ 1 ¼
A comparison of Eqs. (i), (ii) and (6.38) shows that a choice of Fo = ½ in case of onedimensional system gives a simple equation. The above conclusions can be extended to the two and threedimensional systems also. In general, smaller the value of increments Dx and Ds, greater is the accuracy of the result but slower will be the convergence to the solution. One may think that a small value of Dx with a large value of Ds might give a solution with greater accuracy with speed, but the values of the increments cannot be selected arbitrarily. Once the value of Fourier number Fo and the distance increment Dx are selected, the upper limit of the time increment Ds is ﬁxed from the stability criteria. We developed the ﬁnitedifference equations for a node lying within a solid. Following the method outlined above, we can develop the equations for other physical situations as illustrated in Example 6.38.
388
6
Unsteady or Transient Heat Conduction
Example 6.38 Write the ﬁnite difference expression for the temperature of a node on a convective boundary in onedimensional heat flow. Solution Figure 6.29 shows the node (m) on the convective boundary. The rate of heat conduction from node (m – 1) to node (m) under consideration kðDy:1Þ
p Tm1 Tmp Dx
Heat convected into the node (m) is p hðDy:1ÞðT1 Tmp Þ
The change in the internal energy of node (m) due to conduction and convection heat flow is pþ1 Dx T Tmp :Dy:1 m qc 2 Ds Heat balance at the node (m) gives pþ1 Dx T Tmp T p Tmp p :Dy:1 m ¼ kðDy:1Þ m1 þ hðDy:1ÞðT1 Tmp Þ qc 2 Ds Dx Using Dx = Dy and rearranging the terms, we get Tmp þ 1 Tmp ¼ Putting
kDs qcðDxÞ2
= Fo and
2kDs qcðDxÞ
hDx k
2
p ðTm1 Tmp Þ þ
hDx 2kDs p T Tmp : k qcðDxÞ2 1
¼ Bi and solving for Tmp þ 1 , we obtain
p p þ Bi T1 Þ þ ½1 2Fo ð2Bi FoÞTmp ; Tmp þ 1 ¼ 2FoðTm1
which is the desired result.
h ●t∞
m●
m1 ●
=
Fig. 6.29 Example 6.38
6.9 Numerical Method of Solving Transient Conduction Problems
389
6.9.1 The Explicit and Implicit Formulations We developed equations in the previous section following the technique of forwarddifference, pþ1 or Tmp þ 1 is expressed wherein the temperature of a node after the time interval Ds, that is, Tm;n in terms of the temperatures of the surrounding nodes at the beginning of the time interval. pþ1 in terms of previous nodal temperUsing this technique, we calculate the temperature Tm;n atures. Such formulation is termed as explicit formulation because the new temperature at the node under consideration appears explicitly on the left hand of the equation. The solution of the problem, thus, proceeds from one time increment to the next till we obtain the desired state. As already discussed, the stability of the solution in this method is governed by the selection of the values of distance and time increments Dx and Ds according to the stability criteria. Alternatively, the ﬁnite difference equation may be written by using the backwarddifference in which the time derivative moves backward (into the past) as explained below. For the interior node (m, n) of Fig. 6.28, we shall write the equations of the rate of conduction heat flow as pþ1 Tmp þþ 11;n Tm;n Dx pþ1 pþ1 Tm1;n Tm;n kðDy:1Þ Dx pþ1 pþ1 Tm;n þ 1 Tm;n kðDx:1Þ Dy
kðDy:1Þ
kðDx:1Þ
pþ1 pþ1 Tm;n1 Tm;n Dy
The change in the internal energy of the elemental volume represented by the node (m, n) is expressed as earlier, that is, by qcðDx:Dy:1Þ
pþ1 p Tm;n Tm;n Ds
Heat balance equation is kðDy:1Þ
pþ1 pþ1 pþ1 pþ1 pþ1 Tmp þþ 11;n Tm;n Tm1;n Tm;n Tm;n þ 1 Tm;n þ kðDy:1Þ þ kðDx:1Þ Dx Dx Dy
þ kðDx:1Þ
pþ1 pþ1 pþ1 p Tm;n1 Tm;n Tm;n Tm;n ¼ qcðDx:Dy:1Þ Dy Ds
For Dx = Dy, we get pþ1 pþ1 pþ1 pþ1 Tmp þþ 11;n þ Tm1;n þ Tm;n þ 1 þ Tm;n1 4Tm;n ¼
Introducing
kDs qcðDxÞ2
qc pþ1 p ðDxÞ2 Tm;n Tm;n kDs
p ¼ Fo and solving for Tm;n , we get
pþ1 pþ1 pþ1 p pþ1 Tm;n ¼ ð1 þ 4FoÞTm;n Fo Tmp þþ 11;n þ Tm1;n þ Tm;n þ 1 þ Tm;n1
ð6:38Þ
390
6
Unsteady or Transient Heat Conduction
p This equation is called implicit formulation. This formulation expresses temperature Tm;n pþ1 in terms of Tm;n . We get a set of equations when all the nodes are considered, which are to be solved simultaneously by using the method discussed in Chap. 5. The greatest advantage of this formulation is that no restriction is imposed on selection of time and distance increments. Thus much larger time increments can be used for the rapid solution of the problem. This actual selection of Ds and Dx is now based on the tradeoff between the accuracy and the cost of computation.
Example 6.39 Write down the nodal equations using explicit and implicit formulations for the cases shown in Fig. 6.30a–c. Solution Following the procedure outlined in the previous sections, the nodal equations can be easily written as illustrated below. A. Explicit Formulation (a) Interior node with convection The summation of heat flow equations by conduction is p p p p Tm þ 1;n Tm;n Tm1;n Tm;n Dy k :1 þ kðDy:1Þ 2 Dx Dx p p p p Tm;n T T Dx m;n þ1 m;n1 Tm;n :1 þk þ kðDx:1Þ 2 Dy Dy Heat flow by convection is
h
Dx þ Dy p Þ :1 ðT1 Tm;n 2 m, n +1
m, n +1
(h, t∞ ) m1, n
m, n
m + 1, n m, n 1 (h, t∞ )
m1, n
m, n 1
a
b m1, n
(h, t∞ ) m, n
●
m, n 1
c Fig. 6.30 Example 6.39
●m, n
6.9 Numerical Method of Solving Transient Conduction Problems
391
The summation of the conduction and convection terms given above will be equal to the increase in the internal energy of the elemental volume. That is, p p p p Tm þ 1;n Tm;n Tm1;n Tm;n Dy :1 þ kðDy:1Þ k 2 Dx Dx p p p p Tm;n T T Dx þ1 m;n1 Tm;n m;n :1 þk þ kðDx:1Þ 2 Dy Dy pþ1 p Tm;n Tm;n Dx þ Dy Dx:Dy p þh :1 ðT1 Tm;n Þ ¼ qc Dx:Dy :1 2 4 Ds For Dx = Dy, we get 1 p 1 p hDx p p p p Tm þ 1;n þ Tm1;n ðT1 Tm;n þ Tm;n Þ þ 1 þ Tm;n1 3Tm;n þ 2 2 k 3 qc pþ1 p ðDxÞ2 Tm;n Tm;n ¼ 4 kDs pþ1 , we get Introducing the Fourier number and Biot number, and solving for Tm;n
2 p p p pþ1 ¼ FoðTmp þ 1;n þ 2Tm1;n þ 2Tm;n Tm;n þ 1 þ Tm;n1 Þ 3 4 4 p þ Fo BiT1 þ ð1 4Fo Fo BiÞTm;n 3 3 The stability condition is 4 1 4Fo Fo Bi 0 3 or Foð3 þ BiÞ
3 4
(b) Node at plain surface with convection The heat balance gives Conduction heat inflow + convection heat inflow = increase in internal energy or p p p p p p Tm;n þ 1 Tm;n Tm1;n Tm;n Tm;n1 Tm;n Dx Dx k :1 :1 þ kðDy:1Þ þk 2 2 Dy Dx Dy pþ1 p Tm;n Tm;n Dx:Dy p þ hðDy:1ÞðT1 Tm;n Þ ¼ qc :1 2 Ds For Dx = Dy, and introducing Fo and Bi, we get p p p pþ1 p Tm;n ¼ Foð2Tm1;n þ Tm;n þ 1 þ Tm;n1 þ 2Bi T1 Þ þ ð1 4Fo 2Fo BiÞTm;n
392
6
Unsteady or Transient Heat Conduction
The stability condition is Foð2 þ BiÞ
1 : 2
(c) Node at the exterior corner with convection The heat balance equation in this case is Conduction heat inflow + convection heat inflow = increase in internal energy or p p p p Tm1;n Tm;n Tm;n1 Tm;n Dy Dx Dx þ Dy p :1 :1 :1 ðT1 Tm;n þk þh Þ 2 2 2 Dx Dy pþ1 p Tm;n Tm;n Dx:Dy ¼ qc :1 4 Ds
k
Using Dx = Dy, and introducing Fo and Bi and simplifying, we get p p pþ1 p Tm;n ¼ 2FoðTm1;n þ Tm;n1 þ 2Bi T1 Þ þ ð1 4Fo 4Fo BiÞTm;n
The stability condition is ð1 4Fo 4Fo BiÞ 0 or Foð1 þ BiÞ
1 : 4
B. Implicit Formulations (a) Interior node with convection The summation of conduction heat flow terms gives pþ1 pþ1 pþ1 pþ1 Tm þ 1;n Tm;n Tm1;n Tm;n Dy k :1 þ kðDy:1Þ 2 Dx Dx pþ1 pþ1 pþ1 pþ1 Tm;n þ 1 Tm;n Tm;n1 Tm;n Dx :1 þk þ kðDx:1Þ 2 Dy Dy Heat flow by convection is h
Dx þ Dy pþ1 Þ :1 ðT1 Tm;n 2
The summation of the conduction and convection terms given above will be equal to the increase in the internal energy of the elemental volume. That is,
6.9 Numerical Method of Solving Transient Conduction Problems
393
pþ1 pþ1 pþ1 pþ1 Tm þ 1;n Tm;n Tm1;n Tm;n Dy :1 þ kðDy:1Þ k 2 Dx Dx pþ1 pþ1 pþ1 pþ1 Tm;n þ 1 Tm;n Tm;n1 Tm;n Dx :1 þk þ kðDx:1Þ 2 Dy Dy pþ1 p Tm;n Tm;n Dx þ Dy Dx:Dy pþ1 þh :1 ðT1 Tm;n Þ ¼ qc Dx:Dy :1 2 4 Ds For Dx = Dy, and introducing the Fourier and Biot numbers, we get the equation of p of the node (m, n) as present temperature Tm;n 4 2 pþ1 Fo Bi þ 4Fo þ 1 Tm;n FoðTmp þþ 11;n 3 3 4 pþ1 pþ1 pþ1 þ 2Tm1;n þ 2Tm;n þ 1 þ Tm;n1 Þ Fo Bi T1 3
p Tm;n ¼
(b) Node at plain surface with convection The heat balance gives Conduction heat inflow + convection heat inflow = increase in internal energy or pþ1 pþ1 pþ1 pþ1 pþ1 pþ1 Tm1;n Tm;n Tm;n1 Tm;n Dx Tm;n þ 1 Tm;n Dx :1 þ kð :1Þ þk 2 2 Dx Dy Dy pþ1 p Tm;n Tm;n Dx:Dy pþ1 þ hðDy:1ÞðT1 Tm;n Þ ¼ qc :1 2 Ds
kðDy:1Þ
Using Dx = Dy,
kDs 2
qcðDxÞ
¼ Fo,
hDx ¼ Bi and simplifying, we get k
pþ1 p pþ1 Tm;n ¼ ð2Fo Bi þ 4Fo þ 1ÞTm;n Foð2Tm1;n pþ1 pþ1 þ Tm;n þ 1 þ Tm;n1 Þ 2Fo Bi T1 ;
which is the desired result. (c) Node at the exterior corner with convection The heat balance equation in this case is pþ1 pþ1 pþ1 pþ1 Tm1;n Tm;n Tm;n1 Tm;n Dy Dx Dx þ Dy pþ1 :1 :1 :1 ðT1 Tm;n þk þh Þ 2 2 2 Dx Dy pþ1 p Tm;n Tm;n Dx:Dy ¼ qc :1 4 Ds
k
For Dx = Dy,
kDs qcðDxÞ2
¼ Fo,
hDx k
¼ Bi, the equation gives
(m+1, n)
(m, n1)
(m, n)
Δy
Δx
m, n 1
m, n
(h, t∞)
m + 1, n
Δy
Δx
m, n 1
4. Node at exterior corner with convection (h, t∞) m1, n m, n ● Δy
m, n 1
p þ ð1 4Fo 4Fo BiÞTm;n The stability condition is Foð1 þ BiÞ 14
p p pþ1 Tm;n ¼ 2FoðTm1;n þ Tm;n1 þ 2Bi T1 Þ
þ ð1 4Fo 2Fo The stability condition is Foð2 þ BiÞ 12 p BiÞTm;n
p p p pþ1 Tm;n ¼ Foð2Tm1;n þ Tm;n þ 1 þ Tm;n1 þ 2Bi T1 Þ
2 p p p pþ1 Tm;n ¼ FoðTmp þ 1;n þ 2Tm1;n þ 2Tm;n þ 1 þ Tm;n1 Þ 3 4 4 p þ Fo BiT1 þ ð1 4Fo Fo BiÞTm;n 3 3 The stability condition is Foð3 þ BiÞ 34
p þ ð1 4FoÞTm;n The stability condition is Fo 14
Explicit formulation p p p pþ1 Tm;n ¼ Fo Tmp þ 1;n þ Tm1;n þ Tm;n þ 1 þ Tm;n1
4 2 pþ1 Fo Bi þ 4Fo þ 1 Tm;n FoðTmp þþ 11;n 3 3 4 pþ1 pþ1 pþ1 þ 2Tm1;n þ 2Tm;n þ 1 þ Tm;n1 Þ Fo Bi T1 3
pþ1 pþ1 4 2FoðTm1;n þ Tm;n1 Þ 4Fo Bi T1
p pþ1 Tm;n ¼ ð1 þ 4Fo þ 4Fo BiÞTm;n
pþ1 pþ1 pþ1 Foð2Tm1;n þ Tm;n þ 1 þ Tm;n1 Þ 2Fo Bi T1
p pþ1 Tm;n ¼ ð2Fo Bi þ 4Fo þ 1ÞTm;n
p Tm;n ¼
p pþ1 Tm;n ¼ ð1 þ 4FoÞTm;n pþ1 pþ1 pþ1 Fo Tmp þþ 11;n þ Tm1;n þ Tm;n þ 1 þ Tm;n1
Implicit formulation
6
Δx
3. Node at plain surface with convection m, n +1 (h, t∞) Δy ●m, n m1, n
m1, n
m, n +1
2. Node at interior corner with convection
Δx
(m1, n)
(m, n+1)
1. Interior node
Conﬁguration
Table 6.4 Summary of transient, twodimensional ﬁnitedifference equations (Dx = Dy)
394 Unsteady or Transient Heat Conduction
6.9 Numerical Method of Solving Transient Conduction Problems
395
pþ1 pþ1 p pþ1 Tm;n ¼ ð1 þ 4Fo þ 4Fo BiÞTm;n 2FoðTm1;n þ Tm;n1 Þ 4Fo Bi T1 ;
which is the desired equation. Summary of transient, twodimensional ﬁnitedifference equations is given in Table 6.4. Example 6.40 Figure 6.31 shows a plain wall with convection heat transfer on both sides. Write nodal equations for nodes on both sides exposed to convection and for a node lying in the wall following the method of backward differences. Solution It is a onedimensional unsteadystate conduction problem. The wall has been divided by equal distance increments of Dx. The nodes under consideration are (1), (n) and (i) as shown in the ﬁgure. The nodal equation can be written by following the procedure discussed in Sect. 6.9 and Example 6.39. (i) Interior surface node (n) The heat balance equation is kA
pþ1 Tn1 Tnp þ 1 ADx Tnp þ 1 Tnp þ hi AðTip þ 1 Tnp þ 1 Þ ¼ qc 2 Dx Ds
Simpliﬁcation of the equation gives pþ1 Tnp þ 2Fo Bi Tip þ 1 ¼ ð2Fo Bi þ 2Fo þ 1ÞTnp þ 1 2Fo Tn1
(ii) Node (i) inside the wall Differential equation of onedimensional heat conduction is @T ¼ @s
k @2T ; qc @x2
Wall h ● n ● n1 Ti Interior side
Fig. 6.31 Example 6.40
●
i ● ● ● ● i+1 i1
●2 ● 1 Te External side
ðiÞ
396
6
Unsteady or Transient Heat Conduction
which gives ! pþ1 Tipþþ11 þ Ti1 2Tip þ 1 Tip þ 1 Tip ¼a : Ds ðDxÞ2 Alternatively, following the method of the heat balance, we have kA
Tipþþ11 Tip þ 1 T p þ 1 Tip þ 1 T p þ 1 Tip þ kA i1 ¼ qcðADxÞ i Dx Dx Ds
or Tip þ 1 Tip k pþ1 pþ1 pþ1 ¼ T þ T 2T i iþ1 i1 Ds qcðDxÞ2 The required equation is
pþ1 Tip ¼ ð2Fo þ 1ÞTip þ 1 Fo Tipþþ11 þ Ti1
ðiiÞ
(iii) Exterior surface node (1) with convection
kA
T2p þ 1 T1p þ 1 ADx T1p þ 1 T1p þ ho AðTep þ 1 T1p þ 1 Þ ¼ qc 2 Dx Ds
Simpliﬁcation gives T1p þ 2Fo Bi Tep þ 1 ¼ ð2Fo Bi þ 2Fo þ 1ÞT1p þ 1 2FoT2p þ 1
ðiiiÞ
For the solution of this problem, the set of implication equations will consist of Eqs. (i), (iii) and equations similar to Eq. (ii) for each node inside the wall. Example 6.41 Write the ﬁnite difference expression for the temperature of a node on a convective boundary of a onedimensional wall which is suddenly subjected to uniform volumetricheating. Solution Figure 6.32 shows the node (m) on the convective boundary. The rate of heat conduction from node (m – 1) to node (m) under consideration
m1 h, t∞ m ●
●
=
Fig. 6.32 Example 6.41
6.9 Numerical Method of Solving Transient Conduction Problems
kðDy:1Þ
397
p Tm1 Tmp Dx
Heat convected from the node (m) is hðDy:1ÞðT1 Tmp Þ Heat generation in the control volume Dx :Dy:1 qg 2 The change in the internal energy of node (m) is pþ1 Dx T Tmp qc :Dy:1 m 2 Ds Heat balance at the node (m) gives pþ1 p Dx Tm Tmp Tm1 Tmp Dx p :Dy:1 :Dy:1 : ¼ kðDy:1Þ þ hðDy:1ÞðT1 Tm Þ þ qg qc 2 2 Ds Dx Using Dx = Dy and rearranging the terms, we get Tmp þ 1 Tmp ¼
2kDs qcðDxÞ
2
p ðTm1 Tmp Þ þ
Ds hDx 2kDs
qg T1 Tmp þ 2 k qcðDxÞ qc
or Tmp þ 1 Tmp ¼
Putting
hDx 2kDs
kDs ðDxÞ2 p p p qg ðT T Þ þ T T þ 1 m m m1 k qcðDxÞ2 qcðDxÞ2 qcðDxÞ2 k 2kDs
kDs qcðDxÞ
2
¼ Fo and
Tmp þ 1 ¼ 2Fo
hDx ¼ Bi and solving for Tmp þ 1 , we obtain k
p Tm1
! ðDxÞ2 qg þ BiT1 þ þ ð1 2Fo 2Bi:Fo)Tmp ; 2k
which is the desired result. For stability, the coefﬁcient of Tmp must be positive, hence ð1 2Fo 2Bi:Fo) 0 or Fo
1 : 2ð1 þ BiÞ
398
6
Unsteady or Transient Heat Conduction
Emissivity ε 2●
● 0
●1
Fig. 6.33 Example 6.42
Example 6.42 Write the ﬁnite difference expression for the temperature of node 0 (Fig. 6.33) of a thin rod of the area of crosssection Ac and electrical resistivity qe, which is suddenly subjected to heating due to passing of current I. The rod having surface emissivity e is located in a large enclosure with vacuum. The enclosure surface temperature is Tsur. Solution Figure 6.33 shows the node (0) along with nodes 1 and 2. The rate of heat conduction from node (1) to node (0) is kAc
T1p T0p : Dx
Similarly the rate of heat conduction from node (2) to node (0) is kAc
T2p T0p Dx
Heat loss by radiation to the surface of the large space is 4 erðPDxÞðTsur T04;p Þ
Heat generated in the control volume due to the flow of current is qe Dx I Ac 2
The change in the internal energy of node (0) is qcðAc DxÞ
T0p þ 1 T0p Ds
Heat balance at the node (0) gives T0p þ 1 T0p T1p T0p T2p T0p 4;p 4 2 qe Dx ¼ kAc þ kAc þ erPDxðTsur T0 Þ þ I qcðAc DxÞ : Ac Ds Dx Dx Rearranging the terms, we get
6.9 Numerical Method of Solving Transient Conduction Problems
T0p þ 1 T0p þ 1
¼ ¼
Dsk qcðDxÞ Dsk qcðDxÞ þ
Putting
" ðT1p 2
þ T2p Þ
2 "
ðT1p 2
kDs qcðDxÞ
kDs qcðDxÞ2
2
:
þ T2p Þ 2
2
I qe ðDxÞ kA2c
Dsk qcðDxÞ Dsk qcðDxÞ
# 1 T0p þ 2 # 1 T0p þ 2
399
2 DserP 4 I qe Ds ðTsur T04;p Þ þ qcAc qcA2c kDs qcðDxÞ
: 2
erPðDxÞ2 4 ðTsur T04;p Þ kAc
2
¼ Fo, we obtain
T0p þ 1 ¼ FoðT1p þ T2p Þ þ ð1 2FoÞT0p þ Fo:
erPðDxÞ2 4 I 2 qe ðDxÞ2 ðTsur T04;p Þ þ Fo: ; kAc kA2c
which is the desired result. For stability, the coefﬁcient of T0p must be positive, hence 1 Fo : 2
6.10
The Schmidt Graphical Method for OneDimensional Problems
In lieu of the numerical calculations, we can employ the Schmidt graphical technique to solve the onedimensional transient problems. Differential equation governing this case is @T @2T ¼a 2 @s @x
ðiÞ
The corresponding ﬁnite difference equation is Tmp þ 1 Tmp ¼
aDs p p Tm þ 1 þ Tm1 2Tmp 2 ðDxÞ
ðiiÞ
If we choose the increments to give 1 2
ð6:39Þ
1 p p Tm þ 1 þ Tm1 2
ð6:40Þ
aDs ðDxÞ
2
¼
then, Tmp þ 1 ¼
From the above equation, it is clear that the temperature at node m after the time increment Ds is the arithmetic mean of the temperatures of the adjacent nodes (m + 1) and (m – 1) at the beginning of the time increment. This arithmetic mean can be done graphically as shown in
400
6
Unsteady or Transient Heat Conduction
T p
Tm +1
x p +1
Tmp−1
Tm
Fig. 6.34 Temperature at node m after the time increment Ds
Fig. 6.34, wherein the value of Tmp þ 1 has been obtained by drawing a line joining the points p Tmp þ 1 and Tm1 . By continued application of the above procedure, the development of the temperature ﬁeld with time can be determined from a known initial temperature distribution. Thus to ﬁnd the temperature distribution after some speciﬁed time, the solid is divided into equal distance increment Dx. Then using equation Ds ¼ ðDxÞ2 =2a from Eq. (6.39), we can ﬁnd the number of time increments necessary by dividing the total time by Ds. It is to be noted that to use this graphical solution the temperature distribution in the solid must be known at some time and the chosen value of the time increment Ds must satisfy Eq. (6.39). The following example illustrates the application of the method when the boundary (surface) temperatures are maintained at constant values. Example 6.43 A very long 60 mm thick metal plate (a = 2 10–4 m2/min) initially at 50° C is immersed in a fluid ðh ¼ 1Þ of temperature 400°C. By the graphical method, determine the temperature distribution in the plate after 1.5 min. Conﬁrm the result by numerical method. Solution Let Dx = 10 mm. The corresponding nodes are marked in Fig. 6.35. The time increment is Ds ¼
1 1 ðDxÞ2 ¼ ð10=1000Þ2 ¼ 0:25 min: 2a 2 2 104
Hence, the number of necessary time increments = 1.5/0.25 = 6. Figure 6.36 presents the graphical solution. It is to note that due to h ¼ 1, the boundary temperatures (at nodes 0 and 6) are maintained at constant values of 400°C throughout the heating process. Start the process by connecting point 0 with 2, 1 with 3, etc. by straight lines. Thus, a new temperature distribution is obtained. Repeat the process. It is to note that the temperature distribution is symmetrical about node 3.
0 ● 1 ● 2● 3 ● 4 ● 5● 6 ●
Fig. 6.35 Example 6.43
6.10
The Schmidt Graphical Method for OneDimensional Problems
401
Fig. 6.36 Graphical solution, Schmidt plot
Table 6.5 Numerical solution Time (min)
Nodes 0
1
2
3
4
5
6
0
400
50
50
50
50
50
400
0.25
400
225
50
50
50
225
400
0.5
400
225
137.5
50
137.5
225
400
0.75
400
268.75
137.5
137.5
137.5
268.75
400
1.0
400
268.75
203.125
137.5
203.125
268.75
400
1.25
400
301.563
203.125
203.125
203.125
301.563
400
1.5
400
301.563
252.343
203.125
252.343
301.563
400
The numerical solution can be obtained by using Eq. (6.40). Thus, the equations are 1 p T0 þ T2p 2 1
¼ T1p þ T3p 2 1
¼ T2p þ T4p 2
T1p þ 1 ¼ T2p þ 1 T3p þ 1
The solution steps are given in Table 6.5. Example 6.44 A semiinﬁnite solid (k = 30 W/(m K), a = 3 10−6 m2/s), initially at a uniform temperature of 300°C, is suddenly exposed to a convection environment (h = 200 W/(m2 K), T∞ = 30°C). Determine the temperature at the surface and at 30 mm depth after 150 s.
402
6
Unsteady or Transient Heat Conduction
Solution (i) Node 0 on the convective boundary The rate of heat conduction from node (1) to node (0), refer Fig. 6.37, kðDy:1Þ
T1p T0p Dx
Heat convected to the node 0 is hðDy:1ÞðT1 T0p Þ The change in internal energy of node 0 is pþ1 Dx T T0p :Dy:1 0 qc 2 Ds Heat balance at the node 0 gives pþ1 Dx T T0p T p T0p :Dy:1 0 ¼ kðDy:1Þ 1 þ hðDy:1ÞðT1 T0p Þ: qc 2 Ds Dx Using Dx = Dy and rearranging the terms, we get 2kDs
T0p þ 1 T0p ¼
qcðDxÞ
2
ðT1p T0p Þ þ hDx
2Ds qcðDxÞ
2
T1 T0p
or T0p þ 1 ¼
Putting
kDs qcðDxÞ
2
2kDs qcðDxÞ
2
ðT1p T0p Þ þ
hDx 2kDs
T1 T0p þ T0p 2 k qcðDxÞ
hDx ¼ Bi and rearranging the terms, we obtain k
¼ Fo and
T0p þ 1 ¼ 2Fo T1p þ BiT1 þ ð1 2Fo 2Bi:Fo)T0p ; which is the desired equation.
1 h, t∞
0●
●
2 ●
3 ●
=
Fig. 6.37 Example 6.44
ð1Þ
6.10
The Schmidt Graphical Method for OneDimensional Problems
403
For stability, the coefﬁcient of T0p must be positive, hence ð1 2Fo 2Bi:Fo) 0 or Fo
1 : 2ð1 þ BiÞ
(ii) Interior nodes (m = 1, 2, 3….) The rate of heat conduction from node 0 to node 1 is kðDy:1Þ
T0p T1p Dx
The rate of heat conduction from node 2 to node 1 kðDy:1Þ
T2p T1p Dx
The change in the internal energy of node 1 is qcðDx:Dy:1Þ
T1p þ 1 T1p Ds
Heat balance at the node 1 gives qcðDx:Dy:1Þ
T1p þ 1 T1p T p T1p T p T1p ¼ kðDy:1Þ 0 þ kðDy:1Þ 2 : Ds Dx Dx
Using Dx = Dy and rearranging the terms, we get T1p þ 1 ¼
Putting
kDs qcðDxÞ2
kDs qcðDxÞ2
ðT0p T1p Þ þ
kDs qcðDxÞ2
ðT2p T1p Þ þ T1p
¼ Fo and rearranging we obtain
T1p þ 1 ¼ Fo T0p þ T2p þ ð1 2Fo)T1p ;
Similarly, equations for node 2 to 4 can be written. For stability, the coefﬁcient of T1p must be positive, hence ð1 2Fo) 0
ð2Þ
404
6
Unsteady or Transient Heat Conduction
or 1 Fo : 2 Assuming Δx = 15 mm, Biot number is hDx 200 0:015 ¼ ¼ 0:1: k 30
Bi ¼
From stability condition for equation of node 0, Fo
1 2ð1 þ BiÞ 1 2ð1 þ 0:1Þ 0:45:
Let s = 30 s, then Fo ¼
as 2
ðDxÞ
¼
3 106 30 ¼ 0:4\0:45: 0:0152
Condition is satisﬁed. Fo = 0.4 also satisﬁes the stability condition for equations of node 1 to 4. For Bi = 0.1, Fo = 0.4 and T∞ = 30, Eq. (1) reduces to
T0p þ 1 ¼ 0:8 T1p þ 3 þ 0:12T0p
ð3Þ
T1p þ 1 ¼ 0:4 T0p þ T2p þ 0:2T1p
ð4Þ
Equation (2) reduces to
Similarly, for nodes 2, 3 and 4, we have
T2p þ 1 ¼ 0:4 T1p þ T3p þ 0:2T2p
T3p þ 1 ¼ 0:4 T2p þ T4p þ 0:2T3p
T4p þ 1 ¼ 0:4 T3p þ T5p þ 0:2T4p
ð5Þ ð6Þ ð7Þ
For 5 time steps, each of 30 s giving total time 150 s, i.e. p = 0 to 5, solution of Eqs. (3)– (7) is presented in the table below. (Table 6.6) From the table, the surface temperature T0 is 261.3°C and the temperature at a depth of 30 mm T2 is 291.9°C after 150 s. The accuracy of the results from the numerical analysis will increase with a decrease in Δs and Δx.
6.10
The Schmidt Graphical Method for OneDimensional Problems
405
Table 6.6 Numerical solution p
T0
T1
T2
T3
T4
T5
300
300
300
300
300
300
30
278.4
300
300
300
300
300
60
275.8
291.4
300
300
300
300
300
Time (s)
0
0
1 2 3
90
268.6
288.6
296.6
300
4
120
265.5
283.8
294.8
298.6
5
150
261.3
280.9
291.9
The problem can also be solved using Fig. 6.14. The required parameters are x 0:00 pﬃﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0 2 as 2 3:0 106 150 and
x 0:03 pﬃﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0:71; 2 as 2 3:0 106 150 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ h as 200 3:0 106 150 ¼ ¼ 0:14: k 30
i From Fig. 6.14, TðxÞT T1 Ti 0:135 for x = 0 and 0.026 for x = 0.03, which gives T (x = 0) = 263.5°C and T(x = 0.3) = 293.0°C for T∞ = 30°C and Ti = 300°C, which are in reasonable agreement.
Example 6.45 The distribution of temperature t across a large concrete wall 500 mm thick, which is heated from one side, is measured by thermocouples inserted in holes in the wall. It has been found that at a certain instant s the temperature can be represented approximately by the equation t ¼ 90 80x þ 16x2 þ 32x3 25:6x4 where x is in meter and temperature in°C. Assuming that the area of the wall is 2.5 m2, estimate (i) Heat entering and leaving the wall in unit time, (ii) Heat energy stored in the wall in unit time, (iii) Temperature change per unit time at surface (x = 0). Take c = 0.84 kJ/(kg K), q = 2175 kg/m3, and k = 2.7 kJ/(m h K) for the concrete. Solution Temperature distribution across the wall is given as t ¼ 90 80x þ 16x2 þ 32x3 25:6x4
406
6
Unsteady or Transient Heat Conduction
Its derivative is dt ¼ 80 þ 32x þ 96x2 102:4x3 dx The second derivative is d2 t ¼ 32 þ 192x 307:2x2 dx2 (i) The heat entering the wall at x = 0, dt qin ¼ kA dx x¼0 ¼ 2:7 2:5 ð80Þ ¼ 540 kJ=h: The heat leaving the wall at x = 0.5 m, qout
dt ¼ kA dx x¼0:5
¼ 2:7 2:5 ð80 þ 32 0:5 þ 96 0:52 102:4 0:53 Þ ¼ 356:4 kJ=hr: (ii) Heat energy stored in the wall = qin–qout = 540–356.4 = 183.6 kJ/h. (iii) Rate of change of temperature for onedimensional heat flow, is given by 2 dt d t ¼a ds dx2 where a = k/qc = 2.7/(2175 0.84) = 1.48 10−3. Substituting the value of d2 t=dx2 , we have
dt ¼ 1:48 103 32 þ 192x 307:2x2 ds At x = 0, dt ¼ 1:48 103 32 ¼ 0:04736 o C=h ds and at x = 0.5,
dt ¼ 1:48 103 32 þ 192 0:5 307:2 0:52 ¼ 0:0758 o C=hr ds
6.11
6.11
Summary
407
Summary
This chapter has been devoted to the problems of transient state of heat conduction, i.e. the heating or cooling where the temperature of the solid body varies with time as well as in the space. Such problems are of interest because they are encountered in many industrial processes. For bodies with a very high thermal conductivity combined with a low value of the convective heat transfer coefﬁcient, uniform temperature throughout the body at any instant can be assumed. Smaller bodies with lower values of the thermal conductivity may also satisfy this condition. The method of lumped heat capacity analysis has been presented for such cases. The analysis shows that the temperature of the body falls or rises exponentially with time and can be presented in terms of two nondimensional numbers Biot number Bi and Fourier number Fo as (t–t∞)/(ti–t∞) = exp(BiFo). The condition for the applicability of the lumped heat capacity analysis is Bi 0.1. In Sect. 6.5, the transient behaviour of semiinﬁnite solids (which is bounded by plane x = 0 and extends to inﬁnity in the positive xdirection and also in y and zdirections) such as a thick or large block of steel, the earth, etc. has been considered when the solid is heated or cooled for a relatively short period. The temperature of such solids changes only for a short distance from the surface. Treatments for three different cases, namely (a) surface temperature of the body is suddenly lowered and maintained at constant temperature, (b) constant heat flux at the surface and (c) the convective boundary condition, have been presented. For determination of temperature variation with time and spatial position in plates (whose thickness is small compared to the other dimensions), cylinders (whose diameter is small compared to its length), and spheres, solution based on Heisler and Grober charts has been given in Sect. 6.7. In Sect. 6.8, product solutions for the temperature distributions in two or threedimensional transient heat conduction systems are presented. In many problems of practical interest, the geometrical shapes do not conform to regular shapes discussed above. It is not possible to provide a mathematical solution to such problems. The ﬁnitedifference method (the numerical technique) can be used to solve all types of transient conduction problems. The method has been presented in Sect. 6.9 followed by a number of illustrative examples. Using the ﬁnitedifference equation, Schmidt gave a graphical method for the solution of onedimensional problems, which is presented in Sect. 6.10.
Review Questions 6:1 What is the main assumption in the lumped heat capacitance analysis? 6:2. Prove that for a body, whose internal resistance is negligible, the time required for cooling or heating can be obtained from the relation t t1 ¼ expðBiFoÞ ti t1 where Bi and Fo are the Biot and Fourier numbers, respectively, Ti is initial temperature of the body and T∞ is the surrounding fluid temperature.
408
6
Unsteady or Transient Heat Conduction
6:3 Deﬁne Biot and Fourier numbers. 6:4 Draw temperature proﬁle for different values of Biot number in a plane wall symmetrically cooled by convection. 6:5 Write nodal equations for a node at the surface of a plane wall and the node lying inside the wall if the wall is suddenly exposed to a convective atmosphere (h, T∞). 6:6 Explain explicit and implicit formulations for transient heat conduction problems. 6:7 What is the Schmidt graphical technique to solve the onedimensional transient problems?
Problems 6:1 Calculate the Biot number for the following. (a) A triangular ﬁn is 25 mm thick at the base and 100 mm long. It is 75 mm wide and is made of aluminium. The average heat transfer coefﬁcient is 50 W/(m2 K) along the sides. (b) A steel sheet [k = 42 W/(m K)], which is in the form of a rectangular parallelepiped (500 mm 100 mm 3 m), is at a uniform temperature of 800°C. It is cooled to a centreline temperature of 200°C. The heat transfer coefﬁcient is 50 W/(m2 K). Can lump capacity analysis be applied? [Ans. (a) V = ½(25 100) 75 = 93750 mm3; As = 2[√(1002 + 12.52) 75] = 15116.7 mm2; L = V/As = 6.20 mm; Bi = hL/k = 50 6.2/(210 1000) = 0.00148. (b) tav = ½ (800 + 200) = 500°C; k = 42 W/(m K); V = 0.5 0.1 3 = 0.15 m3; As = 2(0.1 0.5 + 0.1 3 + 3 0.5) = 3.70 m2; L = V/As = 0.0405 m; Bi = hL/ k = 50 0.0405 /42 = 0.0483. Since both values of Bi are less than 0.1, the lumped heat capacity analysis is applicable.] 6:2 A solid copper ball of mass 50 g is quenched in a water bath at 20°C. It cools from an initial temperature of 520°C to 320°C in 10 s. What will be the temperature of the ball after 20 s? hAs t t1 [Ans. Lumped heat capacity analysis is applicable. Hence, equation ¼ eðqcV sÞ t0 t1 hAs hAs 320 20 t 20 ¼ eðqcV 10Þ ¼ eðqcV 20Þ and . Solution of the gives 520 20 520 20 s¼10 s¼20 equations gives t = 200°C.] 6:3 A small copper ball of 5 mm diameter at 550 K is dropped into an oil bath whose temperature is 250 K. The thermal conductivity k of the copper is 375 W/(m K), density is 9000 kg/m3 and speciﬁc heat is 400 J/(kg K). If the heat transfer coefﬁcient is 200 W/(m2 K), determine the rate of cooling at the beginning of the cooling. hAs s hAs t t1 hAs [Ans. Differentiation of equation ¼ eðqcV sÞ gives ddst ¼ e qcV ti t1 qcV dt hAs ðti t1 Þ. At the beginning of the cooling (s = 0), ¼ ðti t1 Þ, ds qcV s¼0 where As ¼ pd2 and V ¼ pd 3 =6. Substitution gives ddst ¼ 20 K=s.] s¼0
6.11
Summary
409
6:4 Steel balls 10 mm in diameter are annealed by heating to 880°C and then slowly cooling to 100°C in an air environment at 25°C. If the time required for this cooling process is 16 min, estimate the heat transfer coefﬁcient. For steel, k = 40 W/(m K), c = 450 J/(kg K) and q = 7900 kg/m3. [Ans. We assume that the lumped heat capacity analysis is applicable. Then from hAs t t1 t t1 cqV Eq. (6.2) , or h ¼ ln s ¼ ln ; As s cqV ti t1 ti t1
pD3 =6 D t t1 V ; Hence, h ¼ ln ¼ ¼ cqD On substitution, h ¼ As 6s ; 2 pD 6 ti t1 100 25 450 7900 0:01 ln ¼ 15:0W=ðm2 KÞ; For a sphere, char880 25 6 16 60 acteristic length L = D/6 = 0.01/6 = 1/600 m, Biot number = hL/k = 15/ (40 600) = 6.25 10–4 < 0.1, hence the lumped heat capacity analysis is applicable.] 6:5 A very thick concrete wall is initially at a uniform temperature of 20°C. The surface of the wall is suddenly exposed to uniform heat flux of 150 W/m2. Calculate the temperature of the wall at a depth of 110 mm after 24 h. Take a = 2.4 10–3 m2/hr and k = 1.2 W/(m K). x2 0:112 ¼ ¼ [Ans. Equation (6.22) applies; From given data: 3 4as 4 2:4 10 24 x 0:11 0:0525, q0 ¼ 150 W=m2 , erf pﬃﬃﬃﬃﬃ ¼ erf pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0:254, 2 as 2 2:4 103 24 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ k = 1.2 W/(m K), Using Eq. (6.22), t¼ as=p ¼ 0:1354, A = 1 m2, 2 150 0:1354 150 0:11 eð0:0525Þ ½1 0:254 þ 20 ¼ 41:85 C ] 1:2 1 1:2 1 6:6 A water pipeline is to be placed in the soil. The initial temperature of the soil is 15°C and the surface temperature at the night drops rapidly causing the earth’s surface temperature to fall to –20°C. At what depth the pipeline must be located so that the temperature of the water does not drop below 5°C for 10 h period? The thermal diffusivity of the soil is 1.8 10–3 m2/h. t ts x ¼ 0:7143 ¼ erf pﬃﬃﬃﬃﬃ gives x = 0.203 m.] [Ans. ti ts 2 as 6:7 A large block of steel initially at a uniform temperature of 150°C suddenly has its surface temperature lowered to 50°C. What is the amount of heat removed per unit area of the block when the temperature at a depth of 50 mm has dropped to 100°C? Given: a = 1.2 10–5 m2/s, k = 40 W/(m K). x x s ¼ 0:5, pﬃﬃﬃﬃﬃ 0:48, s = 226 s for x = 0.05 m; Heat [Ans. For erf pﬃﬃﬃﬃﬃ ¼ ttt i ts 2 as 2 as rﬃﬃﬃ Q0 s ¼ 1:128kðts ti Þ removed, ¼ 19:58 106 J=m2 .] a A 6:8 If the slab of Problem 6.7 is suddenly exposed to a fluid at 50°C with heat transfer coefﬁcient of 500 W/(m2 K), calculate the time required for the temperature to reach 100°C at the depth of 50 mm.
410
6:9
6:10
6:11
6:12
6:13
6
Unsteady or Transient Heat Conduction
pﬃﬃﬃﬃﬃ h as x and pﬃﬃﬃﬃﬃ. Read temperature [Ans. Assume a trial value of time s and calculate k 2 as ratio (t – ti)/(t∞ – ti) from Fig. 6.14. Vary value of s till the desired temperature ratio of (100–150)/(50–150) = 0.5 is achieved. The time s = 1650 s corresponding to the temperature ratio of 0.5.] Two large blocks of castiron (k = 52 W/(m K), c = 420 J/(kg K) and q = 7260 kg/m3) and copper (k = 360 W/(m K), c = 380 J/(kg K) and q = 8950 kg/m3), initially at the same uniform temperature, are subjected to a sudden change in surface temperature. Compare the heat fluxes for the two blocks. [Ans. Blocks can be treated as semiinﬁnite solids. From Eq. (6.20b), heat flux q00 is qx¼0 ðt i t s Þ k ¼ k pﬃﬃﬃﬃﬃﬃﬃﬃ ; Hence, for a given time and temperature difference, q00 / pﬃﬃﬃ ¼ A pas a pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃ 00 pﬃﬃﬃﬃﬃﬃﬃﬃ 360 8950 380 ðq Þ ð kqcÞ kqc since a = k/qc; This gives 00 Cu ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ Cu ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 2:78. ðq ÞCI ð kqcÞCI 52 7260 420 The heat flux for the copper block is 2.78 times larger than the heat flux for the castiron block.] A large block of steel initially at a uniform temperature of 150°C is suddenly exposed to a fluid at 50°C with heat transfer coefﬁcient of 500 W/(m2 K), calculate temperature at a depth of 50 mm after 1650 s. Given: a = 1.2 10–5 m2/s, k = 40 W/(m K). pﬃﬃﬃﬃﬃﬃﬃﬃﬃ h ðasÞ h2 ðasÞ 5002 ð1:2 105 1650Þ ¼ 1:76; ¼ ¼ 3:0937; [Ans. k2 402 k hx 500 0:05 x 0:05 pﬃﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0:1777; ¼ ¼ 0:625; From k 40 2 as 2 1:2 105 1650 Eq. (6.23), pﬃﬃﬃﬃﬃ t ti x hx h2 as x h as þ 2 1 erf pﬃﬃﬃﬃﬃ þ ¼ 1 erf pﬃﬃﬃﬃﬃ exp ; Subk k k t1 ti 2 as 2 as t 150 ¼ 1 erf ð0:1777Þ expð3:7187Þ ½1 erf ð1:9377Þ; stitution gives 50 150 o t ¼ 95:18 C. Alternatively Fig. 6.14 may be used for approximate solution.] A large block of metal initially at a uniform temperature of 150°C is suddenly has its surface temperature lowered to 50°C. In 225 s, the temperature at a depth of 50 mm has recorded to be 100°C. Determine thermal conductivity of the slab material. Given: q = 7800 kg/m3 and c = 470 J/(kg K). x t ts x ¼ 0:5, pﬃﬃﬃﬃﬃ 0:48, The gives a = 1.2 10–5 m2/s [Ans. For erf pﬃﬃﬃﬃﬃ ¼ ti ts 2 as 2 as for x = 0.05 m and s = 225 s; k = aqc = 1.2 10–5 7800 470 = 44 W/(m K).] A large plate of stainless steel 40 mm thick and initially at a uniform temperature of 100°C is suddenly exposed to a cooling fluid at 30°C with h = 160 W/(m2 K). Calculate the temperature at a depth of 10 mm from one of the faces 225 s after the exposure. How much heat is removed per unit area of the plate? Given k = 16 W/ (m K), q = 7800 kg/m3 and c = 460 J/(kg K). [Ans. Bi = hL/k = 0.2 > 0.1, Heisler charts may be used. Fo = 2.509, 1/Bi = 5, x/ L = 0.5, ho/hi 0.7, h/ho = 0.975; t = 77.78°C, Bi2Fo = 0.10036, Q/Qo 0.38, Qo/ A = qc(2L)hi = 10.046 106 J/m2 and Q/A = 3.82 106 J/m2.] A large slab of aluminium 100 mm thick is originally at a temperature of 500°C. It is suddenly immersed in a liquid bath at 100°C resulting in a heat transfer coefﬁcient of 1200 W/(m2 K). Determine the time elapsed after the immersion when the surface
6.11
Summary
411
temperature will be 300°C. The properties of the aluminium at the given conditions may be taken as q = 2700 kg/m3, c = 900 J/(kg K) and k = 215 W/(m K). [Ans. L ¼ d=2 ¼ 50 mm; Bi = hL/k = 1200 50 10–3/215 = 0.279 > 0.1; Heisler charts may be used; Correction factor from Fig. 6.20 for 1/Bi = 3.6 and x/L = 1.0 is t t1 300 100 ho to t1 þ 100 ¼ 332:6 o C; ¼ 0:86 gives to ¼ ¼ 0.86. Hence, 0:86 to t1 hi ti t1 332:6 100 ¼ 0:58; For ho/hi = 0.58 and 1/Bi = 3.6, Fo 2.1 from Fig. 6.17; ¼ 500 100 as k s 215 s 2 , which gives 2:1 ¼ We have Fo ¼ 2 ¼ or s ¼ 59:3 s.] L qc L 2700 900 0:052 6:14 The wall of a furnace is 150 mm thick and is adequately insulated from outside. It is initially at a uniform temperature of 20°C. The wall is exposed to hot gases at 900°C. Determine the convective heat transfer coefﬁcient if the outer surface reaches a temperature of 725°C in 10 h. Neglect radiation heat exchange. For the furnace wall, k = 1.2 W/(m K), c = 900 J/(kg K) and q = 2500 kg/m3. ho to t1 725 900 ¼ 0:2, ¼ ¼ [Ans. L = 0.15 m as one side is insulated; 20 900 hi ti t1 as k s 1:2 10 3600 Fo ¼ 2 = ¼ 0:85; For Fo = 0.85 and ho/hi = 2 ¼ L qc L 2500 900 0:152
Bik 10 1:2 ¼ ¼ 80 W= m2 K .] 0.2, 1/Bi = 0.1 from Fig. 6.17; h ¼ L 0:15 6:15 A long glass rod of 40 mm diameter (k = 0.78 W/(m K), c = 840 J/(kg K) and q = 2700 kg/m3), initially at 80°C, is cooled in air [t∞ = 30°C, h = 20 W/(m2 K)] for 2145 s. Then it is removed and isolated from the surroundings. Determine the equilibrium temperature it attains. hL 20 0:02 as [Ans. The Biot number, Bi ¼ ¼ ¼ 0:51; Fourier number, Fo ¼ 2 ¼ k 0:78 L k s 0:78 2145 ¼ 1:844; Bi2 Fo ¼ 1:844 0:512 ¼ 0:48; 2 = qc L 2700 840 0:022 tf t1 For Bi = 0.51 and Bi2Fo = 0.48, Q/Qo = 0.7 from Fig. 6.24; QQo ¼ 1 gives ti t1 equilibrium temperature tf = 45°C for t∞ = 30°C and ti = 80°C.] 6:16 A 2.5 mm thick heated plastic sheet (k = 0.2 W/(m K), a = 1 10–7 m2/s) lies on an insulated surface and is cooled by air flow at 20°C. The convection heat transfer coefﬁcient is 100 W/(m2 K). If the surface of the sheet is found to be 40°C after 2 min, determine the initial temperature of the sheet.
Convective Boundary
Fig. 6.38 Problem 6.18
412
6
Unsteady or Transient Heat Conduction
[Ans. The sheet is equivalent to onehalf of a sheet of thickness d (= 2L) with symmetric convection conditions. The characteristic length L = 0.0025 m; Biot number hL 100 0:0025 as 1 107 120 ¼ ¼ 1:25; Fourier number Fo ¼ 2 ¼ Bi ¼ ¼ k 0:2 L 0:00252 1:92; For 1/Bi = 0.8, and Fo = 1.92, ho/hi 0.205 from Fig. 6.17; hL/ho 0.6 for x/ hi h0 hi h0 L = 1 and 1/Bi = 0.8 from Fig. 6.20; hi ¼ hL ¼ ðtL t1 Þ ¼ h0 hL h0 hL ð40 20Þ ¼ 162:6; ti ¼ hi þ t1 ¼ 162:6 þ 20 ¼ 182:6 C.] 0:205 0:6 6:17 A long square section steel bar (50 mm 50 mm) is initially at a uniform temperature of 800°C. It is heat treated by quenching in an oil tank at 80°C. The average heat transfer coefﬁcient is 400 W/(m2 K). If the temperature at the centreline is to be 300° C, determine the time of quenching. a = 1.4 10–5 m2/s and k = 50 W/(m K). [Ans. The problem is a combination of two inﬁnite plates of equal thicknesses. Hence, #2 " h0 h0 h0 t0 t1 300 80 ¼ 0:306; ; From the given data, ¼ ¼ ¼ hi hi 2Lplate hi ti t1 800 80 sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ﬃ h0 h0 1 k 50 ¼ ¼ ¼ 5; From Fig. 6.17, Hence, ¼ ¼ 0:55; Bi hL 400 0:025 hi 2Lplate hi L2 Fo 0:0252 3:3 ¼ ¼ 147:3 s.] a 1:4 105 6:18 Formulate the ﬁnitedifference equation for the node at an interior corner with convection, as shown in Fig. 6.38, using the explicit and implicit methods. Obtain the stability criterion for the explicit method. Assume equal grid sizes in both x and ydirections. [Hint: Refer Example 6.39.] 6:19 A 250 mm thick concrete wall (a = 0.0066 cm2/s) is split into 5 layers. Determine the time interval Ds for explicit formulation. Assume onedimensional heat flow. [Ans. Dx = 250/5 = 50 mm, Fo = ½ = aDs/(Dx)2 gives Ds = 1894 s = 31.6 min.] Fo 3.3, which gives s ¼
References Grober H, Erk S, Grigull U (1961) Fundamentals of heat transfer. McGrawHill Book Co, New York Heisler MP (1947) Temperature charts for induction and constant temperature heating. Trans ASME 69:227– 236 Holman JP (1992) Adapted for SI units by White PRS, heat transfer. McGrawHill Book Co, New York
7
Convective Heat Transfer
7.1
Introduction
Heat conduction is a process in which the heat is transferred through solids and stagnant layers of fluids (liquids and gases) due to the vibration at the molecular level and electronic effect when a temperature gradient exists. In the case of fluids, the molecules have freedom of motion and heat can also be transferred by the movement of the fluid. This complex process of heat transfer is termed as convection. In convection, there is an observable bulk motion of the fluid and hence can occur only in the case of liquids, gases and multiphase mixtures. The convective heat transfer can be classiﬁed into two categories: (i) natural or free convection and (ii) forced convection.
7.1.1 Natural Convective Heat Transfer In this mode of convection, the fluid motion is entirely because of the density differences in the fluid caused due to the local heating in the gravity ﬁeld. The fluid surrounding the heat source becomes less dense due to the heating and hence moves upwards (known as buoyancy effect). The surrounding colder fluid moves to replace it. This colder fluid is then heated and the process continues, forming the convection currents (the movement of the fluid). Thus, the heat transfers from the bottom of the convection cell to the top. Heating of water in a pot over a stove causing movement of water from the bottom to the top is a typical example of the natural convection. In the nature, convection currents are set up in the air as the solar radiation heats land or water.
7.1.2 Forced Convection Heat Transfer If the motion of the fluid is caused by some external means such as a fan, blower, pump or wind, the mode of heat transfer is termed as the forced convection. Cooling water through a heat exchanger, air propelled through a solar air heater duct and wind blowing over a heated surface are a few examples where forced convective heat transfer takes place.
© Springer Nature Singapore Pte Ltd. 2020 R. Karwa, Heat and Mass Transfer, https://doi.org/10.1007/9789811539886_7
413
414
7 Convective Heat Transfer
However, in any forced convection situation, some amount of natural convection is always present. When the natural convection is not negligible, such flows are typically known as mixed convection. The convection is a combined problem of heat and fluid flow. Hence, the rate of heat transfer is affected by the fluid properties such as viscosity, speciﬁc heat, density, thermal conductivity, coefﬁcient of thermal expansion, etc. The analytical solution of a convection problem involves application of the equation of motion, energy and continuity, and the Fourier’s law of heat conduction. The resulting differential equations that govern the convection are complicated and their exact solution can be given only in a few simple problems of steady flow. There are three basic methods of determining the rate of heat transfer between fluid and a solid by convection. It will be shown in the sections to follow that the fluid is at rest in the immediate vicinity of the solid surface due to the viscous effect. Therefore, the heat flow at the wall is by conduction and not by convection. The ﬁrst method makes use of this observation and the heat transfer rate can be calculated from the Fourier’s law q ¼ kA
dt dy y¼0
ð7:1Þ
where (dt/dy)y = 0 is temperature gradient in the fluid at the wall, refer Fig. 7.1, and k is the thermal conductivity of the fluid. The second method is based on analogy between the mechanisms of transfer of fluid momentum to the wall and the transfer of heat by convection. Using the analogy, the rate of heat transfer by convection can be predicted from the measurement of shear stress between the fluid and the wall. The third method is to experimentally determine the heat transfer coefﬁcient, deﬁned by the Newton’s equation, because the determination of the temperature gradient at the wall and its variation over the entire heattransferring surface is very difﬁcult. The Newton’s equation for the convective heat transfer coefﬁcient is q ¼ hA tw tf
ð7:2Þ
According to this equation, the heat transfer is proportional to the area of the surface A and the temperature difference between the wall and the fluid. The factor of proportionality h in the equation is called the convection heat transfer coefﬁcient. It is the quantity of heat transferred in unit time from unit surface area for a unit temperature difference between the wall and the fluid. Studies have shown that the heat transfer coefﬁcient is a function of many parameters.
t∞ Temperature profile
y
(dt/dy)y = 0 tw
Fig. 7.1 Heat flow at wall
Wall
7.1 Introduction
415
The method of dimensional analysis, based on the dynamic similarity, is used to express the experimental results in the form of a correlation of dimensionless numbers (refer Chap. 8). In this chapter, analytical treatment of simple cases of forced and natural convection will be presented followed by the discussion of analogy between fluid friction and heat transfer.
7.2
Flow of Fluid Past a Flat Plate
Consider a simple case of steadystate flow of fluid with freestream velocity U∞ past a plate of sufﬁcient length as shown in Fig. 7.2. The plate is aligned with the flow so that the xaxis coincides with the flow direction. When the fluid meets the leading edge of the plate, the viscous effect causes it to adhere to the surface. Assuming that there is no slip between the fluid and the surface, the fluid velocity is zero at the plate surface. The fluid velocity increases in the ydirection and approaches the freestream velocity U∞ in a layer of thickness d. This layer is termed as hydrodynamic boundary layer. The reason for the drop of velocity through the boundary layer is the viscosity of the fluid. The fluid flow region may be divided into two regions: the boundary layer (region adjacent to the wall), which exhibits the effect of the viscosity and the main flow or freestream region where the viscosity effect can be neglected. The thickness of the boundary layer increases in xdirection as more and more fluid is included in the boundary layer and is retarded by the friction. In order to accommodate the included fluid, the streamlines must diverge away from the plate so that the flow area is increased. Due to the divergence of the streamlines, the velocity vector is no longer parallel to the plate and a small velocity component v perpendicular to the plate will exist. Thus, the flow in the boundary layer is twodimensional.
Eddies
U∞ , t∞
U∞
U∞ Laminar sublayer
δ Parabolic profile Laminar region
Transition region
Turbulent region
h
0
x
Fig. 7.2 Hydrodynamic boundary layer over a flat plate (the boundary layer thickness is exaggerated for clarity)
416
7 Convective Heat Transfer
U∞ , T∞
Edge of velocity boundary layer
T∞
U∞ δ
Edge of thermal boundary layer Temperature profile
δt Tw
x0
Fig. 7.3 Hydrodynamic (velocity) and thermal boundary layers over a flat plate
The flow along the plate surface (xdirection) is divided into laminar (streamline flow— the fluid flows in distinct fluid layers nearly parallel to the surface) and eddying or turbulent flow (random motion of the fluid in all directions). Initially, the boundary layer development is laminar and the boundary layer thickness increases continuously with the increase in distance x along the plate surface starting from zero at the leading edge of the plate. At some critical distance xc from the leading edge, transition from laminar to turbulent layer with a thin laminar sublayer takes place. Experiments with fluid at different velocities have revealed that the transition occurs at U1 x [ 5 105 m
ð7:3Þ
The group (U∞x/t) is a dimensionless number and is called the Reynolds number. The Reynolds number corresponding to the critical distance xc is known as the critical Reynolds number (Rec = U∞xc/t). The value of the critical Reynolds number is affected by the level of the disturbances present in the free stream approaching the plate and also the plate surface roughness (Rec = 3 105− 5 105 depending on the roughness of the plate surface). For strong disturbances, the critical Reynolds number Rec has been reported to be as low as 8 104. With exceptionally free of disturbances, the flow may remain laminar up to Rec 106 or higher (Schlichting 1979). If the plate is heated or cooled to a uniform temperature Tw, starting at a distance xo from the leading edge of the plate, a thermal boundary layer will also develop as shown in Fig. 7.3. It is the region where temperature gradient is present. The temperature of the fluid varies from T = Tw at the plate surface to T = T∞ at the edge of the thermal boundary layer. The thickness of the thermal boundary layer is designated as dt. The concept of the thermal boundary layer is analogous to that of the velocity or hydrodynamic boundary layer. The relationship between the two boundary layers will be discussed later.
7.3
Flow in Tubes
Consider the fluid flow through a tube. The velocity distribution is uniform at the inlet crosssection if the tube inlet is rounded and the fluid comes from a large space, Fig. 7.4. Because of the friction, the velocity diminishes at the wall and increases at the centre of the tube. The flow does not become fully developed at once but at a certain distance from the tube inlet. Thus, the boundary layer gradually builds up until it reaches the centre of the tube. The velocity distribution curve, now, acquires a stable form and does not vary down the tube. The
7.3 Flow in Tubes
417 Parabolic velocity profile
Boundary layer Uniform inlet flow velocity
u Development length
D
Fully developed flow
Variation of heat transfer coefficient
h
x
Fig. 7.4 Laminar flow in a tube (Re < 2300)
flow is said to be fully developed. The distance for the velocity proﬁle to be fully developed is termed as hydrodynamic development length or entrance length. The development of the thermal boundary layer in a fluid which is heated or cooled in a duct is qualitatively similar to that of the hydrodynamic boundary layer. As the fluid flows along the duct, the heated or the cooled layer increases in the thickness until the heat is transferred to or from the fluid in the centre of the duct. The temperature proﬁle now acquires the fully developed form (known as thermally fully developed) if the velocity proﬁle is fully developed. The shapes of the velocity and temperature proﬁles depend on whether the flow is laminar or turbulent. The hydrodynamic development length for the laminar flow in a circular pipe with a uniform velocity proﬁle at the inlet is approximately given by (Kays and Crawford 1980) L 0:05Re D
ð7:4Þ
where Re is the Reynolds number deﬁned as Re ¼
qUm D l
ð7:5Þ
i.e. it is calculated on the basis of the mean velocity Um and the tube diameter D. The mean velocity is calculated by dividing the volumetric flow rate of the fluid by the crosssectional area of the tube. Thus V 1 Um ¼ ¼ A A
ZA udA 0
For turbulent flow in tubes, the hydrodynamic development length is short (entrance region is typically less than 10–15 tube diameters) and it depends on the geometric character of the entrance (Kays and Crawford 1980).
418
7 Convective Heat Transfer
The thermal entry length for the laminar flow in a tube may be approximated from the following relation (Kays and Crawford 1980): L 0:05Re Pr D
ð7:6Þ
where Pr is the Prandtl number of the fluid, which is a fluid property and equals lcp/k. Thus, the thermal entry length is also dependent on the Prandtl number. For example, for air (Pr = 0.7) flowing with Re = 1000, the development length is 35 tube diameters while for oil (Pr = 100), the same will be 2500 even when flowing with Re = 500. Thus, in an oil exchanger, the fully developed temperature proﬁle is rarely attained.
7.3.1 Laminar Flow Through a Tube When the flow through a straight tube is laminar, the fluid particles move along the path parallel to the tube axis with no rotation of the particles. The velocity proﬁle of the fully developed laminar flow is of parabolic form as shown in Fig. 7.4. In the laminar flow, the velocity proﬁle is given by y y 2 u ¼2 ð7:7Þ Umax R R where u is the velocity at a distance y from the wall and Umax is the velocity at the axis of the tube. Volume flow through annulus of radial width dy at radius (R – y) from the centreline, refer Fig. 7.5, is dV ¼ ½2pðR yÞdyu Hence, the volume flow rate is Z V¼
ZR dV ¼
2puðR yÞdy 0
ZR ¼ 0
ZR ¼ 0
y y 2 2p 2 Umax ðR yÞdy R R y2 y3 2pUmax 2y 3 þ 2 dy R R
1 ¼ pR2 Umax : 2
+
Ry
R
Fig. 7.5 Coordinate system for Eq. (7.7)
y
dy
7.3 Flow in Tubes
419
The mean velocity is deﬁned as Um ¼
V V 1 Umax ¼ 2 ¼ pR2 A pR 2 pR2
or
Um Umax
¼ 0:5
ð7:8Þ
i.e. in the case of laminar flow through a tube, the ratio of the mean velocity to the maximum velocity at the centreline is 0.5.
7.3.2 Turbulent Flow Through a Tube If the flow Reynolds number exceeds a certain critical value (Rec 23001), the flow becomes turbulent. The fully developed velocity proﬁle in this case is shown in Fig. 7.6. In the case of turbulent flow, three distinct regions in the flow have been found. These regions are a laminar sublayer in the immediate vicinity of the wall, a buffer layer (not shown in the ﬁgure) and a prominent turbulent core. The velocity changes abruptly near the wall and takes a somewhat blunter proﬁle in the middle of the tube. The velocity proﬁle becomes nearly flat and the velocity distribution becomes more uniform at very high Reynolds number. The turbulent flows are of great importance because of greater heat transfer rates in this region. The ratio of the mean and the maximum velocities is a function of the Reynolds number, i.e.
Um Umax
¼ f ðReÞ
ð7:9Þ
turbulent
When the fluid is in a turbulent state inside the tube, there is always a layer of fluid at the wall in which the flow is laminar. This layer is known as laminar sublayer. The thickness of this layer increases along the development length from d = 0 at the inlet to its maximum value d. The thickness of the laminar sublayer is very small and can be visualized from the relation of Schlichting (1979), deduced in terms of Reynolds number using Blasius correlation of friction factor for fully developed turbulent flow in smooth circular duct, as d 25 ¼ D Re7=8
1
ð7:10Þ
It is to note that the laminartoturbulent transition is not sudden but occurs over a range of Reynolds number. The numerical value of the critical Reynolds number Rec depends on the duct inlet conditions as well as on the surface roughness of the duct. Disturbances such as vibrations on the exterior of the duct wall and flow pulsation also influence the value of the critical Reynolds number. In a circular duct, the lower limit for the critical Reynolds number is accepted to be 2000 below which the flow remains laminar even in the presence of strong disturbances (Potter et al. 2012). The upper limit of Rec is undeﬁned (Bhatti and Shah 1987), for most practical purposes it is taken as 104. The flow in the range 2300 Re < 104 is termed as transition flow. The velocity proﬁle in the transition regime is neither parabolic nor the usual turbulent proﬁle.
420
7 Convective Heat Transfer Turbulent core
Laminar sublayer
Fig. 7.6 Turbulent flow in a tube
From the relation, it can also be seen that the thickness of the laminar sublayer decreases with the increase in the Reynolds number.
7.4
Equation of Continuity
The equation of continuity is based on the law of conservation of mass. Figure 7.7 shows elemental control volume with dimensions dx, dy and dz in the fluid flow ﬁeld, where the mass can flow in or out of the element. Let us consider the mass balance for this element. The mass of the fluid entering the element through the face ABCD, which is perpendicular to the velocity vector u in the xdirection, in time ds is mx ¼ quðdy:dzÞds Mass leaving the element through face EFGH at distance dx in xdirection is mx þ dx ¼ mx þ
@mx dx @x
The net mass leaving the element in xdirection is dmx ¼ mx þ dx mx ¼
@ ðquÞ ðdx:dy:dzÞds @x
y C B
G dy
F D A
dx
x E
dz
H
z
Fig. 7.7 Elemental control volume for deriving equation of continuity
7.4 Equation of Continuity
421
or dmx ¼
@ ðquÞ dv0 ds @x
ðiÞ
where dv0 = dx.dy.dz is the volume of the element. Similarly, the net fluid amounts leaving in y and zdirections with velocity vectors v and w, respectively, are
@ ðqvÞ dv0 ds dmy ¼ @y @ ðqwÞ dv0 ds dmz ¼ @z
ðiiÞ ðiiiÞ
The total quantity of excess fluid mass leaving the element is the sum of expressions (i), (ii) and (iii), i.e.
@ @ @ ðquÞ þ ðqvÞ þ ðqwÞ dv0 ds dm ¼ @x @y @z
ðivÞ
The net outflow from the element is due to the decrease in the density of the fluid within the element and is dm ¼
@q 0 dv ds @s
ðvÞ
From Eqs. (iv) and (v), we get h
i
@q @ @ @ 0 0 @x ðquÞ þ @y ðqvÞ þ @z ðqwÞ dv ds þ @s dv ds @q @ @ @ @x ðquÞ þ @y ðqvÞ þ @z ðqwÞ þ @s ¼ 0;
¼0
ð7:11Þ
which is the differential equation of continuity in general form. For an incompressible fluid (liquids), q is constant hence @u @v @w þ þ ¼ 0: @x @y @z
ð7:12aÞ
For the twodimensional incompressible fluid flow, the equation is @u @v þ ¼ 0: @x @y
ð7:12bÞ
7.4.1 The Displacement and Momentum Thickness The hydrodynamic or momentum boundary layer is the region in which the fluid velocity changes from zero at the plate surface to its freestream velocity. In fact, there is no precise thickness of boundary layer if this deﬁnition is used because, mathematically, the boundary
422
7 Convective Heat Transfer
layer extends indeﬁnitely in the ydirection. However, the boundary layer thickness is taken to be the distance from the plate surface in which most of the velocity change (from zero to 99 per cent of the freestream velocity) takes place. To avoid the ambiguity in deﬁning the boundary layer thickness, two boundary layer thickness parameters are deﬁned. These parameters are known as velocity displacement thickness dvd and momentum displacement thickness dmd. The displacement thickness is a measure of the displacement of the free stream due to the formation of the boundary layer over the plate while the momentum thickness is a measure of the momentum flux displacement caused by the boundary layer. The velocity displacement thickness dvd can be obtained as follows. The mass flow rate though the boundary layer (for the unit width of the plate) is Z1
Z1 q:uðdy:1Þ ¼
0
q:udy 0
where u is the velocity of the fluid at distance y from the plate surface and q is the density of the fluid. For a frictionless (nonviscous) fluid, the velocity of the fluid will be U∞ and the mass flow rate will be Z1 qU1 dy 0
Thus, the loss in the fluid flow rate due to the formation of the boundary layer is Z1 qðU1 uÞdy 0
If we represent this loss in the fluid flow rate by the flow at U∞ passing through a thickness dvd, then R1 ðdvd :1ÞU1 q ¼ qðU1 uÞdy 0 R1 1 Uu1 dy dvd ¼
ð7:13Þ
0
A geometrical interpretation of the velocity displacement is shown in Fig. 7.8. An inﬁnite boundary layer (d ! ∞) has been substituted by a ﬁnite layer of thickness dvd so that the area under the curve would equal the area of rectangle of height dvd and width U∞. The velocity displacement thickness is a measure of the displacement of the main stream resulting from the presence of the flat plate and its boundary (Kays and Crawford 1980). Similarly, the momentum displacement thickness dmd may be determined as follows.
7.4 Equation of Continuity
423 y
U∞
(U∞  u) u dy y
δvd
Plate surface
Fig. 7.8 Graphical interpretation of the velocity displacement thickness
The mass flux through the boundary layer strip of height dy (for the unit width of the plate) is ¼ qudy Momentum of this mass is ¼ qudy:u The momentum of the same mass before entering the boundary layer is ¼ ðqudyÞU1 Thus, the loss of momentum of this fluid is ¼ qudy:ðU1 uÞ Total loss of the momentum in the complete boundary layer Z1 ¼
quðU1 uÞdy 0
If we represent this momentum loss of fluid by the momentum of the fluid passing through a thickness dmd, then 2 dmd :qU1
Z1 ¼
quðU1 uÞdy 0
dmd ¼
1 2 qU1 Z1
dmd ¼ 0
Z1 quðU1 uÞdy 0
u u 1 dy U1 U1
ð7:14Þ
424
7 Convective Heat Transfer
The momentum displacement thickness is a measure of the momentum flux decrement caused by the boundary layer, which is proportional to the drag of the plate according to the momentum theorem (Kays and Crawford 1980). We can deﬁne energy thickness dke as ‘the thickness of flow moving at the freestream velocity and having energy (the kinetic energy) equal to the deﬁciency of energy in the boundary layer’. Kinetic energy of the mass of the fluid passing through the elemental strip of area (dy.1) is ½qu:ðdy:1Þ
1 2 u 2
¼
1 3 qu dy; 2
If the fluid is nonviscous, the fluid mass (dy.u.q) will have freestream velocity U∞. The total energy of the mass passing through the boundary layer will be Z 1 2 qu U1 dy 2 Thus, the loss of kinetic energy due to the boundary layer is 1 2
Z1
2 qu U1 u2 dy
0
If we represent this energy loss by the energy loss of fluid passing through a thickness dke at the freestream velocity. Then 1 3 1 qU dke ¼ 2 1 2 Z1 dke ¼ 0
Z1
2 qu U1 u2 dy
0
" 2 # u u 1 dy U1 U1
ð7:15Þ
Example 7.1 Determine velocity displacement thickness and momentum thickness if the velocity distribution is linear (u/U∞ = y/d, where u = U∞ at y = d). Solution (i) The velocity displacement thickness, when the limit ∞ is replaced by d, is given by
dvd
Zd u ¼ 1 dy U1 0
Now substituting the equation of the velocity proﬁle, we have
7.4 Equation of Continuity
425
dvd ¼
Zd 0
y 1 dy ¼ d
Zd
Zd dy 0
0
ydy d
¼ d=2 (ii) The momentum displacement thickness is given by Zd dmd ¼ 0
u u 1 dy U1 U1
Zd y y d 1 dy ¼ : ¼ d d 6 0
7.4.2 The Enthalpy and Conduction Thickness For a thermal boundary layer, we can deﬁne integral thickness parameter in an unambiguous way. These parameters are enthalpy thickness and conduction thickness. The enthalpy thickness is deﬁned by the relation R1 de ¼
quedy q1 U1 ew 0
where the enthalpy e = c (T – T∞) and ew= cw (Tw – T∞); c is speciﬁc heat. Substituting the values of e and ew, the thickness de can be expressed in terms of temperature for constant property fluids (q = q∞ and c = cw) as follows: R1 de ¼
uðT T1 Þdy U1 ðTw T1 Þ
0
ð7:16Þ
The conduction thickness is deﬁned as dk ¼
kðT w T1 Þ qw
ðiÞ
The convection heat transfer coefﬁcient is given by h¼
qw ðT w T1 Þ
ðiiÞ
Combining Eqs. (i) and (ii), we get dk ¼
k h
Both the thicknesses have been shown in Fig. 7.9.
ð7:17Þ
426
7 Convective Heat Transfer y dy
y U ∞ (Tw  T ∞ )
δe u (T  T ∞ )
δk 0 T ∞
Tw T
Fig. 7.9 Graphical illustrations of the enthalpy and conduction thicknesses of the boundary layer
7.5
Momentum Equation of Laminar Boundary Layer Over a Flat Plate
The momentum equation of the laminar boundary layer over a flat plate can be derived by making a forceandmomentum balance on the elemental control volume located in the boundary layer as shown in Fig. 7.10. The following assumptions are being made to simplify the analysis: 1. The fluid is incompressible and the flow is steady. 2. There are no pressure variations in the direction perpendicular to the flow, i.e. in the ydirection. 3. The viscosity of the fluid is constant. 4. Viscous shear forces in ydirection are negligible and hence neglected. 5. The effect of gravitational forces is negligible and hence it is neglected. Momentum balance in xdirection for the control volume of Fig. 7.10 gives Viscousshear force þ pressure force ¼ rate of momentum transfer in xdirection
Hydrodynamic boundary layer U∞
y
x μ{∂u/∂y +[∂(∂u/∂y)/∂y]Δy}(Δx.1)
ρu2Δy +[∂(ρu2Δy)/∂x]Δx
ρu(Δy.1)u p (Δy.1)
ρvu Δx +[∂(ρvu Δx)/∂y]Δy
[p + (∂p/∂x)Δx](Δy.1)
Δy Δx
ρv(Δx.1)u
μ(∂u/∂y)(Δx.1) Elemental control volume
Unit depth in zdirection
Fig. 7.10 Elemental control volume for force balance on laminar boundary layer
ð7:18Þ
7.5 Momentum Equation of Laminar Boundary Layer …
427
The rate of momentum transfer in xdirection is the product of mass flow rate through a side of the control volume and the xcomponent of velocity at that point. The mass entering the left face of the element for unit depth in zdirection mx ¼ quðDy:1Þ Hence, the momentum entering this face per unit time is ¼ quðDy:1Þu The momentum leaving the right face is @ 2 qu Dy Dx @x @u ¼ qu2 Dy þ 2qu Dx:Dy @x ¼ qu2 Dy þ
Similarly, the mass entering the bottom face is ¼ qvðDx:1Þ Hence, the momentum entering the bottom face is ¼ qvðDx:1Þu The momentum leaving the top face is
@ ðqvu:DxÞ Dy @y @ ðqvuÞ :Dx:Dy ¼ qvu:Dx þ @y ¼ qvu:Dx þ
The pressure force on the left face is ¼ pðDy:1Þ The pressure force on the right face is
@p Dx ðDy:1Þ ¼ pþ @x Net pressure force in xdirection is
@p ¼ pðDy:1Þ p þ Dx ðDy:1Þ @x @p ¼ ðDx:DyÞ @x
428
7 Convective Heat Transfer
The viscous shear force on the bottom face is ¼ l
@u ðDx:1Þ @y
and the viscous shear force on the top face is @u @ @u ¼l þ Dy ðDx:1Þ @y @y @y The net viscous shear force in xdirection is l
@ @u @2u Dy:ðDx:1Þ ¼ l 2 ðDx:DyÞ @y @y @y
Substitution of various terms in Eq. (7.18) gives l
@2u @p ðDx:DyÞ ðDx:DyÞ ¼ @y2 @x @u @ ðqvuÞ :Dx:Dy qvu:Dx qu2 Dy þ 2qu Dx:Dy qu2 Dy þ qvu:Dx þ @x @y
or l
@ 2 u @p @u @ ¼ 2qu þ ð qvu Þ @y2 @x @x @y @u @v @u ¼ 2qu þ qu þ qv @x @y @y @u @v @u @u þ þ qv ¼ qu þ qu @x @y @x @y
From the continuity equation, @u @v þ ¼0 @x @y
ð7:12Þ
Hence, l
@ 2 u @p @u @u ¼ qu þ qv @y2 @x @x @y
or @u @u @ 2 u @p þv q u ¼l 2 @x @y @y @x This is the momentum equation of the laminar boundary layer over a flat plate.
ð7:19Þ
7.5 Momentum Equation of Laminar Boundary Layer …
429
In the absence of any pressure forces, ∂p/∂x = 0 and the equation becomes @u @u @2u þv q u ¼l 2 @x @y @y
ð7:20Þ
7.5.1 Solution of Momentum Equation (Blasius Solution) The momentum and the continuity equations for the laminar flow over a flat plate have been developed as u
@u @u @2u þv ¼m 2 @x @y @y
ð7:20Þ
and @u @v þ ¼ 0; @x @y
ð7:12Þ
respectively. We introduce a parameter w deﬁned as under, called the stream function, which can replace the components u and v of the velocity by a single function and satisﬁes the continuity equation: u ¼ ð@[email protected]Þ and v ¼ ð@[email protected]Þ The boundary conditions of the flow are u ¼ 0 at y ¼ 0 v ¼ 0 at y ¼ 0 u ¼ U1 at x ¼ 0 u ! U1 as y ! 1 and ð@[email protected]Þ ! 0 as y ! 1: It is assumed that the shape of the velocity proﬁles (Fig. 7.11) is geometrically similar at various xpositions starting from the leading edge of the plate. These velocity proﬁles differ in the ydirection and hence the signiﬁcant variable is the distance variable (y/d), in the nondimensional form. Thus, the velocity proﬁle at any xlocation can be expressed as function of (y/d), i.e. u ¼ gðy=dÞ; U1 where d is the thickness of the boundary layer.
ð7:21Þ
430
7 Convective Heat Transfer U∞
U∞
U∞ U∞
y δ x
Fig. 7.11 Development of laminar boundary layer over a flat plate
Except very close to the wall, the velocity u is of the order of the freestream velocity U∞ and the distance y from the wall is of the order of boundary layer thickness d, i.e. u U1 and y d Using the above orderofmagnitude terms, the continuity equation can be written in an approximate form as U1 v þ 0 d x or v U1 d=x Using the order of magnitude for u, v and y in the momentum equation, u
@u @u @2u þv ¼m 2; @x @y @y
ð7:20Þ
we get U1 U1 d U1 U1 U1 þ m x x d d2 Simpliﬁcation gives rﬃﬃﬃﬃﬃﬃﬃﬃ m:x d U1 It can be expressed in the nondimensional form as d x
rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m 1 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ U1 x Rex
ð7:22Þ
where Rex = (U∞x/t) is the local Reynolds number, which is a nondimensional number. The equation indicates the variation of the boundary layer thickness with the freestream velocity, the viscosity of the fluid and distance x along the plate in the fluid flow direction or with the Reynolds number.
7.5 Momentum Equation of Laminar Boundary Layer …
431
Using orderofmagnitude estimate of d, Eq. (7.21) can be expressed as " # y u y ¼g ¼ g pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ gðgÞ U1 d mx=U1
ð7:23Þ
where "
# rﬃﬃﬃﬃﬃﬃﬃﬃ y U1 g ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ y : mx mx=U1
ð7:24Þ
The variable η is called the similarity variable and g(η) is the function we require as a solution. The stream function w, deﬁned earlier, can be expressed in terms of η as Z w¼
Z
Z
udy ¼
U1 gðgÞdy ¼ rﬃﬃﬃﬃﬃﬃﬃﬃ Z mx ¼ U1 gðgÞ dg U1 Z pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ U1 mx gðgÞdg pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ U1 mxf
U1 gðgÞ
dy dg dg ð7:25Þ
where f = R g(η) dη. The velocity components u and v can be expressed in terms of η and f as u¼
@w @w @g ¼ @y @g @y
@ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ @ y ¼ U1 mx f @g @y rﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ @f U1 ¼ U1 mx @g mx @f ¼ U1 @g
rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 mx
Thus, u ¼ U1 f 0 where (∂f/∂η) = f ′ Similarly,
ð7:26Þ
432
7 Convective Heat Transfer
@w @x @ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ U1 mx f @x pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ @ pﬃﬃﬃ x f ¼ U1 m @x pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ @f @ pﬃﬃﬃ þf x ¼ U1 m x @x @x pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ @f @g f þ pﬃﬃﬃ ¼ U1 m x @g @x 2 x " # rﬃﬃﬃﬃﬃﬃﬃﬃ! pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ 0 @ U1 f þ pﬃﬃﬃ y ¼ U 1 m xf @x mx 2 x " # rﬃﬃﬃﬃﬃﬃﬃﬃ! pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ 0 U1 @ 1 f pﬃﬃﬃ þ pﬃﬃﬃ ¼ U 1 m xf y m @x x 2 x " # rﬃﬃﬃﬃﬃﬃﬃﬃ! pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃ 0 U1 1 3=2 f x ¼ U 1 m xf y þ pﬃﬃﬃ 2 m 2 x rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 U1 m 0 ¼ ðf g f Þ 2 x
v¼
ð7:27Þ
Now the derivatives ∂u/∂x, ∂v/∂x and ∂2u/∂y2 can be determined as follows: @u @ ¼ ðU1 f 0 Þ @x @x @f 0 @f 0 @g ¼ U1 ¼ U1 @x @g @x rﬃﬃﬃﬃﬃﬃﬃﬃ! @ U1 y ¼ U1 f 00 @x mx rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 1 3=2 00 x ¼ U1 f y 2 m rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 U1 00 ¼ y f 2x mx ¼ where (∂f ′/∂η) = f′′.
U1 00 g:f 2x
ð7:28Þ
7.5 Momentum Equation of Laminar Boundary Layer …
@u @ ¼ ðU 1 f 0 Þ @y @y @f 0 @f 0 @g ¼ U1 ¼ U1 @y @g @y rﬃﬃﬃﬃﬃﬃﬃﬃ! @ U1 y ¼ U1 f 00 @y mx rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 ¼ U1 f 00 mx @2u @ @u ¼ @y2 @y @y " rﬃﬃﬃﬃﬃﬃﬃﬃ!# @ U1 00 U1 f ¼ @y mx rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 @f 00 ¼ U1 mx @y rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 @f 00 @g ¼ U1 mx @g @y rﬃﬃﬃﬃﬃﬃﬃﬃ! rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 000 @ U1 y ¼ U1 f @y mx mx rﬃﬃﬃﬃﬃﬃﬃﬃ! rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 000 U1 f ¼ U1 mx mx U1 000 ¼ U1 f mx
433
ð7:29Þ
ð7:30Þ
where (∂f ′′/∂η) = f ′′′. Substitution of various terms in the momentum equation, Eq. (7.20), gives U1 f
0
# " rﬃﬃﬃﬃﬃﬃﬃﬃ!# " rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ U1 00 1 U1 m 0 U1 U1 000 00 g:f þ ðf g f Þ U1 f ¼ m U1 f 2 x 2x mx mx
Simpliﬁcation of the above equation gives 1 1 f 0 f 00 g þ f 00 ðf 0 g f Þ ¼ f 000 2 2 or 2f 000 þ ff 00 ¼ 0:
ð7:31Þ
434
7 Convective Heat Transfer
Table 7.1 Physical and similarity coordinates Physical coordinates
Similarity coordinates
u = 0 at y = 0
f ′ = 0 at η = 0
v = 0 at y = 0
f = 0 at η = 0
u ! U∞ as y ! ∞
f ′ = 1 at η ! ∞
(∂u/∂y) ! 0 as y ! ∞
f ′′ = 0 at η ! ∞
where Z f ¼ f ðgÞ ¼
p gðgÞdg; g ¼ y ½U1 =ðmxÞ; f 0 ¼ ð@f [email protected]Þ ¼ u=U1 ; f 00
¼ ð@f 0 [email protected]Þ and f 000 ¼ ð@f 00 [email protected]Þ: This is an ordinary but nonlinear differential equation of third order for which the boundary conditions are given in Table 7.1. Equation (7.31) has been solved numerically for the function f (η). The ﬁrst solution was obtained by Blasius (1908) in the form of convergent series for small values of η and the asymptotic approximation for the larger values of η. The solution is of the form f ðgÞ ¼
ag2 1 a2 g5 11 a3 g8 375 a4 g11 þ þ . . .. . .. . .:: 2 5! 4 8! 8 11! 2!
ð7:32Þ
where a = 0.33206. Differentiation of the equation gives @f ¼ ag . . .. . .. . .: @g @f 0 f 00 ¼ ¼ a . . .. . .. . .: @g ðf 00 Þg¼0 ¼ a ¼ 0:33206 f0 ¼
ð7:33Þ
The results (values of f, f′ and f′′ as function of η) are presented in Table 7.2. The distribution of the axial velocity u in the ydirection and that of the transverse velocity v along yaxis is presented in Fig. 7.12. From the table and the plots, it is evident that at η = 5, the velocity u is practically equal to the freestream velocity U∞ (f ′ = u/U∞ = 0.99). If we assume that the thickness of the boundary layer d corresponds to u/U∞ = 0.99, then at y = d, η = 5.0 and we have g¼
rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 y mx
y¼d
rﬃﬃﬃﬃﬃﬃﬃﬃ U1 ¼d ¼ 5:0 mx
7.5 Momentum Equation of Laminar Boundary Layer …
435
Table 7.2 Values of functions f(η), etc. for a flat plate η = y[U∞/(mx)]1/2
f(η)
f ′ = u/U∞
f ′′
½ (η f ′ − f)
0
0
0
0.33206
0
0.2
0.00664
0.06641
0.33199
0.00332
0.4
0.02656
0.13277
0.33147
0.01327
0.6
0.05974
0.19894
0.33008
0.02981
0.8
0.10611
0.26471
0.32739
0.05283
1.0
0.16557
0.32979
0.32301
0.08211
1.2
0.23795
0.39378
0.31659
0.11729
1.6
0.42032
0.51676
0.29667
0.20325
2.0
0.65003
0.62977
0.26675
0.30476
2.4
0.92230
0.72899
0.22800
0.41364
2.8
1.23099
0.81152
0.18401
0.52063
3.2
1.56911
0.87609
0.13913
0.61719
3.6
1.92954
0.92333
0.09809
0.69722
4.0
2.30576
0.95552
0.06424
0.75816
4.4
2.69238
0.97587
0.03897
0.80072
4.8
3.08534
0.98779
0.02187
0.82803
5.0
3.28329
0.99155
0.01591
0.83723
5.2
3.48189
0.99425
0.01134
0.84410
5.6
3.88031
0.99748
0.00543
0.85279
6.0
4.27964
0.99898
0.00240
0.85712
6.4
4.67938
0.99961
0.00098
0.85906
6.8
5.07928
0.99987
0.00037
0.85992
7.2
5.47025
0.99996
0.00013
0.86013
7.6
5.87924
0.99999
0.00004
0.86034
8.0
6.27923
1.00000
0.00001
0.86038
8.4
6.67923
1.00000
0.00000
0.86038
8.8
7.07923
1.00000
0.00000
0.86038
or rﬃﬃﬃﬃﬃﬃﬃﬃ mx d ¼ 5:0 U1 or d ¼ 5:0 x
rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m U1 x
ð7:34Þ
436
7 Convective Heat Transfer
0.8604
f′=
v
u
U
U
Re x
= ½ (η f ′ f )
η
Fig. 7.12 Velocity proﬁles in laminar boundary layer
or d 5:0 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ x Rex The skin friction coefﬁcient is deﬁned as sw 2 qU 1 2
Cfx ¼ 1 The shear stress s at the wall is
@u sw ¼ l @y y¼0 where
@u @y
@ u U1 @y U 1 y¼0 0 @f @g ¼ :U1 @g g¼0 @y rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 ¼ ðf 00 Þg¼0 :U1 mx rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 :U1 ¼a mx rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 ¼ 0:33206 :U1 mx ¼
y¼0
ð7:35Þ
7.5 Momentum Equation of Laminar Boundary Layer …
437
Hence, "
# rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 sw ¼ l 0:33206 :U1 mx
ð7:36Þ
Thus, "
# rﬃﬃﬃﬃﬃﬃﬃﬃ! U1 1 0:6641 :U1 1 2 ¼ qﬃﬃﬃﬃﬃﬃﬃ Cfx ¼ l 0:33206 mx U1 x 2 qU1 m
or 0:6641 Cfx ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ Rex
ð7:37Þ
The average value of the skin friction coefﬁcient for the plate length L can be determined by integrating the local value of the coefﬁcient from x = 0 to x = L, and then dividing by the plate length L, i.e. 1 Cf ¼ L ¼
1 L
1 ¼ L
ZL Cfx dx 0
ZL 0
ZL 0
0:6641 pﬃﬃﬃﬃﬃﬃﬃﬃ dx Rex rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ m dx 0:6641 U1 x
rﬃﬃﬃﬃﬃﬃﬃﬃ ZL 1 m 1 pﬃﬃﬃ dx ¼ 0:6641 L U1 x 0 rﬃﬃﬃﬃﬃﬃﬃﬃ 1:3282 m pﬃﬃﬃ 1:3282 ¼ : L ¼ qﬃﬃﬃﬃﬃﬃﬃ L U1 U1 L m
or 1:3282 Cf ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ ReL
ð7:38Þ
where ReL is based on the total length L of the plate in direction of flow. Equations (7.35)–(7.38), valid for the laminar region, indicate that (i) the boundary layer thickness d increases as square root of the distance x from the leading edge and inversely as square root of the freestream velocity U∞,
438
7 Convective Heat Transfer
(ii) the wall shear stress sw is inversely proportional to the square root of the distance x and directly proportional to 3/2 power of the freestream velocity U∞, (iii) the local and average skin friction coefﬁcient Cfx and Cf , respectively, vary inversely as square root of both distance and freestream velocity. Example 7.2 A fluid (kinematic viscosity m = 2.8 10−3 m2/s) flows over a flat plate at 15 m/s. Determine the velocity components u and v at a point located at x = 2 m, y = 40 mm from the leading edge. Solution Flow Reynolds number, Rex ¼
U1 x 15 2 ¼ ¼ 10714 m 2:8 103
Boundary layer thickness at x = 2 m, 5x 52 d ¼ pﬃﬃﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0:0966 [ y Rex 10714 i.e. the point lies within the boundary layer. Stretching factor, rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃ U1 15 g¼y ¼¼ 0:04 ¼ 2:07 2:8 103 2 mx From Table 7.2, we have (corresponding to η = 2.07) f0 ¼
u ¼ 0:63 U1
or u ¼ 0:63 U1 ¼ 0:63 15 ¼ 9:45 m=s; and 1 0 ðgf f Þ ¼ 0:305 2 which gives 1 v¼ 2
rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ U1 m 15 2:8 103 0 0:305 ¼ 0:044 m=s: ðgf f Þ ¼ x 2
Note: It can be seen that v u.
7.6 Integral Momentum Equation of Laminar …
7.6
439
Integral Momentum Equation of Laminar Boundary Layer Over a Flat Plate: von Karman Solution
Figure 7.13 shows the boundary layer flow system. The freestream velocity outside the boundary layer is U∞. The thickness of the boundary layer is d. Consider a control volume at distance x from the leading edge. The control volume is inﬁnitesimal in the xdirection while its height H is such that it is enclosing the boundary layer, i.e. H > d. The mass entering the control volume through plane AA’ for the unit width of the plate is ZH m¼
qudy 0
The mass leaving the control volume through plane BB’ is @m dx ¼ mþ @x
ZH 0
0 H 1 Z @ @ qudy þ qudyAdx @x 0
As no mass can enter the control volume through the solid wall (plane AB), the mass enters the control volume through face A’B’ with the freestream velocity U∞. The momentum (mass velocity) of the mass entering the control volume through the plane AA’ is ZH
ZH qudy:u ¼
0
qdy:u2
0
Similarly, the momentum of the mass leaving the control volume through the planes BB’ and A’B’ is ZH 0
0 H 1 Z @ @ qu2 dyAdx qu2 dy þ @x 0
Control volume B
A U∞
U∞
H y
dy u
δ
x
A
dx
B
Plate
Fig. 7.13 Control volume for integral momentum analysis of laminar boundary layer over a flat plate
440
7 Convective Heat Transfer
and 2
0 H 1 3 Z @ 4 @ qudyAdx5U1 ; @x 0
respectively. If we neglect gravity forces, the shear or drag force at the plate surface (for constant pressure condition) must equal the net momentum change for the control volume. Thus ZH sw dx ¼ 0
qu2 dy
8 H d, i.e. from d to H, the integrand will be zero for y = d to H. Hence, the upper limit of the integrand can be changed to d, which yields 2 d 3 2 d 3 Z Z @ @ sw ¼ U1 4 qudy5 4 qu2 dy5 @x @x 0
0
or 2 d 3 Z @ 4 sw ¼ qðU 1 uÞudy5 @x 0
or 2 d 3 Z @ u u 2 4 sw ¼ qU1 1 dy5 @x U1 U1 0
As the velocity component normal to the plate v is neglected, the equation can be written as 2 d 3 Z d u u 2 4 sw ¼ qU1 1 dy5 dx U1 U1 0
ð7:39Þ
7.6 Integral Momentum Equation of Laminar …
441
The above equation is known as von Karman momentum integral equation for the hydrodynamic boundary layer. It can be used to obtain the expression for the boundary layer thickness if the velocity proﬁle is known. One can represent the unknown velocity proﬁle equation as a polynomial, y y 2 y 3 y 4 u ¼ C1 þ C2 þ C4 þ C5 þ . . .. . .. . .: þ C3 U1 d d d d
ð7:40Þ
The selection of the number of terms depends on the degree of accuracy desired. The constants are determined by application of the conditions at the wall (y = 0) and at the outer limit of the boundary layer (y = d). At the plate surface (y = 0), u = 0. The ﬁrst derivative of u with respect to y at y = 0 determines the viscous shear stress at the wall. The momentum equation of the boundary layer for the zero pressure gradient is u
@u @u @2u þv ¼m 2 @x @y @y
ð7:20Þ
Thus, at the plate surface, where u = v = 0,
@2u @y2
¼0 y¼0
Differentiation of the equation of motion and its evaluation at the plate surface (y = 0) will show that all the higher derivatives of u are also zero at the plate surface. At the outer limit of the boundary layer (y = d), u = U∞. For a smooth transition from the boundary layer to the potential region, all the derivatives of u with respect to y will vanish. These conditions at y = 0 and y = d are summarized below: u/U∞ = 1 at y = δ u/U∞ = 0 at y = 0 ∂u/∂y = 0 at y = δ ∂2u/∂y2 = 0 at y = 0 ∂2u/∂y2 = 0 at y = δ ∂3u/∂y3 = 0 at y = 0 ∂3u/∂y3 = 0 at y = δ
First degree
Second degree
Third degree
(Note: The ﬁrst derivative of u with respect to y at y = 0 determines the viscous shear stress there. Hence, ∂u/∂y at y = 0 may not be speciﬁed.) The number of boundary conditions is selected depending on the requirement of the degree of polynomials. The use of the polynomial as an approximation of the velocity proﬁle means that all the boundary conditions cannot be satisﬁed. Appropriate way is to satisfy as many conditions at the plate surface as at the outer limit of the boundary layer alternating between one and the other as we use higher degrees of polynomial (Chapman 1960). Consider as an example, a third degree polynomial y y 2 y 3 u ¼ C1 þ C2 þ C4 þ C3 U1 d d d
ð7:41Þ
442
7 Convective Heat Transfer
The four constants C1 to C4 can be determined from the four boundary conditions which the velocity function must satisfy. These conditions are u=U1 ¼ 1 at y ¼ d u=U1 ¼ 0 at y ¼ 0 @[email protected] ¼ 0 at y ¼ d @ 2 [email protected] ¼ 0 at y ¼ 0 for constant pressure condition: Applying the boundary conditions, we get C1 ¼ 0 C2 þ C3 þ C4 ¼ 1 C2 þ 2C3 þ 3C4 ¼ 0 2C3 ¼ 0 The values of the constants from the above equations are C1 ¼ 0; C2 ¼ 3=2; C3 ¼ 0; and C4 ¼ 1=2 This gives the equation of the velocity proﬁle as u 3 y 1 y 3 ¼ U1 2 d 2 d
ð7:42Þ
Substitution in Eq. (7.39) gives 8 d 9 1.0, i.e. cooling
µw/µb < 1.0, i.e. heating
m
m
n
n
1
0.12
−0.19
0.092
−0.20
3
0.087
−0.21
0.063
−0.27
10
0.052
−0.22
0.028
−0.36
30
0.029
−0.21
0.000
−0.39
100 1000
0.007
−0.20
−0.04
−0.42
−0.018
−0.20
−0.12
−0.46
a Kays WM, Perkins HC, Forced convection, internal flow in ducts. In: Rohsenow WM, Hartnett JP (eds) Handbook of heat transfer, Chap. 7. McGrawHill, New York. Copyright 1973. The material is reproduced with permission of McGrawHill Education
8.6 Effects of Temperature Varying Properties
601
and n ¼ 0:0 and m ¼ 0:81 0:5\Tw =Tb \1ðfor coolingÞ The values for the turbulent flow of a gas are Tw =Tb [ 1:0 n ¼ 0:55; Tw =Tb \1:0 n ¼ 0:0;
m ¼ 0:1 m ¼ 0:1
Information on the flow through noncircular tubes with variable properties is not complete and hence it is recommended that the circular tube results be used.
8.7
Heat Transfer and Friction in Concentric Circular Tube Annuli and Parallel Plate Duct
Hydraulic or Equivalent Diameter Concept It has been found that nearly the same turbulence intensity and the friction factor prevail in circular and other duct geometries (such as annular, rectangular, square, triangular and irregular passages) if the ratio of flowpassage area to the wetted parameter is kept constant. This ratio is called the hydraulic radius Rh. For a circular passage, this becomes Rh ¼
ðp=4ÞD2 D ¼ 4 pD
It is often convenient to use the term equivalent diameter Dh to signify the diameter of a circular passage, which would have the same hydraulic radius as the passage geometry. Thus Dh ¼
4A P
where A is the flow passage area and P is the wetted parameter. The hydraulic diameter for a circular passage is equal to D. This deﬁnition has been used throughout this book. It must be noted that the use of hydraulic diameter for ducts with sharp corners (e.g., triangular ducts) may lead to error of the order of 35% in turbulent flow friction and heat transfer coefﬁcients determined from the circular duct correlations (Bhatti and Shah 1987).
8.7.1 Laminar Flow Lundgren et al. (1964) determined the pressure drop due to flow development in the entrance region of ducts of arbitrary crosssection. Some data of fully developed Fanning friction factor for laminar flow in concentric annular duct from their results are presented in Table 8.6 along with that for a circular tube. Natarajan and Lakshmanan (in Ebadian and Dong 1998) have presented a simple equation for fRe [fRe = 24(r*)0.035 for 0 < r* 1], which agrees well with the data of Table 8.6. For the friction coefﬁcient in the hydrodynamic entry length for the concentric tube annuli including parallel plate duct, refer Heaton et al. (1964).
602
8
Empirical Relations for Forced Convection Heat Transfer
Table 8.6 Fully developed friction factors for laminar flow in circular tube annuli r* = ri/ro
fRe
0.0 (circular tube)
16
0.01
20.03
0.05
21.57
0.10
22.34
0.20
23.09
0.40
23.68
0.60
23.90
0.80
23.98
1.00 (parallel plate duct)
24.00
Table 8.7 Fully developed Nusselt numbers for laminar flow in circular tube annuli; constant heat ratea r* = ri/ro
Nuii
Nuoo
hi
ho
0.0
∞
4.364
∞
0
0.05
17.81
4.792
2.18
0.0294
0.10
11.91
4.834
1.383
0.0562
0.20
8.499
4.883
0.905
0.1041
0.40
6.583
4.979
0.603
0.1823
0.60
5.912
5.099
0.473
0.2455
0.80
5.58
5.24
0.401
0.299
1.00
5.385
5.385
0.346
0.346
Kays WM, Perkins HC, Forced convection, internal flow in ducts. In: Rohsenow WM, Hartnett JP (eds) Handbook of heat transfer, Chap. 7. McGrawHill, New York. Copyright 1973. The material is reproduced with permission of McGrawHill Education a
In the laminar flow, the fully developed Nusselt number is independent of the Reynolds number and Prandtl number. For the concentric tube flow, there are two Nusselt numbers of interest, one for inner surface (Nuii when inner surface alone is heated) and other for the outer surface (Nuoo when outer surface alone is heated). Their values for constant heat rate condition are given in Table 8.7. For the case of constant heat rate, it is possible to determine Nui and Nuo (the Nusselt number on the inner and outer surfaces, respectively) for any heat flux ratio ðq00o =q00i Þ on the two surfaces, in terms of Nuii and Nuoo, and a pair of influence coefﬁcients hi and ho (refer Table 8.7) using the following equations (Kays and Perkins 1973). Nui ¼
Nuii 1 ðq00o =q00i Þhi
ð8:26aÞ
Nuo ¼
Nuoo 1 ðq00i =q00o Þho
ð8:26bÞ
8.7 Heat Transfer and Friction in Concentric Circular Tube …
603
where Nui ¼
hi dh k
Nuo ¼
ho dh k
q00i ¼ hi ðti tb Þ q00o ¼ ho ðto tb Þ For the thermal entry length data of Nusselt number, readers can refer Kays and Crawford (1980). For the case of constant surface temperature on one surface (with the other surface insulated), the Nusselt number has also been computed by Lundberg et al. (in Kays and Perkins, 1973). The data are presented in Table 8.8. In the case of parallel plates (r* = 1.0), Nu = 7.541 when both surfaces are at the same constant temperature, while it is 8.235 for identical heat rates on the two surfaces of the duct. For the thermal entry length local Nusselt number data for constant and equal wall temperatures and heat rate refer Kays and Perkins (1973). Edwards et al. (in Mills 1995) presented the average Nusselt number relation for the thermal entrance region flow between isothermal plates as Nu ¼ 7:54 þ
0:03ðDh =LÞRe Pr 1 þ 0:016½ðDh =LÞRe Pr2=3
for Re\2800
ð8:27Þ
where the hydraulic diameter is twice the spacing of the plates. Critical Reynolds number of 2800 for transition is appropriate for flow between parallel plates (Mills 1995). For combined thermal and hydrodynamic entry length Nusselt number data relating to the circular tube annulus family (including parallel plates duct) with constant heat rate, refer Kays and Crawford (1980). Example 8.12 Water at a mean bulk temperature of 20°C flows through the annular region formed by two concentric circular tubes. The outer tube has inner diameter of 25 mm, while the inner tube has outer diameter of 12.5 mm. The mass flow rate is 0.015 kg/s. The outer Table 8.8 Fully developed Nusselt numbers for laminar flow in circular tube annuli; constant surface temperature on one surface, the other insulateda r* = ri/ro
Nui
Nuo
0.0
∞
3.66
0.05
17.46
4.06
0.10
11.56
4.11
0.25
7.37
4.23
0.50
5.74
4.43
1.00
4.86
4.86
Kays WM, Perkins HC, Forced convection , internal flow in ducts. In: Rohsenow WM, Hartnett JP (eds) Handbook of heat transfer, Chap. 7. McGrawHill, New York. Copyright 1973. The material is reproduced with permission of McGrawHill Education a
604
8
Empirical Relations for Forced Convection Heat Transfer
surface is insulated and the inner surface is kept at a constant temperature of 50°C. Determine the friction factor and average heat transfer coefﬁcient from the outer surface of the inner tube for the fully developed temperature and velocity proﬁles. Solution At mean bulk temperature of 20°C, the water properties are: q = 998.2 kg/m3, k = 0.601 W/(m K), l = 10.02 10−4 kg/(m s), Pr = 7.0, cp = 4183 J/(kg K). Hydraulic diameter, Dh ¼
4A 4ðp=4ÞðD2i do2 Þ ¼ ¼ ðDi do Þ ¼ 25 12:5 ¼ 12:5 mm: P pðDi þ do Þ
Flow Reynolds number, qUDh m ðDi do Þ ¼ l l ðp=4ÞðD2i do2 Þ 4m 1 4 0:015 1000 1 ¼ ¼ ¼ 508: pðDi þ do Þ l p ð25 þ 12:5Þ 10:02 104
Re ¼
The flow is laminar. Nusselt number Nuii for (Di/do) = 0.5, from Table 8.8 is 5.74. Thus the heat transfer coefﬁcient is hi ¼
Nuk 5:74 0:601 ¼ 276 W/(m2 K): ¼ Dh 12:5=1000
Friction factor, from Table 8.6, is f ¼
23:78 23:78 ¼ ¼ 0:0468: Re 508
Example 8.13 Water at a mean bulk temperature of 20°C and 0.015 kg/s enters the annular region formed by two concentric circular tubes and is heated to 60°C. The outer tube has inner diameter of 25 mm, while the inner tube has outer diameter of 12.5 mm. The mass flow rate is 0.015 kg/s. The outer surface is insulated and the inner surface is subjected to uniform heat rate of 1000 W/m. Determine the tube length required to achieve desired outlet temperature and inner tube surface temperature at outlet. Solution (a) Tube length required to achieve desired outlet temperature At mean bulk temperature of (20 + 60)/2 = 40°C, the water properties are: q ¼ 992:2 kg=m3 ; c ¼ 4179 J=ðkg KÞ
8.7 Heat Transfer and Friction in Concentric Circular Tube …
605
The heat rate for length L is q ¼ q0 L ¼ 1000L W where q’ is heat rate per m. From ﬁrst law equation, q ¼ mcðto ti Þ ¼ 0:015 4179 ð60 20Þ Equating the above equations, we have L¼
0:015 4179 ð60 20Þ ¼ 2:5 m: 1000
(b) Surface temperature at the outlet, At 60°C, the water properties are: k = 0.654 W/(m K), l = 463 10−6 kg/(m s), Pr = 3. Hydraulic diameter, Dh ¼
4A 4ðp=4ÞðD2i do2 Þ ¼ ¼ ðDi do Þ ¼ 25 12:5 ¼ 12:5 mm: P pðDi þ do Þ
Flow Reynolds number, qUDh m ðDi do Þ ¼ 2 2 l l ðp=4ÞðDi do Þ 4m 1 4 0:015 1000 1 ¼ ¼ ¼ 1100: pðDi þ do Þ l p ð25 þ 12:5Þ 463 106
Re ¼
Hence, the flow is laminar, and for (Di/do) = 0.5, it follows from Table 8.7 that the fully developed Nusselt number Nui = Nuii = 6.25 for outer tube insulated, refer Eq. (8.26a). Thus the heat transfer coefﬁcient is hi ¼
Nuk 6:25 0:654 ¼ 327 W/(m2 K): ¼ Dh 12:5=1000
At the outlet, q ¼ q0 ¼ hi ðpdo Þðtso to Þ L or tso ¼ to þ
q0 1000 ¼ 137:9o C: ¼ 60 þ 327 p 12:5=1000 hi ðpdo Þ
606
8
Empirical Relations for Forced Convection Heat Transfer
Since tube surface temperature at the outlet tso > 100°C, local boiling at the tube surface will occur if water is not pressurized, which will increase the local heat transfer coefﬁcient and hence reduce the wall temperature. The length to hydraulic diameter ratio is 2.5/0.0125 = 200. Hydrodynamic development length is (using tube equation) Lhy 0:05Re ¼ 0:05 1100 ¼ 55; Dh and the thermal entry length is Lth 0:05Re Pr ¼ 0:05 1100 3 ¼ 165: D At the outlet, the flow is fully developed. Example 8.14 Air at 25°C and a pressure of 1 atm flows through a parallel plate duct with an average velocity of 0.3 m/s. The duct depth is 5 mm. For fully developed velocity and temperature proﬁles, what is heat transfer coefﬁcient? (a) If both the plates are held at a constant temperature, (b) If identical heat rates are imposed on the two surfaces of the duct. Solution The air properties at 25°C are: q ¼ 1:1868 kg=m3 ; k ¼ 0:02608 W=ðm KÞand l ¼ 1:8363 105 kg=ðm sÞ: The hydraulic diameter of the parallel plate duct is Dh ¼ 2H ¼ 0:01m: Flow Reynolds number, Re ¼
qUDh 1:1868 0:3 0:01 ¼ ¼ 194: 1:8363 105 l
Flow is laminar. In the case of parallel plates, Nu = 7.541 when both surfaces are at the same constant temperature, while it is 8.235 for identical heat rates on the two surfaces of the duct. (a) Heat transfer coefﬁcient (both sides at uniform constant temperature), h¼
Nuk 7:541 0:02608 ¼ ¼ 19:67: D 0:01
8.7 Heat Transfer and Friction in Concentric Circular Tube …
607
(b) Heat transfer coefﬁcient (identical heat rates on the two surfaces of the duct), h¼
Nuk 8:241 0:02608 ¼ ¼ 21:49: D 0:01
Example 8.15 Air at 15°C and a pressure of 1 atm flows through a 200 mm long parallel plate duct whose one side is insulated and the other side is maintained at a constant temperature of 70°C. The duct depth is 5 mm. The pressure drop is recorded to be 1 N/m2. Determine the outlet temperature of the air. Solution The air properties at 25°C (assumed mean bulk temperature) and 1 atm are: q ¼ 1:1868 kg=m3 ; c ¼ 1005:7 J=ðkg KÞ; l ¼ 1:8363 105 kg=ðm sÞ and k ¼ 0:02608 W=ðm KÞ: The hydraulic diameter of the parallel plate duct is Dh ¼ 2H ¼ 0:01 m: Flow Reynolds number, Re ¼
qUm Dh 1:1868 0:01 ¼ Um ¼ 646:3Um : 1:8363 105 l
Assuming flow to be laminar, the friction factor is given by (Table 8.10) f ¼
24 24 0:0371 ¼ ¼ Re 646:3Um Um
Pressure drop is given by Dp ¼
4fLqUm2 2Dh
or 1¼
4 0:0371 0:2 1:1868 Um2 Um 2 0:01
or Um ¼ 0:57 m=s:
608
8
Empirical Relations for Forced Convection Heat Transfer
Hence, Re ¼ 646:3 0:57 ¼ 368: The flow is laminar. From Table 8.10, the Nusselt number for given boundary conditions is 4.861. The heat transfer coefﬁcient is h¼
Nuk 4:861 0:02608 ¼ 12:68 W= m2 K : ¼ Dh 0:01
For isothermal surface, the air outlet temperature is given by tw to P Lh ¼ exp mc tw ti or
P Lh to ¼ tw ðtw ti Þ exp mc The mass flow rate is m ¼ qAc Um ¼ qðwaÞUm ¼ 1:1868 w 0:005 0:57 ¼ 0:00338w
where a is the depth and w (>> a) is width of the duct. Substitution gives w 0:2 12:68 ¼ 43:9 C to ¼ 70 ð70 15Þ exp 0:00338w 1005:7 where P = w. Mean bulk temperature of air is (15 + 43.9)/2 = 29.45°C against assumed temperature of 25°C for the estimate of thermophysical properties. For greater accuracy, retrial may be done.
8.7.2 Turbulent Flow Brighton and Jones (in Kays and Perkins 1973), from their data (0.0625 < r* < 0.562), suggest that the friction factor in circular tube annuli is independent of the radius ratio r*. The friction factors, as determined with water, were found by them to be 6−8% higher than the generally accepted values of smooth circular tube for 4000 < Re < 17000. The same for air flow were found to be about 1–10% higher than circular tube flow values. Kays and Perkins (1973) recommend the following approximate relation, which gives the friction factor in circular annuli about 10% higher than pipe flow values and having little dependence on the radius ratio: f ¼ 0:085Re0:25 : for 6000\Re\300; 000:
ð8:28Þ
8.7 Heat Transfer and Friction in Concentric Circular Tube …
609
However, Bhatti and Shah (1987) recommend the following correlation, which considers the dependence of the friction factor on the radius ratio r* (= ri/ro), f ¼ ð1 þ 0:0925r Þfc
ð8:29Þ
where fc is the friction factor for a circular duct using hydraulic diameter of the annular duct. They report that the predicted friction factors from Eq. (8.29) are within ± 5% of the experimental results. The smooth circular duct friction factor values in Eq. (8.29) may be calculated from the following correlation (Bhatti and Shah 1987): fc ¼ 0:00128 þ 0:1143Re0:311 : for 4000 Re 107 :
ð8:30Þ
Dalle Donne and Meerwald (1973) suggest the following relation for the fully developed isothermal friction factor: f ¼ 0:0014 þ 0:125Re0:32 : for 8000\Re\2 105
ð8:31Þ
` For fully developed flow in parallel plate duct, Beavers et al. (in Ebadian and Dong 1998) obtained following friction factor result from very accurate experimental data: f ¼ 0:1268Re0:3 : for 5000\Re\1:2 106
ð8:32Þ
Dean (in Ebadian and Dong 1998) developed the following equation based on comprehensive survey of the available data: f ¼ 0:0868Re0:25 : for 1:2 104 \Re\1:2 106
ð8:33Þ
In the range of 5000 < Re < 1.2 104, Bhatti and Shah (in Ebadian and Dong 1998) recommend Eq. (8.32); otherwise use of Eq. (8.33) is recommended to obtain the friction factor for the fully developed turbulent flow in parallel plate duct using hydraulic diameter as twice the spacing between two plates. Hart and Lawther (in Kays and Perkins 1973) suggest that annular flow develops more quickly than pipe flow. Olson and Sparrow’s data appear to show that 30 diameter is sufﬁcient for a close approach to fully developed flow (Kays and Perkins 1973). However, Lee (1968) typically gives the thermal development length of only 20 diameters for annular duct of r* = 0.67 at Re = 104 and the development length has been shown to increase with the increase in the Reynolds number. Kays and Leung (1963) presented results of their solution of turbulent flow energy equation of fully developed constant heat rate condition over a wide range of annulus radius ratio, Reynolds number and Prandtl numbers. The solutions are presented for one surface heated and other insulated. Typical results for Pr = 0.7 are given in Table 8.9. For detailed
610
8
Empirical Relations for Forced Convection Heat Transfer
Table 8.9 Fully developed constant heat rate turbulent flow Nusselt number (Pr = 0.7) Re = 104
Re = 3104
Re = 105
Nuoo
Nuii
Nuoo
Nuii
Nuoo
Nuii
Nuoo
48.5
29.8
98
66
235
167
1510
1100
38.6
29.4
79.8
64.3
196
165
1270
1070
0.5
30.9
28.3
66.0
62.0
166
158
1080
1040
0.8
28.5
28
62.3
61
157
156
1050
1020
*
Nuii
0.1 0.2
r
Re = 106
Parallel plates duct 1.0
27.8
61.2
155
1030
results, refer Kays and Leung. However, Bhatti and Shah, and Sparrow and Lin (in Ebadian and Dong 1998) concluded that the Nusselt number for parallel ducts can be determined using the circular duct correlations with hydraulic diameter in Nusselt number and Reynolds number as twice the spacing between the plates. Mills (1995) has extended the reasoning for the same. In the turbulent flow, the viscous sublayer around the perimeter of the duct is very thin, and the velocity and temperature are nearly uniform across the core. Since the viscous sublayer is the major resistance to momentum and heat transfer, the precise shape of the core fluid is not critical. Recently, Gnielinski (2015) has suggested the following correlation: " 2=3 # ðfann =2ÞðRe 1000Þ Pr Dh pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2=3 Nu ¼ 1þ Fann K L 1 þ 12:7 ðfann =2ÞðPr 1Þ
ð8:34Þ
for 0:1 Pr 1000; ðDh =LÞ 1; and Re [ 4000 where the hydraulic diameter Dh of the annulus is do–di. The length of the annular tube is L, do is the inner diameter of the outer tube and di is the outer diameter of the inner tube. The friction factor in Eq. (8.34) depends on the diameter or radius ratio r*(= di/do). Gnielinski suggests 4fann ¼ ð1:8 log10 Re 1:5Þ2
ð8:35Þ
where Re ¼ Re
½1 þ ðr Þ2 ln r þ ½1 ðr Þ2 ½1 r 2 ln r
ð8:36Þ
A factor Fann allows for the different ways of heating or cooling the flow in the two passages of the annulus. By a comparison of large volume of experimental data, the value has been given by Gnielinski (2015) as Fann ¼ 0:75ðr Þ0:17
ð8:37aÞ
8.7 Heat Transfer and Friction in Concentric Circular Tube …
611
for the boundary condition of “hear transfer at the inner wall with the outer wall insulated”, and Fann ¼ 0:9 0:15ðr Þ0:6
ð8:37bÞ
for the boundary condition of “heat transfer at the outer wall with the inner wall insulated,” but no experimental data have been found for heat transfer from both the walls in the annular flow (Gnielinski 2015). The variation of fluid properties with temperature is taken into account by a factor K. As no special studies of annular ducts are known hence correction factors for circular tubes may be adopted (Gnielinski 2015). For gases, K is given as (Gnielinski 2015)
Tb Tw
K¼
n
ð8:38aÞ
where Tb is the bulk temperature of the gas and Tw is the wall temperature in kelvin. n = 0.45 in the temperature range of 0.5 < (Tb/Tw) < 1.0. For carbon dioxide and steam n = 0.15 in the temperature range. For liquids (Gnielinski 2015), K¼
Prb Prw
0:11
ð8:38bÞ
where Prb is the Prandtl number at absolute bulk temperature and Prw is at the absolute wall temperature. For hydrodynamically developing flow, hydrodynamically developed and thermally developing flow, and simultaneously developing flow in parallel plate ducts, refer Ebadian and Dong (1998).
8.8
Heat Transfer and Friction in Rectangular Duct
8.8.1 Laminar Flow Friction Factor Correlation The empirical equation suggested by Shah and London (in Ebadian and Dong 1998) is f Re ¼ 24ð1 1:3553a þ 1:9467a2 1:7012a3 þ 0:9564a4 0:2537a5 Þ
ð8:39Þ
where a* is aspect ratio (ratio of duct height to width). Heat Transfer Coefﬁcient Correlations The fully developed Nusselt numbers NuT for the case of uniform temperature at four walls are approximated by the following equation (Shah and London in Ebadian and Dong 1998): NuT ¼ 7:541ð1 2:610a þ 4:970a2 5:119a3 þ 2:702a4 þ 0:548a5 Þ
ð8:40Þ
612
8
Empirical Relations for Forced Convection Heat Transfer
Table 8.10 Fully developed Nusselt number and friction factor for laminar flow in ducts of different crosssections Boundary conditionsa
Crosssection shape (L/Dh > 100)
2b 2a
α*= 2b/2a = 1 α* = 0.5 α* = 0.25 α* = 0.125
fRe = Cf Re/4
NuT
NuH
3.657
4.354
16.0
2.976
3.608
14.227
3.391
4.123
15.548
4.439
5.331
18.233
5.597
6.490
20.585
7.541
8.235
24.0
4.861
5.385
24.0
α*= 0 Insulated
α*= 0
a NuT: Tw = constant irrespective of x, y, z NuH: Heat rate qw = constant along the length of a tube, but uniform temperature around the periphery)
For rectangular ducts with a uniform heat flux at four walls constant along the length of the duct, but uniform temperature around the periphery NuH can be calculated from the following equation (Shah and London in Ebadian and Dong 1998): NuH ¼ 8:235ð1 2:0421a þ 3:0853a2 2:4765a3 þ 1:0578a4 0:1861a5 Þ
ð8:41Þ
Some typical results for the friction factor and constant heat rate and surface temperature Nusselt numbers are listed in Table 8.10 along with circular and parallel plate ducts for comparison. For Nusselt number with uniform temperature or heat rate at one or more walls, and for the hydrodynamically developing flow, thermally developing flow and simultaneously flow, refer Ebadian and Dong (1998).
8.8.2 Turbulent Flow The entrance conﬁguration (abrupt or smooth) exerts a marked influence on the value of critical Reynolds number Rec for flow in smooth rectangular duct. Results of some experimental measurements are presented in Bhatti and Shah (1987), which shows that Rec varies from 3400 (a* = 0, parallel plate duct) to 4300 (a* = 1.0, square duct) with smooth entrance while these values are 3100 and 2200, respectively, for the abrupt entrance. A unique feature of the fully developed turbulent flow in rectangular ducts is the presence of secondary flow (flow normal to the axis of the duct). Though the secondary flow is small in magnitude (1% of the axial mean velocity), it exerts a signiﬁcant effect on the turbulence
8.8 Heat Transfer and Friction in Rectangular Duct
613
fluid flow characteristics of these ducts and increases the friction factor by approximately 10% (Bhatti and Shah 1987). For fluids with Prandtl number greater than 0.5, the resistance to heat transfer is basically conﬁned to the laminar sublayer and the temperature distribution over most of the turbulent core region is relatively flat. Under these circumstances, the heat transfer from the surface is reasonably independent of the duct shape if the surface temperature around the duct periphery is uniform (Kays and Crawford 1980). Hence, the friction factor and Nusselt number for the fully developed flow in rectangular ducts can be estimated from the circular duct correlations. If laminar equivalent diameter Dl is used in the place of Dh, circular duct correlations provide values within ± 5% of the experimental results (Bhatti and Shah 1987). The laminar equivalent diameter Dl is deﬁned (Bhatti and Shah 1987) as Dl 2 11 a ð2 a Þ ¼ þ Dh 3 24
ð8:42Þ
where Dh ¼
4A 4ab ¼ P ða þ bÞ
ð8:43Þ
and a ¼ ð2b=2aÞ. Bhatti and Shah (1987) have presented the following friction factor correlation for flow in rectangular crosssection smooth duct (0 a* 1) in the range 5000 Re 107: f ¼ ð1:0875 0:1125a Þfo
ð8:44Þ
where fo ¼ friction factor for the circular duct using Dh ¼ 0:0054 þ 2:3 108 Re1:5 and ¼ 1:28 10
3
þ 0:1143Re
0:311
for 2300 Re 3550 for 3550Re 107 :
The prediction of Eq. (8.44) are at par with those determined by substituting Dh by Dl (Bhatti and Shah 1987), i.e. an uncertainty of ± 5% in the predicted friction factors from the above correlation may be considered. For hydrodynamically developing, thermally developing and simultaneously developing flows in rectangular ducts, refer Ebadian and Dong (1998) and Bhatti and Shah (1987). With uniform heating at four walls of a rectangular duct, circular duct Nusselt number correlations provide results with ± 9% accuracy for 0.5 Pr 10 and 104 Re 106 while for the duct with equal heating at two long walls, circular duct correlations provide result with accuracy of ± 10% for 0.5 Pr 10 and 104 Re 105 (Bhatti and Shah 1987). In the case of heating at one long wall only (asymmetrical heating), the circular duct correlations provide values which may be up to 20% higher than the actual experimental values for 0.7 Pr 2.5 and 104 Re 106 and this is true in all probability for 0 a* 1 (Bhatti and Shah 1987). This conclusion is in line with the results of the studies of Sparrow et al. (1966) for Pr = 0.7. Asymmetric heating condition is encountered in solar air heaters. The Nusselt number correlations appropriate for solar air heater ducts have been presented in Chap. 16.
614
8.9
8
Empirical Relations for Forced Convection Heat Transfer
Correlations for External Forced Flow Over a Flat Plate
8.9.1 Laminar Flow Skin Friction Coefﬁcient The skin friction coefﬁcient, 0:6641 Cfx ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ Rex
ð7:37Þ
The average value of the skin friction coefﬁcient for the plate length L is 1:3282 Cf ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ ReL
ð7:38Þ
where ReL is based on the total length L of the plate in direction of flow. Heat Transfer Coefﬁcient Correlations The local Nusselt number equation is Nux ¼ 0:332 Pr 1=3 ðRex Þ1=2 Rex \5 105
ð7:56bÞ
0:6\ Pr \50 The average Nusselt number over the plate length L is Nuav ¼ 0:664 Pr 1=3 ðReL Þ1=2
ð7:57Þ
The Stanton number St ¼ cphqU is used some times as an alternative for Nusselt number when presenting heat transfer data. The above relations can be written in terms of the Stanton number as 0:332 Stx Pr 2=3 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ : Rex
ð7:58Þ
Using Eq. (7.37), we have jH ¼ Stx Pr 2=3 ¼
Cfx : 2
ð7:59Þ
Equation (7.59) is known as ReynoldsColburn analogy. Factor jH is known as Colburn j factor for heat transfer.
8.9 Correlations for External Forced Flow Over a Flat Plate
615
8.9.2 Turbulent Flow Skin Friction Coefﬁcient Case A. Boundary Layer Fully Turbulent from the Leading Edge of the Plate: The local value of the skin friction coefﬁcient, Cfx ¼
0:0583 Re1=5 x
ð7:81Þ
The average value of the skin friction coefﬁcient, Cf ¼
0:0729
ð7:82Þ
1=5
ReL
The above result has been found to be in a good agreement with the experimental results in the range 5 105 < Re < 107. Case B. The boundary layer is laminar up to a certain distance from the leading edge and the transition, as discussed earlier, to turbulent flow takes place at x = xc corresponding to critical Reynolds number Rec = 5 105. Mixed average value of the skin friction coefﬁcient, for Re < 107, Cf ¼
0:074 1=5 ReL
1742 ReL
ð8:45Þ
Another relation suggested for 107 < Re < 109 is Cf ¼
0:455 ðlog10 ReL Þ
2:58
A ReL
ð7:85bÞ
where the value of A depends on the value of the critical Reynolds number at which the transition to turbulent flow takes place. For Rec = 5 105, A = 1700. Heat transfer coefﬁcient correlations The local value of the Nusselt number 1=3 Nux ¼ 0:0296Re4=5 x Pr
for 5 105 \Re\107
ð7:116Þ
Average Nusselt number equation over the plate length L (turbulent flow from the leading edge (x = 0) Nu ¼
hL 1=3 ¼ 0:037Re0:8 L Pr k
ð7:117Þ
Mixed average Nusselt number over the entire laminarturbulent boundary layer when the boundary layer is laminar up to a certain distance from the leading edge and the transition at x = xc corresponds to Rec = 5 105,
616
8
Nu ¼
Empirical Relations for Forced Convection Heat Transfer
1=3 hL ¼ 0:037Re0:8 L 871 Pr k
ð7:118Þ
Example 8.16 Air at a temperature of 300°C flows with a velocity of 3 m/s over a flat plate 0.5 m long. Estimate the cooling rate per unit width of the plate needed to maintain it at a surface temperature of 20°C. The local ﬁlm coefﬁcient may be calculated from the following correlation. Nux ¼ 0:332ðRex Þ0:5 ðPrÞ0:33 for Re 5 105 Nux ¼ 0:0296ðRex Þ0:8 ðPrÞ0:33 for Re [ 5 105
Solution The mean ﬁlm temperature is tfm ¼
300 þ 20 ¼ 160 C: 2
At 160°C, the thermophysical properties of the air are: q ¼ 0:817 kg=m3 ; k ¼ 3:59 102 W=ðm KÞ; l ¼ 2:42 105 kg=ðm sÞ and Pr ¼ 0:6848:
The flow Reynolds number at L = 0.5 m is ReL ¼
qU1 L 0:817 3 0:5 ¼ ¼ 50640: l 2:42 105
Since the Reynolds number at the end of the plate is less than 5 105, the flow is laminar throughout. The applicable Nusselt number correlation is Nux ¼ 0:332ðRex Þ0:5 ðPrÞ0:33 From which, k qxU1 0:5 hx ¼ 0:332 ðPrÞ0:33 x l or 3:59 102 0:817 x 3 0:5 hx ¼ 0:332 ð0:6848Þ0:33 ¼ 3:348x0:5 2:42 105 x
8.9 Correlations for External Forced Flow Over a Flat Plate
617
Hence, the average convective heat transfer coefﬁcient is 1 hav ¼ L ¼
1 L
ZL hx dx 0
ZL
3:348 x0:5 dx
0
0:5 0:5 3:348 x ¼ ¼ 9:47 W/(m2 K): 0:5 0:5 0 Heat flow rate is q ¼ hADT where A DT
heat transfer area = 0.5 1 = 0. 5 m2 tw – tair = 300–20 = 280°C.
This gives q ¼ hADT ¼ 9:47 0:5 280 ¼ 1325:8 W: Example 8.17 Air at atmospheric pressure and 15 m/s flows over a 4 mm 4 mm chip located 100 mm from leading edge on a flat electronic circuit board and removes 20 mW of heat. Determine temperature of the chip surface if the free stream air temperature is 25°C. If flow is turbulent throughout, determine temperature of the chip surface. Solution Thermophysical properties of air at ﬁlm temperature tfm = 50°C (assumed) from Table A5 are: q ¼ 1:0949 kg=m3 ; l ¼ 1:9512 105 kg=ðm sÞ; k ¼ 0:02799 W=ðm KÞ and Pr ¼ 0:703: Local flow Reynolds number, Rex ¼
qU1 x 1:0949 15 0:1 ¼ ¼ 84171: l 1:9512 105
The flow is in the laminar regime. Hence, the local value of the Nusselt number is given by 1=3 Nux ¼ 0:332Re1=2 ; x Pr
618
8
Empirical Relations for Forced Convection Heat Transfer
which gives k k 1=3 hx ¼ Nux ¼ 0:332Re1=2 x Pr x x 0:02799 0:332 ð84171Þ1=2 ð0:703Þ1=3 ¼ 0:1 ¼ 23:97 W= m2 K : Since the chip is very small, hav hx and Newton’s equation q ¼ hx Achip ðts t1 Þ gives ts ¼ t1 þ
q 20 103 ¼ 25 þ ¼ 77:15 C: hx Achip 23:97 4 4 106
i.e., the chip surface temperature is 77.15°C. The ﬁlm temperature tfm = (ts + t∞)/2 = 51.1° C, which is nearly equal to the assumed ﬁlm temperature hence retrial is not required. If flow is turbulent, k k 1=3 hx ¼ Nux ¼ 0:0296Re4=5 x Pr x x 0:02799 0:0296 ð84171Þ4=5 ð0:703Þ1=3 ¼ 0:1 ¼ 64:18 W= m2 K : Since the chip is very small, hav hx and Newton’s equation q ¼ hx Achip ðts t1 Þ gives ts ¼ t1 þ
q 20 103 ¼ 25 þ ¼ 44:48o C hx Achip 64:18 4 4 106
i.e., the chip surface temperature is 44.48°C. The ﬁlm temperature tfm = (ts + t∞)/ 2 = 34.74C° against assumed temperature of 50°C. Retrial is required. Thermophysical properties of air at ﬁlm temperature tfm 35 °C from Table A5 are: q ¼ 1:15 kg=m3 ; l ¼ 1:8823 105 kg=ðm sÞ; k ¼ 0:02684 W=ðm KÞ and Pr ¼ 0:7066: Local flow Reynolds number, Rex ¼
qU1 x 1:15 15 0:1 ¼ ¼ 91643: l 1:8823 105
Local heat transfer coefﬁcient is k k 1=3 hx ¼ Nux ¼ 0:0296Re4=5 x Pr x x 0:02684 0:0296 ð91643Þ4=5 ð0:7066Þ1=3 ¼ 0:1 ¼ 66:0 W= m2 K :
8.9 Correlations for External Forced Flow Over a Flat Plate
619
Since the chip is very small, hav hx and Newton’s equation q ¼ hx Achip ðts t1 Þ gives ts ¼ t1 þ
q 20 103 ¼ 25 þ ¼ 43:94 C: hx Achip 66:0 4 4 106
Further retrial is not required. Example 8.18 Air at atmospheric pressure and a temperature of 25°C is flowing parallel to a 0.8 m long flat plate at a velocity of 10 m/s. If the plate is subjected to uniform heat flux of 1500 W/m2 and flow is turbulent throughout, determine (a) plate surface temperature at x = L and (b) average temperature of the plate surface. Solution The convection heat transfer equation gives q00 ¼ hx ½ts ðxÞ t1 or ts ðxÞ t1 ¼
q00 hx
(a) Plate surface temperature at x = L is ts ðLÞ t1 ¼
q00 hL
For turbulent flow, Nux ¼ 0:0296ðRex Þ0:8 Pr 1=3 Hence, k U1 L 0:8 1=3 Pr hL ¼ 0:0296 L m For air at 50°C and atmospheric pressure, m ¼ 17:82 106 m2 =s; k ¼ 0:02799 W=ðm KÞ and Pr ¼ 0:703: This gives hL ¼
0:8 0:02799 10 0:8 0:0296 0:7031=3 ¼ 30:62 W= m2 K : 6 0:8 17:82 10
620
8
Empirical Relations for Forced Convection Heat Transfer
and ts ðLÞ ¼
q00 1500 þ 25 ¼ 74 C: þ t1 ¼ 30:62 hL
The mean ﬁlm temperature tm = (74 + 25)/2 = 49.5°C, nearly equal to the assumed temperature. (b) Mean plate surface temperature ts t1
1 ¼ L ¼
1 L
ZL ½ts ðxÞ t1 dx 0
ZL 0
q00 ¼ L
q00 dx hx
ZL 0
x dx: kNux
Substituting Nux ¼ 0:0296ðRex Þ0:8 Pr1=3 ¼ 0:0296ðU1 x=mÞ0:8 Pr1=3 , we have ts t1
q00 ¼ L
ZL o
x k 0:0296ðU1 x=mÞ0:8 Pr1=3
q00 1 ¼ L k 0:0296ðU1 =mÞ0:8 Pr1=3
dx
ZL
x0:2 dx
o
q00 1 L1:2 ¼ 0:8 1=3 L k 0:0296ðU1 =mÞ Pr 1:2 ¼
1500 1 0:81:2 0:8 1:2 0:02799 0:0296 ð10=17:82 106 Þ0:8 0:7031=3
¼ 40:83 C: Mean plate surface temperature ts ¼ 40:83 þ t1 ¼ 40:83 þ 25 ¼ 65:83 C:
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
We shall discuss in this section the convection under forced flow condition from a solid body immersed in a stream of fluid. Though the submerged bodies may be of various shapes, we shall consider the flow around cylinder and sphere in cross flow.
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
621
8.10.1 Cylinder in Cross Flow A cylinder in cross flow has been studied extensively from fluid mechanics point of view. Here a brief account of the same is being presented. Readers can refer Schlichting (1960) for the details. Figure 8.9 shows a cylinder in cross flow. Starting from the stagnation point S (/ = 0o), a boundary layer builds up on the front portion of the surface of the cylinder and its thickness increases with angle /. If the body is heated, thermal boundary layers also develops in a similar way. For an ideal or nonviscous flow (known as potential flow) the pressure decreases and the velocity increases from the stagnation point and reaches its maximum at / = 90o. On the back side of the cylinder, the pressure increases again and at / = 180o attains the same value as the forward stagnation point. Figure for this case has been included here. In this case the separation of the flow from the cylinder surface does not take place and the drag (the force resisting the flow) on the cylinder is zero. In a fluid with friction, the flow separates on both sides of the surface of the cylinder at about 80o−100o and on the back side of the cylinder a dead fluid mass is ﬁlled with vortices on the downstream side. The fluid particles outside the boundary layer are able to move against the pressure increase on the back side of the cylinder by changing their kinetic energy into pressure energy. But the fluid particles in the boundary layer do not possess so much kinetic energy. They can therefore move only a small distance into the region of increasing pressure before their kinetic energy is consumed. Then they reverse their flow direction. In this way the flow separates from the surface. At the separation point, the velocity gradient at the surface becomes zero [(∂u/∂y)y = 0 = 0]. Because of this separation, the pressure distribution along the back side of the cylinder is changed also. The extent of the penetration in the region of increasing pressure depends on the kinetic energy of the fluid particles, which is greater in the case of turbulent flow boundary layer than in laminar one. As a consequence, the turbulent boundary layer (at Re greater than critical Reynolds number Rec) separates at about / = 110o and the laminar at / = 82o. The pressure on the rear side of the cylinder is lower than on the front side. This causes a resisting force in the flow direction, which is known as form resistance. To it is
U ∞, T∞
S Wake
Re < Rec
Re > Rec Rec = 2 × 105
Fig. 8.9 Cylinder in cross flow
622
8
Empirical Relations for Forced Convection Heat Transfer
100 Sphere 10 CD
Cylinder 1
0.1 101
1
10
102
103
104
105
106
ReD
Fig. 8.10 Drag coefﬁcient CD versus Reynolds number for cylinder and sphere in cross flow. Schlichting H and Gersten K, Boundary layer theory, 8th Edition, Chap. 1, Figs. 1.12 and 1.19, Copyright 2000. With permission of Springer
added the frictional resistance caused by the skin friction or viscous shear acting tangential on the cylinder surface. Both parts are included in a dimensionless quantity known as the drag coefﬁcient,3 which is deﬁned as FD =A CD ¼ 1 2 2 qU1
ð8:46Þ
where FD A
drag frontal area (area normal to the flow) = LD for a cylinder
It is to note that the total drag on a body is the sum of two components: the pressure or form drag (drag force due to pressure) and the skin friction or viscous drag (drag force due to viscous stresses). The drag coefﬁcient CD is plotted in Fig. 8.10 against the Reynolds number where the Reynolds number is based on the free stream velocity and diameter of the cylinder. In general, the drag depends on the properties of the fluid, its free stream velocity and the shape of the body. At very low values of the Reynolds number ( 1), no separation occurs and the drag is caused only by the shear stresses (viscous friction). From Re > 1, a dead zone of fluid builds up behind the cylinder, which increases with the Reynolds number. With further increase in the Reynolds number from 100, vortices shed alternatively from upper and lower sides from the cylinder surface (known as von Karman vortex street). These vortices are carried away by the flow in a regular pattern. At Re = 4 105, a sudden drop in the drag coefﬁcient is observed due to the fact that the boundary layer becomes turbulent before it separates. At Re > 4 105, no vortex wake forms but only irregular vortices exist in the dead fluid.
If the fluid is flowing in the horizontal direction, the sum of the horizontal components of these forces constitutes the drag, and the sum of their vertical components constitutes the lift. In the study of heat transfer, we are interested only in the drag, which is the resistance to the flow to be overcome in pumping of the fluid.
3
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
623
At small velocities, the drag is mostly due to the frictional resistance. At Re = 10, friction and form drag are of the same order. For Re > 1000 the form resistance predominates. Each of these flow regimes influences the heat transfer coefﬁcient in a distinct way. The heat transfer on the back side occurs in the wake. At low Re, the heat transfer coefﬁcient on the front side is much greater than that on the back. With the increase in the Reynolds number, the heat flow from the back side increases at a greater rate than on the front side and at Re = 5 104, it is as high as on the front side. The reason for this behaviour is that the vortices, while separating alternately from the upper and lower sides of the cylinder sweep the surface of the rear half of the cylinder with an intensity that increases with the Reynolds number. When the Reynolds number is > 4 105, the boundary layer becomes turbulent and the distribution of the heat transfer coefﬁcient along the surface changes. The heat transfer coefﬁcient is reported to be very high at an angle of / = 100o, which indicates the transition in the flow within the boundary layer. Hilpert (1933) for air, and Knudsen and Katz (1958) for other fluids extended the following relation for heat transfer from cylinders: Nu ¼
hD ¼ CRen Pr1=3 k
ð8:47Þ
where Re = U∞D/t. The values of the constant C and exponent n are given in Table 8.11. All the fluid properties are to be taken at ﬁlm temperature. Douglas and Churchill (1956) presented the following correlation for heat transfer from a single cylinder for air: Nu ¼ 0:46Re0:5 þ 0:00128Re for 500\Re\106
ð8:48aÞ
For liquid, they recommended the following equation: Nu ¼ ð0:506Re0:5 þ 0:000141ReÞPr1=3
ð8:48bÞ
For Re < 500, Hsu (1963) suggested the following equation for gases. Nu ¼ 0:43 þ 0:48Re1=2
ð8:49Þ
Table 8.11 Constant C and n in Eq. (8.47)a
a
Re
C
n
0.4−4
0.989
0.330
4−40
0.911
0.385
40−4000
0.683
0.466
4000−40000
0.193
0.618
40000−400000
0.027
0.805
Holman JP, adapted for SI units by White PRS, Heat Transfer, McGrawHill Book Co, New York, Copyright 1992. The material is reproduced with permission of McGrawHill Education (Asia)
624
8
Empirical Relations for Forced Convection Heat Transfer
where the fluid properties are to be taken at the ﬁlm temperature and Re is deﬁned as in Eq. (8.47). Isachenko et al. (1977) presented following correlations based on the extensive studies on circumferential mean rate of heat transfer from the surface of a cylinder for flow of air, water and transformer oil by Zukauskas and coresearchers (1959–1972): Nu ¼ 0:5Re0:5 Pr0:38 ðPr = Prw Þ0:25 Nu ¼ 0:25Re0:6 Pr0:38 ðPr = Prw Þ0:25 Nu ¼ 0:023Re0:8 Pr0:37 ðPr = Prw Þ0:25
for 5\Re\103
ð8:50aÞ
for 103 \Re\2 105
ð8:50bÞ
for 2 105 \Re\2 106
ð8:50cÞ
where the fluid properties are to be taken at the ﬁlm temperature. If the flow approaching the cylinder is artiﬁcially turbulized, the heat transfer coefﬁcient will be larger than given by above correlation. However, the correlations presented in Mikheyev (1964) based on the studies of Zukauskas are: Nu ¼ 0:59Re0:47 Pr0:38 ðPr = Prw Þ0:25 Nu ¼ 0:21Re0:62 Pr0:38 ðPr = Prw Þ0:25
for 10\Re\103
ð8:51aÞ
for 103 \Re\2 105
ð8:51bÞ
Churchill and Bernstein (1977) have presented the following correlation for isothermal cylinder based on the data of various researchers: 0:62ReD Pr1=3
Nu ¼ 0:3 þ
1=2
1 þ ð0:4 Pr Þ
2=3 1=4
1þ
ReD
282000
5=8 4=5
for ReD Pr [ 0:2
ð8:52Þ
All properties in the equation are to be evaluated at the ﬁlm temperature tfm. For the Reynolds number range of 7 104−4 105, the equation predicts Nusselt number values that may be signiﬁcantly lower than those from the direct measurement. The equation can be used for uniform heat flux condition if the perimeteraverage of the temperature difference between the surface and the free stream is used.
8.10.2 Flow Around a Sphere The flow regimes around a sphere are similar to those encountered in the case of a cylinder in cross flow and the heat transfer coefﬁcient is governed by these regimes. Starting from / = 0o, a boundary layer develops on the front portion of the surface of the sphere. At very low Reynolds number ( 1), a laminar flow exists around the sphere and the flow is nearly symmetric. At large Reynolds numbers, the boundary layer separates from the sides of the sphere creating a stagnant region ﬁlled with vortices in the back. On the sides of the sphere the boundary layer may become turbulent at large Reynolds numbers. This influences the position of the separation. The drag coefﬁcient for a sphere versus the Reynolds number is shown in Fig. 8.10. The curve can be divided in four regions. In the ﬁrst region (Re 1), the Stoke’s law is applicable and the drag is
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
FD ¼ 3pDlU1
625
ð8:53Þ
and CD ¼
FD 2 ½ðp=4ÞD2 12 qU1
ð8:54Þ
In the second region (Re = 1–1000), the boundary layer separation begins. The coefﬁcient of drag continues to decrease in this region. In the third region (Re = 103 –2 105), the drag coefﬁcient almost remains constant. The boundary layer is entirely laminar before separation. The fourth region is characterized by sudden drop in CD to almost half at Re = 2.5 105. This is due to the change in the flow pattern in the boundary layer from laminar to turbulent just before the separation point. The Reynolds number where the drop in CD occurs is termed as the critical Reynolds number. Actually the transition occurs within a range of Re = 105 – 106 because it also depends on the free stream turbulence. After the transition, the coefﬁcient of drag increases slightly. Whitaker (1972) proposed the following correlation for the average Nusselt number between an isothermal spherical surface at Tw and free stream fluid (U∞, T∞): 2=3
0:25 0:4 Nu ¼ 2 þ ð0:4Re0:5 D þ 0:06ReD Þ Pr ðl1 =lw Þ
for
3:5\ReD \7:6 104 ;
and
0:71\ Pr \380; 1\l1 =lw \3:2
ð8:55Þ
where all properties except lw are to be evaluated at the free stream temperature t∞. The viscosity lw refers to the wall temperature tw. Noflow limit for this equation is Nu = 2, which agrees with pure radial conduction between sphere surface and the motionless inﬁnite medium that surrounds it (see Example 8.13). Moving liquid drops, such as atomized fuel drops in internal combustion engines, liquid drops in spray drying chambers, or water droplets falling in the cooling tower are also examples of spheres in motion in a fluid experiencing convection in currents. In such cases, of course, mass transfer also takes place simultaneously with the heat transfer. For the falling drops, Ranz and Marshall (1952) proposed the following equation for the convective heat transfer for a system of water drops: Nu ¼ 2 þ 0:6Re0:5 Pr1=3 ½25ðD=xÞ0:7
ð8:56Þ
where x is the distance measured from start of fall of the droplet. Properties are to be evaluated at the free stream temperature. The free fall velocity is given by 2 U1 ¼ 2gx
ð8:57Þ
When x is unknown, average value during the fall can be estimated by omitting the bracketed factor in Eq. (8.56), which accounts for the droplet oscillation.
626
8
Empirical Relations for Forced Convection Heat Transfer
8.10.3 Flow Across Tube Banks The flow across tube banks, the tube array or bundle is frequently encountered in heat exchangers. The arrangements of the tubes frequently used are shown in Fig. 8.11. The arrangement is characterized by the relative longitudinal and transverse pitches, p/D and pt/D of two consecutive tubes, respectively. For both the arrangements, the flow across the ﬁrst transverse row, row 1 in the ﬁgure, of the tubes is practically the same as that for a single tube. The nature of the flow across the tubes of remaining rows depends on the way the tubes are arranged. However, in most applications, the flow is turbulent after a few rows. In the case of inline arrangement, all the tubes in the second and subsequent rows are in the vortex region of the upstream row. Thus, a dead zone with relatively weak circulation of flow is formed in the spaces between the tubes because the fluid flows mainly through the longitudinal passages between the tubes. The points of maximum heat transfer coefﬁcient on the tube circumference are located at the spots where the impact of the main stream occurs. The front part of the tubes is not directly exposed to the flow; therefore the heat transfer coefﬁcient is low here. In the staggered arrangement of the tubes, the flow past the tubes deep inside the array differs little from the nature of flow across the ﬁrst row of tubes. The maximum intensity of the heat flux is at the front portion of the tubes in all rows. In general, the flow turbulence increases as the fluid passes through a few initial rows of the tubes. After a sufﬁcient number of rows, the flow assumes a stable character.
Row 1
pt Umax
U∞ T∞1
T∞2
p
p = pitch pt = transverse pitch D = outer diameter of tubes
D
a Inline tubes
pt U∞ T∞1
T∞2
b Staggered tubes Fig. 8.11 Flow across tube banks
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
627
For the selection of the arrangement of tubes in a bank, the pressure loss (the resistance to the flow) and the rate of clogging is also considered. Zukauskas (1987) reviewed the works published on crossflow through tube bundles and he recommends the following correlations for the average Nusselt number for number of rows n 16. Inline Tube Bundles 0:36 Nu ¼ 0:9Re0:4 ðPr = Prw Þ0:25 D Pr
for 1 Re 100
ð8:58aÞ
for 100\Re 1000
ð8:58bÞ
for 103 \Re 2 105
ð8:58cÞ
0:36 ðPr = Prw Þ0:25 Nu ¼ 0:52Re0:5 D Pr 0:36 ðPr = Prw Þ0:25 Nu ¼ 0:27Re0:63 D Pr 0:25 0:4 Nu ¼ 0:033Re0:8 D Pr ðPr = Prw Þ
for 2 105 \Re 2 106
ð8:58dÞ
Staggered Arrangement 0:36 Nu ¼ 1:04Re0:4 ðPr = Prw Þ0:25 D Pr
for 1 Re 500
ð8:59aÞ
for 500\Re 103
ð8:59bÞ
for 103 \Re 2 105
ð8:59cÞ
for 2 105 \Re 2 106
ð8:59dÞ
0:36 Nu ¼ 0:71Re0:5 ðPr = Prw Þ0:25 D Pr 0:36 ðPr = Prw Þ0:25 ðpt =pÞ0:2 Nu ¼ 0:35Re0:6 D Pr 0:25 0:4 ðpt =pÞ0:2 Nu ¼ 0:031Re0:8 D Pr ðPr = Pr w Þ
In Eqs. (8.58a), (8.58b), (8.58c), (8.58d) and (8.59a), (8.59b), (8.59c), (8.59d), the thermophysical properties except Prw are evaluated at the mean bulk temperature tm [= (t∞1 + t∞2)/2] of the fluid flowing around the tubes in the bundle. The Reynolds number is based on the average velocity of the fluid through the narrowest crosssection of the flow area: ReD ¼
qUmax D l
ð8:60Þ
where Umax = U∞pt/(pt – D), refer Fig. 8.11. The factor (pt/p)0.2 takes account of the effect of the change in the longitudinal and transverse pitches. However, this effect is not so evident in the case of the inline arrangement and in staggered arrangement when Re < 103. In general, with the increase in the number of rows, the turbulence and hence the heat transfer coefﬁcient increases. However, this effect of the number of rows is negligible for n > 16. For n < 16, the heat transfer coefﬁcient is lower due to the lower level of turbulence. For n < 16, the Nusselt number values from Eqs. (8.58a), (8.58b), (8.58c), (8.58d) and (8.59a), (8.59b), (8.59c), (8.59d) must be multiplied by a factor Cn (Fig. 8.12), which is a correction factor to take account of the effect of the number of rows n less than 16. The uncertainty in the predicted values from the above correlations is reported to be within ± 15%. Example 8.19 Compare the predicted values of the Nusselt number from different correlations for flow of fluid (Pr = 1) across a cylinder (L >> D) at Re = 1 102, 1 103, 1 104 and 2 105. Neglect the viscosity correction factor (Prf/Prw).
628
8
Empirical Relations for Forced Convection Heat Transfer
Staggered 100 < Re < 1000
Cn Inline Staggered Re > 1000
Number of rows, n
Fig. 8.12 Correction factor Cn as function of number of rows n
Table 8.12 Example 8.19 Re
Hilpert
Douglas and Churchill
Mikheyev
1102
5.84
4.73
5.14
1103
17.08
15.83
15.17
1104 2105
57.22 499.7
Isachenko et al. 5.0 15.21
58.8
63.42
461.7
406.3
15.81
Churchill and Bernstein 5.89
15.77 62.8 378.9
18.3 61.37 399.6
Solution The results are presented in Table 8.12. Example 8.20 An electrical heater is embedded in a long 30 mm diameter cylinder. Either water at a velocity of 1 m/s or air at atmospheric pressure and a velocity of 10 m/s flowing across the cylinder is to be used to maintain the cylinder surface temperature at 75°C. Both fluids are available at 25°C. Calculate the power of the heater required. Comment on the result. Solution Power required per unit length of the cylinder, q ¼ hðpdÞðts t1 Þ L
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
629
For a cylinder in crossflow, k k h ¼ Nu ¼ CRen Pr1=3 d d
ð8:47Þ
Water as coolant Thermophysical properties of water at ﬁlm temperature tfm [= (75 + 25)/2 = 50°C] from Table A4 are: q ¼ 988:1 kg=m3 ; l ¼ 544 106 kg=ðm sÞ; k ¼ 0:644 W=ðm KÞ and Pr ¼ 3:55: Flow Reynolds number, qUd l 988:1 1:0 0:03 ¼ ¼ 54491: 544 106
Re ¼
From Table 8.11, C = 0.027 and n = 0.805. Hence, heat transfer coefﬁcient, k 0:644 h ¼ CRen Pr1=3 ¼ 0:027 544910:805 3:551=3 d 0:03 ¼ 5745 W= m2 K ; and power required per unit length of the cylinder, q ¼ hðpdÞðTs T1 Þ ¼ 5745 ðp 0:03Þ ð75 25Þ ¼ 27073 W=m: L Air as coolant Thermophysical properties of air at ﬁlm temperature tfm [= (75 + 25)/2 = 50°C] from Table A5 are: q ¼ 1:0949 kg=m3 ; l ¼ 1:9512 105 kg=ðm sÞ; k ¼ 0:02799 W=ðm KÞ and Pr ¼ 0:703: Flow Reynolds number, qUd l 1:0949 10 0:03 ¼ ¼ 16834: 1:9512 105
Re ¼
From Table 8.11, C = 0.193 and n = 0.618. Hence, heat transfer coefﬁcient, k 0:02799 h ¼ CRen Pr1=3 ¼ 0:193 168340:618 0:7031=3 ¼ 65:5 W= m2 K : d 0:03
630
8
Empirical Relations for Forced Convection Heat Transfer
and power required per unit length of the cylinder, q ¼ hðpdÞðts t1 Þ ¼ 65:5 ðp 0:03Þ ð75 25Þ ¼ 308:7 W=m: L Note that despite 10 times velocity of air, hwater = 87.7 hair. Thus water is much more effective heat transfer fluid than air. Note: In general, Nu = C RemPrn gives k k UL m n U m k Prn k Prn m n h ¼ CRe Pr ¼ C Pr ¼ C 1m ; / L L m L mm mm i.e., for given fluid velocity U and characteristic dimension L, the heat transfer coefﬁcient is high for fluids of large k and Pr, and small m. Example 8.21 A long cylindrical ﬁn [k = 40 W/(m K)] of 10 mm diameter is installed on a surface (30 mm 30 mm), refer Fig. 8.13. The surface temperature (ﬁn base temperature) is 100°C. If air at 25°C and 5 m/s flows across the ﬁn, determine the heat transfer rate from the ﬁn and the surface. Solution Thermophysical properties of air at ﬁlm temperature (100 + 25)/2 = 62.5°C from Table A5 are: q ¼ 1:05 kg=m3 ; l ¼ 2:0085 105 kg=ðm sÞ; k ¼ 0:02895 W=ðm KÞ and Pr ¼ 0:7: (a) Heat transfer rate from the ﬁn Reynolds number, ReD ¼
qU1 D 1:05 5 0:01 ¼ ¼ 2614: l 2:0085 105
Nusselt number relation is Nu ¼ CRenD Pr1=3
U∞ = 5 m/s t∞ = 25oC 10
tb = 100oC
Fig. 8.13 Example 8.21
Plate 30 SQ
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
631
where C = 0.683 and n = 0.466 for Re = 2614 from Table 8.11. This gives h¼
k k Nu ¼ 0:683Re0:466 Pr1=3 D D D 0:02895 0:683 ð2614Þ0:466 ð0:7Þ1=3 ¼ 0:01 ¼ 68:7 W= m2 K :
Heat rejection from the ﬁn is given by pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ hPkAc ðtb t1 Þ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ hðpDÞkðp=4ÞD2 ðtb t1 Þ
qfin ¼
Substitution of values of various terms gives pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ qfin ¼ 68:7 p 0:01 40 ðp=4Þ 0:012 ð100 25Þ ¼ 6:18 W: (b) Heat transfer rate from surface Flow is parallel to the surface. Flow Reynolds number, ReL ¼
qU1 L 1:05 5 0:03 ¼ ¼ 7842: l 2:0085 105
The flow is laminar (assuming no appreciable effect of the ﬁn on the flow structure). Relevant correlation of average Nusselt number is 1=3 Nu ¼ 0:664Re0:5 L Pr
This gives hav ¼
k k 1=3 Nu ¼ 0:664Re0:5 L Pr D D 0:02895 0:664 ð7842Þ0:5 ð0:7Þ1=3 ¼ 0:03 ¼ 50:4 W= m2 K :
Heat rejection from the surface is qs ¼ hav As ðts t1 Þ ¼ hav ½W W ðp=4ÞD2 ðts t1 Þ ¼ 50:4 ½0:03 0:03 ðp=4Þ 0:012 ð100 25Þ ¼ 3:11 W:
632
8
Empirical Relations for Forced Convection Heat Transfer
Total heat transfer rate from surface and ﬁn, q ¼ 6:18 þ 3:11 ¼ 9:29 W: Heat transfer rate without ﬁn is qs ¼ hav As ðts t1 Þ ¼ hav ðW WÞðts t1 Þ ¼ 50:4 ð0:03 0:03Þ ð100 25Þ ¼ 3:4 W:
Example 8.22 50 W/m of heat is to be dissipated from a 1 mm diameter electric wire using air flowing across the wire at atmospheric pressure and 20°C so as to maintain surface temperature of the wire below 80°C. Determine velocity to accomplish the task. Neglect radiation heat exchange. Solution For air at ﬁlm temperature (80 + 20)/2 = 50°C and atmospheric pressure, m ¼ l=q ¼ 17:82 106 m2 =s; k ¼ 0:02799 W=ðm KÞ and Pr ¼ 0:703: From convective heat transfer equation, q=L ¼ hpDðts t1 Þ where the heat transfer coefﬁcient from Eq. (8.47) is k k k U1 D n 1=3 n 1=3 h ¼ Nu ¼ CReD Pr ¼ C Pr D D D m Assuming ReD lies in the range 40–4000, Table 8.11 gives C = 0.683 and n = 0.466. Substitution gives h ¼ 0:683
k U1 D 0:466 1=3 Pr D m
and "
# k U1 D 0:466 1=3 q=L ¼ 0:683 Pr pDðts t1 Þ D m " # U1 D 0:466 1=3 Pr ¼ 0:683k pðts t1 Þ m
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
633
Substitution of the values of various parameters gives "
U1 0:001 50 ¼ 0:683 0:02799 17:82 106
0:466
# 0:703
1=3
p ð80 20Þ:
Simpliﬁcation gives U∞ = 6.48 m/s. Reynolds number for this value of U∞ is ReD ¼
U1 D 6:48 0:001 ¼ ¼ 364: m 17:82 106
ReD lies in the range 40−4000. Hence, the assumption of the Re range was correct. Example 8.23 A long 40 mm diameter 18–8 steel rod, initially at a uniform temperature of ti = 30°C, is suddenly exposed to crossflow of air at t∞ = 300°C and U∞ = 50 m/s. Find time for the surface of the rod to reach 170°C. Solution For 18–8 steel, q = 7820 kg/m3, c = 460 J/(kg K) and k = 21 W/(m K). For air at ﬁlm temperature [(30 +170)/2 + 300]/2 = 200°C, q ¼ 0:7474 kg=m3 ; l ¼ 2:57 105 kg=ðm sÞ; k ¼ 0:03859 W=ðm KÞ and Pr ¼ 0:681: Biot number, Bi ¼
hD=4 h 0:04=4 ¼ k 21
Reynolds number, ReD ¼
qU1 D 0:7474 50 0:04 ¼ ¼ 58163: l 2:57 105
The heat transfer coefﬁcient in Eq. (i) can be calculated from Eq. (8.47) h¼
k k Nu ¼ CRen Pr1=3 D D
where C = 0.027 and n = 0.805 for Re = 40000 to 400000 from Table 8.11. Substitution gives h¼
0:03859 0:027 581630:805 0:6811=3 ¼ 156:9 W= m2 K : 0:04
ðiÞ
634
8
Empirical Relations for Forced Convection Heat Transfer
This gives Bi ¼
156:9 0:04=4 ¼ 0:0747\0:1: 21
Lumped capacity analysis is applicable. For this condition, t t1 ¼ expðBiFo) ti t1 where Fo ¼
as k s 21 ¼ ¼ s ¼ 0:0584s 2 2 L qc L 7820 460 ð0:01Þ2
(for a cylinder, characteristic length L = D/4 = 0.04/4 = 0.01 m) This gives 170 300 ¼ expð0:0747 0:0584sÞ 30 300 or s ¼ 167:5 s:
Example 8.24 A thin walled steel pipe, 500 mm outside diameter, carries wet steam at 160° C. Pipe is insulated from outside with 50 mm thick insulation [ki = 0.04 W/(m K)]. If air at 5 m/s and 20°C flows across the pipe, determine the heat loss per m length of the pipe. Solution For the insulated pipe, the heat loss is given by q ts t1 ¼ 1 1 2 L 2pk ln D D1 þ pD2 h i
ðiÞ
per m length of the pipe assuming pipe wall surface temperature is equal to the wet steam temperature because the condensing steam heat transfer coefﬁcient is very high. Heat transfer coefﬁcient at the outer surface can be determined as under. For air at ﬁlm temperature of 25°C (assumed), q ¼ 1:1868 kg=m3 ; l ¼ 1:8363 105 kg=ðm sÞ; k ¼ 0:02608 W=ðm KÞ and Pr ¼ 0:709: Reynolds number, ReD ¼
qU1 D 1:1868 5 0:6 ¼ ¼ 1:94 105 : l 1:8363 105
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
635
ReD lies in the range 40000−400000 for which C = 0.027 and n = 0.805 in Hilpert correlation. Hence, the heat transfer coefﬁcient is h¼
k k Nu ¼ 0:027Re0:805 Pr1=3 D D 0:02608 0:027 ð1:94 105 Þ0:805 0:7091=3 ¼ 0:6 ¼ 18:9 W= m2 K :
Equation (i) gives q 160 20 ¼ 1 ¼ 185:8 W=m: 1 L 2p0:04 ln 0:6 0:5 þ p0:618:9 Insulation surface temperature, q=L þ t1 p D2 h 185:8 þ 20 ¼ 25:2 C: ¼ p 0:6 18:9
to ¼
Film temperature is (20 + 25.2)/2 = 22.6°C, which is nearly equal to the assumed ﬁlm temperature of 25°C. Example 8.25 Show that for pure radial heat conduction between surface of a sphere and motionless inﬁnite fluid medium surrounding the sphere, the Nusselt number Nu = 2. Solution For an isothermal sphere of radius r in inﬁnite medium, the shape factor from Table 5.1 is S ¼ 4pR and q ¼ kSðts t1 Þ ¼ kð4pRÞðts t1 Þ
ðiÞ
From the heat transfer consideration, q ¼ hAðts t1 Þ ¼ h 4pR2 ðts t1 Þ Equating Eqs. (i) and (ii), we get h 4pR2 ðts t1 Þ ¼ kð4pRÞðts t1 Þ
ðiiÞ
636
8
Empirical Relations for Forced Convection Heat Transfer
or hR ¼1 k or hD ¼ Nu ¼ 2 k Example 8.26 A 60 W bulb (ts = 135°C,) is situated in an air stream flowing at 0.5 m/s. Determine the heat loss from the bulb if air temperature is 15°C. The bulb can be approximated to be a sphere of 50 mm diameter. Solution At t∞ = 15°C, air properties are q1 ¼ 1:2323 kg=m3 ; k1 ¼ 0:0253 W=ðm KÞ; l1 ¼ 1:787 105 Ns=m2 and Pr1 ¼ 0:7114: At ts = 135°C, lw = 2.317 10−5 N s/m2. The Reynolds number, Re ¼
qU1 D 1:2323 0:5 50 103 ¼ ¼ 1724: l1 1:787 105
Correlation for the average Nusselt number between an isothermal spherical surface at Tw and free stream fluid (U∞, t∞): 2=3
0:25 0:4 Nu ¼ 2 þ ð0:4Re0:5 D þ 0:06ReD Þ Pr ðl1 =lw Þ
ð8:55Þ
Substitution gives, Nu ¼ 2 þ ½0:4ð1724Þ0:5 þ 0:06ð1724Þ2=3 ð0:7114Þ0:4 ð1:787=2:317Þ0:25 ¼ 22:64: Heat transfer coefﬁcient, h ¼ Nu
k 0:0253 ¼ 22:64 ¼ 11:46 W/(m2 K): D 50 103
The heat transfer rate, q ¼ hAðts t1 Þ ¼ 11:46 4 p 0:0252 ð135 15Þ ¼ 10:8 W: Comments: Heat will also be rejected by free convection and radiation.
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
637
Example 8.27 A 10 mm diameter metal sphere [q = 8500 kg/m3, c = 300 J/(kg K), k = 60 W/(m K)], initially at 500°C, is exposed to 25°C air stream flowing at 5 m/s. Determine time for the sphere to cool to 100°C if radiation is neglected. Solution Assuming lumpedheatcapacity analysis is applicable. From Eq. (6.2), ln
t t1 ti t1
¼
hAs s; cqV
which gives cooling time, s¼
cqV ti t1 ln hAs t t1
For sphere, V/As = D/6. Hence, s¼
cqD ti t1 ln 6h t t1
ðiÞ
The heat transfer coefﬁcient can be estimated from Eq. (8.55): 2=3
0:25 0:4 Nu ¼ 2 þ ð0:4Re0:5 D þ 0:06ReD Þ Pr ðl1 =lw Þ
where ReD ¼ qUl1 D. For air at 25°C, q ¼ 1:1868 kg=m3 ; l ¼ 1:8363 105 kg=ðm sÞ; k ¼ 0:02608 W=ðm KÞ and Pr ¼ 0:709: and at 500°C, lw = 3.564 10−5 kg/(m s). This gives ReD ¼
qU1 D 1:1868 5 0:01 ¼ ¼ 3231 l 1:8363 105
and Nu ¼ 2 þ ð0:4 32310:5 þ 0:06 32312=3 Þ0:7090:4 ð1:8363=3:564Þ0:25 ¼ 28:5: Hence, h¼ and from Eq. (i),
k 0:02608 Nu ¼ 28:5 ¼ 74:3 D 0:01
638
8
Empirical Relations for Forced Convection Heat Transfer
cqD ti t1 s¼ ln 6h t t1 300 8500 0:01 500 25 ¼ ln ¼ 105:6 s: 6 74:3 100 25 Biot number, Bi ¼
hðD=6Þ 74:3 ð0:01=6Þ ¼ ¼ 0:002\0:1: k 60
Lumped heat capacity analysis is applicable. Example 8.28 20 mm diameter brass ball at 100°C is dropped in a water bath at 25°C. Determine the time for the ball surface temperature to reduce to 50°C. Given for brass qb = 8520 kg/m3, cb = 380 J/(kg K), kb= 125 W/(m K). Solution After some time ball will acquire constant speed termed as terminal velocity when the resistance of the water through which it is falling prevents further acceleration. At this instance, the force balance gives 1 CD A qw U 2 ¼ ðqb qw ÞVg 2 or CD
p 4
D2
1 2
qw U 2 ¼ ðqb qw Þ
p 6
D3 g
or sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 4ðqb qw ÞDg U¼ 3CD qw At t∞ = 25°C, qw = 997.0 kg/m3, lw = 890 10−6 kg/(m s), kw = 0.609 W/(m K), and Pr = 6.13. At ts 75°C, ls = 374 10−6 kg/(m s). Substitution gives sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 4 ð8520 997Þ 0:02 9:81 1:4 U¼ ¼ pﬃﬃﬃﬃﬃﬃ 3 CD 997 CD where CD is function of Reynolds number (Fig. 8.10) ReD ¼
qw UD 997 U 0:02 ¼ ¼ 22404U: lw 890 106
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
639
Iterative solution gives U 2 m/s. Hence, ReD ¼ 22404 2 ¼ 44808: Correlation for the average Nusselt number between an isothermal spherical surface at ts and free stream fluid (U, t∞): 2=3
0:25 0:4 Nu ¼ 2 þ ð0:4Re0:5 D þ 0:06ReD Þ Pr ðlw =ls Þ
ð8:55Þ
Substitution gives, Nu ¼ 2 þ ½0:4ð44808Þ0:5 þ 0:06ð44808Þ2=3 ð6:13Þ0:4 ð890=374Þ0:25 ¼ 413:3: Heat transfer coefﬁcient, h ¼ Nu
kw 0:609 ¼ 12585 W/(m2 K): ¼ 413:3 0:02 D
Biot number, Bi ¼
hðD=6Þ 12585 0:02=6 ¼ ¼ 0:34 [ 0:1: k 125
Lumped heat capacity analysis is not applicable. Heisler charts will be used for which Bi ¼
hro 12585 0:01 ¼ 1: ¼ 125 k
From Fig. 6.22, for 1/Bi = 1 and r/ro = 1, h/ho = 0.62. ho h ho 50 25 1 ¼ 0:54: ¼ : ¼ 100 25 0:62 hi hi h For 1/Bi = 1 and ho/hi = 0.54, Fo 0.28 from Fig. 6.22, From Fo = as/r2o, time s¼
Foro2 Foro2 qb cb 0:28 0:012 8520 380 ¼ 0:73 s: ¼ ¼ 125 ab kb
It is to note that the terminal velocity is not reached immediately. Reduced velocity during this period implies reduced Re and h, and hence increased time s. Example 8.29 Air at 30°C and atmospheric pressure flows across the tubes of an inline tube bundle consisting of 8 rows of 25 mm OD tubes in the direction of flow and 10 tubes in each row normal to the flow. The tube spacing is 37.5 mm in both parallel and normal to the flow. Calculate the heat transfer from the bundle per m of length when the tube surface temperature is 90°C. The maximum velocity of flow is 10 m/s.
640
8
Empirical Relations for Forced Convection Heat Transfer
Solution Mean bulk temperature, tm ¼
t11 þ t12 30 þ t12 ¼ 2 2
where t∞2 is the bulk temperature of the air at the outlet of the bundle. t∞2 is not known. Therefore we assume that the mean temperature is 40°C. At tm = 40°C, q = 1.1317 kg/m3, k = 0.0272 W/(m K), l = 1.905 10−5 kg/(m s), cp = 1.0066 kJ/(kg K) and Pr∞ = 0.71. At tw = 90°C, Prw = 0.695. The Reynolds number, Re ¼
qUmax D 1:1317 10 0:025 ¼ ¼ 14852 l 1:905 105
Substitution gives, 0:36 Nu ¼ 0:27Cn Re0:63 ðPr =Prw Þ0:25 D Pr
for 103 \ReD 2 105
¼ 0:27 0:97 ð14852Þ0:63 ð0:71Þ0:36 ð0:71=0:695Þ0:25 ¼ 98:9
ð8:58cÞ
where Cn 0.97 for n = 8 from Fig. 8.12. The heat transfer coefﬁcient, h ¼ Nu k ¼ 98:9 0:0272 ¼ 107:6 W/(m2 K): D 0:025 The heat transfer rate, q 30 þ t12 ¼ hAðts t1 Þ ¼ 107:6 ð8 10 p 0:025Þ 90 L 2
ðiÞ
¼ 676:1 ð75 0:5t12 Þ: From the ﬁrst law of thermodynamics, q ¼ mcp ðt12 t11 Þ L where m ¼ qUmax ½ðpt DÞ 10 ¼ 1:1317 10 ð0:0375 0:025Þ 10 ¼ 1:415 kg/s: Hence, q ¼ 1:415 1006:6 ðt12 30Þ L ¼ 1424:3 ðt12 30Þ
ðiiÞ
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
641
Equating Eqs. (i) and (ii), we get 676:1 ð75 0:5t12 Þ ¼ 1424:3 ðt12 30Þ or t12 ¼ 53:02 C: Revised mean temperature, tm ¼
t11 þ t12 30 þ 53:02 ¼ 41:51 C, ¼ 2 2
which is nearly equal to the assumed value, hence retrial is not required. Note: The effect of factor (Pr/Prw)0.25 is very small in this example. This is practically true for low Prandtl number fluids just like gases when wall and bulk fluid temperatures do not differ too much. Example 8.30 Air at 1 atm and 25°C flows at 7.5 m/s over inline tube bundle with p = pt = 25 mm. The bundle contains 20 rows and 10 tubes per row, refer Fig. 8.11a. The tube diameter is 12.5 mm. If tube surface temperature is 400°C, determine the heat transfer rate per m length. Solution Air properties at the mean bulk temperature tm [= (t∞i + t∞o)/2] of the fluid flowing around the tubes in the bundle (assumed to be 125°C): q =0.8872 kg/m3, c = 1013.8 J/(kg K), l = 2.2776 10−5 N s/m2, k = 0.0335 W/ (m K) and Pr = 0.689. At tw = 400°C, Prw = 0.683. The rate of heat transfer per unit length of the tubes is, refer Example 7.17, q ðtw t1o Þ ðtw t1i Þ ¼ hP L ln tw t1o
ðiÞ
tw t1i
where P is perimeter = pDN. The Reynolds number is based on the velocity of the fluid through the narrowest crosssection of the flow area: ReD ¼
qUmax D l
where Umax ¼ U1
pt 25 ¼ 15 m=s: ¼ 7:5 25 12:5 pt D
642
8
Empirical Relations for Forced Convection Heat Transfer
Hence, ReD ¼
0:8872 15 0:0125 ¼ 7304: 2:2776 105
From Eq. (8.58c), Nusselt number is 0:36 Nu ¼ 0:27Re0:63 ðPr =Prw Þ0:25 D Pr
for 103 \Re 2 105
¼ 0:27 ð7304Þ0:63 0:6890:36 ð0:689=0:683Þ0:25 ¼ 64:3: Heat transfer coefﬁcient, h ¼ k Nu ¼ 0:0335 64:3 ¼ 172:3 W= m2 K : D 0:0125 From equation derived in Example 7.16, t1o
P ¼ tw ðtw t1i Þ exp Lh mcp
or
pDN ¼ tw ðtw t1i Þ exp Lh qU1 Nt pt cp p 0:0125 200 1 172:3 ¼ 400 ð400 25Þ exp 0:8872 7:5 10 0:025 1013:8 ¼ 231:9 C: Thus the heat transfer rate from Eq. (i) is q ð400 231:9Þ ð400 25Þ ¼ 172:3 p 0:0125 200 ¼ 348:95 kW=m: L ln 400231:9 40025 The mean bulk temperature tm = (t∞1 + t∞2)/2 = (25 + 231.9)/2 = 128.45°C, which is nearly equal to the assumed value. Note: The approach used in the previous example is approximate for uniform surface temperature, which can be used when temperature differences between hot and cold streams at inlet and exit [i.e., (tw – ti) and (tw – to)] are not signiﬁcantly different. In the present problem the temperature differences are signiﬁcantly different. It is to note that the approach used in this example may be used in all problems when isothermal heating or cooling surface is speciﬁed. Example 8.31 Determine the value of heat transfer coefﬁcient from the surface of water droplets falling in a cooling tower at a section where the velocity of the droplets relative to the air is 1 m/s and temperature is 80°C. The air temperature at the section is 30°C. The average diameter of droplets is 1.25 mm.
8.10
Forced Convection Laminar and Turbulent Flows Around Submerged Bodies
643
Solution For the falling drops, the average value of Nusselt number is given by: Nu ¼ 2 þ 0:6Re0:5 Pr1=3 ½25ðD=xÞ0:7
ð8:56Þ
At t∞= 30°C, air properties are q ¼ 1:1684 kg=m3 ; l ¼ 1:859 106 kg=ðm sÞ; k ¼ 0:02646 W=ðm KÞ and Pr ¼ 0:708: The Reynolds number, Re ¼
qU1 D 1:1684 1 1:25 103 ¼ ¼ 78:6 l 1:859 105
The distance x from Eq. (8.57) is x¼
2 U1 1 ¼ 0:051 m ¼ 2 9:81 2g
Substitution gives, Nu ¼ 2 þ 0:6ð78:6Þ0:5 ð0:708Þ1=3 ½25 ð1:25 103 =0:051Þ0:7 ¼ 10:84: Heat transfer coefﬁcient, h ¼ Nu
8.11
k 0:02646 ¼ 10:84 ¼ 229:5 W/(m2 K): D 1:25 103
Heat Transfer in Liquid Metals
Figure 8.14 shows a typical variation of the Nusselt number with the Peclet number Pe (= RePr) number for flow of liquid metal through a tube with uniform heat flux (Eckert and Drake 1972). The Nusselt number is having a value of 4.36, as found in Chap. 7, for the laminar flow in tubes. The transition from the laminar to the turbulent flow can be seen to occur at Pe 40 (the critical value). At the start of the turbulent flow regime, the Nusselt number rises sharply to a value of about 7. At moderate value of the Peclet number beyond the critical value, the heat transfer by the turbulent mixing is small compared with the conductive transport. When the Peclet number exceeds a value of about 100, the heat exchange by turbulent mixing becomes appreciable and cannot be neglected. Various empirical correlations for the Nusselt number are available in the literature. However, due to the difﬁculties in assessing the effect of free convection and longitudinal conduction, the experimental results of various investigators deviate from each other in the laminar and nearturbulent regions. NuT and NuH results differ signiﬁcantly and separate equations are presented. Some of the correlations are given in Table 8.13.
644
8
Empirical Relations for Forced Convection Heat Transfer
102 8 6 4 2 Nu
10 8 6 4 2 1 1
2
4
6 8 102
2 Pe
6 8103
4
2
4
6 8 104
Fig. 8.14 Uniform heat flux Nusselt number for flow of liquid metal through a tube (Eckert and Drake 1972)
Table 8.13 Nusselt number for fully developed turbulent flow of liquid metals Investigators
Correlation
Boundary condition 0:8
Constant heat flux
Tubes 1.Lyons (in Eckert and Drake 1972) (referred to as Lyons – Martinelli equation)
Nud ¼ 7 þ 0:025Pe for Pe [ 50
2.Sleicher and Rouse (1975)
Nud ¼ 6:3 þ 0:0167Re0:85 Pr0:93
Constant heat flux
3.Skupinshi et al. (1965)
Nud ¼ 4:82 þ 0:0185Pe0:827
Constant heat flux
4.Sleicher and Rouse (1975)
Nud ¼ 4:8 þ 0:0156Re
0:85
0:8
Pr
0:93
Constant wall temperature
5.Seban and Shimazaki (1951)
Nud ¼ 5:0 þ 0:025Pe
Constant wall temperature
Single Cylinder Grosh and Cess (in Eckert and Drake 1972)
Nud ¼ 1:015Pe0:25 for Pe \500
Heat transfer by vortex motion is assumed to be very small compared to the conduction effect in the back of the cylinder
Sphere Witte (1968)
Nud = 2 + 0.386 Pe0.5 for 3.56104 < Re < 1.525105
Although the Nusselt number tends to be low for the liquid metal, like the laminar flow, the heat transfer coefﬁcient is very high because of the high thermal conductivity of the liquid metals.
8.11
Heat Transfer in Liquid Metals
645
Example 8.32 Liquid mercury at 20°C enters a metal tube of 20 mm internal diameter at the rate of 1 kg/s and is heated to 30°C. The tube wall subjected to uniform heat flux is at an average temperature of 40°C. Determine the length of the tube. Given for the mercury: q = 13560 kg/m3, k = 8.7 W/(m K), l = 1.5 10−3 kg/(m s), Pr = 0.025, cp = 139 J/ (kg K). Solution The heat duty is qt ¼ mcp Dt ¼ mcp ðto ti Þ ¼ 1 139 ð30 20Þ ¼ 1390 W: Flow Reynolds number, Re ¼
qUm d 4m 41 ¼ ¼ ¼ 42441: l pdl p 0:02 1:5 103 Pe ¼ Re Pr ¼ 42441 0:025 ¼ 1061 [ 50:
Heat transfer coefﬁcient from Lyons–Martinelli equation, k k h ¼ Nu ¼ ð7 þ 0:025Pe0:8 Þ d d ¼ ½7 þ 0:025 ð1061Þ0:8
8:7 ¼ 5909 W/m2 K: 0:02
Heat transfer rate per unit length, q ¼ hðpdÞðtw tm Þ ¼ 5909 ðp 0:02Þ ð40 25Þ ¼ 5569 W/m: L Required length of the tube, L¼
8.12
qt 1390 ¼ 0:25 m: ¼ q=L 5569
Influence of Duct Wall Roughness in Turbulent Flow
(a) Friction Factor and Nusselt Number Correlations Nikuradse (1950) carried out experiments to study the relationship of friction factor to the Reynolds number for pipes of various roughnesses. Six different degrees of relative roughness e/r (0.002 < e/r < 0.0679 where e = average height of roughness element and r is pipe radius) were used. Figure 8.15 shows to a logarithmic scale the relation of the friction factor and the Reynolds number for the reciprocal values r/e of the six relative roughnesses and for a smooth pipe. For a given relative roughness, three regions can be seen. At low Reynolds number (within the ﬁrst region termed as hydraulically smooth regime), the roughness has no effect on the resistance and for all values of r/e, the curve coincides with
646
8
Empirical Relations for Forced Convection Heat Transfer
1.1 1.0 0.9 r/e = 15
log (100λ)
0.8 0.7
30.6
0.6
60
0.5
126
0.4
252 507
0.3 0.2
2.6
3.0
3.4
3.8
4.2 log Re
4.6
5.0
5.4
5.8
Fig. 8.15 Relation between log (100k) and log Re; r/e is ratio of radius r to average projection e and k is Darcy friction factor (= 4f) (Nikuradse 1950)
the curve for the smooth pipe and resistance factor is expressed by relation k = 64/Re (i.e., f = 16/Re) for the laminar flow condition. For turbulent flow, up to about Re = 105, the Blasius resistance law (k = 0.361Re−0.25) holds for smooth pipes. The critical Reynolds number for all degrees of relative roughness occurs between 2160 and 2500. Within the second range, termed as transitional range, the influence of the roughness becomes noticeable in an increasing degree; the resistance factor increases with an increasing Reynolds number. The resistance factor depends both on the Reynolds number and the relative roughness. In the third region, termed as fully rough regime, the resistance factor is independent of the Reynolds number and the curve k = f(Re) becomes parallel to the xaxis. The dependency of the friction factor on the relative roughness height in the fully rough regime has been expressed by Nikuradse as 1 2e pﬃﬃﬃ ¼ 3:48 1:737 ln D f
ð8:61Þ
An empirical formula correlating the entire transition regime has been given by Colebrook and White (in Bhatti and Shah 1987): 1 2e 9:35 pﬃﬃﬃ ¼ 3:48 1:7372 ln þ pﬃﬃﬃ D f Re f
ð8:62Þ
For e = 0, i.e. the smooth pipes, Eq. (8.62) transforms to PKN correlation and for very high values of the Reynolds number, it transforms to Nikuradse’s equation for fully rough flow regime. The sand grain roughness used by Nikuradse in his experiment is not similar to the roughness encountered in commercial pipes. Hence, Schlichting introduced the concept of equivalent sandgrain roughness as a means of characterizing other types of roughness elements by referring to the equivalent net effect produced by Nikuradse’s experiments (Kays
8.12
Influence of Duct Wall Roughness in Turbulent Flow
647
and Crawford 1980). Moody (1944) determined the equivalent sandgrain roughness for different types of commercial pipe surfaces (refer Moody 1944). The duct wall roughness has no effect in the case of the laminar flow. If the surface roughness height e is of the order of the magnitude of the laminar sublayer thickness dl, it tends to break up the laminar sublayer. This increases the wall shear stress. The ratio of the surface roughness height e and laminar sublayer thickness dl determine the effect of the roughness. The laminar sublayer thickness dl is proportional to m/u*, where u* is the friction velocity (= √(so/q); so is shear stress fqU2m/2). Thus the ratio e/dl is proportional to eu*/m, which is termed as roughness Reynolds number e+: e
þ
rﬃﬃﬃ rﬃﬃﬃ eu e f e Um D f ¼ ¼ Um ¼ m 2 D m 2 m rﬃﬃﬃ f e ¼ Re 2 D
ð8:63Þ
where D is the tube diameter, Um mean velocity and e/D (or e/Dh for noncircular ducts) is termed as relative roughness height. The roughness Reynolds number has been used to deﬁne the flow regimes identiﬁed by Nikuradse for flow in roughened ducts and the physical interpretation of the observed behaviour in these regimes is as follows. 1. Hydraulically smooth (0 e+ 5): The roughness has no effect on the friction factor because the roughness height e is so small that the roughness elements lie entirely within the laminar sublayer. 2. Transition regime (5 < e+ 70): The heights of the roughness elements in this range are of the same order of magnitude as the thickness of the laminar sublayer. Individual projections may extend through the laminar sublayer and vortices produce an additional loss of energy. As the Reynolds number increases, an increasing number of projections pass through the laminar sublayer because of the reduction in its thickness. The additional loss, then, becomes greater as the Reynolds number increases and the friction factor depends both on relative roughness height e/D and the Reynolds number Re. 3. Fully rough regime (e+ > 70): Finally, with the increase in the Reynolds number, the thickness of the laminar sublayer becomes so small that all projections extend through it. The energy loss due to the vortices now attains a constant value and the friction factor is independent of the Reynolds number. In this regime, the friction factor depends only upon the relative roughness height. Moody (1944) also presented friction factor plot as shown in Fig. 8.16. The horizontal portions of the curves right to the dashed line are represented by Nikuradse’s equation for fully rough regime. The dashed line corresponds to e+ = 70. The downward sloping line for the laminar flow is represented by k = 4f = 64/Re. The lowermost curve for the smooth pipes with turbulent flow is represented by the PKN correlation. It is to note that the artiﬁcial roughness creates turbulence close to the wall, which leads to the enhancement in the heat transfer coefﬁcient also. However, the increase in the heat transfer coefﬁcient is accompanied by a proportionately greater increase in the friction factor (Bhatti and Shah 1987). In the case of fluids with high Prandtl number, the resistance to the heat transfer is mainly concentrated in the laminar sublayer, which is thin compared to the hydrodynamic boundary
648
8
Empirical Relations for Forced Convection Heat Transfer
Moody diagram 0.1
0.05 0.04 0.03 0.02 0.015 0.01
0.06 0.04
0.005
λ = 64/Re, Laminar flow
λ
0.002
0.02
0.01
0.001 5×104 Material
e, mm
Drawn tubing Glass, plastic Iron, cast Steel, structural or forged Steel, rusted
0.0025 0.0025 0.15 0.025 0. 5
103
104
2×104 104 5×105 Smooth pipe 105
106
107
Relative pipe roughness, e/D
Complete turbulence
0.08
105 5×106 106 108
Re
Fig. 8.16 Darcy friction factor (k = 4f) plot for smooth and rough tubes. Moody LF (1944) Friction factors for pipe flow. Trans ASME 66: 671–684. With permission of ASME
layer. For the low Prandtl number fluids, the thermal resistance is distributed over a larger portion of the duct crosssection because the thermal boundary layer is thicker than the hydrodynamic boundary layer. Since the roughness at the wall creates turbulence near the wall, the heat transfer enhancement is greater for the fluids with high Prandtl numbers. As pointed out above, the heat transfer coefﬁcient is not affected as strongly as the friction coefﬁcient. The physical explanation for this is given by Bhatti and Shah (1987) as follows. The friction coefﬁcient is markedly augmented by the proﬁle drag developed by the roughness elements. As regards the heat transfer, there is no mechanism comparable to the proﬁle drag to generate additional heat flux. Consequently, the heat transfer is affected less markedly than the friction coefﬁcient. Several studies on heat transfer coefﬁcient and friction factor behaviour for different roughness types have been carried out by researchers and correlations have been proposed, refer Bhatti and Shah (1987). Some of them are presented in Table 8.14. (b) Roughness and Heat Transfer Functions The experimental results of heat transfer and fluid flow characteristics in roughened ducts have been presented either in the form of direct dependence of friction factor and Nusselt number on the system and operating parameters as presented above or in the form of interrelated roughness and heat transfer functions. This latter method makes it possible to present results in a most general form taking into account the various parameters involved including the roughness parameters. The friction correlation has been based on the law of wall similarity employed by Nikuradse (1950) for sand grain roughness in pipes.
8.12
Influence of Duct Wall Roughness in Turbulent Flow
649
Table 8.14 Nusselt number for fully developed turbulent flow in the fully rough flow regime of a circular ducta Investigators
Correlationsb
Dippery and Sabersky (1963)
Nu ¼
Gowen and Smith (1968)
Nu ¼
Bhatti and Shah (1987)
Nu ¼
1þ
Remarks
pﬃﬃﬃﬃﬃﬃﬃﬃ Re Prðfþ =2Þ 0:2 ðf =2Þ½5:19ðe Þ
This correlation is valid for 0.0024 < e/Dh 0.049, 1.2 Pr 5.94, 1.4 104 Re 5105
Pr0:44 8:48
pﬃﬃﬃﬃﬃﬃﬃﬃ RepPrﬃﬃﬃﬃﬃ ðf =2Þ pﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃ 4:5 þ ½0:155ðRe f =2Þ0:54 þ ð2=f Þ Pr
1þ
=2Þ pﬃﬃﬃﬃﬃﬃﬃﬃRe Prðf þ 0:2 ðf =2Þ½4:5ðe Þ
This correlation is valid for 0.021 < e/Dh 0.095, 0.7 Pr 14.3, 104 Re 5104 This correlation is valid for 0.002 < e/Dh < 0.05, 0.5 < Pr < 10, Re > 104. Its predictions are within ±5% of the available measurements.
Pr0:5 8:48
Bhatti MS, Shah RK, Turbulent and transition flow convective heat transfer. In: Kakac S, Shah RK, Aung W (eds) Handbook of singlephase convective heat transfer, Chap. 4, Wiley, New York, Copyright 1987. Reproduced (abridged) with the permission of John Wiley and Sons Ltd b The friction factor f in these correlations may be calculated from the Nikuradse’s correlation, Eq. (8.61) a
The general expression for friction factor f proposed by Nikuradse is f ¼
2 ½A þ B lnðe þ Þ 2:456 lnð2e=DÞ2
ð8:64Þ
where B = 2.5 is a constant for both smooth and rough tubes and A is a nondimensional parameter, which is discussed below. Nikurdse proposed a law of wall by correlating the measured velocity distribution data for these sand grain roughened tubes by a nondimensional equation of the form u þ ¼ u=u ¼ B lnðy=eÞ þ A
ð8:65Þ
Nikuradse found that the plot of parameter A as a function of log(e+) is similar to the curve pﬃﬃﬃ for the resistance law obtained by plotting b1=ð2 f Þ 2 log10 ðr=eÞc against e+. From this similarity, he deduced the value of A as sﬃﬃﬃ 2 2e þ 2:5 ln þ 3:75 A¼ f D
ð8:66Þ
Thus A ¼ uþ
sﬃﬃﬃ 2 2e þ 2:5 ln þ 3:75 2:5 lnðy=eÞ ¼ f D
The nondimensional parameter A is called by different investigators as the roughness parameter or momentum transfer roughness function or roughness function and has been denoted by R(e+) because Nikuradse found roughness function to be function of e+ only. The usual relation for this function is
650
8
Empirical Relations for Forced Convection Heat Transfer
Roughness function, R(e+) or A
11 10 9 8 From resistance From velocity profile
7
5.5 + 2.5 ln(e+) 6 Smooth 5
1
2
5
Transition 10
20
Fully rough 50
100
200
Roughness Reynolds number, e
500 1000
+
Fig. 8.17 Momentum transfer roughness function versus the roughness Reynolds number (Nikuradse 1950)
sﬃﬃﬃ 2 2e þ 2:5 ln þ E Rðe Þ ¼ f D þ
ð8:67Þ
The constant E is termed as geometric parameter and is dependent on the conﬁguration of the duct. Nikuradse reported a value of this parameter as 3.75 for pipes and the same value has been used by Dippery and Sabersky (1963) in their investigation. However, there is no general agreement on the value of constant E. The plot of the roughness function, A or R(e+) against the roughness Reynolds number e+ obtained by Nikuradse is shown in Fig. 8.17. The variation of the roughness function in the different flow regimes as found by Nikuradse is as follows. (i) In the hydraulically smooth regime (0 < e+ 5), the measured pressure loss data were correlated by Nikuradse in the form: Rðe þ Þ ¼ 5:5 þ 2:5 ln e þ
ð8:68Þ
(ii) In the transition regime (5 < e+ 70), the roughness function is observed at ﬁrst to increase with e+ and then to attain a constant value and ﬁnally the function drops. (iii) In the fully rough regime (e+ > 70), the roughness function can be seen to be independent of the roughness Reynolds number and attains a constant value. (iv) For the sand grain roughness, Nikuradse reported a constant value of 8.48 of the roughness function R(e+) for the fully rough region. However, a number of investigators while investigating the behaviour of wire and rib roughness in tubes and ducts (both square and rectangular) have reported different values of the roughness function for the fully rough flow. Further the investigators have shown that the criterion of hydraulically smooth or fully rough or transitional flow regimes and the values of the roughness function depend not only on the roughness Reynolds number but also on the geometric
8.12
Influence of Duct Wall Roughness in Turbulent Flow
651
parameters of the roughness, shape of the roughness element and channel aspect ratio (ratio of width to height) in the case of flow in rectangular channels. Since the roughness function depends on e+, roughness parameters and channel aspect ratio, it is written as R instead of R(e+). (v) Dippery and Sabersky (1963) hypothesized that the law of wall similarity, which has been shown by Nikuradse to be valid for velocity proﬁle in roughened tubes, also applies to the temperature proﬁle. They developed a heat transfer similarity law for the flow in sand grain roughened tubes. They correlated turbulent data for fluids of various Prandtl numbers and tubes of three different sand grain roughness types using a heat transfer function g termed as gfunction by them, which is function of e+ and the function has also been found to have different values for different fluids. Mathematically, the gfunction is deﬁned by Dippery and Sabersky (1963) as
f 1 gðe þ Þ ¼ 2St
sﬃﬃﬃ 2 þ Rðe þ Þ f
ð8:69Þ
For the fully rough flow (e+ > 70), Dippery and Sabersky found that the heat transfer function is a function of only roughness Reynolds number and Prandtl number. Researchers found gfunction to depend also on the geometrical parameters of roughness and hence have written as g instead of g(e+). Webb et al. (1971) used the law of wall similarity employed by Nikuradse and heat transfermomentum analogy extended by Dippery and Sabersky to correlate data for tubes with transverse repeated rib roughness. This general strategy has been successfully used to correlate friction and heat transfer results for different types of roughness elements by a large number of researchers. For typical correlations developed using this strategy for vdiscrete rib roughness elements, refer Karwa et al. (2005). Example 8.33 A 20 mm diameter smooth surfaced tube is maintained at a constant wall temperature of 100°C. Water enters the tube at 30°C and leaves at 50°C with a velocity of 2 m/s. Determine the length of the tube necessary for the desired heating of the water. If the tube has relative roughness of 0.001, what will be the required length? Solution At mean bulk fluid temperature tm = (ti + to)/2 = (30 + 50)/2 = 40°C, the thermophysical properties of air are q ¼ 992:2 kg=m3 ; k ¼ 0:631 W=ðm KÞ; l ¼ 6:51 104 kg=ðm sÞ; Pr ¼ 4:30 and cp ¼ 4179 J=ðkg KÞ: The heat duty is qt ¼ mcp Dt ¼ ½ðp=4Þd2 Um qcp ðto ti Þ ¼ ðp=4Þ ð0:02Þ2 2 992:2 4179 ð50 30Þ ¼ 52105 W:
652
8
Empirical Relations for Forced Convection Heat Transfer
Flow Reynolds number, Re ¼
qUm d 992:2 2 0:02 ¼ ¼ 60965 [ 10000: l 6:51 104
Flow is turbulent. Smooth tube Heat transfer coefﬁcient from Dittus Boelter relation, k k h ¼ Nu ¼ 0:024Re0:8 Pr0:4 d d ¼ 0:024 ð60965Þ0:8 ð4:3Þ0:4
0:631 ¼ 9134 W/(m2 K): 0:02
Since the change in the temperature of the water is not large, a simple arithmetic mean of the end temperature differences may be used in heat transfer equation instead of the log mean temperature difference when wall temperature is constant. Hence, heat transfer rate per unit length, q ¼ hðpdÞðDtÞ ¼ 9134 ðp 0:02Þ ð70 þ 50Þ=2 ¼ 34434 W/m: L Here the log mean temperature difference comes out to be 59.44oC. Required length of the tube, L¼
qt 52105 ¼ 1:513 m: ¼ q=L 34434
Roughened tube From Colebrook and White relation, 1 2e 9:35 pﬃﬃﬃ ¼ 3:48 1:7372 ln þ pﬃﬃﬃ d f Re f Staring with a trial value of f ¼ 0:0791Re0:25 ¼ 0:005, the solution of the above equation gives f = 0.00585. Alternatively, the Moody Diagram may be used to determine friction factor. Roughness Reynolds number, e
þ
rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃ f e 0:00585 ¼ Re ¼ 60965 0:001 ¼ 3:3: 2 d 2
Gnielinski’s correlation from Table 8.3 with viscosity correction factor may be used, refer Eq. (8.20), which gives 0:11 ðf =2ÞðRe 1000Þ Pr lw pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2=3 Nu ¼ l 1 þ 12:7 ðf =2Þ: Pr 1
8.12
Influence of Duct Wall Roughness in Turbulent Flow
653
Substitution gives ð0:00585=2Þð60965 1000Þ 4:3 2:86 0:11 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 388 Nu ¼ 6:51 1 þ 12:7 ð0:00585=2Þ:ð4:32=3 1Þ for lw = 2.86 at 100°C. Heat transfer coefﬁcient, k 0:631 h ¼ Nu ¼ 388 ¼ 12241 W/(m2 K): d 0:02 Required length of the tube, L ¼ 1:513
9134 ¼ 1:129 m: 12241
It is to note the heat transfer enhancement in roughened tube is accompanied with increase in pumping power requirement because of the increased friction. Example 8.34 Calculate the required length if the tube relative roughness is 0.02 for the flow conditions of the previous example. Solution From the Moody’s diagram, f = 0.05/4 = 0.0125 and the flow is in the fully rough region. Alternatively, Nikuradse’s correlation can be used to determine friction factor, which is 1 2e pﬃﬃﬃ ¼ 3:48 1:737 ln D f
ð8:61Þ
1 pﬃﬃﬃ ¼ 3:48 1:737 lnð2 0:02Þ f which gives f = 0.01215. Roughness Reynolds number, e
þ
rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃ f e 0:01215 ¼ Re ¼ 60965 0:02 ¼ 95: 2 d 2
Using Dippery and Sabersky’s correlation with viscosity correction factor, Table 8.14, Re Prðf =2Þ
lw pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ Nu ¼ 0:2 0:44 þ l 1 þ ðf =2Þ:½5:19ðe Þ Pr 8:48
0:11
or 60965 4:3 ð0:01215=2Þ 2:86 0:11 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 774:7: Nu ¼ 6:51 1 þ ð0:01215=2Þ ½5:19 ð95Þ0:2 ð4:3Þ0:44 8:48
654
8
Empirical Relations for Forced Convection Heat Transfer
Required pipe length is L ¼ 1:129
388 ¼ 0:565 m: 774:7
Miscellaneous Exercises Example 8.35 A fluid at an average bulk temperature of 40°C flows inside a 50 mm ID and 1.5 m long tube with bell mouth inlet. Because of constant heat flux condition, the tube wall is 5°C below the fluid bulk temperature. If the average bulk velocity of the fluid is 2.5 m/s, calculate the heat transfer rate from the fluid. The thermophysical properties of the fluid are given as m ¼ 6:0 107 m2 =s; k ¼ 0:15 W=ðm KÞ and Pr ¼ 5: For the entrance effect, if any, following equation may be considered.
hent
1:4 ¼ h 1þ : L=D
Solution The flow Reynolds number ReL ¼
U1 d 2:5 0:05 ¼ ¼ 2:08 105 : m 6 107
The flow is turbulent. Though the Prandtl number is quite high, the temperature of the wall is only 5°C above the fluid bulk mean temperature, viscosity correction is not required. Using the DittusBoelter equation4 for cooling of fluid, k k h ¼ Nu ¼ 0:026Re0:8 Pr0:3 d d ¼ 0:026 ð2:08 105 Þ0:8 ð5Þ0:3
0:15 ¼ 2271 W/(m2 K): 0:05
L/D = 1.5/0.05 = 30 < 60 hence the corrected value of the heat transfer coefﬁcient considering the entrance effect is 1:4 hent ¼ h 1 þ L=D 1:4 ¼ 2271 1 þ ¼ 2377 W/(m2 K): 30
4
The DittusBoelter equation is widely used because of its simplicity. For greater accuracy, Gnielinski or Petukhov et al. correlation (refer Table 8.3) may be considered.
8.12
Influence of Duct Wall Roughness in Turbulent Flow
655
Heat transfer rate, qt ¼ hADT ¼ 2377 ðp 0:05 1:5Þ 5 ¼ 2800:3 W: Example 8.36 Air at 20°C and a pressure of 5.5 bar flows through a circular crosssection tube (ID = 25 mm and L = 1.6 m) with a mean velocity of 1.5 m/s. The tube is subjected to electric heating with heat flux q” of 1.5 kW/m2. Determine the temperature rise of air and the local heat transfer coefﬁcient at L = 1.0 m. Solution Assuming mean bulk temperature of 40°C, the air properties at atmospheric pressure are q ¼ 1:132 kg=m3 ; cp ¼ 1006:6 J=ðkg KÞ: At 5.5 bar, q ¼ 1:132 5:5 ¼ 6:226 kg/m3 : From heat balance, p 2 D qUcp DT ¼ pDL q00 4 or DT ¼
4Lq00 4 1:6 1500 ¼ 40:85 C: ¼ qUDcp 6:226 1:5 0:025 1006:6
Outlet temperature, to = 20 + 40.85 = 60.85°C. Mean bulk temperature, tm = (60.85 + 20)/2 = 40.43°C, which is nearly the same as assumed. Further iteration is not required. Other thermophysical properties are l ¼ 1:9 105 kg=ðm sÞ; Pr ¼ 0:705 and k ¼ 0:0273 W=ðm KÞ: The flow Reynolds number ReL ¼
qUm d 6:226 1:5 0:025 ¼ ¼ 12288 [ 104 : l 1:9 105
Development length from Eq. (8.19), L ¼ 1:359Re0:25 ¼ 1:359 ð12288Þ0:25 ¼ 14:31: D
656
8
Empirical Relations for Forced Convection Heat Transfer
At x = 1 m, x/D = 1/0.025 = 40, which is much larger than the development length for turbulent flow. Hence, the local heat transfer coefﬁcient at x = 1 m equals the heat transfer coefﬁcient for the fully developed flow. Using DittusBoelter equation, k k h ¼ Nu ¼ 0:024Re0:8 Pr0:4 d d ¼ 0:024 ð12288Þ0:8 ð0:705Þ0:4
0:0273 ¼ 42:6 W/(m2 K): 0:025
Example 8.37 Water at an average bulk temperature of 90°C is flowing through a 25 mm ID circular tube with a mean velocity of 1.5 m/s. The pipe wall temperature is 50°C along the whole length below the local value of the bulk temperature of the water because of the uniform heat flux condition. If pipe is 5 m long, determine the cooling rate. The following turbulent flow equation due to SiederTate with viscosity correction factor may be considered. Nu ¼ 0:027Re0:8 Pr1=3
lb lw
0:14
Solution Thermophysical properties of water at the average bulk temperature of 90°C are q ¼ 965:3 kg=m3 ; Pr ¼ 1:97; k ¼ 0:675 W=ðm KÞ; lb ¼ 311 106 kg=ðm sÞ: At wall temperature of tw = 40°C, lw = 651 10−6 kg/(m s). Due to a signiﬁcant difference in lw and lb, viscosity correction must be considered. Flow Reynolds number, ReL ¼
qUm d 965:3 1:5 0:025 ¼ ¼ 1:16 105 : lb 311 106
Flow is turbulent. Using the given equation, 0:14 k k 0:8 1=3 lb h ¼ Nu ¼ 0:027Re Pr d d lw 5 0:8
¼ 0:027 ð1:16 10 Þ
ð1:97Þ
1=3
311 651
0:14
0:675 0:025
¼ 9280 W/(m2 K): Cooling rate, qt ¼ hADt ¼ hðpdLÞðDtÞ ¼ 9280 ðp 0:025 5Þ 50 ¼ 182:2 kW:
8.12
Influence of Duct Wall Roughness in Turbulent Flow
657
Example 8.38 Water flows through a rectangular duct of width W = 10 mm, height H = 5 mm and 3 m long. The velocity of the water is 0.25 m/s. The mean water temperature in the duct is 60°C and the duct wall temperature is 20°C. Calculate the mean heat transfer coefﬁcient and the heat transfer rate. The following correlation may be used Nu ¼ 0:17Re0:33 Pr0:43 Gr0:1
Pr Prw
0:25
where the subscript w refers to the wall condition. Solution Thermophysical properties of water at the mean temperature of 60°C are (Table A4, Appendix) q ¼ 983:3 kg=m3 ; Pr ¼ 3:0; k ¼ 0:654 W=ðm KÞ; l ¼ 463 106 kg=ðm sÞ; b ¼ 0:521 103 : At t = 20°C, Prw = 7. Hydraulic diameter of the duct, Dh ¼
4WH 4 10 5 ¼ ¼ 6:66 mm: 2ðW þ HÞ 2ð10 þ 5Þ
Reynolds number, Re ¼
qUDh 983:3 0:25 ð6:66 103 Þ ¼ ¼ 3536: l 463 x 106
Grashof number, Gr ¼
q2 bgDTD3h 983:32 0:521 103 9:81 ð60 20Þ 0:006663 ¼ ¼ 2:72 105 : l2 ð463 106 Þ2
Heat transfer coefﬁcient from the given Nusselt number correlation: h ¼ Nu
0:25 k k Pr ¼ 0:17Re0:33 Pr0:43 Gr0:1 Dh Dh Prw
0:25 0:654 3:0 0:33 0:43 5 0:1 0:17 ð3536Þ ¼ 3 ð2:72 10 Þ 0:00666 7:0 ¼ 1122:5 W/(m2 K): Heat transfer rate, qt ¼ hADt ¼ hðPLÞðDtÞ ¼ 1122:5 ð30=1000Þ 3 ð60 20Þ ¼ 4041 W:
658
8
Empirical Relations for Forced Convection Heat Transfer
Example 8.39 The heat transfer coefﬁcient for a gas flowing over a thin flat plate 3 m long and 0.3 m wide varies with the distance from the leading edge according to hx ¼ 10x1=4 W/m2 K: Calculate the (a) average heat transfer coefﬁcient, (b) the rate of heat transfer between the plate and the gas if the plate is at a temperature of 170°C and the gas is at 30°C, and (c) the local heat flux at 2 m from the leading edge. Solution hav ¼
¼ ¼
1 L 1 3
ZL hx dx 0
Z3
10x1=4 dx
0
3 10 x1=4 þ 1 ¼ 10:13 W/(m2 K): 3 1=4 þ 1 0
The rate of heat transfer between the plate and the gas, q ¼ hav ADT ¼ 10:13 ð3 0:3Þ ð170 30Þ ¼ 1276:4 W: The local heat flux at x = 2 m is qx¼2 ¼ hx¼2 DT A ¼ 10 ð2Þ1=4 ð170 30Þ ¼ 1177:3 W/m2 :
q00x¼2 ¼
Example 8.40 What is pressure drop in a 20 m long smooth 20 mm ID tubing when water at 100°C flows through it? A Pitot tube measurement shows centerline velocity of 0.02 m/s. Solution Assuming fully developed laminar flow, Um ¼
Umax ¼ 0:01 m/s: 2
The Reynolds number is Re ¼
qUm D 958:3 0:01 0:02 ¼ ¼ 679 l 2:82 104
where q = 958.3 kg/m3 and l = 2.82 10−4 Pa s for water at 100°C.
8.12
Influence of Duct Wall Roughness in Turbulent Flow
659
The flow is laminar. The hydrodynamic development length, Lhy ¼ 0:05ReD ¼ 0:68 m, which is very small compared to the length of the pipe. Thus the assumption of fully developed laminar flow will introduce only a small error in calculation of the friction factor. From Eq. (8.12), the friction factor is f ¼
16 16 ¼ ¼ 0:0235: Re 679
The pressure drop is Dp ¼
4fLqUm2 4 0:0235 20 958:3 ð0:01Þ2 ¼ ¼ 4:5 Pa: 2D 2 0:02
Example 8.41 What pressure head would be if the tube diameter in the above example is halved maintaining the same volume flow? Solution Effect of the halving of the tube diameter on Um, Re and f are Um ¼
4Q 1 / ; pD2 D2
i.e. the mean velocity Um is four times, Re ¼
qUm D 1 1 / Um D / 2 D / ; l D D
i.e., Re is two times (still laminar), f ¼
16 1 / / D; Re Um D
i.e., f is half of the previous value. In terms of the variables of this problem, 2 1 1 1 Dp / ðDÞ 2 ¼ 4 ; D D D which means a 16fold increase. Example 8.42 Air at atmospheric pressure and 25°C enters a circular tube of 30 mm diameter at mass flow rate of 0.0025 kg/s. The heat transfer coefﬁcient is estimated to be 50 W/(m2 K). The surface heat flux at the tube wall is function of distance and speciﬁed as ax, where a = 700 W/m2. Determine for 2 m length of the tube, the outlet temperature of air and tube surface temperature at x = 0 and x = L.
660
8
Empirical Relations for Forced Convection Heat Transfer
q’’ = 700 x
dqconv
tbi
tbo x
mcp(tb+dtb) dx
L=2m
b
a Fig. 8.18 Example 8.42
Solution Speciﬁc heat cp of air at the mean bulk temperature tm [= (tbi + tbo)/2] (assumed to be 50°C) is 1007.2 J/(kg K). From energy balance at the control volume, Fig. 8.18b, we have dqconv ¼ mcp dtb where dqconv ¼ q00 Pdx ¼ 700xðpDÞdx: Substitution gives 700xðpDÞdx ¼ mcp dtb or dtb ¼
700ðpDÞ xdx mcp
Integration from x = 0 to x = L gives: 700ðpDÞ tbo tbi ¼ mcp
ZL xdx 0
or tbo ¼ tbi þ
L 700ðpDÞ x2 700ðpDÞ L2 ¼ tbi þ mcp mcp 2 o 2
Substitution of the values of various terms gives: tbo
2 700 ðp 0:03Þ 2 ¼ 25 þ ¼ 77:4 C: 0:0025 1007:2 2
This gives mean bulk temperature tm [= (tbi + tbo)/2] = 51.2°C, which is nearly equal to the assumed temperature hence retrial is not required. At x = 0, q’’ = 0, hence tsi= tbi. At x = L = 2 m, q’’ = 700 L = 1400 W/m. Hence, from q ¼ hAðtso tbo Þ;
8.12
Influence of Duct Wall Roughness in Turbulent Flow
tso ¼ tbo þ
661
q 700L pD 700 2 ¼ tbo þ ¼ 77:4 þ ¼ 105:4 C: hA hðpD 1Þ 50
Example 8.43 Water at mean temperature of 60°C flows at mean velocity of 0.8 m/s through a 65 ID and 75 OD steel pipe [k = 40 W/(m K)]. Air at atmospheric pressure and 0° C flows across the pipe at a velocity of 5 m/s. Determine the heat loss from water per unit length of pipe. Solution Heat loss per unit length of the cylinder, q ¼ L
1 hi pdi
twater tair do 1 þ 2pk ln di þ s
1 ho pdo
Thermophysical properties of water at 60°C from Table A4 are: q ¼ 983:3 kg=m3 ; l ¼ 463 106 kg=ðm sÞ; k ¼ 0:654 W=ðm KÞ and Pr ¼ 3:0: The Reynolds number of water flowing through the pipe, qUm di lw 983:3 0:8 0:065 ¼ ¼ 110435: 463 106
Re ¼
Flow is turbulent. The heat transfer coefﬁcient can be determined from Dittus and Boelter equation (for cooling of fluid, Table 8.3): k 0:026Re0:8 Pr0:3 di 0:654 0:026 1104350:8 30:3 ¼ 3938 W= m2 K : ¼ 0:065
hi ¼
Thermophysical properties of air at ﬁlm temperature tfm [= (ts + 0)/2] assumed to be = 25°C from Table A5 are: q ¼ 1:1868 kg=m3 ; l ¼ 1:8363 105 kg=ðm sÞ; k ¼ 0:02608 W=ðm KÞ and Pr ¼ 0:709: The Reynolds number of air flowing across the pipe, qU1 do lair 1:1868 5 0:075 ¼ ¼ 24236: 1:8363 105
Re ¼
662
8
Empirical Relations for Forced Convection Heat Transfer
For a cylinder in crossflow, k k h ¼ Nu ¼ CRen Pr1=3 d d
ð8:47Þ
From Table 8.11, C = 0.193 and n = 0.618. Hence, heat transfer coefﬁcient, k ho ¼ CRen Pr1=3 d 0:02608 0:193 242360:618 0:7091=3 ¼ 30:66 W= m2 K : ¼ 0:075 Hence, the heat loss per unit length of the cylinder, q ¼ 1 L 3938p0:065 þ
60 0
1 75 2p40 ln 65
þ
1 30:66p0:075
¼ 427:84 W:
Pipe outer surface temperature, ts ¼ t1 þ
q=L 427:84 ¼ 59:22 C: ¼ 0þ pdo h p 0:075 30:66
Film temperature tfm = (59.22 + 0)/2 = 29.6°C, which is slightly higher than the assumed values of 25°C. Retrial may be carried out for greater accuracy. However, here it will change the result marginally only. Example 8.44 Air at 1 atm and 30°C enters a 4 m long 40 mm inside diameter smooth tube. The mass flow rate of the air is 0.05 kg/s. The tube surface temperature is maintained at 120°C by condensing steam surrounding the tube. Determine the outlet temperature of the air and pressure drop. Solution Air properties at the mean bulk temperature tm [= (ti + to)/2] of 50°C (assumed) from Table A5: q ¼ 1:0949 kg=m3 ; c ¼ 1007:2 J=ðkg KÞ; l ¼ 1:9512 105 Ns=m2 ; k ¼ 0:02799 W=ðm KÞ and Pr ¼ 0:703: For isothermal tube surface, the air outlet temperature is given by tw to P Lh ¼ exp mc tw ti The Reynolds number based on mass velocity of air through the tube: Re ¼
md 4m 4 0:05 ¼ ¼ ¼ 81568: 2 ðp=4Þd l pdl p 0:04 1:9512 105
ðiÞ
8.12
Influence of Duct Wall Roughness in Turbulent Flow
663
Flow is turbulent. L/D = 4/0.04 = 100. Hence, the flow can be regarded as fully developed. Dittus and Boelter equation may be used for calculation of heat transfer coefﬁcient. k 0:024Re0:8 Pr0:4 d 0:02799 0:024 815680:8 0:7030:4 ¼ 123:9 W= m2 K : ¼ 0:04
h¼
Substitution of values of various parameters in Eq. (i) gives 120 to p 0:04 4 123:9 ¼ exp 0:05 1007:2 120 30 or to ¼ 93:8 C: Mean bulk temperature of water is tm [= (ti + to)/2] = 61.9°C. Retrial with changed air properties is required. Air properties at the mean bulk temperature tm of 62.5°C (assumed) from Table A5: q ¼ 1:05005 kg=m3 ; c ¼ 1008 J=ðkgKÞ; l ¼ 2:0085 105 Ns=m2 ; k ¼ 0:02895 W=ðm KÞ and Pr ¼ 0:7: Revised Reynolds number through the tube: Re ¼
4m 4 0:05 ¼ ¼ 79241: pdl p 0:04 2:0085 105
Revised heat transfer coefﬁcient. k 0:024Re0:8 Pr0:4 d 0:02895 0:024 792410:8 0:70:4 ¼ 125 W= m2 K : ¼ 0:04
h¼
Substitution of values of various parameters in Eq. (i) gives 120 to p 0:04 4 125 ¼ exp 0:05 1008 120 30 or to ¼ 94:1 C: Mean bulk temperature of water is tm [= (ti + to)/2] = 62.05°C. Retrial is not required. From Blausius equation, friction factor is
664
8
Empirical Relations for Forced Convection Heat Transfer
f ¼ 0:0791Re0:25 ¼ 0:0791 792410:25 ¼ 0:0047: Pressure drop, Dp ¼
4fLqU 2 4fLq m 2 ¼ 2d qAc 2d
2 4 0:0047 4 1:05005 0:05 Dp ¼ 2 0:04 1:05005 ðp=4Þ 0:042 ¼ 1417N=m2 ¼ 0:014 atm: Example 8.45 Cold fluid is passing through a thinwalled tube 10 mm in diameter 1 m long which is exposed to crossflow of hot fluid at 100°C. Cold fluid flows at a rate of 0.0025 kg/s and its inlet and outlet temperatures are 30 and 60°C, respectively. Determine the outlet temperature of the cold fluid if its flow rate is increased by 50% with all other conditions remaining the same. Given that dynamic viscosity of the fluid is 0.004 N s/m2. Flow may be assumed to be fully developed. Solution The fluid outlet temperature is given by t1 to P LU ; ¼ exp mc t1 ti which gives U m t1 ti ¼ ln c PL t1 to Substitution of values of various parameters gives the ratio of the overall heat transfer coefﬁcient U and speciﬁc heat c as: U 0:0025 100 30 ¼ ln ¼ 0:0445; c p 0:01 1 100 60 where U ¼
1 hi
þ
1 ho
1
:
Flow Reynolds number is Re ¼
4m 4 0:0025 ¼ ¼ 79:5: pdl p 0:01 0:004
The flow is laminar. For increase in mass flow rate by 50%, the Reynolds number will be 79.5 1.5 = 119. For fully developed laminar flow, the Nusselt number and hence the heat transfer coefﬁcient hi is independent of the Reynolds number. Since the hot fluid condition is unchanged, the overall heat transfer coefﬁcient U will not change.
8.12
Influence of Duct Wall Roughness in Turbulent Flow
665
The fluid outlet temperature for increased mass flow rate is P LU to ¼ t1 ðt1 ti Þ exp mc p 0:01 ¼ 100 ð100 30Þ exp 1 0:0445 1:5 0:0025 ¼ 51:8 C: In the above analysis, it has been assumed that fluid properties c and l are independent of temperature. Example 8.46 Helium at 1 atm is to be heated from 400 K to 800 K while flowing at 0.005 kg/s through a tube of 20 mm diameter and 1 m length. Determine the uniform tube wall temperature required to heat the helium. Thermophysical properties of helium at the mean bulk temperature of 600 K and 1 atm are: q = 0.0818 kg/m3, c = 5190 J/(kg K), l = 32.2 10−6 N s/m2, k = 0.251 W/(m K) and Pr = 0.67. Determine outlet temperature and required mass flow rate to achieve the same heat transfer rate and wall temperature if the air at 1 atm is used in place of helium. Solution (a) Heating of helium Tube surface temperature can be determined from T s To P Lh ¼ exp mc Ts Ti
ðiÞ
Flow Reynolds number is Re ¼
4m 4 0:005 ¼ ¼ 9885: pdl p 0:02 32:2 106
Flow is turbulent. Assuming fully developed flow, the heat transfer coefﬁcient from Dittus Boelter equation is k 0:024Re0:8 Pr0:4 d 0:251 0:024 98850:8 0:670:4 ¼ 403 W= m2 K : ¼ 0:02
h¼
Substituting values of various parameters in Eq. (i) gives Ts 800 p 0:02 1 403 ¼ 0:377 ¼ exp Ts 400 0:005 5190
666
8
Empirical Relations for Forced Convection Heat Transfer
or Ts ¼ 1042 K: Heat transfer rate, q ¼ mcp ðTo Ti Þ ¼ 0:005 5190 ð800 400Þ ¼ 10:38 kW: (b) Heating of air Air properties at the mean bulk temperature of 600 K (327°C) and 1 atm, from Table A5: q ¼ 0:5901 kg=m3 ; c ¼ 1054:5 J=ðkg KÞ; l ¼ 3:011 105 N s=m2 ; k ¼ 0:04647 W=ðm KÞ and Pr ¼ 0:68: From ﬁrst law equation, q ¼ mcp ðTo Ti Þ or To ¼
q 10380 9:84 þ 400 ¼ þ 400 þ Ti ¼ mcp m 1054:5 m
Air outlet temperature can be determined from Eq. (i) Flow Reynolds number is Re ¼
4m 4m ¼ ¼ 2:11 106 m: pdl p 0:02 3:011 105
Assuming the flow to be turbulent and fully developed, the heat transfer coefﬁcient from Dittus Boelter equation is k 0:024Re0:8 Pr0:4 d 0:04647 0:024 ð2:11 106 mÞ0:8 0:680:4 ¼ 5480m0:8 ¼ 0:02
h¼
Substituting values of various parameters in Eq. (i) gives 1042 ð9:84=m þ 400Þ p 0:02 0:8 ¼ exp 1 5480m 1042 400 m 1054:5 0:0153 0:3265 ¼ exp 0:2 1 m m Trial and error solution gives m = 0.032 kg/s. Hence,
8.12
Influence of Duct Wall Roughness in Turbulent Flow
667
Re ¼ 2:11 106 m ¼ 2:11 106 0:032 ¼ 67520: 9:84 9:84 þ 400 ¼ þ 400 ¼ 707:5 K: To ¼ m 0:032 For the same heat transfer rate, air mass flow rate is 6.4 times of that of helium. Example 8.47 0.5 kg/s of air at 25°C is to be heated to 75°C. Available for the service are 40 mm diameter 3 m long tubes. The surface temperature of the tubes will be maintained at 100°C by condensing steam. Determine the number of tubes required for the service. Solution Air properties at the mean bulk temperature tm [= (ti + to)/2] of 50°C from Table A5 are: q ¼ 1:0949 kg=m3 ; c ¼ 1007:2 J=ðkgKÞ; l ¼ 1:9512 105 Ns=m2 ; k ¼ 0:02799 W=ðm KÞ and Pr ¼ 0:703: For isothermal tube surface, the air outlet temperature is given by tw to P Lh ¼ exp mc tw ti The equation can be rewritten as
tw ti ln tw to
¼
P Lh mc
¼
pd Lh mc
or
tw ti ln tw to
where heat transfer coefﬁcient, k h ¼ Nu d or h ¼ k 0:024Re0:8 Pr0:4 d using Dittus Boelter equation for fully developed turbulent flow (assumption). The Reynolds number based on mass velocity of air through the tube: Re ¼
md 4m ¼ 2 ðp=4Þd l pdl
ðiÞ
668
8
Empirical Relations for Forced Convection Heat Transfer
Hence, 0:8 h ¼ k 0:024 4m Pr0:4 d pdl Substitution in Eq. (i) gives
tw ti ln tw to
pd k 4m 0:8 0:4 L 0:024 Pr ¼ mc d pdl
or m0:2 ¼
ln
1 tw ti tw to
pkL 4 0:8 0:4 0:024 Pr c pdl
Substituting values of various terms, we have m0:2 ¼
0:8 1 p 0:02799 3 4 10025 0:024 ð0:703Þ0:4 1007:2 p 0:04 1:9512 105 ln 10075
¼ 0:4638 or m ¼ 0:0215 kg=s: Reynolds number Re ¼
4m 4 0:0215 ¼ ¼ 35074: pdl p 0:04 1:9512 105
Length to diameter ratio L/d = 3/0.04 = 75. Hence, assumption of fully developed turbulent flow is correct. For total mass flow rate of 0.5 kg/s, the required number of tubes is n¼
Total flow 0:5 ¼ ¼ 23:25: Flow per tube 0:0215
Hence, 24 tubes will be used. Example 8.48 Water, while flowing at 15 ml/s through a rectangular crossSection (20 mm 5 mm) tube, is to be heated from 20°C to 50°C. Tube surface is maintained at 60°C. Determine the length of tubing.
8.12
Influence of Duct Wall Roughness in Turbulent Flow
669
Solution Water properties at the mean bulk temperature tm [= (ti + to)/2] of 35°C from Table A4 are: q ¼ 994 kg=m3 ; c ¼ 4178 J=ðkg KÞ; l ¼ 718 106 N s=m2 ; k ¼ 0:624 W=ðm KÞ and Pr ¼ 4:81: For isothermal tube surface, the air outlet temperature is given by tw to P Lh ¼ exp mc tw ti The equation can be rewritten as tw ti mc L ¼ ln tw to Ph
ðiÞ
Mass flow rate, m ¼ 15 106 994 ¼ 0:0149 kg=s: For rectangular crosssection tubing, hydraulic diameter is dh ¼
4WH 4 20 5 1 ¼ ¼ 0:008 m: 2ðW þ HÞ 2 ð20 þ 5Þ 1000
Reynolds number of flow, Re ¼
mdh 0:0149 0:008 ¼ ¼ 1660: ðW HÞl ð20 5Þ 106 718 106
Flow is laminar. Assuming flow to be fully developed, the fully developed Nusselt number NuT for the case of uniform temperature at four walls of a rectangular duct is approximated by the following equation: NuT ¼ 7:541ð1 2:610a þ 4:970a2 5:119a3 þ 2:702a4 þ 0:548a5 Þ
ð8:40Þ
where a* is aspect ratio (ratio of duct height to width). Hence, for a* = 5/20 = 0.25, NuT ¼ 7:541ð1 2:610 0:25 þ 4:970 0:252 5:119 0:253 þ 2:702 0:254 þ 0:548 0:255 Þ ¼ 4:44:
Heat transfer coefﬁcient, h¼
k 0:624 4:44 ¼ 346:3 W= m2 K : NuT ¼ dh 0:008
670
8
Empirical Relations for Forced Convection Heat Transfer
Substitution in Eq. (i) gives L ¼ ln
60 20 0:0149 4178 ¼ 4:984 m: 60 50 ½2ð20 þ 5Þ=1000 346:3
Length to hydraulic diameter ratio, L/Dh = 4.984/0.008 = 623. Hence, the assumption of fully developed flow is valid. Example 8.49 A 0.4 m diameter 100 m long thin walled pipeline transports hot water. Water enters the pipe at 80°C. Water flow rate is 5 kg/s. In order to reduce the heat loss from the pipeline, it is covered with 50 mm thick insulation [ki = 0.05 W/(m K)] and is buried 2.0 m below the Earth’s surface [ke = 1.0 W/(m K)]. If the earth’s surface temperature is 0° C, determine the outlet temperature of the water and heat loss. Solution For a differential control volume in the water, refer Fig. 8.19b, dq ¼ mcdtm ¼ ðts tm Þ=Rt or dtm 1 ¼ ðts tm Þ mcRt where total resistance to heat flow Rt ¼
1 1 Do 1 ln þ þ pDi dxhi 2pki dx ke S Di
where S is shape factor. From Case 8, Table 5.2 (L >> R, Z > 3Ro), S¼
2pdx 2p ¼ 22 dx ¼ 2:27dx ln 0:25 ln 2Z Ro
ts dq
Z tmi x
L
a Fig. 8.19 Example 8.49
tmo
Do
Di
mctm
mc(tm+dtm) dx
b
8.12
Influence of Duct Wall Roughness in Turbulent Flow
671
We can write 1 1 Do 1 R0 þ ln þ =dx ¼ t pDi hi 2pki ke 2:27 Di dx
Rt ¼ Hence,
dtm dx ¼ mcR0t ðtm ts Þ Integration for length L gives ts tmo L ln ¼ mcR0t ts tmi or tmo
L ¼ ts ðts tmi Þ exp mcR0t
ðiÞ
The thermophysical properties of water at 80°C are: q ¼ 971:8 kg=m3 ; c ¼ 4198 J=ðkg KÞ; l ¼ 351 106 kg=ðm sÞ; k ¼ 0:670 W=ðm KÞ and Pr ¼ 2:23: The Reynolds number of flow in the pipe, Re ¼
qUm Di m Di 4m 45 ¼ ¼ ¼ ¼ 45343: pDi l p 0:4 351 106 l ðp=4ÞD2i l
Flow is turbulent. Length to diameter ratio of the pipe is 50/0.4 = 125. Flow can be assumed to be fully developed. From DittusBoelter relation, heat transfer coefﬁcient is Nuk k 0:67 0:024 453430:8 2:230:4 ¼ 0:024Re0:8 Pr0:4 ¼ Di Di 0:4 ¼ 294:3 W= m2 K :
hi ¼
Hence, R0t
1 1 Do 1 ¼ þ ln þ pDi hi 2pki ke 2:27 Di 1 1 0:5 1 þ ln ¼ þ p 0:4 294:3 2p 0:05 0:4 1:0 2:27 ¼ 1:154 K m=W:
Substitution in Eq. (i) gives
672
8
tmo
Empirical Relations for Forced Convection Heat Transfer
L ¼ ts ðts tmi Þ exp mcR0t 100 ¼ 0 ð0 80Þ exp ¼ 79:67 C: 5 4198 1:154
Heat loss, q ¼ mcðtmi tmo Þ ¼ 5 4198 ð80 79:67Þ ¼ 6927 W:
8.13
Summary
There are a large number of convection problems for which the analytical solutions for determination of heat transfer coefﬁcient have not met the success especially the problems involving turbulent flow or flows where detachment occurs, for example around cylinders, spheres or other curved bodies. Thus the direct measurement of the heat transfer coefﬁcient (experimental study) has been the main approach for the solution of the most of the heat transfer problems. The experimental results are presented in the form of generalized correlations using the method of dimensional analysis termed as empirical relations because they rely on the observations and experiments not on theory. Based on the dimensional analysis, the generalized correlation for forced convection is written as Nu = w(Re) /(Pr). The form of functions w(Re) and /(Pr) are speciﬁed for different conditions of heat transfer by convection on the basis of theoretical analysis or experimental investigations. Physical signiﬁcance of the dimensionless numbers has also been explained in Sect. 8.3.3. Experimental scheme for the determination of forced convection heat transfer coefﬁcient for uniform temperature and uniform heat flux conditions has been discussed in Sect. 8.4. Correlations of friction factor and heat transfer have been presented for tubes, annuli, rectangular and parallel plate ducts, flat plates, submerged bodies and tube banks for different boundary conditions. Fluid flow and heat transfer characteristics of circular crosssection smooth ducts (tubes) have been investigated in a greater detail as this geometry is having wide applications. Both laminar and turbulent flows in circular tubes have been divided into four categories: fully developed, hydrodynamically developing, hydrodynamically developed but thermally developing, and both hydrodynamically and thermally developing (i.e., simultaneously developing). The fully developed laminar flow of a constant property fluid in a smooth circular tube has been studied analytically as presented in Chap. 7. The Fanning friction factor correlation is f = 16/Re. The hydrodynamic development length Lhy in laminar flow, the distance required for the friction factor to decrease to within 5% of its fully developed value, is given by Lhy/ D 0.05Re. Friction factor in the hydrodynamic entrance region is higher than that for the fully developed case and is deﬁned as apparent mean friction factor fapp , refer Fig. 8.2. For the laminar flow with fully developed velocity and temperature proﬁles i.e., hydrodynamically and thermally fully developed flow, the Nusselt number is a function of the type of heating boundary condition. From the analytical results as given in Chap. 7, Nu = 4.364 for constant heat rate (uniform heat flux) and 3.658 for constant surface temperature condition.
8.13
Summary
673
The thermal entrance length (temperature proﬁle developing and fully developed velocity proﬁle) Lth in laminar flow required for the Nusselt number to decrease to within 5% of its fully developed value is given by Lhy/D 0.05RePr. For the constant surface temperature (known as Graetz problem), the local Nusselt number Nux and the mean Nusselt number Num in the terms of nondimensional tube length x+ = (x/R)/(RePr) are given in Fig. 8.3. Dimensionless group RePr(D/x) is known as Graetz number Gz. For the constant heat rate (uniform heat flux), the values of the local Nusselt number are given in Fig. 8.4. Empirical relation of average Nusselt number over the entire length of the tube for fully developed laminar flow in tubes at constant wall temperature is given in Eq. (8.15). If the velocity and temperature (thermal) proﬁles develop simultaneously, the resulting Nusselt numbers in the entry region are always higher than the preceding case. Constant surface temperature and constant heat rate results are presented in Fig. 8.5a, b, respectively, in the form of mean Nusselt number Num or local Nusselt number Nux as a function of x+. For constant surface temperature condition, Sieder and Tate equation may be used for both liquids and gases in the laminar region where the thermal and velocity proﬁles are developing simultaneously. The equation must be used only for entrance region. It is to be noted that for the turbulent duct flow, the hydrodynamic and thermal entrance lengths are much shorter than the corresponding lengths in the laminar flow. Hence, the results of fully developed turbulent flow friction factor and heat transfer are frequently used in design calculations neglecting the effect of the entrance regions. However, for low Prandtl number fluids, and for the heat exchangers short in length, the entrance region effects must be considered. Several experimental friction factor correlations have been developed for fully developed turbulent flow in smooth ducts. Some of these correlations are presented in Table 8.2, which must be critically studied for application. In comparison to the laminar flow, the hydraulic entrance length is much shorter for turbulent flow and its dependence on the Reynolds number is weaker. In many pipe flow of practical interest, the entrance effects are not signiﬁcant beyond a pipe length of 10 diameters. For fully developed velocity and temperature proﬁles, two boundary conditions of interest are uniform heat flux or constant heat rate along tube length (UHF) termed as H boundary condition and a uniform wall temperature (UWT) termed as T boundary condition. The constant heat rate Nusselt number NUH is always greater than the constant surface temperature Nusselt number NuT, but with exception of very low Prandtl number fluids, the difference in NH and NuT is much smaller than for the laminar flow. At Pr = 0.7, it is only a few percent. A large number of Nusselt number correlations, both theoretical and empirical (based on experimental data), have been developed for fully developed turbulent flow in smooth tubes. Some of them are presented in Table 8.3. Remarks in third column of the table are worth consideration before selecting a correlation for use. Researchers have also presented correlations covering laminar, transition and turbulent flow regimes, which are, in general, difﬁcult to use. When the velocity and temperature proﬁles develop simultaneously in turbulent flow, entry region correction is required, which is strongly affected by the shape at the inlet. If L/D > 60, entry region correction is not required. Effects of temperature varying properties on heat transfer and friction factor values have been discussed and the empirical relations have been presented in Sect. 8.6. When the heat transfer takes place, the temperature of the fluid varies both over the crosssection and along the tube length. Since the transport properties of most fluids vary with temperature and hence
674
8
Empirical Relations for Forced Convection Heat Transfer
will vary over the crosssection of the tube. This causes distortion of the laminar velocity proﬁle due to the heating or cooling of fluids. In the case of the turbulent flow, the effect of the variation of viscosity at the wall is mainly conﬁned to the laminar sublayer. The laminar sublayer tends to be thinner for a decrease in the viscosity at the wall due to the heating of liquids or cooling of gases. The reverse is true when the liquid is being cooled or the gas is being heated. The method to take account of the effects of the variation of the fluid properties with the temperature is to evaluate all the properties at the bulk temperature, and then the variable properties effects on Nusselt number and friction factor are lumped into functions of ratio of the wall to bulk temperatures (Tw/Tb)n and (Tw/Tb)m, respectively, for gases or viscosities (lw/lb)n and (lw/lb)m, respectively, for liquids. The exponents m and n are functions of the geometry and type of flow. It is to note that information on the flow through noncircular tubes with variable properties is not complete and hence it is recommended that the circular tube results be used. It has been found that nearly the same turbulence intensity and the friction factor prevail in circular and other duct geometries (such as annular, rectangular, square, triangular and irregular passages) if the ratio of flowpassage area to the wetted parameter is kept constant. This ratio is called the hydraulic radius Rh = A/P. However, it has been found convenient to use the term equivalent or hydraulic diameter Dh = 4A/P. It has been reported that the use of hydraulic diameter for ducts with sharp corners (e.g., triangular ducts) may lead to the error of the order of 35% in turbulent flow friction and heat transfer coefﬁcients determined from circular duct correlations. Fully developed Fanning friction factor for laminar flow in concentric annular duct results are presented in Table 8.6 as function of radius ratio. Fully developed laminar flow Nusselt number in concentric annular duct is independent of the Reynolds number and Prandtl number. There are two Nusselt numbers of interest, one for inner surface (when inner surface alone is heated) and other for the outer surface (when outer surface alone is heated). Their values for constant heat rate condition are given in Table 8.7. For the case of constant surface temperature on one surface (with the other surface insulated), the Nusselt number data are presented in Table 8.8. In the case of parallel plates, Nu = 7.541 when both surfaces are at the same constant temperature, while it is 8.235 for identical heat rates on the two surfaces of the duct for laminar flow. Average Nusselt number relation for the thermal entrance region flow between isothermal parallel plates for Re < 2800 has been presented as Eq. (8.27). For parallel plates hydraulic diameter is taken as twice the spacing between the plates. Discussion of various correlations and data of friction factor and Nusselt number for turbulent flow in circular tube annuli are presented in Sect. 8.7.2. The friction factors, as determined with water, were found to be 6−8% higher than the generally accepted values of smooth circular tube for 4000 < Re < 17000. The same for air flow were found to be about 1–10% higher than circular tube flow values. Approximate correlation of Kays and Perkins, f = 0.085 Re−0.25 for 6000 < Re < 300,000, which gives the friction factor in circular annuli about 10% higher than pipe flow values and having little dependence on the radius ratio may be used. It is reported that the annular flow develops more quickly than pipe flow. The development length is reported to be 20–30 diameters for a close approach to the fully developed flow. Equation (8.32), f = 0.1268 Re−0.3 for 5000 < Re < 1.2 106, and Eq. (8.33), f = 0.0868 −0.25 Re for 1.2 104 < Re < 1.2 106, are recommended to obtain the friction factor for the fully developed turbulent flow in the parallel plate duct.
8.13
Summary
675
Fully developed constant heat rate turbulent flow Nusselt number results for Pr = 0.7 as function of radius ratio of the annular duct are presented in Table 8.9 The solutions are presented for one surface heated and other insulated. It is recommended that the Nusselt number for turbulent flow between parallel plates can be determined using the circular duct correlations with hydraulic diameter in Nusselt number and Reynolds number as twice the spacing between the plates. Correlations for laminar flow friction factor and fully developed Nusselt numbers NuT for the case of uniform temperature at four walls and uniform heat flux at four walls constant along the length of the duct, but uniform temperature around the periphery NuH as function of duct aspect ratio (ratio of duct height to width) of rectangular duct have been presented in Sect. 8.8. Some typical results for the friction factor, and constant heat rate and surface temperature Nusselt numbers are listed in Table 8.10 along with circular and parallel plate ducts for comparison. The entrance conﬁguration (abrupt or smooth) exerts a marked influence on the value of critical Reynolds number Rec for flow in smooth rectangular duct. Another unique feature of the fully developed turbulent flow in rectangular ducts is the presence of secondary flow (flow normal to the axis of the duct), which exerts a signiﬁcant effect on the turbulence fluid flow characteristics of these ducts and increases the friction factor by approximately 10%. With uniform heating at four walls of a rectangular duct, circular duct Nusselt number correlations provide results with ± 9% accuracy while for the duct with equal heating at two long walls, circular duct correlations provide result with accuracy of ± 10% for 0.5 Pr 10 and 104 Re 105. In the case of heating at one long wall only (asymmetrical heating), the circular duct correlations provide values which may be up to 20% higher than the actual experimental values for 0.7 Pr 2.5 and 104 Re 106. Friction factor and Nusselt number correlations for laminar and turbulent external forced flow over a flat plate have been developed in Chap. 7 and are also presented in this chapter in Sect. 8.9. Forced laminar and turbulent flows across cylinders and spheres have been discussed in Sect. 8.10. Plots of drag coefﬁcient CD versus the Reynolds number for cylinder and sphere in cross flow have been presented in Fig. 8.10. Nusselt number correlations for heat transfer from cylinders have been presented in Eqs. (8.47)–(8.52). Correlation for the average Nusselt number between an isothermal spherical surface and free stream fluid is presented in Eq. (8.55). The flow across tube banks, the tube array or bundle is frequently encountered in heat exchangers. The inline and staggered arrangements of the tubes are frequently used. For both the arrangements, the flow across the ﬁrst transverse row of the tubes is practically the same as that for a single tube. The nature of the flow across the tubes of remaining rows depends on the way the tubes are arranged. However, in most applications, the flow is turbulent after a few rows. Correlations for the average Nusselt number for number of rows n 16 are presented in Eqs. ((8.58a), (8.58b), (8.58c), (8.58d) and (8.59a), (8.59b), (8.59c), (8.59d) for inline and staggered arrangements, respectively. For n < 16, the heat transfer coefﬁcient is lower due to the lower level of turbulence and the Nusselt number values from Eqs. (8.58a), (8.58b), (8.58c), (8.58d) and (8.59a), (8.59b), (8.59c), (8.59d) is multiplied by a correction factor.
676
8
Empirical Relations for Forced Convection Heat Transfer
Typical variation of the Nusselt number with the Peclet number Pe (= RePr) for flow of liquid metal through a tube with uniform heat flux is presented in Fig. 8.14. The Nusselt number is having a value of 4.36 for the laminar flow in tubes. The transition from the laminar to the turbulent flow occurs at critical value of Pe 40. At moderate value of the Peclet number beyond the critical value, the heat transfer by the turbulent mixing is small compared with the conductive transport. When the Peclet number is greater than 100, the heat exchange by turbulent mixing becomes appreciable and cannot be neglected. Some of the Nusselt number correlations for fully developed turbulent flow of liquid metals are given in Table 8.13. Although the Nusselt number tends to be low for the liquid metal , the heat transfer coefﬁcient is very high because of the high thermal conductivity of the liquid metals. Roughness is encountered in commercial pipes and artiﬁcial roughness, which creates turbulence close to the wall, has been used to enhance heat transfer. Nikuradse carried out experiments to study the relationship of friction factor to the Reynolds number for pipes of various roughnesses. At low Reynolds number (within the ﬁrst region termed as hydraulically smooth regime), the roughness has no effect on the resistance. Nikuradse noticed that the critical Reynolds number for all degrees of relative roughness occurs between 2160 and 2500. With an increasing Reynolds number, if the surface roughness height e is of the order of the magnitude of the laminar sublayer thickness dl, it tends to break up the laminar sublayer. This increases the wall shear stress and the ratio of the surface roughness height e and laminar sublayer thickness dl determines the effect of the roughness. The roughness Reynolds number e+ = √(f/2)Re(e/D) has been used to deﬁne three flow regimes in roughened ducts, namely (i) hydraulically smooth (0 e+ 5), (ii) transition regime (5 < e+ 70) and (iii) fully rough regime (e+ > 70). Moody determined the equivalent sandgrain roughness for different types of commercial pipe surfaces and also presented plot of friction factor versus the Reynolds number as function of relative pipe roughness e/D. Moody’s plot may be used to determine friction factor for flow in tubes. Several studies on heat transfer coefﬁcient behaviour for different roughness types have been carried out by researchers and correlations have been proposed. Some of them are presented in Table 8.14 for fully developed turbulent flow in the fully rough flow regime of a circular duct. It is to note that in the case of fluids with high Prandtl number, the resistance of the heat transfer is mainly concentrated in the laminar sublayer, which is thin compared to the hydrodynamic boundary layer. For the low Prandtl number fluids, the thermal resistance is distributed over a larger portion of the duct crosssection because the thermal boundary layer is thicker than the hydrodynamic boundary layer. Since the roughness at the wall creates turbulence near the wall, the heat transfer enhancement due to artiﬁcial roughness is greater for the fluids with high Prandtl numbers. The experimental results of heat transfer and fluid flow characteristics in roughened ducts have been presented either in the form of direct dependence of frition factor and Nusselt number on the system and operating parameters or in the form of interrelated roughness and heat transfer functions. This method makes it possible to present results in a most general form taking into account the various parameters involved including the roughness parameters.
8.13
Summary
677
Looking to the importance of this chapter in the design of heat exchangers, a signiﬁcant number of numerical problems have been included in the chapter to illustrate application of various correlations presented in the chapter.
Review Questions 8:1 Explain the principle of the dimensional analysis. What are the limitations of the dimensional analysis? 8:2 What are fundamental and derived dimensions? 8:3 Express the variables in Table 8.1 in terms of MLTh and MLThQ systems of fundamental dimensions. 8:4 List the variables that affect the forced convection heat transfer coefﬁcient. 8:5 Using the technique of dimensional analysis establish the following relation for forced convection heat transfer. Nu ¼ wðReÞ/ðPrÞ: 8:5 With the help of Buckingham pi theorem, show that for the forced convection Nu ¼ f ðRe; PrÞ: 8:6 At low flow velocities, the free convection effect may be present in forced convection heat transfer case. Using dimensional analysis, show that the following form of correlation is obtained. f ðNu; Re; Pr; GrÞ ¼ 0: 8:7 Derive the relationship (for forced convection) Nu ¼ f ðRe; Pr; EcÞ: 8:8 What is the physical interpretation of the following nondimensional numbers? Re; Pr; Nu; Pe; St; Ec: 8:9 Deﬁne hydraulic diameter. 8:10 Discuss the effect of the heating or cooling of a fluid on velocity distribution under laminar flow condition in a tube. How does this affect the heat transfer and friction factor and how this effect is taken into account? 8:11 Enlist various heat transfercoefﬁcient and frictionfactor correlations developed for laminar and turbulent flows in circular tubes. Discuss them. 8:12 What are the effects of thermal and hydrodynamic entry lengths on the heat transfer coefﬁcient and friction factor in both laminar and turbulent tube flows? How is this effect taken into account?
678
8
Empirical Relations for Forced Convection Heat Transfer
8:13 Obtain a relation between the convective heat transfer coefﬁcient and the friction factor when Pr = 1.
Problems 8:1 What is the effect of the following on the average value of the heat transfer coefﬁcient in the case of fully developed turbulent flow in tubes? (a) Two fold increase in the fluid mass flow rate; all other parameters remaining the same. (b) Two fold increase in the tube diameter; all other parameters including the flow velocity are the same. [Ans. For fully developed turbulent flow in tubes, Dittus Boelter relation may be 0:8 used, which gives Nu / Re0:8 ¼ qUlm d ; h ¼ Nu dk / Um0:8 d 0:2 ; Two fold increase in velocity will increase h h / Um0:8 by 20.8 = 1.741 times. A two fold increase in the diameter will reduce h h / d10:2 by 20.2 i.e., 14.86%.] 8:2 Water flows through a long electrically heated smooth tube of 30 mm diameter. If velocity and temperature proﬁles are fully developed at a location x = a, determine the heat transfer coefﬁcient and heat flux at the location. The mass flow rate of water is 0.5 kg/s. At x = a, water temperature is 30°C and wall temperature is 40°C. [Ans. Thermohydraulic properties of water at bulk temperature 30°C are: q = 995.6 4m kg/m3, k = 0.617 W/(m K), l = 7.97 10−4 N s/m2, Pr = 5.4; Re ¼ qUlm d ¼ pdl ¼
26625; Turbulent flow; h ¼ Nu dk ¼ 0:024Re0:8 Pr0:4 dk ¼ 3362 W/(m2 K); q 2 A ¼ hðtw tb Þ ¼ 33:62 kW/m :] 8:3 One kg/s of water at 35°C flows through a 25 mm diameter tube whose surface is maintained at a uniform temperature of 100°C. Determine the required length of the tube for 65°C water outlet temperature. [Ans. Water properties at the mean bulk temperature tm = (ti + to)/2 = 50°C from Table A4: q =988.1 kg/m3, c = 4182 J/(kg K), l = 544 10−6 N s/m2, k = 0.644 W/(m K) and md 4m 41 Pr = 3.55. Re ¼ ðp=4Þd 2 l ¼ pdl ¼ p0:025544106 ¼ 93620; From Dittus and Boelter 936200:8 3:550:4 ¼ 9735 W/(m2 equation, h ¼ dk 0:024Re0:8 Pr0:4 ¼ 0:644 0:025 0:024 10065 tw to P 14282 o K); From ttwwt L ¼ mc ln ln ¼ p0:0259735 10035 ¼ ti ¼ exp mc Lh , t t w i Ph
3:46m.] 8:4 Cold fluid is passing through a thinwalled tube 10 mm in diameter 2 m long whose surface is maintained at 100°C. The cold fluid flows at a rate of 0.05 kg/s and its inlet and outlet temperatures are 30 and 60°C, respectively. Determine the outlet temperature of the cold fluid if its flow rate is increased by 50% with all other conditions remaining the same. Given that dynamic viscosity of the fluid is 0.0004 N s/m2. Flow may be assumed to be fully developed. [Ans. tw to ¼ exp P Lh gives h ¼ m ln tw ti ¼ 0:05 ln 10030 ¼ 0:445; tw ti
mc
Reynolds number Re ¼
4m pdl
¼
PL c 40:05 p0:010:0004
tw to
p0:012
10060
¼ 15915: The flow is turbulent. For increase
in mass flow rate by 50%, the Reynolds number will increase by 50%. Since h / Re0:8
8.13
Summary
679
from DittusBoelter relation for fully developed turbulent flow, new 0:445 1:50:8 ¼ 0:616. The fluid outlet temperature for increased mass flow rate h=c h i h 100 ð100 30Þ exp p0:012 to ¼ tw ðtw ti Þ exp PL m c 1:50:05 0:616 ¼ 58:2 C:]
8:5 Air at 1 atmospheric pressure and 40°C is heated while it passes at a velocity of 8 m/s through a tube 1 m long and 40 mm in diameter whose surface is maintained at 120°C. Determine the outlet temperature of the air. [Ans. Mean bulk temperature, tm ¼ ti þ2 to ¼ 40 2þ to ; Assuming a trial value of tm = 50°C, the thermophysical properties of the air are: q = 1.0949 kg/m3, k = 0.02799 W/(m K), l = 1.9512 10−5 kg/(m s), cp = 1007.2 J/(kg K) and Pr = 0.703; Re ¼ qUlm D ¼ 17957; Flow is turbulent. DittusBoelter equation gives Nu = 0.024Re0.8Pr0.4 = 52.77; h ¼ Nu Dk ¼ 36:93 W/(m2 K); m ¼ q p4 D2 Um ¼ 0:011 kg/s; outlet temperature to ¼ tw h i h 36:93 ðtw ti Þ exp PL ¼ 120 ð120 40Þ exp p0:041 m c 0:011 1007:2 ¼ 67:38 C; Revised mean temperature tm = 53.69°C, retrial with this estimate of tm may be carried out.] 8:6 Air at atmospheric pressure and 25°C mean bulk temperature flows through a rectangular duct (height H = 400 mm and width W = 800 mm) with a mean velocity of 5 m/s. The duct is at an average temperature of 40°C. Determine the heat loss per unit length of the duct. [Ans. At 25°C bulk temperature, thermophysical properties of the air are: q = 1.1868 kg/m3, k = 0.02608 W/(m K), l = 1.8363 10−5 N s/m2, Pr = 0.709; 4WH Re ¼ qUlm Dh ¼ 1:71 105 ;Nu ¼ 0:024Re0:8 Pr0:4 ¼ 321:3; Dh ¼ 2ðW þ HÞ ¼ 0:53 m;
h ¼ Nu Dkh ¼ 15:81 W/(m2 K); Lq ¼ h½2ðW þ HÞðtw tb Þ ¼ 569:2 W/m:] 8:7 Air at atmospheric pressure and 100°C is flowing through a 20 mm diameter tube at a velocity of 20 m/s. The wall temperature is 30°C above the air temperature all along the tube length. Calculate the heat transfer rate per unit length of the tube. Assume fully developed flow condition. [Ans. At 100°C bulk temperature, thermophysical properties of the air are: q = 0.9452 kg/m3, k = 0.0317 W/(m K), l = 2.172 10−5 N s/m2, Pr = 0.693, cp = 1.0113 kJ/(kg K); Re ¼ qUlm d ¼ 17407; Flow is turbulent; h ¼ Nu dk ¼
0:024Re0:8 Pr0:4 dk ¼ 81:11 W/(m2 K); Lq ¼ hðpdÞðtw tb Þ ¼ 152:85 W/m:] 8:8 For flow of 0.04 m3/s of oil at 20°C bulk temperature through a 1.5 m long tube 125 mm in diameter kept at 30°C, determine the average heat transfer coefﬁcient. The property values are: k = 0.14 W/(m K), lb = 1.2 kg/(m s), lw = 0.6 kg/(m s), Pr = 20000, cp = 2000 J/(kg K), q = 890 kg/m3. qVd 4qV [Ans. ReL ¼ qUlm d ¼ ðp=4d Flow is laminar; 2 Þl ¼ pdl ¼ 302; Lhy ¼ 0:05Red ¼ 1:89m > 1.5 m; Lth ¼ 0:05Re Pr d ¼ 37:75 km [ 1:5 m; Flow is simultaneously developing; SiederTate relation may be used; d 0:33 l 0:14 2 b Nu ¼ 1:86 L Re Pr ¼ 156;h ¼ Nu dk ¼ 174:7 W/(m K):] l w
680
8
Empirical Relations for Forced Convection Heat Transfer
8:9 Air at atmospheric pressure and 20°C flows across a long cylinder of 50 mm diameter at a velocity of 40 m/s. The cylinder surface temperature is maintained at 100°C. Calculate the heat transfer rate per unit length of the cylinder. [Ans. At mean temperature tm = 60°C, air properties: q = 1.059 kg/m3, k = 0.02875 W/(m K), l = 1.997 10−5 kg/(m s), Pr = 0.701, cp = 1008 J/(kg K); Rex ¼ qUl1 d ¼ 106059; From Table 8.11, C = 0.027, n = 0.805; h ¼ Nu dk ¼ 8:10
8:11
8:12
8:13
0:027Re0:805 Pr1=3 dk ¼ 153:1 W/(m2 K); Lq ¼ hðpdÞðtw tb Þ ¼ 1923:9 W=m.] Air at 25°C flows at 10 m/s parallel to the surface of a highly polished aluminium plate flat plate maintained at a uniform temperature of ts = 75°C by a series of segmented heaters. Determine the heat removed from the section between x1 = 0.3 m and x2 = 0.4 m. The plate width is 0.3 m. The flow is turbulent throughout. [Ans. For air at ﬁlm temperature of 50°C from Table A5, q = 1.0949 kg/m3, l = 1.9512 10−5 kg/(m s), k = 0.02799 W/(m K) and Pr = 0.703; 0:8 1=3 ¼kx 0:0296 qUl1 x Pr1=3 ; Hence, hx1 ¼ 0:02799 hx ¼ kx Nux ¼ kx 0:0296Re0:8 x Pr 0:3 1:0949100:30:8 2 0:7031=3 ¼ 37:25 W/(m K), Similarly h2 = 35.17 W/(m2 0:0296 1:9512105 K); The average heat transfer coefﬁcient h12 for distance x1 = 0.3 m and x2 = 0.4 m is (37.25 + 35.17)/2 = 36.21; Heat transfer rate q = hx12 (x2 – x1) W (ts t∞) = 54.3 W.] Air at 1 atm, 25°C and 5 m/s is in cross flow over a long cylinder of 30 mm diameter. Determine the drag force per unit length of the cylinder. [Ans. For air at 25°C from Table A5, q = 1.1868 kg/m3, l = 1.8363 10−5 kg/(m s); ReD ¼ qUl1 D ¼ 1:186850:03 5 ¼ 9694; From Fig. 8.10, CD 1; From Eq. (8.46), FD ¼ 1 2 1:836310 C A qU ; where A = LD; Substitution gives FD ¼ 1 ð1 0:03Þ 1D 2 1 2 ¼ 0:445 N=m.] 2 1:1868 5 Water at 25°C and 10 m/s flows over a sphere of 10 mm diameter. Surface temperature of the sphere is 75°C. Determine the drag force. What will be the drag force if fluid is air? [Ans. For water at ﬁlm temperature of 50°C from Table A4, q = 988.1 kg/m3, ¼ 1:8 105 ; From Fig. 8.10, l = 544 10−6 kg/(m s); ReD ¼ qUl1 D ¼¼ 988:1100:01 6 1 54410 2 ; where A = frontal area = (p/4)D2; CD 0.4; From Eq. (8.46), FD ¼ CD A 2 qU1 1 p 2 Substitution gives FD ¼ 0:4 4 0:01 2 988:1 102 ¼ 1:55N. For air at ﬁlm temperature of 50°C from Table A5, q = 1.0949 kg/m3, l = 19.512 10−6 kg/(m s); ¼ 5:6 103 ; From Fig. 8.10, CD 0.4; From Eq. (8.46), ReD ¼ qUl1 D ¼1:0949100:01 6 1 19:51210 2 FD ¼ CD A 2 qU1 ; where A = frontal area = (p/4)D2; Substitution gives FD ¼ 0:4 p4 0:012 12 1:0949 102 ¼ 0:0017N. Comment: In the present case, drag force associated with water is signiﬁcantly higher than for air because of higher density of water.] Air at 1 atm and 25°C flows at 7.5 m/s over inline tube bundle with p = pt = 25 mm. The bundle contains 10 tubes per row. The tube diameter is 12.5 mm. If air outlet temperature is 225°C and tube surface temperature is 400°C, determine the number of rows of the tubes per m length.
8.13
Summary
681
[Ans. Air properties at the mean bulk temperature tm [= (t∞i + t∞o)/2 = 125°C] are: q = 0.8872 kg/m3, c = 1013.8 J/(kg K), l = 2.2776 10−5 kg/(m s), k = 0.0335 W/ t (m K) and Pr = 0.689. At tw = 400°C, Prw = 0.683; Umax ¼ U1 pt pD ¼ 15 m=s; D ¼ 7304; ReD ¼ qUmax l
Assuming
N > 16, from Eq. (8.58 0:36 c),Nu ¼ 0:27Re0:63 ðPr = Prw Þ ¼ 64:3; h ¼ Dk Nu ¼ 172:3 W=ðm2 KÞ; From 7. D Pr pDN 1o 16, t1o ¼ tw ðtw t1i Þ exp mcP p Lh , or ln ttwwt t1i ¼ qU1 Nt pt cp Lh, which gives t1i qU1 Nt pt cp N ¼ ln ttwwt ¼ 190 for L = 1 m; No. of rows = N/Nt = 190/10 = 19 > 16.] 1o pDLh 0:25
8:14 Liquid mercury at 20°C enters a metal tube of 20 mm internal diameter at the rate of 1 kg/s and is to be heated to 30°C. The tube wall is at a constant temperature of 40°C. Determine the length of the tube. Given for the mercury: q = 13560 kg/m3, k = 8.7 W/ (m K), l = 1.5 10−3 kg/(m s), Pr = 0.025, cp = 139 J/(kg K). 4m 41 [Ans. Reynolds number, Re ¼ qUlm d ¼ pdl ¼ p0:021:510 3 ¼ 42441; Pe ¼ Re Pr¼ 42441 0:025 ¼ 1061; For constant wall temperature, Seban and Shimazaki equation, Table 8.13, gives 8:7 h ¼ Nu dk ¼ ð5:0 þ 0:025Pe0:8 Þ dk ¼ ½5:0 þ 0:025 ð1061Þ0:8 0:02 ¼ 5039 W/(m2 K); tw ti mcp PL o For isothermal case, ln ttwwt h, which gives L ¼ ln ¼ mcp ti tw to Ph 1:0139 ¼ ln 4020 ¼ 0:304m.] 4030 p0:025039
References Bejan A, Kraus AD (eds) (2003) Heat transfer handbook. Wiley, New York Bhatti MS, Shah RK (1987) Turbulent and transition flow convective heat transfer. In: Kakac S, Shah RK, Aung W (eds) Handbook of singlephase convective heat transfer, Chap. 4, Wiley, New York Churchill SW (1977) Comprehensive correlating equations for heat, mass and momentum transfer in fully developed flow in smooth tubes. Ind Eng Chem Fundam 16(1):109–116 Churchill SW, Bernstein M (1977) A correlating equation for forced convection from gases and liquids to a circular cylinder in crossflow. J Heat Transfer 99:300–306 Dalle Donne M, Meerwald E (1973) Heat transfer and friction coefﬁcients for turbulent flow in smooth annuli at high temperatures. Int J Heat Mass Transfer 16:787–809 Dippery DF, Sabersky RH (1963) Heat and momentum transfer in smooth and rough tubes at various Prandtl numbers. Int J Heat Mass Transfer 6:329–353 Douglas WJM, Churchill SW (1956) Recorrelation of data for convective heat transfer between gases and single cylinders with large temperature differences. Chem Eng Prog Symp Ser 52(18):23–28 Ebadian MA, Dong ZF (1998) Forced convection, internal flow in ducts. In: Rohesenow WM, Hartnett JP, Cho YI (eds) Handbook of heat transfer. McGraw Hill Book Co, New York Eckert ERG, Drake RM Jr (1972) Analysis of heat and mass transfer. McGrawHill Kogakusha, Tokya Gnielinski V (2015) Turbulent heat transfer in annular spaces a new comprehensive correlation. Heat Transfer Eng 36:787–789 Gowen RA, Smith JW (1968) Heat transfer from smooth and rough surfaces. Int J Heat Mass Transfer 11:1657–1673 Heaton HS, Reynolds WC, Kays WM (1964) Heat transfer in annular passages: simultaneous development of velocity and temperature ﬁelds in laminar flow. Int J Heat Mass Transfer 7:763–781 Hilpert R (1933) Heat transfer from cylinders. Forsch Geb Ingenieurwes 4:215–224 Holman JP (1992) Adapted for SI units by White PRS. Heat transfer. McGrawHill Book Co, New York Hsu ST (1963) Engineering heat transfer. D. Van Nostrand, New York Isachenko VP, Osipova VA, Sukomel AS (1977) Heat transfer, 2nd edn. MIR Publishers, Moscow
682
8
Empirical Relations for Forced Convection Heat Transfer
Kays WM, Crawford ME (1980) Convective heat and mass transfer. Tata McGrawHill Publishing Co., Ltd, New Delhi Kays WM, Leung EY (1963) Heat transfer in annular passageshydrodynamically developed turbulent flow with arbitrarily prescribed heat flux. Int J Heat Mass Transfer 6:537–557 Kays WM, Perkins HC (1973) Forced convection, internal flow in ducts. In: Rohsenow WM, Hartnett JP (eds) Handbook of heat transfer, Chap. 7. McGrawHill, New York Knudsen JD, Katz DL (1958) Fluid dynamics and heat transfer. McGrawHill, New York Langhaar HL (1942) Steady flow in the transition length of a straight tube. J Appl Mech Trans ASME 64: A55–A58 Lee Y (1968) Turbulent heat transfer from the core tube in thermal entrance regions of concentric annuli. Int J Heat Mass Transfer 11:509–522 Lundgren TS, Sparrow EM, Starr JB (1964) Pressure drop due to the entrance region in ducts of arbitrary crosssection. J Basic Eng 86(3):620–626 Mikheyev M (1964) Fundamental of heat transfer. Mir Publishers, Moscow Mills AF (1995) Heat and mass transfer. Richard D, Irwin, Chicago Moody LF (1944) Friction factors for pipe flow. Trans ASME 66:671–684 Nikuradse J (1950) Laws of flow in rough pipes. NACA, Technical Memorandum 1292 Rajendra Karwa, Bairwa RD, Jain BP, Karwa N (2005) Experimental study of the effect of rib angle and discretization on heat transfer and friction in an asymmetrical heated rectangular duct. J Enhanced Heat Transfer 12(4):343–355 Ranz WE, Marshall WR Jr (1952a) Evaporation from drops. Parts I. Chem Eng Progr 48:141–146 Ranz WE, Marshall WR Jr (1952b) Evaporation from drops. Parts II. Chem Eng Progr 48:173–180 Rogers GFC, Mayhew YR (1967) Engineering thermodynamics work and heat transfer. English Language Book Co and Longmans, Green and Co, London Schlichting H (1960) Boundary layer theory (trans: Kestin J), 4th edn. McGrawHill, New York Schlichting H, Gersten K (2000) Boundarylayer theory (trans: Katherine Mayes). Springer, Heidelberg Seban RA, Shimazaki TT (1951) Heat transfer to a fluid flowing turbulently in a smooth pipe with walls at constant temperature. Trans ASME 73:803–807 Sieder EN, Tate CE (1936) Heat transfer and pressure drop of liquids in tubes. Ind Eng Chem 28:1429–1435 Skupinshi E, Tortel J, Vautrey L (1965) Determination des coefﬁcients de convection a’un alliage sodiumpotassium dans un tube circulaire. Int. J. Heat Mass Transfer 8:937–951 Sleicher CA, Rouse MW (1975) A convenient correlation for heat transfer to constant and variable property fluids in turbulent pipe flow. Int J Heat Mass Transfer 18:677–683 Sleicher Jr. CA (1955) Heat transfer in a pipe with turbulent flow and arbitrary walltemperature distribution. Ph.D. Thesis, University of Michigen, p. 67 Sparrow EM, Lloyd JR, Hixon CW (1966) Experiments on turbulent heat transfer in an asymmetrically heated rectangular duct. J Heat Transfer 88:170–174 Tam LM, Ghajar AJ (2006) Transitional heat transfer in plain horizontal tubes. Heat Transfer Eng 27(5):23– 28 Webb RL, Eckert ERG, Goldstein RJ (1971) Heat transfer and friction in tubes with repeatedrib roughness. Int J Heat Mass Transfer 14:601–617 Whitaker S (1972) Forced convection heat transfer correlations for flow in pipes, past flat plates, single cylinders, single spheres, and flow in packed beds and tube bundles. AIChE J 18(2):361–371 Winterton RHS (1998) Where did the Dittus and Boelter equation come from? Int J Heat Mass Transfer 41:809–810 Witte LC (1968) An experimental study of forcedconvection heat transfer from a sphere to liquid sodium. J Heat Transfer 90:9–12 Zukauskas AA (1987) Convective heat transfer in crossflow. In: Kakac S, Shah RK, Aung W (eds) Handbook of singlephase convective heat transfer, Chap. 6. Wiley, New York
9
Empirical Relations for Natural or Free Convection
9.1
Introduction
The development of boundary layer along a flat vertical plate under natural or free convection condition has been presented in Chap. 7. Figure 9.1 illustrates the nature of flow in freeconvection near heated horizontal plates and tubes of different dimensions. In case of a large horizontal plate with heated surface facing upwards, the central portion of the plate is practically isolated due to the flow established at the edges of the plate. The central portion receives the cold fluid by flow of fluid from above. If the heated surface of the plate is facing downward, the fluid coming in contact with the heated surface cannot move directly in the upwards direction but only along the edges of the plate. Therefore, only a diminished level of convection flow is possible in this case. The flow structure near the horizontal cylinders differs signiﬁcantly from the flow along a flat vertical wall. For thin wires (d < 0.2–1.0 mm), the laminar flow is preserved throughout. At very small temperature differences, almost a stationary ﬁlm of heated fluid may form around the wire.
9.2
Buoyancy Force in Natural Convection
In the free convection, the motion of the fluid is caused by the buoyancy force arising from the change in the density of the fluid due to the rise in the temperature. Consider any fluid element, see Fig. 9.2, which is at a temperature DT above the surrounding fluid. If qo is the density of the surrounding fluid and q is the density of the fluid element, then qo ¼ qð1 þ bDTÞ
ðiÞ
where b is the coefﬁcient of cubical expansion. The buoyancy force acting upwards is qo Vg
© Springer Nature Singapore Pte Ltd. 2020 R. Karwa, Heat and Mass Transfer, https://doi.org/10.1007/9789811539886_9
ðiiÞ
683
684
9 Empirical Relations for Natural or Free Convection
Heated side Plate
Heated side
Fig. 9.1 Free convection flow
ρoVg
Surrounding fluid
ρ ρo
Hot fluid element
mg = ρVg
Fig. 9.2 Forces on a fluid element under free convection
The gravitational force qVg acts downwards. The net force F acting on the fluid element is F ¼ qo Vg qVg ¼ Vgðqo qÞ
ðiiiÞ
Using Eq. (i), we obtain F ¼ qVgbðDTÞ
ð9:1Þ
Thus the net force acting upwards per unit mass is bgDT. This force causes the upwards motion of the fluid element.
9.3
Dimensional Analysis Applied to Natural Convection
From the experience, it has been found that the heat transfer coefﬁcient in natural convection depends on the viscosity l, thermal conductivity k, density q, temperature difference DT and the characteristic dimension d. Since the net buoyancy force per unit mass is gbDT, the product gb is also included as a variable. Hence we may write
9.3 Dimensional Analysis Applied to Natural Convection
685
h ¼ f ðl; q; k; cp ; DT; bg; dÞ
ð9:2aÞ
h ¼ Cla qb kc cdp DT e bg f d g
ð9:2bÞ
or
where C is a constant of proportionality and a, b, c, etc. are arbitrary indices.
9.3.1 Rayleigh’s Method The dimensions of different physical quantities have been listed in Table 8.1. The dimensions of quantities b and g are given in Table 9.1. Substituting the dimensions of different variables in Eq. (9.1), we have a b Q M M Q c Q d e L f g ¼C ð hÞ ðLÞ L2 Th LT L3 LTh Mh hT 2 Equating the indices of the fundamental dimensions on both sides, M L T h Q
0 ¼ aþb d 2 ¼ a 3b c þ f þ g 1 ¼ a c 2f 1 ¼ c d þ e f 1 ¼ cþd
There are seven unknowns and ﬁve equations. Therefore the values of ﬁve of them, say a, b, c, e, and g may be determined in terms of the other two unknowns d and f, which gives a ¼ d 2f ; b ¼ 2f ; c ¼ 1 d; e ¼ f ; and g ¼ 1 þ 3f Substituting the values in Eq. (9.2b), we get h ¼ Cld2f q2f k1d cdp DT f ðbgÞ f d1 þ 3f 2 f k q bgDTd 3 lcp d ¼C d l2 k
Table 9.1 Dimensions of different physical quantities Variable Coefﬁcient of cubical expansion Acceleration due to gravity
Symbol (units) b (1/K) g (m/s2)
Dimensions ML Th Q system −1
h LT−2
MLTh system h −1 LT−2
686
9 Empirical Relations for Natural or Free Convection
or hd ¼f k
"
q2 bgDTd3 l2
The dimensionless group
q2 bgDTd3 l2
f # lcp d k
is termed as the Grashof number Gr. Hence Nu ¼ f ðGr; PrÞ
The generalized correlation can be written as Nu ¼ wðGrÞ/ðPrÞ
ð9:3Þ
The form of functions w(Gr) and /(Pr) may be speciﬁed for different conditions of heat transfer by convection on the basis of theoretical analysis or experimental investigations.
9.3.2 Buckingham’s Pi Method f ðl; q; k; cp ; DT; bg; d; hÞ ¼ 0 There are eight variables and ﬁve fundamental units hence we expect (8 – 5), i.e., 3 pterms. Taking l, k, bg, DT and d as repeated variables, the pterms can be established as follows. p1 ¼ la kb ðbgÞc DT d de h
ðiÞ
or 1¼
M LT
a
Q LTh
b
L hT 2
c
ðhÞd ðLÞe
Q L2 Th
Equating the indices of the fundamental dimensions on both sides, we obtain M: L: T: h: Q:
0¼a 0 ¼ a b þ c þ e 2 0 ¼ a b 2c 1 0 ¼ b c þ d 1 0 ¼ bþ1 Solution gives a ¼ 0; b ¼ 1; c ¼ 0; d ¼ 0; and e ¼ 1: Substitution in Eq. (i) gives
9.3 Dimensional Analysis Applied to Natural Convection
687
hd ¼ Nu k
ðiiÞ
p2 ¼ la kb ðbgÞc DT d d e q
ðiiiÞ
p1 ¼ k1 dh ¼ Following the above procedure,
or 1¼
M LT
a
Q LTh
b
L hT 2
c
ðhÞd ðLÞe
M L3
Equating the indices, M: L: T: h: Q:
0 0 0 0 0
= = = = =
a+1 −a − b + c + e  3 −a − b – 2c −b − c + d b
Solution gives a ¼ 1; b ¼ 0; c ¼ 1=2; d ¼ 1=2; and e ¼ 3=2: This gives 1
p2 ¼ l ðbgÞ
1=2
DT
1=2 3=2
d
q¼
q2 bgDTd 3 l2
1=2 ðivÞ
Similarly, p3 ¼ la kb ðbgÞc DT d d e cp
ðvÞ
or 1¼ Equating the indices, M: L: T: h: Q:
0 0 0 0 0
= = = = =
a−1 −a − b + c + e −a − b – 2c −b − c + d − 1 b+1
M LT
a
Q LTh
b
L hT 2
c
ðhÞd ðLÞe
Q Mh
688
9 Empirical Relations for Natural or Free Convection
Solution of the above equations gives a ¼ 1; b ¼ 1; c ¼ 0; d ¼ 0; and e ¼ 0: This gives p3 ¼ lk1 cp ¼
lcp ¼ Pr k
ðviÞ
Thus the functional relation is f ðNu; Gr; PrÞ ¼ 0 The generalized correlation can be written as Eq. (9.3): Nu ¼ wðGrÞ/ðPrÞ
ð9:3Þ
Example 9.1 Using M, L, T and h system of fundamental units, develop the functional relation for free convection heat transfer. Solution Buckingham’s pi method Analysis in previous section has shown that (bg) and (ΔT) terms appear together in the ﬁnal functional relation for free convection hence we write f ðl; q; k; cp ; bgDT; d; hÞ ¼ 0
ðiÞ
There are seven variables and four fundamental units hence we expect (7 – 4), i.e., 3 pterms. Taking l, q, k, and d as repeated variables, the pterms can be established as follows. p1 ¼ la qb kc dd ðbgDTÞ or 1¼
M LT
a b M ML c d 1 L ð L Þ h L3 hT 3 h T2
Equating the indices of the fundamental dimensions on both sides, we obtain M L T h
0 0 0 0
= = = =
a+b+c −a − 3b + c + d + 1 −a − 3c − 2 −c
ðiiÞ
9.3 Dimensional Analysis Applied to Natural Convection
689
Solution gives a ¼ 2; b ¼ 2; c ¼ 0 and d ¼ 3: Substitution gives p1 ¼ l2 q2 d3 bgDT ¼
q2 ðbgDTÞd3 ¼ Gr l2
ðiiiÞ
Following the above procedure, p2 ¼ la qb kc dd cp
ðivÞ
or 1¼
M LT
a b M ML c d L2 ð L Þ ; L3 hT 3 hT 2
which gives p2 ¼
lcp ¼ Pr k
ðvÞ
Similarly, p3 ¼ la qb kc dd h;
ðviÞ
which gives p3 ¼
hd ¼ Nu k
ðviiÞ
Thus the functional relation is f ðNu; Gr; PrÞ ¼ 0 The generalized correlation can be written as Eq. (9.3) Nu ¼ wðGrÞ/ðPrÞ
ð9:3Þ
h ¼ Cla qb kc cdp ðbgDTÞe d f
ðviiiÞ
Rayleigh’s Method
or
M hT 3
M ¼C LT
e a b d M ML c L2 1 L h ðLÞ f L3 hT 3 h T2 hT 2
690
9 Empirical Relations for Natural or Free Convection
Equating the indices of the fundamental dimensions on both sides, we obtain M L T h
1=a+b+c 0 =  a  3b + c + 2d + e + f 3 =  a  3c  2d 2e 1 =  c  d
We have four equations but six unknowns, so we shall determine the values of a, c, e and f in the terms of b and d. Thus a ¼ d b; c ¼ 1 d; e ¼ b=2 and f ¼ ð3=2Þb 1: Substitution gives h ¼ Cldb qb k1d cdp ðbgDTÞb=2 d3=2b1 b=2 k q2 ðbgDTÞd3 lcp d ¼C d l2 k or 2 b=2 hd q ðbgDTÞd3 lcp d ¼C 2 k l k Thus the functional relation is Nu ¼ wðGrÞ/ðPrÞ
ð9:3Þ
9.3.3 Physical Interpretation of Grashof Number As mentioned in Chap. 7, the Grashof number Gr can be interpreted physically as a dimensionless number representing the ratio of the buoyancy to viscous forces in free convection, i.e. Gr ¼
q2 gbDTd3 l2
¼
Buoyancy force Viscous force
The relation Nu = f (Gr, Pr) suggest that the Grashof number plays a part in the free convection analogous to the part played by the Reynolds number in forced convection. The product (Gr Pr), called the Rayleigh number Ra, serves as a criterion of turbulence in free convection. However, the value of the product for the onset of turbulence depends on the geometric conﬁguration. In the case of low velocity flows, the free convection may play a signiﬁcant role in the forced convection hence the functional relation is presented as (refer Chap. 8) Nu ¼ wðRe; Gr; PrÞ
ð8:6Þ
9.4 Experimental Determination of Natural Convection Heat Transfer Coefficient
9.4
691
Experimental Determination of Natural Convection Heat Transfer Coefficient
A simple experimental setup for the experimental investigation of natural convection heat transfer from a long horizontal tube is shown in Fig. 9.3a. The heat of the electric heater flows in radial direction to the surface of the tube. The heat loss from the tube surface occurs both by convection (qc) and radiation (qr) to the surroundings. The total heat loss rate from the tube is q ¼ qc þ qr
ðiÞ
The convection heat transfer is given by qc ¼ hAðTw T1 Þ
ðiiÞ
and the radiation heat transfer is
Copper tube
Side insulation
Heater V A Variac
Supply
a Edge insulation Plate
Heater Insulation Plate
c Thermocouples
b Fig. 9.3 Experimental setup for determination of natural convection heat transfer coefﬁcient
692
9 Empirical Relations for Natural or Free Convection
qr ¼ erAðTw4 Ts4 Þ
ðiiiÞ
where q h A Do L Tw T∞ Ts e
electric power input heat transfer coefﬁcient surface area of the tube = pDoL tube outer diameter length of the tube tube wall temperature surrounding air temperature surrounding surface temperature for radiation heat exchange emissivity of the tube surface
Substitution in Eq. (i) from Eqs. (ii) and (iii) gives q ¼ hAðTw T1 Þ þ erAðTw4 Ts4 Þ:
ð9:4Þ
Equation (9.4) is solved for the heat transfer coefﬁcient. In the above analysis, the heat loss from the ends of the tube has not been considered assuming it to be negligible. This condition can be fulﬁlled when the tube is sufﬁciently long (say 2–3 m) and the ends are properly insulated. In the case of short tubes, guard heaters are installed at the ends of the tubes. The wall temperature Tw is average of the readings of sufﬁcient number of thermocouples afﬁxed axially and circumferentially on the tube wall surface. All readings are taken when the steady state condition is achieved. The experiments are conducted at different heating rates resulting in different values of wall temperature. The result is presented in the form of relation Nu = f(Gr Pr). The experimental scheme presented above can be extended for determination of heat transfer coefﬁcient from vertical and horizontal plates, refer Fig. 9.3b, c. Example 9.2 An electrically heated 300 mm 300 mm square metal plate is used for an experimental determination of natural convection heat transfer coefﬁcient. The plate is placed in vertical position and is exposed to room temperature at 25°C. The plate is heated to a uniform temperature of 60°C. The electric power input is measured as 40 W. The emissivity of the surface has been estimated as 0.1. Determine the heat transfer coefﬁcient. Solution In the equilibrium, the heat lost from one side of the plate is 20 W. Hence, 4 q ¼ hAðTw T1 Þ þ eArðTw4 T1 Þ
or 20 ¼ h 0:3 0:3 ð333 298Þ þ 0:1 0:3 0:3 5:67 108 ð3334 2984 Þ; assuming surrounding surface temperature for radiation heat exchange to be equal to the surrounding air temperature.
9.4 Experimental Determination of Natural Convection Heat Transfer Coefficient
693
Solution of the equation gives h ¼ 5:635 W/m2 K:
9.5
Empirical Relations for Free or Natural Convection
9.5.1 Vertical Plate and Cylinders (i) Uniform Temperature Condition (Isothermal Surface): A vertical cylinder can also be treated as a vertical plate under laminar flow conditions if the boundary layer thickness is much less than the diameter of the cylinder. This condition is satisﬁed when D C 1=4 L GrL
ð9:5Þ
where D is the diameter of the cylinder, and C = 35 for fluids of Pr 0.72 and C 25 for fluids of Pr 6 [Cebeci (1974) in Popiel (2008)]. However, some references (Holman 1992; Cengel 2007. Incropera et al. 2006) use a single value of C = 35. The plate results are accurate to within 5%. The reason for deviation is the curvature effect. When the thermal boundary layer thickness in much thinner than the radius of the cylinder, the curvature can be neglected and plate relation may be used. The vertical plane relation for laminar flow with length L as characteristic dimension (McAdams, 1954) is Num ¼ 0:59Ra1=4 for 104 Ra 109
ð9:6Þ
Various terms in the above relation are deﬁned as Num ¼ Ra ¼ GrPr ¼
hm L k
bgðTw T1 ÞL3 Pr m2
ð9:7Þ ð9:8Þ
where hm is average heat transfer coefﬁcient, Tw is surface or wall temperature and T∞ is free stream temperature. b and thermophysical properties of the fluid are taken at ﬁlm temperature. The subscript ‘m’ refers to the mean or average value. For turbulent flow (Ra 109), the correlation suggested (Bailey in Kays and Crawford 1980) is Num ¼ 0:1Ra1=3 for 109 Ra 1013
ð9:9aÞ
694
9 Empirical Relations for Natural or Free Convection
For turbulent flow over a slender vertical cylinder (Ra 4 109), the mean Nusselt number for air can be approximated from the following correlation of McAdams (1954) as suggested by Popiel (2008) for L/D from about 70 to 136. Num ¼ 0:13Ra1=3
ð9:9bÞ
Churchill and Chu (1975b) gave the following relations of mean Nusselt number for a wide range of Ra, which is a complex relation but more accurate.
Num ¼
8 > < > :
0:825 þ h
0:387Ra
92 > =
1=6
1 þ ð0:492= PrÞ9=16
i8=27 > ;
ð9:10Þ
for 101 Ra 1012 Equation (9.10) is also valid for constant heat flux when factor 0.492 is replaced by 0.437 (Baehr and Stephan 2011). In the laminar region, the following relation gives better accuracy. Num ¼ 0:68 þ h
0:670Ra1=4 1 þ ð0:492= PrÞ9=16
i4=9
ð9:11Þ
for Ra 109
(ii) Uniform Heat Flux (qw = Constant) The suggested equation of local Nusselt number is (Baehr and Stephan 2011) Nux ¼
0:616Ra1=5 x
Pr 0:8 þ Pr
1=5 ð9:12Þ
for Pr 0:1; Rax 109 where Nux = hxx/k; Rax = GrxPr = [bgqwx4/(m2k)].Pr. Alternative equation for local Nusselt number in laminar region (Holman 1992) is ðNux Þf ¼
1=5 hx ¼ 0:60 Grx Prf kf for 105 Grx Prf 1011
4
ð9:13Þ
wx where Grx ¼ Grx Nux ¼ gbq km2 . The transition begins between Grx Pr = 3 1012 − 4 1013 and ends between 2 1013− 4 1014. Fully developed turbulent flow is present beyond Grx Pr = 1014 (Holman 1992). For the turbulent region, the recommended relation is
9.5 Empirical Relations for Free or Natural Convection
1=4 ðNux Þf ¼ 0:17 Grx Prf for 2 1013 Grx Pr 1016
695
ð9:14Þ
Subscript f in Eqs. (9.13) and (9.14) indicates that the fluid properties are to be evaluated at the local ﬁlm temperature Tf [= (Tw + T∞)/2]. The average or mean heat transfer coefﬁcient is hm ¼
1 L
Z
L
hx dx
0
Thus, for the laminar flow, Eqs. (9.12) and (9.13) yield 5 hm ¼ hx¼L 4
ð9:15Þ
It can be shown that in the case of turbulent flow condition, the local heat transfer coefﬁcient is constant with x.
9.5.2 Inclined Plate When a plate is inclined, the buoyancy force parallel to the surface is reduced and there is reduction in the fluid velocity along the plate causing reduction in the convection heat transfer. For 0 h 60° correlation of Churchill and Chu (1975b), Eq. (9.10), is valid for the inclined plate when the acceleration due to gravity g is replaced by its component parallel to the wall g cosh, where h is the angle of inclination with the vertical. This approach is satisfactory only for the top and bottom surfaces of cooled and heated plates, respectively (Incropera et al. 2012). For opposite surfaces, refer Holman (1992).
9.5.3 Horizontal Plate For a horizontal plate, the characteristic length L is deﬁned as A P
L¼
ð9:16Þ
where A the heat transfer area and P is the perimeter of the plate. When the upper side of the plate is heated or lower side is cooled, refer Fig. 9.1, 1=4
Num ¼ 0:54RaL
for 104 Ra 107
ð9:17Þ
1=3
Num ¼ 0:15RaL
for 107 Ra 1011
ð9:18Þ
696
9 Empirical Relations for Natural or Free Convection
The equation of mean Nusselt number, when upper side of the plate is cooled or lower side is heated is 1=4
Num ¼ 0:27RaL
ð9:19Þ
for 105 Ra 1011
(i) Uniform Heat Flux Condition (qw = Constant) (Fujii and Imura 1972 in Holman 1992) For the heated surface facing upwards, 1=3
Num ¼ 0:13RaL
ð9:20Þ
for Ra 2 108 1=3
Num ¼ 0:16RaL for 2 108 Ra 1011
ð9:21Þ
For the heated surface facing downwards, 1=5
Num ¼ 0:58RaL
ð9:22Þ
for 106 Ra 1011
All properties except b, are evaluated at reference temperature Te = Tw – 0.25(Tw – T∞). Tw is the wall temperature related to the heat flux by hm = qw/(Tw  T∞).
9.5.4 Horizontal Cylinder of Diameter D and Length L > > d The mean Nusselt number equation for this case is given by 1=4
Num ¼ 0:53Rad
ð9:23Þ
for 104 Ra 109 1=3
Num ¼ 0:13Rad
ð9:24Þ
for 109 Ra 1012
Churchill and Chu (1975a) presented the following correlation of average or mean Nusselt number for the laminar flow (Rad 109) Num ¼ 0:36 þ h
1=4
0:518Rad
1 þ ð0:559= PrÞ9=16
i4=9
ð9:25Þ
where characteristic dimension for the Nusselt number and Rayleigh number is the diameter of the cylinder.
9.5 Empirical Relations for Free or Natural Convection
697
For turbulent flow (Rad 109), the following correlation may be used.
Num ¼
8 > < > :
0:60 þ h
92 > =
1=6 0:387Rad
1 þ ð0:559= PrÞ9=16
i8=27 > ;
ð9:26Þ
However, Churchill and Chu (1975a,b) recommend use of Eq. (9.26) for all Pr and Ra values.
9.5.5 Sphere of Diameter d The characteristic dimension in the equations of the mean Nusselt number and Rayleigh number is the diameter of the sphere and the suggested equation for the average Nusselt number for Pr 0.7 is (Churchill 1990) Num ¼ 2 þ h
1=4
0:589Rad
1 þ ð0:469= PrÞ9=16
i4=9
ð9:27Þ
for Rad 1011 Example 9.3 For a vertical wall at 50°C exposed to still air at 20°C, what is the maximum height for laminar free convection? Solution At tm = (50 + 20)/2 = 35°C, the air properties are m ¼ l=q ¼ 16:36 106 m2 =s; Pr ¼ 0:7066 and b ¼ 1=ð35 þ 298Þ K1 : Grx Pr ¼
gbðtw t1 Þx3 : Pr m2
or x¼
Rax m2 gbðtw t1 Þ Pr
1=3 ðiÞ
The limit for laminar flow is Rax 109. Substitution of the values of various terms in Eq. (i) gives "
109 ð16:36 106 Þ2 x¼ 9:81 ð1=303Þ ð50 20Þ 0:7066
#1=3 ¼ 0:731 m:
698
9 Empirical Relations for Natural or Free Convection
Example 9.4 Determine the coefﬁcient of heat transfer from a vertical plate of 2 m height to the surrounding still air, which at a distance from the plate is at 20°C. The mean surface temperature of the plate is measured to be 120°C. Solution At the mean fluid temperature tm = (120 + 20)/2 = 70°C, the physical properties of air are: k ¼ 0:0295 W=ðm KÞ; m ¼ l=q ¼ 2 105 m2 =s; Pr ¼ 0:698 and b ¼ 1=T ¼ 1=343 K1 : Hence, Gr Pr ¼
bgDTL3 1=343 9:81 ð120 20Þ 23 Pr ¼ 0:698 ¼ 3:99 1010 109 : m2 ð2 105 Þ2
Equation (9.9a) applies, which gives Nu ¼ 0:1 ð3:99 1010 Þ1=3 ¼ 341:4 and heat transfer coefﬁcient, h¼
Nu k 341:4 2:95 102 ¼ ¼ 5:04 W/(m2 K): L 2
Example 9.5 A vertical plate is rejecting heat by free convection from one of its face which is at an average temperature of 90°C. The surrounding air is at 30°C. The height of the plate is 200 mm. Calculate the heat rejected from the plate if the width of the plate is 600 mm. The following correlation for the local Nusselt number may be used. Nu x ¼ 0:52
Pr 0:95 þ Pr
0:25
ðGr PrÞ0:25 :
Solution From the given data, tw þ t1 tm ¼ ¼ 2
90 þ 30 ¼ 60o C: 2
The thermophysical properties of air are: q ¼ 1:059 kg=m3 ; k ¼ 0:02875 W=ðmKÞ; l ¼ 2 105 N s=m2 ; Pr ¼ 0:7:
9.5 Empirical Relations for Free or Natural Convection
b¼
699
1 1 ¼ 3:003 103 ð1=KÞ: ¼ Tm ð60 þ 273Þ
From the given correlation, " 0:25 # 0:25 2 k Pr q bgDTx3 hx ¼ 0:52 Pr : x 0:95 þ Pr l2 Substitution of values of various terms gives 0:25 0:02875 0:7 hx ¼ 0:52 x 0:95 þ 0:7 1:0592 3:003 103 9:81 60 x3
ð2 105 Þ2
!0:25 0:7
¼ 2:928x0:25 : The mean heat transfer coefﬁcient 1 h¼ L
Z
L 0
0:2 1 x0:75 2:928 hx dx ¼ ¼ 5:84 W/(m2 K): 0:2 0:75 0
Heat transfer rate q ¼ hADT ¼ 5:84 0:2 0:6 60 ¼ 42:05 W: Example 9.6 A vertical rectangular plate 1 m 0.6 m has one of its surface insulted and the other face is maintained at a uniform temperature of 50°C. It is exposed to quiescent air at 20°C. Calculate the heat transfer coefﬁcient if (i) 1 m side of the plate is in vertical direction (ii) 0.6 m side of the plate is in vertical direction. Solution For the range 104 < Ra < 109, Nu ¼ 0:59Ra0:25 and for the range 109 Ra < 1013, Nu ¼ 0:1 Ra1=3 The thermophysical properties of the air at the mean ﬁlm temperature of 35°C from Table A5 are:
700
9 Empirical Relations for Natural or Free Convection
q = 1.15 kg/m3, cp = 1006 J/(kg K), k = 2.68 10−2 W/(m K), µ = 1.883 10−5 kg/(m s), Pr = 0.7066 and the coefﬁcient of cubical expansion, b = 1/T = 1/(35 + 273) = 1/308 K−1. Hence, the Rayleigh number is q2 bgDTL3 q2 bgðtw t1 ÞL3 Pr ¼ Pr l2 l2 1:152 1=308 9:81 ð50 20Þ L3 0:7066 ¼ ð2:52 109 ÞL3 : ¼ ð1:883 105 Þ2
Ra ¼ Gr Pr ¼
(i) 1 m side is vertical Characteristic dimensions, L = 1 m. Rayleigh number, Ra ¼ 2:52 109 L3 ¼ 2:52 109 1:03 ¼ 2:52 109 [ 109 Nusselt number, Nu ¼ 0:1Ra1=3 ¼ 0:1 ð2:52 109 Þ1=3 ¼ 136: Heat transfer coefﬁcient, Nuk 136 0:0268 ¼ ¼ 3:645 W/(m2 K): L 1:0
h¼
(ii) 0.6 m side is vertical Characteristic dimensions, L = 0.6 m Rayleigh number, Ra ¼ 2:52 109 L3 ¼ 2:52 109 0:63 ¼ 5:4 108 \109 : Nusselt number, Nu ¼ 0:59Ra0:25 ¼ 0:59 ð5:4 108 Þ0:25 ¼ 89:94 Heat transfer coefﬁcient, h¼
Nuk 89:94 0:0268 ¼ ¼ 4:02 W/(m2 K): L 0:6
9.5 Empirical Relations for Free or Natural Convection
701
Example 9.7 A horizontal plate (1.0 m 1.0 m) is maintained at 140°C by electric heating. Estimate the convective heat transfer rate to the surrounding still air at 20°C. The lower surface of the plate is insulated. Solution At mean ﬁlm temperature of (140 + 20)/2 = 80°C, the air properties are v ¼ l=q ¼ 20:87 106 m2=s ; k ¼ 0:0303W=ðmKÞ; Pr ¼ 0:696; and b ¼ 1=ð50 þ 273ÞK1 : For a horizontal plate, L¼
A 1:0 1:0 ¼ ¼ 0:25 m: P 4
The Rayleigh number, RaL ¼ GrL Pr ¼ ¼
gbðtw t1 ÞL3 : Pr m2
9:81 ð1=353Þ ð140 20Þ 0:253 ð20:87 106 Þ2
0:696 ¼ 8:33 107 :
Equation (9.17) applies, which gives NuL ¼ 0:15ðRaL Þ1=3 ¼ 0:15 ð8:33 107 Þ1=3 ¼ 65:5: Heat transfer coefﬁcient, h¼
Nu k 65:5 0:0303 ¼ ¼ 7:94 W/(m2 K): L 0:25
Heat transfer rate, q ¼ hADT ¼ 7:94 ð1:0 1:0Þ ð140 20Þ ¼ 952:8 W: Example 9.8 If the heated surface of the plate of Example 9.7 is facing downwards, estimate the convective heat transfer rate. Solution For a horizontal plate with heated surface facing downwards, Eq. (9.19) applies, which gives NuL ¼ 0:27ðRaL Þ1=4 ¼ 0:27 ð8:33 107 Þ1=4 ¼ 25:8
702
9 Empirical Relations for Natural or Free Convection
Heat transfer coefﬁcient, h¼
Nu k 25:8 0:0303 ¼ ¼ 3:13 W/(m2 K): L 0:25
Heat transfer rate, q ¼ hADT ¼ 3:13 ð1:0 1:0Þ ð140 20Þ ¼ 375:6 W:
Comments: The heat rejection rate is signiﬁcantly reduced, which is expected because of the difference in the flow structure of the two cases as presented in Fig. 9.1. Example 9.9 A 400 mm square plate is inclined from vertical at an angle of 30°. The surface temperature of the plate is 330 K. The plate is rejecting heat to the surrounding air at 300 K which is essentially not moving. Determine the convective heat transfer rate from the plate. Solution At mean ﬁlm temperature of (330 + 300)/2 = 315 K = 42°C, the air properties are q ¼ 1:124 kg=m3 ; l ¼ 1:91 105 kg=ðm sÞ; k ¼ 0:0274 W=ðm KÞ; Pr ¼ 0:705 and b ¼ 1=315 K1 For an inclined plate, ðg cos hÞbðTw T1 ÞL3 : Pr m2 ð9:81 cos 30Þ ð1=315Þ ð330 300Þ 0:43
RaL ¼ GrL Pr ¼ ¼
ð1:91 105 =1:124Þ2
0:705
¼ 1:26 108 \109 : The flow is laminar. Equation (9.6) applies: NuL ¼ 0:59 ðRaL Þ1=4 ¼ 0:59 ð1:26 108 Þ1=4 ¼ 62:5: Heat transfer coefﬁcient, h¼
Nuk 62:5 0:0274 ¼ ¼ 4:28 W/(m2 K): L 0:4
Heat transfer rate, q ¼ hADT ¼ 4:28 ð0:4 0:4Þ ð330 300Þ ¼ 20:54 W:
9.5 Empirical Relations for Free or Natural Convection
703
Example 9.10 Determine the hourly loss of heat from a bare horizontal steam pipe whose diameter is 100 mm and the length is 5 m. The pipe wall temperature is 450 K and the temperature of the surrounding air is 300 K. Assume that the radiation heat loss is negligible. Solution The mean ﬁlm temperature, Tm ¼
Tw þ T1 450 þ 300 ¼ 375 K: ¼ 2 2
From Table A5 of “Thermophysical Properties of Air”, we have q ¼ 0:941 kg=m3 ; Pr ¼ 0:693; k ¼ 0:032 W=ðm KÞ; l ¼ 2:18 105 kg=ðm sÞ: Grashof number, Gr ¼
bgDTd3 ½1=ð375Þ 9:81 ð450 300Þ 0:13 ¼ ¼ 7:31 106 : m2 ð2:18 105 =0:941Þ2
From Eq. (9.23), Nusselt number Nu ¼ 0:53ðGrPrÞ0:25 ¼ 0:53 ð7:31 106 0:693Þ0:25 ¼ 25:1:
The heat transfer coefﬁcient at the outer surface, h¼
Nuk 25:1 0:032 ¼ ¼ 8:0 W/(m2 K): d 0:1
The hourly heat loss, q ¼ hADT ¼ hðpdLÞðTw T1 Þ ¼ 8:0 ðp 0:1 5Þ ð450 300Þ 3600 ¼ 6785:8 kJ: Example 9.11 If the pipe of Example 9.10 is vertical, determine the hourly loss. Solution Grashof number, Gr ¼
bgDTL3 ½1=ð375Þ 9:81 ð450 300Þ 53 ¼ ¼ 9:14 1011 : m2 ð2:18 105 =0:941Þ2
Since Gr > 4 109, the flow is turbulent and L/D = 50, Eq. (9.9b) may be used, which gives
704
9 Empirical Relations for Natural or Free Convection
Num ¼ 0:13ðRaÞ1=3 ¼ 0:13 ð9:14 1011 0:693Þ1=3 ¼ 1116:
The heat transfer coefﬁcient at the outer surface of the cylinder is hm ¼
Num k 1116 0:032 ¼ ¼ 7:14 W/(m2 K): L 5
The hourly heat loss, q ¼ hm AðDTÞ ¼ hm ðpdLÞðTw T1 Þ ¼ 7:14 ðp 0:1 5Þ ð450 300Þ 3600 ¼ 6056 kJ: Example 9.12 A thin walled vertical cylindrical tank, 0.3 m in diameter and 0.8 m in height, contains water at 70°C. The tank is placed in a larger tank with water at 10°C. Determine heat transfer rate across the tank wall. Solution Assuming wall temperature as mean of the water temperatures, i.e. 40°C, the temperature of the ﬁlm on the inner side is (70 + 40)/2 = 55°C. Thermophysical properties of water at the ﬁlm temperature of 55°C from Table A4 are q ¼ 985:2 kg=m3 ; l ¼ 501 106 N s=m2 ; k ¼ 0:648 W=ðm KÞ; Pr ¼ 3:24 and b ¼ 0:484 103 1=K:
The Grashof number based on cylinder height, Gr ¼ ¼
q2 gbDTL3 l2 985:22 9:81 0:484 103 ð70 40Þ 0:83 ð501 106 Þ2
D/L = 0.375 and 35/Gr0.25 = 0.048. Hence, condition
D L
¼ 2:82 1011 : 35
Gr1=4 L
cylinder can be treated as a vertical plate. The Rayleigh number is Ra ¼ GrPr ¼ 2:82 1011 3:24 ¼ 9:14 1011 : From Eq. (9.10),
is satisﬁed and the
9.5 Empirical Relations for Free or Natural Convection
Num ¼
¼
8 > < > : 8 > < > :
0:825 þ h
0:387Ra
1=6
1 þ ð0:492= PrÞ
9=16
705
92 > =
1 12 i8=27 > for 10 Ra 10 ; 11 1=6
92 > =
0:387 ð9:14 10 Þ 0:825 þ h i8=27 > ¼ 1276:9: ; 1 þ ð0:492=3:24Þ9=16
Heat transfer coefﬁcient (inner side), k 0:648 1276:9 ¼ 1034:3 W=ðm2 KÞ: hi ¼ Num ¼ L 0:8 The ﬁlm temperature on the outer side is (40 + 10)/2 = 25°C. Properties of water at 25°C from Table A4 are: q ¼ 997 kg=m3 ; l ¼ 890 106 N s=m2 ; k ¼ 0:609 W=ðm KÞ; Pr ¼ 6:13 and b ¼ 0:253 103 1=K: The Grashof number based on cylinder height, Gr ¼ ¼
q2 gbDTL3 l2 2 997 9:81 0:253 103 ð40 10Þ 0:83 ð890
106 Þ2
D/L = 0.375 and 35/Gr0.25 = 0.0748. Hence, condition
D L
¼ 4:78 1010 : 35
Gr1=4 L
is satisﬁed and the
cylinder can be treated as a vertical plate. The Rayleigh number is Ra ¼ GrPr ¼ 4:78 1010 6:13 ¼ 2:93 1011 : From Eq. (9.10), 8 >
=
0:387Ra 1 12 0:825 þ h i8=27 > for 10 Ra 10 9=16 ; 1 þ ð0:492= PrÞ 92 8 > > < 11 1=6 = 0:387 ð2:93 10 Þ ¼ 0:825 þ h i8=27 > ¼ 924:3: > 9=16 ; : 1 þ ð0:492=6:13Þ
Num ¼
> :
Heat transfer coefﬁcient,
706
9 Empirical Relations for Natural or Free Convection
k 0:609 924:3 ¼ 703:6 W=ðm2 KÞ: ho ¼ Num ¼ L 0:8 Overall heat transfer coefﬁcient, U¼
1 1 þ hi ho
1
¼
1 1 þ 1034:3 703:6
1
¼ 418:7 W=ðm2 KÞ:
Heat transfer rate, q ¼ UADT ¼ UpDLðtw ta Þ ¼ 418:7 p 0:3 0:8 ð70 10Þ ¼ 18941:5 W: Wall temperature: q ¼ ho pDLðts to Þ or ts ¼
q 18941:5 þ 10 ¼ 45:7o C, þ to ¼ h0 pDL 703:6 p 0:3 0:8
as against the assumed value of 40°C. Retrial with ts = 45°C may be carried out. Example 9.13 A horizontal steel pipeline, having inside diameter of 50 mm and outside diameter of 60 mm, carries water flowing with a velocity of 0.2 m/s. The mean water temperature is 80°C. The pipeline is lagged with asbestos. The outside diameter of the lagging is 90 mm. Determine the loss of heat from 1 m length of the pipeline if the temperature of the still air surrounding the pipeline is 20°C. Neglect the radiation heat loss from the outer surface of the lagging. Also determine the surface temperature of the pipeline and the lagging. For steel, ks = 46 W/(m K) and for asbestos ka = 0.11 W/(m K). Solution Thermophysical properties of water at 80°C (the mean temperature of water) are q ¼ 971:8 kg=m3 ; Pr ¼ 2:21; k ¼ 0:67 W=ðm KÞ and l ¼ 351 106 kg=ðm sÞ: (i) The coefﬁcient of heat transfer from water to the inner surface of the pipe: Re ¼
qUd 971:8 0:2 0:05 ¼ ¼ 2:77 104 : l 351 106
The flow is in the turbulent regime and the heat transfer coefﬁcient can be determined from the following correlation applicable to Re > 10000.
9.5 Empirical Relations for Free or Natural Convection
707
tw2 °
ti
50
°
60 90
tw1 • tw3
• to
Fig. 9.4 Example 9.13
Nu ¼ 0:024Re0:8 Pr0:4 ¼ 0:024 ð2:77 104 Þ0:8 ð2:21Þ0:4 ¼ 118:0: This gives the heat transfer coefﬁcient: hi ¼
Nuk 118:0 0:67 ¼ ¼ 1581 W/(m2 K): d 0:05
(ii) The coefﬁcient of heat transfer from water to the inside surface of the tube is very high. Further the pipe is lagged and the heat transfer coefﬁcient from the outer surface of the lagging is likely to be low because of natural convection condition at the surface of the lagging. Hence, the pipe inside surface temperature will be nearly equal to the temperature of the water, i.e. tw1 ti. (iii) Since the free convection heat transfer rate is a function of temperature difference DT, we assume the temperature of the outer surface of the lagging tw3 as 50°C for ﬁrst trial. Refer Fig. 9.4. With this assumption, the Grashof number is Gr ¼
bgDTL3 1=308 9:81 ð50 20Þ 0:093 ¼ ¼ 2:6 106 m2 ð1:88 105 =1:15Þ2
where tm b d DT l q
(to + tw3)/2 = (20 + 50)/2 = 35°C 1/Tm = 1/(35 + 273) = 1/308 K−1 outer diameter of the lagging = 0.09 m 50–20 = 30°C 1.88 10−5 kg/(m s) for the air at 35°C 1.15 kg/m3 for air at 35°C
This gives Ra ¼ GrPr ¼ 2:6 106 0:707 ¼ 1:84 106
708
9 Empirical Relations for Natural or Free Convection
where Pr= 0.707 for the air. For 104 Ra 109, Eq. (9.23) applies, which gives Nu ¼ 0:53 ðRaÞ0:25 ¼ 0:53 ð1:84 106 Þ0:25 ¼ 19:5: The heat transfer coefﬁcient at the outer surface, ho ¼
Nuk 19:5 0:02684 ¼ ¼ 5:82 W/(m2 K): d 0:09
The heat transfer rate for the unit length of the pipeline is q¼
1 2pri hi
þ
1 2pks
t to i r3 r2 1 ln r1 þ 2pk ln r2 þ a
1 2pro ho
or q¼
1 2p0:0251581
þ
80 20 1 ln 0:045 þ 2p0:11 0:03 þ
1 0:03 2p46 ln 0:025
1 2p0:0455:82
¼ 50:04 W: This gives tw3 ¼ to þ
q 50:04 ¼ 50:4 C: ¼ 20 þ 2pro ho 2p 0:045 5:82
Since the calculated temperature is nearly the same as the assumed one, further trial is not needed. The inside pipesurface temperature is tw1 ¼ ti
q 50:04 ¼ 79:8 C; ¼ 80 2pri hi 2p 0:025 1588
which is nearly equal to the water temperature assumed earlier. The temperature of the outside surface of the pipe, tw2 ¼ tw1
q r2 50:04 0:03 ln ln ¼ 79:8 ¼ 79:77o C, 2pks 2p 46 0:025 r1
which indicates that the temperature drop across the wall of the pipe is negligible. This is because of the fact that the resistance offered by the metal wall to the heat flow is very small. The main resistances here are that of the insulation and outer ﬁlm.
9.5 Empirical Relations for Free or Natural Convection
709
Example 9.14 If wind blows at a speed of 2 m/s across the pipe of Example 9.13, determine the increase in the heat loss. Solution The Reynolds number, Red ¼
qU1 d 1:15 2 0:09 ¼ ¼ 11010: lf 1:88 105
From Table 8.11, for Re = 4000–40000, C = 0.193 and n = 0.618. The average heat transfer coefﬁcient, h ¼ k C ðRed Þn Pr1=3 d 0:02684 0:193 ð11010Þ0:618 ð0:707Þ1=3 ¼ 16:1 W/(m2 K): ¼ 0:09 Heat flow rate q¼
1 2pri hi
þ
1 2pks
t to i r3 r2 1 ln r1 þ 2pk ln r2 þ a
1 2pro ho
or q¼
1 2p0:0251581
þ
80 20 1 ln 0:045 þ 2p0:11 0:03 þ
1 0:03 2p46 ln 0:025
1 2p0:04516:1
¼ 74:0 W:
Alternatively using Eq. (8.52), we have " 5=8 #4=5 0:62Re1=2 Pr1=3 Re Nu ¼ 0:3 þ h 2=3 i1=4 1 þ 282000 1 þ 0:4 Pr " #4=5 0:62 ð11010Þ1=2 ð0:707Þ1=3 11010 5=8 ¼ 0:3 þ 1þ ¼ 56:46: h 0:4 2=3 i1=4 282000 1 þ 0:707 Hence, h ¼ k Nu ¼ 0:02684 56:46 ¼ 16:84 W/(m2 K) d 0:09
710
9 Empirical Relations for Natural or Free Convection
This gives q¼
1 2pri hi
t to i r3 r 1 1 2 þ 2pk ln ln þ 2pk r r2 þ s 1 a
1 2pro ho
Substitution gives q¼
1 2p0:0251581
þ
80 20 1 þ 2p0:11 ln 0:045 0:03 þ
1 0:03 2p46 ln 0:025
1 2p0:04516:74
¼ 74:8 W: The heat loss due to the wind increases by about 50%.
9.6
Free Convection in Parallel Plate Channels
Array of circuit boards or ﬁn arrays employed to enhance free convection heat transfer may be regarded as parallel vertical plates forming a channel that is open to the ambient at its ends as shown in Fig. 9.5. The heated fluid rises and exits from the top of the channel. Boundary layers develop beginning at x = 0 on each wall (if both the walls of the channel are heated) and increase in thickness with height. If the ratio of the spacing S and height L is small, the boundary layers eventually merge and the flow, thereafter, is fully developed. When the ratio S/L is large (i.e. the plates are far apart), they behave like isolated vertical plates in inﬁnite, quiescent fluid. Because of the effect of the ratio of spacing to height S/L on the flow structure, this ratio appears in the heat transfer correlations presented by researchers. Surface thermal conditions have been classiﬁed as isothermal or isoflux and symmetrical (Tw1 = Tw2; q00w1 = q00w2 ) or asymmetrical (Tw1 6¼ Tw2; q00w1 6¼ q00w2 ), where q00w is heat flux.
x=L Tw1 or
S
q w" 1
Tw2 or
q w" 2
x
●T∞ Fig. 9.5 Free convection flow between heated parallel plates
9.6 Free Convection in Parallel Plate Channels
711
9.6.1 Vertical Channels BarCohen and Rohsenow (1984) presented correlations for symmetric and asymmetric isothermal and isoflux (uniform heat flux) conditions of vertical parallel plate channels. They have also presented relations for optimum spacing, which is the spacing to provide maximum heat transfer from an array in a given volume, i.e. spacing for the maximum volumetric heat dissipation. It is to note that as the number of plates in a given space is increased, the heat transfer coefﬁcient h decreases with decreasing spacing and the reduction in the fluid flow rate because of the increase in the viscous force. But the increased total area A of the larger number of plates yields a maximum of the product hA. Spacing needed to maximize the heat transfer from each plate in the array has also been deﬁned and relations with respect to the spacing for the maximum volumetric heat dissipation are given. This spacing for maximum heat transfer from each plate in the array precludes the overlap of the adjoining boundary layers and is found to be larger than the spacing for the maximum volumetric heat dissipation.
9.6.2 Inclined Channels Experimental data for inclined channels formed by symmetric isothermal plates and isothermalinsulated plates in water (inclination 0 to 45o with vertical) for isolated plate limit of [RaS(S/L) > 200] have been correlated within ± 10% by Azevedo and Sparrow (1985) as ðNuS Þm ¼ 0:645½RaS ðS=LÞ 1=4
ð9:28Þ
where the Rayleigh number is based on the spacing S and the fluid properties are evaluated at average temperature Tm = (Tw + T∞)/2.
9.7
Empirical Correlations for Enclosed Spaces
In some engineering applications, heat transfer takes place between two surfaces that are at different temperatures and a fluid is present in the space enclosed by these surfaces. Typical examples of such enclosures are double pane windows, cavities in the walls of the buildings, solar collectors, etc.
S T1
Boundary layer T2
L
a Fig. 9.6 Cellular flow in a vertical enclosure
b
712
9 Empirical Relations for Natural or Free Convection
Here the correlations are being presented only for the rectangular enclosures. (i) Vertical Enclosure Consider the rectangular enclosure shown in Fig. 9.6. The left wall is at a higher temperature than the right wall. Top and bottom surfaces of the enclosure are insulated. The heat transfers by the free convection currents in the enclosed fluid. The Rayleigh number for this case is deﬁned as RaS ¼
gbðT1 T2 ÞS3 Pr m2
ð9:29Þ
where S is the spacing between the heat transferring surfaces. If the Rayleigh number is less than about 103, viscous forces dominate over the buoyancy forces and the fluid motion does not occur. This can happen when the spacing S is small and the viscosity is very large. In such a situation, the heat transfer takes place by conduction only through the stagnant fluid and Nu = 1. At Ra 104, the flow changes to a boundary layer type (Mills 1995). The fluid at the hot surface moves upwards along the wall and after giving heat to the cold face of the enclosure moves downwards along the wall if the spacing S is sufﬁciently large. A cellular flow establishes, Fig. 9.6(a), which is encountered in the thin boundary layers adjoining the vertical surfaces. The core is nearly stagnant. With small spacing, the flows affect each other which may result in internal circulation as if partitions are present, Fig. 9.6(b). At Ra 106, the flow in the core becomes turbulent (Mills 1995). Proposed empirical correlations for average Nusselt number are (MacGregor and Emery 1969) NuS ¼ 0:42ðRaS Þ
1=4
Pr
0:012
0:3 L S
10 L=S 40 4
10 RaS 10
ð9:30Þ 7
1 Pr 2 104 NuS ¼ 0:046ðRaS Þ1=3 1 L=S 40 106 RaS 109 1 Pr 20 where the Nusselt number is deﬁned as
T2
L
S T1
Fig. 9.7 Bernard cells
ð9:31Þ
9.7 Empirical Correlations for Enclosed Spaces
713
NuS ¼
hS k
ð9:32Þ
and the heat transfer rate is given by q ¼ hAðT1 T2 Þ
ð9:33Þ
(ii) Horizontal Enclosure Consider the horizontal enclosure with the lower wall heated and upper wall cold. Side h i T2 ÞS3 walls are insulated. For Rayleigh number RaS ¼ gbðT1am less than a critical value of 1708, the viscous forces are very strong than the buoyancy forces and there is no movement of the fluid. The heat transfer takes place by conduction only and Nu = 1. For 1708 < RaS 5 104, the fluid motion consists of regularly spaced rotating cells (called as Bernard cells) as shown in Fig. 9.7. At higher Rayleigh numbers, the cells break down and turbulent motion occurs. The correlation proposed by Globe and Dropkin (1959) for average Nusselt number is hm S 1=3 ¼ 0:069RaS Pr0:074 k for 3 105 RaS 7 109
ðNuS Þm ¼
ð9:34Þ
where all fluid properties are to be evaluated at the mean temperature Tm= (T1 + T2)/2. If the upper wall is heated, then the heat transfer takes place by conduction only (Nu = 1) irrespective of the value of the Rayleigh number. It is to note that radiation heat exchange between the hot and cold surfaces is always present when a gas is ﬁlled in the cavity. (iii) Tilted Enclosure (Fig. 9.8)
S
T2 L T1 θ Fig. 9.8 Tilted enclosure Table 9.2 Critical tilt angle* L/S h
*
1
3 o
25
6 o
53
60
12 o
67
>12 o
70o
*Incropera Frank P, DeWitt David P, Bergman Theodore L, Lavine Adrienne S, Fundamentals of heat and mass transfer, 6th edn, John Wiley & Sons, New York, Copyright 2006. Reproduced with the permission of John Wiley & Sons Ltd
714
9 Empirical Relations for Natural or Free Convection
Some applications involve tilted enclosures. Typical example is that of a flat plate solar collector where the heat flows from the heated absorber plate to the glass cover placed at a distance. The tilt angle may vary from 0o (horizontal) to about 70o at high latitude applications. In inclined enclosures, the fluid motion consists of a combination of the rolling structure of horizontal cavities and cellular structure of vertical cavities. Typically transition between the two types of motions occurs at a critical tilt angle h* given in Table 9.2. For large aspect ratio duct of solar collectors (L/S > 12) and the tilt angle less than the critical value h°, the following correlation of the average Nusselt number has been proposed by Hollands et al. (1976). # " #þ þ " 1708 1708ðsin1.8hÞ1:6 RaS cosh 1=3 þ ðNum ÞS ¼ 1 þ 1:44 1 1 1 RaS cosh 5830 RaS cosh for 0 h 60 and 0\Ra 105 : ð9:35Þ Either of the terms in + bracket goes to zero when negative. This implies that if the Rayleigh number is less than a critical value of 1708/cos h, there is no flow within the cavity (Incropera et al. 2012). This equation can be expected to give values of Nu with a maximum error of 5%. Equation (9.35) may also be used for h up to 75° but error of up to 10 percent may be expected (Hollands et al. 1976). Buchberg et al. (1976) presented the following threeregion correlation for the estimate of the convective heat transfer coefﬁcient between the absorber plate and glass cover of flat plate solar air heaters. 1708 þ Nu ¼ 1 þ 1:446 1 Ra0
ð9:36aÞ
0
for 1708 Ra 5900 (the + bracket goes to zero when negative) Nu ¼ 0:229ðRa0 Þ
0:252
for 5900\Ra0 9:23 104 Nu ¼ 0:157ðRa0 Þ
ð9:36bÞ
0:285
for 9:23 104 \Ra0 106
ð9:36cÞ
where Ra′ (= Ra cosh) is Rayleigh number for the inclined air layers. Example 9.15 A horizontal enclosure consists of two square plates (0.5 m 0.5 m) separated by a distance of 10 mm. The lower plate is maintained at a uniform temperature of 40° C and the upper plate is at a uniform temperature of 10°C. Water at atmospheric pressure is ﬁlled in the enclosure. Calculate the heat loss rate from the lower plate if the side walls are insulated. Given gbq2c/(lk) = 1.85 1010 m−3K−1 at the mean temperature.
9.7 Empirical Correlations for Enclosed Spaces
715
Solution At the mean temperature of (40 + 10)/2 = 25°C, k = 0.61 W/(m K) for water. The Rayleigh number = gbq2c/(lk) DT S3 = 1.85 1010 (40 3 5 10) 0.01 = 5.55 10 . For 3 105 RaS 7 109, Eq. (9.34) applies:
–
1=3 Nu ¼ 0:069ðRas Þ1=3 ðPrÞ0:074 ¼ 0:069 5:55 105 ð6:13Þ0:074 ¼ 6:48: Mean heat transfer coefﬁcient, h ¼ Nu
k 0:61 ¼ 6:48 ¼ 395:3 W/(m2 K): S 0:01
Heat loss rate, q ¼ hAðT1 T2 Þ ¼ 395:3 ð0:5 0:5Þ ð40 10Þ ¼ 2964:7 W:
Example 9.16 Two horizontal surfaces of an enclosure with air between them are separated by 10 mm. Calculate the heat flow rate per m2 of the plate surface if the upper plate is at 50° C and lower is at 20°C. The vertical sides are insulated. Solution In the present case, the convection currents are suppressed and the heat transfer is by conduction only. Hence, q ¼ kA
T1 T2 50 20 ¼ 81 W/m2 ; ¼ 0:027 1 0:01 S
where k = 0.027 W/(m K) at the mean temperature (50 + 20)/2 = 35°C. Example 9.17 Absorber plate and cover plate combination of a flat plate solar collector can be treated as a tilted enclosure (Fig. 9.8). If T1 = 70°C, T2 = 30°C, L = 1.5 m, S = 50 mm and tilt h = 45o determine heat flux due to free convection. Solution For large aspect ratio duct of solar collectors (L/S > 12) and the tilt angle less than the critical value h* given in Table 9.2, Eq. (9.35) of the average Nusselt number can be used:
1708 ðNum ÞS ¼ 1 þ 1:44 1 RaS cosh
þ "
#þ # " 1708ðsin1.8hÞ1:6 RaS cosh 1=3 1 1 þ 5830 RaS cosh
for 0 h 60 and 0\Ra 105 :
Either of the terms in + bracket goes to zero when negative.
716
9 Empirical Relations for Natural or Free Convection
Thermophysical properties of air at mean temperature = (70 + 30)/2 = 50°C from Table A5 are: q ¼ 1:0949 kg=m3 ; l ¼ 1:9512 105 kg=ðm sÞ; k ¼ 0:02799 W=ðm KÞ; a ¼ 0:257 104 m2 /s and Pr ¼ 0:703: Rayleigh number, gbðT1 T2 ÞS3 am 9:81 1=ð50 þ 273Þ ð75 25Þ 0:053 ¼ ¼ 4:14 105 : 0:257 104 ð1:9512 105 =1:0949Þ
RaS ¼
Hence, # þ " 1708 1708 ðsin1.8 45Þ1:6 ðNum ÞS ¼ 1 þ 1:44 1 1 4:14 105 cos45 4:14 105 cos45 " 1=3 # þ 4:14 105 cos45 þ 1 5830
¼ 1 þ 1:44 0:994 0:994 þ 2:689 ¼ 4:79: Heat transfer coefﬁcient, k 0:02799 4:79 ¼ 2:68 W=ðm2 KÞ: h ¼ ðNum Þs ¼ L 0:05 Hence, heat flux is q ¼ hðT1 T2 Þ ¼ 2:68 50 ¼ 134 W=m2 : Here the radiation heat transfer has not been considered. Example 9.18 A cover plate with 50 mm air gap is installed over a horizontal surface (0.4 m 0.4 m in area) maintained at T1= 75°C to reduce heat transfer from the surface as shown in Fig. 9.9. Determine the heat transfer rate to the still air at 25°C above the cover plate. Neglect radiation heat transfer and contact resistance between the surface and cover. Solution In equilibrium, the heat transfer across the air gap of the cover equals the heat rejection from the top surface of the cover, i.e. h1 ðT1 T2 Þ ¼ h2 ðT2 T1 Þ
ðiÞ
where T2 is not known which is required to estimate h1 and h2. T2 Cover
S T1
Fig. 9.9 Example 9.18
9.7 Empirical Correlations for Enclosed Spaces
717
For trial, we assume T2 = 50°C. For heat transfer through air gap of the cover, mean air temperature is Tm1 = (T1 + T2)/2 = (75 + 50)/2 = 6