BooksAudiobooksComicsSheet Music### Categories

### Categories

### Categories

Editors' Picks Books

Hand-picked favorites from

our editors

our editors

Editors' Picks Audiobooks

Hand-picked favorites from

our editors

our editors

Editors' Picks Comics

Hand-picked favorites from

our editors

our editors

Editors' Picks Sheet Music

Hand-picked favorites from

our editors

our editors

Top Books

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Audiobooks

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Comics

What's trending, bestsellers,

award-winners & more

award-winners & more

Top Sheet Music

What's trending, bestsellers,

award-winners & more

award-winners & more

Welcome to Scribd! Start your free trial and access books, documents and more.Find out more

Fundamentals of Momentum,

Heat, and Mass Transfer

5th Edition

Fundamentals of Momentum,

Heat, and Mass Transfer

5th Edition

James R. Welty

Department of Mechanical Engineering

Charles E. Wicks

Department of Chemical Engineering

Robert E. Wilson

Department of Mechanical Engineering

Gregory L. Rorrer

Department of Chemical Engineering

Oregon State University

John Wiley & Sons, Inc.

ASSOCIATE PUBLISHER

ACQUISITIONS EDITOR

MARKETING MANAGER

CREATIVE DIRECTOR

DESIGNER

SENIOR MEDIA EDITOR

SENIOR PRODUCTION EDITOR

PRODUCTION MANAGEMENT SERVICES

Daniel Sayre

Jennifer Welter

Christopher Ruel

Harry Nolan

Michael St. Martine

Lauren Sapira

Patricia McFadden

Thomson Digital

This book was set in by Thomson Digital and printed and bound by Hamilton Printing. The cover was

printed by Lehigh Press, Inc.

**This book is printed on acid free paper.
**

1

Copyright # 2008 John Wiley & Sons, Inc. All rights reserved. No part of this publication may be reproduced,

stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying,

recording, scanning or otherwise, except as permitted under Sections 107 or 108 of the 1976 United States

Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the

appropriate per-copy fee to the Copyright Clearance Center, Inc. 222 Rosewood Drive, Danvers, MA 01923,

website www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions

Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030-5774, (201)748-6011, fax (201)7486008, website http://www.wiley.com/go/permissions.

To order books or for customer service please, call 1-800-CALL WILEY (225-5945).

ISBN-13 978-0470128688

Printed in the United States of America

10 9 8 7 6 5 4 3 2 1

Preface to the 5th Edition

**The ﬁrst edition of Fundamentals of Momentum, Heat, and Mass Transfer, published in
**

1969, was written to become a part of what was then known as the ‘‘engineering science

core’’ of most engineering curricula. Indeed, requirements for ABET accreditation have

stipulated that a signiﬁcant part of all curricula must be devoted to fundamental subjects.

The emphasis on engineering science has continued over the intervening years, but the

degree of emphasis has diminished as new subjects and technologies have entered the

world of engineering education. Nonetheless, the subjects of momentum transfer (ﬂuid

mechanics), heat transfer, and mass transfer remain, at least in part, important components

of all engineering curricula. It is in this context that we now present the ﬁfth edition.

Advances in computing capability have been astonishing since 1969. At that time, the

pocket calculator was quite new and not generally in the hands of engineering students.

Subsequent editions of this book included increasingly sophisticated solution techniques as

technology advanced. Now, more than 30 years since the ﬁrst edition, computer competency

among students is a fait accompli and many homework assignments are completed using

computer software that takes care of most mathematical complexity, and a good deal of

physical insight. We do not judge the appropriateness of such approaches, but they surely

occur and will do so more frequently as software becomes more readily available, more

sophisticated, and easier to use.

In this edition, we still include some examples and problems that are posed in English

units, but a large portion of the quantitative work presented is now in SI units. This is

consistent with most of the current generation of engineering textbooks. There are still some

subdisciplines in the thermal/ﬂuid sciences that use English units conventionally, so it

remains necessary for students to have some familiarity with pounds, mass, slugs, feet, psi,

and so forth. Perhaps a ﬁfth edition, if it materializes, will ﬁnally be entirely SI.

We, the original three authors (W3), welcome Dr. Greg Rorrer to our team. Greg is a

member of the faculty of the Chemical Engineering Department at Oregon State University

with expertise in biochemical engineering. He has had a signiﬁcant inﬂuence on this

edition’s sections on mass transfer, both in the text and in the problem sets at the end of

Chapters 24 through 31. This edition is unquestionably strengthened by his contributions,

and we anticipate his continued presence on our writing team.

We are gratiﬁed that the use of this book has continued at a signiﬁcant level since the

ﬁrst edition appeared some 30 years ago. It is our continuing belief that the transport

phenomena remain essential parts of the foundation of engineering education and practice.

With the modiﬁcations and modernization of this fourth edition, it is our hope that

Fundamentals of Momentum, Heat, and Mass Transfer will continue to be an essential

part of students’ educational experiences.

Corvallis, Oregon

March 2000

J.R. Welty

C.E. Wicks

R.E. Wilson

G.L. Rorrer

v

This page intentionally left blank

Contents

1.

**Introduction to Momentum Transfer
**

1.1

1.2

1.3

1.4

1.5

1.6

2.

3.

5.

29

**Fundamental Physical Laws
**

29

Fluid-Flow Fields: Lagrangian and Eulerian Representations

Steady and Unsteady Flows

30

Streamlines

31

Systems and Control Volumes

32

Integral Relation

34

Speciﬁc Forms of the Integral Expression

Closure

39

35

43

**Integral Relation for Linear Momentum
**

43

Applications of the Integral Expression for Linear Momentum

Integral Relation for Moment of Momentum

52

Applications to Pumps and Turbines

53

Closure

57

**Conservation of Energy: Control-Volume Approach
**

6.1

6.2

29

34

**Newton’s Second Law of Motion: Control-Volume Approach
**

5.1

5.2

5.3

5.4

5.5

6.

**Pressure Variation in a Static Fluid
**

16

Uniform Rectilinear Acceleration

19

Forces on Submerged Surfaces

20

Buoyancy

23

Closure

25

**Conservation of Mass: Control-Volume Approach
**

4.1

4.2

4.3

5

16

**Description of a Fluid in Motion
**

3.1

3.2

3.3

3.4

3.5

4.

**Fluids and the Continuum
**

1

Properties at a Point

2

Point-to-Point Variation of Properties in a Fluid

Units

8

Compressibility

9

Surface Tension

11

Fluid Statics

2.1

2.2

2.3

2.4

2.5

1

**Integral Relation for the Conservation of Energy
**

Applications of the Integral Expression

69

46

63

63

vii

viii

Contents

6.3

6.4

7.

**Shear Stress in Laminar Flow
**

7.1

7.2

7.3

7.4

7.5

8.

8.2

8.3

99

113

**Fluid Rotation at a Point
**

113

The Stream Function

114

Inviscid, Irrotational Flow about an Inﬁnite Cylinder

Irrotational Flow, the Velocity Potential

117

Total Head in Irrotational Flow

119

Utilization of Potential Flow

119

Potential Flow Analysis—Simple Plane Flow Cases

Potential Flow Analysis—Superposition

121

Closure

123

116

120

125

Dimensions

125

Dimensional Analysis of Governing Differential Equations

The Buckingham Method

128

Geometric, Kinematic, and Dynamic Similarity

131

Model Theory

132

Closure

134

Viscous Flow

12.1

12.2

99

**The Differential Continuity Equation
**

Navier-Stokes Equations

101

Bernoulli’s Equation

110

Closure

111

137

Reynolds’s Experiment

Drag

138

137

88

92

**Fully Developed Laminar Flow in a Circular Conduit of Constant
**

92

Cross Section

Laminar Flow of a Newtonian Fluid Down an Inclined-Plane Surface

Closure

97

**Dimensional Analysis and Similitude
**

11.1

11.2

11.3

11.4

11.5

11.6

12.

**Newton’s Viscosity Relation
**

81

Non-Newtonian Fluids

82

Viscosity

83

Shear Stress in Multidimensional Laminar Flows of a Newtonian Fluid

Closure

90

**Inviscid Fluid Flow
**

10.1

10.2

10.3

10.4

10.5

10.6

10.7

10.8

10.9

11.

81

**Differential Equations of Fluid Flow
**

9.1

9.2

9.3

9.4

10.

72

**Analysis of a Differential Fluid Element in Laminar Flow
**

8.1

9.

**The Bernoulli Equation
**

Closure

76

126

95

Contents

12.3

12.4

12.5

12.6

12.7

12.8

12.9

12.10

12.11

12.12

12.13

12.14

12.15

12.16

13.

**Flow in Closed Conduits
**

13.1

13.2

13.3

13.4

13.5

13.6

14.

16.

185

Centrifugal Pumps

186

Scaling Laws for Pumps and Fans

194

Axial and Mixed Flow Pump Conﬁgurations

Turbines

197

Closure

197

197

201

Conduction

201

Thermal Conductivity

202

Convection

207

Radiation

209

Combined Mechanisms of Heat Transfer

Closure

213

**Differential Equations of Heat Transfer
**

16.1

16.2

16.3

16.4

165

168

**Fundamentals of Heat Transfer
**

15.1

15.2

15.3

15.4

15.5

15.6

146

**Dimensional Analysis of Conduit Flow
**

168

Friction Factors for Fully Developed Laminar, Turbulent,

and Transition Flow in Circular Conduits

170

Friction Factor and Head-Loss Determination for Pipe Flow

173

Pipe-Flow Analysis

176

Friction Factors for Flow in the Entrance to a Circular Conduit

179

Closure

182

Fluid Machinery

14.1

14.2

14.3

14.4

14.5

15.

**The Boundary-Layer Concept
**

144

The Boundary-Layer Equations

145

Blasius’s Solution for the Laminar Boundary Layer on a Flat Plate

Flow with a Pressure Gradient

150

von Ka´rma´n Momentum Integral Analysis

152

Description of Turbulence

155

Turbulent Shearing Stresses

157

The Mixing-Length Hypothesis

158

Velocity Distribution from the Mixing-Length Theory

160

The Universal Velocity Distribution

161

Further Empirical Relations for Turbulent Flow

162

The Turbulent Boundary Layer on a Flat Plate

163

Factors Affecting the Transition From Laminar to Turbulent Flow

Closure

165

209

217

**The General Differential Equation for Energy Transfer
**

217

Special Forms of the Differential Energy Equation

220

Commonly Encountered Boundary Conditions

221

Closure

222

ix

x

Contents

17.

**Steady-State Conduction
**

17.1

17.2

17.3

17.4

17.5

18.

19.

21.

21.1

21.2

21.3

22.

**Fundamental Considerations in Convective Heat Transfer
**

274

Signiﬁcant Parameters in Convective Heat Transfer

275

Dimensional Analysis of Convective Energy Transfer

276

Exact Analysis of the Laminar Boundary Layer

279

Approximate Integral Analysis of the Thermal Boundary Layer

Energy- and Momentum-Transfer Analogies

285

Turbulent Flow Considerations

287

Closure

293

Natural Convection

297

Forced Convection for Internal Flow

Forced Convection for External Flow

Closure

318

**Boiling and Condensation
**

Boiling

323

Condensation

Closure

334

22.1

22.2

22.3

22.4

22.5

22.6

297

305

311

323

328

Heat-Transfer Equipment

266

274

**Convective Heat-Transfer Correlations
**

20.1

20.2

20.3

20.4

230

252

Analytical Solutions

252

Temperature-Time Charts for Simple Geometric Shapes

261

Numerical Methods for Transient Conduction Analysis

263

An Integral Method for One-Dimensional Unsteady Conduction

Closure

270

**Convective Heat Transfer
**

19.1

19.2

19.3

19.4

19.5

19.6

19.7

19.8

20.

**One-Dimensional Conduction
**

224

One-Dimensional Conduction with Internal Generation of Energy

Heat Transfer from Extended Surfaces

233

Two- and Three-Dimensional Systems

240

Closure

246

**Unsteady-State Conduction
**

18.1

18.2

18.3

18.4

18.5

224

336

**Types of Heat Exchangers
**

336

Single-Pass Heat-Exchanger Analysis: The Log-Mean Temperature

Difference

339

Crossﬂow and Shell-and-Tube Heat-Exchanger Analysis

343

The Number-of-Transfer-Units (NTU) Method of Heat-Exchanger

Analysis and Design

347

Additional Considerations in Heat-Exchanger Design

354

Closure

356

283

Contents

23.

**Radiation Heat Transfer
**

23.1

23.2

23.3

23.4

23.5

23.6

23.7

23.8

23.9

23.10

23.11

23.12

23.13

24.

25.5

399

407

428

27.4

27.5

436

452

**One-Dimensional Mass Transfer Independent of Chemical Reaction
**

452

One-Dimensional Systems Associated with Chemical Reaction

463

Two- and Three-Dimensional Systems

474

Simultaneous Momentum, Heat, and Mass Transfer

479

Closure

488

**Unsteady-State Molecular Diffusion
**

27.1

27.2

27.3

433

**The Differential Equation for Mass Transfer
**

433

Special Forms of the Differential Mass-Transfer Equation

Commonly Encountered Boundary Conditions

438

Steps for Modeling Processes Involving Molecular

Diffusion

441

Closure

448

**Steady-State Molecular Diffusion
**

26.1

26.2

26.3

26.4

26.5

27.

**Molecular Mass Transfer
**

The Diffusion Coefﬁcient

Convective Mass Transfer

Closure

429

398

**Differential Equations of Mass Transfer
**

25.1

25.2

25.3

25.4

26.

Nature of Radiation

359

Thermal Radiation

360

The Intensity of Radiation

361

Planck’s Law of Radiation

363

Stefan-Boltzmann Law

365

Emissivity and Absorptivity of Solid Surfaces

367

Radiant Heat Transfer Between Black Bodies

370

Radiant Exchange in Black Enclosures

379

Radiant Exchange in Reradiating Surfaces Present

380

Radiant Heat Transfer Between Gray Surfaces

381

Radiation from Gases

388

The Radiation Heat-Transfer Coefﬁcient

392

Closure

393

**Fundamentals of Mass Transfer
**

24.1

24.2

24.3

24.4

25.

359

496

**Unsteady-State Diffusion and Fick’s Second Law
**

496

Transient Diffusion in a Semi-Inﬁnite Medium

497

Transient Diffusion in a Finite-Dimensional Medium Under Conditions of

Negligible Surface Resistance

500

Concentration-Time Charts for Simple Geometric Shapes

509

Closure

512

xi

xii

Contents

28.

**Convective Mass Transfer
**

28.1

28.2

28.3

28.4

28.5

28.6

28.7

28.8

29.

Equilibrium

551

Two-Resistance Theory

Closure

563

569

**Mass Transfer to Plates, Spheres, and Cylinders
**

569

Mass Transfer Involving Flow Through Pipes

580

Mass Transfer in Wetted-Wall Columns

581

Mass Transfer in Packed and Fluidized Beds

584

Gas-Liquid Mass Transfer in Stirred Tanks

585

Capacity Coefﬁcients for Packed Towers

587

Steps for Modeling Mass-Transfer Processes Involving Convection

Closure

595

**Mass-Transfer Equipment
**

31.1

31.2

31.3

31.4

31.5

31.6

31.7

551

554

**Convective Mass-Transfer Correlations
**

30.1

30.2

30.3

30.4

30.5

30.6

30.7

30.8

31.

**Fundamental Considerations in Convective Mass Transfer
**

517

Signiﬁcant Parameters in Convective Mass Transfer

519

Dimensional Analysis of Convective Mass Transfer

521

Exact Analysis of the Laminar Concentration Boundary Layer

524

Approximate Analysis of the Concentration Boundary Layer

531

Mass, Energy, and Momentum-Transfer Analogies

533

Models for Convective Mass-Transfer Coefﬁcients

542

Closure

545

**Convective Mass Transfer Between Phases
**

29.1

29.2

29.3

30.

517

588

603

**Types of Mass-Transfer Equipment
**

603

Gas-Liquid Mass-Transfer Operations in Well-Mixed Tanks

605

Mass Balances for Continuous Contact Towers: Operating-Line Equations

Enthalpy Balances for Continuous-Contact Towers

620

Mass-Transfer Capacity Coefﬁcients

621

Continuous-Contact Equipment Analysis

622

Closure

636

Nomenclature

611

641

APPENDIXES

A.

Transformations of the Operators = and =2 to Cylindrical Coordinates

B.

Summary of Differential Vector Operations in Various Coordinate Systems

C.

Symmetry of the Stress Tensor

D.

The Viscous Contribution to the Normal Stress

E.

**The Navier–Stokes Equations for Constant r and m in Cartesian,
**

Cylindrical, and Spherical Coordinates

657

F.

Charts for Solution of Unsteady Transport Problems

654

655

659

648

651

Mass-Transfer Diffusion Coefﬁcients in Binary Systems K. The Error Function 697 M. Standard Tubing Gages Author Index 703 Subject Index 705 694 700 678 691 xiii . Physical Properties of Gases and Liquids J. Physical Properties of Solids 672 675 I. Lennard–Jones Constants L.Contents G. Properties of the Standard Atmosphere H. Standard Pipe Sizes 698 N.

This page intentionally left blank .

Fundamentals of Momentum. Heat. and Mass Transfer 5th Edition .

This page intentionally left blank .

Some substances such as glass are technically classiﬁed as ﬂuids.and twentieth century analytical work in hydrodynamics with the empirical knowledge in hydraulics that man has collected over the ages. Excluding action-at-a-distance forces such as gravity. for example.1 FLUIDS AND THE CONTINUUM A ﬂuid is deﬁned as a substance that deforms continuously under the action of a shear stress. are composed of molecules whose numbers stagger the imagination.Chapter 1 Introduction to Momentum Transfer M omentum transfer in a ﬂuid involves the study of the motion of ﬂuids and the forces that produce these motions. In most cases it is convenient to think of a ﬂuid as a continuous distribution of matter or a continuum. the number of molecules per unit volume would be time-dependent for the microscopic volume even though the macroscopic volume had a constant number of 1 . In a cubic inch of air at room conditions there are some 1020 molecules. An important consequence of this deﬁnition is that when a ﬂuid is at rest. However. or momentum transfer. the number of molecules in a small volume of a gas at rest. which is historically ﬂuid mechanics. may be shown to be the result of microscopic (molecular) transfer of momentum. Thus the subject under consideration. like all matter. Fluids. If the volume were taken small enough. the rate of deformation in glass at normal temperatures is so small as to make its consideration as a ﬂuid impractical. Each area of study has its phraseology and nomenclature. certain instances in which the concept of a continuum is not valid. Both liquids and gases are ﬂuids. Any theory that would predict the individual motions of these many molecules would be extremely complex. which was veriﬁed by experiment. From Newton’s second law of motion it is known that force is directly related to the time rate of change of momentum of a system. The mating of these separately developed disciplines was started by Ludwig Prandtl in 1904 with his boundary-layer theory. Most engineering work is concerned with the macroscopic or bulk behavior of a ﬂuid rather than with the microscopic or molecular behavior. 1. such as those resulting from pressure and shear stress. is both analytical and experimental. may equally be termed momentum transfer. Momentum transfer being typical. far beyond our present abilities. of course. Consider. The history of ﬂuid mechanics shows the skillful blending of the nineteenth. There are. the forces acting on a ﬂuid. the basic deﬁnitions and concepts will be introduced in order to provide a basis for communication. there can be no shear stresses. Modern ﬂuid mechanics. Concept of a Continuum.

2 PROPERTIES AT A POINT When a ﬂuid is in motion. Molecular domain Continuum domain ∆m ∆V δV ∆V Figure 1. may vary from point to point in a ﬂuid and may also vary with respect to time as in a punctured automobile tire. and dV is the smallest volume surrounding the point for which statistical averages are meaningful. The treatment of ﬂuids as continua is valid whenever the smallest ﬂuid volume of interest contains a sufﬁcient number of molecules to make statistical averages meaningful. The validity of the continuum approach is seen to be dependent upon the type of information desired rather than the nature of the ﬂuid. The concept of a continuum would be valid only for the latter case. as it allows us to describe ﬂuid ﬂow in terms of continuous functions.2 Chapter 1 Introduction to Momentum Transfer molecules in it.1. taking r ¼ limDV!dV (Dm/DV) is extremely useful. at a particular point in the ﬂuid is deﬁned as r ¼ lim DV!dV Dm DV where Dm is the mass contained in a volume DV. The density. that is. The deﬁnition of some ﬂuid variables at a point is presented below.1 Density at a point. The concept of the density at a mathematical point. . Our immediate task is to deﬁne these properties at a point. Under ﬂow conditions. particularly in gases. 1. however. r. Density at a Point. the density may vary greatly throughout the ﬂuid. The macroscopic properties of a continuum are considered to vary smoothly (continuously) from point to point in the ﬂuid. in general. the quantities associated with the state and the motion of the ﬂuid will vary from point to point. at DV ¼ 0 is seen to be ﬁctitious. The density of a ﬂuid is deﬁned as the mass per unit volume. The density. The limit is shown in Figure 1.

3. DFn ¼ s ii DA lim DA!dA DFs ¼ tij DA Here sii is the normal stress and tij the shear stress. have densities that remain almost constant over wide ranges of pressure and temperature. On the contrary. pressure and frictional forces require physical contact for transmission. The student will recall that normal stress is positive in tension. the double-subscript stress notation as used in solid mechanics will be employed. for example. Forces acting on a ﬂuid are divided into two general groups: body forces and surface forces. Body forces are those which act without physical contact. Consider the force DF acting on an element DA of the body shown in Figure 1.1. The force per unit area or stress at a point is deﬁned as the limit of DF/DA as DA ! dA where dA is the smallest area for which statistical averages are meaningful ∆A lim DA!dA Figure 1. From a static viewpoint.1 1 Mathematically. air is a compressible ﬂuid and water incompressible. The force DF is resolved into components normal and parallel to the surface of the element. compressibility effects are considered a property of the ﬂow. and the student is hereby alerted to the importance of this concept.2 Properties at a Point 3 Fluid Properties and Flow Properties. Instead of being classiﬁed according to the ﬂuid. ∆F ∆Fn ∆Fs Stress at a Point. . as it requires magnitude. direction. Fluids which exhibit this quality are usually treated as being incompressible. As a surface is required for the action of these forces they are called surface forces. are more a property of the situation than of the ﬂuid itself. the ﬂow of air at low velocities is described by the same equations that describe the ﬂow of water. The limiting process for the normal stress is illustrated in Figure 1. and orientation with respect to a plane for its determination. often subtle.2. gravity and electrostatic forces. is made between the properties of the ﬂuid and the properties of the ﬂow. For example. particularly liquids. Stress is therefore a surface force per unit area. Molecular domain Continuum domain ∆Fn ∆A δA ∆A Figure 1. The effects of compressibility. In this text.2 Force on an element of ﬂuid. however. stress is classed as a tensor of second order.3 Normal stress at a point. A distinction. Some ﬂuids.

Thus. It may be recalled that there can be no shearing stress in a static ﬂuid. In the x direction DFx DFs sin u ¼ 0 Since sin u ¼ Dy/Ds. by evaluating the above equation s xx ¼ s ss In the y direction. the only surface forces present will be those due to normal stresses. one has DFy DFs Dx Dx Dy Dz rg ¼0 Ds 2 Dividing through by Dx Dz and taking the limit as before.4 Element in a static ﬂuid. is acted upon by gravity and normal stresses. SF ¼ 0. This element. Consider the element shown in Figure 1. ∆Fs y ∆Fx x ∆s ∆y z q ∆z ∆x ∆Fy Figure 1. the above equation becomes DFx DFs Dy ¼0 Ds Dividing through by Dy Dz and taking the limit as the volume of the element approaches zero. applying SF ¼ 0 yields DFy DFs cos u rg Dx Dy Dz ¼0 2 Since cos u ¼ Dx/Ds.4 Chapter 1 Introduction to Momentum Transfer Pressure at a point in a Static Fluid. we obtain. we obtain DFy DFs rgDy ¼0 lim 2 Ds Dz DV!0 Dx Dz (1-1) . The weight of the ﬂuid element is rg(Dx Dy Dz/2). For a static ﬂuid. we obtain DFx DFs lim ¼0 Ds Dz DV!0 Dy Dz Recalling that normal stress is positive in tension. the normal stress at a point may be determined from the application of Newton’s laws to a ﬂuid element as the ﬂuid element approaches zero size.4. For a body at rest. while at rest.

3 POINT-TO-POINT VARIATION OF PROPERTIES IN A FLUID In the continuum approach to momentum transfer. a tensor. The pressure varies continuously throughout the region. may also be shown for the case of zero shear stress in a ﬂowing ﬂuid. Scalars in this book will be set in regular type. the normal stress components at a point may not be equal. the concept of a gravitational ﬁeld has been introduced. and one may observe the pressure levels and infer the manner in which the pressure varies by examining such a map. When shearing stresses are present. we would observe that it acts inward or places the element in compression. let us investigate the manner in which ﬂuid properties vary from point to point. If we were to measure the force per unit area acting on a submerged element. In previous studies. the reduction of stress. a scalar. which in light of the preceding development. In this book. temperature. is a vector. vectors will be written in boldfaced type. the lines drawn are the loci of points of equal pressure.5. one being ﬂow in shock waves. This important simpliﬁcation. and thus a gravitational ﬁeld is a vector ﬁeld.3 Point-to-Point Variation of Properties in a Fluid which becomes s yy þ s ss or 5 rg ð0Þ ¼ 0 2 s yy ¼ s ss (1-2) It may be noted that the angle u does not appear in equation (1-1) or (1-2). of course. density. Weather maps illustrating the pressure variation over this country are published daily in our newspapers. thus the normal stress at a point in a static ﬂuid is independent of direction. use will be made of pressure. As pressure is a scalar quantity. and is therefore a scalar quantity. pressure. Figure 1. must be the negative of the normal stress.1. the only surface forces acting are those due to the normal stress. The quantity measured is. and stress ﬁelds. As the element is at rest. Gravity. to pressure. of course. Now that certain properties at a point have been discussed. In Figure 1. 1. the pressure is still equal to the average normal stress. . however. velocity. that is P ¼ 13ðs xx þ s yy þ s zz Þ with very few exceptions. such maps are an illustration of a scalar ﬁeld.5 Weather map—an example of a scalar ﬁeld.

and its functional relation describes the rate of change of P in the s direction. we have sin a @P/@x ¼ tan a ¼ cos a dP/ds¼0 @P/@y dP/ds¼0 . however. the partial derivatives represent the manner in which P changes along the x and y axes. respectively. respectively. may be written as dP @P @P ¼ cos a þ sin a ds @x @y (1-5) y Path s dy = sin α ds ds dy α dx = cos α ds dx x Figure 1. A small portion of the pressure ﬁeld depicted in Figure 1.5 by x and y.6 Chapter 1 Introduction to Momentum Transfer Of speciﬁc interest in momentum transfer is the description of the point-to-point variation in the pressure. The arbitrary path s is shown. The path for which the directional derivative is zero is quite simple to ﬁnd.6 Path s in the xy plane. y). and it is easily seen that the terms dx/ds and dy/ds are the cosine and sine of the path angle. Denoting the directions east and north in Figure 1. The change in P. a.5 is shown in Figure 1. between two points in the region separated by the distances dx and dy is given by the total differential dP ¼ @P @P dx þ dy @x @y (1-3) In equation (1-3). therefore. The directional derivative. with respect to the x axis. Along an arbitrary path s in the xy plane the total derivative is dP @P dx @P dy ¼ þ ds @x ds @y ds (1-4) In equation (1-4). the term dP/ds is the directional derivative.6. There are an inﬁnite number of paths to choose from in the xy plane. written as dP. we may represent the pressure throughout the region by the general function P(x. two particular paths are of special interest: the path for which dP/ds is zero and that for which dP/ds is maximum. Setting dP/ds equal to zero.

we have dP ¼ 0. grad P. The magnitude of the directional derivative when the directional derivative is maximum is dP @P @P cos a þ sin a ¼ ds max @x @y where cos a and sin a are evaluated along the path given by equation (1-7). we see that the two directions deﬁned by these equations are perpendicular. Thus the gradient of P. Paths along which a scalar is constant are called isolines. The directional derivative along the path of maximum value is frequently encountered in the anlaysis of transfer processes and is given a special name. In order to ﬁnd the direction for which dP/ds is a maximum. since tan a ¼ dy/dx. is @P @P ex þ ey grad P ¼ @x @y . or d dP @P @P ¼ sin a þ cos a ¼0 da ds @x @y or tan a dP/ds is max ¼ @P/@y @P/@x (1-7) Comparing relations (1-6) and (1-7).3 or. As the cosine is related to the tangent by 1 cos a ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ tan2 a we have cos a dP/ds is max @P=@x ¼ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ð@P/@xÞ2 þ ð@P/@yÞ2 Evaluating sin a in a similar manner gives sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 2 2 dP ð@P/@xÞ2 þ ð@P/@yÞ2 @P @P ¼ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ þ ds max @x @y ð@P/@xÞ2 þ ð@P/@yÞ2 (1-8) Equations (1-7) and (1-8) suggest that the maximum directional derivative is a vector of the form @P @P ex þ ey @x @y where ex and ey are unit vectors in the x and y directions. we must have the derivative (d/da)(dP/ds) equal to zero. respectively.1. the gradient. and thus P is constant. we have Point-to-Point Variation of Properties in a Fluid dy @P/@x ¼ dxdP/ds¼0 @P/@y 7 (1-6) Along the path whose slope is defined by equation (1-6).

is then equal to one kilogram meter per newton per second per second (1 kg m/N s2). The corresponding unit of force is the newton (N). y. and spherical coordinate systems are listed in Appendix B. giving =P ¼ @P @P @P ex þ ey þ ez @x @y @z where =¼ @ @ @ ex þ ey þ ez @x @y @z (1-10) Equation (1-10) is the deﬁning relationship for the = operator in Cartesian coordinates. The relationship between force and mass may be expressed by the following statement of Newton’s second law of motion: F¼ ma gc where gc is a conversion factor which is included to make the equation dimensionally consistent. length. and time in seconds. force. the geometric meaning of the gradient remains the same. length. This symbol indicates that differentiation is to be performed in a prescribed manner. The conversion factor. Using the SI system of units has greatly reduced these difﬁculties. length. such as cylindrical and spherical coordinates. Through the arbitrary choice of fundamental dimensions. it is a vector having the direction and magnitude of the maximum rate of change of the dependent variable with respect to distance. the gradient takes on a different form. and time are taken as basic units. and time.2 However. One newton is the force required to accelerate a mass of one kilogram at a rate of one meter per second per second (1 m/s2).4 UNITS In addition to the International Standard (SI) system of units. In the SI system. In engineering practice. 1. z).8 Chapter 1 Introduction to Momentum Transfer where P ¼ P(x. force is expressed in pounds force (lbf). gc. and time have been frequently chosen as deﬁning fundamental units. This concept can be extended to cases in which P ¼ P(x. 2 The forms of the gradient operator in rectangular. In other coordinate systems. y). mass. some confusion has occurred in the use of the English systems of units. In deﬁning each term of this law. mass. The corresponding unit of mass will be that which will be accelerated at the rate of 1 ft/(s)2 by 1 lbf. there are two different English systems of units commonly used in engineering. These systems have their roots in Newton’s second law of motion: force is equal to the time rate of change of momentum. and time in seconds (s). cylindrical. length in meters (m). length in feet. The basic units are mass in kilograms (kg). For this more general case grad P ¼ @P @P @P ex þ ey þ ez @x @y @z (1-9) Equation (1-9) may be written in more compact form by use of the operation = (pronounced ‘‘del’’). a direct relationship has been established between the four basic physical quantities used in mechanics: force. With this system. .

will correctly relate the units corresponding to a system. 3 In subsequent calculations in this book.1. it will be the reader’s responsibility to use units that are consistent with every term in the equation. The conversion factor. the unit of mass is 1 lbm. gc.1 System Length Time Force Mass 1 meter second newton kilogram 2 foot second lbf slug gc 1 kg m N s2 1 ðslugÞðftÞ ðlbf ÞðsÞ2 3 foot second lbf lbm 32:174 ðlbm ÞðftÞ ðlbf ÞðsÞ2 As all three systems are in current use in the technical literature. Therefore the conversion factor. gc. Careful checking for dimensional consistency will be required in all calculations. The bulk modulus of elasticity. The unit of force is 1 lbf. Liquids are generally considered to be incompressible whereas gases are certainly compressible. and mass. for this system is 32.5 Compressibility 9 This unit of mass having the dimensions of (lbf)(s)2/(ft) is called the slug.174 (ft)/(s)2. the student should be able to use formulas given in any particular situation. It is deﬁned according to b dP dV/V (1-11a) or as b dP dr/r and has the dimensions N/m2. force. The conversion factor. time. 1.5 COMPRESSIBILITY A ﬂuid is considered compressible or incompressible depending on whether its density is variable or constant. The force exerted by gravity on 1 lbm at sea level is deﬁned as 1 lbf. length and time are given in units of feet and seconds. gc will be rounded off to a value of 32. instead. is then a multiplying factor to convert slugs into (lbf)(s)2/(ft). its acceleration will be 32. often referred to as simply the bulk modulus. (1-11b) . gc. A third system encountered in engineering practice involves all four fundamental units. is a ﬂuid property that characterizes compressibility. and its value is 1 (slug)(ft)/(lbf)(s)2. There will be no attempt by the authors to incorporate the conversion factor in any equations.3 A summary of the values of gc is given in Table 1. Table 1. along with the units of length.1 for these three English systems of engineering units. respectively. When 1 lbm at sea level is allowed to fall under the inﬂuence of gravity.174 (lbm)(ft)/(lbf)(s)2.2 lbmft/2 lbf.

R ¼ 0. v. For air. M. The velocity is designated the acoustic velocity. in turn. where the air temperature is 239 K. be evaluated. undergoing an isentropic process where PV k ¼ C. a constant. It can be shown that the acoustic velocity is related to changes in pressure and density according to ½ dP C¼ (1-12) dr Introducing equation (1-11b) into this relationship yields b ½ C¼ r (1-13) For a gas.4. to the speed of sound. that is. (b) At v ¼ 650 km/hr ð180:5 m/sÞ M¼ v 180:5 m/s ¼ ¼ 0:582 C 310 m/s Compressible effects must be accounted for. in the ﬂuid: v M¼ (1-16) C A general rule of thumb is that when M < 0. The Mach number. k ¼ 1. C. may be treated in a ﬂow situation as incompressible. . a dimensionless parameter. the speed of sound in the ﬂuid. requires that the acoustic velocity at each airspeed. we have ½ kP C¼ (1-14) r or C ¼ ðkRTÞ½ (1-15) The question arises concerning when a gas. A common criterion for such a consideration involves the Mach number. when density variations are negligibly small. It is symbolized C. is deﬁned as the ratio of the ﬂuid velocity. Determine whether compressibility effects are signiﬁcant at airspeeds of (a) 220 km/h and (b) 650 km/h.287 kJ/kgK. which. EXAMPLE 1 A jet aircraft is ﬂying at an altitude of 15. and C ¼ ðkRTÞ½ ¼ ½1:4 ð0:287 kJ/kg KÞð239 KÞð1000 N m/kJÞðkg m/N s2 Þ½ ¼ 310 m/s (a) At v ¼ 220 km/hr ð61:1 m/sÞ M¼ v 61:1 m/s ¼ ¼ 0:197 C 310 m/s The ﬂow may be treated as incompressible.2 the flow may be treated as incompressible with negligible error. which is compressible.10 Chapter 1 Introduction to Momentum Transfer Disturbances introduced at some location in a ﬂuid continuum will be propagated at a ﬁnite velocity.500 m. The test for compressibility effects requires calculating the Mach number. that is.

Particles near the surface. Cleveland.028 0.2 lists values of s for several ﬂuids in air at 1 atm and 208C.2 Surface tensions of some ﬂuids in air at latm and 20 8C Fluid s(N/m) Ammonia Ethyl alcohol Gasoline Glycerin Kerosene Mercury Soap solution SAE 30 oil 0. DP.7 we show a free body diagram of a hemispherical drop of liquid with the pressure and surface tension forces in balance.1. in reality. Table 1. an interface between two phases.6 SURFACE TENSION The situation where a small amount of unconﬁned liquid forms a spherical drop is familiar to most of us. The phenomenon is the consequence of the attraction that exists between liquid molecules. The condition examined is typically used for this analysis as a sphere represents the minimum surface area for a prescribed volume. between the inside and outside of the hemisphere produces a net pressure force that is balanced by the surface tension force.02S 0.6 Surface Tension 11 1. A surface is. 2prs pr2∆P Figure 1.028 0. but a much stronger function of temperature.7 A free body diagram of a hemispherical liquid droplet. 62nd Ed. In Figure 1. Thus both phases will have the property of surface tension.440 0.035 Source: Handbook of Chemistry and Physics. Within a drop a molecule of liquid is completely surrounded by many others. but many others are also possible. Given this behavior. Nm/m2 or force per unit length of interface in N/m. will experience an imbalance of net force because of the nonuniformity in the numbers of adjacent particles. Chemical Rubber Publishing Co. For water in air the surface tension is expressed as a function of temperature according to s ¼ 0:123 ð1 0:00139 TÞN/m (1-17) where T is in Kelvins. s. it is evident that some work must be done when a liquid particle moves toward the surface. on the contrary. The most common materials involving phase interfaces are water and air. For a given interfacial composition. the surface tension property is a function of both pressure and temperature. The extreme condition is the density discontinuity at the surface. This force balance can be expressed as pr2 DP ¼ 2prs and the pressure difference is given by DP ¼ 2s r (1-18) . As more ﬂuid is added the drop will expand creating additional surface. 1980. Particles at the surface experience a relatively strong inwardly directed attractive force.021 0. symbolized. The pressure difference. Quantitatively.022 0.063 0. s is the work per unit area. The work associated with creating this new surface is the surface tension.. Table 1. OH.

The limit of this relationship is the case of a fully wetted surface where r ﬃ 1. Water in contact with a clean glass surface will completely wet the surface and. This is the result of attraction between the liquid molecules and the tube wall being greater than the attraction between water molecules at the liquid surface. is associated with u > 908. as shown in wetting gas–liquid–solid interface.8 Contact angle for a nonthrough the liquid. For mercury in contact with a clean glass tube u ﬃ 1308. Liquid A consequence of the pressure difference resulting from surface tension is the phenomenon q q of capillary action. deﬁned as illustrated in Figure 1. for this case. u. u ﬃ 0.12 Chapter 1 Introduction to Momentum Transfer For the case of a soap bubble. Illustrated in Figure 1. the intermolecular forces at the liquid surface are greater than the attractive forces between . The indicator for Solid wetting or nonwetting is the contact angle.9 is the case of a small glass tube inserted into a pool of (a) water and (b) mercury. the ﬁgure. and the Gas pressure difference due to surface tension is zero. For the mercury case. This effect is related to how well a liquid wets a solid boundary. With u measured Figure 1. there are two interfaces and the pressure difference will be DP ¼ 4s r (1-19) Equations (1-18) and (1-19) indicate that the pressure difference is inversely proportional to r. Note that water will rise in the tube and that in mercury the level in the tube is depressed. For a wetting case u < 908.9 Capillary effects with a tube inserted in (a) water and (b) mercury. h h (a) (b) Figure 1.8. which has a very thin wall. the liquid rises a distance h above the level in the pool. For the water case. a nonwetting case.

thus.10 Free body diagram of a wetting liquid in a tube. The upward force. h Figure 1. Equation (1-20) applies.11 Capillary depression of mercury in a glass tube. for mercury and glass. For mercury at 208C r ¼ 13. depressed a distance h below the level of the pool.6 2prs Surface Tension 13 liquid mercury and the glass surface. Equating these forces we obtain 2pr s cos u ¼ rgpr 2 h and the value of h becomes pr2h Figure 1.580 kg/m3. and for mercury in air s ¼ 0. so we have h¼ 2s cos u rgr Recall that.11). EXAMPLE 2 h¼ 2s cos u rgr (1-20) Determine the distance h that mercury will be depressed with a 4-mm-diameter glass tube inserted into a pool of mercury at 208C (Figure 1.44 N/m (Table 1.2) giving h¼ 2ð0:44 N/mÞðcos130 Þ ð13580kg/m3 Þð9:81m/s2 Þð2 103 mÞ ¼ 2:12 103 m ð2:12 mmÞ .10.1. due to surface tension 2prs cos u will be equal to the downward force due to the weight of liquid having volume V ¼ pr2h. A free body diagram of the wetting liquid is shown in Figure 1. The mercury is. u ¼ 1308.

and b are constants. 1. b) at time t ¼ (L2/a) ln e when the temperature ﬁeld is given by 2 x y T ¼ T0 eat=4L sin cosh a b where T0. v 1. x). . V1 .11 Find the pressure gradient at point (a.3 For a ﬂuid of density r in which solid particles of density rs are uniformly dispersed. a. Determine its bulk modulus of elasticity.1 and the properties of the standard atmosphere given in Appendix G.7 and 1.9 Using the geometric relations given below and the chain rule for differentiation. there are 4 1020 molecules per in. 5). At standard conditions. y. pressure density speciﬁc heat temperature stress pressure gradient velocity speed of sound 1.11)? 1. u) when a2 c ¼ A r sin u 1 2 r 1.15 If the ﬂuid of density r in Problem 1.5 What pressure change is required to change the density of air by 10% under standard conditions? 1. b. 5).16 Using the expression for the gradient in polar coordinates (Appendix A). a. Will this result be valid if a liquid is present instead of a solid? 1.12 dimensionally homogeneous? What must the units of r1 be in order that the pressure be in pounds per square foot when v 1 is given in feet per second (problem 1. ﬁnd the gradient of c(r.000 ft. r.3.11 and 1.50 MPa. 1.1 The number of molecules crossing a unit area per unit time in one direction is given by N ¼ 14 nv where n is the number of molecules per unit volume and v is the average molecular velocity. a.6 Using the information given in Problem 1. As the average molecular velocity is approximately equal to the speed of sound in a perfect gas. How much water will overﬂow if the water is heated to 808C? 1.10 Transform the operator = to cylindrical coordinates (r. show that @ sin u @ @ ¼ þ cos u @x r @u @r 1.19 A liquid in a cylinder has a volume of 1200 cm3 at 1.17 Given the following expression for the pressure ﬁeld where x. 1. estimate the number of molecules crossing a circular hole 103 in. Find the component of rf that makes a 608 angle with the x axis at the point (3. show that if x is the mass fraction of solid in the mixture. 1. in diameter. 1. rm. Determine the pressure in psi required to increase water density by 1% above its nominal value.b) when the pressure ﬁeld is given by x y x P ¼ r1 v 21 sin sin þ 2 a b a where r1. z). 1. Assume that the gas is at standard conditions. that is. estimate the number of molecules per cubic inch at an altitude of 250. and b are constants. and z are space coordinates. show that der/du ¼ eu and deu /du ¼ er .25 MPa and a volume of 1188 cm3 at 2.9. ﬁnd the pressure gradient x 2 V t xyz 1 2 1 þ P ¼ P0 þ 2rV1 2 3 þ 3 L L L 1.18 A vertical cylindrical tank having a base diameter of 10 m and a height of 5 m is ﬁlled to the top with water at 208C.2 Which of the quantities listed below are ﬂow properties and which are ﬂuid properties? and @ cos u @ @ ¼ þ sin u @y r @u @r when r2 ¼ x2 þ y2 and tan u ¼ y/x. and P0. a. t is time.7 Show that the unit vectors er and eu in a cylindrical coordinate system are related to the unit vectors ex and ey by Where is the gradient maximum? The terms A and a are constant.7.4 rs r rx þ rs ð1 xÞ An equation linking water density and pressure is 7 PþB r ¼ P1 þ B r1 where the pressure is in atmospheres and B ¼ 3000 atm. using the results of Problems 1. er ¼ ex cos u þ ey sin u and eu ¼ ex sin u þ ey cos u 1.14 Chapter 1 Introduction to Momentum Transfer PROBLEMS 1.14 A scalar ﬁeld is given by the function f ¼ 3x2y þ 4y2. Find rf at the point (3. (RT/M). 1. 1.13 Are the ﬁelds described in Problems 1.8 Using the results of Problem 1. obtain the equation of state of the mixture. the density is given by rmixture ¼ 1.12 Find the temperature gradient at point (a.3 obeys the perfect gas law. u. and L are constants. P ¼ f (rs.

Contact angles are 08 for water and 1308 for mercury.47 N/m. Determine the change in pressure required to reduce a given volume by 0. The water is subjected to a pressure of 120 MPa.2875 cm. . determine how high the water will rise.27 Determine the difference in pressure between the inside and outside of a soap ﬁlm bubble at 20 8C if the diameter of the bubble is 4 mm.30 Determine the diameter of the glass tube necessary to keep the capillary-height change of water at 308C less than 1 mm. causing a volume reduction of 0.205 GPa.0662 N/m and that of mercury is 0.20 A pressure of 10 MPa is applied to 0. having a diameter of 3 mm. 1. Determine how far the column of mercury in the tube will be depressed for a contact angle of 1308.25 A glass tube having an inside diameter of 0. clean glass tube. 1. are dipped in water. Determine the percentage decrease in its volume.625 mm. 1. 1. If s ¼ 0.23 Determine the height to which water at 688C will rise in a clean capillary tube having a diameter of 0. 1. Determine the upward force on the glass.22 Water in a container is originally at 100 kPa. separated by a gap of 1. Determine the bulk modulus of elasticity.26 Determine the capillary rise for a water–air–glass interface at 408C in a clean glass tube having a radius of 1 mm.55 mm.25 mm and an outside diameter of 0.75%.21 The bulk modulus of elasticity for water is 2.Problems 15 1. 1. 1.25 m3 of a liquid. 1.29 At 608C the surface tension of water is 0. is inserted vertically into a dish of mercury at 208C.0735 N/m.28 An open.005 m3. 1. 1.24 Two clean and parallel glass plates.35 mm is inserted into a pool of mercury at 208C such that the contact angle is 1308. Determine the capillary height changes in these two ﬂuids when they are in contact with air in a glass tube of diameter 0.

the point-to-point variation of a particular variable. If. we are well within normal limits of accuracy to neglect the absolute acceleration of the coordinate system that. A frequently encountered static situation exists for a ﬂuid that is stationary on Earth’s surface. . In this chapter. y P6 P3 P1 P2 ∆y P4 ∆z ∆x P5 x z 16 Figure 2. This means that the only forces acting on the ﬂuid are those due to gravity and pressure. a ﬂuid is stationary with respect to a coordinate system that has some signiﬁcant absolute acceleration of its own. is the element of ﬂuid Dx Dy Dz with a corner at the point xyz. Newton’s law may be satisﬁed by applying it to an arbitrary free body of ﬂuid of differential size. The application of Newton’s second law of motion to a ﬁxed mass of ﬂuid reduces to the expression that the sum of the external forces is equal to the product of the mass and its acceleration. whereas the more general statement åF ¼ ma must be used for the noninertial case.1. The free body selected. Although Earth has some motion of its own. In the case of an inertial reference. An example of this latter situation would be the ﬂuid in a railroad tank car as it travels around a curved portion of track.1 Pressure forces on a static ﬂuid element. shown in Figure 2.Chapter 2 Fluid Statics The deﬁnition of a ﬂuid variable at a point was considered in Chapter 1. As the sum of the forces must equal zero throughout the ﬂuid. The coordinate system xyz is an inertial coordinate system. Such a coordinate system is said to be an inertial reference. it is known that there can be no shear stress in a ﬂuid at rest. we would naturally have the relation åF ¼ 0. the reference is said to be noninertial. will be considered for the special case of a ﬂuid at rest. on the contrary. in this situation.1 PRESSURE VARIATION IN A STATIC FLUID From the deﬁnition of a ﬂuid. 2. pressure. would be ﬁxed with reference to Earth.

Choosing the y axis in the direction shown. As the size of the element approaches zero. and Dz approach zero and the element approaches the point (x. In the limit PjyþDy Pjy PjxþDx Pjx PjzþDz Pjz ex þ ey þ ez rg ¼ lim Dx Dy Dz Dx. and so on. EXAMPLE 1 The manometer.Dz!0 or rg ¼ @P @P @P ex þ ey þ ez @x @y @z (2-1) Recalling the form of the gradient. along with the force due to gravity acting on the element rg Dx Dy Dz. we have Patm PC ¼ rm gd2 and then integrating between B and A in the tank ﬂuid. may be analyzed from the previous discussion. To ﬁnd the sum of the forces on the element. a pressure measuring device. we see that equation (2-2) becomes dP ey ¼ rgey dy Integrating between C and D in the manometer ﬂuid. z). constant pressure lines are perpendicular to the gravitational vector. Dy. we obtain PA PB ¼ rT gd1 .2. P2 ¼ PjxþDx . Evaluating the forces acting on each face. The simplest type of manometer is the U-tube manometer shown in Figure 2. the pressure on each face must ﬁrst be evaluated. as isolines are perpendicular to the gradient. The point-to-point variation in pressure may be obtained by integrating equation (2-2). we may write equation (2-1) as rg ¼ =P (2-2) Equation (2-2) is the basic equation of ﬂuid statics and states that the maximum rate of change of pressure occurs in the direction of the gravitational vector. y. The pressure in the tank at point A is to be measured.Dy. The ﬂuid in the tank extends into the manometer to point B. P1 ¼ Pjx . For example.1 Pressure Variation in a Static Fluid 17 The pressures that act on the various faces of the element are numbered 1 through 6.2. we ﬁnd that the sum of the forces is rg(Dx Dy Dz) þ (Pjx PjxþDx ) Dy Dzex þ (Pjy PjyþDy )Dx Dzey þ (Pjz PjzþDz )Dx Dyez ¼ 0 Dividing by the volume of the element Dx Dy Dz. Dx. We shall designate the pressure according to the face of the element upon which the pressure acts. In addition. we see that the above equation becomes rg PjyþDy Pjy PjxþDx Pjx Pj Pjz ex ey zþDz ez ¼ 0 Dx Dy Dz where the order of the pressure terms has been reversed.

18 Chapter 2 Fluid Statics Patm g D A y d2 d1 B C Manometer fluid – rM Fluid in tank – rT Figure 2. we see that equation (2-2) becomes dP PMg ¼ rg ¼ dy RT Separating variables. As elevations B and C are equal. rockets. a . a relation between the pressure and density is required to integrate equation (2-2). PB and PC. We may. thus. we see that the above differential equation becomes dP Mg ¼ dy P RT Integration between y ¼ 0 (where P ¼ Patm) and y ¼ y (where the pressure is P) yields ln or P Mgy ¼ Patm RT P Mgy ¼ exp Patm RT In the above examples. This difference is called the gage pressure and is frequently used in pressure measurement. which is constant for this case. and density.2 A U-Tube manometer. must be the same. where P ¼ rRT/M. the atmospheric pressure and a model of pressure variation with elevation have appeared in the results. Here R is the universal gas constant. the pressures. temperature. As performance of aircraft. The simplest case is that of the isothermal perfect gas. and T is the temperature. Selecting the y axis parallel to g. EXAMPLE 2 In the ﬂuid statics of gases. M is the molecular weight of the gas. combine the above equations to obtain PA Patm ¼ rm gd2 rT gd1 The U-tube manometer measures the difference between the absolute pressure and the atmospheric pressure. and many types of industrial machinery varies with ambient pressure.

however. Thus. Hg ¼ 14. In the case of uniform rectilinear acceleration. Choosing the y axis parallel to g a.1 is not an inertial coordinate system.2. The point-to-point variation in pressure is obtained from integration of equation (2-3). standard atmospheric conditions are P ¼ 29:92 in: Hg ¼ 2116:2 lbf /ft2 ¼ 14:696 lbf /in:2 ¼ 101 325 N/m2 T ¼ 519 R ¼ 59 F ¼ 288 K r ¼ 0:07651 lbm /ft3 ¼ 0:002378 slug/ft3 ¼ 1:226 kg/m3 A table of the standard atmospheric properties as a function of altitude is given in Appendix G. pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ dP ey ¼ rjg ajey ¼ r g2 þ a2 ey dy Vent Fuel g B To fuel pump Figure 2. and lines of constant pressure are perpendicular to g a. .2 Uniform Rectilinear Acceleration 19 standard atmosphere has been established in order to evaluate performance. therefore the surface of the ﬂuid will be perpendicular to this direction. the ﬂuid will be at rest with respect to the accelerating coordinate system.92 in. At sea level.4.2 UNIFORM RECTILINEAR ACCELERATION For the case in which the coordinate system xyz in Figure 2. 1 These performance standard sea-level conditions should not be confused with gas-law standard conditions of P ¼ 29. If the tank is given a uniform acceleration to the right.3. as required by Newton’s second law of motion. With a constant acceleration. we ﬁnd that equation (2-3) may be integrated between point B and the surface. The result is =P ¼ r(g a) (2-3) The maximum rate of change of pressure is now in the g a direction.2 ¼ 101 325 Pa. The pressure gradient becomes dP/dy ey with the selection of the y axis parallel to g a as shown in Figure 2. what will be the pressure at point B? From equation (2-3) the pressure gradient is in the g a direction. we may apply the same analysis as in the case of the inertial coordinate system except that åF ¼ ma ¼ r Dx Dy Dza.1 2. T ¼ 4928R ¼ 328F ¼ 273 K. EXAMPLE 3 A fuel tank is shown in Figure 2. equation (2-2) does not apply.696 lb/in.3 Fuel tank at rest.

5 A submerged plane surface.20 Chapter 2 Fluid Statics y q g–a g a d –a Figure 2. B Integrating between y ¼ 0 and y ¼ d yields pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ Patm PB ¼ r g2 þ a2 ( d) or PB Patm ¼ r pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ g2 þ a2 (d) The depth of the ﬂuid.5 is inclined at an angle a to the surface of the ﬂuid. The plane surface illustrated in Figure 2. use will be made of the relations describing the point-topoint variation in pressure which have been developed in the previous sections. at point B is determined from the tank geometry and the angle u. y a a g a C η η PG Centroid of area b b C View C–C η η Figure 2.4 Uniformly accelerated fuel tank. d. As these forces are due to pressure.3 FORCES ON SUBMERGED SURFACES Determination of the force on submerged surfaces is done frequently in ﬂuid statics. . 2. and the density of the ﬂuid is r. The area of the inclined plane is A.

we must find the point at which the total force on the plate must be concentrated in order to produce the same moment as the distributed pressure. Figure 2. .5 ft. The force on the window is F ¼ rg sin a Ah where a ¼ p/2 and (2-6) 1. The point at which this force acts (the center of pressure) is not the centroid of the area. giving dF ¼ rgh sin a dA Integration over the surface of the plate yields Z F ¼ rg sin a h dA A The deﬁnition or the centroid of the area is h Z 1 A h dA A and thus F ¼ rg sin a hA (2-4) Thus. In order to find the center of pressure.6 Submerged window. PG ¼ rgy ¼ rgh sin a.5 ft belowthe surface of a tank as shown in Figure 2. the magnitude and location of the force acting Find on the window. hc:p: h ¼ EXAMPLE 4 A circular viewing port is to be located 1. h ¼ 1:5 ft. we have hc:p: ¼ 1 Ah Z h2 dA ¼ A Iaa Ah (2-5) The moment of the area about the surface may be translated from an axis aa located at the ﬂuid surface to an axis bb through the centroid by Iaa ¼ Ibb þ h2 A and thus Ibb Ah The center of pressure is located below the centroid a distance Ibb /Ah.2. the force due to the pressure is equal to the pressure evaluated at the centroid of the submerged area multiplied by the submerged area. or Z Fhc:p: ¼ hPG dA A Substitution for the pressure yields Z Fhc:p: ¼ rg sin a h2 dA A and since F ¼ rg sin a hA. 1 ft.3 Forces on Submerged Surfaces 21 The magnitude of the force on the element dA is PG dA. where PG is the gage pressure.6.

7.2) acts as a fluid. The force on a unit width of the wall is obtained by integrating the pressure difference between the right and left sides of the wall. the force due to pressure is Z F ¼ (P Patm )(1) dy Water 1m Soil 3m Figure 2.22 Chapter 2 Fluid Statics the force is (62:4 lbm /ft3 )ð32:2 ft/s2 )(p/4 ft2 )(1:5 ft) 32:2 lbm ft/s2 lbf ¼ 73:5 lbf (327 N) F ¼ rgAh ¼ The force F acts at h þ EXAMPLE 5 Icentroid . Z Fyc:p: ¼ rH2 O g 0 yc:p: 1 y2 dy þ Z 4 y½1 þ 2:2ðy 1Þ dy 1 1 ¼ (1000 kg/m3 )(9:807 m/s2 )(1 m)(37:53 m3 ) ¼ 2:80 m (9:19 ft) (131 414 N) .7 Retaining wall. For a circular area. Taking the origin at the top of the wall and measuring y downward. If the watersaturated soil (specific gravity ¼ 2. so we obtain Ah pR4 hc:p: ¼ 1:5 þ ft ¼ 1:542 ft 4pR2 1:5 Rainwater collects behind the concrete retaining wall shown in Figure 2. determine the force and center of pressure on a 1-m width of the wall. Icentroid ¼ pR4 /4. The pressure difference in the region in contact with the water is P Patm ¼ rH2 O gy and the pressure difference in the region in contact with the soil is P Patm ¼ rH2 O g(1) þ 2:2 rH2 O g(y 1) The force F is F ¼ rH2 O g Z 1 y dy þ rH2 O g 0 Z 4 ½1 þ 2:2(y 1) dy 0 F ¼ (1000 kg/m3 )(9:807 m/s2 )(1 m)(13:4 m2 ) ¼ 131 414 N(29 546 lbf ) The center of pressure of the force on the wall is obtained by taking moments about the top of the wall.

or simply dA. the force on a curved submerged surface may be obtained from the weight on the volume BCO and the force on a submerged plane surface.4 Buoyancy 23 The force on a submerged curved surface may be obtained from knowledge of the force on a plane surface and the laws of statics. By consideration of the equilibrium of the ﬁctitious body BCO. Thus.2. The element of volume h dA has gravity and pressure forces acting on it. .9 Forces on submerged volume. Consider the curved surface BC illustrated in Figure 2.4 BUOYANCY The body shown in Figure 2. Since åF ¼ 0.9 is submerged in a ﬂuid with density r. we have FCB ¼ W FCO (2-7) The force of the fluid on the curved plate is the negative of this or W þ FCO. the force of the curved plate on body BCO may be evaluated. The component of the force due to the pressure on the top of the element is P2 dS2 cos a ey . The net F P2 y dS2 h x dA a dS1 z P1 Figure 2.8. 2. The resultant force F holds the body in equilibrium. The product dS2 cos a then is the projection of dS2 onto the xz plane.8 Submerged curved surface. O B g Weight of fluid in volume COB W FCO Force on plane surface OC C FBC Figure 2. where a is the angle between the plane of the element dS2 and the xz plane.

assuming constant densities. Expressing each of the pressures as Patm þ rw gh. we have Fy ¼ rw g½ð11Þð1Þ 10ð1Þ þ rc gV ¼ rw gV þ rc gV The ﬁrst term is seen to be a buoyant force. In the case of a ﬂoating body. we obtain. The resultant force F is composed of two parts. yields F ¼ (r rB )gVey (2-8) where V is the volume of the body. the buoyant force is rgVsey. Summing forces on the vertical direction. for our force balance rc gV þ rw gð11 ftÞð1 ft2 Þ rw gð10 ftÞð1 ft2 Þ þ Fy ¼ 0 Solving for Fy. solving for Fy. equal to the weight of displaced water. and W as rc gV. The body experiences an upward force equal to the weight of the displaced ﬂuid. where Vs is the submerged volume.24 Chapter 2 Fluid Statics pressure force on the element is (P1 P2 ) dA ey . The pressure forces on all lateral surfaces of the cube cancel. we obtain åFy ¼ W þ P(1)jbottom P(1)jtop þ Fy ¼ 0 where Fy is the additional force required to hold the cube in position. Finally. the resultant force will cause the body to ﬂoat on the surface. as they are at different depths. This is the well-known principle of Archimedes. EXAMPLE 6 A cube measuring 1 ft on a side is submerged so that its top face is 10 ft below the free surface of water. The difference in pressure P1 P2 may be expressed as rgh. and the resultant force on the element is dF ¼ (P1 P2 ) dA ey rB gh dA ey where rB is the density of the body. Determine the magnitude and direction of the applied force necessary to hold the cube in this position if it is made of (a) cork (r ¼ 10 lbm /ft3) (b) steel (r ¼ 490 lbm /ft3). Those on the top and bottom do not. When r > rB . we obtain (a) rc ¼ 10 lbm /ft3 Fy ¼ ð62:4 lbm /ft3 Þð32:2 ft/s2 Þð1 ft3 Þ ð10 lbm ft3 Þð32:2 ft/s2 Þð1 ft3 Þ þ 32:2 lbm ft/s2 lbf 32:2 lbm ft/s2 lbf ¼ 52:4 lbf ðdownwardÞð233 NÞ . the weight rB gVey and the buoyant force rgVey . so dF ¼ (r rB )gh dA ey Integration over the volume of the body.

Estimate as accurately as possible the altitude of the airplane above sea level. determine the pressure at this point in atmospheres. Our next task is to examine the behavior of ﬂuids in motion to describe the effect of that motion.2 above. to what depth h will the tank submerge? The air barometric pressure is 14. 2. PROBLEMS 2. 2. and an upward force was required. 2.7 psia.6 psia. In the second case. The application of Newton’s laws of motion led to the description of the point-to-point variation in ﬂuid pressure. Fundamental laws other than Newton’s laws of motion will be necessary for this analysis. 2. 2. where b is the compressibility. Hg. If the tank weights 250 lb. The thickness of tank wall may be neglected. determine the pressure and density at a point 32.5 CLOSURE In this chapter the behavior of static ﬂuids has been examined. and the temperature is 708F. Assuming sea water to have a constant density of 1050 kg/m3.00 psi. What effect do temperatureinduced density changes have on the lifting power of a rigid lighter-than-air craft? 2.1 in.025.7 Matter is attracted to the center of Earth with a force proportional to the radial distance from the center. compute the pressure at Earth’s center.3 In Problem 2. including manometry. the weight exceeded the buoyant force. and the buoyancy of ﬂoating objects. assuming the material behaves like a liquid.1 On a certain day the barometric pressure at sea level is 30. Using the known value of g at the surface where the radius is 6330 km. What is the gage pressure in sea water at that depth? The speciﬁc gravity of sea water is 1. and the temperature gage shows the air temperature to be 468F. and that the mean speciﬁc gravity is 5. from which force relations were developed. the buoyant force exceeded the weight of the cube. ﬁnd the depth at which the net force on the tank is zero. forces on plane and curved submerged surfaces. Assume b ¼ 300. . thus to keep it submerged 10 ft below the surface.5 The change in density due to temperature causes the takeoff and landing speeds of a heavier-than-air craft to increase as the square root of the temperature.000 ft below the surface of the sea. a downward force of over 52 lb was required. The pressure gage in an airplane in ﬂight indicates a pressure of 10.2 The open end of a cylindrical tank 2 ft in diameter and 3 ft high is submerged in water as shown. Speciﬁc applications have been considered.67.8 The deepest known point in the ocean is 11.Problems 25 rc ¼ 490 lbm /ft3 (b) (62:4 lbm /ft3 )ð32:2 ft/s2 )(1 ft3 ) (490 lbm ft3 )(32:2 ft/s2 )(1 ft3 ) þ 32:2 lbm ft/s2 lbf 32:2 lbm ft/s2 /lbf ¼ þ 427:6 lbf (upward)(1902 N) Fy ¼ In case (a). 2.6 The practical depth limit for a suited diver is about 185 m.4 If the density of sea water is approximated by the equation of state r ¼ r0 exp½ð p patm Þ/bÞ. What additional force is required to bring the top of the tank ﬂush with the water surface? Air 2 ft h 3 ft 2. The static analyses that have been considered will later be seen as special cases of more general relations governing ﬂuid behavior.034 m in the Mariana Trench in the Paciﬁc.

10 2.3 in.? d2 H2O d1 h Hg A d3 B H2O 2. Water temp.7 ft. what is the value of h if the vapor pressure of the liquid is 3 psia? 2. if d1 ¼ 2 ft.14 What is the pressure difference between points A and B if d1 ¼ 1. and (c) mercury (SG ¼ 13. 2. Kerosene A a d3 Hg d1 Mercury d2 2.26 Chapter 2 Fluid Statics 2. 15' A 10 in.13 Find the difference in pressure between tanks A and B.2 and inverting a tube full of this material as shown in the ﬁgure.6).8.0250). Find the pressure at point A.. A 2 in. H2O 1' Hg Hg . Which section has the higher pressure? H2O pA Open Oil 10' B 4 in. Air B b d4 45° H2O 5 in. d3 ¼ 2. and d4 ¼ 4 in.4 in.11 Using a liquid having a speciﬁc gravity of 1. d2 ¼ 6 in.. and d3 ¼ 6. d2 ¼ 1 in. Determine the pressure difference between points A and B in psi.9 Determine the depth change to cause a pressure increase of 1 atm for (a) water. (b) sea water (SG ¼ 1..15 A differential manometer is used to measure the pressure change caused by a ﬂow constriction in a piping system as shown. Air 2. = 150°F 7 in.12 What is the pressure pA in the ﬁgure? The speciﬁc gravity of the oil is 0.

6 m in diameter and to be centered 2 m below the surface level. 4 ft 4 ft. Neglect bearing friction. Each window is to be 0.22 A dam spillway gate holds back water of depth h.16 The car shown in the ﬁgure is accelerated to the right at a uniform rate.) 2. 4 ft at the base 4 ft high. If the point of contact is at O. Water r o a 2.75-m diameter beach ball to stop a drain in a swimming pool. 2.Problems 2. Obtain an expression that relates the drain diameter D and the minimum water depth h for which the ball will remain in place.23 It is desired to use a 0. has its 6in. side half buried at the bottom of a lake 23 ft deep. Water H2O 6 psig D h 4 ft Air Gate (a) (b) 2. Calculate the total horizontal load per unit . The top 12 ft behind the bulkhead consists of sea water with a density of 2 slugs/ft3.24 A watertight bulkhead 22 ft high forms a temporary dam for some construction work. but the bottom 10 ft begin a mixture of mud and water can be considered a ﬂuid of density 4 slugs/ft3. 10 ft 2. determine the required density of the log. (b) triangular.21 A rectangular block of concrete 3 ft 3ft 6 in.20 A circular log of radius r is to be used as a barrier as shown in the ﬁgure below.19 Find the minimum value of h for which the gate shown will rotate counterclockwise if the gate cross section is (a) rectangular.18 Glass viewing windows are to be installed in an aquarium. The gate weights 500 1b/ft and is hinged at A. What way will the balloon move relative to the car? 27 2.17 The tank is accelerated upward at a uniform rate. At what depth of water will the gate rise and permit water to ﬂow under it? 15 ft CG A h 6 ft 60° 2. What force is required to lift the block free of the bottom? What force is required to maintain the block in this position? (Concrete weights 150 1b/ft3. Find the force and location of the force acting on the window. Does the manometer level go up or down? P = 2 psig 2.

r. in terms of R. Develop an expression for x. spaced 1 m apart. the fraction of the ﬂoat submerged. are held together by cables. 8m M 2. An upward buoyant force F is required to shut the ballcock valve. The speciﬁc gravity of the wood is 0.29 A cubical piece of wood with an edge L in length ﬂoats in the water. Neglecting atmospheric pressure.90. AB and AC. F. Water at 27°C Pa Water 128 m Pa h 96 m A 2. 1m B 1m C P 1m 2.28 Chapter 2 Fluid Statics width and the location of the center of pressure measured from the bottom.28 The ﬂoat in a toilet tank is a sphere of radius R and is made of a material with density r.27 The dam shown below is 100 m wide. g. Determine the cable tension. The density of water is designated rw. 2.26 The ﬁgure below shows an open triangular channel in which the two sides. 2.25 The circular gate ABC has a 1 m radius and is hinged at B. between B and C. Determine the magnitude and location of the force on the inclined surface. Cable B C Water A 6m . determine the force P just sufﬁcient to keep the gate from opening when h ¼ 12 m. What moment M is required to hold the cube in the position shown? The right-hand edge of the block is ﬂush with the water. and rw .

’’ Already in our previous studies. apply to each and every ﬂow independently of the nature of the ﬂuid under consideration.2 FLUID-FLOW FIELDS: LAGRANGIAN AND EULERIAN REPRESENTATIONS The term ﬁeld refers to a quantity deﬁned as a function of position and time throughout a given region. The ﬁrst law of thermodynamics Equation Continuity equation Momentum theorem Energy equation The next three chapters will be devoted exclusively to the development of a suitable working form of these laws. we shall present the pertinent physical laws and discuss the methods used to describe a ﬂuid in motion. while some of the auxiliary relations that will be used will be called laws. the student will be responsible for noting the difference in scope between the fundamental physical laws and the auxiliary relations. However accurate these ‘‘laws’’ may be over a restricted range. The law of conservation of mass 2. 3. Hooke’s law. This is the familiar approach of particle and 1 The second law of thermodynamics is also fundamental to ﬂuid-ﬂow analysis. Unfortunately.1 FUNDAMENTAL PHYSICAL LAWS There are three fundamental physical laws that. the perfect gas law. The difference between these approaches lies in the manner in which the position in the ﬁeld is identiﬁed. most of these auxiliary relations have also been termed ‘‘laws. with the exception of relativistic and nuclear phenomena. and others have been encountered. In the Lagrangian approach. An analytic consideration of the second law is beyond the scope of the present treatment. Lagrange’s form and Euler’s form. Thus.1 In addition to the above laws. their validity is entirely dependent upon the nature of the material under consideration. the physical variables are described for a particular element of ﬂuid as it traverses the ﬂow. These laws are listed below with the designations of their mathematical formulations. Newton’s second law of motion 3. Law 1. 3.Chapter 3 Description of a Fluid in Motion The development of an analytical description of ﬂuid ﬂow is based upon the expression of the physical laws related to ﬂuid ﬂow in a suitable mathematical form. There are two different forms of representation for ﬁelds in ﬂuid mechanics. Accordingly. 29 . certain auxiliary or subsidiary relations are employed in describing a ﬂuid. These relations depend upon the nature of the ﬂuid under consideration.

the determination of the force on a stationary body in a flow field requires that we know the pressure and shear stress at every point on the body. y. equation (3-2) gives the velocity of the fluid at that location at time t1. The Lagrangian approach is seldom used in fluid mechanics. as shown in Figure 3. For a particular point (x1. In this text the Eulerian approach will be used exclusively. being functions of the same coordinates. t) (3-2) where x. we note that the ﬂuid ﬂow will. in general. y. may be represented in a similar manner. z. will vary as the vehicle approaches it. If the ﬂow at every point in the ﬂuid is independent of time. Consider an airplane ﬂying through the air at constant speed v 0 . for example. y. The coordinates (x. y P v0 x z Figure 3. The ﬂuid element is identiﬁed by its position in the ﬁeld at some arbitrary time. If the ﬂow at a point varies with time. z2) and t1. t). For example. the ﬂow is termed steady. z. b. In functional form the velocity ﬁeld is written as v ¼ v(x. When observed from the stationary x. 3. as the type of information desired is usually the value of a particular fluid variable at a fixed point in the flow rather than the value of a fluid variable experienced by an element of fluid along its trajectory. b. The Eulerian approach gives the value of a ﬂuid variable at a given point and at a given time.1. as such. z coordinate system. are functions of time. The ﬂow at the point P illustrated. the ﬂow pattern is unsteady. .1 Unsteady ﬂow with respect to a ﬁxed coordinate system.30 Chapter 3 Description of a Fluid in Motion rigid-body dynamics. The velocity ﬁeld in this case is written in functional form as v ¼ v(a. y1. y. z) are the coordinates of the element of ﬂuid and. z. The other fluid-flow variables. and t are all independent variables. It is possible in certain cases to reduce an unsteady ﬂow to a steady ﬂow by changing the frame of reference. usually t ¼ 0. The Eulerian form provides us with this type of information. t) (3-1) where the coordinates (a. the ﬂow is termed unsteady. y. be a function of the four independent variables (x.3 STEADY AND UNSTEADY FLOWS In adopting the Eulerian approach. c) refer to the initial position of the fluid element. The coordinates (x. z) are therefore dependent variables in the Lagrangian form. y. c.

Whenever a body moves through a ﬂuid with a constant velocity. Thus. It is obvious that path lines and streamlines are coincident only in steady ﬂow.3. Figure 3. The actual trajectory of a ﬂuid element as it traverses the ﬂow is designated as a path line. z0 coordinate system which is moving at constant velocity v 0 as illustrated in Figure 3. use is made of this concept. streamline patterns change from instant to instant. the ﬂow ﬁeld may be transformed from an unsteady ﬂow to a steady ﬂow by selecting a coordinate system that is ﬁxed with respect to the moving body. and thus the ﬂow is steady when viewed from the moving coordinate system. We shall make use of this transformation whenever applicable. hence a streamline is the trajectory of an element of ﬂuid in such a situation. the trajectory of a ﬂuid element will be different from a streamline at any particular time. In unsteady ﬂow.4 Streamlines 31 y P y' x' v0 z' x z Figure 3.3 shows the streamline pattern for ideal ﬂow past a football-like object. In the wind-tunnel testing of models.2 Steady ﬂow with respect to a moving coordinate system. In steady ﬂow. the path of a ﬂuid particle follows a streamline. 3.3 Illustration of streamlines. y0. The ﬂow conditions are now independent of time at every point in the ﬂow ﬁeld. A streamline is deﬁned as the line-drawn tangent to the velocity vector at each point in a ﬂow ﬁeld. Figure 3.2. The physical as well as the analytical simpliﬁcations afforded by this transformation are considerable. . Now consider the same situation when observed from the x0 . Data obtained for a static model in a moving stream will be the same as the data obtained for a model moving through a static stream.4 STREAMLINES A useful concept in the description of a ﬂowing ﬂuid is that of a streamline. as all velocity vectors are invariant with time.

Some additional discussion is provided in Chapter 10 regarding the mathematical description of streamlines around solid objects. F ¼ ma. The selection of the system for the application of these laws is quite ﬂexible and is. A system is deﬁned as a collection of matter of ﬁxed identity. or it may be accelerating (noninertial). Any analysis utilizing a fundamental law must follow the designation of a speciﬁc system. . m ¼ the mass of the system. consider Newton’s second law. and the difﬁculty of solution varies greatly depending on the choice made. Such a region is a control volume. 2 In the piston-and-cylinder arrangement shown in Figure 3. Primary consideration here will be given to inertial control volumes. Figure 3.5 Control volume for analysis of ﬂow through a nozzle. in many cases. a convenient system to analyze. a ¼ the acceleration of the center of mass of the system. In the case of the nozzle shown in Figure 3. As an illustration. In three dimensions this becomes dx dy dz ¼ ¼ (3-4) vx vy vz The utility of the above relations is in obtaining an analytical relation between velocity components and the streamline pattern. The basic laws give the interaction of a system with its surroundings.5. Thus different systems occupy the nozzle at different times. readily identiﬁed by virtue of its conﬁnement. a complex problem. 3. System Figure 3.5 SYSTEMS AND CONTROL VOLUMES The three basic physical laws listed in Section 3.2 A control volume may be ﬁxed or moving uniformly (inertial). A more convenient method of analysis of the nozzle would be to consider the region bounded by the dotted line. the ﬂuid occupying the nozzle changes from instant to instant.4. A control volume is a region in space through which ﬂuid ﬂows. is the mass of material enclosed within the cylinder by the piston.32 Chapter 3 Description of a Fluid in Motion The streamline is useful in relating that ﬂuid velocity components to the geometry of the ﬂow ﬁeld. For two-dimensional ﬂow this relation is v y dy (3-3) ¼ v x dx as the streamline is tangent to the velocity vector having x and y components v x and v y . The terms represented are as follows: F ¼ the resultant force exerted by the surroundings on the system.1 are all stated in terms of a system.4 An easily identiﬁable system.

5 Systems and Control Volumes 33 The extreme mobility of a ﬂuid makes the identiﬁcation of a particular system a tedious task. By developing the fundamental physical laws in a form that applies to a control volume (where the system changes from instant to instant). Succeeding chapters will convert the fundamental physical laws from the system approach to a control-volume approach. the analysis of ﬂuid ﬂow is greatly simpliﬁed. .3. The control-volume approach circumvents the difﬁculty in system identiﬁcation. The control volume selected may be either ﬁnite or inﬁnitesimal.

the rate of mass efﬂux ¼ (rv)(dA cos u). v. > : control volume volume volume Consider now the general control volume located in a ﬂuid ﬂow ﬁeld. From vector algebra. For the small element of area dA on the control surface. to dA. : . the law of conservation of mass may be simply stated as 9 9 8 9 8 8 rate of rate of mass > rate of mass > > > > > > > > > > > = = < = < < accumulation flow into efflux from ¼0 þ control > > of mass within > > control > > > > > > .1. n.1 Fluid ﬂow through a control volume.1 INTEGRAL RELATION The law of conservation of mass states that mass may be neither created nor destroyed. With respect to a control volume. we recognize the product rv dA cos u ¼ r dAjvj jnj cos u as the ‘‘scalar’’ or ‘‘dot’’ product r(v: n) dA 34 . Streamlines at time t dA q v n Figure 4. The integral relation thus developed will be applied to some often-encountered ﬂuid-ﬂow situations. we shall develop an integral relationship that expresses the law of conservation of mass for a general control volume.Chapter 4 Conservation of Mass: Control-Volume Approach The initial application of the fundamental laws of ﬂuid mechanics involves the law of conservation of mass. In this chapter. and u is the angle between the velocity vector. as shown in Figure 4. and the outward directed unit normal vector. > : . where dA cos u is the projection of the area dA in a plane normal to the velocity vector. v. 4.

this product represents the amount of mass ﬂowing through a unit cross-sectional area per unit time. @/@t c:v: r dV. . For incompressible ﬂow the density. if the integral is positive. the accumulation term. the mass within the control volume is constant. hence the partial derivative with respect to time is zero. The product rv is the mass ﬂux. the density term in the surface integral may be canceled. will be zero. ﬂowing inward across the control surface.4. the product v: n ¼ jvj jnj cos u is negative. and cos u is therefore negative. Additionally. the properties of a ﬂow ﬁeld are invariant with time. This is readily seen when one recalls that. The conservation-of-mass expression for incompressible ﬂow of this nature thus becomes ZZ (v: n) dA ¼ 0 (4-3) c:s: The following examples illustrate the application of equation (4-1) to some cases that recur frequently in momentum transfer. If we now integrate this quantity over the entire control surface. by the deﬁnition of steady ﬂow. for this situation the applicable form of the continuity expression is ZZ r(v: n) dA ¼ 0 (4-2) c:s: Another important case is that of an incompressible ﬂow with ﬂuid ﬁlling the control volume. Note that if mass is entering the control volume. is constant. we have ZZ r(v: n) dA c:s: which is the net outward ﬂow of mass across the control surface. If ﬂowRRR is steady relative to coordinates ﬁxed to the control volume. G. there is a net inﬂux of mass. since u > 90 . negative.2 SPECIFIC FORMS OF THE INTEGRAL EXPRESSION Equation (4-1) gives the mass balance in its most general form. Thus. Thus. that is. r.2 Speciﬁc Forms of the Integral Expression 35 which is the form we shall use to designate the rate of mass efﬂux through dA. or the net mass efﬂux from the control volume. We now consider some frequently encountered situations where equation (4-1) may be applied. often called the mass velocity. Physically. The rate of accumulation of mass within the control volume may be expressed as ZZZ @ r dV @t c:v: and the integral expression for the mass balance over a general control volume becomes ZZ ZZZ @ r(v: n) dA þ r dV ¼ 0 (4-1) @t c:s: c:v: 4. hence the accumulation term involving the partial derivative with respect to time is again zero. zero. there is a net efﬂux of mass.

this is the beauty of the control-volume approach. as the velocity vectors and outwardly directed normal vectors are collinear both at (1) and (2).2. the two vectors are opposite in sense. Speciﬁcally.2 Steady one-dimensional ﬂow into and out of a control volume. our expression is ZZ ZZ ZZ r(v: n) dA ¼ r(v: n) dA þ r(v: n) dA ¼ 0 c:s: A1 A2 The absolute value of the scalar product (v: n) is equal to the magnitude of the velocity in each integral. for which the ﬂow area is circular and the velocity proﬁle is parabolic (see Figure 4. let us consider the common situation of a control volume for which mass efﬂux and inﬂux are steady and one dimensional. In solving Example 1. As mass crosses the control surface at positions (1) and (2) only. thus this product is positive.2 is deﬁned for analytical purposes. but a more general case is one in which the velocity varies over the cross-sectional area. 2 Equation (4-2) applies. EXAMPLE 2 Let us consider the case of an incompressible ﬂow. as it should be for an efﬂux of mass.36 Chapter 4 EXAMPLE 1 Conservation of Mass: Control-Volume Approach As our ﬁrst example. varying according to the expression r 2 v ¼ v max 1 R . At (2) these vectors have the same sense. The box-shaped control volume illustrated in Figure 4. This situation may be approached physically. the actual system contained in this box could be as simple as a pipe or as complex as a propulsion system or a distillation tower. consider the control volume indicated by dashed lines in Figure 4. hence the sign is negative. the ﬂow inside the control volume can be analyzed from information (measurements) obtained on the surface of the control volume. it is noted that the ﬂow situation inside the control volume was unspeciﬁed. At (1). We may now express the continuity equation in scalar form ZZ ZZ ZZ r(v: n) dA ¼ rv dA þ rv dA ¼ 0 c:s: A1 A2 Integration gives the familiar result r1 v 1 A1 ¼ r2 v 2 A2 (4-4) In obtaining equation (4-4). In fact. where mass ﬂows into the control volume. we assumed a constant velocity at sections (1) and (2).3). v1 A1 r1 v2 A2 r2 1 Figure 4.

e. It will also be derived theoretically in Chapter 8 for the case of laminar ﬂow in a circular conduit. A tank initially contains 1000 kg of brine containing 10% salt by mass. Find the amount of salt in the tank at any time t. and the elapsed time when the amount of salt in the tank is 200 kg. and R is the radial distance to the inside surface of the circular area considered. at r ¼ 0). where v max is the maximum velocity which exists at the center of the circular passage (i. the mass rate of ﬂow is ZZ (rv)avg A ¼ rv dA A For the present case of incompressible ﬂow. we have v avg ¼ ¼ 1 A ZZ 1 pR2 v dA A Z 0 2p Z R 0 r 2 r dr du v max 1 R v max ¼ 2 In the previous examples. EXAMPLE 3 Let us now examine the situation illustrated in Figure 4. At the station where this velocity proﬁle exists.2 R cL Speciﬁc Forms of the Integral Expression 37 vmax r cL Figure 4. in this case. Brine is removed from the tank via an outlet pipe at a rate of 10 kg/min.4. As the average velocity is of particular interest in engineering problems. An inlet stream of brine containing 20% salt by mass ﬂows into the tank at a rate of 20 kg/min. This type application is common to chemical processes in particular. . r. we will now consider the means of obtaining the average velocity from this expression.3 A parabolic velocity proﬁle in a circular ﬂow passage.4. salt. Solving for the average velocity. The above velocity-proﬁle expression may be obtained experimentally. Equation (4-1) applies to ﬂuid streams containing more than one constituent as well as to the individual constituents alone.. the density is constant. from the center of the ﬂow section. This expression represents the velocity at a radial distance. Our ﬁnal example will use the law of conservation of mass for both the total mass and for a particular species. we were not concerned with the composition of the ﬂuid streams. The mixture in the tank is kept uniform by stirring.

Writing the complete expression. we may now apply equation (4-1) to the salt. initial content 1000 kg Control volume Figure 4. The concentration by weight of salt may be expressed as S S ¼ M 1000 þ 10t kg salt kg brine Using this deﬁnition. obtaining ZZ c:s: r(v: n) dA ¼ 10S (0:2)(20) 1000 þ 10t kg salt min and @ @t ZZZ r dV ¼ c:v: d dt Z S dS ¼ S0 dS dt kg salt min The complete expression is now ZZ c:s: r(v: n) dA þ @ @t ZZZ r dV ¼ c:v: S dS 4þ ¼0 100 þ t dt . For the control volume shown ZZ r(v: n) dA ¼ 10 20 ¼ 10 kg=min c:s: @ @t ZZZ r dV ¼ c:v: d dt Z M dM ¼ 1000 d (M 1000) dt where M is the total mass of brine in the tank at any time.38 Chapter 4 Conservation of Mass: Control-Volume Approach 20 kg/min Salt content 20% by mass Tank. 10 kg/min We ﬁrst apply equation (4-1) to express the total amount of brine in the tank as a function of time.4 A mixing process. we have ZZ c:s: r(v: n) dA þ @ @t ZZZ r dV ¼ 10 þ c:v: d (M 1000) ¼ 0 dt Separating variables and solving for M gives M ¼ 1000 þ 10t (kg) We now let S be the amount of salt in the tank at any time.

0) and (2.4 Water enters a 4-in.3 Water is ﬂowing through a large circular conduit with a velocity proﬁle given by the equation v ¼ 9(1 r 2 =16) fps. This approach may seem needlessly tedious at the outset. 4. square channel as shown at a velocity of 10 fps. 2).2 Using the velocity vector of the previous problem. Similar integral expressions for conservation of energy and of momentum for a general control volume will be developed and used in subsequent chapters.5 ft . Thus the ﬁrst part of the answer.6 min. expressing the amount of salt present as a function of time.000.1 The velocity vector in a two-dimensional ﬂow is given by the expression v ¼ 10ex þ 7xey m=s when x is measured in meters. Find the exiting water’s average velocity and total rate of ﬂow. . (b) the volume of ﬂow that crosses a plane surface connecting points (1. square conﬁguration as shown at the discharge end. PROBLEMS 4. The channel converges to a 2-in. The integral expression developed for this case was seen to be quite general in its form and use. 4 in. 2). 2 in .5-ft pipe? 4. Determine the component of the velocity that makes a 30 angle with the x axis at the point (2. Such temptations should be overcome. using the initial condition that S ¼ 100 at t ¼ 0 to give C ¼ 10. The outlet section is cut at 308 to the vertical as shown. but the mean velocity of the discharging water remains horizontal. r 30° 8 ft 4 in d = 1. 4. 4.Problems 39 This equation may be written in the form dS S þ ¼4 dt 100 þ t which we observe to be a ﬁrst-order linear differential equation. but it will always ensure a complete analysis of a problem and circumvent any errors that may otherwise result from a too-hasty consideration. is S¼ 10 000 þ 400t þ 2t2 100 þ t The elapsed time necessary for S to equal 200 kg may be evaluated to give t ¼ 36. determine (a) the equation of the streamline passing through point (2. There will be a strong temptation simply to write down an equation without considering each term in detail. 1). 2 in .3 CLOSURE In this chapter we have considered the ﬁrst of the fundamental laws of ﬂuid ﬂow: conservation of mass. The student should now develop the habit of always starting with the applicable integral expression and evaluating each term for a particular problem. What is the average water velocity in the 1. The general solution is S¼ 2t(200 þ t) C þ 100 þ t 100 þ t The constant of integration may be evaluated.

82 6. 4. 0. diameter were measured as follows: Distance from center (in.2 lb/min.16 7. r1.48 6.45 6.11 A shock wave moves down a pipe as shown below. The ﬂuid properties change across the shock. ﬂows out at a ﬁxed rate of 19..) Velocity (fps) 0 7.40 Chapter 4 Conservation of Mass: Control-Volume Approach 4. The solution.5 Water enters one end of a perforated pipe 0.40 7. and v w. the large piston has a velocity of 2 fps and an acceleration of r1 v1 = 0 vx = v0 6d d v0 vx = v0 v0 . the x-direction velocity proﬁles are shown for a control volume surrounding a cylinder. Hint: Use a control volume that is moving to the right at velocity. 5 fps2.5 7.10 4. Write the continuity equation and obtain the relation between r2. 4.92 lb/gal of salt ﬂows at a ﬁxed rate of 2 gal/min into a 100-gal tank. The mass in the control volume at any time is M ¼ r2 Ax þ r1 Ay. what must the rate of ﬂow be across the horizontal control-volume surface? x How many pounds of salt will there be in the tank at the end of 1 h and 40 min? (b) What is the upper limit for the number of pounds of salt in the tank if the process continues indeﬁnitely? (c) How much time will elapse while the quantity of salt in the tank changes from 100 to 150 lb? (a) 4.37 5. Velocity (fps) Distance from center (in.. show that equation (4-6) may be written ZZ @M þ d m_ ¼ 0 @t c:s: 4.81 .75 5.7 Salt water containing 1.) d = 0. The discharge through the pipe wall is approximated by a linear proﬁle.2 m in diameter with a velocity of 6 m/s.33 6. The density of the incoming solution is 71.5 m v Oil 4.10 8. (b) the ﬂow rate in cubic feet per second. Find the average velocity in the pipe in terms of v max.. .47 3. If the ﬂow is incompressible.9 Show that in a one-dimensional steady ﬂow the following equation is valid: 6 m/s v dA dv dr þ þ ¼0 A v r 2 4.13 In the ﬁgure below. v w.42 9.00 2. y 4. 4. Determine the velocity and acceleration of the smaller piston. initially ﬁlled with fresh water.8 lb/ft3.94 4.8 In the piston and cylinder arrangement shown below.50 5..5 in.49 3.10 Using the symbol M for the mass in the control volume.12 The velocity proﬁle in circular pipe is given by v ¼ v max (1 r=R)1=7 .6 The velocities in a circular duct of 20-in. 2 fp s 5 fp s2 v 2 d = 4 in. where R is the radius of the pipe.75 8. Shock vw r2 v2 x Find (a) the average velocity. If the ﬂow is steady. v 2. but they are not functions of time. The velocity of the shock is v w. kept uniform by stirring.07 5. ﬁnd the discharge velocity.15 10.

14 Two very long parallel plates of length 2L are separated a distance b. Suppose that h ¼ 2 cm and the ﬂow rate is 2 L/min. Estimate v 0.0013 m3/s. The check valve on the left (pleated) end is closed 4. Derive expressions for (a) the rate of change dh/dt and (b) the time required for the surface to rise from h1 to h2. Determine the mass ﬂow rate and maximum velocity: 41 during the stroke. Determine the average velocity of the ﬂuid escaping between the rod and the tank walls (a) relative to the bottom of the tank and (b) relative to the advancing rod. has a velocity proﬁle v x v 0 (2y=h y2 =h2 ).6 cm pipe and entrains a secondary ﬂow of water V2 ¼ 3 m/s in the annular region around the small pipe. derive an expression for outlet mass ﬂow m_ 0 as a function of stroke u(t). (a) If the exit velocity is uniform.19 The jet pump injects water at V1 ¼ 40 m/s through a 7. determine the volume rate of ﬂow in the ﬁlm. A portion of the ﬂow is diverted through the showerhead. A ﬂuid ﬁlls the space between the plates. where V3 is approximately constant. For steady incompressible ﬂow.16 The V-shaped tank has width w into the paper and is ﬁlled from the inlet pipe at volume ﬂow Q. estimate the exit velocity from the showerhead jets.1 m/s. A cylindrical rod of diameter di (less than tank diameter.17 A bellows may be modeled as a deforming wedge-shaped volume.18 Water ﬂows steadily through the piping junction. If w is the bellows width into the paper. (b) If the exit velocity is parabolic. compute V3. If the plane has width 10 cm into the paper. 2 cm y 1 1. x 7. Assuming uniform shower ﬂow. Fluid is squeezed out between the plates. where v 0 is the surface velocity.5 cm vx h 2 4 cm 4. The average velocity at section 2 is 2.20 A vertical.6 cm V2 4. which contains 100 holes of 1-mm diameter. draining from an inclined plane. entering section 1 at 0. V1 V3 25 cm h 20° 20° Q 4. The upper plate moves downward at a constant rate V. cylindrical tank closed at the bottom is partially ﬁlled with an incompressible liquid.15 A thin layer of liquid. The two ﬂows become fully mixed downstream.Problems 4. d0) is lowered into the liquid at a velocity V. V m0 q L x 2L 4. 4. .

how fast should the plunger be advanced (a) if leakage in the plunger clearance is neglected and (b) if leakage is 10% of the needle ﬂow? 4.21 The hypodermic needle shown below contains a liquid serum (r ¼ 1 g/cm3).14 if the plates are circular and have radius L. while downstream.24 Rework Problem 4. V0 ¼ 8 cm/s. with a uniform inlet proﬁle and a polynomial exit proﬁle 3h h3 y vx ¼ v0 where h ¼ d 2 Compute the volume ﬂow Q across the top surface of the control volume. .42 Chapter 4 Conservation of Mass: Control-Volume Approach 4. V 0. where a is a constant. as in the ﬁgure below.8 mm 2 cm 4. the ﬂow develops into the parabolic proﬁle v x ¼ az(z0 zÞ. If the serum is to be injected steadily at 6 cm3/s.22 Incompressible steady ﬂow in the inlet between parallel plates is uniform. What is the maximum value of v x? y Q v0 v0 d z x V0 z0 x 4. The plate has width b into the paper.23 An incompressible ﬂuid ﬂows past a ﬂat plate.

it will be necessary to recast its statement into a form applicable to control volume which contains different ﬂuid particles (i.1. a different system) when examined at different times. Applications of these expressions to physical situations will be considered. In Figure 5. we see that Region I is occupied by the system only at time t. 43 . Region II is occupied by the system at t þ Dt. The system considered is the material occupying the control volume at time t. observe the control volume located in a ﬂuid-ﬂow ﬁeld. We note at the outset two very important parts of this statement: ﬁrst. and second.. Referring to the ﬁgure. this law refers to a speciﬁc system. In order to use this law.e. Boundary of system at time t Streamlines at time t I III II Boundary of system at time t + ∆t Stationary control volume Figure 5. we shall develop integral relations for linear and angular momentum.Chapter 5 Newton’s Second Law of Motion: Control-Volume Approach The second of the fundamental physical laws upon which ﬂuid-ﬂow analyses are based is Newton’s second law of motion. Region III is common to the system both at t and at t þ Dt. and its position is shown both at time t and at time t þ Dt. 5. it includes direction as well as magnitude and is therefore a vector expression.1 Relation between a system and a control volume in a ﬂuid-ﬂow ﬁeld. Starting with Newton’s second law.1 INTEGRAL RELATION FOR LINEAR MOMENTUM Newton’s second law of motion may be stated as follows: The time rate of change of momentum of a system is equal to the net force acting on the system and takes place in the direction of the net force.

> : . we may write the following word equation for the conservation of linear momentum with respect to a control volume: 9 8 9 > 9 8 9 8 8 rate of > > > sum of > > > > > rate of > > rate of > > > > > = < accumulation > = > < = > < = > < momentum momentum forces acting þ of momentum ¼ on control > > > > into control > > out of control > > > > > > within control > > > . : volume |ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ} net rate of momentum efflux from control volume (5-3) . as Dt!0. the linear momentum of the system that now occupies regions II and III may be expressed as PjtþDt ¼ PII jtþDt þ PIII jtþDt and at time t we have Pjt ¼ PI jt þ PIII jt Subtracting the second of these expressions from the ﬁrst and dividing by the time interval Dt gives PjtþDt Pjt PII jtþDt þ PIII jtþDt PI jt PIII jt ¼ Dt Dt We may rearrange the right-hand side of this expression and take the limit of the resulting equation to get lim Dt!0 PjtþDt Pjt PIII jtþDt PIII jt PII jtþDt PI jt ¼ lim þ lim Dt Dt Dt Dt!0 Dt!0 (5-2) Considering each of the limiting processes separately. At time t þ Dt. we have d d åF ¼ (mv) ¼ P dt dt (5-1) where the symbols F. The next limiting process PII jtþDt PI jt lim Dt Dt!0 expresses the net rate of momentum efﬂux across the control surface during the time interval Dt. m. and v have their usual meanings and P represents the total linear momentum of the system. The ﬁrst limit on the right-hand side of equation (5-2) may be evaluated as lim Dt!0 PIII jtþDt PIII jt d ¼ PIII dt Dt This we see to be the rate of change of linear momentum of the control volume itself. region III becomes the control volume. equation (5-1). for the left-hand side lim Dt!0 PjtþDt Pjt d ¼ P dt Dt which is the form speciﬁed in the statement of Newton’s second law. regions I and II become coincident with the control-volume surface. As Dt approaches zero. > : . we have. equation (5-1). Considering the physical meaning of each of the limits in equation (5-2) and Newton’s second law.44 Chapter 5 Newton’s Second Law of Motion: Control-Volume Approach Writing Newton’s second law for such a situation. > : > > volume volume volume . since.

we have ZZ vr(v: n) dA c:s: which is the net momentum efﬂux from the control volume. dA projected in a direction normal to the velocity vector. v: n and the momentum efﬂux term becomes rv(v: n) dA Integrating this quantity over the entire control surface.2 and evaluate the various terms. v. In its integral form the momentum ﬂux term stated above includes the rate of momentum entering the control volume as well as that leaving. where u is the angle between v and the outwardly directed normal vector. n. We will designate the total force acting on the control volume as åF. and of body forces resulting from the location of the control volume in a force ﬁeld.5. c:s: volume volume . Conversely. we may write rate of momentum efflux ¼ v(rv)(dA cos u) Observe that the product (rv)(dA cos u) is the rate of mass efﬂux from the control volume through dA. The gravitational ﬁeld and its resultant force are the most common examples of this latter type. If the small area dA on the control surface is considered. as discussed in Chapter 4. and its surroundings through direct contact. a positive sign of the product v: n is associated with a momentum efﬂux from the control volume. If mass is entering the control volume. Recall further that dA cos u is the area. : . Thus. From vector algebra this product may be written as v(rv)(dA cos u) ¼ v(r dA)½jvj jnj cos u The term in square brackets is the scalar or dot product. and the associated momentum ﬂux is an input. the sign of the product v: n is negative. The total force acting on the control volume consists both of surface forces due to interactions between the control-volume ﬂuid. the ﬁrst two terms on the right-hand side of equation (5-3) may be written 8 9 8 9 < rate of momentum = < rate of momentum = Z Z vr(v: n) dA out of control into control ¼ : . We may then multiply the rate of mass efﬂux by v to give the rate of momentum efﬂux through dA.1 Integral Relation for Linear Momentum 45 Streamlines at time t dA q v n Figure 5.2 Fluid ﬂow through a control volume. We shall now apply equation (5-3) to a general control volume located in a fluid-flow field as shown in Figure 5.

The control volume chosen in this manner is designated in Figure 5. EXAMPLE 1 Consider ﬁrst the problem of ﬁnding the force exerted on a reducing pipe bend resulting from a steady ﬂow of ﬂuid in it. that equation (5-4) is a vector expression opposed to the scalar form of the overall mass balance considered in Chapter 4. observe. showing the external forces imposed upon it. (5-4). it must be remembered that each term has a sign with respect to the positively deﬁned x. but experience in handling problems of this type will enable such a choice to be made readily.46 Chapter 5 Newton’s Second Law of Motion: Control-Volume Approach The rate of accumulation of linear momentum within the control volume may be expressed as ZZZ @ vr dV @t c:v: and the overall linear-momentum balance for a control volume becomes ZZ ZZZ @ vr(v: n) dA þ rv dV åF ¼ @t c:s: c:v: (5-4) This extremely important relation is often referred to in ﬂuid mechanics as the momentum theorem. There are no general rules to aid in this deﬁnition. and z directions. 5. It should also be remembered that as equations (5-5a–c) are written for the ﬂuid in the control volume. the overall momentum balance. the forces to be employed in these equations are those acting on the ﬂuid. The combination of the proper sign associated with each of these terms will give the correct sense to the integral. it is ﬁrst necessary to deﬁne the control volume that will make possible the simplest and most direct solution to the problem at hand. The external . as both the velocity component (v x) and the scalar product (v: n) have signs.2 APPLICATIONS OF THE INTEGRAL EXPRESSION FOR LINEAR MOMENTUM In applying equation (5-4). of the several available. In rectangular coordinates the single-vector equation.4. Note the great similarity between (5-4) and (4-1) in the form of the integral terms.3. One choice for the control volume. is all ﬂuid in the pipe at a given time. may be written as three scalar equations ZZZ ZZ @ åFx ¼ v x r(v: n) dA þ rv x dV (5-5a) @t c:s: c:v: ZZZ ZZ @ v y r(v: n) dA þ rv y dV (5-5b) åFy ¼ @t c:s: c:v: ZZZ ZZ @ v z r(v: n) dA þ rv z dV (5-5c) åFz ¼ @t c:s: c:v: When applying any or all of the above equations. however. A detailed study of the example problems to follow should aid in the understanding of. A diagram of the pipe bend and the quantities signiﬁcant to its analysis are shown in Figure 5. and afford facility in using. The ﬁrst step is the deﬁnition of the control volume. The determination of the sign of the surface integral should be considered with special care. y.

The resultant force on the ﬂuid (due to Pw and tw) by the pipe is symbolized as B. The actual signs for these components. and its x and y components as Bx and By . the body force due to the weight of ﬂuid in the control volume. Evaluating the surface integral in both the x and y directions. for the problem considered. forces imposed on the ﬂuid include the pressure forces at sections (1) and (2). ﬂow is steady. as. Pw and tw.2 Applications of the Integral Expression for Linear Momentum 47 1 v1 W 2 y v2 x q Figure 5. we have ZZ v x r(v: n) dA ¼ (v 2 cos u)(r2 v 2 A2 ) þ (v 1 )(r1 v 1 A1 ) ZZ c:s: c:s: v y r(v: n) dA ¼ (v 2 sin u)( p2 v 2 A2 ) The accumulation term is zero in both equations. of the overall momentum balance. will indicate whether or not this assumption is correct. and the forces due to pressure and shear stress.3 Flow in a reducing pipe bend.5. Pw τw 1 P1 W 2 P2 Pw τw q Figure 5. respectively.4 Control volume deﬁned by pipe surface. the external forces acting on the ﬂuid in the control volume are åFx ¼ P1 A1 P2 A2 cos u þ Bx and åFy ¼ P2 A2 sin u W þ By Each component of the unknown force B is assumed to have a positive sense. (5-5a) and (5-5b). . exerted on the ﬂuid by the pipe wall. Considering the x.and y-directional component equations. when a solution is obtained.

. This control volume is bounded simply by the straight planes cutting through the pipe at sections (1) and (2). we have Bx ¼ v 22 r2 A2 cos u v 21 r1 A1 P1 A1 þ P2 A2 cos u and By ¼ v 22 r2 A2 sin u P2 A2 sin u þ W Recall that we were to evaluate the force exerted on the pipe rather than that on the ﬂuid. The force sought is the reaction to B and has components equal in magnitude and opposite in sense to Bx and By. exerted on the pipe are Rx ¼ v 22 r2 A2 cos u þ v 21 r1 A1 þ P1 A1 P2 A2 cos u and Ry ¼ v 22 r2 A2 sin u þ P2 A2 sin u W Some simpliﬁcation in form may be achieved if the ﬂow is steady. The components of the reaction force. The ﬁnal solution for the components of R may now be written as _ 1 v 2 cos u) þ P1 A1 P2 A2 cos u Rx ¼ m(v Ry ¼ mv _ 2 sin u þ P2 A2 sin u W The control volume shown in Figure 5.4 for which the above solution was obtained represents only possible choice. that the results of complicated processes occurring internally may be analyzed quite simply by considering only those quantities of transfer across the control surface.5. Another is depicted in Figure 5. 1 By Bx P1 W 2 P2 Figure 5. R.5 Control volume including ﬂuid and pipe. Applying equation (4-3).48 Chapter 5 Newton’s Second Law of Motion: Control-Volume Approach The complete momentum expressions in the x and y directions are Bx þ P1 A1 P2 A2 cos u ¼ (v 2 cos u)(r2 v 2 A2 ) þ v 1 (r1 v 1 A1 ) and By þ P2 A2 sin u W ¼ ðv 2 sin u)(r2 v 2 A2 ) Solving for the unknown force components Bx and By. The fact that a control volume such as this can be used indicates the versatility of this approach. we have r1 v 1 A1 ¼ r2 v 2 A2 ¼ m_ where m_ is the mass ﬂow rate. that is.

a correct solution may be obtained for each of several chosen control volumes so long as they are analyzed carefully and completely. h The logical choice for a control volume in this case is the water-tank/scoop combination. . hence Newton’s second law and the momentum theorem may be employed directly. the x. v0 Figure 5.2 Applications of the Integral Expression for Linear Momentum 49 For this control volume.and y-directional momentum equations are Bx þ P1 A1 P2 A2 cos u ¼ (v 2 cos u)(r2 v 2 A2 ) þ v 1 (v 1 r1 A1 ) and By þ P2 A2 sin u W ¼ (v 2 sin u)(r2 v 2 A2 ) where the force having components Bx and By is that exerted on the control volume by the section of pipe cut through at sections (1) and (2). As the train is moving with a uniform velocity. Let us ﬁrst analyze the system by using a moving coordinate system. there are two possible choices of coordinate systems. Our control-volume boundary will be selected as the interior of the tank and scoop. The pressures at (1) and (2) in the above equations are gage pressures. EXAMPLE 2 As our second example of the application of the control-volume expression for linear momentum (the momentum theorem). v 0.7 with the xy coordinate system moving at velocity v0.6 Schematic of locomotive tender scooping water from a trough. The moving control volume is shown in Figure 5. Thus. The force on the train due to the water is to be obtained.6.5. All velocities are determined with respect to the x and y axes. Note that the resulting equations for this control volume are identical to those obtained for the one deﬁned previously. Recall that a uniformly translating coordinate system is an inertial coordinate system. which obtains water from a trough by means of a scoop. We may select a coordinate system either ﬁxed in space or moving1 with the velocity of the train. y Fx h 1 v0 x Figure 5.7 Moving coordinate system and control volume. as the atmospheric pressures acting on all surfaces cancel. consider the steam locomotive tender schematically illustrated in Figure 5.

v x ¼ v 0 ¼ constant. as the entering RRR ﬂuid has zero velocity. Fx is the total force exerted on the ﬂuid by the train and scoop. The momentum ﬂux term is ZZ v x r (v: n) dA ¼ r(v 0 )(1)(v 0 )ðhÞ (per unit length) c:s: and the rate of change of momentum within the control volume is zero. where m_ is the mass of ﬂuid entering the control volume 0 c:v: at the rate @m=@t ¼ rv 0 h so that Fx ¼ rv 20 h as before. care must be exercised in the interpretation of the momentum ﬂux ZZ vr(v: n) dA c:s: Regrouping the terms.50 Chapter 5 Newton’s Second Law of Motion: Control-Volume Approach The applicable expression is equation (5-5a) ZZ åFx ¼ c:s: v x r(v: n) dA þ @ @t ZZZ v x r dV c:v: In Figure 5. The force exerted by the ﬂuid on the train is the opposite of this. x The student should note that. it is obvious that while v is the velocity relative to ﬁxed coordinates. Fx ¼ rv 20 h This is the force exerted by the train on the ﬂuid. y v0 Fx h Figure 5. as the velocity. As the forces due to pressure and shear are to be neglected. or rv 20 h. åFx is represented as Fx and is shown in the positive sense. may be written as v 0 @=@t r dV or v (@m=@t). . Thus. no ﬂuid leaving the controlRRR volume.8 Stationary coordinate system and moving control volume. There is.7. we obtain ZZ c:s: vrðv: nÞ dA ZZ v d m_ c:s: Thus. v: n is the velocity relative to the control-volume boundary.8). The terms @=@t c:v: v x r dV. Now let us consider the same problem with a stationary coordinate system (see Figure 5. Employing once again the control-volume relation for linear momentum ZZZ ZZ @ åFx ¼ v x r(v: n) dA þ v x r dV @t c:s: c:v: we obtain Fx ¼ 0 þ @ @t ZZZ v x r dV c:v: where the momentum ﬂux is zero. of course. as the ﬂuid in the control volume has zero velocity in the x direction. in the case of a stationary coordinate system and a moving control volume.

for parts (a) and (b) of this example.9. and the governing equation is F ¼ rA j v 2j We may now introduce the appropriate numerical values and solve for F. The x-directional component form of the momemtum theorem will yield the expression F ¼ r A j (v j v 0 )2 Substitution of appropriate numerical values yields F ¼ (1000 kg/m3 )(0:005 m2 )(12 4 m/sÞ2 ¼ 320 N .9 A ﬂuid jet striking a vertical plate. Plate Aj = 0.2 EXAMPLE 3 51 Applications of the Integral Expression for Linear Momentum A jet of ﬂuid exits a nozzle and strikes a vertical plane surface as shown in Figure 5. The control volume to be used in this analysis is shown Figure 5. The coordinates are ﬁxed with the control volume which. . we have ZZ åFx ¼ c:s: v x r(v: n) dA þ @ @t ZZZ 2 C. is stationary. Writing the x-directional scalar form of the momentum theroem. from the perspective of an observer moving the control volume. (a) Determine the force required to hole the plate stationary if the jet is composed of i. In this case.V. For case (a) (i) rw ¼ 1000 kg/m3 F ¼ (1000 kg/m3 )(0:005 m2 )(12 m/s)2 ¼ 720 N (ii) rw ¼ 1:206 kg/m3 F ¼ (1:206 kg/m3 )(0:005 m2 )(12 m/s)2 ¼ 0:868 N For case (b).5. y v x r dv 1 c:v: x F Evaluation of each term in this expression yields åFx ¼ F ZZ c:s: @ @t v x r(v: n) dA ¼ v j r(v j A j ) ZZZ 3 v x r dv ¼ 0 c:v: Figure 5.10 Control volume for Example 3. the control volume and the coordinate system are moving to the right at a velocity of 4 m/s.10. Figure 5. the same control volume will be used. the velocity of the incoming water jet is (v j v 0 ) ¼ 8 m/s.005 m2 F Vj = 12 m/s (b) Determine the magnitude of the rest the restraining force for a water jet when the plate is moving to the right with a uniform velocity of 4 m/s. air. water ii. however.

the total moment about the origin of all forces acting on the system.11. The complete expression is now r åM ¼ d H dt (5-7) As with its analogous expression for linear momentum.11) åF ¼ d d (mv) ¼ P dt dt (5-1) y mv ΣF r x Figure 5. Clearly. this term is also the time rate of change of the moment of momentum of the system. By the same limit process as that used for linear momentum. We shall use the symbol H to designate moment of momentum. due to all forces applied to the system. Starting with equation (5-1). This we can write as d d d d mv ¼ (r mv) ¼ (r P) ¼ H dt dt dt dt Thus. r SF. r. again. åM. is the resultant moment.52 Chapter 5 Newton’s Second Law of Motion: Control-Volume Approach 5. z we take the vector or ‘‘cross’’ product of a position vector. equation (5-1).11 A system and its displacement vector r. about the origin as shown in Figure 5. which is a mathematical expression of Newton’s second law of motion applied to a system of particles (Figure 5.3 INTEGRAL RELATION FOR MOMENT OF MOMENTUM The integral relation for the moment of momentum of a control volume is an extension of the considerations just made for linear momentum. we may write r åF ¼ år F ¼ åM where åM is. . with each term and get r åF ¼ r d d (mv) ¼ r P dt dt (5-6) The quantity on the left-hand side of equation (5-6). equation (5-7) applies to a speciﬁc system. The right-hand side of equation (5-6) is the moment of the time rate of change of linear momentum.

4 Applications to Pumps and Turbines 53 we may recast this expression into a form applicable to a control volume and achieve a word equation 9 8 rate of > > 9 9 8 8 > > 9 > 8 > > accumulation > rate of sum of > > > > > > > > > > > > rate of > > > > > > > > > > = < of moment > = > < = > < moment of > = > < moments > moment of þ of momen momentum acting on ¼ momentum into > > > > > > > > > > > > tum within > control > > > .5. in this section. The following two examples illustrate how moment-of-momentum analysis is used to generate expressions for evaluating turbine performance. Similar approaches will be used in Chapter 14 to evaluate operating characteristics of fans and pumps. it is designated a turbine. The terms on the right-hand side represent the net rate of efﬂux of moment of momentum through the control surface and the rate of accumulation of moment of momentum within the control volume. generally classiﬁed as pumps and turbines.4 APPLICATIONS TO PUMPS AND TURBINES The moment-of-momentum expression is particularly applicable to two types of devices. : |ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ} volume net rate of efflux of moment of momentum from control volume (5-8) Equation (5-8) may be applied to a general control volume to yield the following equation: ZZ ZZZ @ åM ¼ (r v)r(v: n) dA þ (r v)r dV (5-9) @t c:s: c:v: The term on the left-hand side of equation (5-9) is the total moment of all forces acting on the control volume. : . If energy is derived from a ﬂuid acting on a rotating device. 5. whereas a pump adds energy to a ﬂuid. . > : > > > > control volume volume > > . consider those having rotary motion only. and z as ZZ åMx ¼ ZZ c:s: åMy ¼ c:s: (r v)x r(v: n) dA þ @ @t (r v)y r(v: n) dA þ @ @t ZZZ ZZZ (r v)x r dV (5-10a) (r v)y r dV (5-10b) (r v)z r dV (5-10c) c:v: c:v: and ZZ åMz ¼ c:s: @ (r v)z r(v: n) dA þ @t ZZZ c:v: The directions associated with Mx and (r v) are those considered in mechanics in which the right-hand rule is used to determine the orientation of quantities having rotational sense. > : > > out of control > > > > > > > > control volume > > . We shall. This single-vector equation may be expressed as three scalar equations for the orthogonal inertial coordinate directions x. The rotating part of a turbine is called a runner and that of a pump an impeller. y. respectively.

w v0 w×r Bottom view of bucket Figure 5.12 Pelton wheel. a jet of ﬂuid. All rotation is in the xy plane. the z direction. usually water. for the external moment åMz ¼ Mshaft . y Pat x Figure 5. in turn. has a sense normal to the xy plane. we have.54 Chapter 5 EXAMPLE 4 Newton’s Second Law of Motion: Control-Volume Approach Let us ﬁrst direct our attention to a type of turbine known as the Pelton wheel. It encloses the entire runner and cuts the jet of water with velocity v0 as shown. In this turbine. The dashed lines in Figure 5. or a tendency to produce angular motion. ZZ åMz ¼ c:s: (r v)z r(v: n) dA þ @ @t ZZZ r (r v)z dV c:v: Evaluating each term separately. is directed from a nozzle striking a system of buckets on the periphery of the runner. The control surface also cuts through the shaft on both sides of the runner. cause rotation.13 Control volume for analysis of Pelton wheel. Such a device is represented in Figure 5.12. The applicable scalar form of the general moment-of-momentum expression is equation (5-10c) written for the z direction. and—according to the right-hand rule—the vector representation of a quantity having angular motion. The buckets are shaped so that the water is diverted in such a way as to exert a force on the runner which will. Recall that a positive angular sense is that conforming to the direction in which the thumb on the right hand will point when the ﬁngers of the right hand are aligned with the direction of counterclockwise angular motion. v0 We must initially deﬁne our control volume. that is. we may determine the torque resulting from such a situation. Using the moment-of-momentum relation.13 illustrates the control volume chosen.

4 Applications to Pumps and Turbines 55 where Mshaft. and that of the exiting ﬂuid relative to the bucket and leaving at an angle u to the direction of motion of the bucket.14 Velocity vectors for turbine bucket. Thus our ﬁnal result is Torque ¼ Mshaft ¼ r(v 0 rv)(1 þ cos u)rQ . we have ZZZ ZZ @ åMz ¼ Mshaft ¼ (r v)z r (v: n) dA þ r(r v)z dV @t c:s: c:v: ¼ r½rv (v 0 rv) cos urQ rv 0 rQ ¼ r(v 0 rv)(1 þ cos u)rQ The torque applied to the shaft is equal in magnitude and opposite in sense to Mshaft.14. Replacing each term in the complete expression by its equivalent. rv. The ﬁnal expression for the surface integral is now ZZ c:s: (r v)z r(v: n) dA ¼ r½rv (v 0 rv) cos urQ rv 0 rQ The last term. The leaving velocity is the vector sum of the velocity of the turbine bucket. of the wheel is constant. the moment applied to the runner by the shaft. The x-direction component of the ﬂuid leaving the control volume is frv (v 0 rv) cos ugex Turbine bucket y Nozzle w×r v0 v0 – ⎜w × r ⎜ x q Figure 5. RRR the term expressing the time derivative of moment of momentum of the control volume. with a volumetric ﬂow rate Q. As the problem under consideration is one in which the angular velocity. These velocity vectors are shown in the ﬁgure. The ﬂuid leaving the control volume is illustrated in Figure 5. The surface integral ZZ (r v)z r(v: n) dA c:s: is the net rate of efﬂux of moment of momentum. rv 0rQ. (v 0 rv) cos u.5. is the only such moment acting on the control volume. v. is the moment of momentum of the incoming ﬂuid stream of velocity v 0 and density r. @/@t c:v: (r v)z r dV ¼ 0. Here it is assumed that the z components of the velocity are equal and opposite.

The outer boundary of the control volume is at radius r1. The control volume to be used is illustrated below in Figure 5. The width of the runner is h.16 Radial-ﬂow turbine-runner control volume. where T is the shaft torque. We will use equation (5-9) in order to determine the torque. we have. If we express the velocity of the water in polar coordinates v ¼ v r er þ v u eu . so that (r v) ¼ rer (v r er þ v u eu ) ¼ rv u ez .15 Radial-ﬂow turbine.56 Chapter 5 EXAMPLE 5 Newton’s Second Law of Motion: Control-Volume Approach The radial-ﬂow turbine illustrated in Figure 5. is given by ZZ (r v)r(v: n) dA ¼ fr1 v u1 r(v r1 ) 2pr1 h þ r2 v u2 rv r2 2pr2 hgez c:s: The general result is . For steady ﬂow this equation becomes ZZ åM ¼ c:s: Figure 5. w h a r1 v1 r2 (r v)r(v: n) dA y Evaluating each term separately.16. In this device. which impart a tangential velocity and hence angular momentum to the ﬂuid before it enters the revolving runner which reduces the angular momentum of the ﬂuid while delivering torque to the runner. the ﬂuid (usually water) enters the guide vanes. Thus the surface integral. The surface integral requires the evaluation of the vector product (r v) at the outer boundary r1 and at the inner boundary r2. and the inner boundary is at r2.15 may be analyzed with the aid of the moment-of-momentum expression. for the external moment of the runner on the ﬂuid åM ¼ Mfluid ez ¼ Tez x Figure 5. assuming uniform velocity distribution.

however. The velocity of the water v 2 is the vector sum of the velocity with respect to the runner v 20 and the runner velocity r 2 v. requires knowledge of ﬂow conditions on the runner. The velocity at r 2 may be determined by the following analysis.15 and 5.17. Tez ¼ rv r1 v u1 2pr12 h þ rv r2 v u2 2pr22 h ez The law of conservation of mass may be used rv r1 2pr1 h ¼ m_ ¼ rv r2 2pr2 h so that the torque is given by T ¼ m(r _ 1 v u 1 r2 v u 2 ) The velocity at r1 is seen from Figures 5. . In Figure 5. the ﬂow conditions at the outlet of the runner are sketched. The velocity at r2.16 to be determined by the ﬂow rate and the guide vane angle a.

whereas in a hasty consideration certain terms might be evaluated incorrectly or neglected completely. the basic relation involved has been Newton’s second law of motion. ﬁnd the pressure difference between inlet and outlet sections. The radial component of the ﬂow may be determined from conservation of mass m_ v r2 ¼ v 02 cos b ¼ 2prr2 h Thus. the total drag on the object is measured to be 800 N/m of length normal to the direction of ﬂow.1 if the exit velocity proﬁle is given by py v ¼ v 2 1 cos 4 when y is measure vertically from the center line of the water tunnel. as developed. v 1 . in the system for Problem 5. applies to an inertial control volume only.0805 lbm/ft3 enters as shown. The ﬂuid is assumed to ﬂow in the same direction as the blade. and moment of momentum. is given by 57 w v u2 ¼ r2 v v 00 sinb where b is the blade angle as shown. The student is again urged to start always with the complete integral expression when working a problem. it should be noted that the momentum theorem expression. was recast so that it could apply to a control volume.3 2 ft 4 ft Rework Problem 5. v1 = 20 fps v2 5. 5. 5. Equation (5-4) is often referred to as the momentum theorem of ﬂuid mechanics.5 CLOSURE In this chapter. and frictional forces at the walls are neglected. the tangential velocity of the water leaving the runner. This law. is uniform across the cross section. equation (5-9). PROBLEMS 5. ﬁnd the value of v 2 . m_ tan b T ¼ m_ r1 v u1 r2 r2 v 2prr2 h v2 r2w v2′ b Figure 5. as written for a system. 5.4 A stationary jet engine is shown. Air with a density of 0.Problems The velocity v u2 .1 A two-dimensional object is placed in a 4-ft-wide water tunnel as shown. For the downstream velocity proﬁle as shown. A term-by-term analysis from this basis will allow a correct solution. equation (5-4). This equation is one of the most powerful and often-used expressions in this ﬁeld. The inlet and outlet . As a ﬁnal remark. The result of a consideration of a general control volume led to the integral equations for linear momentum. In practice.17 Velocity at runner exit (only one blade is shown).2 If.1. The upstream velocity. the guide vanes are adjustable to make the relative velocity at the runner entrance tangent to the blades.

6 The pump in the boat shown pumps 6 ft3/s of water through a submerged water passage. which entrains a secondary stream of water having a velocity v s ¼ 10 fps in a constant-area pipe of total area A ¼ 0. P = 5 psig D = 2. ﬁnd the magnitude and velocity of the water jet leaving the blade. (a) 25 fps 30° v2 vj 5. and the rate of discharge is 400 gal/min. and the breadths a.10 If the plate shown is inclined at an angle as shown. ¼ 0.7 Oil (sp.) 2 V h P = 50 psig D = 12 in. What are the magnitude and direction of the force necessary to hold the nozzle to the pipe? a l 0 b a . what are the forces Fx and Fy necessary to maintain its position? The ﬂow is frictionless. For these conditions. v = 100 fps 3 Fx • m = 2 slugs/s Fy 30° 5. velocity v.58 Chapter 5 Newton’s Second Law of Motion: Control-Volume Approach cross-sectional areas are both 10. 5. Assuming one-dimensional ﬂow and neglecting wall shear (a) ﬁnd the average velocity of mixed ﬂow at section 2.p. 4 5.5 Determine the magnitude of the x and y components of the force exerted on the ﬁxed blade shown by a 3-ft3/s jet of water ﬂowing at 25 fps. b. frictionless. (b) Determine the distance 1 to the center of pressure (c. which has an area of 0.) along the plate from the point 0. gr. (b) If the blade is moving to the right at 15 fps.15 ft2 at the stern. 1 Aj Fuel 2 vs 5.11 A steady.5 in. Pump 5.8 At the end of a water pipe of 3-in. of the two branches. (b) ﬁnd the pressure rise ðP2 P1 Þ. 1 (a) Determine the total forceon the plate.6 ft2.8) ﬂows smoothly through the circular reducing section shown at 3 ft3/s.8 ft2. v1 v2 300 fps 900 fps 5. The pressure in the pipe is 60 psig (pounds per square inch gage). estimate the force which must be applied to the reducer to hold it in place. the water is thoroughly mixed. (The center of pressure is the point at which the plate can be balanced without requiring an additional moment. diameter is a nozzle that discharges a jet having a diameter of 1½ in.9 A water jet pump has an area A j ¼ 0:06 ft2 and a jet velocity v j ¼ 90 fps. assuming the pressure of the jet and secondary stream to be the same at section 1. Determine the tension in the restraining rope. two-dimensional jet of ﬂuid with breadth h. If the entering and leaving velocity proﬁles are uniform. incompressible. and unit width impringes on a ﬂat plate held at an angle a to its axis. At section 2. Gravitational forces are to be neglected. The mass of fuel consumed is 2% of the mass of air entering the test section. assuming that the inlet and exit pressures are equal.25 ft2 at the bow of the boat and 0. calculate the thurst developed by the engine tested. into the open atmosphere.

5. 2. The properties in front and in back of the shock are not a function of time. Assume frictionless ﬂow. the diameter is 0. A1 5. the diameter is 0. 5. deterine the pressure change causing a velocity change of 10 fps in v D P (a) air at standard conditions.13 is approximated by the speed of sound. 26 psia 5. the velocity is 12 m/s. vx = v0 2 ∆s 1 P2. 1 x 2 3 Thickness = 3/8 in. Determine the forces Fx and Fz necessary to hold the pipe bend stationary.18 The pressure on the control volume illustrated below is constant.15 Consider the differential control volume shown below.17 The rocket nozzle shown below consists of three welded sections. Welds y 5. 990 psia 530 psia 6700 fps 24 in. At station 5. 5. r1 v0 vx = v0 v0 v1. Determine the pressure rise in the pipe. Determine the force exerted on the cylinder by the ﬂuid.20 A dam discharge into a channel of constant width as shown. The velocity and height of the ﬂow in the channel are given as v and h.3 m. and the pressure is 128 kPa gage.12 A plate moves perpendicularly toward a discharging jet at the rate of 5 fps. By using the illustrated control volume.Problems 59 5.38 m and the pressure is 145 kPa gage. The mass ﬂow rate is 770 lbm/s. show that dP þ rv dv þ g dy ¼ 0 900 fps 3400 fps 18 in.14 If the shock-wave velocity in Problem 5.19 Water ﬂows in a pipe at 3 m/s. respectively. A valve at the end of the pipe is suddenly closed.16 Water ﬂows steadily through the horizontal 308 pipe bend shown below. (b) water. r2 v2. Assume incompressible ﬂow. Find the force of the ﬂuid on the plate and compare it with what it would be if the plate were stationary. show that the pressure difference across the shock is x 30° z 2 P2 P1 ¼ r1 v w v 2 1 P2 r2 v2 vw P1 r1 v1 = 0 5. The jet discharges water at the rate of 3 ft3/s and a speed of 30 fps. It is observed that a region of still water backs up behind the jet to a height H.13 The shock wave illustrated below is moving to the right at v w fps. By applying the conservation of mass and the momentum theorem. Determine the axial stress at junctions 1 and 2 when the rocket is operating at sea level. The x components of velocity are as illustrated. At station 1. A2 6d d ∆y P1. and the density of . 12 in.

collects ﬂuids from an overhead sprinkler which directs ﬂuid downward with velocity v s. Assuming that the sprinkler ﬂow is uniform over the car area. At the instant shown the car passes under a jet of water issuing from a stationary 0.60 Chapter 5 Newton’s Second Law of Motion: Control-Volume Approach the water is r.24 An open tank car as shown travels to the right at a uniform velocity of 4. If v 1 ¼ 5 m/s and the manometer reading Aj . necessary to hold the sluice gate in place. D2 ¼ 5 cm. P = Patm 4. low-energy condition (v 2 . The aire pressure in the cavity below the crest of falling water is to be taken as atmospheric. A jet of area Aj exhausts ﬂuid of density r at a velocity v j fps relative to the car.25 An open tank L ft long as shown below travels to the right at a velocity v c fps. v 1 ). a high-speed channel ﬂow (v 1 . (b) Determine the force per unit width.5 m/s R V1 h P = Patm V2 L 1 2 5. Using the momentum theroem and the control surface indicted. What force is exerted on the tank by the water jet? 45° (a) Determine the velocity at station 2. is h ¼ 58 cm. estimate the total force resisted by the ﬂange bolts. R. at the same time. determine H. vs V2 h1 vj L vc 5.23 For the pipe-ﬂow-reducing section D1 ¼ 8 cm.21 A liquid of density r ﬂows through a sluice gate as shown. Use the continuity and momentum relations to ﬁnd h2 and v 2 in terms of (h1. and p2 ¼ 1 atm. The upstream and downstream ﬂows are uniform and parallel.22 As can often be seen in a kitchen sink when the faucet is running. h1) may ‘‘jump’’ to a low-speed.5 m/s. The tank car. determine the net force of the ﬂuid on the tank car. The pressure at sections 1 and 2 is approximately hydrostatic. so that the pressure variations at stations 1 and 2 may be considered hydrostatic. and wall friction is negligible. h2 V1 5. Neglect the horizontal momentum of the ﬂow that is entering the control volume from above and assume friction to be negligible.1-m-diameter pipe with a velocity of 20 m/s. h2). 5. Ac. 1 2 Water V1 h Hg H v h 5.

V(t) h(t) 5. w v ¼ 1180 rpm t2 ¼ 0:6 in: u2 ¼ 135 r1 ¼ 2 in: r2 ¼ 8 in: t1 ¼ 0:8 in: vr R a t1 t2 q2 q1 r2 5. If the water 5. derive the differential equation for the downward motion v(t) of the liquid. diameter jets at the ends of a rotating hollow rod as shown.29 A water sprinkler consists of two 1/2-in.27 determine (a) the angle u1 such that the entering ﬂow is parallel to the vanes. and the effective discharge area is A.32 Water ﬂows at 30 gal/min through the 0.75-in. 61 leaves at 20 fps. and the pressure at the pump inlet is atmospheric. Using a control-volume analysis of mass and vertical momentum. Assume that the absolute velocity of the water entering the impeller is radial. Find the moment on the tube about the axis BB from the ﬂow of water inside the pipe system. The pressures are p1 ¼ 30 lbf/in. Determine the torque exerted on the impeller by the ﬂuid and the power required to drive the pump. (b) the axial load on the shaft if the shaft diameter is 1 in. Find an expression for the speed of the sprinkler in terms of the signiﬁcant variables. where vr is the velocity of the water relative to the rotating pipe. the stopper is suddenly removed. The dimensions are as follows: 5.2 and .26 A liquid column of height h is conﬁned in a vertical tube of cross-sectional area A by a stopper.31 The pipe shown below has a slit of thickness 1/4 in. The velocity is constant along the pipe as shown and a ﬂow rate of 8 ft3/s enters at the top.28 In Problem 5. 5. thus the water is discharged at a rate Q ¼ 2vyr A. what torque would be necessary to hold the sprinkler in place? 30° 12 in. issues out radially from the pipe. A constant friction torque Mf resists the motion of the sprinkler.27 Sea water. The sprinkler rotates with an angular velocity v. B r1 A = 4 in. exposing the bottom of the liquid to atmospheric pressure. r ¼ 64 lbm/ft3. At t ¼ 0.2 B 3 ft 6 ft w 5. Assume one-dimensional. incompressible.-diameter double-pipe bend.30 A lawn sprinkler consists of two sections of curved pipe rotating about a vertical axis as shown. so shaped that a sheet of water of uniform thickness 1/4 in. ﬂows through the impeller of a centrifugal pump at the rate of 800 gal/min. frictionless ﬂow.Problems 5.

Compute the torque T at point B necessary to keep the pipe from rotating. Show that the power transmitted to the cart is maximum when v c /v ¼ 1/3 (b) Assuming that the are a large number of such vanes attached to a rotating wheel with peripheral speed. The vane receives a jet that leaves a ﬁxed nozzle with speed v.33 The illustration below shows a vane with a turning angle u which moves with a steady speed v c . v c . B 1 (a) Assume that the vane is mounted on rails as shown in the sketch. v vc . 50° 3 ft q 2 5.62 Chapter 5 Newton’s Second Law of Motion: Control-Volume Approach p2 ¼ 24 lbf/in.2. show that the power transmitted is maximum when v c /v ¼ 1/2.

Keenan. p. is used with a ‘‘point’’ function. H. that is.17 ft lb/Btu in engineering units. the reader is referred to G. Hatsopoulos and J. and examples of the application of the integral expression will be shown.1 INTEGRAL RELATION FOR THE CONSERVATION OF ENERGY The ﬁrst law of thermodynamics may be stated as follows: If a system is carried through a cycle. and the student is reminded that all equations must be dimensionally homogeneous. In the SI system. recasting this statement into a form applicable to a control volume which contains different ﬂuid particles at different times.Chapter 6 Conservation of Energy: Control-Volume Approach T he third fundamental law to be applied to ﬂuid-ﬂow analyses is the ﬁrst law of thermodynamics. 1 For a more complete discussion of properties. Thermodynamic properties are. An integral expression for the conservation of energy applied to a control volume will be developed from the ﬁrst law of thermodynamics. The procedure will then be similar to that used in Chapter 5. and the integrals of such functions may be evaluated without a knowledge of the path by which the change in the property occurs between the initial and final states. d. and thus. This factor will not be written henceforth. The more familiar differential operator. is used as both heat transfer and work are path functions and the evaluation of integrals of this type requires a knowledge of the path. 1965.’’ numerically equal to 778. respectively. point functions. 6. point functions and path fuctions. J ¼ 1 N m/J.1 The quantity J is the so-called ‘‘mechanical equivalent of heat. New York. the equations resulting from the ﬁrst law of thermodynamics will be scalar in form. The symbols dQ and dW represent differential heat transfer and work done. The statement of the ﬁrst law given above may be written in equation form as %dQ ¼ 1J %dW (6-1) where the symbol % refers to a ‘‘cyclic integral’’ or the integral of the quantity evaluated over a cycle. The differential operator. the total heat added to the system from its surroundings is proportional to the work done by the system on its surroundings. unlike the momentum equations considered in Chapter 5. 14. by definition. 63 . Note that this law is written for a speciﬁc group of particles—those comprising the deﬁned system. Wiley. Principles of General Thermodynamics. N. d. The statement of the ﬁrst law of thermodynamics involves only scalar quantities however.

we may write. dQ – dW. The system under consideration. a general control volume ﬁxed in inertial space located in a ﬂuid-ﬂow ﬁeld. which is any path other than a between these points. however.64 Chapter 6 Conservation of Energy: Control-Volume Approach We now consider a general thermodynamic cycle.2. is equal to a point function or a property. as in Chapter 5. as shown in Figure 6. Utilizing equation (6-1). it follows that the quantity. equation (6-4) may be written as dQ dW dE ¼ dt dt dt (6-5) Consider now. the total energy of the system. An alternate expression for the ﬁrst law of thermodynamics may be written dQ dW ¼ dE (6-4) The signs of dQ and dW were speciﬁed in the original statement of the ﬁrst law. designated by dashed lines. occupies the control volume at time t. The cycle a occurs between points 1 and 2 by the paths indicated. and its position is also shown after a period of time Dt has elapsed. This property is designated dE.1 Reversible and irreversible thermodynamic cycles. Again equation (6-1) allows us to write Z 2 Z 1 Z 2 Z 1 dQ þ dQ ¼ dW þ dW (6-2b) 1a 2b 1a 2b Subtracting equation (6-2b) from equation (6-2a) gives Z 1 Z 1 Z 1 Z dQ dQ ¼ dW 2a 2b 2a 1 dW 2b which may be written Z 1 2a Z ðdQ dWÞ ¼ 1 ðdQ dWÞ (6-3) 2b As each side of equation (6-3) represents the integrand evaluated between the same two points but along different paths. dW is positive when work is done by the system. the cycle is completed by path b between points 2 and 1. A new cycle between points 1 and 2 is postulated as follows: the path between points 1 and 2 is identical to that considered previously. as shown in Figure 6. . for cycle a Z 2 Z 1 dQ þ dQ 1a 2a (6-2a) Z 2 Z 1 ¼ dW þ 1a 2 a a P b 1 dW r 2a Figure 6.1. dQ is positive when heat is added to the system. For a system undergoing a process occurring in time interval dt.

1 65 Integral Relation for the Conservation of Energy Boundary of system at time t Streamlines at time t Stationary control volume III I II Boundary of system at time t + ∆t Figure 6. Dt. equation (6-5).2 Relation between a system and a control volume in a ﬂuidﬂow ﬁeld. we have lim Dt ! 0 EjtþDt Ejt dE ¼ dt Dt which corresponds to the right-hand side of the ﬁrst-law expression. In this ﬁgure. region II is occupied by the system at t þ Dt. region I is occupied by the system at time t. we have EjtþDt Ejt EIII jtþDt þ EII jtþDt EIII jt EI jt ¼ Dt Dt Rearranging and taking the limit as Dt ! 0 gives lim Dt ! 0 EjtþDt Ejt EIII jtþDt EIII jt EII jtþDt EI jt ¼ lim þ lim Dt Dt Dt Dt ! 0 Dt ! 0 (6-6) Evaluating the limit of the left-hand side. . and region III is common to the system both at t and at t þ Dt. as the volume occupied by the system as Dt ! 0 is the control volume under consideration. The second limit on the right of equation (6-6) EII jtþDt EI jt Dt Dt ! 0 lim represents the net rate of energy leaving across the control surface in the time interval Dt. At time t þ Dt the total energy of the system may be expressed as EjtþDt ¼ EII jtþDt þ EIII jtþDt and at time t Ejt ¼ EI jt þ EIII jt Subtracting the second expression from the ﬁrst and dividing by the elapsed time interval.6. On the right-hand side of equation (6-6) the ﬁrst limit becomes EIII jtþDt EIII jt dEIII ¼ Dt dt Dt ! 0 lim which is the rate of change of the total energy of the system.

accounts both for efﬂux and for inﬂux of mass across the control surface as considered . we may now recast the ﬁrst law of thermodynamics into a form applicable to a control volume expressed by the following word equation: 9 8 9 8 8 9 > rate of energy > rate of addition > > > > > > rate of work done = = < = < < out of control of heat to control by control volume ¼ volume due to > volume from > : . The speciﬁc energy includes the potential energy. The integral of this quantity over the control surface ZZ er(v: n)dA c:s: represents the net efﬂux of energy from the control volume. > > > > > > on its surroundings . dA. Theta (u) is the angle between v and the outwardly directed normal vector.3. We may now write e(rv)(dA cos u) ¼ er dA½jvj jnj cos u ¼ er(v: n)dA which we observe to be similar in form to the expressions previously obtained for mass and momentum. as discussed in the previous chapters. projected normal to the velocity vector. : .3 Fluid ﬂow through a control volume. The sign of the scalar product. gy. Streamlines at time t dA q v n Figure 6. v. v: n. of the ﬂuid due to its thermal state. to fluid flow control volume Equation (6-7) will now be applied to the general control volume shown in Figure 6. u. due to its velocity. The quantity e is the speciﬁc energy or the energy per unit mass. and the internal energy. The rates of heat addition to and work done by the control volume will be expressed as dQ/dt and dW/dt. : . The quantity dA cos u represents the area. due to the position of the ﬂuid continuum in the gravitational ﬁeld. Consider now the small area dA on the control surface. The rate of energy leaving the control volume through dA may be expressed as rate of energy efflux ¼ e(rv)(dA cos u) The product (rv)(dA cos u) is the rate of mass efﬂux from the control volume through dA. n. : its surroundings (6-7) 8 9 8 9 fluid flow < rate of energy into = < rate of accumulation = control volume due þ of energy within : . the kinetic energy of the ﬂuid. v 2 /2.66 Chapter 6 Conservation of Energy: Control-Volume Approach Having given physical meaning to each of the terms in equation (6-6).

The third type of work is designated shear work. Vector S is the force intensity (stress) having components sii and tij in the directions normal and tangential to the surface. as shown in Figure 6.6.4. Thus. the ﬁrst two terms on the right-hand side of equation (6-7) may be evaluated as 8 9 8 9 < rate of energy = < rate of energy = Z Z out of control into control ¼ er(v: n)dA : . . In terms of S. another effect on the elemental portion of control surface. dA. which is performed on the surroundings to overcome shear stresses at the control surface. Ws . Examining our control volume for ﬂow and shear work rates.4 Flow and shear work for a general control volume. and the rate of work done by the ﬂuid ﬂowing through dA is S dA: v. There are three types of work included in the work-rate term. respectively. c:s: volume volume The rate of accumulation of energy within the control volume may be expressed as ZZZ @ er dV @t c:v: Equation (6-7) may now be written as ZZ ZZZ dQ dW @ ¼ er(v: n) dA þ er dV dt dt @t c:s: c:v: (6-8) A ﬁnal form for the ﬁrst-law expression may be obtained after further consideration of the work-rate or power term. The ﬁrst is the shaft work. Ws.1 Integral Relation for the Conservation of Energy 67 previously. dW/dt. we have. S dA v q n Figure 6. A second kind of work done is ﬂow work. which is that done by the control volume on its surroundings that could cause a shaft to rotate or accomplish the raising of a weight through a distance. the force on dA is S dA. : . which is that done on the surroundings to overcome normal stresses on the control surface where there is ﬂuid ﬂow. Wt . The net rate of work done by the control volume on its surroundings due to the presence of S is ZZ c:s: v: S dA where the negative sign arises from the fact that the force per unit area on the surroundings is –S.

equation (6-8). we obtain. The remaining part of the normal stress term. that associated with pressure. we have ZZZ ZZ ZZ dQ dWs dWt @ þ ¼ s ii ðv: nÞdA þ erðv: nÞdA þ er dV dt @t dt dt c:s: c:s: c:v: The term involving normal stress must now be presented in a more usable form. to follow. is the negative of the thermodynamic pressure. The use of the overall energy balance will be illustrated in the following example problems. the work rate accomplished in overcoming viscous effects at the control surface. is used to make this distinction. The shear and ﬂow work terms may now be written as follows: ZZ ZZ dWt dWm : ¼ s ii (v n)dA P(v: n) dA dt dt c:s: c:s: Combining this equation with the one written previously and rearranging slightly will yield the ﬁnal form of the ﬁrst-law expression: ZZZ ZZ dQ dWs P @ dWm : ¼ (6-10) eþ er dV þ r(v n) dA þ dt r @t dt dt c:s: c:v: Equations (6-10).68 Chapter 6 Conservation of Energy: Control-Volume Approach The ﬁrst-law expression. This portion of the required work rate. and (5-4) constitute the basic relations for the analysis of ﬂuid ﬂow via the control-volume approach. m. For the present. A complete expression for sii is stated in Chapter 9. A thorough understanding of these three equations and a mastery of their application places at the disposal of the student very powerful means of analyzing many commonly encountered problems in ﬂuid ﬂow. is included in the derivative form given above and its analysis is included in Example 3. We shall thus combine the work associated with the viscous portion of the normal stress with the shear work to give a single term. dWt /dt. sii. may be written in slightly different form if we recall that the bulk stress. may now be written as ZZZ ZZ ZZ dQ dWs @ þ v: S dA ¼ er(v: n) dA þ er dV dt @t dt c:s: c:s: c:v: (6-9) where dWs /dt is the shaft work rate. P. Writing the normal stress components of S as siin. we may say simply that the normal stress term is the sum of pressure effects and viscous effects. (4-1). for the net rate of work done in overcoming normal stress Z Z ZZ ZZ : : v S dA ¼ v s ii n dA ¼ s ii ðv: nÞ dA c:s: normal c:s: c:s: The remaining part of the work to be evaluated is the part necessary to overcome shearing stresses. . the work done to overcome the viscous portion of the normal stress is unavailable to do mechanical work. Just as with shear work. is transformed into a form that is unavailable to do mechanical work. The subscript. The work rate now becomes ZZ dW dWs dWs dWt dWs dWt ¼ þ þ ¼ s ii (v: n) dA þ dt dt dt dt dt dt c:s: Substituting into equation (6-9). dWm/dt. This term. representing a loss of mechanical energy.

Considering now the surface integral. dependent upon the sense of v: n. the mass balance for this same situation was found to be m_ ¼ r1 v 1 A1 ¼ r1 v 2 A2 If each term in the above expression is now divided by the mass ﬂow rate. we recognize the product r(v: n) dA to be the mass ﬂow rate with the sign of this product indicating whether mass ﬂow is into or out of the control volume. A2. The speciﬁc total energy. equation (6-10). the surface integral becomes 2 ZZ v P P2 (v: n) dA ¼ 2 þgy2 þ u2 þ (r v 2 A2 ) r eþ r 2 r2 2 c:s: 2 v P1 1 þgy1 þ u1 þ (r v 1 A1 ) 2 r1 1 The energy expression for this example now becomes 2 2 v v dQ dWs P2 P1 ¼ 2 þ gy2 þ u2 þ (r2 v 2 A2 ) 1 þgy1 þ u1 þ (r v 1 A1 ) dt dt 2 r2 2 r1 1 In Chapter 4. we have 2 2 _s v v qW P2 P1 ¼ 2 þ gy2 þ u2 þ 1 þ gy1 þ u1 þ m_ 2 r2 2 r1 . e þ P/r.2 Applications of the Integral Expression 69 6. so that eþ P v2 P ¼ gy þ þ u þ r r 2 As mass enters the control volume only at section (1) and leaves at section (2). The factor by which the mass-ﬂow rate is multiplied. represents the types of energy that may enter or leave the control volume per mass of ﬂuid. e. r2 dWs dt Figure 6. becomes ZZ er dV þ dWm dt ! ZZZ P @ (v: n)dA þ r eþ r @t c:s: c:v: ! dQ dWs ¼ dt dt 0steady flow 0 dQ dt 1 2 υ1. may be expanded to include the kinetic.6.2 APPLICATIONS OF THE INTEGRAL EXPRESSION EXAMPLE 1 As a ﬁrst example. For the speciﬁed conditions the overall energy expression.5 Control volume with one-dimensional ﬂow across boundaries. r1 υ2. A1. and internal energy contributions. let us choose a control volume as shown in Figure 6. potential.5 under the conditions of steady ﬂuid ﬂow and no frictional losses.

6 A control volume for pump analysis. which is equal to the sum of these quantities by deﬁnition h u þ P/r. that is.70 Chapter 6 Conservation of Energy: Control-Volume Approach or. u þ P/r. The stagnation pressure is. Such a pressure is designated the stagnation pressure. 6 in. Hg Figure 6. which is greater than the static pressure by an amount equivalent to the change in kinetic energy of the ﬂow. Shaft work 2 12 in. where the velocity has been reduced to zero. If water ﬂows under steady conditions in which the pump delivers 3 horsepower to the ﬂuid. in more familiar form _s v 21 q v2 W þ gy1 þ h1 þ ¼ 2 þ gy2 þ h2 þ m_ m_ 2 2 where the sum of the internal energy and ﬂow energy. Deﬁning the control volume as shown by the dashed lines. 1 Pump 6 in. ﬁnd the mass ﬂow rate if frictional losses may be neglected.6. EXAMPLE 2 As a second example. has been replaced by the enthalpy. expressed as 1 Pstagnation ¼ P0 ¼ Pstatic þ rv 2 2 . we may evaluate equation (6-10) term by term as follows: dQ ¼0 dt dWs ¼ (3 hp)(2545 Btu / hph)(778 ftlbf / Btu)(h /3600 s) dt ¼ 1650 ft lbf /s ZZ eþ c:s: ZZ ZZ P P P r(v: n)dA ¼ r(v: n)dA þ r(v: n) dA eþ eþ r r r A2 A1 2 v2 P2 þ gy2 þ u2 þ ¼ (r2 v 2 A2 ) 2 r2 2 v1 P1 þ gy1 þ u1 þ (r1 v 1 A1 ) 2 r1 2 v v 21 P2 P1 þ g( y2 y1 ) þ (u2 u1 ) þ (rvA) ¼ 2 2 r2 r1 Here it may be noted that the pressure measured at station (1) is the static pressure while the pressure measured at station (2) is measured using a pressure port that is oriented normal to the oncoming ﬂow. consider the situation shown in Figure 6. h. thus.

The potential energy change is zero between sections (1) and (2) and as we consider the ﬂow to be isothermal. Find the rate at which energy must be removed from the bearing in order that the lubricating D oil between the rotating shaft and the stationary bearing d surface remains at constant temperature. The sum of these two quantities is known as the impact or stagnation pressure. Hence. the variation in internal energy is also zero.7 Bearing and control volume for bearing analysis.2 71 Applications of the Integral Expression for incompressible ﬂow.The control volume selected consists of a unit length of the ﬂuid surrounding the shaft as shown in Figure 6.6. the manometer opening is normal to the ﬂowing ﬂuid stream. the surface integral reduces to the simple form indicated. however. The pressure sensed at section (1) is the static pressure. hence the energy ﬂux term may be rewritten as ZZ P P02 P1 v 21 r(v: n) dA ¼ (rvA) eþ r r 2 c:s: ( 6(1 1/13:6) in: Hg(14:7 lb/in:2 )(144 in:2 /ft2 ) ¼ (62:4 lbm /ft3 )(29:92 in: Hg) ) v 21 f(62:4 lbm /ft3 )(v 1 )(p/4 ft2 )g 64:4(lbm ft /s2 lbf ) v2 ¼ 6:30 1 (49v 1 ) ft lbf /s 64:4 ZZZ @ er dV ¼ 0 @t c:v: dWm ¼0 dt In the evaluation of the surface integral the choice of the control volume coincided with the location of the pressure taps at sections (1) and (2). as the manometer opening is parallel to the ﬂuid-ﬂow direction. The shaft is assumed to be lightly loaded and concentric with the journal. . The shaft diameter is d and the shear stress acting on the shaft is t. The pressure measured by such an arrangement includes both the static ﬂuid pressure and the pressure resulting as a ﬂuid ﬂowing with velocity v 2 is brought to rest. At section (2). The ﬁrst law of thermodynamics for the control volume is ZZ dQ dWs P (v : n)dA ¼ r eþ dt r dt c:s: ZZZ @ dWm re dV þ þ dt @t c:v: w dQ dt Figure 6. The solution is v 1 ¼ 16:59 fps (5:057 m/s) m_ ¼ rAv ¼ 813 lbm /s (370 kg/s) EXAMPLE 3 A shaft is rotating at constant angular velocity v in the bearing shown in Figure 6. The ﬂow rate of water necessary for the stated conditions to exist is achieved by solving the resulting cubic equation.7.7.

with constant speciﬁc heat c c dT 2tv d 2 ¼ dt r(D2 d 2 ) where D is the outer bearing diameter. on the surroundings. the expression of the ﬁrst law of thermodynamics applied to a control volume reduces to an extremely useful relation known as the Bernoulli equation. The ﬂow is steady. If energy is not removed from the system then dQ/dt ¼ 0. and in which no heat transfer or change in internal energy occurs.72 Chapter 6 Conservation of Energy: Control-Volume Approach From the ﬁgure we may observe the following: 1. At the outer boundary. RR thus the viscous work is c:s: t(v et ) dA. 2. If equation (6-10) is applied to a control volume as shown in Figure 6. t. incompressible. When the velocity on the surface of the control volume is zero. 2. v ¼ 0 and at the inner boundary c:s: t(v: et )dA ¼ t(vd/2)A. the viscous-work term is zero. The resulting sign is consistent with the concept of work being positive when done by a system on its surroundings.8.3 THE BERNOULLI EQUATION Under certain ﬂow conditions. a term-by-term evaluation of equation (6-10) gives the following: dQ ¼0 dt dWs ¼0 dt . in which ﬂow is steady. 6. dQ v d2 p ¼ t dt 2 which is the heat transfer rate required to maintain the oil at a constant temperature. where et indicates the sense of the shear stress. and inviscid. No shaft work crosses the control surface. and ZZZ @ dWm er dV ¼ @t dt c:v: As only the internal energy of the oil will increase with respect to time @ @t ZZZ 2 D d 2 dm dWm d2 p ¼ t er dV ¼ rp ¼v dt 2 4 dt c:v: or. No ﬂuid crosses the control surface. The viscous work rate must be determined. In this example the use of the viscous-work term has been illustrated.RRIn this case all of the : viscous work is done to overcome shearing stresses. Note that 1. Thus dQ/dt = dWm/dt = dWt/dt. The viscous-work term involves only quantities on the surface of the control volume. 3. Thus.

The quantities are often designated ‘‘heads’’ due to elevation. One such situation of practical value is for ﬂow into and out of a stream . ZZ ZZ P P r eþ r eþ (v: n) dA ¼ (v: n) dA r r c:s: A1 ZZ P r eþ þ (v: n) dA r A2 v 21 P1 ¼ gy1 þ þ ( r1 v 1 A1 ) 2 r1 v 22 P2 (r2 v 2 A2 ) þ gy2 þ þ 2 r2 ZZZ @ er dV ¼ 0 @t c:v: The ﬁrst-law expression now becomes v 22 P2 v 21 P1 (rv 2 A2 ) gy1 þ þ (rv 1 A1 ) 0 ¼ gy2 þ þ 2 r 2 r As ﬂow is steady. isothermal ﬂow.3 The Bernoulli Equation 73 A2 Streamlines 2 Control volume A1 1 Figure 6. steady. or approached. but they are met. Note that each term in equation (6-11b) has the unit of length. inviscid. incompressible. velocity. in many physical systems.6. These terms. Equation (6-11) may be interpreted physically to mean that the total mechanical energy is conserved for a control volume satisfying the conditions upon which this relation is based. inviscid. indicate the quantities which may be directly converted to produce mechanical energy. These conditions may seem overly restrictive. and pressure. that is. with no heat transfer or work done. both individually and collectively.8 Control volume for steady. incompressible. respectively. we have y1 þ Either of the above expressions is designated the Bernoulli equation. the continuity equation gives r 1 v 1 A1 ¼ r 2 v 2 A2 which may be divided through to give gy1 þ v 21 P1 v 2 P2 þ ¼ gy2 þ 2 þ 2 r 2 r (6-11a) v 21 P1 v 2 P2 þ ¼ y2 þ 2 þ 2g rg 2g rg (6-11b) Dividing through by g. isothermal ﬂow.

the pressure acting at section (1) is considered uniform with value P1. and pressure head from point-to-point in a ﬂuid-ﬂow ﬁeld. and A2. the Bernoulli equation can actually describe the variation in elevation.10. r1 A2 1 Figure 6. A classic example of the application of the Bernoulli equation is depicted in Figure 6.10 Flow through a sudden enlargement. u2 The control volume is deﬁned as shown by dashed lines in the ﬁgure.9 Control volume for Bernoulli equation analysis. u1 A1. Find the change in internal energy between stations (1) and (2) for steady. and thus can be considered to be at the same height as the ﬂuid. Pd = P1 u1.74 Chapter 6 Conservation of Energy: Control-Volume Approach tube. Under these conditions. velocity. There is ﬂuid ﬂow across this surface. Constant level maintained in tank 1 y1 2 Figure 6. we have y1 þ Patm v 22 Patm ¼ þ rg 2g rg from which the exiting velocity may be expressed in the familiar form pﬃﬃﬃﬃﬃﬃﬃ v2 ¼ 2gy As a ﬁnal illustration of the use of the control-volume relations. Neglect shear stress at the walls and express u2 u1 in terms of v 1. The control volume selected is indicated by the dotted line. EXAMPLE 4 In the sudden enlargement shown below in Figure 6. the Bernoulli equation. A1. 2 Conservation of Mass ZZ c:s: r(v: n)dA þ @ @t ZZZ r dV ¼ 0 c:v: . the proper form of the ﬁrst law of thermodynamics is equation (6-11).9. Applying equation (6-11). an example using all three expressions is presented below. As stream tubes may vary in size. but the surface area is large enough that the velocity of this ﬂowing ﬂuid may be considered negligible. in which it is desired to ﬁnd the velocity of the ﬂuid exiting the tank as shown. The upper boundary of the control volume is just below the ﬂuid surface. incompressible ﬂow.

(6-13) ZZZ P @ dWm r eþ re dV þ (v: n) dA þ dt r @t c:s: c:v: ZZ e1 þ P1 P2 (rv 1 A1 ) ¼ e2 þ (rv 2 A2 ) r r or. we obtain 2 A1 A1 v 21 v 21 A1 2 u2 u1 ¼ v 21 v 21 þ A2 A2 2 2 A2 " # v2 v2 A1 A1 2 A1 2 ¼ 1 1 ¼ 1 12 þ (6-15) 2 A2 A2 2 A2 . for steady. incompressible ﬂow. we have u2 u1 ¼ P1 P2 v 21 v 22 þ þ g( y1 y2 ) r 2 (6-14a) Substituting equation (6-13) for (P1 P2 )/r and equation (6-12) for v 2 and noting that y1 ¼ y2. becomes r1 v 1 A1 ¼ r2 v 2 A2 or v2 ¼ v1 Momentum åF ¼ ZZ c:s: A1 A2 rv(v: n)dA þ @ @t (6-12) ZZZ rv dV c:v: and thus P1 A2 P2 A2 ¼ rv 22 A2 rv 21 A1 or P1 P2 A1 ¼ v 22 v 21 r A2 Energy dQ dWs ¼ dt @t Thus. e1 þ P1 P2 ¼ e2 þ r r The speciﬁc energy is e¼ v2 þ gy þ u 2 Thus our energy expression becomes v 21 P1 v 22 P2 þ gy1 þ u1 þ ¼ þ gy2 þ u2 þ 2 r 2 r (6-14) The three control-volume expressions may now be combined to evaluate u2 u1.6. since rv 1 A1 ¼ rv 2 A2 . the continuity expression. a considerable distance downstream from the sudden enlargement.3 The Bernoulli Equation 75 If we select station (2). From equation (6-14).

but from equation (6-14a) it can be seen that the change in total head. a Bourdon pressure gage just outside the casing of the 12-in.-diameter suction pipe reads 6 psig (i. the third of the fundamental relations upon which ﬂuid-ﬂow analyses are based. has been used to develop an integral expression for the conservation of energy with respect to a control volume. 6. this expression has broad application to physical situations. On the 10-in.e. Develop an expression for the ﬂuid velocity. is. PROBLEMS 6. The outlet pressure is 175 kPa. 6. P1 v 21 P2 v 22 þ þ gy1 þ þ gy2 r 2 r 2 is equal to the internal energy change. The resulting expression.76 Chapter 6 Conservation of Energy: Control-Volume Approach Equation (6-15) shows that the internal energy increases in a sudden enlargement. vacuum). The heat transfer to the pipe can be neglected. that is.1 The velocity proﬁle in the annular control volume of Example 3 is approximately linear. v(r). and the energy equation for steady.-diameter pipe.. in conjunction with equations (4–1) and (5–4). the internal energy change in an incompressible ﬂow is designated as the head loss.25 m in diameter. is 1. 0.2 Sea water. A special case of the integral expression for the conservation of energy is the Bernoulli equation.8 m above the inlet. equation (6-11).152 m in diameter. hL.-diameter horizontal pipe at a ﬂow rate of 35 gal/min. how much power does the pump add to the ﬂuid? 6. At the inlet the pressure is –0. r ¼ 1025 kg/m3 . ﬂows through a pump at 0:21 m3 /s.15 m of mercury.5 During the ﬂow of 200 ft3/s of water through the hydraulic turbine shown. Assume the incoming air is at reservoir pressure and ﬂows through a 8-in. The pump inlet is 0.6 During the test of a centrifugal pump. one of the fundamental expressions for the control-volume analysis of ﬂuid-ﬂow problems. and frictional forces result in a pressure drop of 10 psi. Accordingly. adiabatic.4 CLOSURE In this chapter the ﬁrst law of thermodynamics. P + rv 2/2 for an incompressible ﬂuid. Flows into a 10-ft3 reservoir at a velocity of 110 fps.3 Air at 708F. A Turbine D = 2 ft 15 ft B 6. incompressible ﬂow in a stream tube is written as P1 v 21 P2 v 22 þ þ y1 ¼ hL þ þ þ y2 (6-16) rg 2g rg 2g Note the similarity to equation (6-11). The temperature change corresponding to this change in internal energy is insigniﬁcant. What is the temperature change of the water? 6. 6.4 Water ﬂows through a 2-in. If the reservoir pressure is 14 psig and the reservoir temperature 708F. If the inlet and exit temperature are equal. The pump outlet. varying from a velocity of zero at the outer boundary to a value of vd/2 at the inner boundary. ﬁnd the rate of temperature increase in the reservoir.-diameter discharge . where r is the distance from the center of the shaft. What should gage B read if the turbine is delivering 600 hp at 82% efﬁciency? Gage B is designed to measure the total pressure. equation (6-10). Although simple in form and use. the pressure indicated by gage A is 12 psig.

Assuming that the air pressure is maintained constant.30-mdiameter round duct that has a smoothly rounded entrance.10 m of alcohol in a differential manometer connected to the inlet and the throat? The speciﬁc gravity of alcohol may be taken as 0. 3.10 In Problem 6.60 m 6. P1 ¼ P3. Determine the volume rate of air ﬂow in the duct in cubic feet per second. What is the required throat diameter if this ﬂow is to give a reading of 0. .6 m from the nozzle as shown. The inlet pipe is submerged 6 ft into the water and is vertical.7 A fan draws air from the atmosphere through a 0.6. If the barometer reads 29 in.5 cm 6.5 cm of water. 1 2 6. cv.. Compute the horsepower input of the test pump. A differential manometer connected to an opening in the wall of the duct shows a vacuum pressure of 2.13 An automobile is driving into a 45-mph headwind at 40 mph. The density of air is 1. A3.30 m 2 in. what is the pressure at a point on the auto where the wind velocity is 120 fps with respect to the auto? 6. v 3.11 A Venturi meter with an inlet diameter of 0. The internal energy is given by cvT. Use the results of Problem 6. Estimate the pressure inside the pipe at the pump inlet.0-cm-diameter nozzle that is inclined at a 308 angle above the horizontal. What is the horsepower output of the fan? 77 6.12 The pressurized tank shown has a circular cross section of 6 ft in diameter.14 Water is discharged from a 1. P – Patm = 4 in. The discharge of water through the pump is measured to be 4 ft3/s.22 kg/m3. Hg and the temperature is 408F.6 m and a vertical distance of 0. 6. Explain why you cannot proﬁtably use Bernoulli’s equation here for a force calculation. in diameter in the side of the tank.) q 30° Top view 3 0.45 ft of ﬂowing ﬂuid.85 and that of mercury is 13. v 1. The ﬂuid is water and T1 ¼ T3.6 m is designed to handle 6 m3/s of standard air.95 in. The discharge pipe is 5 ft above the suction pipe.9 A liquid ﬂows from A to B in a horizontal pipe line shown at a rate of 3 ft3/s with a friction loss of 0. and it is 10 ft long. A B 6 in.8 Find the change in temperature between stations (1) and (2) in terms of the quantities A1.26 as well as any other data given in that problem that you may need. what is the rate of ﬂow in cubic meters per second? What is the total head of the jet? (See equation (6-11b).6 m 6.15 The pump shown in the ﬁgure delivers water at 598F at a rate of 550 gal/min. 6. If the jet strikes the ground at a horizontal distance of 3. what will be the pressure head at A? PA 24 in. how long does it take to lower the oil surface in the tank by 2 ft? The speciﬁc gravity of the oil in the tank is 0.8. 6 ft 2. 12 in. The inlet pipe has an inside diameter of 5. For a pressure head at B of 24 in. compute the upward force on the device from water and air. Oil is drained through a nozzle 2 in.Problems pipe another gage reads 40 psig. Hg Fan Oil 5 ft 0. 6.26. and u.

8 in. The ﬂow is upward.24 A 1968 Volkswagen sedan is driving over a 7300-ft-high mountain pass at a speed of v m/s into a headwind of W m/s.23 Residential water use. The pump inlet is located far enough away from the outlet that the inlet and outlet do not interact. 15 m long. Assume that friction causes a head loss of 4 ft. The local air density is 0.17 Using the data of Problem 5. Estimate the change in water temperature caused by this drop. Compute the gage pressure in mPa at a point on the auto where the velocity relative to the auto is v – W m/s.16 In the previous problem. 6. 6. If the skirt has a rectangular shape 3 9 m. 6.27 Water ﬂows through the pipe contraction shown at a rate of 1 ft3/s. 6.19 Determine the head loss between stations (1) and (2) in Problem 5. supplied by a compressor. The inlet is also vertical so that the net thrust of the jets on the ship is independent of the inlet velocity and pressure.984 kg/m3. If the water is delivered to a residence at 60 psig. the pressure 340. What pressure rise would result from such a velocity head? 6. Neglect line losses and elevation changes.247 psi.22 is " 1=2 # 8v 2 h1 1 1þ 1 h2 ¼ 2 gh1 6.26 Water ﬂows steadily up the vertical pipe and is then deﬂected to ﬂow outward with a uniform radial velocity.18 In order to maneuver a large ship while docking.78 Chapter 6 Conservation of Energy: Control-Volume Approach 6. determine the velocity head of the ﬂuid leaving the impeller. lowpressure pump? Jet Side The solution to Problem 5. a low-volume. determine the airﬂow rate needed to maintain the cushion and the power given by the 1 2 in. what is the ﬂow rate of water through the pipe if the pressure at A is 10 psig? 12 in. Top 6. estimate the monthly energy required to pump the water from atmospheric pressure to the delivery pressure. If the increase in internal energy is 200. determine the ﬂow rate at which the pump inlet pressure is equal to the vapor pressure of the water. pumps are used to issue a jet of water perpendicular to the bow of the ship as shown in the ﬁgure. 6. Assume that the inlet and outlet are at the same depth.25 A liquid is heated in a vertical tube of constant diameter. Assume that the air speeds within the cushion are very low. Assume the pumps are 75% efﬁcient and are driven by electric motors with 90% efﬁciency. At the entrance the average velocity is 1 m/s. the vehicle mass is 8100 kg and the ground clearance is 3 cm. 6. 6. compressor to the air. Air. high-pressure pump or a high-volume. The vapor pressure of water at 598F is 0. runs about 80 gallons per person per day. ﬁnd the heat added to the ﬂuid. escapes through the clearing between the ground and the skirt of the vehicle.21 An ‘‘air cushion’’ vehicle is designed to traverse terrain while ﬂoating on a cushion of air. 5 ft A 6. Which will produce more thrust per horsepower. Calculate the differential manometer reading in inches of . exclusive of ﬁre protection.7.27. If friction is neglected. Derive all expressions for the change in total head across a hydraulic jump. Determine the pump horsepower required per pound of thrust. and the density is 1001 kg/m3.000 Pa.22 Show that Bernoulli’s equation applied between sections 1 and 2 does not give this result.20 Multnomah Falls in Oregon has sheer drop of 165 m.000 J/kg.

Can Bernoulli’s equation be used across section (1)? u=? 0. 79 4 ft ? 6. and the pressure head at B if the pipe has a uniform diameter of 1 in. 2 in.90 m Compressed air H2O 10 ft 6.2 v 2/g where v is the ﬂow velocity in the pipe. 1. B 4 in. Neglecting friction and treating the ﬂow processes as incompressible except for the heating. compute the discharge velocity v of the air–water mixture. Neglecting friction and the unsteadiness of the ﬂow. determine the velocity. Neglecting any pressure drop across section (1).8 m 6.Problems mercury. Diameter 4 in.86 L2 (b) 6.32 In Problem 6. and D.30 Air of density 1. The ﬂuid then escapes through a vertical chimney that has a height L. B. 6. Find the rate of discharge in cubic feet per second. ﬁnd the time required for the water in the tank to drop from a level of 28 ft above the nozzle to the 4-ft level.G. determine the readings on manometers (a) and (b) in the ﬁgures below. assuming no energy loss in the ﬂow.28 The ﬁgure illustrates the operation of an air lift pump.31. Diameter r1 Heating section . 6.34 Water in an open cylindrical tank 15 ft in diameter discharges into the atmosphere through a nozzle 2 in. assume the ﬂow to be frictionless in the siphon. Be sure to give the correct direction of the manometer reading. 6 in. in diameter. (b) the pressure and velocity at points A.35 A ﬂuid of density r1 enters a chamber where the ﬂuid is heated so that the density decreases to r2. The ﬂuid velocity entering the heating chamber may be neglected and the chimney is immersed in ﬂuid of density r1. entrance.29 Rework Problem with the assumption that the momentum of the incoming air at section (1) is zero. If v ¼ 15 m/s. in the stack.33 Assume that the level of water in the tank remains the same and that there is no friction loss in the pipe. and the magnitude of the pressure drop at section (1). v.5.31 Referring to the ﬁgure. ﬁnd the rate of discharge if the frictional head loss in the pipe is 3. or nozzle. Determine 1 (a) the volumetric discharge rate from the nozzle. v. C. Air u H2O (a) 20 ft 20 ft 23 ft B D 3 ft A 6. u V 24 m/s r2 Oil r1 S. = 0. How long will it take for the water level to decrease by 3 ft? The tank diameter is 10 ft. C 6. Determine the exit velocity. Compressed air is forced into a perforated chamber to mix with the water so that the speciﬁc gravity of the air–water mixture above the air inlet is 0.21 kg/m3 is ﬂowing as shown.

6. what is the mass ﬂow rate in each line? .36 Repeat the previous problem without the assumption that the velocity in the heating section is negligible. Assuming frictionless ﬂow.80 Chapter 6 Conservation of Energy: Control-Volume Approach 6.37 Consider a 4-cm pipe that runs between a tank open to the atmosphere and a station open to the atmosphere 10 m below the water surface in the tank. Assuming frictionless ﬂow. what will be the mass ﬂow rate? Repeat the problem if a head loss of 3 v 2 /g occurs in the pipe where v is the ﬂow velocity in the pipe. The ratio of the ﬂow area of the heating section to the chimney ﬂow area is R. The exits of both lines are open to the atmosphere. 6. also 4 cm in diameter runs from the tank to a station 20 m below the water level in the tank. a 4-cm pipe that exits 10 m below the water level in the tank and a second line. what will be the mass ﬂow rate? If a nozzle with a 1-cm diameter is placed at the pipe exit.38 The tank in the previous problem feeds two lines.

From Figure 7. but it has not been related to the ﬂuid or ﬂow properties.Dy.1. 7. observable ﬂuctuations in ﬂuid and ﬂow properties. The criteria for laminar and turbulent ﬂows will be discussed in Chapters 12 and 13. dd/dt ¼ dv/dy ¼ rate of shear strain: 81 .Dy. the viscosity depends upon the temperature.Chapter 7 Shear Stress in Laminar Flow In the analysis of ﬂuid ﬂow thus far. the ﬂuid ﬂows in smooth layers or lamina. and the shear stress is the result of these ﬂuctuations. The ﬂow parallel to the x axis will deform the element if the velocity at the top of the element is different than the velocity at the bottom. the viscosity is the property of a fluid to resist the rate at which deformation takes place when the fluid is acted upon by shear forces.1 NEWTON’S VISCOSITY RELATION In a solid. there exists a relation similar to (7-1). The shear stress acting on a ﬂuid depends upon the type of ﬂow that exists. composition. Turbulent ﬂow is characterized by the large scale. The rate of deformation in a simple ﬂow is illustrated in Figure 7. We shall now investigate this relation for laminar ﬂow. but is independent of the rate of shear strain.1. it may be seen that djtþDt djt dd ¼ lim Dt Dx.Dt!0 dt fp/2 arctan ½(vjyþDy vjy )Dt/Dyg p/2 ¼ lim Dt Dx. the resistance to deformation is the modulus of elasticity. shear stress has been mentioned. The rate of shear strain at a point is deﬁned as dd/dt. The shear stress in turbulent ﬂow will be discussed in Chapter 13.Dt!0 (7-3) In the limit. and the shear stress is the result of the (nonobservable) microscopic action of the molecules. As a property of the fluid. and pressure of the fluid. which relates the shear stress in a parallel. laminar ﬂow to a property of the ﬂuid. This relation is Newton’s law of viscosity viscosity ¼ shear stress rate of shear strain (7-2) Thus. The shear modulus of an elastic solid is given by shear modulus ¼ shear stress shear strain (7-1) Just as the shear modulus of an elastic solid is a property of the solid relating shear stress and shear strain. In the so-called laminar ﬂow.

u.2. In Newtonian ﬂuids. we may write Newton’s law of viscosity as dv (7-4) t¼m dy The velocity profile and shear stress variation in a fluid flowing between two parallel plates is illustrated in Figure 7.3.2 NON-NEWTONIAN FLUIDS Newton’s law of viscosity does not predict the shear stress in all ﬂuids. The ‘‘ideal plastic’’ has a 1 t tic las lp a Ide c sti l ea pla c sti R o ud pla e Ps nian Yield stress to New fluid tant Dila Rate of strain Figure 7. While ﬂuids deform continuously under the action of shear stress. the shear stress varies in a linear manner. t 7.1 Deformation of a ﬂuid element. The velocity profile1 in this case is parabolic. Fluids are classiﬁed as Newtonian or nonNewtonian. In non-Newtonian ﬂuids. the shear stress depends upon the rate of shear strain.82 Chapter 7 Shear Stress in Laminar Flow y (u ⎜y+∆y – u ⎜y) ∆t ∆y d ∆x Element at time t Element at time t + ∆t Figure 7. h h u Figure 7. plastics will sustain a shear stress before deformation occurs. x Combining equations (7-2) and (7-3) and denoting the viscosity by m. as the shear stress is proportional to the derivative of the velocity.2 Velocity and shear stress proﬁles for ﬂow between two parallel plates.3 Stress rate-of-strain relation for Newtonian and non-Newtonian ﬂuids. the relation is linear. The derivation of velocity proﬁles is discussed in Chapter 8. depending upon the relation between shear stress and the rate of shearing strain. . as shown in Figure 7.

4 Molecular motion at the surface of a control volume. hence the name no-slip (boundary) condition. y ux = ux(y) t ∆y ∆x t x Figure 7. the layer of ﬂuid next to the wall is at rest.3 VISCOSITY The viscosity of a ﬂuid is a measure of its resistance to deformation rate. Thixotropic substances such as printer’s ink have a resistance to deformation that depends upon deformation rate and time. In ﬂow situations in which the viscous effects are neglected—the so called inviscid ﬂows—only the component of the velocity normal to the boundary is zero. An understanding of the existence of the viscosity requires an examination of the motion of ﬂuid on a molecular basis. The no-slip condition is the result of experimental observation and fails when the ﬂuid no longer can be treated as a continuum. Tar and molasses are examples of highly viscous ﬂuids. are examples of ﬂuids with relatively low viscosities. The no-slip condition is a result of the viscous nature of the ﬂuid.4. they are similar in their action at a boundary. The No-Slip Condition Although the substances above differ in their stress rate-of-strain relations. which are the subject of frequent engineering interest. When the boundary is a stationary wall. the layer of ﬂuid moves at the velocity of the boundary. the layer of ﬂuid adjacent to the boundary has zero velocity relative to the boundary. Consider the control volume shown in Figure 7.3 Viscosity 83 linear stress rate-of-strain relation for stresses greater than the yield stress. If the boundary or wall is moving. In both Newtonian and non-Newtonian ﬂuids. air and water. . The mechanism by which a gas resists deformation may be illustrated by examination of the motion of the molecules on a microscopic basis. The molecular motion of gases can be described more simply than that of liquids. 7.7.

The paths of the molecules between collisions are represented by the random arrows. we have Z t ¼ å mn (v x jy v x jyþ ) (7-6) n¼1 We shall treat shear stress exclusively as a force per unit area. and also noting from the same source that d ¼ 2/3l. l. for the shear stress Z dv x t ¼ 2 å mn dn dy n¼1 y0 In the above expression d is the y component of the distance between molecular collisions. as the average distance between collisions. If the molecular portion of the total momentum ﬂux is to be treated as a force. (v x jy v x jyþ ) in equation (7-6). where y ¼ y0 d. it must be placed on the left-hand side of equation (5-4). Because the top of the control volume is a streamline. 4 dv x t ¼ mlZ (7-7) 3 dy y0 as the shear stress. Using a similar expression for yþ. we shall write the x-directional average velocity of the upward molecular ﬂux as v x jy . The bracketed term. we obtain. If Z molecules cross the plane per unit time. . Denoting the y coordinate of the top of the control surface as y0. Thus the molecular momentum ﬂux term changes sign. this term includes both the macroscopic and molecular momentum ﬂuxes. we obtain. hence the upward molecular ﬂux must equal the downward molecular ﬂux. Denoting the negative of the molecular momentum ﬂux as t. Borrowing from the kinetic theory of gases. The x-directional momentum carried across the top of the control surface is then mv x jy per molecule. where m is the mass of the molecule. the net molecular ﬂux across this surface must be zero. The relation between the molecular momentum ﬂux and the shear stress may be seen from the control-volume expression for linear momentum ZZZ ZZ @ åF ¼ rvðv: nÞ dA þ rv dV (5-4) @t c:s: c:v: The ﬁrst term on the right-hand side of equation (5-4) is the momentum ﬂux.84 Chapter 7 Shear Stress in Laminar Flow The top of the control volume is enlarged to show that. When a control volume is analyzed on a molecular basis. for a pure gas. individual molecules cross this plane. even though the top of the element is a streamline of the ﬂow. the concept of the mean free path. The molecules that cross the control surface in an upward direction have average velocities in the x direction corresponding to their points of origin. may be evaluated by noting that v x jy ¼ v x jy0 (dv x /dyjy0 )d. then the net x-directional momentum ﬂux will be Z å mn (v x jy v x jyþ ) (7-5) n¼1 The ﬂux of x-directional momentum on a molecular scale appears as a shear stress when the ﬂuid is observed on a macroscopic scale. where the minus sign signiﬁes that the average velocity is evaluated at some point below y0.

H. Kennard. . Bird. B. New York. Hirschfelder. 3 J. Part I. McGraw-Hill Book Company.7. New York. we see that 4 m ¼ mlZ 3 (7-8) The kinetic theory gives Z ¼ NC/4. Curtiss. where N ¼ molecules per unit volume C ¼ average random molecular velocity and thus 1 rlC m ¼ NmlC ¼ 3 3 or. Wiley. and Bird3 for the details of this approach. A more realistic molecular model utilizing a force ﬁeld rather than the rigidsphere approach will yield a viscosity-temperature relationship much more consispﬃﬃﬃﬃ tent with experimental data than the T result. Even though the preceding development was somewhat crude in that an indeﬁnite property. Wiley. New York. and R. Chapter 24. Molecular Theory of Gases and Liquids. the interpretation of the viscosity of a gas being due to the microscopic momentum ﬂux is a valuable result and should not be overlooked. to be essentially true for pressures up to approximately 10 atmospheres. 1954. 1938. O. we have pﬃﬃﬃﬃﬃﬃﬃﬃﬃ mkT 2 m ¼ 3=2 d2 3p (7-9) Equation (7-9) indicates that m is independent of pressure for a gas. was introduced. The most acceptable expression for nonpolar molecules is based upon the Lennard–Jones potential energy function. It is also important to note that equation (7-9) expresses the viscosity entirely in terms of ﬂuid properties. The expression for viscosity of a pure gas that results is pﬃﬃﬃﬃﬃﬃﬃﬃ 6 MT (7-10) m ¼ 2:6693 10 s 2 Vm 2 In order of increasing complexity. F. Halliday. Physics. experimentally. the molecular diameter. the expressions for mean free path are presented in R. The constant-diameter rigid-sphere model for the gas molecule is responsible for the less-than-adequate viscosity-temperature relation. Curtiss. Kinetic Theory of Gases. Experimentalpﬃﬃﬃ evidence indicates that at low temperatures the viscosity ﬃ varies more rapidly than T . Chapter 2. This has been shown. using2 1 l ¼ pﬃﬃﬃ 2pNd 2 and C¼ rﬃﬃﬃﬃﬃﬃﬃﬃ 8kT pm where d is the molecular diameter and k is the Boltzmann constant. C.3 Viscosity 85 Comparing equation (7-7) with Newton’s law of viscosity. The interested reader may refer to Hirschfelder. 1966. and E. Resnick and D. This function and the development leading to the viscosity expression will not be included here.

New York. Hence. k is the Boltzmann constant. we have fi j ¼ 1. E. C. N. the molecular theory of liquids is much less advanced. For polar molecules. see R. in A (Angstroms). Wilke. and (7-12) are for nonpolar gases and gas mixtures at low density. F. an approximate theory has been developed by Eyring. As a liquid heats up. Wiley.. s is the ‘‘collision diameter. The difﬁculties in the analytical treatment of a liquid are largely inherent in nature of the liquid itself.86 Chapter 7 Shear Stress in Laminar Flow where m is the viscosity. (7-11). For multicomponent gas mixtures at low density.’’ a Lennard–Jones parameter. the molecules become more mobile. and e is the characteristic energy of interaction between molecules. Mj are the molecular weights of species i and j. Transport Phenomena. J. Vm is the ‘‘collision integral. Values of s and e for various gases are given in Appendix K. Curtiss. in pascal-seconds. O. and E. xj are mole-fractions of species i and j in the mixture. 5 . Experimental evidence for the viscosity of liquids shows that the viscosity decreases with temperature in agreement with the concept of intermolecular adhesive forces being the controlling factor. 1.’’ a Lennard–Jones parameter that varies in a relatively slow manner with the dimensionless temperature kT/e. Molecular Theory of Gases and Liquids. Units of Viscosity The dimensions of viscosity may be obtained from Newton’s viscosity relation m¼ 4 t dv/dy C. This results in less restraint from intermolecular forces.38 1016 ergs/K. Bird. the preceding relation must be modiﬁed. Whereas in gases the distance between molecules is so great that we consider gas molecules as interacting or colliding in pairs. and a table of Vm versus kT/e is also included in Appendix K. In spite of these difﬁculties. 517–519 (1950). 6 For a description of Eyring’s theory. B. B. R. J. W. Chem. mj are the viscosities of species i and j.5 Although the kinetic theory of gases is well developed. New York. and mi. Note that when i ¼ j. the major source of knowledge concerning the viscosity of liquids is experiment. Wilke4 has proposed this empirical formula for the viscosity of the mixture: n mmixture ¼ å x i mi i¼1 åx j fi j where xi. and 2 !1/2 32 1/2 M j 1/4 5 1 Mi 41 þ mi fi j ¼ pﬃﬃﬃ 1 þ Mj mj Mi 8 (7-11) (7-12) where Mi. in K. 18. Wiley. 1. T is absolute temperature. Equations (7-10). Phys. Bird. and the more sophisticated models of molecular interaction accurately predict viscosity in a gas. Lightfoot. Hirschfelder. Chap. the close spacing of molecules in a liquid results in the interaction of several molecules simultaneously. and R. 1960. This situation is somewhat akin to an N-body gravitational problem.6 The viscosity of a liquid can be considered due to the restraint caused by intermolecular forces. which illustrates the relation of the intermolecular forces to viscosity. Stewart. 1954. M is the ˚ molecular weight.

. Pa·s 200 100 80 60 50 40 30 Water 20 H2 10 8 6 5 4 3 Air 2 CO2 1 0 250 300 350 Temperature. Kinematic viscosity in the metric system is expressed in (meters)2 per second (1 m2/s ¼ 104 stokes ¼ 10. L ¼ length. m × 105.02089 lbf s/ft2 ¼ 0. is given the name kinematic viscosity and is denoted by the symbol v. we ﬁnd that the dimensions of viscosity in the mass–length–time system become M/Lt.6720 lbm/ft s).5 Viscosity–temperature variation for some liquids and gases.5 for three common gases and two liquids as functions of temperature. v. This ratio. in dimensional form F/L2 Ft ¼ (L/t)(1/L) L2 where F ¼ force. Either of the two names. Absolute and kinematic viscosities are shown in Figure 7. is frequently employed to distinguish m from the kinematic viscosity. t ¼ time.3 Viscosity 87 or. In the SI system.02089 slugs/ft s ¼ 0. m/r. dynamic viscosity is expressed in pascal-seconds (1 pascal-second ¼ 1 N n s/m2 ¼ 10 poise ¼ 0. 1000 800 600 500 400 300 Kerosene Viscosity. absolute viscosity or dynamic viscosity.7.76 ft2/s). The origin of the name kinematic viscosity may be seen from the dimensions of v: v m M/Lt L2 ¼ r M/L3 t The dimensions of v are those of kinematics: length and time. Using Newton’s second law of motion to relate force and mass (F ¼ ML/t2 ). K 400 450 Figure 7. A more extensive listing is contained in Appendix I. The ratio of the viscosity to the density occurs frequently in engineering problems.

Rate of Shear Strain The rate of shear strain for a three-dimensional element may be evaluated by determining the shear strain rate in the xy. the shear strain rate is again dd/dt. The tensor component.6. laminar ﬂows.6(a) is positive. In addition to the double subscript. as the element moves from position 1 to position 2 in time Dt djtþDt djt dd ¼ lim dt Dx. Stokes extended the concept of viscosity to three-dimensional laminar ﬂow. The stress tyx acts in the positive x direction. The student may apply similar reasoning to tyx acting on the bottom of the element and conclude that tyx is also positive as illustrated. the shear stress tyx at the top of the element acts on surface Dx Dz.88 Chapter 7 Shear Stress in Laminar Flow 7.4 SHEAR STRESS IN MULTIDIMENSIONAL LAMINAR FLOWS OF A NEWTONIAN FLUID Newton’s viscosity relation.Dy. are indicated in the positive sense. in Figure 7. The usual method of identiﬁcation of the shear stress involves a double subscript. For example.7. A shear stress component is positive when both the vector normal to the surface of action and the shear stress act in the same direction (both positive or both negative). however.Dt!0 arctanf[(v y jxþDx v y jx )Dt]/Dxg p/2 Dt . direction. As in the mechanics of solids. the element may deform in both the x and the y directions.6(a). yz. a sense is required. t i j ¼ t ji (see Appendix C). The basis of Stokes’ viscosity relation is equation (7-2) viscosity ¼ shear stress rate of shear strain (7-2) where the shear stress and rate of shear strain are those of a three-dimensional element. tij. The deﬁnition of positive shear stress can be generalized for use in other coordinate systems. and orientation with respect to a plane for identiﬁcation. Accordingly. is valid for only parallel. illustrated in Figure 7.Dt!0 Dt p/ 2 arctanf[(v x jyþDy v x jy )Dt]/Dyg ¼ lim Dt Dx. discussed previously. and xz planes.Dy. The vector normal to this area is in the positive y direction. In the xy plane illustrated in Figure 7. hence tyx as illustrated in Figure 7. Hence. is identiﬁed as follows: t ¼ magnitude first subscript ¼ direction of axis to which plane of action of shear stress is normal second subscript ¼ direction of action of the shear stress: Thus txy acts on a plane normal to the x axis (the yz plane) and acts in the y direction. Shear Stress The shear stress is a tensor quantity requiring magnitude. The shear stresses acting on an element Dx Dy Dz. we must examine shear stress and strain rate for a three-dimensional body. such as txy.

the shear strain rates in the yz and xz planes may be evaluated as ddyz @v y @v z ¼ þ dt @z @y ddxz @v x @v z ¼ þ dt @z @x y uy (ux ⎜y+∆y – ux ⎜y) ∆t ux t ∆y t + ∆t (uy ⎜x+∆x – uy ⎜x) ∆t d 1 ∆x 2 x Figure 7. it will be subscripted xy.7 Shear strain in the xy plane. In a similar manner. ddxy /dt ¼ @v x /@y þ @v y /@x.4 Shear Stress in Multidimensional Laminar Flows of a Newtonian Fluid 89 y tyx txy ∆y txy ∆x tyx x (a) z x tzy tyz txz ∆z tyz tzx ∆x tzx ∆y ∆z tzy txz y z Figure 7.7. As the shear strain evaluated above is in the xy plane. In the limit. .6 Shear stress acting in a positive sense.

P. . is more difﬁcult to express than in the case of shear strain. and the rate of shear strain. is the negative of the pressure. 7. For this reason the development of normal stress. viscosity. (7-14b).90 Chapter 7 Shear Stress in Laminar Flow Stokes’s Viscosity Relation (A) Shear Stress. The shear stress in other coordinate systems may be obtained from evaluating the shear-strain rate in the associated coordinate systems. we have. the strain rate.5 CLOSURE The shear stress in laminar ﬂow and its dependence upon the viscosity and kinematic derivatives has been presented for a cartesian coordinate system. The normal stress in rectangular coordinates written for a newtonian ﬂuid is given by s xx s yy @v x 2 =: v P ¼m 2 @x 3 @v y 2 : = v P ¼m 2 @y 3 (7-14a) (7-14b) and @v z 2 =: v P s zz ¼ m 2 @z 3 (7-14c) It is to be noted that the sum of these three equations yields the previously mentioned result: the bulk stress. Stokes’s viscosity relation for the shear-stress components in laminar ﬂow may now be stated with the aid of the preceding developments for rate of shear strain. of course. is included in detail in Appendix D. with only the result expressed below in equations (7-14a). The normal stress may also be determined from a stress rate-of-strain relation. Several problems of this nature are included at the end of this chapter. and it is to be noted that equation (7-2) forms the general relation between shear stress. will occur frequently. however. Using equation (7-2). The shear stress in other coordinate systems. for the shear stresses written in rectangular coordinate form @v x @v y þ @y @x @v y @v z þ tyz ¼ tzy ¼ m @z @y txy ¼ t yx ¼ m (7-13a) (7-13b) and tzx ¼ txz ¼ m @v z @v x þ @x @z (7-13c) (B) Normal Stress. on the basis of a generalized Hooke’s law for an elastic medium. s ¼ (s xx þ s yy þ s zz ) /3. and (7-14c).

7.2 For a two-dimensional. If the rate of travel of the ram is 0. Determine the frictional moment as a function of the angle a.1-in. what is the net force on ship one when the above velocities exist? 7. If the average velocity is 2 fps. The shaft is rotating at 1700 rpm. determine how much heat must be removed to maintain the bearing at constant temperature.11 An automobile crankshaft is 3.183 cm in diameter and 2.17? 7.8 At what temperature is the kinematic viscosity of glycerin the same as the kinematic viscosity of helium? 7. For a parabolic velocity proﬁle in a circular tube (see Example 4. 7.Problems 91 PROBLEMS 7.3 Show that the axial strain rate in a one-dimensional ﬂow. and determine the rate of volume change. A bearing on the shaft is 3. incompressible ﬂow with velocity v x ¼ v x (y). 7. Assuming that the shaft is centrally located in the bearing.02-cm-diameter ram that slides in a 36. The bearing is lubricated with SAE 30 oil at a temperature of 365 K.44 m of the ram engaged. the viscosity.11.17 For water ﬂowing in a 0. sketch a three-dimensional ﬂuid element and illustrate the magnitude. the gap distance.1 Sketch the deformation of a ﬂuid element for the following cases: (a) @v x /@y is much larger than @v y /@x.-diameter tube.175 cm in diameter. the velocity distribution is parabolic (see Example 4.2).14 An auto lift consists of 36. 7.10 Repeat the preceding problem for air.13 Two ships are traveling parallel to each other and are connected by ﬂexible hoses.14 m of the ram is engaged in the cylinder. 7. 7. 7. (b) @v y /@x is much larger than @v x /@y. and at a given instant the ﬁrst ship is making 4 m/s whereas the second ship is making 3. The annular region is ﬁlled with oil having a kinematic viscosity of 0. and the viscosity of the oil is 0. and surface of action of each stress component. v x ¼ v x (x).6 Calculate the viscosity of oxygen at 3508K and compare with the value given in Appendix I.12 If the speed of the shaft is doubled in Problem 7. is given by @v x /@x: What is the rate of volume change? Generalize for a three-dimensional element.18 What pressure drop per foot of tube is caused by the shear stress in Problem 7. w 7. Assume 2. the angular velocity. what will be the percentage increase in the heat transferred from the bearing? Assume that the bearing remains at constant temperature.7 What is the percentage change in the viscosity of water when the water temperature rises from 60 to 1208F? D 7. and the shaft diameter. estimate the maximum sinking speed of the ram and rack when gravity and viscous friction are the only forces acting.85. direction. Fluid is transferred from one ship to the other for processing and then returned.2).1 m/s. If the ﬂuid is ﬂowing at 100 kg/s. determine the magnitude of the shear stress at the tube wall. 7. show that Stokes’s viscosity relation yields the following shear stress components. What percentage change in volumetric ﬂow rate occurs in a laminar ﬂow as the water temperature changes from near freezing to 1408F? 2a h 7.19 The rate of shear work per unit volume is given by the product tv.01 Pa s.16 The conical pivot shown in the ﬁgure has angular velocity v and rests on an oil ﬁlm of uniform thickness h.04-cm-diameter cylinder.8 cm long.15 If the ram and auto rack in the previous problem together have a mass of 680 kg. the volumetric ﬂow rate is inversely proportional to the viscosity. determine the distance from the wall at which the shear work is maximum.9 According to the Hagen–Poiseuille laminar ﬂow model. @ v u 1 @v r t ru ¼ t ur ¼ m r þ @r r r @u @v u 1 @v z þ t uz ¼ t zu ¼ m @z r @u @v z @v r þ t zr ¼ t rz ¼ m @r @z 7.00037 m2/s and a speciﬁc gravity of 0. estimate the frictional resistance when 3. 7. 7.4 Using a cylindrical element. 7. .5 Estimate the viscosity of nitrogen at 175 K using equation (7-10).15 m/s.

r. but it can give much greater insight into the mechanisms of mass. we shall direct our attention to elements of ﬂuid as they approach differential size. we may 1 This transformation may be accomplished by a variety of methods. Applying Newton’s second law to this control volume. rather easily. We now consider the cylindrical control volume of ﬂuid having an inside radius. Such information may be of less interest to the engineer needing overall design information. it is not inﬂuenced by entrance effects and represents a steady-ﬂow situation. It is possible to change from one form of analysis to the other. one is concerned only with gross quantities of mass. Fully developed ﬂow is deﬁned as that for which the velocity proﬁle does not vary along the axis of ﬂow. the resulting expressions from such analyses will be differential equations. and length Dx.1. Our goal is the estimation and description of ﬂuid behavior from a differential point of view. A more general differential approach will be discussed in Chapter 9. and energy transfer. 8. momentum. Changes occurring within the control volume by each differential element of ﬂuid cannot be obtained from this type of overall analysis. that is. The solution to these differential equations will give ﬂow information of a different nature than that achieved from a macroscopic examination. from a differential analysis to an integral analysis by integration and vice versa. In this chapter. In analyzing a problem from the standpoint of a macroscopic control volume.Chapter 8 Analysis of a Differential Fluid Element in Laminar Flow The analysis of a ﬂuid-ﬂow situation may follow two different paths.1 A complete solution to the differential equations of ﬂuid ﬂow is possible only if the ﬂow is laminar. that is. We shall now analyze this situation for the case of incompressible laminar ﬂow. among which are the methods of vector calculus. thickness Dr. momentum. we have a section of pipe in which the ﬂow is laminar and fully developed. We shall use a limiting process in this text. One type of analysis has been discussed at length in Chapters 4–6 in which the region of interest has been a deﬁnite volume. In Figure 8. for this reason only laminar-ﬂow situations will be examined in this chapter. the macroscopic control volume. 92 . and energy crossing the control surface and the total change in these quantities exhibited by the material under consideration.1 FULLY DEVELOPED LAMINAR FLOW IN A CIRCULAR CONDUIT OF CONSTANT CROSS SECTION Engineers are often confronted with ﬂow of ﬂuids inside circular conduits or pipes.

respectively. we have ZZ c:s: åFx ¼ P(2pr Dr)jx P(2pr Dr)jxþDx þ t rx (2pr Dx)jrþDr trx (2pr Dx)jr v x r(v: n) dA ¼ (rv x )(2pr Drv x )jxþDx (rv x )(2pr Drv x )jx and @ @t ZZZ v x r dV ¼ 0 c:v: in steady ﬂow. the pressure gradient. evaluate the appropriate force and momentum terms for the x direction. by the original stipulation that ﬂow is fully developed. we ﬁnd that this expression reduces to the form r PjxþDx Pjx (rt rx )jrþDr (rt rx )jr þ ¼0 Dx Dr Evaluating this expression in the limit as the control volume approaches differential size. The convective momentum ﬂux (rv x )(2pr Drv x )jxþDx (rv x )(2p Drv x )jx is equal to zero as. dP/dx.1 Fully Developed Laminar Flow in a Circular Conduit of Constant Cross Section 93 ∆r P ⎜x x P ⎜x+∆x ∆x Figure 8. as Dx and Dr approach zero.8. we have r dP d þ (rtrx ) ¼ 0 dx dr (8-1) Note that the pressure and shear stress are functions only of x and r.1 Control volume for ﬂow in a circular conduit. The variables in equation (8-1) may be separated and integrated to give dP r C1 þ trx ¼ dx 2 r . that is. Starting with the control-volume expression for linear momentum in the x direction ZZZ ZZ @ åFx ¼ rv x (v: n) dA þ rv x dV (5-5a) @t c:s: c:v: and evaluating each term as it applies to the control volume shown. all terms are independent of x. is constant. In a region of fully developed flow. and thus the derivatives formed are total rather than partial derivatives. Substitution of the remaining terms into equation (5-5a) gives [P(2pr Dr)jxþDx P(2pr Dr)jx ] þ trx (2pr Dx)jrþDr trx (2pr Dx)jr ¼ 0 Canceling terms where possible and rearranging.

upon integration.2 that 2 v max dP R ¼ (8-8) v avg ¼ dx 8m 2 . will be inﬁnite. the only realistic value for C1 is zero. We may. 2 dP R C2 ¼ dx 4m and the velocity distribution becomes dP 1 2 (R r 2 ) dx 4m (8-4) 2 r 2 dP R 1 dx 4m R (8-5) vx ¼ or vx ¼ Equations (8-4) and (8-5) indicate that the velocity profile is parabolic and that the maximum velocity occurs at the center of the circular conduit where r ¼ 0. v x . use the result obtained in Example 4. C2. Further information may be obtained if we substitute the Newtonian viscosity relationship. vx ¼ dP r 2 þ C2 dx 4m The second constant of integration. Thus.2. that is. may be evaluated. using the boundary condition that the velocity. Thus. As this is physically impossible. assuming the ﬂuid to be Newtonian and recalling that the ﬂow is laminar t rx ¼ m dv x dr (8-3) Substituting this relation into equation (8-2) gives dv x dP r ¼ m dx 2 dr which becomes. r ¼ R. 2 dP R (8-6) v max ¼ dx 4m and equation (8-5) may be written in the form r 2 v x ¼ v max 1 R (8-7) Note that the velocity profile written in the form of equation (8-7) is identical to that used in Example 4. therefore. the inside surface of the conduit. r ¼ 0. Thus. where for any ﬁnite value of C1. to a maximum at r ¼ R. the shear stress. trx . Such a condition is known at the center of the conduit.94 Chapter 8 Analysis of a Differential Fluid Element in Laminar Flow The constant of integration C1 may be evaluated by knowing a value of t rx at some r. is zero at the conduit surface (the no-slip condition). the shear-stress distribution for the conditions and geometry speciﬁed is dP r (8-2) t rx ¼ dx 2 We observe that the shear stress varies linearly across the conduit from a value of 0 at r ¼ 0.

in terms of v avg dP 8mv avg 32mv avg ¼ ¼ (8-9) dx R2 D2 Equation (8-9) is known as the Hagen–Poiseuille equation. that of a Newtonian ﬂuid in laminar ﬂow down an inclined-plane surface.2 Laminar Flow of a Newtonian Fluid Down an Inclined-Plane Surface 95 Equation (8-8) may be rearranged to express the pressure gradient. 8. which is ZZZ ZZ @ åFx ¼ v x r(v n) dA þ rv x dV (5-5a) @t c:s: c:v: Evaluating each term in this expression for the fluid element of volume ðDxÞðDyÞð1Þ as shown in the figure. The ﬂuid (a) is Newtonian. in honor of the two men credited with its original derivation. This expression may be integrated over a given length of conduit to find the pressure drop and associated drag force on the conduit resulting from the flow of a viscous fluid.1 will now be applied to a slightly different situation. (d) incompressible. 2. We will examine the two-dimensional case. The ﬂow is (a) laminar. (c) fully developed.2 LAMINAR FLOW OF A NEWTONIAN FLUID DOWN AN INCLINED-PLANE SURFACE The approach used in Section 8. that is. The conditions for which the preceding equations were derived and apply should be remembered and understood.8. dP/dx. This conﬁguration and associated nomenclature are depicted in Figure 8. we have åFx ¼ P Dyjx P DyjxþDx þ t yx DxjyþDy t yx Dxjy þ rg Dx Dy sin u ZZ rv x (v: n) dA ¼ rv 2x DyjxþDx rv 2x Dyjx c:s: . (b) steady. (b) behaves as a continuum.2. y L x ∆y ∆x g q Figure 8.2 Laminar ﬂow down an inclined-plane surface. The analysis again involves the application of the control-volume expression for linear momentum in the x direction. They are as follows: 1. we consider no signiﬁcant variation in the z direction.

96 Chapter 8 Analysis of a Differential Fluid Element in Laminar Flow and @ @t ZZZ rv x dV ¼ 0 c:v: Noting that the convective-momentum terms cancel for fully developed ﬂow and that the pressure–force terms also cancel because of the presence of a free liquid surface. v x ¼ 0 at y ¼ 0. Additional calculations may be performed to determine the average velocity as was indicated in Section 8. is seen to be zero. the volume of the element considered. we see that the equation resulting from the substitution of these terms into equation (5-5a) becomes tyx DxjyþDy tyx Dxjy þ rg Dx Dy sin u ¼ 0 Dividing by (Dx)(Dy)(1). is zero at the free surface.1. upon separation of variables and integration. y ¼ L. may be evaluated by using the boundary condition that the shear stress. y ¼ L. to be made for tyx. gives t yx jyþDy tyx jy Dy þ rg sinu ¼ 0 In the limit as Dy ! 0. C2. The ﬁnal expression for the velocity proﬁle may now be written as rgL2 sinu y 1 y 2 vx ¼ (8-12) m L 2 L The form of this solution indicates the velocity variation to be parabolic. becomes rgL sinu y2 y vx ¼ þ C2 m 2L Using the no-slip boundary condition. The reason for this is the . the constant of integration. Note that there will be no counterpart in this case to the Hagen– Poiseuille relation. equation (8-9). yielding dv x rgL sinu h yi 1 ¼ m L dy which. for the pressure gradient. Thus the shear-stress variation becomes h yi (8-11) tyx ¼ rgL sinu 1 L The consideration of a Newtonian fluid in laminar flow enables the substitution of m(dv x /dy). reaching the maximum value v max ¼ rgL2 sinu 2m (8-13) at the free surface. that is. tyx . we get the applicable differential equation d t yx þ rg sinu ¼ 0 dy (8-10) Separating the variables in this simple equation and integrating we obtain for the shear stress tyx ¼ rg sinuy þ C1 The integration constant. C1.

3 CLOSURE The method of analysis employed in this chapter.273 cm3/s.6 A thin rod of diameter d is pulled at constant velocity through a pipe of diameter D.3 A 0.635-cm hydraulic line suddenly ruptures 8 m from a reservoir with a gage pressure of 207 kPa. 8.9 Determine the velocity proﬁle for ﬂuid ﬂowing between two parallel plates separated by a distance 2h.5 Derive the expressions for the velocity distribution and for the pressure drop for a Newtonian ﬂuid in fully developed laminar ﬂow in the annular space between two horizontal. what will be the new capacity of this network? Flow in both cases is laminar and the pressure drop remains 3:45 106 Pa. Oil 8 cm 55 cm 8. 8. where the element may be subjected to an energy or a mass balance. If oil of constant density ﬂows out of the viscosimeter shown at the rate of 0. what is the kinematic viscosity of the ﬂuid? The tube diameter is 0. that of applying the basic relation for linear momentum to a small control volume and allowing the control volume to shrink to differential size. such as oils. Apply the momentum theorem to an annular ﬂuid shell of thickness Dr and show that the analysis of such a control volume leads to d DP (rt) ¼ r dr L The desired expressions may then be obtained by the substitution of Newton’s viscosity law and two integrations.7 The viscosity of heavy liquids. Velocity and shear-stress proﬁles are examples of this type of information. 8. concentric pipes. .18 cm. The ﬂuid ﬁlling the space between the rod and the inner pipe wall has density r and viscosity m. for our present case.Problems 97 presence of a free liquid surface along which the pressure is constant. the methods introduced in this chapter will be used to derive differential equations of ﬂuid ﬂow for a general control volume. In Chapter 9. PROBLEMS 8. If the wire is at the center of the pipe. Compare the laminar and inviscid ﬂow rates from the ruptured line in cubic meters per second.1 Express equation (8-9) in terms of the ﬂow rate and the pipe diameter. The annular region between these cylinders is ﬁlled with liquid and the torque required to rotate the inner cylinder at constant speed is computed. This method has direct counterparts in heat and mass transfer. the rate of outﬂow being determined by timing the fall in the surface level.4 A common type of viscosimeter for liquids consists of a relatively large reservoir with a very slender outlet tube. The behavior of a ﬂuid element of differential size can give considerable insight into a given transfer process and provide an understanding available in no other type of analysis. 8. If a parallel line of the same size is laid along the last 18 km of the line.2 A 40-km-long pipeline delivers petroleum at a rate of 4000 barrels per day. For what ratio of cylinder diameters is the assumption of a linear proﬁle accurate within 1% of the true proﬁle? 8. ﬁnd the drag per unit length of wire. The resulting pressure drop is 3:45 106 Pa. what percentage change will occur in the ﬂow rate? 8. Thus. ﬂow is not the result of a pressure gradient but rather the manifestation of the gravitational acceleration upon a ﬂuid. The pressure drop is constant. 8. a linear velocity proﬁle being assumed. Express the boundary conditions at the interface between the two ﬂuids. If the pipe diameter is doubled at constant pressure drop. enables one to ﬁnd information of a sort different from that obtained previously.8 Two immiscible ﬂuids of different density and viscosity are ﬂowing between two parallel plates. 8. is frequently measured with a device that consists of a rotating cylinder inside a large cylinder.

and b. (a) State clearly the boundary conditions at y ¼ 0 and y ¼ h to be satisﬁed by the velocity.11 Derive the equation of motion for a one-dimensional. a distance h apart. The volumetric ﬂow rate per unit length is Q and the plates remain a constant distance. r. The pressure at B is higher than that at A. so that the ﬂow in the annulus is equivalent to the ﬂow between two ﬂat plates. The case and the drum are concentric. The width of the annulus h is very small compared to the diameter of the drum. Find the pressure rise and efﬁciency as a function of the ﬂow rate per unit depth. 8. r.15 Aviscous ﬁlm drains uniformly down the side of a vertical rod of radius R. The upper plate moves at velocity. ﬂows through the annulus between the case and the drum.14 Oil is supplied at the center of two long plates. determine the velocity distribution in the ﬁlm. (c) Determine the rate at which ﬂuid is being dragged up with the belt in terms of m.13 The device in the schematic diagram below is a viscosity pump.12 A continuous belt passes upward through a chemical bath at velocity v 0 and picks up a ﬁlm of liquid of thickness h. v 0 . density. the difference being Dr: The length of the annulus is L. the ﬁlm approaches a terminal or fully developed ﬂow such that the ﬁlm thickness. Fluid enters at A. and leaves at B. m. 8.16 Determine the maximum ﬁlm velocity in Problem 8. unsteady compressible ﬂow in a pipe of constant crosssectional area neglect gravity. and that the atmosphere produces no shear at the outer surface of the ﬁlm.98 Chapter 8 Analysis of a Differential Fluid Element in Laminar Flow 8. apart.10 Fluid ﬂows between two parallel plates. At some distance down the rod. L. is constant and v z ¼ f(r). h. b. v 0 : 8. and viscosity m. the lower plate is stationary. Gravity tends to make the liquid drain down. For what value of pressure gradient will the shear stress at the lower wall be zero? Assume the ﬂow to be laminar. inviscid. Neglecting the shear resistance due to the atmosphere. Assume that the ﬂow is a well-developed laminar ﬂow with zero pressure gradient. A 8.15. (b) Calculate the velocity proﬁle. B h R Ω Case Drum 8. h. Determine the vertical force per unit length as a function of the Q. . but the movement of the belt keeps the ﬂuid from running off completely. It consists of a rotating drum inside of a stationary case. 8.

may also be expressed in mathematical form for a special type of control volume. In Chapter 9. incompressible laminar ﬂows.1 Mass ﬂux through a differential control volume. z The control volume expression for the conservation of mass is ZZ ZZZ @ rðv nÞdA þ r dV ¼ 0 @t which states that (4-1) 8 9 8 9 <net rate of mass= <rate of accumulation= flux out of þ of mass within ¼0 : . These differential equations of ﬂuid ﬂow provide a means of determining the point-to-point variation of ﬂuid properties. steady.1. 9.1 THE DIFFERENTIAL CONTINUITY EQUATION The continuity equation to be developed in this section is the law of conservation of mass expressed in differential form. The basic tools used to derive these differential equations will be the control-volume developments of Chapters 4 and 5.Chapter 9 Differential Equations of Fluid Flow The fundamental laws of ﬂuid ﬂow. Consider the control volume Dx Dy Dz shown in Figure 9. the differential element. which have been expressed in mathematical form for an arbitrary control volume in Chapters 4–6. ruy ⎜y+∆y y ruz ⎜z rux ⎜x ruz ⎜z+∆z rux ⎜x+∆x ∆y ∆x ∆z x ruy ⎜y Figure 9. : . we shall express the law of conservation of mass and Newton’s second law of motion in differential form for more general cases. control volume control volume 99 . Chapter 8 involved the differential equations associated with some one-dimensional.

r ¼ r(x. In the limit as Dx. z. so we may divide the above equation by Dx Dy Dz. The divergence of a vector is the dot product with = div A = A The student may verify that the ﬁrst three terms in equation (9-1) may be written as = rv. when flow is incompressible.100 Chapter 9 Differential Equations of Fluid Flow The mass flux r(v n) at each face of the control volume is illustrated in Figure 9. Dy. three-dimensional flow. Substituting into equation (4-1) yields ðrv x jxþDx rv x jx ÞDy Dz þ ðrv y jyþDy rv y jy ÞDx Dz þ ðrv z jzþDz rv z jz ÞDx Dy þ @ ðr Dx Dy DzÞ ¼ 0 @t The volume does not change with time. that is. The net mass ﬂux out of the control volume in the x direction is ðrv x jxþDx rv x jx ÞDy Dz in the y direction ðrv y jyþDy rv y jy ÞDx Dz and in the z direction ðrv z jzþDz rv z jz ÞDx Dy The total net mass ﬂux is the sum of the above three terms. we have @r @r @r @r @v x @v y @v z þ vx þ vy þ vz þr þ þ ¼0 @t @x @y @z @x @y @z . and thus the time rate of change of mass within the control volume is @ ðr Dx Dy DzÞ @t The student is reminded that the density in general can vary from point to point. Carrying out the differentiation indicated in (9-1). this equation reduces to = v ¼ 0 (9-3) whether the flow is unsteady or not.1. and Dz approach zero. It is apparent that. Equation (9-2) may be arranged in a slightly different form to illustrate the use of the substantial derivative. and thus a more compact statement of the continuity equation becomes = rv þ @r ¼0 @t (9-2) The continuity equation above applies to unsteady. The mass within the control volume is r Dx Dy Dz. t). we obtain @ @ @ @r ðrv x Þ þ ðrv y Þ þ ðrv z Þ þ ¼0 @x @y @z @t (9-1) The first three terms comprise the divergence of the vector rv. y.

for instance. bearing only coincidental relationship to the air currents. symbolized as Dr/Dt. If. dy/dt. The substantial derivative will be applied to both scalar and vector variables in subsequent sections. giving dP @P dx @P dy @P dz @P ¼ þ þ þ dt @t dt @x dt @y dt @z (9-6) As a first approach. dz/dt are the x. falls. A second approach involves the pressure-measuring instrument housed in an aircraft which. is equal to the local derivative with respect to time @P/@t. we wish to evaluate the change in atmospheric pressure. P. The continuity equation may. the coefﬁcients dx/dt. which is.9. and v z. In this case. dy/dt. the total differential written in rectangular coordinates is dP ¼ @P @P @P @P dt þ dx þ dy þ dz @t @x @y @z where dx. v y.2 NAVIER–STOKES EQUATIONS The Navier–Stokes equations are the differential form of Newton’s second law of motion.2 Navier–Stokes Equations 101 The ﬁrst four terms of the above equation comprise the substantial derivative of the density. and z velocities of the aircraft. and they are arbitrarily chosen. dz/dt are those of the ﬂow and they may be designated v x. the instrument to measure pressure is located in a weather station. Thus. or ﬂy in any chosen x. three different approaches may be taken.1. This latter situation corresponds to the substantial derivative and the terms may be grouped as designated below dP DP @P @P @P @P ¼ ¼ þ vx þ vy þ vz dt Dt @t @x @y @z |ﬄﬄﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄﬄﬄ} |ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ} local rate of change rate of of pressure change of due to motion pressure (9-7) The derivative D/Dt may be interpreted as the time rate of change of a fluid or flow variable along the path of a fluid element. can be made to climb or descend. 9. y. z direction. . The rate of pressure change is obtained by dividing through by dt. and dz are arbitrary displacements in the x. and drifts as inﬂuenced by the ﬂow of air in which it is suspended. The third situation is one in which the pressure indicator is in a balloon that rises. fixed on Earth’s surface. respectively. where D @ @ @ @ ¼ þ vx þ vy þ vz Dt @t @x @y @z (9-4) in cartesian coordinates. of course. and for a fixed point of observation the total derivative. the coefficients dx/dt. Here the coefﬁcients dx/ dt. at the pilot’s discretion. dy. and z directions. dP/dt. y. y. thus be written as Dr þ r= v ¼ 0 Dt (9-5) When considering the total differential of a quantity. dz/dt are all zero. Consider the differential control volume illustrated in Figure 9. dy/dt.

Figure 9. in the prior case. Equation (5-4) can also be written RRR RR P F @/@t rv dV rvðv nÞ dA ¼ lim lim þ lim (9-8) Dx. and body forces such as that due to gravity.2 Forces acting on a differential control volume. each will be evaluated separately and then substituted into equation (5-4).Dz ! 0 Dx Dy Dz Dx Dy Dz Dx Dy Dz Dx.Dz ! 0 2 1 1 1 3 1 1 Sum of the external forces.Dy. z .Dy. The development may be further simpliﬁed by recalling that we have. The forces acting on the control volume are those due to 1 the normal stress and to the shear stress. divided by the volume of the control volume and taken the limit as the dimensions approach zero.2 y syy ⎜y+∆y tyx ⎜y+∆y txy ⎜x+∆x sxx ⎜x sxx ⎜x+∆x txy ⎜x tyx ⎜y syy ⎜y x z x szz ⎜z+∆z sxx ⎜x+∆x tzy ⎜z+∆z txz ⎜x+∆x tyz ⎜y+∆y syy ⎜y syy ⎜y+∆y tyz ⎜y tzy ⎜z tzx ⎜z+∆z szz ⎜z szz ⎜z+∆z tzx ⎜z txz ⎜x szz ⎜z sxx ⎜x y Figure 9.Dz ! 0 Dx.Dy.102 Chapter 9 Differential Equations of Fluid Flow The basic tool we shall use in developing the Navier–Stokes equations is Newton’s second law of motion for an arbitrary control volume as given in Chapter 5 ZZZ ZZ P @ rvðv nÞdA þ rvdV F¼ (5-4) @t c:s: c:v: which states that ( ) ( time rate of change ) sum of the external net rate of linear forces acting on the ¼ þ of linear momentum momentum efflux c:v: within the c:v: As the mathematical expression for each of the above terms is rather lengthy.

2 Navier–Stokes Equations 103 illustrates the various forces acting on the control volume.Dy. we obtain P Fx ¼ðs xx jxþDx s xx jx ÞDy Dz þ ðtyx jyþDy t yx jy ÞDx Dz þ ðt zx jzþDz tzx jz ÞDx Dy þ gx r Dx Dy Dz where gx is the component of the gravitational acceleration in the x direction. The net momentum ﬂux through 1 the control volume illustrated in Figure 9.Dz ! 0 Dx Dy Dz P Fz @txz @tyz @s zz lim ¼ þ þ þ rgz (9-11) @x @y @z Dx.9.Dy.Dz ! 0 Dx Dy Dz 2 Net momentum ﬂux through the control volume.Dy.3 is RR . In the limit as the dimensions of the element approach zero this becomes P Fx @s xx @tyx @tzx ¼ þ þ þ rgx (9-9) lim @x @y @z Dx.Dz ! 0 Dx Dy Dz Similar expressions are obtained for the force summations in the y and z directions P Fy @txy @s yy @tzy ¼ þ þ þ rgy lim (9-10) @x @y @z Dx. Summing the forces in the x direction.

Dy.Dy.Dz ! 0 Dx. rvv x jxþDx rvv x jx Dy Dz rvðv nÞdA ¼ lim lim Dx Dy Dz Dx Dy Dz Dx.Dz ! 0 þ ðrvv y jyþDy rvv y jy ÞDx Dz .

Dx Dy Dz rvv z jzþDz rvv z jz Dx Dy þ Dx Dy Dz @ @ @.

Dy.3 Momentum ﬂux through a differential control volume.Dz ! 0 @v @v @v þ r v x þv y þv z @x @y @z rvuy ⎜y + ∆y y rvuz ⎜z rvux ⎜x rvux ⎜x + ∆x rvuz ⎜z + ∆z x rvuy ⎜y z Figure 9. . rvv y þ ðrvv z Þ ¼ ðrvv x Þ þ @x @y @z (9-12) Performing the indicated differentiation of the right-hand side of equation (9-12) yields RR rvðv nÞdA @ @ @ lim ¼ v ðrv x Þ þ ðrv y Þ ðrv z Þ Dx Dy Dz @x @y @z Dx.

we obtain three differential equations that are the statements of Newton’s second law in the x. whereas the rate-of-change-ofmomentum terms are expressed as vectors. : @x @y @z RR 2 1 3 1 lim Dx. The time rate of 1 change of momentum within the control volume may be evaluated directly RRR @/@t vr dV ð@/@tÞrv Dx Dy Dz @ @v @r ¼ rv ¼ r þ v (9-14) ¼ lim Dx Dy Dz @t @t @t Dx Dy Dz Dx.Dy. upon substitution.Dz ! 0 lim ð9-9Þ ð9-10Þ ð9-11Þ (9-13) (9-14) It can be seen that the forces are expressed in components.Dy. and z directions @v x @v x @v x @v x @s xx @t yx @tzx r þ vx þ vy þ vz þ þ ¼ rgx þ @t @x @y @z @x @y @z @v y @v y @v y @v y @txy @s yy @t zy þ vx þ vy þ vz þ þ r ¼ rgy þ @t @x @y @z @x @y @z @v z @v z @v z @v z @t xz @t yz @s zz r þ vx þ vy þ vz þ þ ¼ rgz þ @t @x @y @z @x @y @z (9-15a) (9-15b) (9-15c) .Dz ! 0 We have now evaluated all terms in equation 9-8 1 1 8 9 @s xx @tyx @tzx > > > þ þ þrgx ex > > > > > > > @x @y @z > > > > > > > > > > P = < @txy @s yy @tzy F lim þ þ þrgy ey ¼ > > @x @y @z Dx.Dy. y.Dy. When the momentum terms are expressed as components. yields RR rvðv nÞdA @r @v @v @v þ r vx þ vy þ vz ¼ v lim @t @x @y @z Dx Dy Dz Dx.104 Chapter 9 Differential Equations of Fluid Flow The above term may be simplified with the aid of the continuity equation @r @ @ @ þ ðrv x Þ þ ðrv y Þ þ ðrv z Þ ¼ 0 @t @x @y @z which.Dy.Dz ! 0 rvðv nÞdA @r @v @v @v þ r vx þ vy þ vz ¼ v @t @x @y @z Dx Dy Dz RRR @/@t rvdV @v @r ¼r þv @t @t Dx Dy Dz Dx.Dz ! 0 (9-1) (9-13) 3 Time rate of change of momentum within the control volume.Dz ! 0 Dx Dy Dz > > > > > > > > > > > > @t > > @t @s yz xz zz > > > þ þ þrgz ez > .

. regardless of the nature of the stress rate-of-strain relation.. Navier. The sum of these two bracketed terms is the total acceleration. the terms on the left-hand side represent the time-rate of change of momentum. When the substantial derivative notation is used. Cambridge Phys. are used for the stress components. Focusing our attention on the left-hand terms in equation (9-15a). involves the time rate of change of v x at a point and is called the local acceleration. equations (7-13) and (7-14). M.2 Navier–Stokes Equations 105 It will be noted that in equations (9-15) above. @v x =@t. 398 (1822). G. If Stokes’s viscosity relations. Mem. Soc. de l’Acad. Me´moire sur les Lois du Mouvements des Fluides. 6. v y. C. the convective acceleration. Stokes. The reader may verify that the terms on the left-hand side of equations (9-15) are all of the form @ @ @ @ þ vx þ vy þ vz vi @t @x @y @z where v i ¼ v x.9. d. Sci. equations (9-15) become r Dv x @s xx @tyx @tzx ¼ rgx þ þ þ Dt @x @y @z (9-16a) r Dv y @txy @s yy @tzy ¼ rgy þ þ þ Dt @x @y @z (9-16b) r Dv z @t xz @t yz @s zz ¼ rgz þ þ þ Dt @x @y @z (9-16c) and Equations (9-16) are valid for any type of ﬂuid.. On the Theories of the Internal Friction of Fluids in Motion. The remaining terms involve the velocity change from point to point. As no assumptions 1 L. or v z. 8 (1845). Trans. The above term is the substantial derivative of v i. and the terms on the right-hand side represent the forces. equations (9-16) become Dv x @P @ 2 @v m= v þ = m ¼ rgx r (9-17a) þ = ðm=v x Þ @x @x 3 @x Dt Dv y @P @ 2 @v m= v þ = m ¼ rgy (9-17b) þ = ðm=v y Þ r @y @y 3 @y Dt and r Dv z @P @ 2 @v m= v þ = m ¼ rgz þ = ðm=v z Þ @z @z 3 @z Dt (9-17c) The above equations are called the Navier–Stokes1 equations and are the differential expressions of Newton’s second law of motion for a Newtonian fluid. that is. H. we see that @v x @v x @v x @v x @ @ @ @ þ vx þ vy þ vz þ vy þ vz þ vx vx ¼ @t @x @y @z @t @x @y @z |ﬄﬄﬄ{zﬄﬄﬄ} |ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ} local rate of change in of change v x due to motion of v x The ﬁrst term.

In an incompressible flow. Equations (9-17) thus become r 2 Dv x @P @ vx @ 2 vx @ 2vx ¼ rgx þm þ þ Dt @x2 @y2 @z2 @x 2 Dv y @ vy @ 2 vy @ 2vy @P ¼ rgy þm þ þ Dt @x2 @y2 @z2 @y 2 Dv z @P @ vz @2vz @2vz ¼ rgz þm r þ þ Dt @x2 @y2 @z2 @z r (9-18a) (9-18b) (9-18c) These equations may be expressed in a more compact form in the single vector equation r Dv ¼ r g =P þ m=2 v Dt (9-19) The above equation is the Navier–Stokes equation for an incompressible ﬂow. The Navier–Stokes equations are written in cartesian. cylindrical. the limitations of equation (9-19). the governing equation 2 Strictly speaking. vertical surfaces. As the development has been lengthy. incompressible ﬂow. the Navier–Stokes equation becomes r Dv ¼ r g =P Dt (9-20) which is known as Euler’s equation. 2. that being inviscid flow.106 Chapter 9 Differential Equations of Fluid Flow relating to the compressibility of the fluid have been made. these equations are valid for both compressible and incompressible flows.4. If the ﬂow is inviscid ðm ¼ 0Þ. This will be illustrated in Chapter 12. 3. Euler’s equation has only one limitation. = v ¼ 0. A situation for which a solution can be obtained is illustrated in Figure 9. If we consider the ﬂuid Newtonian and the ﬂow laminar. and other information of the type achieved in Chapter 8. let us review the assumptions and. as the turbulent stress is included in the momentum ﬂux term. pressure gradients. One surface. . Many situations are of sufﬁcient complexity to make the solution extremely difﬁcult and are beyond the scope of this text. The assumptions are 1. In our study of momentum transfer we shall restrict our attention to incompressible flow with constant viscosity. whereas the other is moving upward at a constant velocity v 0.4 shows the situation of an incompressible ﬂuid conﬁned between two parallel.2 All of the above assumptions are associated with the use of the Stokes viscosity relation. Figure 9. therefore. equation (9-19) is valid for turbulent ﬂow. constant viscosity. shown to the left. EXAMPLE 1 Equation (9-19) may be applied to numerous ﬂow systems to provide information regarding velocity variation. is stationary. and spherical coordinate forms in Appendix E. laminar ﬂow.

2 Navier–Stokes Equations 107 of motion is the Navier–Stokes equation in the form given by equation (9-19).4 Fluid between two vertical plates with This differential equation is separable. The reduction of each term in the vector equation into its applicable form is shown below. r Dv @v @v @v @v ¼r þ vx þ v y þv z ¼0 Dt @t @x @y @z r g ¼ rgey dP =P ¼ ey dy Fluid where dP/dy is constant.9. The ﬁrst integration yields the one on the left stationary and the other on the right moving dv y x dP þ rg ¼ C1 vertically upward with dx m dy velocity v 0 : Integrating once more. or zero. the terms may be added to yield the complete velocity proﬁle. using the boundary conditions that v y ¼ 0 at x ¼ 0.5. in equation (9-21). downward. . Equation (9-21) is valid whether v 0 is upward. which are 1 and 1 It is interesting to note. The ﬁrst term is the equation for a symmetric parabola. and v y ¼ v 0 at x ¼ L. the effect of the terms labeled 1 added. These results are indicated in Figure 9. The resulting velocity proﬁle obtained by superposing the two parts is shown in each case. In each case. and m=2 v ¼ m L d2 v y ey dx2 y The resulting equation to be solved is x u0 0 ¼ rg d2 v y dP þm 2 dx dy Figure 9. we obtain vy þ x2 dP rg ¼ C1 x þ C2 2m dy The integration constants may be evaluated. The constants thus become C1 ¼ v0 L dP þ rg L 2m dy and C2 ¼ 0 The velocity proﬁle may now be expressed as vy ¼ 1 dP x rg Lx x2 þ v 0 2m dy L |ﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ{zﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄﬄ} |{z} 1 1 (9-21) 2 1 2 . the second for a straight line.

6. or stationary. Streamlines Figure 9. causes the ﬂuid to move in circular streamlines with a velocity that is inversely proportional to the distance from the shaft.7.6 Rotating shaft in a ﬂuid. EXAMPLE 2 A rotating shaft. downward. Assuming that there is no slip between the ﬂuid and the shaft at the surface of the =P ¼ rg r . when using the coordinate system shown in Figure 9. as illustrated in Figure 9.5 Velocity proﬁles for one surface moving upward. u0 Euler’s equation may also be solved to determine velocity proﬁles. as will be shown in Chapter 10. where A is a constant. we may observe that the free surface is perpendicular to the pressure gradient. The vector properties of Euler’s equation are illustrated by the example below. As the pressure along the free surface will be constant. in which the form of the velocity proﬁle is given. Rearranging equation (9-20). Determination of the pressure gradient.108 Chapter 9 Differential Equations of Fluid Flow u0 + + = = = + Figure 9. therefore. Find the shape of the free surface if the ﬂuid can be considered inviscid. will enable us to evaluate the slope of the free surface. we have Dv (9-20) Dt The velocity v ¼ Aeu /r.

so dv Dv vR2 v2 R4 ¼ ¼ 2 ver ¼ 3 er r r dt fluid Dt This result could have been obtained in a more direct manner by observing that Dv/Dt is the local ﬂuid acceleration. we have vðRÞ ¼ vR ¼ A R and thus A ¼ vR2 and v¼ vR2 eu r The substantial derivative Dv/Dt may be evaluated by taking the total derivative dv vR2 vR2 deu ¼ 2 eu r_ þ r r dt dt where deu /dt ¼ u_er .8. and u˙ for the ﬂuid is v/r. which for this case is v 2 er /r. it can be seen that the free surface makes an angle b with the r axis so that tan b ¼ ¼ rv2 R4 r 3 rg v2 R4 gr 3 . shaft. The total derivative becomes dv vR2 vR2 _ u er ¼ 2 r_eu r r dt Now the ﬂuid velocity in the r direction is zero.2 Navier–Stokes Equations 109 ez w eq z r er q Figure 9.9. The pressure gradient becomes =P ¼ rgez þ r v2 R4 er r3 From Figure 9.7 Cylindrical coordinate system for rotating shaft and ﬂuid.

Thus. the use of streamline coordinates is extremely helpful.9. In integrating Euler’s equation. n. the substantial derivative of the velocity in streamline coordinates is Dv @v @v ¼ þv Dt @t @s (9-22) . 9.9 Streamline coordinates. n. The s direction is parallel to the streamline and the n direction is perpendicular to the streamline.3 BERNOULLI’S EQUATION Euler’s equation may be integrated directly for a particular case. z Following the form used in equations (9-6) to obtain the substantial derivative. y en v es Streamlines x Figure 9. and P ¼ P(s. Streamline coordinates s and n are illustrated in Figure 9. v ¼ v(s. t). we have dv @v @v @v ¼ þ s˙ þ n˙ dt @t @s @n As the velocity of the ﬂuid element has components s˙ ¼ v. t).8 Free-surface slope. The ﬂow and ﬂuid properties are functions of position and time. directed away from the instantaneous center of curvature. The substantial derivatives of the velocity and pressure gradients in equation (9-20) must be expressed in terms of streamline coordinates so that equation (9-20) may be integrated. n˙ ¼ 0.110 Chapter 9 Differential Equations of Fluid Flow z Free surface ∇P –rgez b rw2R4 r3 er r Figure 9. ﬂow along a streamline.

we have @v @ v2 @P es ds þ ds ds ¼ rg es ds r @t @s 2 @s (9-24) Selecting g to act in the y direction. The limitations are 1. therefore. steady ﬂow. we have g es ds ¼ g dy. the equation applies along a streamline. that the viscosity in some way will effect a change in internal energy. 2. we obtain @v @v @P @P es ds þ v es ds r es þ en es ds ¼ rg es ds @t @s @s @n or. 3. For steady incompressible flow. as the other assumptions were the same. inviscid ﬂow.9. We may note. Limitation 4 will be relaxed for certain conditions to be investigated in Chapter 10. equation (9-24) may be integrated to yield v2 P þ gy þ ¼ constant 2 r (9-25) which is known as Bernoulli’s equation. It is interesting to note that the constant internal energy assumption and the inviscid ﬂow assumption must be equivalent. incompressible ﬂow. Bernoulli’s equation was also developed in Chapter 6 from energy considerations for steady incompressible ﬂow with constant internal energy. 9. 4.4 Closure 111 The pressure gradient in streamline coordinates may be written as =P ¼ @P @P es þ en @s @n (9-23) Taking the dot product of equation (9-20) with es ds. and using equations (9-22) and (9-23). These equations may be subdivided into two special groups @r þ = rv ¼ 0 @t (9-26) (continuity equation) Inviscid flow r Dv ¼ rg =P Dt (9-27) (Euler’s equation) Incompressible. as @v/@s es ¼ @/@sðv es Þ ¼ @v/@s. viscous flow = r v¼0 Dv ¼ rg =P þ m=2 v Dt (9-28) (continuity equation) (9-29) (Navier–Stokes equation for incompressible flow) .4 CLOSURE We have developed the differential equations for the conservation of mass and Newton’s second law of motion.

18. 9. If v ¼ 20. the continuity equation is 1 @ 1 @v u ðrv r Þ þ ¼0 r @r r @u Show that @ @ @ vx þ vy þ vz @x @y @z may be written ðv =Þ.20 Determine the velocity proﬁle in a ﬂuid situated between two coaxial rotating cylinders. Using continuity. Using the continuity equation.11 Derive equation (2-3) from equation (9-27).18 Using the Navier–Stokes equations in Appendix E. tÞer þ v u ðr.10 Write equations (9-17) in component form for cartesian coordinates. from the Navier– Stokes equation. together with an equation of state and the energy equation. 2L). PROBLEMS 9. Let the inner cylinder have radius R1. For ﬂows at high velocity and small viscosity. 400[(y/L)2 e 2 x þ (x/L) ey ] fps. (a) if v u ¼ 0. 9. (a) 9. In component form. show that the ﬂuid volume change is zero. u. 9.14 Obtain the equations for a one-dimensional steady. 9. 9. unsteady. (c) a static ﬂuid will always start to move in the direction of decreasing pressure. it is not proper to delete the viscous term v=2 v. (b) in the absence of viscous forces the ﬂuid accelerates in the direction of decreasing pressure. for example. 9.1 Apply the law of conservation of mass to an element in a polar coordinate system and obtain the continuity equation for a steady. 9. obtain an expression for the velocity proﬁle between two ﬂat. it is possible to delete the inertia terms. solve Problem 8.17 Using the Navier–Stokes equations. tÞeu : Remember that the unit vectors have derivatives. then v r ¼ F(u)/r. the volume of the ﬂuid is constant.9 In a velocity ﬁeld where v ¼ determine the pressure gradient at the point (L.13.5 For ﬂow at very low speeds and with large viscosity (the so-called creeping ﬂows) such as occur in lubrication. compressible ﬂow in the x direction from the Navier– Stokes equations. u. show that the solution to the equation does not involve viscosity. evaluate the rate of density change. show that 9.000 fps at 100. then v u ¼ f (r). parallel plates.13. Explain this. viscous.000 ft.000 ft. where b ¼ 22.19 For the ﬂow described in Problem 8. The y axis is vertical. (b) if v r ¼ 0. and v r ¼ f (r). 9. let the outer cylinder have radius R2 and angular velocity V2. and angular velocity V1 . 9.13 Using the laws for the addition of vectors and equation (). and viscous force all lie in the same plane. (These equations. 9.3 In an incompressible ﬂow. ﬁnd the differential equation for a radial ﬂow in which v z ¼ v u ¼ 0. show that in the absence of gravity. 9. incompressible ﬂow.15 Obtain the equations for one-dimensional inviscid. .) 9.4 Find Dv/Dt in polar coordinates by taking the derivative of the velocity. the student should note the physical meaning of the substantial derivative and appreciate the compactness of the vector representation. two-dimensional. t).112 Chapter 9 Differential Equations of Fluid Flow In addition. compressible ﬂow. obtain the differential equation of motion if v u ¼ f (r.8 The atmospheric density may be approximated by the relation r ¼ r0 exp( y/b).7 Does the velocity distribution in Example 2 satisfy continuity? 9.2 In cartesian coordinates. 9.12 In polar coordinates. equation (9-29) comprises some 27 terms in cartesian coordinates. may be solved for the case of weak shock waves. = v ¼ 0. 9.) the ﬂuid acceleration. pressure force.16 Using the Navier–Stokes equations as given in Appendix E work Problems 8. 9.17 and 8.6 Using the Navier–Stokes equations and the continuity equation. Dv/Dt. (Hint: v ¼ v r ðr. What is the physical meaning of the term ðv =Þ? 9. the density is 64:4 lbm /ft3 and the ﬂow may be considered inviscid. Determine the rate at which the density changes with respect to body falling at v fps.

analytical solutions to the differential equations of ﬂuid ﬂow are possible.1. Although the element may deform. The subject of inviscid ﬂow has particular application in aerodynamics and hydrodynamics and general application to ﬂow about bodies—the so called external ﬂows. We have discussed the deformation previously in Chapter 7. by virtue of the absence of shear stress. in which. we shall introduce the fundamentals of inviscid ﬂow analysis.Dz. From Figure 10.Dy. In time Dt the element will move in the xy plane as shown. In addition to translation. the element may also deform and rotate. In this chapter. we see that vz ¼ 1 Dx.Chapter 10 Inviscid Fluid Flow An important area in momentum transfer is inviscid ﬂow.1 FLUID ROTATION AT A POINT Consider the element of ﬂuid shown in Figure 10.Dt ! 0 2 ! . Now let us focus our attention on the rotation of the element. 10.1. the orientation will be given by the average rotation of the line segments OB and OA or by denoting the rotation by d aþb vz ¼ dt 2 where the counterclockwise sense is positive.

1 Rotation of a ﬂuid element. arctanf ðv x jyþDy v x jy ÞDt =Dyg arctan v y jxþDx v y jx Dt =Dx þ Dt Dt lim y B b B A y a O x O t t + t A x Figure 10. 113 .

The use of this relation is the subject of the next section. The kinematic condition = v ¼ 0 is not the ﬁrst time we have encountered a kinematic relation that satisﬁes one of the fundamental physical laws of ﬂuid mechanics. For irrotational ﬂow = v ¼ 0. equation (9-29). the flow must be rotational. The signiﬁcance of ﬂuid rotation at a point may be examined by a different approach. incompressible ﬂow. in the limit. The law of conservation of mass for an incompressible ﬂow. if v x ¼ F(x. is also expressed as a kinematic relation. As the student may verify @v y @v x @v z @v y @v x @v z ex þ ey þ ez =v¼ @y @z @z @x @x @y and thus = v ¼ 2v (10-4) The vector = v is also known as the vorticity. = v ¼ 0. may also be written in the form r Dv ¼ =P þ rg m½= ð= vÞ Dt (9-29) It may be observed from the above equation that if viscous forces act on a fluid. y). 10.114 Chapter 10 Inviscid Fluid Flow which becomes. as can be seen from equation (10-4). In the xz and yz planes the rotation at a point is given by 1 @v x @v z vy ¼ @x 2 @z (10-2) 1 @v z @v y @z 2 @y (10-3) and vx ¼ The rotation at a point is related to the vector cross product of the velocity. 1 @v y @v x vz ¼ @y 2 @x (10-1) The subscript z indicates that the rotation is about the z axis. the continuity equation is = v ¼ @v x @v y þ ¼0 @x @y (9-3) Equation (9-3) indicates that v x and v y are related in some way so that @v x /@x ¼ ð@v y /@yÞ: Perhaps the easiest way to express this relation is by having v x and v y both related to the same function. Consider the function F(x.2 THE STREAM FUNCTION For a two-dimensional. then @v y @F ¼ @y @x Z or v y ¼ @F dy @x . The Navier–Stokes equation for incompressible ﬂow. When the rotation at a point is zero the ﬂow is said to be irrotational. y).

y)/@y]. v x and v y. y) thus represents the streamlines. dC ¼ 0. The differential equation that governs C is obtained by consideration of the ﬂuid rotation. . and thus if Figure 10. and thus equation (10-5) becomes vy dy ¼ (10-6) dx C¼constant v x The slope of the path C ¼ constant is seen to be the same as the slope of a streamline as discussed in Chapter 3. Figure 10. v. the selection of v x ¼ F(x. y). C. The physical signiﬁcance of C can be seen from the following considerations. at a point. if F(x. we may write @v y @ @C ¼ @x @y @y @C @y @ @C vy þ ¼0 @y @x or for this to be true in general vy ¼ @C @x Instead of having two unknowns.2 illustrates the streamlines and velocity components for ﬂow about an airfoil. we now have only one unknown. the total derivative is dC ¼ @C @C dx þ dy @x @y Also @C ¼ v y @x and @C ¼ vx @y and thus dC ¼ v y dx þ v x dy (10-5) Consider a path in the xy plane such that C ¼ constant: Along this path. The function C(x. As C ¼ C(x. y) ¼ (@C(x. The unknown.2 The Stream Function 115 Unfortunately. then vx ¼ As @v x /@x ¼ (@v y /@y). We can easily remove the integral sign if we make the original F(x.y) equal to the derivative of some function with respect to y. y) results in an integral for v y. In a two-dimensional ﬂow.10.2 Streamlines and the stream function. For example. vz ¼ ½ [(@v y /@x) (@v x /@y)]. C. is called the stream function.

v r jr¼a ¼ 0 or @C/@ujr¼a ¼ 0: 2. The magnitude of the velocity as r ! 1 is v 1 . The circle r ¼ a must be a streamline.3 Cylinder in a uniform ﬂow. the line u ¼ 0 must also be a streamline. 4. A stationary circular cylinder of radius a is situated in uniform. As the velocity normal to a streamline is zero. As there is cylindrical symmetry. for an incompressible. irrotational ﬂow pattern about a cylinder of inﬁnite length will be examined. As r ! 1 the velocity must be ﬁnite. y r u q a x Figure 10. Hence v u ju¼0 ¼ 0 or @C/@rju¼0 ¼ 0: 3. From symmetry. a constant. polar coordinates are employed. equation (10-7) becomes Laplace’s equation =2 C ¼ @2C @2C þ 2 ¼0 @x2 @y (10-8) 10. we obtain. In polar coordinates. parallel ﬂow in the x direction. These are as follows: 1. . IRROTATIONAL FLOW ABOUT AN INFINITE CYLINDER In order to illustrate the use of the stream function.1 equation (10-8) becomes @ 2 C 1 @C 1 @ 2 C þ þ ¼0 @r 2 r @r r 2 @u2 (10-9) where the velocity components v r and v u are given by vr ¼ 1 @C r @u vu ¼ @C @r (10-10) The solution for this case must meet four boundary conditions.3 INVISCID. 1 The operator =2 in cylindrical coordinates is developed in Appendix A. steady ﬂow 2vz ¼ @2C @2C þ 2 @x2 @y (10-7) When the flow is irrotational.116 Chapter 10 Inviscid Fluid Flow the velocity components v y and v x are expressed in terms of the stream function C. the inviscid.3. The physical situation is illustrated in Figure 10.

10.4

Irrotational Flow, the Velocity Potential

117

The solution to equation (10-9) for this case is

a2

Cðr; uÞ ¼ v 1 r sinu 1 2

r

The velocity components v r and v u are obtained from equation (10-10),

1 @C

a2

¼ v 1 cosu 1 2

vr ¼

r @u

r

(10-11)

(10-12)

and

vu ¼

@C

a2

¼ v 1 sinu 1 þ 2

@r

r

(10-13)

**By setting r ¼ a in the above equations, the velocity at the surface of the cylinder may be
**

determined. This results in

vr ¼ 0

and

v u ¼ 2v 1 sinu

(10-14)

**The velocity in the radial direction is, of course, zero, as the cylinder surface is a
**

streamline. The velocity along the surface is seen to be zero at u ¼ 0 and u ¼ 180 : These

points of zero velocity are known as stagnation points. The forward stagnation point is at

u ¼ 180 ; and the aft or rearward stagnation point is at u ¼ 0 : The student may verify

that each of the boundary conditions for this case are satisfied.

**10.4 IRROTATIONAL FLOW, THE VELOCITY POTENTIAL
**

In a two-dimensional irrotational ﬂow = v ¼ 0; and thus @v x /@y ¼ @v y ; /@x: The

similarity of this equation to the continuity equation suggests that the type of relation

used to obtain the stream function may be used again. Note, however, that the order of

differentiation is reversed from the continuity equation. If we let v x ¼ @f(x; y)/@x, we

observe that

@v y

@v x

@2f

¼

¼

@x@y

@y

@x

or

@ @f

vy ¼ 0

@x @y

**and for the general case
**

vy ¼

@f

@y

**The function f is called the velocity potential. In order for f to exist, the ﬂow must be
**

irrotational. As the condition of irrotationality is the only condition required, the velocity

potential can also exist for compressible, unsteady ﬂows. The velocity potential is

commonly used in compressible ﬂow analysis. Additionally, the velocity potential, f exists

for three-dimensional ﬂows, whereas the stream function does not.

118

Chapter 10

Inviscid Fluid Flow

**The velocity vector is given by
**

v ¼ v x ex þ v y ey þ v z ez ¼

@f

@f

@f

ex þ

ey þ

ez

@x

@y

@z

**and thus, in vector notation
**

v ¼ =f

(10-15)

**The differential equation defining f is obtained from the continuity equation. Considering
**

a steady incompressible flow, we have = v ¼ 0; thus, using equation (10-15) for v, we

obtain

= =f ¼ =2 f ¼ 0

(10-16)

**which is again Laplace’s equation; this time the dependent variable is f. Clearly, C and f
**

must be related. This relation may be illustrated by a consideration of isolines of C and f.

An isoline of C is, of course, a streamline. Along the isolines

@C

@C

dx þ

dy

dC ¼

@x

@y

or

vy

dy

¼

dx C¼constant v x

and

df ¼

Accordingly

@f

@f

dx þ

dy

@x

@y

dy/dxf¼constant ¼

dy

vx

¼

vy

dx df¼0

1

dy/dxC¼constant

(10-17)

**and thus C and f are orthogonal. The orthogonality of the stream function and the
**

velocity potential is a useful property, particularly when graphical solutions to

equations (10-8) and (10-16) are employed.

Figure 10.4 illustrates the inviscid, irrotational, steady incompressible ﬂow about an

inﬁnite circular cylinder. Both the streamlines and constant-velocity potential lines are

shown.

**Figure 10.4 Streamlines and constant
**

velocity potential lines for steady,

incompressible, irrotational, inviscid ﬂow

about a cylinder.

10.6

Utilization of Potential Flow

119

**10.5 TOTAL HEAD IN IRROTATIONAL FLOW
**

The condition of irrotationality has been shown to be of aid in obtaining analytical solutions

in ﬂuid ﬂow. The physical meaning of irrotational ﬂow can be illustrated by the relation

between the rotation or vorticity, = v, and the total head, P/r þ v 2 /2 þ gy: For an inviscid

ﬂow we may write

and

Dv

=P

¼g

Dt

r

(Euler’s equation)

2

Dv @v

v

v ð= vÞ

¼

þ=

2

Dt

@t

(Vector identity)

**As the gradient of the potential energy is g, Euler’s equation becomes, for incompressible
**

ﬂow

P v2

@v

=

þ þ gy ¼ v ð= vÞ :

(10-18)

r 2

@t

If the ﬂow is steady, it is seen from equation (10-18) that the gradient of the total head

depends upon the vorticity, = v: The vector ð= vÞ is perpendicular to the velocity

vector; hence, the gradient of the total head has no component along a streamline. Thus,

along a streamline in an incompressible, inviscid, steady ﬂow,

P v2

þ þ gy ¼ constant

r 2

(10-19)

**This is, of course, Bernoulli’s equation, which was discussed in Chapters 6 and 9. If the
**

ﬂow is irrotational and steady, equation (10-18) yields the result that Bernoulli’s equation is

valid throughout the ﬂow ﬁeld. An irrotational, steady, incompressible ﬂow, therefore, has a

constant total head throughout the ﬂow ﬁeld.2

**10.6 UTILIZATION OF POTENTIAL FLOW
**

Potential ﬂow has great utility in engineering for the prediction of pressure ﬁelds, forces, and

ﬂow rates. In the ﬁeld of aerodynamics, for example, potential ﬂow solutions are used to

predict force and moment distributions on wings and other bodies.

An illustration of the determination of the pressure distribution from a potential ﬂow

solution may be obtained from the solution for the ﬂow about a circular cylinder presented in

Section 10.3. From the Bernoulli equation

P v2

þ ¼ constant

r 2

(10-20)

**We have deleted the potential energy term in accordance with the original assumption of
**

uniform velocity in the x direction. At a great distance from the cylinder the pressure is

P1 , and the velocity is v 1, so equation (10-20) becomes3

Pþ

2

rv 2

rv 2

¼ P 1 þ 1 ¼ P0

2

2

(10-21)

A more general result, Crocco’s theorem, relates the vorticity to the entropy. Thus it can be shown that a

steady, inviscid, irrotational ﬂow, either compressible or incompressible, is isentropic.

3

The stagnation pressure as given in equation (10-21) applies to incompressible ﬂow only.

120

Chapter 10

Inviscid Fluid Flow

**where P0 is designated the stagnation pressure (i.e., the pressure at which the velocity
**

is zero). In accordance with equation (10-19), the stagnation pressure is constant

throughout the field in an irrotational flow. The velocity at the surface of the body is

v u ¼ 2v 1 sinu, thus the surface pressure is

P ¼ P0 2rv 21 sin2 u

(10-22)

**A plot of the potential flow pressure distribution about a cylinder is shown in Figure
**

10.5.

2

0

2

(P – P) / 1 ru2

q

P

u

1

–1

–2

–3

0

30

60

90

q

120

150

180

**Figure 10.5 Pressure distribution
**

on a cylinder in an inviscid,

incompressible, steady ﬂow.

**10.7 POTENTIAL FLOW ANALYSIS—SIMPLE PLANE FLOW CASES
**

In this section, a number of cases will be considered in which solutions are achieved for

two-dimensional, incompressible irrotational ﬂow. We begin with some very straightforward ﬂow situations.

Case 1. Uniform ﬂow in the x direction.

For a uniform ﬂow parallel to the x axis, with velocity v 1 ¼ constant; the stream

function and velocity potential relationships are

vx ¼ v1 ¼

vy ¼ 0 ¼

@ C @f

¼

@y

@x

@ C @f

¼

@x

@y

**which integrate to yield
**

C ¼ v1y

f ¼ v1x

Case 2. A line source or sink.

A line source, in two dimensions, is a ﬂow which is radially outward from the source

which is the origin in this example. The reverse, or sink ﬂow, has the ﬂow directed inward.

10.8

Potential Flow Analysis—Superposition

121

The source strength is the volume ﬂow rate per unit depth, Q ¼ 2prv r : The radial velocity

associated with a source is

Q

vr ¼

2pr

and the azimuthal velocity is given by v u ¼ 0: The stream function and velocity potential

are evaluated from the expressions

Q

1 @C @f

¼

¼

vr ¼

2pr r @u

@r

@C 1 @f

¼

vu ¼ 0 ¼

@r

r @u

Integrating these expressions, we obtain for the line source

Q

u

C¼

2p

Q

ln r

f¼

2p

For sink ﬂow, the sign of the radial velocity is negative (inward) and thus Q is negative.

The expressions for a line source or sink present a problem at r ¼ 0, the origin, which

is a singular point. At r ¼ 0, the radial velocity approaches inﬁnity. Physically this is

unrealistic, and we use only the concept of line source or sink ﬂow under conditions where

the singularity is excluded from consideration.

Case 3. A line vortex.

Vortex ﬂow is that which occurs in a circular fashion around a central point, such as a

whirlpool. A free vortex is one where ﬂuid particles are irrotational, i.e., they do not rotate as they

move in concentric circles about the axis of the vortex. This would be analogous to people sitting

in cabins on a ferris wheel. For an irrotational ﬂow in polar coordinates (see Appendix B), the

product rv u must be constant. The stream function and velocity potential can be written directly,

1 @C @f

vr ¼ 0 ¼

¼

r @u

@r

K

@C 1 @f

¼

¼

vu ¼

2pr

@r

r @u

which, upon integration, become

C¼

f¼

K

ln r

2p

K

u

2p

**where K is referred to as the vortex strength. When K is positive, the ﬂow is observed to be
**

counterclockwise about the vortex center.

**10.8 POTENTIAL FLOW ANALYSIS—SUPERPOSITION
**

It was shown earlier that both the stream function and the velocity potential satisfy Laplace’s

equation for two-dimensional, irrotational, incompressible ﬂow. As Laplace’s equation is

linear we can use known solutions to achieve expressions for both C and w for more complex

situations using the principle of superposition. Superposition, simply put, is the process of

adding known solutions to achieve another, i.e., if C1 and C2 are solutions to =2 C ¼ 0, then

so is C3 ¼ C1 þ C2 a solution.

122

Chapter 10

Inviscid Fluid Flow

**The reader is reminded that the solutions obtained for these very specialized ﬂow
**

conditions are idealizations. They apply for inviscid ﬂow which is a reasonable approximation for conditions outside the region, near a solid body, where viscous effects are

manifested. This region, the boundary layer, will be considered in some depth in Chapter 12.

Some cases will now be considered where the elementary plane ﬂows of the previous

section give some interesting and useful results through the process of superposition.

Case 4. The doublet.

A useful case is achieved from considering a source-sink pair on the x axis as the

separation distance, 2a, approaches zero. Geometrically, we can note that the streamlines

and velocity potential lines are circles with centers on the y and x axes but with all circles

passing through the origin that is a singular point.

The strength of a double, designated l, is deﬁned as the ﬁnite limit of the quantity

2aQ as a ! 0. For our case, the source is placed on the x axis at a and the sink is placed on

the x axis at þa. The resulting expressions for C and f in polar coordinates are

l sinu

r

l cosu

f¼

r

C¼

**Case 5. Flow past a half body—superposition of uniform ﬂow and a source.
**

The stream function and velocity potentials for uniform ﬂow in the x direction and for a

line source are added together, yielding

C ¼ Cuniform flow þ Csource

Q

Q

u

¼ v 1 y þ u ¼ v 1 r sinu þ

2p

2p

f ¼ funiform flow þ fsource

Q

Q

ln r ¼ v 1 r cosu þ ln r

¼ v1x þ

2p

2p

Case 6. Flow past a cylinder—superposition of uniform ﬂow and a doublet.

As a ﬁnal illustration of the superposition method, we will consider a case of

considerable utility. When the solutions for uniform ﬂow and the doublet are superposed,

the result, similar to the past case, deﬁnes a streamline pattern inside and around the outside

surface of a body. In this case the body is closed and the exterior ﬂow pattern is that of ideal

ﬂow over a cylinder. The expressions for C and w are

C ¼ Cuniform flow þ Cdoublet

l sinu

l sinu

¼ v 1 r sinu

¼ v1y

r

r

l

¼ v 1 r sinu

r

f ¼ funiform flow þ fdoublet

l cosu

l cosu

¼ v 1 r cosu þ

¼ v1x þ

r

r

l

¼ v 1 r þ cosu

r

10.9

Closure

123

**It is useful, at this point, to examine the above expressions in more detail. First, for the
**

stream function

l

C ¼ v 1 r sinu

r

l/v 1

¼ v 1 r 1 2 sinu

r

where, as we recall, l is the doublet strength. If we choose l such that

l

¼ a2

v1

where a is the radius of our cylinder, we obtain

a2

Cðr; uÞ ¼ v 1 r sinu 1 2

r

which is the expression used earlier, designated as equation (10-11).

10.9 CLOSURE

In this chapter, we have examined potential ﬂow. A short summary of the properties of the

stream function and the velocity potential is given below.

Stream function

1. A stream function C(x; y) exists for each and every two-dimensional, steady,

incompressible ﬂow, whether viscous or inviscid.

2. Lines for which C(x; y) ¼ constant are streamlines.

3. In cartesian coordinates

@C

@C

vx ¼

vy ¼

@y

@x

(10-23a)

and in general

vs ¼

@C

@n

(10-23b)

**where n is 908 counterclockwise from s.
**

4. The stream function identically satisﬁes the continuity equation.

5. For an irrotational, steady incompressible ﬂow

=2 C ¼ 0

(10-24)

Velocity potential

1. The velocity potential exists if and only if the ﬂow is irrotational. No other

restrictions are required.

2. =f ¼ v:

3. For irrotational, incompressible ﬂow, =2 f ¼ 0:

4. For steady, incompressible two-dimensional ﬂows, lines of constant velocity

potential are perpendicular to the streamlines.

124

Chapter 10

Inviscid Fluid Flow

PROBLEMS

10.1

In polar coordinates, show that

1 @ðrv u Þ @v r

ez

=v¼

r

@r

@u

For this ﬂow ﬁeld, plot several streamlines for 0

u

p=3:

10.12 For the case of a source at the origin with a uniform freestream plot the streamline c ¼ 0.

10.2 Determine the ﬂuid rotation at a point in polar coordinates, using the method illustrated in Figure 10.1.

**10.13 In Problem 10.12, how far upstream does the ﬂow from
**

the source reach?

10.3 Find the stream function for a ﬂow with a uniform freestream velocity v 1 : The free-stream velocity intersects the x axis

at an angle a.

**10.14 Determine the pressure gradient at the stagnation point
**

of Problem 10.10(a).

**10.4 In polar coordinates, the continuity equation for steady
**

incompressible ﬂow becomes

1 @

1 @v u

¼0

ðrv r Þ þ

r @r

r @u

Derive equations (10-10), using this relation.

10.5 The velocity potential for a given two-dimensional ﬂow

ﬁeld is

5 3

f¼

x 5xy2

3

Show that the continuity equation is satisﬁed and determine the

corresponding stream function.

10.6 Make an analytical model of a tornado using an irrotational vortex (with velocity inversely proportional to distance

from the center) outside a central core (with velocity directly

proportional to distance). Assume that the core diameter is 200 ft

and the static pressure at the center of the core is 38 psf below

ambient pressure. Find

(a) the maximum wind velocity;

(b) the time it would take a tornado moving at 60 mph to lower

the static pressure from 10 to 38 psfg;

(c) the variation in stagnation pressure across the tornado.

Euler’s equation may be used to relate the pressure gradient in the core to the ﬂuid acceleration.

10.7 For the ﬂow about a cylinder, ﬁnd the velocity variation

along the streamline leading to the stagnation point. What is the

velocity derivative @v r /@r at the stagnation point?

10.8 In Problem 10.7, explain how one could obtain @v u /@u at

the stagnation point, using only r and @v r /@r:

10.9 At what point on the surface of the circular cylinder in a

potential ﬂow does the pressure equal the free-stream pressure?

10.10 For the velocity potentials given below, ﬁnd the stream

function and sketch the streamlines

x 3 3xy2

3

(a) f ¼ v 1 L

L

L

xy

(b) f ¼ v 1

L

v1 L

2

(c) f ¼

ln x þ y2 :

2

10.11 The stream function for an incompressible, two-dimensional ﬂow ﬁeld is

c ¼ 2r 3 sin 3u

**10.15 Calculate the total lift force on the Arctic hut shown
**

below as a function of the location of the opening. The lift force

results from the difference between the inside pressure and the

outside pressure. Assume potential ﬂow and the hut is in the

shape of a half-cylinder.

0pening

u

q

**10.16 Consider three equally spaced sources of strength m
**

placed at (x, y) ¼ ( a, 0),(0, 0), and (a, 0). Sketch the resulting

streamline pattern. Are there any stagnation points?

10.17 Sketch the streamlines and potential lines of the ﬂow due

to a line source of at (a, 0) plus an equivalent sink at (a, 0).

10.18 The stream function for an incompressible, twodimensional ﬂow ﬁeld is

c ¼ 3x2 y þ y

For this ﬂow ﬁeld, sketch several streamlines.

10.19 A line vortex of strength K at (x, y) ¼ (0, a) is combined

with opposite strength vortex at (0, a): Plot the streamline pattern

and ﬁnd the velocity that each vortex induces on the other vortex.

10.20 A source of strength 1:5 m2 /s at the origin is combined

with a uniform stream moving at 9 m/s in the x direction. For the

half-body which results, ﬁnd

(a) the stagnation point;

(b) the body height as it crosses the y axis;

(c) the body height at large x;

(d) the maximum surface velocity and its position (x, y).

10.21 When a doublet is added to a uniform stream so that the

source part of the doublet faces the stream, a cylinder ﬂow

results. Plot the streamlines when the doublet is reversed so that

the sink faces the stream.

10.22 A 2-m-diameter horizontal cylinder is formed by bolting

two semicylindrical channels together on the inside. There are

12 bolts per meter of width holding the top and bottom together.

The inside pressure is 60 kPa (gage). Using potential theory for

the outside pressure, compute the tension force in each bolt if the

free stream ﬂuid is sea-level air and the free-stream wind speed is

25 m/s.

Chapter

11

Dimensional Analysis

and Similitude

An important consideration in all equations written thus far has been dimensional

homogeneity. At times it has been necessary to use proper conversion factors in order

that an answer be correct numerically and have the proper units. The idea of

dimensional consistency can be used in another way, by a procedure known as

dimensional analysis, to group the variables in a given situation into dimensionless

parameters that are less numerous than the original variables. Such a procedure is very

helpful in experimental work in which the very number of signiﬁcant variables presents

an imposing task of correlation. By combining the variables into a smaller number of

dimensionless parameters, the work of experimental data reduction is considerably

reduced.

This chapter will include means of evaluating dimensionless parameters both in

situations in which the governing equation is known, and in those in which no equation

is available. Certain dimensionless groups emerging from this analysis will be familiar,

and some others will be encountered for the ﬁrst time. Finally, certain aspects of

similarity will be used to predict the ﬂow behavior of equipment on the basis of

experiments with scale models.

11.1 DIMENSIONS

In dimensional analysis, certain dimensions must be established as fundamental, with all

others expressible in terms of these. One of these fundamental dimensions is length,

symbolized L. Thus, area and volume may dimensionally be expressed as L2 and L3,

respectively. A second fundamental dimension is time, symbolized t. The kinematic

quantities, velocity and acceleration, may now be expressed as L/t and L/t2, respectively.

Another fundamental dimension is mass, symbolized M. An example of a quantity

whose dimensional expression involves mass is the density that would be expressed as

M/L3. Newton’s second law of motion gives a relation between force and mass and

allows force to be expressed dimensionally as F ¼ Ma ¼ ML/t2 . Some texts reverse this

procedure and consider force fundamental, with mass expressed in terms of F, L, and t

according to Newton’s second law of motion. Here, mass will be considered a fundamental

unit.

The signiﬁcant quantities in momentum transfer can all be expressed dimensionally in

terms of M, L, and t; thus these comprise the fundamental dimensions we shall be concerned

with presently. The dimensional analysis of energy problems in Chapter 19 will require the

addition of two more fundamental dimensions, heat and temperature.

125

126

Chapter 11

Dimensional Analysis and Similitude

**Some of the more important variables in momentum transfer and their dimensional
**

representations in terms of M, L, and t are given in Table 11.1.

Table 11.1 Important variables in momentum transfer

Variable

Mass

Length

Time

Velocity

Gravitational acceleration

Force

Pressure

Density

Viscosity

Surface tension

Sonic velocity

Symbol

Dimension

M

L

t

v

g

F

P

r

m

s

a

M

L

t

L/t

L/t2

ML/t2

M/Lt2

M/L3

M/Lt

M/t2

L/t

**11.2 DIMENSIONAL ANALYSIS OF GOVERNING DIFFERENTIAL EQUATIONS
**

The differential equations that describe ﬂuid behavior as developed in Chapter 9 are

powerful tools for analyzing and predicting ﬂuid phenomena and their effects. The Navier–

Stokes equations have been solved analytically for a few simple situations. For more

complex applications, these relationships provide the basis for a number of sophisticated

and powerful numerical codes.

In this section, we will use the differential forms of the continuity and momentum

(Navier–Stokes) equations to develop some useful dimensionless parameters that will be

valuable tools for subsequent analysis. This process will now be illustrated as we examine

two-dimensional incompressible ﬂow.

The governing differential equations are the following.

Continuity:

@v x @v y

þ

¼0

(9-3)

@x

@y

Momentum:

2

@v

@v

@v

@ v @2v

r

þ

þ vx

þ vy

¼ rg =r þ m

@t

@x

@y

@x2 @y2

(9-19)

We now stipulate the reference values for length and velocity

reference length

L

reference velocity v 1

and, accordingly, specify nondimensional quantities for the variables in equations (9-3) and

(9-19) as

x
¼ x/L

v
x ¼ v x /v 1

y
¼ y/L

v
y ¼ v y /v 1

tv 1

v
¼ v/v 1

t
¼

L

=
¼ L =

11.2

Dimensional Analysis of Governing Differential Equations

127

**The last quantity in this list, =
, is the dimensionless gradient operator. As = is composed
**

of ﬁrst derivatives with respect to space coordinates, the product L= is seen to be

dimensionless.

The next step is to nondimensionalize our governing equations by introducing the

speciﬁed dimensionless variables. This process involves the chain rule for differentiation;

for example, the two terms in equation (9-3) are transformed as follows:

@v x @v
x @v x @x
@v
x

v 1 @v
x

¼

¼
ðv 1 Þð1=LÞ ¼

@x

@x @v x @x

@x

L @x

@v y @v
y @v y @y
v 1 v
x

¼

¼

@y

@y @v y @y

L @x

Substitution into equation (9-3) gives

@v
x @v y

þ

¼0

@x
@y

(11-1)

**and we see that the continuity equation has the same form in terms of dimensionless
**

variables as it had originally.

Utilizing the chain rule in the same manner as just discussed, the equation of motion

becomes

rv 21 @v

1

mv 1 @ 2 v
@ 2 v

@v

@v

(11-2)

þ v x
þ v y
¼ rg þ = P þ 2

þ

L

@t

@x

@y

L

@x
2 @y
2

L

In equation (11-2), we note that each term has the units M/L2 t2 or F/L3. Also, it

should be observed that each term represents a certain kind of force, that is

rv 21

is an inertial force

L

mv 21

is a viscous force

L

rg is a gravitational force

P/L is a pressure force

**If we next divide through by the quantity, rv 21 /L, our dimensionless equation becomes
**

2

@v

L

=
P

m

@ v

@ 2 v

@v

@v

(11-3)

þ vx
þ vy
¼ g 2 2 þ

þ

@t

@x

@y

v 1 rv 1 Lv 1 r @x
2 @y
2

This resulting dimensionless equation has the same general characteristics as its original

except that, as a result of its transformation into dimensionless form, each of the original force

terms (those on the right-hand side) has a coefficient composed of a combination of variables.

An example of these coefficients reveals that each is dimensionless. Additionally, because of

the manner in which they were formed, the parameters can be interpreted as a ratio of forces.

Consideration of the ﬁrst term, gL/v 21 , reveals that it is, indeed, dimensionless. The

choice of gL/v 21 or v 21 /gL is arbitrary; clearly both forms are dimensionless.

The conventional choice is the latter form. The Froude number is deﬁned as

Fr v 21 /gL

(11-4)

**This parameter can be interpreted as a measure of the ratio of inertial to gravitational
**

forces. The Froude number arises in analyzing flows involving a free liquid surface. It is

an important parameter when dealing with open-channel flows.

many situations of interest in which the governing equation is not known. (11-8) . is the coefficient of drag CD ¼ F/A rv 21 =2 (11-6) which. It is then necessary to determine the number of dimensionless parameters into which the variables may be combined. has application to both internal and external flows. we discuss a more general approach for generating dimensionless groups of variables. If direct solution is not possible then one must resort to numerical modeling or experimental determination of these functional relationships. also clearly dimensionless. where r is the rank of the dimensional matrix of the variables. 2. we will see directly. i¼nr where i ¼ the number of independent dimensionless groups n ¼ the number of variables involved and r ¼ the rank of the dimensional matrix 1 E. the results will provide the functional relationships between applicable dimensionless parameters. The third dimensionless ratio that has been generated is the Reynolds number. Buckingham. however. which states The number of dimensionless groups used to describe a situation involving n variables is equal to n – r. 345 (1914). 11. In these cases. is observed to be the ratio of pressure forces to inertial forces. The initial step in applying the Buckingham method requires the listing of the variables signiﬁcant to a given problem. we need an alternative method for dimensional analysis. which is conventionally expressed as Re Lv 1 r=m (11-7) In this form the Reynolds number is observed to represent the ratio of inertial forces to viscous forces. quite powerful when one knows the differential equation that pertains to a speciﬁc ﬂuid ﬂow process. In this form it is designated the Euler number. It is ubiquitous in all of the transport processes. This number may be determined using the Buckingham pi theorem. We will encounter it frequently throughout the remainder of this text. If equation (11-3) can be solved. It is generally referred to as the Buckingham method. In this section.128 Chapter 11 Dimensional Analysis and Similitude The next parameter. There are. This procedure was proposed by Buckingham1 in the early part of the twentieth century.3 THE BUCKINGHAM METHOD The procedure introduced in the previous section is. Phys. Thus. obviously. Eu P=rv 21 (11-5) A modified form of equation (11-5). Rev. The Reynolds number is generally considered the most important dimensionless parameter in the field of fluid mechanics. P/rv 21 .

L. r. hence the exponents 1. and t among them. Let us arbitrarily let the viscosity be the other exclusion from the core. The rank is 3 in this case. hence it will not be in the core. which. Initially. For example. The matrix is then the array of numbers shown below 0 1 1 0 1 1 0 @ 1 1 3 1 1 A 2 1 0 1 0 The rank. and t. and L (a signiﬁcant dimension of the body). The dimensional matrix that applies is formed from the following tabulation: F v r m L M 1 0 1 1 0 L 1 1 3 1 1 2 1 0 1 0 t The numbers in the table represent the exponents of M. .11. EXAMPLE 1 Determine the dimensionless groups formed from the variables involved in the ﬂow of ﬂuid external to a solid body. The two dimensionless parameters will be symbolized p1 and p2 and may be formed in several ways. L. as well as the application of the Buckingham method. r. and L. and t. r. the number of dimensionless parameters to be formed may be found by applying equation (11-4). among them. follows. Our core group now consists of the remaining variables v. Thus. respectively. In the present problem it would be desirable to have the drag force in only one dimensionless group. The force exerted on the body is a function of v. L. A usual ﬁrst step is to construct a table of the variables and their dimensions. a core group of r variables must be chosen. contain all of the fundamental dimensions. One way to choose a core is to exclude from it those variables whose effect one desires to isolate. m. we must know r. include M. which appear in each of the variables involved. and t in the dimensional expression for each variable involved. L. we observe. An example of the evaluation of r and i. and 2 are tabulated versus M. the dimensional expression of F is ML /t2 . which will consist of those variables that will appear in each pi group and. Variable Force Velocity Density Viscosity Length Symbol F v r m L Dimensions ML/t2 L/t M/L3 M/Lt L Before determining the number of dimensionless parameters to be formed.3 The Buckingham Method 129 The dimensional matrix is simply the matrix formed by tabulating the exponents of the fundamental dimensions M. In this example i ¼ 5 3 ¼ 2. the quantities with which they are associated. of a matrix is the number of rows (columns) in the largest nonzero determinant that can be formed from it. 1.

L. e ¼ 1 and f ¼ 1. we have. that one of them includes F and the other m. and t on both sides of this expression. and v . Writing p1 ¼ v a rb Lc F and p2 ¼ v d re L f m we shall evaluate the exponents as follows. In order that each be dimensionless. in dimensional form 1¼ d e L M M (L) f t L3 Lt and for exponents of M 0¼eþ1 for L 0 ¼ d 3e þ f 1 and for t 0 ¼ d 1 giving d ¼ 1. b ¼ 1. and that they are both dimensionless. Considering each p group independently. and c ¼ 2. for our second dimensionless group we have p2 ¼ m/rvL ¼ 1/Re Dimensional analysis has enabled us to relate the original ﬁve variables in terms of only two dimensionless parameters in the form Eu ¼ f(Re) (11-9) CD ¼ f (Re) (11-10) .130 Chapter 11 Dimensional Analysis and Similitude We now know that p1 and p2 both include r. we write p1 ¼ v a rb Lc F and dimensionally M 0 L0 t0 ¼ 1 ¼ a b L M ML (L)c 2 t L3 t Equating exponents of M. the variables must be raised to certain exponents. L. giving p1 ¼ F L2 rv 2 ¼ F/L2 ¼ Eu rv 2 Similarly for p2 we have. for M 0¼bþ1 for L 0 ¼ a 3b þ c þ 1 and for t 0 ¼ a 2 From these we ﬁnd that a ¼ 2. Thus.

2 lists several dimensionless groups that pertain to ﬂuid ﬂow. Cf P/rv 2 F/A rv 2 /2 Pressure Force Inertial force Flows involving pressure differences due to frictional effects Froude number. Three types of similarity are important in this regard.4 Geometric.2 is in using experimental results obtained using models to predict the performance of full-size prototypical systems. Fr v 2 /gL Inertial force Gravitational force Flows involving free liquid surfaces Weber number. kinematic. Geometric similarity exists between two systems if the ratio of all signiﬁcant dimensions is the same for each system.1 is equal in magnitude to the ratio a/b for the larger section. Kinematic. and Dynamic Similarity 131 The two parameters. For example. they are geometrically similar. geometric similarity would be achieved when all geometric ratios between model and prototype are equal. KINEMATIC. they are geometric. The validity of such scaling requires that the models and prototypes possess similarity.11. The functions f(Re) and f (Re) must be determined by experiment. In this example. We rv 2 L s Inertial force Surface tension force Flows with signiﬁcant surface tension effects Mach number. Re Lvr/m Inertial force Viscous force Widely applicable in a host of ﬂuid ﬂow situations Euler number. AND DYNAMIC SIMILARITY An important application and use of the dimensionless parameters listed in Table 11. . For more complex geometries. Similar tables will be include in later chapters that list dimensionless parameters common to heat transfer and to mass transfer. there are only two signiﬁcant dimensions. and dynamic similarity.4 GEOMETRIC. Table 11. 1 a b 2 Figure 11. Eu and CD. Table 11. M v/C Inertial force Compressibility force Flows with signiﬁcant compressibility effects Area of application 11.2 Common dimensionless parameters in momentum transfer Name/Symbol Dimensionless group Physical meaning Reynolds number.1 Two geometrically similar objects. if the ratio a/b for the diamond-shaped section in Figure 11. were also generated in the previous section by an alternate method. Eu Coefﬁcient of skin friction.

it is customary to build small models geometrically similar to the larger prototypes. dynamic similarity. EXAMPLE 2 A cylindrical mixing tank is to be scaled up to a larger size such that the volume of the larger tank is ﬁve times that of the smaller one.1 requires that Da Db ¼ ha hb Db or Da hb Db ¼ ha Da ha The volumes of the two tanks are Va ¼ p 2 D ha 4 a and Vb ¼ p 2 D hb 4 b A The scaling ratio between the two is stipulated as. What will be the ratios of diameter and height between the two? Geometric similarity between tanks a and b in Figure 11. exists when. Experimental data achieved for the models are then scaled to predict the performance of full-sized prototypes according to the requirements of geometric. the ratios of signiﬁcant forces are equal between model and prototype. The process of scaling using these similarity requirements will be presented in Section 11. Vb Va ¼ 5.2 Cylindrical mixing tanks for Example 2. The following examples will illustrate the manner of utilizing model data to evaluate the conditions for a full-scale device. 2 Db hb ¼5 Da ha . 11.5. in geometrically similar systems 1 2 . the velocities at the same locations are related according to 1 vx vx vx vx ¼ ¼ vy 1 vy 2 vz 1 vz 2 The third type of similarity. in geometrically and kinematically similar systems. thus. kinematic.132 Chapter 11 Dimensional Analysis and Similitude 1 and Kinematic similarity similarly exists when.2.5 MODEL THEORY In the design and testing of large equipment involving ﬂuid ﬂow. These force ratios that are important in ﬂuid ﬂow applications include the dimensionless parameters listed in Table 11. and dynamic similarity. Vb (p/4)D2b hb ¼ ¼5 Va (p/4)D2b ha and we get hb B Figure 11.

Model Characteristic length Velocity Viscosity Density Speed of sound L v m r 275 m/s Prototype 24.11. (2) The ratio of forces between the model and the full-scale aircraft. Conditions of dynamic similarity should prevail.5 Model Theory 133 We now substitute the geometric similarity requirement that gives Db Da 3 ¼ 3 Lb ¼5 La and the two ratios of interest become Db Lb ¼ ¼ 51/3 ¼ 1:71 Da La EXAMPLE 3 Dynamic similarity may be obtained by using a cryogenic wind tunnel in which nitrogen at low temperature and high pressure is employed as the working ﬂuid. For dynamic similarity to exist. we know that both model and prototype must be geometrically similar and that the Reynolds number and the Mach number must be the same. Equating Mach numbers we obtain Mm ¼ M p v¼ 275 60 ¼ 48:5 m/s 340 Equating the Reynolds numbers of the model and the prototype we obtain Rem ¼ Re p r 48:5L 1:225 60 24:38 ¼ 1:002 108 ¼ m 1:789 105 . The speed of sound in nitrogen at 183 K is 275 m/s. If nitrogen at 5 atm and 183 K is used to test the low speed aerodynamics of a prototype that has a 24.38 m 60 m/s 1:789 105 Pa s 1:225 kg/m3 340 m/s The conditions listed for the prototype have been obtained from Appendix I. determine (1) The scale of the model to be tested. A table such as the following is helpful.38 m wing span and is to ﬂy at standard sea-level conditions at a speed of 60 m/s.

we obtain L ¼ 3:26 m ð10:7 ftÞ The ratio of the forces on the model to the forces experienced by the prototype may be determined equating values of Eu between the model and the prototype.134 Chapter 11 Dimensional Analysis and Similitude Using equation (7-10). Hence F rV 2 AR ¼ model F rV 2 AR prototype where AR is a suitable reference area. Substituting numbers Fm 7:608 48:5 2 3:26 2 ¼ ¼ 0:0726 24:38 F p 1:225 60:0 The forces on the model are seen to be 7.m (rV 2 )m lm 2 ¼ ¼ F p r p V 2p AR. p (rV 2 )p l p where the ratio of reference areas can be expressed in terms of the scale ratio. It should be kept in mind that dimensional analysis cannot predict which variables are important in a given situation. . The ratio of model force to prototype force is then given by Fm rm Vm2 AR. e/k ¼ 91:5 K and ˚ for nitrogen so that kT/ e ¼ 2 and Vm ¼ 1:175 (Appendix K). which are fewer in number than the original variables. From Appendix K. nor does it give any insight into the physical transfer mechanism involved.6 CLOSURE The dimensional analysis of a momentum-transfer problem is simply an application of the requirement of dimensional homogeneity to a given situation. The indicated relations between dimensionless parameters are then useful in expressing the performance of the systems to which they apply. dimensional analysis techniques are a valuable aid to the engineer. this reference area is the projected wing area. we may evaluate m for nitrogen. Even with these limitations. By dimensional analysis the work and time required to reduce and correlate experimental data are decreased substantially by the combination of individual variables into dimensionless p groups.26% the prototype forces. s ¼ 3:681 A m ¼ 2:6693106 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 28183 ¼ 1:200105 Pas (3:681)2 (1:175) The density may be approximated from the perfect gas law r¼ P M T1 r P1 M1 T 1 so that 28 288 1:225 ¼ 7:608 kg/m3 r¼5 28:96 183 Solving for the wing span of the model. For an aircraft. 11. Thus.

. N. The requirements of geometric. Model theory is thus an important application of the parameters obtained in a dimensional analysis.surface tension. g ¼ acceleration of gravity. kinematic. may be used. m. and the ﬂuid viscosity m. and the acceleration of gravity. D 2g in which f is a dimensionless coefﬁcient that depends on (a) the average velocity v of the pipe ﬂow. the density r of water.4 The maximum pitching moment that is developed by the water on a ﬂying boat as it lands is noted as cmax The following are the variables involved in this action: a ¼ angle made by ﬂight path of plane with horizontal. r ¼ density of water. and dynamic similarity enable one to use model date to predict the behavior of a prototype or full-size piece of equipment. (b) the pipe diameter D. the lubricant viscosity. and the efﬁciency h of the turbine. and s. the surface tension of the lubricating oil. This is a very general approach but gives no physical meaning to the dimensionless parameters obtained from such an analysis. liquid density. the mass density r. Suggest appropriate parameters to be used in correlating experimental data for such a system.Neglect any effects of viscosity. the impeller diameter D. and D are kept in separate groups. an empirical method. and the applied tensile force T. This method also gives physical meaning to the groups thus obtained. choosing them so that P. How would you accumulate and present the experimental data for this system? 11. generate a set of appropriate dimensionless groups. no equation applies. the volumetric rate of ﬂow Q. ﬁnd a dimensionless function for the coefﬁcient f. the angular velocity v. R ¼ radius of gyration of plane about axis of pitching. on the contrary. Suggest a set of dimensionless parameters relating these variables. If. b ¼ angle deﬁning attitude of plane. This process is believed to be controlled by the following variables: k ¼ mass-transfer coeﬁcient D ¼ diffusion coefﬁcient d ¼ disk diameter a ¼ angular velocity r ¼ density m ¼ viscosity Dimensions L/t L2/t L 1/t M/L3 M/Lt Obtain the set of dimensionless groups for these variables where k. then by the efﬁciency of the pump. the gravitational acceleration g. and (e) the average pipe wall uneveness e (length). 11. 11. PROBLEMS 11.5 The rate at which metallic ions are electroplated from a dilute electrolytic solution onto a rotating disk electrode is usually governed by the mass diffusion rate of ions to the disk.6 The performance of a journal bearing around a rotating shaft is a function of the following variables: Q.1 The power output of a hydraulic turbine depends on the diameter D of the turbine. 11. L ¼ length of hull. 11. Darcy derived an equation for the friction loss in pipe ﬂow as hL ¼ f L v2 . Find the pertinent dimensionless groups. the discharge Q of water through the turbine.3 The pressure rise across a pump P (this term is proportional to the head developed by the pump) may be considered to be affected by the ﬂuid density r.8 The functional frequency n of a stretched string is a function of the string length L. its diameter D. how many independent dimensionless groups should there be which characterize this problem? (b) What is the dimensional matrix of this problem? What is its rank? (c) Evaluate the appropriate dimensionless parameters for this problem. Using the Buckingham p theorem. the shaft speed in revolutions per minute. the Buckingham method. r. the height H of water surface above the turbine. Find similar expressions. (c) the ﬂuid density r. m. H. By dimensional analysis. M ¼ mass of plane. the angular velocity v of the turbine wheel. and m each appear in one group only. (d) the ﬂuid viscosity m. Q. 11. D. 11. the rate of ﬂow lubricating oil to the bearing in volume per unit time.2 Through a series of tests on pipe ﬂow. the bearing diameter. replacing the pressure rise ﬁrst by the power input to the pump.7 The mass M of drops formed by liquid discharging by gravity from a vertical tube is a function of the tube diameter D. Determine the independent dimensionless groups that would allow the surface-tension effect to be analyzed. the number of dimensionless groups is automatically determined by taking ratios of the various terms in the expression to one another.Problems 135 If the equation describing a given process is known. the lubricant density. (a) According to the Buckingham p theorem.

and the viscosity m.11 The size d of droplets produced by a liquid spray nozzle is thought to depend upon the nozzle diameter D. it is desired to test a 10% scale model in a towing tank to determine the drag characteristics of the hull. If the lift F1 is a function of the density r of the ﬂuid. If the fullscale vehicle experiences the unsteady effects at a Mach number of 1 at an altitude of 40.18 An estimate is needed on the lift provided by a hydrofoil wing section when it moves through water at 60 mph. 11.13 A car is traveling along a road at 22. Find the air speed required to test the model and ﬁnd the ratio of the model drag to the fullscale drag. In addition.19 A model of a harbor is made on the length ratio of 360:1. The scaling of time can be made with the aid of the length scale and velocity scaling factors. m. Experimental evidence suggests the radius r of the high-pressure blast wave depends on time t as well as the energy E and r the density of the ambient air. jet velocity V. but only on the Froude number (based on forward velocity V and propeller diameter d). what is the prototype resistance? 11. ﬁnd the equation for r as a function of t. Develop a relation between these variables by dimensional analysis. 11. and g. (b) based on the diameter of the radio antenna. it is thought that the ratio of forward to rotational speed of the propeller must be constant (the ratio V/Nd.21 A model ship propeller is to be tested in water at the same temperature that would be encountered by a full-scale propeller. and that the kinematic viscosity of the water is approximately 1:0 105 ft2 /s: Take the density of water to be 1:94 slugs/ft3 : (a) Using the Buckingham method. Take D. Determine how the model is to be tested if the Froude number is to be duplicated. and E. angular velocity v. respectively. what should be the tidal period in the model? 11.45-m diameter prototype? (b) A torque of 20 N m is required to turn the model. Rewrite this relation in dimensionless form. and s. Over the speed range considered. the equation of motion becomes r Dv TH 1 ¼ =P þ m=2 v þ rg T0 Dt Show that the ratio of gravity (buoyancy) to inertial forces acting on a ﬂuid element is Lg TH 1 V02 T0 where L and V0 are reference lengths and velocity.000 ft.20 A 40% scale model of an airplane is to be tested in a ﬂow regime where unsteady ﬂow effects are important. a forward speed of 2. what velocity of ﬂow in the wind tunnel would correspond to the hydrofoil velocity for which the estimate is desired? Assume the same angle of attack in both cases. What model velocity corresponds to a torpedo velocity of 20 knots? If the model resistance is 10 lb. r.8 m and the antenna diameter is 6. 11. volume ﬂow rate Q. 11. 11. where ﬂuid viscosity and angular velocity appear in only one dimensionless parameter.14 In natural-convection problems. 11. Signiﬁcant variables are the length scale. the angle of attack u. what pressure must the model be tested at to produce an equal Reynolds number? The model is to be tested in air at 708F. ﬂuid density r. 11. the chord length D.2 m/s Calculate the Reynolds number (a) based on the length of the car. the velocity v of the ﬂow. 11. that its kinematic viscosity is 8:0 105 ft2 /s. and the properties of the liquid r.16 During the development of a 300-ft ship. what should be the size and speed of the waves in the model? (b) If the time between tides in the prototype is 12 h.136 Chapter 11 Dimensional Analysis and Similitude 11. If a warm gas at TH moves through a gas at temperature T0 and if the density change is only due to temperature changes.4 mm. (a) With a model 041 m in diameter. What will the timescale of the ﬂow about the model be relative to the full-scale vehicle? 11. Storm waves of 2 m amplitude and 8 m/s velocity occur on the breakwater of the prototype harbor. and V as repeating variables. r. and ﬂuid viscosity m. velocity. 11. the variation of density due to the temperature difference DT creates an important buoyancy term in the momentum equation. the acceleration of gravity. The car length is 5. that the density of the air in the pressurized tunnel is 5:0 103 slugs/ft3 . and the model thrust is measured to be 245 N. it is assumed that there is no dependence on the Reynolds or Euler numbers.10 A large amount of energy E is suddenly released in the air as in a point of explosion.17 A 25% scale model of an undersea vehicle that has a maximum speed of 16 m/s is to be tested in a wind tunnel with a pressure of 6 atm to determine the drag characteristics of the fullscale vehicle. where N is propeller rpm). The model is 3 m long. What are the torque and thrust for the prototype? .9 The power P required to run a compressor varies with compressor diameter D. What are the forward and rotational speeds corresponding to a 2. Test data are available for this purpose from experiments in a pressurized wind tunnel with an airfoil section model geometrically similar to but twice the size of the hydrofoil.12 Identify the variables associated with Problem 8.13 and ﬁnd the dimensionless parameters. (b) Show that the speed of the wave front decreases as r increases.58 m/s and a rotational speed of 450 rpm is recorded.15 A 1/6-scale model of a torpedo is tested in a water tunnel to determine drag characteristics. (a) Neglecting friction.

giving it a ﬂuctuating nature. and in order for us to measure the viscosity. in which small packets of ﬂuid particles are transferred between layers. Our task in this chapter is to consider viscous ﬂuids and the role of viscosity as it affects the ﬂow. m. The existence of laminar and turbulent ﬂow.1 Reynolds’s experiment. The second ﬂow regime. Similar behavior may be observed for water ﬂowing slowly from a faucet. At high ﬂow rates. The smoke emanating from a lighted cigarette is seen to ﬂow smoothly and uniformly for a short distance from its source and then change abruptly into a very irregular. is called the turbulent ﬂow regime. all ﬂuids are viscous. however. 12. as shown.1. The well-ordered type of ﬂow occurs when adjacent ﬂuid layers slide smoothly over one another with mixing between layers or lamina occurring only on a molecular level. Clearly. A dye having the same speciﬁc gravity as water was introduced at the pipe opening and its pattern observed for progressively larger ﬂow rates of water. although recognized earlier.1(a). unstable pattern. at a rate controlled by a valve. Water was allowed to ﬂow through a transparent pipe. was ﬁrst described quantitatively by Reynolds in 1883. His classic experiment is illustrated in Figure 12. this laminar ﬂow must exist. a ﬂuid may be considered ideal or inviscid. 137 .1 REYNOLDS’S EXPERIMENT The existence of two distinct types of viscous ﬂow is a universally accepted phenomenon. It was for this type of ﬂow that Newton’s viscosity relation was derived. making possible an analysis by the methods of Chapter 10. At low rates of ﬂow. but in certain situations and under certain conditions. the dye pattern was regular and formed a single line of color as shown in Figure 12. Of particular interest is the case of ﬂow past solid surfaces and the interrelations between the surfaces and the ﬂowing ﬂuid.Chapter 12 Viscous Flow The concept of ﬂuid viscosity was developed and viscosity deﬁned in Chapter 7. the dye became dispersed throughout the pipe cross section because of Water Dye (a) Re < 2300 Valve Water Dye Valve (b) Re > 2300 Figure 12.

138 Chapter 12 Viscous Flow the very irregular ﬂuid motion. Cf which is deﬁned by equation (12-2). additionally CD ¼ the drag coefficient and AP ¼ the projected area of the surface The value of AP used in expressing the drag for blunt bodies is normally the maximum projected area for the body. and v 1 are as described above and. in honor of Osborne Reynolds and his important contributions to fluid mechanics. small disturbances will cause a transition to turbulent ﬂow whereas below this value disturbances are damped out and laminar ﬂow prevails. The transition from laminar to turbulent ﬂow in pipes is thus a function of the ﬂuid velocity.. is deﬁned as F rv 2 CD 1 AP 2 (12-3) where F. combined into the single dimensionless parameter Re Drv m (12-1) form the Reynolds number. A is the area of contact between the solid body and the fluid. Above a Reynolds number of 2300. CD. In such a situation another coefﬁcient. Frictional drag is evaluated by using the expression F rv 2 Cf 1 A 2 (12-2) where F is the force. and indeed. and ﬂuid viscosity. The critical Reynolds number for pipe ﬂow thus is 2300. of course. the others being pipe diameter. For ﬂow in circular pipes. due to the orderly nature of laminar ﬂow in the ﬁrst case and to the ﬂuctuating character of turbulent ﬂow in the latter case. . and v 1 is the free-stream fluid velocity. r. it is found that below a value for Reynolds number of 2300 the ﬂow is laminar. 12. Above this value the ﬂow may be laminar as well. ﬂuid density.e. Another manner of illustrating these different ﬂow regimes and their dependence upon Reynolds number is through the consideration of drag. These four variables. The drag force due to friction is caused by the shear stresses at the surface of a solid object moving through a viscous ﬂuid.2 DRAG Reynolds’s experiment clearly demonstrated the two different regimes of ﬂow: laminar and turbulent.000 in experiments wherein external disturbances were minimized. laminar ﬂow has been observed for Reynolds numbers as high as 40. r is the fluid density. ﬂow around a body as opposed to ﬂow inside a conduit). Actually. A particularly illustrative case is that of external ﬂow (i. Reynolds found that ﬂuid velocity was the only one variable determining the nature of pipe ﬂow. is dimensionless. The difference in the appearance of the dye streak was. Cf is the coefficient of skin friction. The total drag on an object may be due to pressure as well as frictional effects. symbolized Re. The coefﬁcient of skin friction.

2 Drag coefﬁcient for circular cylinders as a function of Reynolds number. being less than 1. Recalling the physical signiﬁcance of the Reynolds number from Chapter 11 as the ratio of the inertia forces to the viscous forces.2. The features of each regime will be examined. Shaded regions indicate areas inﬂuenced by shear stress.01 10 –1 10 – 0 II 10 1 III 10 2 10 3 10 4 rvD Reynolds number = m IV 10 5 10 6 Figure 12. or drag due to lift. A simple geometric shape that illustrates the drag dependence upon the Reynolds number is the circular cylinder. In an incompressible ﬂow. may be subdivided into four regimes. The variation in the drag coefﬁcient with the Reynolds number for a smooth cylinder is shown in Figure 12. 100 10 CD 1 0. Regime 1 In this regime the entire ﬂow is laminar and the Reynolds number small. a net rearward force develops. 1 A third source of pressure drag. This deviation in streamline pattern prevents the pressure over the rest of a body from reaching the level it would attain otherwise. and hence the effects of shear stress on the ﬂow. we may say that in regime 1 the viscous forces predominate. is associated with shock waves. the ﬂow adheres to the body. produced no drag. The ﬂow pattern and general shape of the curve suggest that the drag variation. wave drag. As the pressure at the front of the body is now greater than that at the rear. and in some cases to separate from the body altogether.12. . which arises from the fact that the shear stress causes the streamlines to deviate from their inviscid ﬂow paths.2 Drag 139 The quantity rv 21 /2 appearing in equations (12-2) and (12-3) is frequently called the dynamic pressure. The ﬂow pattern in this case is almost symmetric.1 Regimes I 0.1 One is induced drag. The ﬂow pattern about the cylinder is illustrated for several different values of Re. Pressure drag arises from two principal sources. as there existed neither frictional nor pressure drag. The inviscid ﬂow about a circular cylinder was examined in Chapter 10. The other source is wake drag. The inviscid ﬂow about a cylinder of course. the drag coefﬁcient depends upon the Reynolds number and the geometry of a body.

This change in the character of the wake from a steady to an unsteady nature is accompanied by a change in the slope of the drag curve. The minimum point on the curves for Reynolds numbers of 105 and 6 105 are both at the point of ﬂow 2 v P 1 θ Re = 105 2 (P – P) / 1 ρv2 0 –1 –2 Inviscid flow Re = 6 105 –3 –4 0 30 60 90 120 q. As the Reynolds number is increased. In this regime of the so-called creeping ﬂow. The wake is no longer characterized by large eddies. Regime 2 Two illustrations of the ﬂow pattern are shown in the second regime.5.3) up to the point of separation is fairly close to the inviscid ﬂow pressure distribution depicted in Figure 10. Regime 4 At a Reynolds number near 5 105. The pattern of eddies shown in regime 2 is called a von Ka´rma´n vortex trail. The paramount features of this regime are (a) the unsteady nature of the wake and (b) ﬂow separation from the body.140 Chapter 12 Viscous Flow and the wake is free from oscillations. In addition. these eddies grow to the point at which they separate from the body and are swept downstream into the wake. the drag coefﬁcient suddenly decreases to 0. .3. In the ﬁgure it will be noticed that the pressure variation about the surface is a changing function of Reynolds number. viscous effects predominate and extend throughout the ﬂow ﬁeld. degrees 150 180 Figure 12.3 Pressure distribution on a circular cylinder at various Reynolds numbers. The drag coefﬁcient levels out at a near-constant value of approximately 1. and the shear stress in this interval is appreciable only in a thin layer near the body. the pressure distribution about the cylinder (shown in Figure 12. small eddies form at the rear stagnation point of the cylinder. The ﬂow on the surface of the body from the stagnation point to the point of separation is laminar. At higher values of the Reynolds number. although it remains unsteady. Regime 3 In the third regime the point of ﬂow separation stabilizes at a point about 808 from the forward stagnation point. it is observed that the point of separation has moved past 908. When the ﬂow about the body is examined.

The density of glass is 2250 kg/m3.000 Figure 12.000.4 shows the variation in the drag coefﬁcient with the Reynolds number for a sphere. This is again due to the change from laminar to turbulent ﬂow in the boundary layer. As the Reynolds number is large in this regime.5 0.05 0. In regimes 3 and 4. .5 1 2 5 10 20 50 200 1000 10. undergoing transition from laminar ﬂow close to the forward stagnation point. (b) water at 300 K. a turbulent ﬂow resists ﬂow separation better than a laminar ﬂow. In general. The marked decrease in drag is due to the change in the point of separation.12. Figure 12.1 0. and (c) glycerin at 300 K.2. agreement with inviscid-ﬂow predictions at a given Reynolds number increases as the slenderness of the body increases.2 Drag 141 separation. From this ﬁgure it is seen that separation occurs at a larger value of u for Re ¼ 6 105 than it does for Re ¼ 105. the ﬂow pattern over the forward part of the cylinder agrees well with the inviscid ﬂow theory. as might be expected.000 Reynolds number = rvD/m 100. one may observe the same sharp decrease in CD to a minimum value near a Reynolds number value of 5 105 . The layer of ﬂow near the surface of the cylinder is turbulent in this regime. The majority of cases of engineering interest involving external ﬂows have ﬂow ﬁelds similar to those of regimes 3 and 4.2 0. for inﬁnite plates. Note the similarity in form of the curve of CD for the sphere to that for a cylinder in Figure 12. a similar variation in the realm of inﬂuence of viscous forces is observed and.4 Drag coefﬁcient versus Reynolds number for various objects. and for circular disks and square plates. EXAMPLE 1 Evaluate the terminal velocity of a 7. Speciﬁcally. For other geometries.02 0.1 0. it may be said that the inertial forces predominate over the viscous forces. The four regimes of ﬂow about a cylinder illustrate the decreasing realm of inﬂuence of viscous forces as the Reynolds number is increased.2 Circular disks and square plates 0.000 1. 200 100 50 v D Spheres 20 v Infinite plates Circular D disk and square plates 10 5 CD Strokes drag CD = 24/Re v Sphere Infinite plates D 2 1 0.5-mm-diameter glass sphere falling freely through (a) air at 300 K.

142 Chapter 12 Viscous Flow The terminal (steady state) velocity of a falling body is reached when the force due to ﬂuid drag matches the body’s weight. A trial-and-error solution is. 0:4 < CD < 0:5. In this case. Red curve.4. that is. . thus. which would permit the use of Stokes’ law in expressing CD. For air at 300 K n ¼ 1:569 105 m2 /s r ¼ 1:177 kg/m3 dv 1 (7:5 103 m)v 1 ¼ n 1:569 105 m2 /s ¼ 478:0 v 1 Red ¼ (A) Inserting known values into our force balance expression. we are unable to solve explicitly for v 1 unless Red < 1.4. is plotted as a function of Reynolds number. The conditions to be satisﬁed are our force balance expression and the graphical relation between CD and Red in Figure 12. however. and solve equation (B) for v 1: v1 187:5 2 2 1/2 ¼ ¼ 21:65 m/s m /s 0 Equation (b) then yields Red ¼ (478:0)(21:65) ¼ 1:035 104 These results are compatible with Figure 12. over a range in Red between 500 < Red < 105 . Speciﬁcally. the shape of the CD vs.4 although the absolute accuracy is obviously not great. over 3 orders of magnitude! In such a speciﬁc case. we will assume CD ﬃ 0:4. in Figure 12. we have CD v 21 ¼ 4 2250 kg/m3 (7:5 103 m)(9:81 m/s2 ) 3 1:177 kg/m3 (B) ¼ 187:5 m /s 2 2 Normally the trial-and-error procedure to achieve a solution would be straightforward. Red. poses a bit of a problem. the value of CD remains nearly uniform. given in Figure 12.4. CD. As CD is a function of v 1. In this case the weight of the glass sphere can be expressed as rs pd 3 g 6 The ﬂuid drag force is given by CD rf v 21 pd 2 2 4 and a force balance yields CD v 21 ¼ 4 rs dg 3 rf The drag coefﬁcient. required.

and achieve the result that v 1 ¼ 0:495 m/s Red ¼ 4220 These results. approximately v 1 ﬃ 21:6 m/s (a) For water at 300 K n ¼ 0:879 106 m2 /s r ¼ 996:1 kg/m3 Red ¼ CD v 21 ¼ (7:5 103 m)v 1 ¼ 8530 v 1 0:879 106 m2 /s 4 2250 kg/m3 (7:5 103 m)(9:81 m/s2 ) 3 996 kg/m3 ¼ 0:0981 As in part (a).12.2 Drag 143 Finally. we determine the terminal velocity to be. The terminal velocity in glycerin thus v 1 ¼ 0:0773 m/s (c) . we have 24n 2 CD v 21 ¼ v ¼ 0:1752 m2 /s2 dv 1 1 v1 ¼ (0:1752 m2 /s2 )(7:5 103 m) 24(7:08 104 m2 /s) ¼ 0:0773 m/s To validate the use of Stokes’ law. for glycerin at 300 K n ¼ 7:08 104 m2 /s r ¼ 1260 kg/m3 Red ¼ CD v 21 ¼ (7:5 103 m)v 1 ¼ 10:59 v 1 7:08 104 m2 /s 4 2250 kg/m3 (7:5 103 m)(9:81 m/s2 ) 3 1260 kg/m3 ¼ 0:1752 In this case we suspect the Reynolds number will be quite small. thus CD ¼ 24/Re: Solving for v 1 for this case. satisfy Figure 12.4. we will initially assume CD ﬃ 0:4. As an initial guess we will assume Stokes’ law applies. we check the value of Reynolds number and get Red ¼ (7:5 103 m)(0:0773 m/s) 7:08 104 m2 /s ¼ 0:819 which is in the allowable range. in water v 1 ¼ 0:495 m/s (b) Finally. for air. again. Thus.

Region of transition Laminar boundary-layer region Turbulent boundary-layer region n Streamline δ x Laminar sublayer Figure 12. a very thin ﬁlm of ﬂuid called the laminar sublayer. (The thickness is exaggerated for clarity. and above. for example. d. known as the local Reynolds number.) Figure 12. The thickness of the boundary layer. At relatively small values of x. and this is designated as the laminar boundary-layer region. At larger values of x the transition region is shown where ﬂuctuations between laminar and turbulent ﬂows occur within the boundary layer.5 Boundary layer on a ﬂat plate. Rex. In the region in which the boundary layer is turbulent.3 THE BOUNDARY-LAYER CONCEPT The observation of a decreasing region of inﬂuence of shear stress as the Reynolds number is increased led Ludwig Prandtl to the boundary-layer concept in 1904.5 illustrates how the thickness of the boundary layer increases with distance x from the leading edge. Thus the only unknowns are the velocity components. This means that the pressure in the boundary layer is the same as the pressure in the inviscid ﬂow outside the boundary layer. ﬂow within the boundary layer is laminar. there exists. Further. as shown. the effects of ﬂuid friction at high Reynolds numbers are limited to a thin layer near the boundary of a body. The local Reynolds number is deﬁned as xvr (12-4) Rex m For ﬂow past a ﬂat plate. as shown in Figure 12. experimental data indicate that for (a) Rex < 2 105 (b) 2 105 < Rex < 3 106 (c) 3 106 < Rex the boundary layer is laminar.144 Chapter 12 Viscous Flow 12. The criterion for the type of boundary layer present is the magnitude of Reynolds number. based on the distance x from the leading edge. According to Prandtl’s hypothesis. The boundary layer on a ﬂat plate is shown in Figure 12. wherein ﬂow is still laminar and large velocity gradients exist. Finally. is arbitrarily taken as the distance away from the surface where the velocity reaches 99% of the free-stream velocity. the boundary layer is turbulent . The thickness is exaggerated for clarity. The pressure.5. hence the term boundary layer. there is no signiﬁcant pressure change across the boundary layer.5. for a certain value of x. The signiﬁcance of the Prandtl theory lies in the simpliﬁcation that it allows in the analytical treatment of viscous ﬂows. the boundary layer may be either laminar or turbulent. may be obtained from experiment or inviscid ﬂow theory. the boundary layer will always be turbulent.

4 THE BOUNDARY-LAYER EQUATIONS The concept of a relatively thin boundary layer at high Reynolds numbers leads to some important simpliﬁcations of the Navier–Stokes equations.4 The Boundary-Layer Equations 145 12. the equations for ﬂow over a ﬂat plate become @v x @v x @v x @P @ 2 vx þm 2 þ vx þ vy r (12-7) ¼ @x @t @x @y @y and r @v y @v y @v y @2vy @P þm þ vx þ vy ¼ @y @t @x @y @x2 (12-8) Furthermore.5. The shear stress in a thin boundary layer is closely approximated by m(@v x /@y). hence @P/@x ¼ dP/dx. and thus @v x /@y @v y /@x: The normal stress at a large Reynolds number is closely approximated by the negative of the pressure as m(@v x /@x) O(mv 1 /x) ¼ O(rv 21 /Rex ). For example. For incompressible. twodimensional ﬂow over a ﬂat plate.12. for a relatively thin boundary layer. therefore s xx ’ s yy ’ P: When these simpliﬁcations in the stresses are incorporated. s xx ¼ P þ 2m(@v x /@x) and s yy ¼ Pþ 2m(@v y / @y). Then @v y v y jd @v x v x jd O O @y d @x x so x2 @v x /@y O @v y /@x d which. is a large number. and thus @P/@y ’ 0. the Navier–Stokes equations are @v x @v x @v x @s xx @tyx þ vx þ vy þ (12-5) r ¼ @t @x @y @x @y and @v y @v y @v y @t xy @s yy þ vx þ vy þ r ¼ @t @x @y @x @y (12-6) where txy ¼ tyx ¼ m(@v x /@y þ @v y /@x). which according to Bernoulli’s equation is equal to rv 1 dv 1 /dx: The ﬁnal form of equation (12-7) becomes @v x @v x @v x dv 1 @2vx þ vx þ vy ¼ v1 þn 2 @t @x @y dx @y 2 (12-9) The order of magnitude of each term may be considered as above. where O signiﬁes the order of magnitude. we may write v x jd /v y jd O(x/d).2 the terms in the second equation are much smaller than those in the ﬁrst equation. This can be seen by considering the relative magnitudes of @v x /@y and @v y /@x From Figure 12. v x (@v y /@x) O(v 1 (v 1 /x) (d/x)) ¼ O(v 21 d/x2 ): .

Grenzshichten in Flu¨ssigkeiten mit kleiner Reibung. The equations to be solved are now the following: vx @v x @v x @2vx þ vy ¼n 2 @x @y @y (12-11a) and @v x @v y þ ¼0 @x @y (12-11b) with boundary conditions v x ¼ v y ¼ 0 at y ¼ 0. v 1 (x) ¼ v 1 and dP/dx ¼ 0. y) ¼ (12-12) 2 nx and f (h) ¼ C(x. This set of equations may be reduced to a single ordinary differential equation by transforming the independent variables x. For ﬂow parallel to a ﬂat surface. 1908. and v x ¼ v 1 at y ¼ 1. as described in Chapter 10. The following expressions will result. @C v 1 0 ¼ vx ¼ f (h) (12-14) @y 2 @C 1 nv 1 1/2 ¼ vy ¼ (h f 0 f ) (12-15) @x 2 x @v x v 1 h 00 f ¼ (12-16) 4x @x @v x v 1 v 1 1/2 00 ¼ f (12-17) @y 4 nx @ 2 v x v 1 v 1 000 f ¼ (12-18) @y2 8 nx 3 H. y) (nxv )1/2 (12-13) 1 The appropriate terms in equation (12-11a) may be determined from equations (12-12) and (12-13). which automatically satisﬁes the twodimensional continuity equation. Phys. Blasius. Math. 12. The reader may wish to verify the mathematics involved.146 Chapter 12 Viscous Flow The above equation. to h and the dependent variables from C(x. equation (12-11b). . Sci.5 BLASIUS’S SOLUTION FOR THE LAMINAR BOUNDARY LAYER ON A FLAT PLATE One very important case in which an analytical solution of the equations of motion has been achieved is that for the laminar boundary layer on a ﬂat plate in steady ﬂow. and the continuity equation @v x @v y þ ¼0 @x @y (12-10) are known as the boundary-layer equations. y) to f(h) where y v 1 1/2 h(x. Blasius3 obtained a solution to the set of equations (12-11) by ﬁrst introducing the stream function.. 1. Z. y. C. according to the Bernoulli equation. U.

6 3.2796 4.7876 1.6298 0.8792 6.4580 1.0096 0.9223 1.4819 3.6924 3. p.2 2.5294 0. Oxford Univ.9996 0.4 1. London. Soc. as a single ordinary differential equation f 000 þ f f 00 ¼ 0 (12-19) with the appropriate boundary conditions f ¼ f0 ¼ 0 at h ¼ 0 0 at h ¼ 1 f ¼2 Observe that this differential equation.7360 0.3096 1. two are initial values and the third is a boundary value. although ordinary.9980 1.9555 0. Howarth4 later performed essentially the same work but obtained more accurate results. using a series expansion to express the function.0000 1.0000 4 L.9878 0.3938 0.9950 1.1867 1.0000 2. Roy. at the origin and an asymptotic solution to match the boundary condition at h ¼ 1.12.9885 1. 1938.8 3.1 presents the signiﬁcant numerical results of Howarth.9998 2.8115 0. This equation was solved ﬁrst by Blasius.’’ Proc.1061 0.9990 0.9295 2.8 4.0670 0.9999 1.1328 0.0 1.3058 2.0002 0.7522 1.7290 0.6230 1.9110 1. A simpler way of solving equation (12-19) has been suggested in Goldstein5 who presented a scheme whereby the boundary conditions on the function f are initial values.6.5168 0.5691 1.0454 0. f 0 .0875 0.8 2. Press.6 0.1558 0.0336 1.4 3.0039 0.0217 0.0 3.2792 8.0000 2. Howarth.0000 0 0.0 5.2570 0.3924 0.6794 5.8761 0.6500 0.2647 0.2380 0.6 2. is nonlinear and that.2310 1.4 0. ‘‘On the solution of the laminar boundary layer equations.0015 0.9756 1.0853 3. London.0 0 0.9233 0.9915 0.8 1.2664 1.4 2. f 00 . Table 12.32824 1.2655 0.8803 4.1 Values of f.2 1.9518 1.0 2. A164 547 (1938).2 3.2792 0 0. and vx /v1 for laminar ﬂow parallel to a ﬂat plate (after Howarth) rﬃﬃﬃﬃﬃﬃﬃ vx y v1 h¼ f f0 f 00 v1 2 nx 0 0.0000 2.0000 1.9992 1.9943 0.8466 1.2 0.0266 0.5565 0. .5 Blasius’s Solution for the Laminar Boundary Layer on a Flat Plate 147 Substitution of (12-14) through (12-18) into equation (12-11a) and cancellation gives.4793 5.0000 1. Modern Developments in Fluid Dynamics.0005 0. Table 12. A plot of these values is included in Figure 12.0793 5.3260 1. of the end conditions on the variable f (h).4203 0.0000 1. f (h).0000 0.6 1.9759 0. 5 S.9124 0. 135.2596 1. Goldstein.

0 0. Berlin. then f00 (j ¼ 0) ¼ A/C3 : The constant A must have a certain value to . Berichtswesen. Zentrale F.0 h = 1 2 y 3.0 v nx Figure 12. wiss. 1942) for the Reynolds number range from 1:08 105 to 7:28 105 . If we deﬁne two new variables in terms of the constant. so that f ¼ f /C (12-20) j ¼ Ch (12-21) and then the terms in equation (12-19) become f (h) ¼ Cf(j) (12-22) f 0 ¼ C 2 f0 (12-23) 00 3 00 f ¼C f (12-24) f 000 ¼ C4 f000 (12-25) and The resulting differential equation in f(j) becomes f000 þ ff00 ¼ 0 and the initial conditions on f are f¼0 f0 ¼ 0 (12-26) f00 ¼ ? at j ¼ 0 The other boundary condition may be expressed as follows: f0 (j) ¼ f 0 (h) 2 ¼ 2 2 C C at j ¼ 1 An initial condition may be matched to this boundary condition if we let f 00 (h ¼ 0) equal some constant A.0 2.8 nx /v 148 0. C. Nikuradse (monograph.4 Blasius theory 0.6 0.Chapter 12 Viscous Flow 1.2 0 0 1. Experimental data by J.6 Velocity distribution in the laminar boundary layer over a ﬂat plate.

we have v x /v 1 ﬃ 0:99 thus. giving A ¼ 2C 3 : Thus initial values of f. The estimate on f00 (0) requires that 2/3 2 2 0 f (1) ¼ 2 ¼ 2 (12-27) C A Thus equation (12-26) may be solved as an initial-value problem with the answer scaled according to equation (12-27) to match the boundary condition at h ¼ 1. we have rﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃ y v1 d v1 h¼ ¼ ¼ 2:5 2 nx 2 nx and thus rﬃﬃﬃﬃﬃﬃﬃ nx d¼5 v1 or d 5 5 ¼ rﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ x v1x Rex n (b) The velocity gradient at the surface is given by equation (12-27): rﬃﬃﬃﬃﬃﬃﬃ @v x v 1 v 1 1/2 00 v1 ¼ f (0) ¼ 0:332 v 1 @y y¼0 4 nx nx (12-28) (12-29) As the pressure does not contribute to the drag for flow over a flat plate. all the drag is viscous. The signiﬁcant results of Blasius’s work are the following: (a) The boundary thickness. For this reason the symbol Cfx is used.12.1. d. and f00 are now speciﬁed. is obtained from Table 12. f0 . When h ¼ 2:5. . the x subscript indicating a local coefﬁcient. As an estimate we let f00 (j ¼ 0) ¼ 2.5 Blasius’s Solution for the Laminar Boundary Layer on a Flat Plate 149 satisfy the original boundary condition on f 0 . The shear stress at the surface may be calculated as @v x t0 ¼ m @y y¼0 Substituting equation (12-29) into this expression. we have rﬃﬃﬃﬃﬃﬃﬃ v1 t0 ¼ m 0:332 v 1 nx (12-30) The coefﬁcient of skin friction may be determined by employing equation (12-2) as follows: rﬃﬃﬃﬃﬃﬃﬃ v1 0:332mv 1 t Fd /A nx Cfx 2 ¼ 2 ¼ rv 1 /2 rv 1 /2 rv 21 /2 rﬃﬃﬃﬃﬃﬃﬃﬃﬃ n ¼ 0:664 xv 1 0:664 Cfx ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ Rex ð12-31Þ Equation (12-31) is a simple expression for the coefﬁcient of skin friction at a particular value of x. designating y ¼ d at this point.

is related to Cfx by Z 1 CfL ¼ Cfx dA A A For the case solved by Blasius. consider a plate of uniform width W. the pressure gradient was zero. most often one wishes to calculate the total drag resulting from viscous ﬂow over some surface of ﬁnite size. The pressure gradient plays a major role in ﬂow separation. Figure 12. for which rﬃﬃﬃﬃﬃﬃﬃ Z Z 1 L 1 L n 1/2 CfL ¼ Cfx dx ¼ 0:664 dx x L 0 L 0 v1 rﬃﬃﬃﬃﬃﬃﬃﬃﬃ n ¼ 1:328 Lv 1 1:328 CfL ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ (12-32) ReL 12.7 Variation in velocity and velocity derivatives across the laminar boundary layer when dP/dx ¼ 0. y y vx y ∂vx ∂y – ∂ 2 vx + ∂ y2 Figure 12. and length L. it is seldom that a local value is useful. A much more common ﬂow situation involves ﬂow with a pressure gradient.6 FLOW WITH A PRESSURE GRADIENT In Blasius’s solution for laminar ﬂow over a ﬂat plate. conditions at the wall v x ¼ v y ¼ 0. at y ¼ 0 equation 12-7 becomes @ 2 v x dP (12-33) m 2 ¼ @y y¼0 dx which relates the curvature of the velocity profile at the surface to the pressure gradient. . If we make use of the boundary.150 Chapter 12 Viscous Flow While it is of interest to know values of Cfx.7 illustrates the variation in v x. designated CfL. The mean coefﬁcient of skin friction that is helpful in this regard may be determined quite simply from Cfx according to Z rv 21 rv 21 ¼ Fd ¼ ACfL Cfx dA 2 2 A or the mean coefﬁcient. as can be seen with the aid of the boundarylayer equation (12-7). and @ 2 v x /@y2 across the boundary layer for the case of a zero-pressure gradient. @v x /@y.

the velocity in the layer of ﬂuid adjacent to the wall must be zero or negative. the variation in v x and its derivatives is shown in Figure 12.9 Velocity proﬁles in separatedﬂow region.8. We may now turn our attention to the subject of ﬂow separation.8 Variation in v x and its derivatives across the boundary layer for various pressure gradients. This type of velocity proﬁle is seen Separated region d Separation point Figure 12. the second derivative of the velocity at the wall must also be zero. and approaching zero at the outer edge of the boundary layer. as shown in Figure 12. The decrease in the velocity gradient means that the second derivative of the velocity must be negative. A positive value of dP/dx. A negative pressure gradient is seen to produce a velocity variation somewhat similar to that of the zero-pressure-gradient case. The derivative @ 2 v x /@y2 is shown as being zero at the wall. requires a positive value of @ 2 v x /@y2 at the wall.6 Flow with a Pressure Gradient 151 When dP/dx ¼ 0. y y y dP dx dP dx dP dx vx dP dx dP dx dP dx ∂ vx ∂y – ∂ 2 vx + ∂ y2 Figure 12. is associated with an inﬂection point. negative within the boundary layer. In order for ﬂow separation to occur. however. hence the velocity proﬁle is linear near the wall. As this derivative must approach zero from the negative side. For values of dP/dx 6¼ 0. Further.8. A zero second derivative. It is important to note that the second derivative must approach zero from the negative side as y ! d. out in the boundary layer. at some point within the boundary layer the second derivative must equal zero. the velocity gradient becomes smaller and gradually approaches zero. The inﬂection point is shown in the velocity proﬁle of Figure 12. it will be recalled.9.12. .

cannot cause ﬂow separation. to the present time. the y axis. For this reason a positive pressure gradient is called an adverse pressure gradient.10 Control volume for integral analysis of the boundary layer. and a line parallel to the y axis a distance Dx away. The control volume to be analyzed is of unit depth and is bounded in the xy plane by the x axis.10. Therefore. a negative pressure gradient is called a favorable pressure gradient.8.7 VON KA The Blasius solution is obviously quite restrictive in application. thus dP/dx > 0 is a necessary but not a sufﬁcient condition for separation. here drawn tangent to the surface at point 0. The velocity gradient at the wall increases as the pressure gradient becomes more favorable. proved inferior to experiment. Flow can remain unseparated with an adverse pressure gradient. In contrast a negative pressure gradient. We shall consider the case of two-dimensional. it may be concluded that a positive pressure gradient is necessary for ﬂow separation. incompressible steady ﬂow. applying only to the case of a laminar boundary layer over a ﬂat surface. in the absence of sharp corners. The presence of a pressure gradient also affects the magnitude of the skin friction coefﬁcient. the edge of the boundary layer. The above terms represent the x-directional pressure forces on the . As the only type of boundary-layer ﬂow that has an inﬂection point is ﬂow with a positive pressure gradient. ´ RMA ´ N MOMENTUM INTEGRAL ANALYSIS 12. A momentum analysis of the deﬁned control volume involves the application of the x-directional scalar form of the momentum theorem ZZZ ZZ @ : v x r(v n) dA þ v x r dV (5-5a) åFx ¼ @t c:s: c:v: A term-by-term analysis of the present problem yields the following: Pj Pjx (djxþDx djx ) t0 Dx åFx ¼ Pdjx PdjxþDx þ Pjx þ xþDx 2 where d represents the boundary-layer thickness. An approximate method providing information for systems involving other types of ﬂow and having other geometries will now be considered. vy Control volume y vx Stream lines d 0 x x Figure 12. as can be inferred from Figure 12.152 Chapter 12 Viscous Flow to require a point of inﬂection. Consider the control volume in Figure 12. and both forces are assumed negligible. Any situation of practical interest more complex than this involves analytical procedures that have.

may be evaluated as Z d Z d rv x dy rv x dy (12-34) m_ top ¼ xþDx 0 0 x The momentum expression. The surface integral term becomes ZZ c:s: Z v x r(v: n)dA ¼ d 0 rv 2x dy Z xþDx d 0 rv 2x dy v 1 m_ top x and the accumulation term is @ @t ZZZ v x r dV ¼ 0 c:v: as this is a steady-ﬂow situation. including equation (12-34). An application of the integral equation for conservation of mass will give ZZ ZZZ @ r(v: n) dA þ r dV ¼ 0 @t c:s: c:v: ZZ c:s: r(v: n) dA ¼ Z @ @t d 0 Z rv x dy xþDx d 0 (4-1) rv x dy m_ top x ZZZ r dV ¼ 0 c:v: and the mass-ﬂow rate into the top of the control volume. right. we get PjxþDx Pjx PjxþDx Pjx djxþDx djx Pdjx Pdjx djxþDx þ þ Dx 2 Dx Dx ! ! Rd 2 Rd Rd 2 Rd rv dyj rv dyj rv dyj rv dyj x x x x xþDx x xþDx x 0 0 0 ¼ v1 0 þ t0 Dx Dx Taking the limit as Dx!0 we obtain d dP d ¼ t0 þ dx dx Z d 0 rv 2x dy v 1 d dx Z d rv x dy 0 (12-35) . and the frictional force on the bottom.Von Ka´rma´n Momentum Integral Analysis 12. respectively. now becomes PjxþDx Pjx þ Pjx (djxþDx djx ) t 0 Dx 2 Z d Z d Z d Z d 2 2 ¼ rv x dyxþDx rv x dy v 1 rv x dy rv x dy (PdjxþDx Pdjx ) þ 0 0 x xþDx 0 0 x Rearranging this expression and dividing through by Dx. m_ top .7 153 left. and top sides of the control volume.

@ 2 v x /@y2 ¼ 0. with proper rearrangement. and d from these conditions.154 Chapter 12 Viscous Flow The boundary-layer concept assumes inviscid ﬂow outside the boundary layer. get the result t0 ¼ r d v1 dx Z d 0 d (v 1 v x ) dy þ dx Z d v x (v 1 v x ) dy (12-37) 0 Equation (12-37) is the von Ka´rma´n momentum integral expression. c. As the pressure is constant in this case. Equation (12-37) is a general expression whose solution requires a knowledge of the velocity. therefore (d/dx)v 1 ¼ 0 and equation (12-36) simpliﬁes to t0 d ¼ dx r Z d v x (v 1 v x ) dy (12-38) 0 An early solution to equation (12-38) was achieved by Pohlhausen. In this case the free-stream velocity is constant. which states that the second derivative at the wall is equal to the pressure gradient. a situation for which an exact answer is known. let us consider the case of laminar ﬂow over a ﬂat plate. we get a¼0 b¼ 3 v1 2d c¼0 d¼ v1 2d3 . v x . named in honor of Theodore von Ka´rma´n who ﬁrst developed it. Solving for a. as a function of distance from the surface. As an example of the application of equation (12-37). who assumed for the velocity proﬁle a cubic function v x ¼ a þ by þ cy2 þ dy3 (12-39) The constants a. and d may be evaluated if we know certain boundary conditions that must be satisﬁed in the boundary layer. c. for which we may write Bernoulli’s equation dP dv 1 þ rv 1 ¼0 dx dx which may be rearranged to the form d dP d d ¼ (dv 2 ) v 1 (dv 1 ) r dx dx 1 dx (12-36) Notice that the left-hand sides of equations (12-35) and (12-36) are similar. b. The accuracy of the ﬁnal result will depend on how closely the assumed velocity proﬁle approaches the real one. b. We may thus relate the right-hand sides and. These are (1) vx ¼ 0 at y ¼ 0 (2) vx ¼ v1 @v x ¼0 @y at y ¼ d (3) at y ¼ d and (4) @2vx ¼ 0 at y ¼ 0 @y2 Boundary condition (4) results from equation (12-33). y.

Figure 12. yet the analytical treatment of turbulent ﬂow is not nearly well developed as that of laminar ﬂow.8 DESCRIPTION OF TURBULENCE Turbulent ﬂow is the most frequently encountered type of viscous ﬂow. we observe a difference of about 7% in d and 3% in Cf . In a turbulent ﬂow the ﬂuid and ﬂow variables vary with time. (12-42). yields d 4:64 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ x Rex (12-41) The local skin-friction coefficient. equations (12-28). for example. give the form of the velocity proﬁle vx 3 y 1 y3 ¼ 2 d v1 2 d (12-40) Upon substitution. is given by t0 2v 3 v 1 0:646 ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 2 2 v Rex rv 1 2 d 2 1 Cfx 1 (12-42) Integration of the local skin-friction coefficient between x ¼ 0 and x ¼ L as in equation (12-32) yields 1:292 CfL ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ ReL (12-43) Comparing equations (12-41). The instantaneous velocity vector. This difference could. of course. 12. upon integration. (12-31). have been smaller had the assumed velocity profile been a more accurate representation of the actual profile.8 Description of Turbulence 155 which.11 illustrates the type of time dependence experienced by the . In this section.12. particularly with respect to the mechanism of turbulent contributions to momentum transfer. and (12-32). The momentum integral method may also be used to determine the shear stress from the velocity proﬁle. will differ from the average velocity vector in both magnitude and direction. Cfx . after integrating 3 v1 39 d 2 v (v d) ¼ 2 d 280 dx 1 As the free-stream velocity is constant. This comparison has shown the utility of the momentum integral method for the solution of the boundary layer and indicates a procedure that may be used with reasonable accuracy to obtain values for boundary-layer thickness and the coefﬁcient of skin friction where an exact analysis is not feasible. when substituted in equation (12-39). we examine the phenomenon of turbulence. and (12-43) with exact results obtained by Blasius for the same situation. equation (12-38) becomes 3v v 1 d ¼ 2 d1 dx Z d 0 v 21 3 y 1 y3 2d 2 d 3 y 1 y3 þ 1 dy 2d 2 d or. a simple ordinary differential equation in d results d dd ¼ 140 v dx 13 v 1 This.

t)dt. these ﬂuctuations contribute to the mean value of certain ﬂow quantities.156 Chapter 12 Viscous Flow v v _ v(x. For example. y. small random ﬂuctuations in velocity occur about the mean value. y. Q. t) dt ¼ 0 (12-46) t1 0 x Hereafter. y. as expressed by Z 1 t1 0 0 vx ¼ v (x. z) þ v 0x (x. While the mean value of the turbulent ﬂuctuations is zero. axial component of the velocity for turbulent ﬂow in a tube. y. t) is zero. t) t t (a) (b) Figure 12. y. y. z) represents the time-averaged velocity at the point (x. Accordingly.11 Time dependence of velocity in a turbulent ﬂow: (a) steady mean ﬂow. y.11a is seen to be steady in its mean value. z. we may express the ﬂuid and ﬂow variables in terms of a mean value and a ﬂuctuating value. The mean value of v 0x (x. z. t according to Q ¼ 1/t1 01 Q(x. z. Q will Rbe used to designate the time average of the general property. z. the mean kinetic energy per unit volume is 1 KE ¼ r½(v x þ v x0 )2 þ (v y þ v y0 )2 þ (v z þ v z0 )2 2 The average of a sum is the sum of the averages. z. z. z) _ v(x. (b) unsteady mean ﬂow. hence the kinetic energy becomes . the x-directional velocity is expressed as v x ¼ v x (x. While the velocity in Figure 12. y. t) Here vx (x. y. For example. y. t) dt t1 0 (12-44) (12-45) where t1 is a time that is very long in comparison with the duration of any fluctuation. z) Z 1 t1 vx ¼ v x (x.

.

o 1 n.

KE ¼ r v 2x þ 2v x v x0 þ v x02 þ v 2y þ 2v y v y0 þ v y02 þ v 2z þ 2v z v z0 þ v z02 2 or. 1 . since v x v x0 ¼ v x v x0 ¼ 0.

The level or intensity of turbulence is defined as qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ . (v x02 þ v y02 þ v z02 )1/2 is a significant quantity. KE ¼ r v 2x þ v 2y þ v 2z þ v x02 þ v y02 þ v z02 2 (12-47) A fraction of the total kinetic energy of a turbulent flow is seen to be associated with the magnitude of the turbulent fluctuations. It can be shown that the rms (root mean square) value of the fluctuations.

v x02 þ v y02 þ v z02 /3 (12-48) I v1 .

Although the turbulent fluctuations are functions .12 Turbulent motion at the surface of a control volume. the measurement of turbulence is seen to be a necessity in many applications.9 Turbulent Shearing Stresses 157 where v 1 is the mean velocity of the flow. y v–x = – vx ( y) vy' vx' y x x Figure 12. the time derivative in equation (5-4) is zero. 12.9 TURBULENT SHEARING STRESSES In Chapter 7. We shall now turn our attention to the effect of the turbulent ﬂuctuations on momentum transfer. will also result in a net transfer of momentum. The random nature of turbulence lends itself to statistical analysis. The general discussion so far has indicated the ﬂuctuating nature of turbulence. If the (molecular) random motions give rise to momentum transfer. such as those present in a turbulent ﬂow.3. The intensity of turbulence is a measure of the kinetic energy of the turbulence and is an important parameter in flow simulation. The relation between the macroscopic momentum ﬂux due to the turbulent ﬂuctuations and the shear stress may be seen from the control-volume expression for linear momentum ZZ ZZZ @ åF ¼ vr(v: n) dA þ vr dV (5-4) @t c:s: c:v: The flux of x-directional momentum across the top of the control surface is ZZ ZZ vr(v: n) dA ¼ v 0y r(v x þ v 0x ) dA top (12-49) top If the mean value of the momentum flux over a period of time is evaluated for the case of steady mean flow. Using an approach similar to that of Section 7.12. Thus. the random molecular motion of the molecules was shown to result in a net momentum transfer between two adjacent layers of ﬂuid. thus 0 ZZ ZZ ZZ 0 0 0 åFx ¼ v y r(v x þ v x )dA ¼ v y rv x dA þ rv 0y v 0x dA (12-50) % The presence of the turbulent fluctuations is seen to contribute a mean x directional momentum flux of rv x0 v y0 s per unit area. simulation of turbulent flows requires not only duplication of Reynolds number but also duplication of the turbulent kinetic energy. In model testing.12. it seems reasonable to expect that large-scale ﬂuctuations. let us consider the transfer of momentum in the turbulent ﬂow illustrated in Figure 12.

the Reynolds stress becomes (tyx )turb ¼ At d vx dy where At is the eddy viscosity. While Reynolds stresses exist for multidimensional ﬂows. eM . the turbulent contribution is expressed solely in terms of the ﬂuctuating properties of the ﬂow. . An early attempt to formulate a theory of turbulent shear stress was made by Boussinesq. even for the simplest case. v x . Transposing this term to the left-hand side of equation (5-4) and incorporating it with the shear stress due to molecular momentum transfer. Mem. The analog to the mean free path in molecular momentum exchange for the turbulent case is the mixing length proposed by Prandtl8 in 1925.. 12. and thus (tyx )turb ¼ reM d vx dy (12-52) The difficulties in analytical treatment still exist. these ﬂow properties are not expressible in analytical terms. ZAMM. momentum and continuity. The reason for these difﬁculties may be seen by examining the number of equations and the number of unknowns involved. par div. Pre. their analytical description has not been achieved. it may be observed that the units of the eddy diffusivity are L2/t. we see that the total shear stress becomes t yx ¼ m d vx rv x0 v y0 dy (12-51) The turbulent contribution to the shear stress is called the Reynolds stress. is a property of the flow and not of the fluid. as the eddy diffusivity. for example.10 THE MIXING-LENGTH HYPOTHESIS A general similarity between the mechanism of transfer of momentum in turbulent ﬂow and that in laminar ﬂow permits an analog to be made for turbulent shear stress. 7 J. Consider the simple turbulent ﬂow shown in Figure 12. Whereas the molecular contribution is expressed in terms of a property of the ﬂuid and a derivative of the mean ﬂow. An important difference between the molecular and turbulent contributions to the shear stress is to be noted. 6 The existence of the Reynolds stresses may also be shown by taking the time average of the Navier–Stokes equations.6 the difﬁculties in analytically predicting even the one-dimensional case have proved insurmountable without the aid of experimental data. v x0 . and v y0 . The shear stress in laminar ﬂow is tyx ¼ m(dv x /dy). (1877).13. Boussinesq introduced the concept relating the turbulent shear stress to the shear strain rate.158 Chapter 12 Viscous Flow of position and time. The close analogy between the molecular exchange of momentum in laminar flow¯ and the macroscopic exchange of momentum in turbulent flow suggests that the term rv 0x v 0y be regarded as a shear stress. 5. and four unknowns. thus by analogy. v y . Boussinesq. By analogy with the kinematic viscosity in a laminar flow.7 By analogy with the form of Newton’s viscosity relation. In the incompressible turbulent boundary layer. however. Prandtl. 136 (1925). there are two pertinent equations. Subsequent reﬁnements have led to the introduction of the eddy diffusivity of momentum. XXIII. 8 L. Further. In turbulent flows it is found that the magnitude of the Reynolds stress is much greater than the molecular contribution except near the walls. eM At /r. Sav.

giving 2 d v x d v x 0 0 (12-54) v x v y ¼ L dy dy Comparison with Boussinesq’s expression for the eddy diffusivity yields d v x eM ¼ L2 dy (12-55) . The important differences are its magnitude and dependence upon ﬂow properties rather than ﬂuid properties. With an expression for v 0x at hand. the velocity difference would be v x jyþL v x jy .12. Both the continuity equation and experimental data show that there is some degree of proportionality between v x0 and v y0 .13 The Prandtl mixing length. as 2 d v x d v x 0 0 (12-53) v x v y ¼ (constant)L dy dy The constant represents the unknown proportionality between v x0 and v y0 as well as their correlation in taking the time average. the mixing length. depending on the point of origin with respect to y. which is also unknown. The velocity ﬂuctuation v x0 is hypothesized as being due to the y-directional motion of a ‘‘lump’’ of ﬂuid through a distance L. In undergoing translation the lump of ﬂuid retains the mean velocity from its point of origin. which is unknown. If the lump of ﬂuid originated at y þ L. a distance L from the point of origin. The constant in (12-53). v x0 v y0 . may be incorporated into the mixing length. is assumed to be small enough to permit the velocity difference to be written as and thus v x jyL v x jy ¼ L dv x dy and thus v 0x ¼ L dv x dy (12-52) The concept of the mixing length is somewhat akin to that of the mean free path of a gas molecule. although ﬁnite. then the time average of their product would be zero. the sign of L. rv x0 v y0 . The minus sign and the absolute value were introduced to make the quantity v x0 v y0 agree with experimental observations. an expression for v 0y is necessary to determine the turbulent shear stress. If v x0 and v y0 were completely independent. Upon reaching a destination. The instantaneous value of v 0x jy is then v x jyL v x jy . Prandtl expressed the time average. Further. of course. Prandtl assumed that v 0x must be proportional to v y0 .10 The Mixing-Length Hypothesis 159 y L y vx Figure 12. Using the fact that v y0 v x0 . the lump of ﬂuid will differ in mean velocity from that of the adjacent ﬂuid by an amount v x jyL v x jy .

12. The empirical nature of the preceding discussion cannot be overlooked. Integration of the above equation pﬃﬃﬃﬃﬃﬃﬃﬃﬃ t0 /r In y þ C vx ¼ K (12-56) where C is a constant of integration. 356. Congr.11 VELOCITY DISTRIBUTION FROM THE MIXING-LENGTH THEORY One of the important contributions of the mixing-length theory is its use in correlating velocity proﬁles at large Reynolds numbers.. however. 62. and thus L ¼ Ky. 10 . Zurich (1927). in that assumptions regarding the nature and variation of the mixing length may be made on an easier basis than assumptions concerning the eddy viscosity.14. The agreement of experimental data for turbulent ﬂow in smooth tubes with equation (12-57) is quite good. we may write the turbulent shear stress as t yx d vx 2 ¼ rK y ¼ t0 (a constant) dy 2 2 or pﬃﬃﬃﬃﬃﬃﬃﬃﬃ d vx t0 /r ¼ Ky dy The quantity yields pﬃﬃﬃﬃﬃﬃﬃﬃﬃ t 0 /r is observed to have units of velocity. The velocity v x is assumed to increase in the y direction. Mech. Nikuradse. Appl. The shear stress is assumed to be entirely due to turbulence and to remain constant over the region of interest. it is remarkable that equation (13-15) describes the velocity proﬁle so well. There is an advantage. VDI-Forschungsheft. whereby v x max v x 1h yi pﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ ln K h t0 /r (12-57) The constant K was evaluated by Prandtl9 and Nikuradse10 from data on turbulent ﬂow in tubes and found to have a value of 0. namely that the shear stress is not constant and that the geometry was treated from a twodimensional viewpoint rather than an axisymmetric viewpoint. In view of these obvious difﬁculties. where K remains a dimensionless constant to be determined via experiment. Intern. and thus d v x /dy ¼ jd v x /dyj. as can be seen from Figure 12. In the neighborhood of the wall the mixing length is assumed to vary directly with y..4. Prandtl. Several assumptions regarding the ﬂow are known to be incorrect for ﬂow in tubes. 2nd Congr. Consider a turbulent ﬂow as illustrated in Figure 12. 9 L. Proc.13. Using these assumptions. This constant may be evaluated by setting v x ¼ v x max at y ¼ h. J. 1932.160 Chapter 12 Viscous Flow At ﬁrst glance it appears that little has been gained in going from the eddy viscosity to the mixing length.

of course.2 0. A pseudo-Reynolds number is found useful in this regard.15) of v þ versus ln y+. .12.12 THE UNIVERSAL VELOCITY DISTRIBUTION For turbulent ﬂow in smooth tubes. Equation (12-61) indicates that for ﬂow in smooth tubes v þ ¼ f (yþ ) or pﬃﬃﬃﬃﬃﬃﬃﬃﬃ vx y t 0 /r v þ pﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ f ln n t0 /r (12-61) (12-62) The range of validity of equation (13-19) may be observed from a plot (see Figur 12. we may pﬃﬃﬃﬃﬃﬃﬃﬃﬃ introduce a dimensionless velocity v x / t0 /r. 12. Recalling that the term t0 /r has the units of velocity. therefore the right-hand side of this equation must also be dimensionless.12 The Universal Velocity Distribution 161 12 11 Data at Re = 106 10 9 7 t0 /r (vx max – vx) 8 6 5 Equation (12-57) 4 3 2 1 0 0 0. Defining pﬃﬃﬃﬃﬃﬃﬃﬃﬃ t0 /r þ y (12-60) y n we find that equation (12-59) becomes vþ ¼ 1 nyþ 1 ln pﬃﬃﬃﬃﬃﬃﬃﬃﬃ þ C ¼ (ln yþ þ ln b) K K t0 /r where the constant b is dimensionless. dimensionless.4 0. using the data of Nikuradse and Reichardt.6 0.14 Comparison of data for ﬂow in smooth tube with equation (12-57). equation (12-57) pﬃﬃﬃﬃﬃﬃﬃﬃﬃ may be taken as a basis for a more general development. Deﬁning vx v þ pﬃﬃﬃﬃﬃﬃﬃﬃﬃ (12-58) t 0 /r we may write equation (12-56) as vþ ¼ 1 ½ln y þ C K (12-59) The left-hand side of (12-59) is.8 1.0 y/h Figure 12.

The constant b in equation (13-19) becomes ln b ¼ pﬃﬃﬃﬃﬃﬃﬃﬃ 3:4 ln ½(e t0 /r)/n for rough tubes. Because of the empirical nature of these equations. it is important to note that wall roughness affects the magnitude of the shear stress in a turbulent ﬂow. and a laminar sublayer. 30 for the laminar sublayer. there are. inconsistencies. a buffer layer.15 Velocity correlation for ﬂow in circular smooth tubes at high Reynolds number (H.13 FURTHER EMPIRICAL RELATIONS FOR TURBULENT FLOW Two important experimental results that are helpful in studying turbulent ﬂows are the power-law relation for velocity proﬁles and a turbulent-ﬂow shear-stress relation due to Blasius. the scale of the roughness e is found to affect the ﬂow in the turbulent core. The velocity is correlated as follows: for turbulent core. . 1943). The velocity gradient. these equations are extremely useful for describing ﬂow in smooth tubes. NACA TM1047. at the center of the tube predicted by (12-63) is not zero. As the wall shear stress appears in the revised expression for ln b. Three distinct regions are apparent: a turbulent core. v þ ¼ 5:5 þ 2:5 ln yþ (12-63) v þ ¼ 3:05 þ 5 ln yþ (12-64) yþ 5 5 > yþ > 0 v þ ¼ yþ (12-65) Equations (12-63) through (12-65) deﬁne the universal velocity distribution.162 Chapter 12 Viscous Flow 25 Laminar layer Buffer layer Turbulant layer 20 Equation (12-63) 15 v+ Equation (12-65) 10 Equation (12-64) Nikuradse Reichardt 5 0 1 2 5 10 20 50 100 200 500 1000 y+ Figure 12. yþ 30 for the buffer layer. of course. but not in the ﬃ laminar sublayer. 12. Both of these relations are valid for ﬂow adjacent to smooth surfaces. In rough tubes. In spite of this and other inconsistencies. for example. Reichardt.

it is found that over much of the cross section the velocity proﬁle may be correlated by y 1/n vx ¼ (12-66) v x max R where R is the radius of the tube and n is a slowly varying function of Reynolds number. The manner of approximation involved in a turbulent analysis differs from that used previously.200. we have seen that the velocity proﬁle depends upon the wall shear stress and that no single function adequately represents the velocity proﬁle over the entire region. The power-law profile has also been found to represent the velocity distribution in boundary layers. The procedure we shall follow in using the von Ka´rma´n integral relation in a turbulent ﬂow is to utilize a simple proﬁle for the integration with the Blasius correlation for the shear stress. For a zero pressure gradient the von Ka´rma´n integral relation is Z t0 d d ¼ v x (v 1 v x ) dy (12-38) r dx 0 Employing the one-seventh-power law for v x and the Blasius relation. the wall shear stress in a turbulent ﬂow is given by 1/4 n 2 (12-68) t0 ¼ 0:0225rv x max v x max ymax where ymax ¼ R in pipes and ymax ¼ d for flat surfaces. is written in place of v x max . the power law is extremely useful in connection with the von Ka´rma´n-integral relation. as we shall see in Section 12. At Reynolds numbers of 105 the value of n is 7. This expression indicates that the velocity gradient at the wall is inﬁnite and that the velocity gradient at d is nonzero. The exponent n is found to vary from a value of 6 at Re ¼ 4000 to 10 at Re ¼ 3. we obtain n 1/4 7 dd (12-70) ¼ 0:0225 v1d 72 dx . In a laminar ﬂow. Performing the indicated integration and differentiation. a simple polynomial was assumed to represent the velocity proﬁle. This leads to the frequently used one-seventh-power law. v 1 . the power law is written y1/n vx ¼ (12-67) v x max d The power-law proﬁle has two obvious difﬁculties: the velocity gradients at the wall and those at d are incorrect.14 The Turbulent Boundary Layer on a Flat Plate 163 For ﬂow in smooth circular tubes. 12.12. v x /v x max ¼ (y/R)1/7 .000. In a turbulent ﬂow. Another useful relation is Blasius’s correlation for shear stress. equation (12-68) for t0. For boundary layers of thickness d. In spite of these inconsistencies.14.14 THE TURBULENT BOUNDARY LAYER ON A FLAT PLATE The variation in boundary-layer thickness for turbulent ﬂow over a smooth ﬂat plate may be obtained from the von Ka´rma´n momentum integral. we see that equation (12-38) becomes Z n 1/4 d d 2 y1/7 y2/7 2 dy (12-69) 0:0225v 1 ¼ v v1d dx 0 1 d d where the free-stream velocity. For pipe-ﬂow Reynolds numbers up to 105 and ﬂat-plate Reynolds numbers up to 107.

a turbulent boundary layer is desired because it resists separation better than a laminar boundary layer.16.16 Comparison of velocity proﬁles in laminar and turbulent boundary layers. they are limited to values of Rex < 107 . they apply only to smooth ﬂat plates.Chapter 12 Viscous Flow which becomes.2 0. . the above equation may be rearranged to give d 0:376 ¼ x Re1/5 (12-72) x The local skin-friction coefﬁcient may be computed from the Blasius relation for shear stress. and is associated with a larger skin friction coefﬁcient. and (12-73). 5 dT = 3. We shall retain the assumption of a completely turbulent boundary layer for the simplicity it affords.4 0.9 dL 4 Turbulent y/d laminar 164 3 2 dT 1 dL Laminar 0 0 0. to give Cfx ¼ 0:0576 Re1/5 (12-73) x Several things are to be noted about these expressions. At the same Reynolds number. The Reynolds number is 500.0 Figure 12. that this assumption introduces some error in the case of a boundary layer that is not completely turbulent. The boundary layer is known to be laminar initially and to undergo transition to turbulent ﬂow at a value of Rex of about 2 105 .6 vx /v• 0.8 1. the turbulent boundary layer is observed to be thicker. equation (12-67). In most cases of engineering interest. Second. First. the reverse is generally true. however. While it would appear then that a laminar boundary layer is more desirable. it is recognized. A comparison of a laminar and a turbulent boundary layer can be made from Blasius’s laminar-ﬂow solution and equations (12-28). x ¼ 0 (a poor assumption).000. by virtue of the Blasius relation. a major assumption has been made in assuming the boundary layer to be turbulent from the leading edge. (12-72). The velocity proﬁles in laminar and turbulent boundary layers are compared qualitatively in Figure 12. Last. upon integration 1/4 n x ¼ 3:45 d5/4 þ C v1 (12-71) If the boundary layer is assumed to be turbulent from the leading edge.

while a variety of factors other than Re actually inﬂuence transition. u. So far the occurrence of transition has been expressed in terms of the Reynolds number alone. Laminar ﬂow has also been seen to undergo transition to turbulent ﬂow at certain Reynolds numbers. The prevailing procedure is to equate the momentum thicknesses.2 indicates the inﬂuence of some of these factors on the transition Reynolds number. Consider a ﬂat plate with transition from laminar ﬂow to turbulent ﬂow occurring on the plate. That is. Concepts of skin friction and drag coefﬁcients were introduced and quantitative relationships were developed for both internal and external ﬂows.15 FACTORS AFFECTING THE TRANSITION FROM LAMINAR TO TURBULENT FLOW The velocity proﬁles and momentum-transfer mechanisms have been examined for both laminar and turbulent ﬂow regimes and found to be quite different.2 Factors affecting the Reynolds number of transition from laminar to turbulent ﬂow Factor Inﬂuence Pressure gradient Favorable pressure gradient retards transition. an integral method was employed. at the transition point.12. Two approaches were employed for analyzing laminar boundary layer ﬂows—exact analysis using boundary layer equations and approximate integral methods. the principal parameter for predicting transition. Concave curvature decreases it Cool walls increase transition Re. however. For turbulent boundary layer analysis along a plane surface. The general approach for turbulent boundary layers with a pressure gradient involves the use of the von Ka´rma´n momentum integral as given in equation (12-46). is equal to the momentum thickness at the end of the laminar portion of the boundary layer. a problem arises in how to join the laminar boundary layer to the turbulent layer at the point of transition.16 Closure 165 It can be seen that the turbulent boundary layer has a greater mean velocity. The Reynolds number remains. The greater momentum and energy permit the turbulent boundary layer to remain unseparated for a greater distance in the presence of an adverse pressure gradient than would be the case for a laminar boundary layer. equation (12-44). Hot walls decrease it Free-stream turbulence Roughness Suction Wall curvatures Wall temperature 12. the momentum thickness. If transition from laminar ﬂow to turbulent ﬂow is assumed to occur abruptly (for computational purposes). decreases transition in external ﬂow Suction greatly increases transition Re Convex curvature increases transition Re. Numerical integration is required. . Table 12. hence both greater momentum and energy than the laminar boundary layer. at the start of the turbulent portion of the boundary layer. 12.16 CLOSURE Viscous ﬂow has been examined in this chapter for both internal and external geometries. Table 12. unfavorable pressure gradient hastens it Free-stream turbulence decreases transition Reynolds number No effect in pipes.

14 What diameter circular plate would have the same drag as the auto of Problem 12. Determine the speed at the critical Reynolds number for a 42-mm-diameter sphere in air.12 The lift coefﬁcient is deﬁned as CL (lift force)/ 1rv 2 A . For a typical subsonic jet.11? 12.-diameter sphere as a function of velocity.18 Golfball ‘‘dimples’’ cause the drag drop (see Figure 12. At a yaw angle of 208. 12.4 0. velocity for a 1. The concepts developed in this chapter will be used to develop important expressions for momentum ﬂow application in the next chapter. 12. Derive the von Ka´rma´n momentum relation. 12. At 95 mph determine a. Determine the friction drag on such an aircraft 12.5 0. the lift coefﬁcient increased to 1.5 Evaluate and compare with the exact solution d. the Reynolds number.4 Find a velocity proﬁle for the laminar boundary layer of the form vx ¼ c1 þ c2 y þ c3 y2 þ c4 y3 v xd when the pressure gradient is not zero.65-in. culminating in expressions for the ‘‘universal’’ velocity distribution. The table below gives the drag coefﬁcient for a rough sphere as a function of the Reynolds number.1 If Reynolds’s experiment was performed with a 38-mmID pipe.5 12. b. 0.16 A 1998 Lexus LS400 has a drag coefﬁcient of 0.6 There is ﬂuid evaporating from a surface at which v x jy¼0 ¼ 0.-diameter sphere in air between velocities of 50 fps and 400 fps.65-in. c. Show several comparison points for a smooth sphere. 12.7 The drag coefﬁcient for a smooth sphere is shown below.17 A baseball has a circumference of 9 14 inches and a weight of 5 14 ounces.4 and the illustration for Problem 12. using a reference area of 2.2 0.13 The auto in Problems and showed a sensitivity to yaw angle. Determine the horsepower required to overcome drag when driving at 70 mph at sea level. 12. Determine the horsepower required to overcome drag at a velocity of 30 m/s.11 A 2007 Toyota Prius has a drag coefﬁcient of 0.3 CD 0.38 15 0.7-mm-diameter cable be in the unsteady wake region of Figure 12. at a speed of 60 mph.22 20 0. the type of ﬂow (see the illustration for Problem ). on a cold winter day T ﬃ 0 F 12.26 at road speeds. 12.0.10 . At what distance from the leading edge will transition occur? 12. 12. the transition or buffer layer.48 10 0.1 0 103 104 105 Re = Dv/n 106 107 Re 104 CD 7. and the turbulent core. PROBLEMS 12.12 25 0.011.28 and a reference area of 2. but v x jy¼0 6¼ 0.4 m2. Compare this ﬁgure with the case of head and tail winds of 6 m/s.8 Plot a curve of drag vs.166 Chapter 12 Viscous Flow Approaches to modeling turbulent ﬂows were introduced. the parasite drag coefﬁcient based on wing area is 0. 12. Similar applications will be developed in the sections related to both heat and mass transfer in later sections of this text. b. at 200 mph at sea level.10 Estimate the drag force on a 3-ft-long radio antenna with an average diameter of 0. at 500 mph at 35 000 ft. The wing area is 2400 ft2.15 Estimate the normal force on a circular sign 8 ft in diameter during a hurricane wind (120 mph).2 in. Plot the drag for1. 0. b. the drag force. using the velocity proﬁle v x ¼ a sin by: 12.2? a.3 Consider the ﬂow of air at 30 m/s along a ﬂat plate. a. What is the lift force at 100 mph for this case? 12. determine the lift force at a road speed of 100 mph.9 For what wind velocities will a 12.21. and CfL for the laminar boundary layer over a ﬂat plate. This approach considers turbulent ﬂows to be described in three parts: the laminar sublayer.2 Modern subsonic aircraft have been reﬁned to such an extent that 75% of the parasite drag (portion of total aircraft drag not directly associated with producing lift) can be attributed to friction along the external surfaces. Cfx.7) to occur at a lower Reynolds number. on a hot summer day T ﬃ 100 F. 12.33 m2. If the lift coefﬁcient for the auto in the previous problem 2 x r is 0. what ﬂow velocity would occur at transition? 12.

12. 167 12.19 The lift coefﬁcient on a rotating sphere is very approximately given by RV RV CL ﬃ 0:24 0:05 over the range of 1:5 > > 0:2: V V Here R is the sphere radius and V is rotation rate of the sphere. 12. determine the thickness of a. 12. dimension. make a dimensionless plot of the momentum and kinetic energy proﬁles in the boundary layer a Reynolds number of 105. near the wall. turbulent ﬂow.26 For a thin a plate 6 in. The ﬂow is parallel to the 6-in.? 12. develop an expression for the local skin-friction coefﬁcient in pipes. In pipes.33 Using the Blasius shear-stress relation (12-68) and the power-law velocity proﬁle. water ﬂows through a copper tube with a 0.20 If the vertical velocity at the wall is not zero such as would be the case with suction or blowing. hot water (T ﬃ 120 F): b. Assume both boundary layers to originate at the leading edge of the ﬂat plate. the buffer layer.28 Estimate the friction drag on a wing by considering the following idealization. 7 ft by 40 ft.15-m pipe at a rate of 0:006 m3 /s. assuming a.17. assuming a.27 Using a sine proﬁle the laminar ﬂow and a one-seventhpower law for turbulent ﬂow. Determine the Reynolds number for a.31 The turbulent shear stress in a two-dimensional ﬂow is given by (t yx )turb ¼ reM @v x ¼ rv x v y @y Expanding v x0 and v y0 in a Taylor series in x and y near the wall and with the aid of the continuity equation @v x0 @v y0 þ ¼0 @x @y show that.24 For the fully developed ﬂow of water in a smooth 0. with a smooth surface. determine the boundary-layer thickness on a ﬂat plate as a function of the Reynolds number and the exponent n. for the powerlaw velocity proﬁle at y ¼ 0 and y ¼ R: 12. the laminar sublayer. at a ﬂow rate of 2 gpm.25 Using Blasius’ correlation for shear stress (equation 1268).22 In a house. Determine the drag. Consider the wing to be a rectangular ﬂat plate. laminar ﬂow.29 Compare the boundary-layer thicknesses and local skinfriction coefﬁcients of a laminar boundary layer and a turbulent boundary layer on a smooth ﬂat plate at a Reynolds number of 106. b. the turbulent core. the average velocity is used for the friction coefﬁcient and the Reynolds number. b.23 Plot the boundary-layer thickness along a ﬂat plate for the ﬂow of air at 30 m/s assuming a.Problems 12. How does this compare with the mixing-length theory? 12. b. For the baseball in Problem 12.32 Evaluate the velocity derivative. cold water (T ﬃ 45 F): 12. eM y3 þ higher order terms in y. @v x /@y. what fraction of the total kinetic energy of the ﬂow is due to the turbulence? 12. what modiﬁcations occur to equation (12-33)? 12. .21 If the turbulence intensity is 10%. a turbulent boundary layer. b. laminar ﬂow. estimate the friction force in air at a velocity of 40 fps. 12.75in.ID. turbulent ﬂow. c. determine the rotation rate for a baseball thrown at 110 mph to have the lift equal the weight. 12. a laminar boundary layer. What would the drag be if laminar ﬂow could be maintained over the entire surface? 12. 12. How many rotations would such a ball make in 60 ft 6 in. Assume turbulent ﬂow to exist over the entire length of the plate. wide and 3 ft long. The wing is ﬂying at 140 mph at 5000 ft.30 Use the 1/7 power-law proﬁle and compute the drag force and boundary layer thickness on a plate 20 ft long and 10 ft wide (for one side) if it is immersed in a ﬂow of water of 20 ft/s velocity. Indicate the probable transition point. Use the one-seventh-power law.

In this chapter. both laminar and turbulent. circular pipe of constant cross section. Some experimental correlations were introduced in Chapter 12 for turbulent ﬂow in or past surfaces of simple geometry. The roughness is included to represent the condition of the pipe surface and may be thought of as the characteristic height of projections from the pipe wall. incompressible ﬂow. D. horizontal. hence the dimension of length. then the groups to be formed are as follows: p1 ¼ v a Db rc DP p2 ¼ v d De r f L p3 ¼ v g Dh ri e p4 ¼ v j Dk rl m 168 . and r. through closed conduits. The signiﬁcant variables and their dimensional expressions are as represented in the following table: Variable Pressure drop Velocity Pipe diameter Pipe length Pipe roughness Fluid viscosity Fluid density Symbol Dimension DP v D L e m r M/Lt2 L/t L L L M/Lt M/L3 Each of the variables is familiar. If the core group consists of the variables v. namely ﬂuid ﬂow. symbolized e. the number of independent dimensionless groups to be formed with these variables is four. an application of the material that has been developed thus far will be considered with respect to a situation of considerable engineering importance. with the exception of the pipe roughness.1 DIMENSIONAL ANALYSIS OF CONDUIT FLOW As an initial approach to conduit ﬂow. we shall utilize dimensional analysis to obtain the signiﬁcant parameters for the ﬂow of an incompressible ﬂuid in a straight. and the like.Chapter 13 Flow in Closed Conduits Many of the theoretical relations that have been developed in the previous chapters apply to special situations such as inviscid ﬂow. 13. According to the Buckingham p theorem.

thus p1 becomes hL v 2 /g The third p group.1 Dimensional Analysis of Conduit Flow 169 Carrying out the procedure outlined in Chapter 11 to solve for the unknown exponents in each group. deﬁned by equation 13-4. hL ¼ f D L v2 D 2g (13-4) Quite obviously. Expressing the head loss from equation (13-2) in terms of f. which varies with the relative roughness and Reynolds number. Another friction factor in common use is the Darcy friction factor. Our task now becomes that of determining suitable relations for ff from that theory and experimental data. will be used exclusively in this text. giving hL L e ¼ f . The student may easily verify that the Fanning friction factor is the same as the skin friction coefficient Cf. we see that the dimensionless parameters become DP rv 2 L p2 ¼ D e p3 ¼ D p1 ¼ and p4 ¼ vDr m The ﬁrst p group is the Euler number. . The fourth p group is the Reynolds number.13. Re 2 v 2 /g D D (13-2) The function f2. fD ¼ 4 ff . A functional expression resulting from dimensional analysis may be written as hL L e ¼ f . This ratio may. Re (13-1) 1 v 2 /g D D Experimental data have shown that the head loss in fully developed ﬂow is directly proportional to the ratio L/D. The student should be careful to note which friction factor he is using to properly calculate frictional head loss by either equation (13-3) or (13-4). this parameter is often written with DP/r replaced by ghL where hL is the ‘‘head loss’’. is the so-called relative roughness. . The Fanning friction factor. be removed from the functional expression. ff. is designated f. then. we have hL ¼ 2 f f L v2 D g (13-3) With the factor 2 inserted in the right-hand side. fD. As the pressure drop is due to ﬂuid friction. equation (13-3) is the deﬁning relation for ff. the friction factor. the Fanning friction factor. Re. the ratio of pipe roughness to diameter.

All development will be based on circular conduits. Forming an expression for frictional head loss from equation (13-5). we obtain ff ¼ 16 m 16 ¼ Dv avg r Re (13-7) This very simple result indicates that ff is inversely proportional to Re in the laminar ﬂow range. No easily derived relation such as the Hagen– Poiseuille law applies. we should expect no difﬁculty in obtaining a functional relationship for ff in the case of laminar ﬂow.170 Chapter 13 Flow in Closed Conduits 13. AND TRANSITION FLOW IN CIRCULAR CONDUITS Laminar Flow Some analysis has been performed already for incompressible laminar ﬂow. Turbulent Flow In the case of turbulent ﬂow in closed conduits or pipes. Recall that. damping out any irregularities in the ﬂow caused by protrusions from a rough surface. for closed conduits. This result has been experimentally veriﬁed and is the manifestation of the viscous effects in the ﬂuid. conduit ﬂow mv avg dP ¼ 32 2 (8-9) dx D Separating variables and integrating this expression along a length. From Chapter 8. TURBULENT. the defining relation for ff hL ¼ 32 mv avg L L v2 ¼ 2 ff 2 D g grD and solving for ff. L. thus we . the relation for ff is not so simply obtained or expressed as in the laminar case. the friction factor is not a function of pipe roughness for values of Re < 2300. but varies only with the Reynolds number. laminar. of the passage. the ﬂow may be considered laminar for values of the Reynolds number less than 2300. As ﬂuid behavior can be described quite well in this regime according to Newton’s viscosity relation. we get Z P Z mv avg L dP ¼ 32 2 dx D P0 0 and DP ¼ 32 mv avg L D2 (13-5) Recall that equation (8-9) held for the case of fully developed flow. we have hL ¼ mv avg L DP ¼ 32 rg grD2 (13-6) Combining this equation with equation (13-3).2 FRICTION FACTORS FOR FULLY DEVELOPED LAMINAR. thus v avg does not vary along the length of the passage. the Hagen–Poiseuille equation was derived for incompressible. some use can be made of the velocity proﬁles expressed in Chapter 12 for turbulent ﬂow. however.

Smooth Tubes.2 from experimental data.2 Friction Factors for Fully Developed Laminar. obtained the equation pﬃﬃﬃﬃﬃ 1 pﬃﬃﬃﬃﬃ ¼ 4:0 log10 Re ff 0:40 (13-12) ff which is very similar to equation (13-11). 356. 1 2 T. we obtain v avg ¼ 2:5 pﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃ t 0 /rR t 0 /r ln þ 1:75 t 0 /r v (13-8) pﬃﬃﬃﬃﬃﬃﬃﬃﬃ The functions t0 /r and Cf are related according to equation (12-2). 1931. As Cf and ff are equivalent. 1932.1 Nikuradse. J. and changing to log10. VDI-Forschungsheft. Turbulent 171 are primarily concerned with pipes or tubes. NACA TM 611. The velocity proﬁle in the turbulent core has been expressed as v þ ¼ 5:5 þ 2:5 ln yþ where the variables v þ and yþ are defined according to the relations v v þ pﬃﬃﬃﬃﬃﬃﬃﬃﬃ t 0 /r and pﬃﬃﬃﬃﬃﬃﬃﬃﬃ t0 /r y y v þ (12-63) (12-58) (12-60) The average velocity in the turbulent core for flow in a tube of radius R can be evaluated from equation (12-63) as follows: RA v dA v avg ¼ 0 A pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃ Z R t0 /ry t0 /r 2:5 ln þ 5:5 2pr dr v 0 ¼ pR2 Letting y ¼ R r. von Ka´rma´n.13. Nikuradse. In turbulent ﬂow a distinction must be made between smooth. The preceding development was first performed by von Ka´rma´n. . we see that equation (13-10) reduces to pﬃﬃﬃﬃﬃ 1 pﬃﬃﬃﬃﬃ ¼ 4:06 log10 Re ff 0:60 (13-11) ff This expression gives the relation for the friction factor as a function of Reynolds number for turbulent flow in smooth circular tubes.and rough-surfaced tubes. we may write v avg 1 pﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ pﬃﬃﬃﬃﬃﬃﬃﬃﬃ (13-9) t0 /r ff /2 The substitution of equation (13-9) into equation (13-8) yields pﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 R pﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 2:5 ln v avg ff /2 þ 1:75 v ff /2 (13-10) Rearranging the argument of the logarithm into Reynolds number form.

(13-14) ! (13-15) . J. The difference is. (London) II. Beyond some value of Re. 133 (1938–39). roughness. An empirical equation describing the variation of ff in the transition region has been proposed by Colebrook. Colebrook. then either equation (13-12) or (13-14) must be used to obtain the proper value for ff. of course.3 1 D D/e pﬃﬃﬃﬃﬃ ¼ 4 log10 þ 2:28 4 log10 4:67 pﬃﬃﬃﬃﬃ þ 1 (13-15) e ff Re ff Equation (13-15) is applicable to the transition region above a value of (D/e)/ p ﬃﬃﬃﬃﬃ (Re ff ) ¼ 0:01. had nothing to do with the transition from laminar to turbulent flow. F. By an analysis similar to that used for smooth tubes. To summarize the development of this section. Inst. that the former equation is for smooth tubes and the latter for rough tubes. von Ka´rma´n developed equation (13-13) for turbulent ﬂow in rough tubes 1 D pﬃﬃﬃﬃﬃ ¼ 4:06 log10 þ 2:16 (13-13) e ff which compares very well with the equation obtained by Nikuradse from experimental data 1 D pﬃﬃﬃﬃﬃ ¼ 4:0 log10 þ 2:28 (13-14) e ff Nikuradse’s results for fully developed pipe flow indicated that the surface condition.172 Chapter 13 Flow in Closed Conduits Rough Tubes. D/e)/(Re ff ) < 0:01) (13-12) 1 D pﬃﬃﬃﬃﬃ ¼ 4:0 log10 þ 2:28 e ff And for transition ﬂow 1 D D/e pﬃﬃﬃﬃﬃ ¼ 4 log10 þ 2:28 4 log10 4:67 pﬃﬃﬃﬃﬃ þ 1 e ff Re ff 3 C. Below this value. that is. and the ﬂow is said to be fully turbulent. this variation deviates from the smooth-tube equation and achieves a constant value dictated by the tube roughness as expressed by equation (13-14). the friction factor is independent of the Reynolds number. Civil Engr. (Re > 3000. the following equations express the friction-factor variation for the surface and ﬂow conditions speciﬁed: For laminar ﬂow (Re < 2300) ff ¼ 16 Re (13-7) For turbulent ﬂow (smooth pipe. The question that naturally arises at this point is ‘‘what is ‘rough’?’’ It has been observed from experiment that equation (13-12) describes the variation in ff for a range in Re. Once the Reynolds number becomes large enough so that flow is fully turbulent. These two equations are quite different in that equation (13-12) expresses ff as a function of Re only and equation (13-14) gives ff as a function only of the relative roughness. Re > 3000) pﬃﬃﬃﬃﬃ 1 pﬃﬃﬃﬃﬃ ¼ 4:0 log10 Re ff 0:40 ff pﬃﬃﬃﬃﬃ For turbulent ﬂow (rough pipe. The region wherein ff varies both with Re and e/D is called the transition region. even for rough tubes.

008 0. e/D Fanning fraction factor.0008 0.003 5 10–5 0. ff 0.0002 0.008 0.006 0.004 0. ASME.004 0.009 0.13.006 0. Trans.002 0.0025 5 10–6 3 10–5 2 10–5 Smooth pipe (e = 0) 0.002 1 10–5 68 2 3 4 68 2 3 4 68 2 3 4 68 2 3 4 68 2 3 4 68 3 4 5 6 7 8 10 10 10 10 10 10 Reynolds number = Dvavg /n Figure 13.0006 0.0015 0.02 0.03 0.3 FRICTION FACTOR AND HEAD-LOSS DETERMINATION FOR PIPE FLOW Friction Factor A single friction-factor plot based upon equations (13-7).001 0.0004 0.01 0.05 0.3 173 Friction Factor and Head-Loss Determination for Pipe Flow 13.1 is a plot of the Fanning friction factor vs.04 0. Relative roughness.1 The Fanning friction factor as a function of Re and D/e. and (13-15) has been presented by Moody.003 Uncertain region 0.025 0.015 0. F.01 .005 0.007 0. 4 L.025 Laminar flow Complete turbulence rough pipe 0. 0. (13-14). 671 (1944). (13-13).0001 0.015 0. the Reynolds number for a range of values of the roughness parameter e/D. Moody.4 Figure 13. 66.02 0.

0005 0.01 0.04 0.0003 e 0.06 0. Figure 13.005 0. et 03 cr 0.000.0008 0.008 5 00 0.000.014 e st ee 0.0002 0.009 e 04 00 0. in in.07 0. its roughness may change considerably.000. (From L. Figure 13.0006 0.008 0. e/D 174 = 0.018 e 00 3 08 5 d al te ci al er m ph 00 0. making the determination of e/D quite difﬁcult.1.008 0.005 1 2 3 4 5 6 8 10 20 30 40 60 80 100 200 300 Pipe diameter.01 0. (1944).000. Moody has presented a chart. After a pipe or tube has been in service for some time.000. D.05 0. Moody. Trans. 00 ro e = rw iro lo D 0.03 0. reproduced in Figure 13. The combination of these two plots enables the frictional head loss for a length.000. 0.04 0.000. F.05 0.016 4f for complete turbulence. 0. L. 5 0.025 = 01 0.2. by which a value of e/D can be determined for a given size tube or pipe constructed of a particular material. tu = n e w ra 05 e 0. 00 0.006 0.012 00 06 ug 0. ro va ti al e as e C G 0. it is necessary to know the value of the roughness parameter that applies to a pipe of given size and material.02 d ze ni n 0.02 = 01 ng 00 bi 0.05 0.03 e C = on 0.2 Roughness parameters for pipes and tubes.000.03 0.000.) Values of e are given in feet.004 0. 00 n iro = 0. = = As om C ca st = e 0.06 0. 0. rough pipes Chapter 13 Relative roughness.003 0.006 0.002 n iro 0.001 0.0001 0. n 1 ht 0.035 Wood stave 0. ASME.04 0.08 0.Flow in Closed Conduits When using the friction-factor plot. using the relation hL ¼ 2 f f L v2 D g (13-3) . of pipe having diameter D to be evaluated.000.0004 0.01 0.02 Riveted steel 0.

Head Losses Due to Fittings The frictional head loss calculated from equation (13-3) is only a part of the total head loss that must be overcome in pipe lines and other ﬂuid-ﬂow circuits. open Standard 908 elbow Short-radius 908 elbow Long-radius 908 elbow Standard 458 elbow Tee. and the roughness. E. Other losses may occur due to the presence of valves. An equivalent method of determining the head loss in ﬁttings is to introduce an equivalent length. and thus the total head loss for a piping system may be determined by adding the equivalent lengths for the fittings to the pipe length to obtain the total effective length of the pipe. the head loss may be evaluated as hL ¼ DP v2 ¼K r 2g (13-16) where K is a coefficient depending upon the fitting. through side outlet Tee. wide open Gate valve. to a ﬁrst approximation. JFE. As the losses in ﬁttings.5%) as 1 6:9 e 10/9 pﬃﬃﬃﬃﬃ ¼ 3:6 log10 þ (13-15a) Re 3:7D ff This expression allows explicit calculation of the friction factor. 1 2 1 4 open Gate valve. The head losses resulting from such ﬁttings are functions of the geometry of the ﬁtting.15 0. 900 32 41 20 15 67 20 75 .4 200 Gate valve. wide open Gate valve. such that hL ¼ 2 f f Leq v 2 D g (13-17) where Leq is the length of pipe that produces a head loss equivalent to the head loss in a particular fitting. Trans.6 S.9 0.5 0.3 Friction Factor and Head-Loss Determination for Pipe Flow 175 Recently Haaland5 has shown that over the range 108 Re 4 104 .13.1 Friction loss factors for various pipe ﬁttings Fitting K Leq /D Globe valve. the Reynolds number. 0:05 e/D 0. straight through 1808 Bend 5 20 0. 105. Table 13. wide open Angle valve. and any other ﬁttings that involve a change in the direction of ﬂow or in the size of the ﬂow passage. Haaland. have been found to be independent of the Reynolds number. 34 open 7.35 1. Leq. the friction factor may be expressed (within 1.8 0.5 3.85 350 170 7 40 4. Equation (13-17) is seen to be in the same form as equation (13-3).4 1. ASME.4 0. 89 (1983).7 0. elbows.

is dependent only upon the roughness of the ﬁtting. The assumption made in both equations (13-16) and (13-17) is that the Reynolds number is large enough so that the ﬂow is fully turbulent.176 Chapter 13 Flow in Closed Conduits Comparison of equations (13-16) and (13-17) shows that the constant K must be equal to 4 ff Leq /D. An equivalent diameter is calculated according to Deq ¼ 4 cross-sectional area of flow wetted perimeter (13-18) The ratio of the cross-sectional area of flow to the wetted perimeter is called the hydraulic radius.1. These equations may be used to estimate the head loss in a closed conduit of any conﬁguration if an ‘‘equivalent diameter’’ for a noncircular ﬂow passage is used. The friction coefﬁcient for a given ﬁtting. The equivalent diameter for this conﬁguration is determined as follows: p. The reader may verify that Deq corresponds to D for a circular ﬂow passage. then. Typical values for K and Leq /D are given in Table 13. with the result given in equation (6-13). One type of noncircular ﬂow passage often encountered in transfer processes is the annular area between two concentric pipes. Although equation (13-17) appears to be dependent upon the Reynolds number because of the appearance of the Fanning friction factor. Recall that the head loss due to a sudden expansion is calculated in Chapter 6. Equivalent Diameter Equations (13-16) and (13-17) are based upon a circular ﬂow passage. it is not.

using the relations and methods developed previously for circular conduits.4 PIPE-FLOW ANALYSIS Application of the equations and methods developed in the previous sections is common in engineering systems involving pipe networks. of the types of problems found in engineering practice. Such analyses are always straightforward but may vary as to the complexity of calculation. 2 D0 D2i 4 Wetted perimeter ¼ pðD0 þ Di Þ Cross-sectional area ¼ yielding p/4 ðD20 D2i Þ ¼ D0 Di (13-19) p ðD0 þ Di Þ This value of Deq may now be used to evaluate the Reynolds number. The following three example problems are typical. Deq ¼ 4 13. the friction factor. and there is an increase in elevation of 2 ft from the inlet of the pipe to its exit. Find the power required to produce this ﬂow rate for the speciﬁed conditions. we obtain ZZZ ZZ dQ dWs dWm P @ ðv : nÞ dA þ ¼ r eþ re dV (6-10) dt r @t dt dt c:s: c:v: . Applying the energy equation to this control volume. EXAMPLE 1 Water at 598F ﬂows through a straight section of a 6-in. The pipe is 120 ft long. The control volume in this case is the pipe and the water it encloses.-ID cast-iron pipe with an average velocity of 4 fps. but by no means all-inclusive. and the frictional head loss.

4 Pipe-Flow Analysis 177 An evaluation of each term yields dQ ¼0 dt dWs _ ¼W dt 2 ZZ v v2 P P2 P1 r eþ ðv: nÞ dA ¼ rAv avg 2 þ gy2 þ þ u2 1 gy1 u1 2 r 2 r r c:s: @ @t ZZZ re dV ¼ 0 c:v: and dWm ¼0 dt The applicable form of the energy equation written on a unit mass basis is now 2 2 _ m_ ¼ v 1 v 2 þ gðy1 y2 Þ þ P1 P2 þ u1 u2 W/ 2 r and with the internal energy change written as ghL . and for a pipe of constant cross section (v 21 v 22 )/2 ¼ 0. _ m_ (P1 P2 )/r ¼ 0. the frictional head loss. giving for W/ _ m_ ¼ g(y1 y2 ) ghL W/ Evaluating hL. 000 1:22 105 e ¼ 0:0017 (from Figure 14:2) D f f ¼ 0:0059 (from equation (14-15a)) yielding hL ¼ 2ð0:0059Þð120 ftÞð16 ft2 /s2 Þ ¼ 1:401 ft ð0:5 ftÞð32:2 ft/s2 Þ The power required to produce the speciﬁed ﬂow conditions thus becomes " # gðð2 ftÞ 1:401 ftÞ 62:3 lbm /ft3 p 1 2 ft _ ft 4 W¼ 550 ft lbf /hp-s s 32:2 lbm ft/s2 lbf 4 2 ¼ 0:300 hp . we have Re ¼ ð12Þð4Þ ¼ 164.13. the expression for w becomes 2 2 _ m_ ¼ v 1 v 2 þ gðy1 y2 Þ þ P1 P2 ghL W/ 2 r Assuming the ﬂuid at both ends of the control volume to be at atmospheric pressure.

). we see that a term by term evaluation gives dQ ¼0 dt dWs dWm ¼0 ¼0 dt dt 2 ZZ v v2 P P2 P1 ðv: nÞ dA ¼ rAv avg 2 þ gy2 þ þ u2 1 gy1 u1 r eþ r 2 r 2 r c:s: ZZZ @ re dV ¼ 0 @t c:v: and the applicable equation for the present problem is 0¼ P2 P1 þ ghL r The quantity desired. solve the above equation for D. 3.132 m (5. is included in the head-loss term but cannot be solved for directly. Calculate Re with this D. as the governing equation 0¼ P2 P1 þ ghL r .-ID pipe. What size pipe is required for this application? Once again.0567 m3/s of water through a smooth pipe with an equivalent length of 122 m. EXAMPLE 3 An existing heat exchanger has a cross section as shown in Figure 13. For a 5-ft length of heat exchanger. applying equation (6-10).-OD tubes inside a 5-in.178 Chapter 13 EXAMPLE 2 Flow in Closed Conduits A heat exchanger is required. which will be able to handle 0. as the friction factor also depends on D. Carrying out these steps for the present problem. A possible procedure is the following: 1. The total pressure drop is 103.2 in. what ﬂow rate of water at 608F can be achieved in the shell side of this unit for a pressure drop of 3 psi? An energy-equation analysis using equation (6-10) will follow the same steps as in example 13. 4.3 with nine 1-in. the diameter. Using this ff. yielding. check the assumed value of ff.000 Pa. Assume a value for ff. we obtain 0¼ 103 000 Pa 0:0567 2 m2 122 m g þ 2 f f 1000 kg/m3 pD2 /4 s2 D m g or 0 ¼ 103 þ 1:27 ff D5 The solution to this problem must now be obtained by trial and error. Inserting numerical values into the above equation and solving. Using e/D and the calculated Re. Repeat this procedure until the assumed and calculated friction factor values agree. the required pipe diameter is 0. 5. 2.2.

13. Calculate v avg from the above expression. 5 ft Figure 13.5 Friction Factors for Flow in the Entrance to a Circular Conduit 179 1 in. and the methods just described will be adequate to evaluate and predict the signiﬁcant ﬂow parameters. giving a ﬂow rate for this problem of 2:06 ft3 /min ð0:058 m3 /sÞ: Notice that in each of the last two examples in which a trial-and-error approach was used. 2. in both cases a value for ff could be assumed within a much closer range than either D or v avg. This was not. Evaluate Re from this value of v avg.13. the assumption of ff was made initially. 4.3 Shell-and-tube head-exchanger conﬁguration.5 FRICTION FACTORS FOR FLOW IN THE ENTRANCE TO A CIRCULAR CONDUIT The development and problems in the preceding section have involved ﬂow conditions that did not change along the axis of ﬂow. we ﬁnd the velocity to be 23. which is a function of v avg. 3. Assume a value for ff. the only way to approach these problems. repeat this procedure until they do. 5 in. a simple trial-and-error procedure such as the following might be employed: 1. however. . of course. 5. Check the assumed value of ff using equation (13-15a). The equivalent diameter for the shell is evaluated as follows: p Flow area ¼ ð25 9Þ ¼ 4p in:2 4 Wetted perimeter ¼ pð5 þ 9Þ ¼ 14p in: thus Deq ¼ 4 4p ¼ 1:142 in: 14p Substituting the proper numerical values into the energy equation for this problem reduces it to 0¼ 3 lbf /in:2 ð144 in:2 /ft2 Þ 5 ft g þ 2 ff v 2avg ft2 /s2 ð1:142/12Þ ft g 1:94 slugs/ft3 or 0 ¼ 223 þ 105 ff v 2avg As ff cannot be determined without a value of Re. This condition is often met.6 fps. Employing this method. If the assumed and calculated values for ff do not agree.

The reader is referred to the work of Deissler7 for experimentally obtained turbulent velocity proﬁles in the entrance region of the circular pipes. Deissler. Observe that the ﬂuid velocity outside the boundary layer increases with x. . ASME. A boundary layer forms on the surface of a pipe. A-55 (1942). Langhaar. The buildup of the boundary layer in pipe ﬂow is depicted in Figure 13. As the friction factor is a function of dv/dy at the pipe surface.4 Boundary-layer buildup in a pipe. G. At some value of x. L. v x Figure 13. derived analytically. the distance downstream from the pipe entrance. as is required to satisfy continuity. NACA TN 2138 (1950). R.4. 64. The velocity proﬁle will not change the downstream from this point. A boundary layer forms on the inside surface and occupies a larger amount of the ﬂow area for increasing values of x. symbolized as Le. Trans. There is no relation available to predict the entrance length for a fully developed turbulent velocity proﬁle. The distance downstream from the pipe entrance to where ﬂow becomes fully developed is called the entrance length. The entrance length required for a fully developed velocity proﬁle to form in laminar ﬂow has been expressed by Langhaar6 according to Le ¼ 0:0575 Re D (13-20) where D represents the inside diameter of the pipe. the boundary layer ﬁlls the ﬂow area. we are also interested in this starting length. A general conclusion of the results of Deissler and others is that the turbulent velocity proﬁle becomes fully developed after a minimum distance of 50 diameters downstream from the entrance. The gradient decreases in the downstream direction. which cause the friction factor to be greater than in fully developed ﬂow. becoming constant before the velocity proﬁle becomes fully developed. and the ﬂow is said to be fully developed. and its thickness increases in a similar manner to that of the boundary layer on a ﬂat plate as described in Chapter 12. The ﬁrst of these is the extremely large wall velocity gradient right at the entrance. This relation.180 Chapter 13 Flow in Closed Conduits In many real ﬂow systems this condition is never realized. Two conditions exist in the entrance region. An additional factor that affects the entrance length in turbulent ﬂow is the nature of the entrance itself. has been found to agree well with experiment. The velocity at the center of the pipe ﬁnally reaches a value of 2v 1 for fully developed laminar ﬂow. The reader should realize that the entrance length for the velocity proﬁle differs considerably from the entrance length for the velocity gradient at the wall. The other factor is the existence of a ‘‘core’’ of ﬂuid outside the viscous layer whose velocity must increase as 6 7 H.

0620 1.5 is a qualitative representation of this variation. there will be some portion of the entrance over which the boundary layer is laminar.000205 0. NACA TN 3016 (1953).1413 0. The entrance conﬁguration. a distance x from the entrance.0341 0. Table 13.5 Friction Factors for Flow in the Entrance to a Circular Conduit Fully developed laminar flow 1 0 x/D 181 Figure 13.2 Average friction factor for laminar ﬂow in the entrance to a circular pipe x x/D ff D Re 0. A plot similar to Figure 13. affects the length of the pipe over which the laminar boundary layer exists before becoming turbulent.00535 0. For turbulent ﬂow in the entrance region.2075 0. The ﬂuid in the core is thus being accelerated. Deissler9 has analyzed this situation and presented his results graphically. . the friction factor as well as the velocity proﬁle is difﬁcult to express.6 for turbulent-ﬂow friction factors in the entrance region. as well as the Reynolds number.01788 0.547 0. thereby producing an additional drag force whose effect is incorporated in the friction factor.308 0.538 8 9 Op cit.00838 0.000830 0.340 0.Friction-factor ratio 13. dictated by continuity.2 gives the results of Langhaar for the average friction factor between the entrance and a location.003575 0.5 is presented in Figure 13.8 His results indicated the friction factor to be highest in the vicinity of the entrance. Figure 13.0530 0.02368 0.0449 1. Even for very high free-stream velocities.5 Velocity proﬁle and friction-factor variation for laminar ﬂow in the region near a pipe entrance.001805 0. G.2605 0. R.0760 1. then to decrease smoothly to the fully developed ﬂow value. The friction factor for laminar ﬂow in the entrance to a pipe has been studied by Langhaar.0965 0. Deissler. Table 13.845 0.461 0.028 0.659 0.01373 0.

07 0.06 4x(r v2/2) D (P1 – Px ) 0.Chapter 13 Flow in Closed Conduits Friction-factor ratio 182 1 Laminar boundary layer 0 0 Fully developed flow (turbulent) Turbulent boundary layer Figure 13. thus the friction factor will be higher than that predicted from the equations for fully developed ﬂow or the friction-factor plot. The student will ﬁnd that he is able to apply much of the information learned in momentum transfer to counterparts in the areas of heat and mass transfer.04 0.6 CLOSURE The information and techniques presented in this chapter have included applications of the theory developed in earlier chapters supported by correlations of experimental data. It is important to realize that in many situations ﬂow is never fully developed. . One speciﬁc type of transfer.02 Reynolds number = Dvavg /n 104 0. circular tube (Deissler). x /D The foregoing description of the entrance region has been qualitative.05 0. has been considered up to this point.7 Static pressure drop due to friction and momentum change in the entrance to a smooth.01 3 104 6 104 105 0 0 2 4 6 8 10 x /D 12 14 16 18 20 Figure 13. horizontal.7 may be utilized. Deissler’s results portrayed in Figure 13. 0.03 0. 13. The chapters to follow will be devoted to heat and mass transfer. For an accurate analytical consideration of a system involving entrance-length phenomena. momentum transfer.6 Velocity proﬁle and friction-factor variation in turbulent ﬂow in the region near a pipe entrance.

13.1 in. diameter $11. and the velocity there is negligible. what is the most economical pipe diameter? The pump efﬁciency is 80%. Determine the tube diameter. 13. If the relative roughness e/D ¼ 0:004. If the ﬂow is fully turbulent.71 m diameter at an average velocity of 1. determine the power required to pump the oil.42 m. The exit of the pipe is submerged at distance h3 ¼ 8 m from the reservoir surface. c.40 per ft $14.11 The siphon of Problem 6. 6 ft long.-ID hose. Determine the ﬂow rate and the pressure at point B. 13. Determine the ﬂow rate. the discharge station is 250 m lower in elevation than the upstream station.31 is made of smooth rubber hose and is 23 ft long. 13. b.55 psi. 13. Determine the total pumping power required when using the modiﬁed pipeline. 13. The pressure in the main line is 60 psig (virtually independent of ﬂow). A pressure drop of 13 psi is obtained at a ﬂow rate of 28:3 lbm /s. a. Neglect changes in elevation throughout the system. 13. diameter 12-in.24 in. Use the (incorrect) assumption that the ﬂow is fully developed. and is 30 in. The oil has a kinematic viscosity of 4:5 106 m2 /s and a density of 810 kg/m3 . Six 908 standard elbows.07 per kilowatt hour over the 20-year life of the pipeline.33 is 35 m long and made of commercial steel. Determine the ﬂow rate through the pipe if the pipe is 80 m long and the friction factor ff ¼ 0:004. A 3/4-in.13 A cast-iron pipeline 2 m long is required to carry 3 million gal of water per day. in diameter at the rate of 10 gal/h.4 A 280-km-long pipeline connects two pumping stations.-ID hose. The pressure drop is 4. square is 25 ft long and carries 600 ft3/min of standard air.10 The pipe in Problem 6. Determine the pressure drop in 50 ft of tube. d. and the water inlet temperature is expected to be constant at 428F. The ﬂow rate through the pump is 500 gal/min. determine the ﬂow rate of the oil.62-m-diameter line. One wide-open angle valve (with no obstruction). b.1 An oil with kinematic viscosity of 0:08 103 ft2 /s and a density of 57 lbm /ft3 ﬂows through a horizontal tube 0. The outlet is 175 ft higher than the inlet.Problems 183 PROBLEMS 13. Consider the faucet to be made up of two parts: (1) a conventional globe valve and (2) a nozzle having a cross-sectional area of 0. The pipe inlet is set ﬂush with the wall. Find the maximum rate of discharge from the faucet. The inlet pressure may be taken as atmospheric.1.3 The pressure drop in a section of a pipe is determined from tests with water. The pipe is constructed of commercial steel. and the discharge pressure is to be maintained at 300. The roughness of the pipe is equivalent to that of a commercial steel pipe. and made of h1 h3 L = 80 m h2 .6 Oil having a kinematic viscosity of 6:7 106 m2 /s and density of 801 kg/m3 is pumped through a pipe of 0.7 The cold-water faucet in a house is fed from a water main through the following simpliﬁed piping system: a.15 Two water reservoirs of height h1 ¼ 60 m and h2 ¼ 30 m are connected by a pipe that is 0. Use the properties given in Problem 13.35 m in diameter. determine the friction factor and ﬂow rate. long. The total pipeline length remains 280 km. A 160 ft length of 3/4-in.8 Water at the rate of 118 ft3/min ﬂows through a smooth horizontal tube 250 ft long. As a ﬁrst try. what will be the pressure drop when liquid oxygen (r ¼ 70 lbm /ft3 ) ﬂows through the pipe at the rate of 35 lbm /s? 13.2 A lubricating line has an inside diameter of 0. If pumping stations are 320 km apart. If the line can be bonded with 6.70 per ft $16. The faucet.1 m/s. commercial steel. diameter 14-in. 13.2.000 Pa.80 per ft 13.0% annual interest. a 10-km-long section of the pipeline is replaced during a repair process with a pipe with internal diameter of 0. The pipe is 6 in.14 Estimate the ﬂow rate of water through 50 ft of garden hose from a 40-psig source for 13. If 0:56 m3 /s are to be pumped through a 0. 13. 13. in diameter.5 In the previous problem.-ID copper pipe leading from the main line to the base of the faucet.12 A galvanized rectangular duct 8 in. A 1/2-in. b. The costs of three sizes of pipe when in place are as follows: 10-in.9 Calculate the inlet pressure to a pump 3 ft above the level of a sump. assume for the pipe ff ¼ 0:007. Power costs are estimated at $0. a.10 in. ﬁnd the head loss (in meters of oil) between the pumping stations and the power required. 13. If the pressure drop is 15 psi. Determine the pressure drop in inches of water.

valves. 13. Assuming a level pipe.16 m.500 m per 300 m of pipe. m Length. 1 2 3 13. 1/2 closed. 1.19 A 2. mm 0.6 m diameter and 30 m long. 1/4 closed. and the pipe discharges to the atmosphere at an elevation 30 m.5 m and a relative roughness of 0. Data for the three pipes are as follows: 13.17 Determine the ﬂow rate through a 0. Data for the three pipes are as follows: Pipe 1 2 3 h Length. The ﬂow rate of water at 208C through the pipes is 0. 5-m-diameter headrace tunnel at the Paute river hydroelectric project in Ecuador supplies a power station 668 m below the entrance of the tunnel.16 An 8-km-long. bends. c. Pipe 1 is 900 m long and has a diameter of 0. 13.18 Water at 208C ﬂows through a cast-iron pipe at a velocity of 34 m/s. mm 125 150 100 8 6 4 0.21 A 15-cm diameter wrought-iron pipe is to carry water at 208C. determine the volumetric ﬂow rate at the discharge if the pressure loss is not permitted to exceed 30.2 m. The pipe is 400 m long and has a diameter of 0. and both run from the same pump to a reservoir.20 m.20 Water at 208C is being drained from an open tank through a cast-iron pipe 0. The pressure drop is 210 kPa and the lines are 150 m long. and ﬁttings. 13. The pipe diameter is 0. determine the difference in surface elevations.30 m. 13.30 cm L For water at 208C. Determine the ﬂow rate of water in each line. Determine the mass ﬂow rate in the pipe in kg/s.120 0.18 m. Each pipe has a length of 312. cm 8 6 4 Roughness.240 0. Pipe 2 has a length of 1500 m and a diameter of 0.240 0. Determine the volumetric ﬂow rate of the water leaving the pipe.200 1.30 cm commercial steel pipe.28 A system consists of three pipes in parallel with a total head loss of 24 m.9 m. The manometers indicate pressure head of 1.2-m-diameter cast-iron pipe and a 67-mmdiameter commercial steel pipe are parallel. Diameter. 3/4 closed.2 m and the pipe roughness is 0.0 kPa per 100 m.18 m3/s. if one pipe has a diameter of 0. cm Roughness. The surface of the water in the pipe is at atmospheric pressure and at an elevation of 46. . 13. The head loss due to friction is 0.22 A level 10-m-long water pipe has a manometer at both the inlet and the outlet. The total pressure drop is 180 kPa. with a total head loss of 18 m for both pipes.184 Chapter 13 Flow in Closed Conduits 13. neglect minor losses and determine the volumetric ﬂow rate in the system.23 Determine the depth of water behind the dam in the ﬁgure that will provide a ﬂow rate of 5:675 104 m3 /s through a 20-m-long. Pipe 13. d. 13.25 m3/s from reservoir 1 to reservoir 2 through three concrete pipes connected in series.27 A 0. open.20-m diameter pipe carries water at 158C. respectively. determine the diameter of the other. determine the volumetric ﬂow rate of the water leaving the pipe. Neglecting minor losses due to conﬁguration.120 0. Neglecting minor losses. ﬁnd the pressure at the end of the tunnel if the ﬂow rate is 90 m3 /s.25 A system consists of three pipes in series.24 Water ﬂows at a volumetric ﬂow rate of 0. Determine the head loss due to friction.200 13. Neglecting minor losses.18 m.0035 m.26 Two concrete pipes are connected in series.0004 m. If the tunnel surface is concrete. 13. and the decrease in elevation is 5 m. 13. Pipe 3 is 800 m long and the diameter is 0.5 and 0. b. m 100 150 80 Diameter.2-m gate valve with upstream pressure of 236 kPa when the valve is a.

a ﬂuid is conﬁned in a chamber whose volume is varied.Chapter 14 Fluid Machinery In this chapter we will examine the operating principles of mechanical devices that exchange ﬂuid energy and mechanical work. thereby generating ﬂow.1. There are two principal types of ﬂuid machines—positive displacement machines and turbomachines. 185 . Examples of positive-displacement-type machines are shown in Figure 14. A pump is a machine whose purpose is to apply mechanical energy to a ﬂuid.1 Some examples of positive-displacement conﬁgurations. or producing a higher pressure. A turbine does just the opposite—producing work through the application of ﬂuid energy. In positive displacement machines. Aorta To body Right atrium Pulmonary artery To lungs Oxygenated blood from lungs Left atrium Suction Discharge Venous blood Right ventricle Left ventricle Figure 14. or both.

and their compatibility with piping systems. In the impulse turbine. The term pump is generally used when the working ﬂuid is a liquid.2. as previously stated. Window fans and aircraft propellers are examples of unshrouded turbomachines. In reaction turbines the ﬂuid ﬁlls the blade passages and a pressure decrease occurs as it ﬂows through the impeller. involve rotary motion. Consideration will be given to general pump and fan performances.2 Turbomachines. The energy transfer in such devices involves some thermodynamic considerations beyond simple momentum analysis. They are of two primary types. ﬂuid enters the pump casing axially. the following terms are used: Fans are associated with relatively small pressure changes. 14. the jet ﬂow is essentially at a constant pressure. If the ﬂuid is a gas or vapor. which convert ﬂuid energy into mechanical work in different ways. In this conﬁguration. In this conﬁguration. The two general types of pumps in this category are shown in Figure 14.186 Chapter 14 Fluid Machinery Outlet Rotor Rotor Inlet Inlet w Stator Outlet Housing or casing Radial or centrifugal flow Housing or casing Axial flow Figure 14. It then encounters the impeller blades that direct the ﬂow tangentially and radially outward into the outer part of the casing and is then discharged. with DP up to 2:8 m of H2 O (40 psi): Compressors are of both positive and variable conﬁgurations having delivery pressures as high as 69 MPa (103 psi): Turbines. The designations radial ﬂow and axial ﬂow refer to the direction of ﬂuid ﬂow relative to the axis of rotation of the rotating element.3 shows two cutaway views of a typical centrifugal pump. extract energy from high-pressure ﬂuids. into a high-velocity jet. on the order of DP 35 cm of H2 O (0:5 psi): Blowers are of both positive and variable displacement types. scaling laws. Turbomachines. The basic analysis of these devices is examined in Chapter 5. as the name implies. The ﬂuid experiences an increase in velocity and pressure as it passes through .1 CENTRIFUGAL PUMPS Figure 14. Pumps used with liquids generally have shrouds that conﬁne and direct the ﬂow. impulse and reaction. This jet then strikes the turbine blades as they pass. The remainder of this chapter will be devoted entirly to pumps and fans. the high-energy ﬂuid is converted. by means of a nozzle.

becomes v 22 v 21 dWs _ ¼ m h 2 h1 þ þ g(y2 y1 ) dt 2 It is customary to neglect the small differences in velocity and elevation between sections one and two.14. designated h. which is doughnut shaped. For a centrifugal pump the efﬁciency.3 by dashed lines. causes the ﬂow to decelerate and the pressure to increase further. the impeller. Note that ﬂow enters at section one and leaves at two.3 Cutaway views of a centrifugal pump. Pump Performance Parameters We now focus on the control volume designated in Figure 14. the efﬁciency. adiabatic flow with no viscous work. can now be expressed in broad terms as the ratio of actual output to required input. which is the most common conﬁguration. is h¼ power added to the fluid shaft power to the impeller . Applying the ﬁrst law of thermodynamics to this control volume we have ZZ ZZZ dQ dWs P @ dWm : ¼ (6-10) eþ er dv þ r(v n) dA þ dt dt dt r @t c:s: c:v: which for steady. we write u2 u1 ¼ h L The net pressure head produced in the pump is P2 P 1 1 dWs ¼ hL _ dt m r (14-1) An important performance parameter. thus v 22 v 21 0 and y2 y1 0 and the remaining expression is dWs P P1 _ 2 h1 Þ ¼ m _ u2 u 1 þ 2 ¼ mðh dt r Recalling that the term u2 u1 represents the loss due to friction and other irreversible effects. The impeller blades shown have a backward-curved shape. The discharge section.1 Centrifugal Pumps 187 2 Casing Impeller 1 w Control volume Expanding area scroll Figure 14.

b2 are the angles made between the blade and tangent directions at r1 and r2. respectively v n2 is the normal velocity of the ﬂow at r2 v t2 is the tangential velocity of the ﬂow at r2 b1 . The governing equation for this analysis is ZZ ZZZ @ (r v)z r(v n) dA þ (r v)z r dv (5-10c) åMz ¼ @t c:s: c:v: The axis of rotation of the rotor depicted in Figure 14. In Figure 14. We now wish to solve for Mz by applying equation (5-10c) to the control volume in Figure 14. Recall that the rotor contains backward-curved blades. v b2 represent velocities along the blade at r1 and r2. (14-2). hL. The blade is attached to the rotor hub at distance r1 from the z axis. To develop actual performance information for centrifugal pumps. the outer dimension of the blade has the value r2. In this ﬁgure v b1. and (14-3) provide general relationships for important pump performance parameters.4. we show a detailed view of a single rotor blade. υt2 = r2w υn2 υb2 υt1 = r1w υb1 B2 r2 r1 b1 Figure 14.188 Chapter 14 Fluid Machinery The power added to the ﬂuid is given by equation (14-1) dW P 2 P1 _ ¼m r dt fluid (14-2) and the efficiency can be expressed as h¼ _ 2 P1 ) m(P r(dWs /dt)c:v: (14-3) The difference between dWs /dtjc:v: and dWs /dtjliquid is clearly the head loss. we must examine our control volume once again from a moment of momentum perspective.3 for one-dimensional steady ﬂow.4 Velocity diagram for ﬂow exiting a centrifugal pump impeller. Equations (14-1).3 has been chosen as the z direction. . The coordinate system will be ﬁxed with the z direction along the axis of rotation. hence our choice of equation (5-10c). respectively.

v u2 . we evaluate the following: the normal velocity of ﬂow at r2 v n2 ¼ V_ 2pr2 L the velocity of ﬂow along the blade at r2 v n2 sinb2 v b2 ¼ (14-5) (14-6) the blade tip velocity v t2 ¼ r2 v (14-7) The velocity we want.1 Centrifugal Pumps 189 Equation (5-10c) can now be written as 2 er 4 _ r Mz ¼ m vr eu 0 vu ez 0 v z 2 er r vr eu 0 vu 3 ez 0 5 v z 1 z which becomes _ Mz ¼ m[(rv u )2 0] _ 2 v u2 ¼ r Vr (14-4) The velocity. can now be evaluated as v u2 ¼ v t2 v b2 cosb2 Substitution from equations (14-6) and (14-7) yields v n2 cosb2 sinb2 ¼ r2 v v n 2 cotb2 v u2 ¼ r2 v Finally. introducing the expression for v n2 from equation (14-5) we have v u ¼ r2 v V_ cotb2 2pr2 L (14-8) and the desired moment is _ Mz ¼ rVr2 r2 v V_ cotb2 2pr2 L (14-9) . The quantities shown in Figure 14.4 will be useful in evaluating v u2 . v 2. normal to the plane of Figure 14. is the vector sum of the velocity relative to the impeller blade and the velocity of the blade tip relative to our coordinate system. v u2 . For blade length.4. L. is the tangential component of the fluid stream exiting the rotor relative to the fixed coordinate system.14. The absolute velocity of the existing ﬂow.

v. when v r1 ¼ r1 v tanb1 (14-11) Typical performance curves. with mass flow rate rV: This expression may be related to equations (14-2) and (14-3) to evaluate the imparted pressure head and the pump efﬁciency. equivalently. h Head Efficiency Brake horsepower Normal or design flow rate 0 0 Figure 14.5. V Example 1 illustrates how the analysis presented above relates to centrifugal pump performance. Referring to Figure 14. such that inlet ﬂow is along the blade surface. when v r1 ¼ v b1 sinb1 ¼ r1v sinb1 cosb1 and. Shutoff head Head. b2 . for a centrifugal pump.5 Centrifugal pump performance curves. ﬁnally.190 Chapter 14 Fluid Machinery The power delivered to the fluid is.4. bhp Efficiency. by definition. the design point for minimum losses is achieved when v b1 cosb1 ¼ r1 v or. are shown in Figure 14. This is accomplished by conﬁguring the angle. · Flow rate. ha Brake horsepower. EXAMPLE 1 Water ﬂow is produced by a centrifugal pump with the following dimensions: r1 ¼ 6 cm r2 ¼ 10:5 cm L ¼ 4:75 cm b1 ¼ 33 b2 ¼ 21 . and L. the radial location at which ﬂow enters the impeller. Pressure head. brake horsepower. b1 . It is reasonable to choose operating conditions at or near the flow rate where maximum efficiency is achieved. operating at angular velocity. thus V_ dWs _ _ cotb2 W¼ ¼ Mz v ¼ rVr2 v r2 v 2prL dt (14-10) Equation (14-10) expresses the power imparted to the fluid for an impeller with _ dimensions r2 . and efficiency are all shown as functions of volumetric flow rate. It is a standard practice to minimize friction loss at r1. Mz v.

1) expresses the net pressure head as _ W P2 P1 hL ¼ _ mg rg (b) (14-1) The maximum value.1 Centrifugal Pumps 191 at a rotational speed of 1200 rpm determine (a) the design ﬂow rate. To experience minimum losses. . equation (14-11) must be satisﬁed. (b) the power added to the ﬂow.14. with negligible loses. (c) the maximum pressure head at the pump discharge. thus v r1 ¼ r1 v tanb1 rev 2p rad min ¼ (0:06 m) 1200 (tan 33 ) min rev 60 s (14-11) ¼ 4:896 m/s The corresponding ﬂow rate is V_ ¼ 2pr1 Lv r1 ¼ 2p(0:06 m)(0:0475 m)(4:896 m/s) ¼ 0:0877 m3 /s (1390 gpm) (a) The power imparted to the ﬂow is expressed by equation (14-10): _ _ ¼ rVr _ 2 v r2 v V cotb2 W 2prL Evaluating the following: v¼ rev rad min 2p ¼ 125:7 rad/s 1200 min rev 60 s _ 2 v ¼ (1000 kg/m3 )(0:0877 m3 /s)(0:105 m)(125:7 rad/s) rVr ¼ 1157 kg m/s V_ 0:0877 m3 /s ¼ 2pr2 L 2p(0:105 m)(0:0475 m) ¼ 2:80 m/s we obtain _ ¼ (1157 kg m/s)[ð0:105 m)(125:7 rad/s) (2:80 m/s)(cot 21)] W ¼ 6830 W ¼ 6:83 kW Equation (14. will be P2 P1 6830 W ¼ rg (1000 kg/m3 )(0:0877 m3 /s)(9:81 m/s2 ) ¼ 7:94 m H2 O gage For P1 ¼ 1 atm ¼ 14:7 psi ¼ 10:33 m H2 O P2 ¼ (7:94 þ 10:33)m H2 O ¼ 18:3 m H2 O (26 psi) (c) The actual discharge pressure will be less than this owing to friction and other irreversible losses.

y.192 Chapter 14 Fluid Machinery Net Positive Suction Head A major concern in pump operation is the presence of cavitation. Figure 14. The parameter designated net positive suction head (NPSH) characterizes the likelihood for cavitation to occur. Cavitation occurs when a liquid being pumped vaporizes or boils. often. structural damage to the pump that may lead to catastrophic failure.7.6. 2 NPSH (m or ft) y2 P 1 = P atm Reference plane 1 · V (m3/s. An energy balance between the pump inlet and the reservoir level yields Patm P2 v 2 ¼ y2 þ þ 2 þ åhL rg rg 2g (14-13) where the term åhL represents head losses between locations 1 and 2 as discussed in Chapter 13. determined experimentally over a range in flow rates. the NPSH can be expressed as NPSH þ Pv v 2i Pi ¼ þ rg 2g rg (14-12) where v i and Pi are evaluated at pump inlet and Pv is the liquid vapor pressure. the value of NPSH evaluated using equation (14-14) should be greater than the value obtained from a pump performance plot at the same flow rate. Example 2 illustrates the use of NPSH. . a representative pump installation is shown with the liquid being drawn from a reservoir located a distance. for a given pump. the vapor bubbles that have been formed cause a decrease in efﬁciency and. Values of NPSH are. If this occurs. At the suction side of the impeller. V_ is shown in Figure 14. y2. below the pump inlet. A typical variation of NPSH vs. In Figure 14.7 Pump installation at a level y above a supply reservoir. The principal use of these ideas is to establish a maximum value for the height. 9 gpm) Figure 14. Combining this relationship with equation (14-12) we get v 22 P2 Pv þ 2g rg rg Patm Pv y2 åhL ¼ rg rg NPSH ¼ (14-14) For proper pump installation. where pressure is lowest thus the location where cavitation will ﬁrst occur.6 Typical variation of NPSH ˙ with V . in general.

The inlet pipe to the centrifugal pump is 12 cm in diameter and the desired ﬂow rate is 0. 1 and 1 2 as shown in the ﬁgure. y2 operating efﬁciency. are r ¼ 997 kg/m3 Pv ¼ 3598 Pa and we have v¼ 0:0025 M3 /s V_ ¼ 2:21 m/s ¼ p A (0:12 m)2 4 åhL ¼ KL v 2 12(2:21 m/s2 )2 ¼ ¼ 2:99 m 2g 2(9:81 m/s2 ) We can now complete the solution y¼ (101360 3598)Pa 2:99 m 4:2 m (997 kg/m3 )(9:81 m/s2 ) ¼ 2:805 m Combined Pump and System Performance (9:2 ft) 2 As depicted in Figure 14. Piping system perfor. all being ﬂow-rate-dependent.Figure 14. and NPSH values. an energy With the two reservoir surfaces designated 1 balance between these two locations yields _ P 2 P1 W þ (u2 u1 ) ¼ g( y2 y1 ) þ (14-15) _ m r . A simple ﬂow system is illustrated in Figure 14. to the performance of the system in which the pump produces ﬂow. The quantity desired.14.7 is to be assembled to pump water. The minor loss coefﬁcient for the system may be taken as K ¼ 12: Water properties are to be evaluated at 300 K. a pump has 1 the capability of operating over a range in ﬂow rates with its delivered head. At this ﬂow rate. the distance between pump inlet and reservoir level.5.025 m3/s. mance is discussed in Chapter 13. with its known operating characteristics. at 300 K. is given by y¼ Patm Pv åhL NPSH rg (14-14) Water properties required.8 Pumping system conﬁguration. y. the speciﬁcations for this pump show a value of NPSH of 4. Determine the maximum value of y. An y1 Pump important task of the engineer is to match a given pump.2 m.8 where a pump is used to produce ﬂow between two reservoirs at different elevations.1 EXAMPLE 2 Centrifugal Pumps 193 A system like the one shown in Figure 14.

2 SCALING LAWS FOR PUMPS AND FANS The concepts of similarity and scaling are introduced in Chapter 11. % Total head. the ﬂow rate of maximum pump efﬁciency. is a useful tool in generating the dimensionless parameters that apply to rotating ﬂuid machines. or to pump operating conditions. which is generally a difﬁcult process.50 m3/s NPSH. we have _ W ¼ y2 y1 þ åhL _ mg (14-16) From Chapter 13.0 Figure 14. .9. A system designer would. 14. changes must be made either to the system.9 Combined pump and system performance. elbows. m Pump efficiency. size. 150 System operating line 100 11 h 10 50 9 NPSH 8 0 0 0. or as near as possible to. naturally.50 1. The operating line for system performance is now expressed by åhL ¼ åK _ W v2 ¼ y2 y1 þ åK _ 2g mg (14-17) Plotting the system operating line together with the plot of pump performance yields the combined performance diagram as shown in Figure 14. which is introduced in Chapter 11. want the system to operate at. In this section we will develop the ‘‘fan laws’’ that are used to predict the effect of changing the ﬂuid. Dimensional Analysis of Rotating Machines The Buckingham method of dimensional analysis. If the operating point corresponds to an undesirable efﬁciency value. one can read the corresponding efﬁciency from the chart. we can write for the head loss v2 2g where the quantity åK accounts for frictional pipe loss as well as minor losses due to valves.0 Flow rate. kinematic. m Head 2. and ﬁttings.194 Chapter 14 Fluid Machinery Observing that P1 ¼ P2 ¼ Patm and expressing u2 u1 ¼ åhL . which are in a geometrically similar family. or speed of rotating machines. We note that the two operating lines intersect at a ﬂow rate where the required head for system operation matches that which the particular pump can produce. The requirements of geometric. At this operating ﬂow rate. V 1. and dynamic similarity ﬁnd important applications in the scaling of rotating ﬂuid mechanics.

08 CQ 0. v.1 Pump performance variables Variable Total head Flow rate Impeller diameter Shaft speed Fluid density Fluid viscosity Power Symbol Dimensions gh V_ L2 /t2 L3 /t L 1/t M/L3 M/Lt ML2 /t3 D v r m _ W Without repeating all details regarding the Buckingham method. for a representative centrifugal pump family.25 1 0 CP and efficiency 1. we can establish the following: i¼nr ¼73¼4 with a core group including the variables D. the efficiency. h.50 2 the head coefficient (14-18) 3 _ p2 ¼ V/vD ¼ CQ the flow coefficient (14-19) 3 5 _ p3 ¼ W/rv D ¼ CP the power coefficient (14-20) Figure 14.75 CP 3 0. The efﬁciency is related to the other parameters deﬁned above according to 0. Table 14.14. and is included as one of the dependent variables in Figure 14.12 0 0. one additional dimensionless performance parameter. CQ. the efﬁciency. CH CQ CP (14-21) As the parameters on the right-hand-side of the equation are functionally related to CQ.10 Dimensionless performance curves for a typical centrifugal pump. Table 14. There is. The other three groups here are designated. h 4 CH 0.2 Scaling Laws For Pumps And Fans 195 As discussed earlier. as 6 CH 5 Efficiency. the ﬂow coefﬁcient.1 lists the variables of interest along with their symbols and dimensional representation in the MLt system. . r. the ﬁrst step to be undertaken is to develop a table of variables that are important to our application. of course.00 p1 ¼ gh/D2 v2 ¼ CH h¼ 0 0.10 is a plot of the dimensionless parameters CH and CP vs.10. by the pump community. the dimensionless pi groups become p1 ¼ gh/D2 v2 3 _ p2 ¼ V/vD 3 5 _ p3 ¼ W/rv D p4 ¼ m/D2 vr The group p4 ¼ m/D2 vr is a form of Reynolds number. is also a function of CQ.04 0.16 Figure 14.

2 2 h2 v2 D2 ¼ h1 v1 D1 1300 rpm 2 1:3 D1 2 ¼ 120 m H2 O 1100 rpm D1 ¼ 283 m H2 O (b) (14-22) (c) . (a) For a geometrically similar pump.196 Chapter 14 Fluid Machinery The three coefﬁcients CH . what ﬂow rate will be achieved? (b) If the new larger pump described in part (a) is also operated at 1300 rpm. we may write.85 m3/s. then for the larger pump. we have. we obtain V_ 2 v2 D2 3 ¼ (14-23) V_ 1 v1 D1 P2 r2 v2 3 D2 5 ¼ (14-24) P1 r1 v1 D1 These three equations comprise the ‘‘fan laws’’ or ‘‘pump laws’’ that are used extensively for scaling rotating machines as well as predicting their performance. designated 1 and 2. CQ . from equation (14-23) 1300 rpm 1:3 D1 3 V_ 2 ¼ 0:85 m3 /s 1100 rpm D1 ¼ 2:207 m3 /s The new head is determined using equation (14-22). D2 ¼ 1:3 D1 . Example 3 illustrates the use of these expressions. using equation (14-23) V_ 2 v2 D2 3 ¼ (14-23) V_ 1 v1 D1 3 1:3 D1 V_ 2 ¼ 0:85 m3 /s D1 (a) ¼ 1:867 m3 /s For the case with D2 ¼ 1:3 D1 and v2 ¼ 1300 rpm. for CH CH1 ¼ CH2 or gh1 gh2 ¼ 2 2 2 2 v1 D1 v2 D2 thus 2 2 h2 v2 D2 ¼ (14-22) h1 v1 D1 Performing the same equations on CQ and CP . operating at the same speed but with an impeller diameter 30% greater than the original. EXAMPLE 3 A centrifugal pump. D ¼ D1 . what will be the new values of ﬂow rate and total head? Specifying for pump 1. For similar pumps. thus the new ﬂow rate will be. and CP provide the basis for the fan laws. operating at 1100 rpm against a head of 120 m H2O produces a ﬂow of 0.

0 00 . or mixed ﬂow conﬁgurations depends on the desired values of ﬂow rate and head needed in a speciﬁc application. 14. ﬂow is in the direction of the axis of rotation. other than that presented in Chapter 5. The momentum exchange produced as the ﬂuid changes direction generates power at the rotor shaft. designated mixed ﬂow. 1. A detailed discussion of turbine operation. producing power from a high-energy ﬂuid.5 Closure 197 14. ﬂow is turned 90 to the axis of rotation.ft)3/4 00 . It is deﬁned as NS ¼ 1/2 CQ C 3/4 (14-25) H Figure 14. The choice of centrifugal. in the axial ﬂow case. where the ﬂow has both normal and axial components.0 hmax 0. The reader is referred to Section 5. External power applied to pumps and fans produces higher pressure. increased ﬂow. (1998).11 Optimum pump efﬁciency as a function of speciﬁc speed. whereas lower heads and higher ﬂow rates require mixed ﬂow or axial ﬂow pumps. Numerous treatises are available to the interested reader.7 Axial flow (rpm) (gal/min)1/2(H. The designation centrifugal ﬂow or axial ﬂow relates to the direction of ﬂuid ﬂow in the pump. is beyond the scope of this book. is included in the text by Munson et al. A good introductory discussion.4 TURBINES Analysis of turbines follows the same general steps as has been done for pumps.3 AXIAL AND MIXED FLOW PUMP CONFIGURATIONS Our examination of pumps.9 Mixed flow 0.8 Centrifugal pump 0. emanating from a nozzle. In the centrifugal case. The values of NS shown in this plot correspond to the somewhat unusual units shown. designated the rotor. There is an intermediate case.11 is that higher delivery head and lower ﬂow rate combinations dictate the use of centrifugal pumps.6 Figure 14. along with extensive references.5 CLOSURE This chapter has been devoted to the examination of rotating ﬂuid machines.4 in Chapter 5 for a review of the analysis of an impulse turbine. axial. The single parameter that includes both head and ﬂow rate effects is designated NS. . or both.14. the speciﬁc speed. The basic message conveyed by Figure 14. The other basic conﬁguration is axial ﬂow. 14. Turbines operate in the reverse. Turbine operation uses the energy of a ﬂuid. to interact with blades attached to the rotating unit.0 15 Ns 10 00 00 50 40 00 20 00 10 50 0 0. thus far. has focused on centrifugal pumps.11 is a plot of optimum efﬁciencies of the three pump types as functions of NS.

2 A centrifugal pump is used with gasoline ðr ¼ 680 kg/m3 Þ: Relevant dimensions are as follows: d1 ¼ 15 cm. Determine (a) the ﬂow rate. The resulting ‘‘fan laws’’ that relate two similar systems are 2 2 h2 v2 D2 ¼ (14-22) h1 v1 D1 V_ 2 v2 D2 3 ¼ (14-23) V_ 1 v1 D1 P2 r2 v2 3 D2 5 ¼ (14-24) P1 r1 v1 D1 PROBLEMS 14. The dimensions are as the following: V1 28 cm 8 cm 1020 rpm 1. Standard performance plots for a family of geometrically similar pumps or fans show the head.5 A centrifugal pump is being used to pump water at a ﬂow rate of 0.45 m. and the head generated is 52 m of water.5 kW. 14. In centrifugal pumps. 14. determine the power required to drive the pump. and the exiting velocity may be assumed to be tangent to the vane at its trailing edge. Assuming radial entry ﬂow.3 A centrifugal pump has the following dimensions: d2 ¼ 42 cm. (b) the power delivered to the gasoline. power. determine the head generated by the pump. determine the theoretical values for (a) the ﬂow rate and (b) the power. (b) the water horsepower. and (c) the head in meters.4 A centrifugal pump has the conﬁguration and dimensions shown below.6 A centrifugal pump having the dimensions shown develops a ﬂow rate of 0. Determine (a) the torque and power required to drive the pump and (b) the maximum pressure increase across the pump. and (c) the proper blade angle at the impeller inlet. Scaling laws were developed using parameters generated from dimensional analysis.2 m3/s of water when operating at 850 rpm. b1 ¼ 25 . and (c) the discharge head. blade length ¼ 50 cm. Machines with both centrifugal and axial ﬂow components are designated mixed-ﬂow pumps. and b2 ¼ 40 : The gasoline enters the pump parallel to the pump shaft when the pump operates at 1200 rpm. and blade exit angle ¼ 24 . and b2 ¼ 33 : It rotates at 1200 rpm. ﬂow is parallel to the ﬂow axis in axial ﬂow pumps. 14. efﬁciency. Relevant impeller dimensions are as follows: outside diameter ¼ 0. d2 ¼ 28 cm. The inlet ﬂow is directed radially outward. L ¼ 5 cm. L ¼ 9 cm. 14. 6.0071 m3/s and an impeller speed of 1020 rpm.032 m3/s when pumping gasoline ðr ¼ 680 kg/m3 Þ: The inlet ﬂow may be assumed to be radial.198 Chapter 14 Fluid Machinery Types of pumps or fans are characterized by the direction of ﬂow through the rotor. (b) the head increase. and NPSH as functions of ﬂow rate for a designated speed of rotation. For water ﬂowing at a rate of 0. 5 cm 35° 10 cm 1650 rpm 14.0071 m /s 55° 14. Estimate (a) the theoretical horsepower.85 cm r1 ¼ 12 cm r2 ¼ 20 cm L ¼ 4:2 cm b1 ¼ 32 b2 ¼ 20 Determine (a) the design point discharge rate. . If the pump efﬁciency is 63%.018 m3/s and the required power is measured to be 4. the ﬂow is turned 90 to the ﬂow axis.5 cm 3 V = 0.7 A centrifugal water pump operates at 1500 rpm.1 A centrifugal pump delivers 0.

2 0 0. for a desired ﬂow rate of 14.8 for a pump diameter of 0.1 0. and (c) discharge ﬂow rate should be expected at maximum efﬁciency? 14. (a) the pump speed required and (b) the rate of discharge.14 is used to pump water from one reservoir to another that is 95 m higher in elevation.8 h 0 0. (c) the pressure rise. 14.6 CP 3 CP 0. h 4 0. and (d) the brake horsepower. at these new conditions.Problems 6 CH 5 1.2 Head. For a pump from this family with a characteristic diameter of 0. 199 0.20 Discharge.19 The pump having the characteristics shown in Problem 14.17 The pump having the characteristics shown in Problem 14.16 For the pump having the characteristics shown in Problem 14.18 If the pump having the characteristics shown in Problem 14.40 m operating at 2200 rpm. 14. Estimate (a) the impeller diameter and (b) the maximum pressure rise.14 Performance curves for an operating centrifugal pump are shown below in both conventional units and in dimensionless form.05 0. what will be (a) the new discharge ﬂow rate and (b) the power required at this new speed? 14. 14.13 Rework Problem 14.45 m operating at maximum efﬁciency and pumping water at 15 C with a rotational speed of 1600 rpm.00 Efficiency.4 0. Power.75 CP 3 0. m3/s 0.30 m3/s at 1800 rpm.16 0. What discharge rate is to be expected if the head developed is 410 m? 14.14.0 h.10 Rework Problem 14.04 0.11 Rework Problem 14.3 CQ Power input 150 Efficiency.9 0.14 was used as a model for a prototype that is to be six times larger.14 is to be operated at 800 rpm. estimate (a) the head.35 m operating at 2400 rpm.8 The ﬁgure below represents performance. 14.50 2 1.201 m3/s at 1800 rpm.9 0 h 0. (b) head.7 7 0. m.25 14. (b) the discharge rate.12 Rework Problem 14. Determine.8 0.2 m3/s when operating at best efﬁciency and a rotational speed of 1400 rpm.12 0 0. what (a) power. % 5 0 0 0.10 0. what will be the discharge rate and head when operating at maximum efﬁciency? 14.14 is tripled in size but halved in rotational speed. kW 200 80 Efficiency 60 Head 100 40 w = 360rpm 50 20 d = 39cm 0.15 0. Cp CH 14.6 6 CH 250 100 CH 4 0. operating at maximum efﬁciency with the speed increased to 1000 rpm. If this prototype operates at 400 rpm. The pump is used to pump water at maximum efﬁciency at a head of 90 m.8 for a pump diameter of 0. for a family of centrifugal pumps. in nondimensional form. 14.9 A pump having the characteristics described in the previous problem is to be built that will deliver water at a rate of 0.08 CQ 0.9 for a desired ﬂow rate of 0.25 1 0. The water will ﬂow through a steel pipe .15 The pump having the characteristics shown in Problem 14.

06 m diameter commercial steel.20 A pump whose operating characteristics are described in Problem 14. or radial ﬂow machine? .9 41. mixed ﬂow.27 A pump is required to deliver 60. Determine (a) the discharge rate and (b) power required.23 A 0. The pump must deliver 2400 gpm against a head of 18 m. 14. Determine the ﬂow rate through the system. Determine the corresponding NPSH.760 m3/s.6 35. Determine the required operating rpm of the pump.2 0 1. m3/m 14.6 0. Determine the discharge rate. The performance curves are shown below.2 39.90 m Steel pipe 14. At a ﬂow rate of 0. Would you expect cavitation to occur? 0.1 m/s. What types of pumps are they? 14. 14. What type of pump is this? 22 m 3m 14.8 m above the surface of the supply reservoir.24 For the pumping system described in Problem 14. Diameter = 0.0.23.4 0.30 A pump operating at 2400 rpm delivers 3.0 1. m 14.25 A centrifugal pump with an impeller diameter of 0.80 m of water.5 m Elevation = 24 m Elevation = 18 m 3m Water 14.200 Chapter 14 Fluid Machinery that is 0.4 Capacity.2 NPSH.28 An axial ﬂow pump has a speciﬁed speciﬁc speed of 6. The discharge line consists of 60 m of 0.8 1. 0.26 Pumps used in an aqueduct operate at 400 rpm and deliver a ﬂow of 220 m3/s against a head of 420 m. Is this pump an axial ﬂow. 14.3 m3/s of water ﬂow against a head of 16 m.06 m diameter steel pipe. % 00 10 20 30 40 50 36.000 gpm against a head of 300 m when operating at 2000 rpm.36 m 0.7 kPa and the inlet velocity is 6. 15 cm diameter 0 20 15 50 0 Capacity.1 31. All valves are fully open globe valves.1 32.25 m pump delivers 20 C water ðPv ¼ 2:34 kPaÞ at 0:065 m3 /s and 2000 rpm: The pump begins to cavitate when the inlet pressure is 82.22 Water at is to be pumped through the system shown.29 A pump operating at 520 rpm has the capability of producing 3.2 m3/s of water against a head of 21 m.2 27.5 23.18 m is to be used to pump water ðr ¼ 1000 kg/m3 Þ with the pump inlet located 3. m 20 C 150 Head. m3 /s 104 ω = 3300rpm 20 cm diameter 10 5 NPSH 0.28 m in diameter and 550 m long.7 The inlet pipe to the pump is 0. What type of pump should be speciﬁed? 14.21 For the same pump and system operation described in Problem determine (a) the discharge rate and (b) power required when the pump operates at 900 rpm. how will the maximum elevation above the surface of the reservoir change if the water temperature is 80 CðPv ¼ 47:35 kPaÞ? 14.5 m in length.3 00 19.9 34. The operating data for this motor-driven pump data are as follows: 100 Developed head.14 is to be used in the system depicted below.6 42. 8. m Efﬁciency. the head loss between the reservoir surface and the pump inlet is 1.

in which the greater motion of a molecule at a higher energy level (temperature) imparts energy to adjacent molecules at lower energy levels. Certainly.1 CONDUCTION Energy transfer by conduction is accomplished in two ways. The ability of solids to conduct heat varies directly with the concentration of free electrons. As heat conduction is primarily a molecular phenomenon. The second mechanism of conduction heat transfer is by ‘‘free’’ electrons. liquid. thus it is not surprising that pure metals are the best heat conductors. as our experience has indicated. Not only must the equipment accomplish its required mission but it must also be economical to purchase and to operate. Our immediate goal is to examine the basic mechanisms of energy transfer and to consider the fundamental equations for evaluating the rate of energy transfer. All heat-transfer processes involve one or more of these modes. The overriding consideration is. we might expect the basic equation used to describe this process to be similar to the expression used in the molecular 201 . and radiation. and the auxiliary equipment required for its utilization are all important considerations for the engineer. The ﬁrst mechanism is that of molecular interaction. This type of transfer is present. Considerations of an engineering nature such as these require both a familiarity with the basic mechanisms of energy transfer and an ability to evaluate quantitatively these rates as well as the important associated quantities. in all systems in which a temperature gradient exists and in which molecules of a solid. the concentration of free electrons varies considerably for alloys and becomes very low for nonmetallic solids. The result of a ﬁrst-law analysis is only a part of the required information necessary for the complete evaluation of a process or situation that involves energy transfer. 15. The freeelectron mechanism is signiﬁcant primarily in pure-metallic solids. There are three modes of energy transfer: conduction. Gross quantities of heat added to or rejected from a system may be evaluated by applying the controlvolume expression for the ﬁrst law of thermodynamics as discussed in Chapter 6. in designing a plant in which heat must be exchanged with the surroundings. convection. The remainder of this chapter will be devoted to an introductory description and discussion of these types of transfer. in many instances. or gas are present. to some degree. the materials of which it is to be constructed.Chapter 15 Fundamentals of Heat Transfer The next nine chapters deal with the transfer of energy. the size of heat-transfer equipment. the rate at which energy transfer takes place.

equation (15-2) is the vector form of the Fourier rate equation. in K/m or F/ft. in which energy transfer in the y direction is on a molecular scale only. The proportionality constant is seen to be the thermal conductivity. Applying equation (6-10) and considering transfer only across the top face of the element considered ZZ ZZZ dQ dWs dWm P @ ¼ eþ er dV (6-10) r(v n) dA þ dt dt dt r @t c:s: c:v: .2 THERMAL CONDUCTIVITY As the mechanism of conduction heat transfer is one of the molecular interaction. A is the area normal to the direction of heat flow. in Watts or Btu/h. Such an equation was ﬁrst stated in 1822 by Fourier in the form qx dT ¼ k (15-1) A dx where qx is the heat-transfer rate in the x direction. The ratio qx /A. like the viscosity. A more general relation for the heat flux is equation (15-2) q ¼ k=T A (15-2) which expresses the heat flux as proportional to the temperature gradient. Mass transfer across the top of this control volume is considered to occur only on the molecular scale. varying signiﬁcantly with pressure only in the case of gases subjected to high pressures. having the dimensions of W/m2 or Btu/h ft2. in m2 or ft2. 15. this expression applies to an isotropic medium only. which plays a role similar to that of the viscosity in momentum transfer. it will be illustrative to examine the motion of gas molecules from a standpoint similar to that in Section 7. k. The negative sign in equation (15-2) indicates that heat flow is in the direction of a negative temperature gradient. The thermal conductivity is a property of a conducting medium and. often referred to as Fourier’s first law of heat conduction. thus. is referred to as the heat flux in the x direction. is primarily a function of temperature.3.1 Molecular motion at the surface of a control volume.1. Wood is a good example of an anisotropic material where the thermal conductivity parallel to the grain may be greater than that normal to the grain by a factor of 2 or more. in W/ðmKÞ or Btu/h ft F. This criterion is met for a gas in laminar ﬂow. dT/dx is the temperature gradient in the x direction. Most materials of engineering interest are isotropic. is assumed independent of direction in equation (15-2). and k is the thermal conductivity. Considering the control volume shown in Figure 15. y y T = T ( y) x x Figure 15. equation (7-4). which is deﬁned by equation (15-1).202 Chapter 15 Fundamentals of Heat Transfer transfer of momentum. The thermal conductivity. we may utilize the ﬁrst-law analysis of Chapter 6 as follows.

becomes 4 k ¼ rcp Zl 3 Utilizing further the results of the kinetic theory of gases. The right-hand term is the summation of the energy ﬂux associated with the molecules crossing the control surface. and that a similar expression may be written for Tjyþ. Some . where l is the mean free path of a molecule. we have qy 4 T ¼ rcp Zl (15-5) A 3 y y0 Comparing equation (15-5) with the y component of equation (15-6) qy T ¼ k A y it is apparent that the thermal conductivity. we may rewrite equation (15-3) in the form Z qy T X ¼ 2 mn cp d (15-4) A y y0 n¼1 where d represents the y component of the distance between collisions. as previously in Chapter 7. and cp ¼ 3k 2N giving. respectively. pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 8kT/pm (k being the Boltzmann 1 l ¼ pﬃﬃﬃ 2pNd 2 where d is the molecular diameter. k. Using this relation and summing over Z molecules. Z is the frequency with which molecules will cross area Dx Dz. and Tjy . even though some oversimplifications were used in its development. is significant in that it shows the thermal conductivity of a gas to be independent of pressure. we may make the following substitutions: Z¼ NC 4 where C is the average random molecular velocity.2 Thermal Conductivity 203 For Z molecules crossing the plane Dx Dz per unit time. C ¼ constant). The significance of this result should not be overlooked. this equation reduces to qy ¼ Z X n¼1 mn c p (Tjy Tjyþ )Dx Dz (15-3) where mn is the mass per molecule. where y ¼ y0 d. and to vary as the 1/2 power of the absolute temperature. Noting now that Tjy ¼ T T/yjy0 d. applying specifically to monatomic gases. that d ¼ (˜¯)l . Tjyþ are the temperatures of the gas slightly below and slightly above the plane considered. ﬁnally k¼ ﬃ 1 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ k3 T/m p3/2 d 2 (15-6) This development. cp is the molecular heat capacity of the gas.15. We note.

A more complete tabulation of thermal conductivity may be found in Appendices H and I. to relate the two conductivities in a crude way. Bird. Physik und Chemie (Poggendorffs). The following two examples illustrate the use of the Fourier rate equation in solving simple heat-conduction problems. E. The thermal conductivity of a liquid is not amenable to any simpliﬁed kinetic-theory development. New York. This realization led Wiedemann and Franz. Figure 15. Lightfoot. thermal conductivity is attributed both to molecular interaction. 1 R. Ann. may be found in Bird. and vice versa.3. Some empirical correlations have met with reasonable success. and in 1872. in 1853. Stewart. Lorenz3 presented the following relation. For a discussion of molecular theories related to the liquid phase and some empirical correlations of thermal conductivities of liquids. the recommended equation is pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ k ¼ 0:0829 (T/M)/s 2 Vk (15-7) where k is in W/m K. known as the Wiedemann. are available. N. as in other phases. and Vk is the Lennard–Jones collision integral. liquid. B. the reader is referred to Reid and Sherwood. 429 (1872). New York. 1960. and solid phases. The thermal properties of most solids of engineering interest have been evaluated. McGraw-Hill Book Company. based upon more sophisticated molecular models. In the solid phase. 7. and L is the Lorenz number. 2 . chap. The numerical values of the quantities in equation (15-8) are of secondary importance at this time. One problem in experimentally determining values of the thermal conductivity in a liquid is making sure the liquid is free of convection currents. Lorenz equation: L¼ k ¼ constant ke T (15-8) where k is the thermal conductivity. and extensive tables and charts of these properties. that those materials that are good conductors of electricity are likewise good heat conductors. identical with V as discussed in Section 7. 147. s is in Angstroms. Both s and Vk may be evaluated from Appendices J and K. The free-electron mechanism of heat conduction is directly analogous to the mechanism of electrical conduction. The signiﬁcant point to note here is the simple relation between electrical and thermal conductivities and. 8. speciﬁcally. as the molecular behavior of the liquid phase is not clearly understood and no universally accurate mathematical model presently exists. 3 L.2 A general observation about liquid thermal conductivities is that they vary only slightly with temperature and are relatively independent of pressure.204 Chapter 15 Fundamentals of Heat Transfer relations for thermal conductivity of gases. Reid and Sherwood. Franz.2 illustrates the thermal conductivity variation with temperature of several important materials in gas. Lorenz. including thermal conductivity. and Lightfoot. chap. and E. Wiley. For a monatomic gas. The solid phase is amenable to quite precise measurements of thermal conductivity.1 The Chapman–Enskog theory used in Chapter 7 to predict gas viscosities at low pressures has a heat-transfer counterpart. The Properties of Gases and Liquids. and to free electrons. Stewart. 1958. W. ke is the electrical conductivity. as there is no problem with convection currents. which are present primarily in pure metals. but these are so specialized that they will not be included in this book. M is the molecular weight. T is the absolute temperature. Transport Phenomena.

2 500 400 Silver Copper 300 Gold Aluminum Aluminum alloy 2024 Tungsten 200 0.2 Helium 0.8 Water 100 Platinum 50 Thermal conductivity (W/mK) Thermal conductivity (W/mK) Thermal Conductivity Iron Stainless steel.15.4 Glycerine 0.3 Thermal conductivity (W/mK) Hydrogen 0.1 Water (steam.2 Pyroceram Engine oil 2 1 100 Fused quartz Freon 12 300 500 1000 Temperature (K) 2000 4000 0 200 300 (a) Solid materials 400 Temperature (K) (b) Liquids 0.2 Thermal conductivity of several materials at various temperatures.6 Ammonia 0. 1 atm) Carbon dioxide Air 0 0 Freon 12 200 400 600 Temperature (K) 800 (c) Gases and vapors Figure 15. AISI 304 20 10 Aluminum oxide 5 0. 1000 500 205 .

The ﬁrst law of thermodynamics applied to this problem will reduce to the form dQ/dt ¼ 0. respectively (see Figure 15.3). . we see that the equation becomes qr ¼ k(2prL) dT dr ro r ri Ti To Figure 15. Find the heat ﬂow rate per meter of pipe length.88 cm and a wall thickness of 0. where qr. Note that for the same amount of heat ﬂow. we obtain 2p(42:90 W/m K)(367 344)K ln(2:66/1:88) ¼ 17 860 W/m(18 600 Btu/hr ft) qr ¼ The inside and outside surface areas per unit length of pipe are giving Ai ¼ p(1:88)(102 )(1) ¼ 0:059 m2 /m(0:194 ft2 /ft) Ao ¼ p(2:662)(102 )(1) ¼ 0:084 m2 /m(0:275 ft2 /ft) qr 17 860 ¼ 302:7 kW/m2 (95 900 Btu/hr ft2 ) ¼ 0:059 Ai qo 17 860 ¼ 212:6 kW/m2 (67 400 Btu/hr ft2 ) ¼ 0:084 Ai One extremely important point to be noted from the results of this example is the requirement of specifying the area upon which a heat-ﬂux value is based. is constant. and the proper form for the Fourier rate equation is qr ¼ kA dT dr Writing A ¼ 2pL.3 Heat conduction in a radial direction with uniform surface temperatures. the ﬂuxes based upon the inside and outside surface areas differ by approximately 42%.391 cm is subjected to inside and outside surface temperature of 367 and 344 K. that is Q ¼ q ¼ constant: As the heat ﬂow will be in the radial direction. the independent variable is r.206 Chapter 15 EXAMPLE 1 Fundamentals of Heat Transfer A steel pipe having an inside diameter of 1. and also the heat ﬂux based on both the inside and outside surface areas. indicating that the rate of heat transfer into the control volume is equal to the rate leaving. which may be separated and solved as follows: Z To Z Ti dr ¼ 2pkL dT ¼ 2pkL dT ri r Ti To ro ¼ 2pkL(Ti To ) qr ln ri 2pkL (Ti To ) qr ¼ ln ro /ri Z ro qr (15-9) Substituting the given numerical values.

using the above relation for the thermal conductivity. If the thermal-conductivity variation may be described as a linear function of temperature according to k ¼ k0 (1 þ bT) calculate the steady-state heat-transfer rate in the radial direction. . Figure 15. and free or natural convection wherein warmer (or cooler) ﬂuid next to the solid boundary causes circulation because of the density difference resulting from the temperature variation throughout a region of the ﬂuid.3 applies. The equation to be solved is now qr ¼ [ko (1 þ bT)](2prL) dT dr which. becomes Z ro Z ro dr ¼ 2pko L (1 þ bT)dT qr ri r ri Z Ti (1 þ bT)dT ¼ 2pko L To T 2pko L bT 2 i Tþ 2 To ln ro /ri 2pko L b qr ¼ 1 þ (Ti þ To ) (Ti To ) ln ro /ri 2 qr ¼ (15-10) Noting that the arithmetic average value of k would be b kavg ¼ ko 1 þ (Ti þ To ) 2 we see that equation (15-10) could also be written as qr ¼ 2pkavg L (Ti To ) ln ro /ri Thus. whether a different geometrical conﬁguration or a different thermal-conductivity expression would make the results of the two types of solutions different.3 EXAMPLE 2 Convection 207 Consider a hollow cylindrical heat-transfer medium having inside and outside radii of ri and ro with the corresponding surface temperatures Ti and To. A distinction must be made between forced convection.15. and compare the result with that using a k value calculated at the arithmetic mean temperature. The student may ﬁnd it instructive to determine what part of the problem statement of this example is responsible for this interesting result. upon separation and integration.3 CONVECTION Heat transfer due to convection involves the energy exchange between a surface and an adjacent ﬂuid. the two methods give identical results. 15. that is. wherein a ﬂuid is made to ﬂow past a solid surface by an external agent such as a fan or pump.

1 are average convective heat-transfer coefficients. The ﬁlm coefﬁcients associated with these two kinds of transfer are quite high. Two types of heat transfer that differ somewhat from free or forced convection but are still treated quantitatively by equation (15-11) are the phenomena of boiling and condensation. A substantial portion of our work in the chapters to follow will involve the determination of this coefficient. in m2 or ft2. water Boiling water Condensing water vapor 0001–10 0005–50 0050–3000 0500–5000 1000–20. and the coefﬁcient h is often referred to as the ﬁlm coefﬁcient. Table 15. a function of system geometry. in K or 8F. We will designate the local coefﬁcient hx. sometimes extremely thin. It is. this is indeed the case. W/(m2 K) Free convection. . As this is always true. those that apply at a point. is related to hx according to the relation Z q¼ hx DT dA ¼ hA DT (15-12) A The values given in Table 15. From our previous experience we should also recall that even when a ﬂuid is ﬂowing in a turbulent manner past a surface. also.208 Chapter 15 Fundamentals of Heat Transfer The rate equation for convective heat transfer was ﬁrst expressed by Newton in 1701. in W/m2 K or Btu/h ft2 F. Equation (15-11) is not a law but a definition of the coefficient h. air Forced convection. A is the area normal to direction of heat flow. that is. the mechanism of heat transfer between a solid surface and a ﬂuid must involve conduction through the ﬂuid layers close to the surface.000 2500–25.000 5000–100. close to the surface where ﬂow is laminar. and is referred to as the Newton rate equation or Newton’s ‘‘law’’ of cooling. according to equation (15-11) dq ¼ hx DT dA Thus the average coefﬁcient. the ﬂuid particles next to the solid boundary are at rest. and h is the convective heat transfer coefficient.000 It will also be necessary to distinguish between local heat transfer coefﬁcients.1 Approximate values of the convective heat-transfer coefﬁcient Mechanism h. This equation is q/A ¼ h DT (15-11) where q is the rate of convective heat transfer. h. As ﬂow properties are so important in the evaluation of the convective heat transfer coefﬁcient.000 0005–50 0025–250 0250–15. and the magnitude of DT. there is still a layer. in W or Btu/h.1 represents some order-of-magnitude values of h for different convective mechanisms. air Forced convection. in general. This ‘‘ﬁlm’’ of ﬂuid often presents the controlling resistance to convective heat transfer. Table 15. we may expect many of the concepts and methods of analysis introduced in the preceding chapters to be of continuing importance in convective heat transfer analysis. fluid and flow properties. and total or average values of h that apply over a given surface area. DT is the temperature difference between surface and fluid. Btu/h ft2 F h.

15. we have L T1 qx dT ¼ k dx A x T2 (15-1) Solving this equation for qx subject to the boundary conditions T ¼ T1 at x ¼ 0 and T ¼ T2 at x ¼ L. Certain modiﬁcations will be made in equation (15-13) to account for the net energy transfer between two surfaces. Consider the case depicted in Figure 15. and s is the Stefan–Boltzmann constant. We wish to express the steady-state heat-transfer rate per unit area between a hot gas at temperature Th on one side of this wall and a cool gas at . in m2 or ft2. 15. in actual situations.5 COMBINED MECHANISMS OF HEAT TRANSFER The three modes of heat transfer have been considered separately in Section 15. It will be instructive to look at some situations in which heat transfer is accomplished by a combination of these mechanisms. The rate of energy emission from a perfect radiator or blackbody is given by q ¼ sT 4 A (15-13) where q is the rate of radiant energy emission. which is equal to 5:676 108 W/m2 K4 or 0:1714 108 Btu/h ft2 R4. Equation (15-13) is most often referred to as the Stefan–Boltzmann law of thermal radiation. who derived this relation theoretically in 1884.4. proposed equation (15-13) in 1879. and Boltzmann. in W or Btu/h.4. indeed energy transfer by radiation is maximum when the two surfaces that are exchanging energy are separated by a perfect vacuum. in K or 8R. the degree of deviation of the emitting and receiving surfaces from blackbody behavior.5 Combined Mechanisms of Heat Transfer 209 15. It is rare. The proportionally constant relating radiant-energy flux to the fourth power of the absolute temperature is named after Stefan who. Consider the composite plane wall constructed of three materials in layers with dimensions as shown in Figure 15. we obtain Figure 15. that of steady-state conduction through a plane wall with its surfaces held at constant temperatures T1 and T2. T is the absolute temperature.4 Steady-state conduction through a plane wall. from experimental observations. for only one mechanism to be involved in the transfer of energy. A is the area of the emitting surface. Writing the Fourier rate equation for the x direction.4 RADIATION Radiant heat transfer between surfaces differs from conduction and convection in that no medium is required for its propagation. These considerations are discussed at length in Chapter 23. qx A Z L Z dx ¼ k 0 T2 Z dT ¼ k T1 T1 dT T2 or qx ¼ kA (T1 T2 ) L (15-14) Equation (15-14) bears an obvious resemblance to the Newton rate equation qx ¼ hA DT (15-11) We may utilize this similarity in form in a problem in which both types of energy transfer are involved. and geometrical factors associated with radiant exchange between a surface and its surroundings.5.

210 Chapter 15 Fundamentals of Heat Transfer Tc on the other side. Each temperature difference is expressed in terms of qx as follows: Th T1 ¼ qx (1/hh A) T1 T2 ¼ qx (L1 /k1 A) T2 T3 ¼ qx (L2 /k2 A) T3 T4 ¼ qx (L3 /k3 A) T4 Tc ¼ qx (1/hc A) Adding these equations. The following relations for qx arise from the application of equations (15-11) and (15-14): qx ¼ hh A(Th T1 ) ¼ ¼ k1 Th k1 A k2 A (T1 T2 ) ¼ (T2 T3 ) L1 L2 k3 A (T3 T4 ) ¼ hc A(T4 Tc ) L3 k2 k3 T2 T1 T4 T3 L1 L2 L3 Tc Figure 15. solving for qx. we have qx ¼ Th Tc 1/hh A þ L1 /k1 A þ L2 /k2 A þ L3 /k3 A þ 1/hc A (15-15) Note that the heat-transfer rate is expressed in terms of the overall temperature difference. If a series electrical circuit R1 R2 R3 R4 R5 V1 V2 is considered. we obtain 1 L1 L2 L3 1 þ þ þ þ Th Tc ¼ qx h h A k 1 A k 2 A k 3 A hc A and ﬁnally.5 Steady-state heat transfer through a composite wall. Temperature designations and dimensions are as shown in the ﬁgure. we may write I¼ DV DV ¼P Ri R1 þ R2 þ R3 þ R4 þ R5 .

The surrounding air is at 294 K.8 cm thickness of 85% magnesia insulation on its outer surface. respectively. they can be utilized in the form indicated by equation (15-16). DV ! DT I ! qx Ri ! 1/hA. Equation (15-15) thus becomes a heat-transfer analog to Ohm’s law. relating heat ﬂow to the overall temperature difference divided by the total thermal resistance between the points of known temperature.5 Combined Mechanisms of Heat Transfer 211 The analogous quantities in the expressions for heat ﬂow and electrical current are apparent. In the case of the bare pipe there are three thermal resistances to evaluate: R1 ¼ Rconvection inside ¼ 1/hi Ai R2 ¼ Rconvection outside ¼ 1/ho Ao R3 ¼ Rconduction ¼ ln(ro =ri )/2pkL For conditions of this problem. EXAMPLE 3 Saturated steam at 0.15.09 cm and an outside diameter of 2.67 cm. With specific reference to equation (15-9). The thermal-resistance terms will change in form for cylindrical or spherical systems. these resistances have the values R1 ¼ 1/[(5680 W/m2 K)(p)(0:0209 m)(1 m)] h R ¼ 0:00268 K/W 0:00141 Btu R2 ¼ 1/[(22:7 W/m2 K)(p)(0:0267 m)(1 m)] h R ¼ 0:525 K/W 0:277 Btu . but once evaluated. Equation (15-15) may now be written simply as DT (15-16) q¼P Rthermal This relation applies to steady-state heat transfer in systems of other geometries as well. it may be noted that the thermal resistance of a cylindrical conductor is ln(ro /ri ) 2pkL Another common way of expressing the heat-transfer rate for a situation involving a composite material or combination of mechanisms is with the overall heat-transfer coefﬁcient deﬁned as qx (15-17) U A DT where U is the overall heat-transfer coefficient having the same units as h. in W/m2 K or Btu/h ft2 F.276 MPa ﬂows inside a steel pipe having an inside diameter of 2. L/kA and each term in the denominator of equation (15-15) may be thought of as a thermal resistance due to convection or conduction. Find the heat loss per meter of bare pipe and for a pipe having a 3. The convective coefﬁcients on the inner and outer pipe surfaces may be taken as 5680 and 22:7 W/m2 K.

which would be. plus additional resistances to account for the insulation.212 Chapter 15 Fundamentals of Heat Transfer and ln(2:67/2:09) 2p(42:9 W/m K)(1 m) h R ¼ 0:00091 K/W 0:00048 Btu R3 ¼ The inside temperature is that of 0. Example 3 could also have been worked by using an overall heat-transfer coefﬁcient. for instance. For the insulation ln(10:27/2:67) 2p(0:0675 W/m K)(1 m) h R ¼ 3:176 K/W 1:675 Btu R4 ¼ and for the outside surface of the insulation R5 ¼ 1/[(22:7 W/m2 K)(p)(0:1027 m)(1 m)] h R ¼ 0:1365 K/W 0:0720 Btu Thus. an increased rate of heat transfer were desired. in general P U¼ qx DT/ R 1 ¼ ¼ P A DT A DT A R or. The heat transfer rate per meter of pipe may now be calculated as DT 404 294 K q¼P ¼ R 0:528 K/W Btu ¼ 208 W 710 h In the case of an insulated pipe. the heat loss for the insulated pipe becomes DT DT 404 294 K q¼P ¼ ¼ R R1 þ R2 þ R4 þ R5 3:316 K/W Btu ¼ 33:2 W 113 h A reduction of approximately 85%! It is apparent from this example that certain parts of the heat-transfer path offer a negligible resistance. 404 K or 2678F.276 MPa saturated steam. for the speciﬁc case considered U¼ 1 Af1/Ai hi þ [ln(ro /ri )]/2pkL þ 1/Ao ho g (15-18) . If. the obvious approach would be to alter the outside convective resistance. in the case of the bare pipe. which is almost 200 times the magnitude of the next-highest thermalresistance value. the total thermal resistance would include R1 and R3 evaluated above.

A q¼k DT L and for a cylinder 2pL q¼k DT ln(ro /ri ) Each of the bracketed terms is the shape factor for the applicable geometry. may have a different numerical value. when specifying an overall coefﬁcient. we have 1 Uo ¼ Ao /Ai hi þ [Ao ln(ro /ri )]/2pkL þ 1/ho Thus. convection. symbolized as S. Ao. The transport property. has been discussed and some consideration given to energy transfer in a monatomic gas at low pressure. Considering the steady-state relations developed for plane and cylindrical shapes kA DT (15-14) q¼ L and q¼ 2pkL DT ln(ro /ri ) (15-9) if that part of each expression having to do with the geometry is separated from the remaining terms. we have. to relate it to a speciﬁc area. depending on which area it is based upon. and radiation—have been introduced.6 CLOSURE In this chapter. The rate equations for heat transfer are as follows: Conduction: the Fourier rate equation q ¼ k =T A Convection: the Newton rate equation q ¼ h DT A . then the shape factor may be calculated and q determined for various materials displaying a range of values of k. If. U.15. A general relation utilizing this form is q ¼ kS DT (15-19) Equation (15-19) offers some advantages when a given geometry is required because of space and conﬁguration limitations. the basic modes of energy transfer—conduction. If this is the case. U is based upon the outside surface area of the pipe. One other means of evaluating heat-transfer rates is by means of the shape factor. along with the simple relations expressing the rates of energy transfer associated therewith. thermal conductivity. it is necessary. 15. for instance. for a plane wall.6 Closure 213 Equation (15-18) indicates that the overall heat-transfer coefficient.

214

Chapter 15

Fundamentals of Heat Transfer

**Radiation: the Stefan–Boltzmann law for energy emitted from a black surface
**

q

¼ sT 4

A

Combined modes of heat transfer were considered, speciﬁcally with respect to the

means of calculating heat-transfer rates when several transfer modes were involved. The

three ways of calculating steady-state heat-transfer rates are represented by the equations

DT

q¼P

RT

where

P

(15-16)

**RT is the total thermal resistance along the transfer path;
**

q ¼ UA DT

(15-17)

**where U is the overall heat transfer coefficient; and
**

q ¼ kS DT

(15-19)

**where S is the shape factor.
**

The equations presented will be used throughout the remaining chapters dealing with

energy transfer. A primary object of the chapters to follow will be the evaluation of the heattransfer rates for special geometries or conditions of ﬂow, or both.

Note: Effects of thermal radiation are included, along with convection, in values of

surface coefﬁcients speciﬁed in the following problems.

PROBLEMS

15.1 An asbestos pad is square in cross section, measuring

5 cm on a side at its small end increasing linearly to 10 cm on a

side at the large end. The pad is 15 cm high. If the small end is

held at 600 K and the large end at 300 K, what heat-ﬂow rate

will be obtained if the four sides are insulated? Assume onedimensional heat conduction. The thermal conductivity of

asbestos may be taken as 0:173 W/mK:

15.2 SolveProblem forthecaseofthelargercrosssectionexposed

to the higher temperature and the smaller end held at 300 K.

15.3 Solve Problem 15.1 if, in addition to a varying crosssectional area, the thermal conductivity varies according to k ¼

k0 (1 þ bT); where k0 ¼ 0:138, b ¼ 1:95 104 , T ¼ temperature in Kelvin, and k is in W/m K. Compare this result to that

using a k value evaluated at the arithmetic mean temperature.

15.4 Solve Problem 15.1 if the asbestos pad has a 1.905-cm

steel bolt running through its center.

15.5 A sheet of insulating material, with thermal conductivity

of 0:22 W/m K is 2 cm thick and has a surface area of 2.97 m2.

If 4 kW of heat are conducted through this sheet and the outer

(cooler) surface temperature is measured at 55 C(328 K), what

will be the temperature on the inner (hot) surface?

15.6 For the sheet of insulation speciﬁed in Problem 15.5, with

a heat rate of 4 kW, evaluate the temperature at both surfaces if

**the cool side is exposed to air at 308C with a surface coefﬁcient of
**

28:4 W/m2 K:

15.7 Plate glass, k ¼ 1:35 W/m K; initially at 850 K, is cooled

by blowing air past both surfaces with an effective surface coefﬁcient of 5 W/m2 K: It is necessary, in order that the glass does not

crack, to limit the maximum temperature gradient in the glass to

15 K/mm during the cooling process. At the start of the cooling process,whatisthelowesttemperatureofthecoolingairthatcanbeused?

15.8 Solve Problem 15.7 if all speciﬁed conditions remain the

same but radiant energy exchange from glass to the surroundings

at the air temperature is also considered.

15.9 The heat loss from a boiler is to be held at a maximum of

900 Btu/h ft2 of wall area. What thickness of asbestos (k ¼

0:10 Btu/h ft F) is required if the inner and outer surfaces of the

insulation are to be 1600 and 5008F, respectively?

15.10 If, in the previous problem, a 3-in.-thick layer of kaolin

brick (k ¼ 0:07 Btu/h ft F) is added to the outside of the asbestos, what heat ﬂux will result if the outside surface of the kaolin is

2508F? What will be the temperature at the interface between the

asbestos and kaolin for this condition?

15.11 A composite wall is to be constructed of 1/4-in. stainless

steel (k ¼ 10 Btu/h ft F), 3 in. of corkboard (k ¼ 0:025 Btu/

h ft F) and 1/2 in. of plastic (k ¼ 1:5 Btu/h ft F):

Problems

a. Draw the thermal circuit for the steady-state conduction

through this wall.

b. Evaluate the individual thermal resistance of each material

layer.

c. Determine the heat ﬂux if the steel surface is maintained at

2508F and the plastic surface held at 808F.

d. What are the temperatures on each surface of the corkboard

under these conditions?

15.12 If, in the previous problem, the convective heat-transfer

coefﬁcients at the inner (steel) and outer surfaces are 40 and

5 Btu/h ft F, respectively, determine

a. the heat ﬂux if the gases are at 250 and 708F, adjacent to the

inner and outer surfaces;

b. the maximum temperature reached within the plastic;

c. which of the individual resistances is controlling.

15.13 A 1-in.-thick steel plate measuring 10 in. in diameter is

heated from below by a hot plate, its upper surface exposed to air

at 808F. The heat-transfer coefﬁcient on the upper surface is

5 Btu/h ft F and k for steel is 25 Btu/h ft F:

**wide with a depth of 0.5 m. Determine the thickness of styrofoam
**

insulation (k ¼ 0:30 W/m K) needed to limit the heat loss to

400 W if the inner and outer surface temperatures are 10 and

338C, respectively.

15.20 Evaluate the required thickness of styrofoam for the

freezer compartment in the previous problem when the inside

wall is exposed to air at 108C through a surface coefﬁcient of

16 W/m2 K and the outer wall is exposed to 338C air with a

surface coefﬁcient of 32 W/m2 K: Determine the surface temperatures for this situation.

15.21 The cross section of a storm window is shown in the

sketch. How much heat will be lost through a window measuring

1.83 m by 3.66 m on a cold day when the inside and outside air

temperatures are, respectively, 295 and 250 K? Convective

coefﬁcients on the inside and outside surfaces of the window

are 20 and 15 W/m2 K, respectively. What temperature drop

will exist across each of the glass panes? What will be the

average temperature of the air between the glass panes?

**a. How much heat must be supplied to the lower surface of the
**

steel if its upper surface remains at 1608F? (Include radiation.)

b. What are the relative amounts of energy dissipated from the

upper surface of the steel by convection and radiation?

Air

space

0.8 cm

wide

**15.14 If, in Problem 15.13, the plate is made of asbestos, k ¼
**

0:10 Btu/h ft F; what will be the temperature of the top of the

asbestos if the hot plate is rated at 800 W?

Window glass 0.32 cm thick

**15.15 A 0.20-m-thick brick wall (k ¼ 1:3 W/m K) separates
**

the combustion zone of a furnace from its surroundings at 258C.

For an outside wall surface temperature of 1008C, with a

convective heat transfer coefﬁcient of 18 W/m2 K; what will

be the inside wall surface temperature at steady-state conditions?

15.16 Solve for the inside surface temperature of the brick

wall described in Problem 15.15, but with the additional

consideration of radiation from the outside surface to surroundings at 258C.

15.17 The solar radiation incident on a steel plate 2 ft square is

400 Btu/h. The plate is 1.4 in. thick and lying horizontally on an

insulating surface, its upper surface being exposed to air at 908F.

If the convective heat-transfer coefﬁcient between the top surface and the surrounding air is 4 Btu/h ft F; what will be the

steady-state temperature of the plate?

15.18 If in Problem 15.17, the lower surface of the plate is

exposed to air with a convective heat transfer coefﬁcient of

3 Btu/h ft F, what steady-state temperature will be reached

a. if radiant emission from the plate is neglected;

b. if radiant emission from the top surface of the plate is

accounted for?

15.19 The freezer compartment in a conventional refrigerator

can be modeled as a rectangular cavity 0.3 m high and 0.25 m

215

**15.22 Compare the heat loss through the storm window
**

described in Problem 15.21 with the same conditions existing

except that the window is a single pane of glass 0.32 cm thick.

15.23 The outside walls of a house are constructed of a 4-in. layer

of brick, 1/2 in. of celotex, an air space 3 5=8 in. thick, and 1/4 in. of

wood panelling. If the outside surface of the brick is at 308F and the

inner surface of the panelling at 758F, what is the heat ﬂux if

a. the air space is assumed to transfer heat by conduction

only?

b. the equivalent conductance of the air space is 1:8 Btu/

h ft2 F?

c. the air space is ﬁlled with glass wool?

kbrick ¼ 0:38 Btu/h ft F

kcelotex ¼ 0:028 Btu/h ft F

kair ¼ 0:015 Btu/h ft F

kwood ¼ 0:12 Btu/h ft F

kwool ¼ 0:025 Btu/h ft F:

15.24 Solve Problem 15.23 if instead of the surface temperatures being known, the air temperatures outside and inside are 30

and 758F, and the convective heat-transfer coefﬁcients are 7 and

2 Btu/h ft2 F; respectively.

216

Chapter 15

Fundamentals of Heat Transfer

**15.25 Determine the heat-transfer rate per square meter of wall
**

area for the case of a furnace with inside air at 1340 K. The

furnace wall is composed of a 0.106-m layer of ﬁreclay brick and

a 0.635-cm thickness of mild steel on its outside surface. Heat

transfer coefﬁcients on inside and outside wall surfaces are 5110

and 45 W/m2 K, respectively; outside air is at 295 K. What will

be the temperatures at each surface and at the brick-steel interface?

15.26 Given the furnace wall and other conditions as speciﬁed

in Problem 15.25, what thickness of celotex (k ¼ 0:065 W/

m K) must be added to the furnace wall in order that the

outside surface temperature of the insulation not exceed 340 K?

15.27 A 4-in.-OD pipe is to be used to transport liquid metals

and will have an outside surface temperature of 14008F under

operating conditions. Insulation is 6 in. thick and has a thermal

conductivity expressed as

k ¼ 0:08(1 0:003 T)

where k is in Btu/h ft 8F and T is in 8F, is applied to the outside

surface of the pipe.

a. What thickness of insulation would be required for the

outside insulation temperature to be no higher than 3008F?

b. What heat-ﬂow rate will occur under these conditions?

15.28 Water at 408F is to ﬂow through a 11=2-in. schedule 40 steel

pipe. The outside surface of the pipe is to be insulated with a 1-in.thick layer of 85% magnesia and a 1-in.-thick layer of packed glass

wool, k ¼ 0:022 Btu/h ft F: The surrounding air is at 1008F.

a. Which material should be placed next to the pipe surface to

produce the maximum insulating effect?

**b. What will be the heat ﬂux on the basis of the outside pipe
**

surface area? The convective heat-transfer coefﬁcients for

the inner and outer surfaces are 100 and 5 Btu/h ft F:

respectively.

15.29 A 1-in.-nominal-diameter steel pipe with its outside

surface at 4008F is located in air at 908F with the convective

heat-transfer coefﬁcient between the surface of the pipe and the

air equal to 1:5 Btu/h ft F: It is proposed to add insulation having

a thermal conductivity of 0:06 Btu/h ft F to the pipe to reduce

the heat loss to one half that for the bare pipe. What thickness of

insulation is necessary if the surface temperature of the steel pipe

and ho remain constant?

15.30 If, for the conditions of Problem 15.29, ho in Btu/h ft F

varies according to ho ¼ 0:575/D1o/4 , where Do is the outside

diameter of the insulation in feet, determine the thickness of

insulation that will reduce the heat ﬂux to one half that of the

value for the bare pipe.

15.31 Liquid nitrogen at 77 K is stored in a cylindrical container having an inside diameter of 25 cm. The cylinder is made

of stainless steel and has a wall thickness of 1.2 cm. Insulation

is to be added to the outside surface of the cylinder to reduce

the nitrogen boil-off rate to 25% of its value without insulation.

The insulation to be used has a thermal conductivity of

0:13 W/m K: Energy loss through the top and bottom ends

of the cylinder may be presumed negligible.

Neglecting radiation effects, determine the thickness of

insulation when the inner surface of the cylinder is at 77 K, the

convective heat-transfer coefﬁcient at the insulation surface

has a value of 12 W/m2 K, and the surrounding air is at

258C.

Chapter

16

Differential Equations

of Heat Transfer

P

**aralleling the treatment of momentum transfer undertaken in Chapter 9, we shall now
**

generate the fundamental equations for a differential control volume from a ﬁrst-law-ofthermodynamics approach. The control-volume expression for the ﬁrst law will provide

our basic analytical tool. Additionally, certain differential equations already developed

in previous sections will be applicable.

**16.1 THE GENERAL DIFFERENTIAL EQUATION FOR ENERGY TRANSFER
**

Consider the control volume having dimensions Dx, Dy, and Dz as depicted in Figure 16.1.

Refer to the control-volume expression for the ﬁrst law of thermodynamics

dQ dWs dWm

¼

dt

dt

dt

ZZ

ZZZ

P

@

:

eþ

er dV

r(v n)dA þ

r

@t

c:s:

c:v:

**The individual terms are evaluated and their meanings are discussed below.
**

The net rate of heat added to the control volume

will include all conduction effects, the net release

of thermal energy within the control volume due to

volumetric effects such as a chemical reaction or

induction heating, and the dissipation of electrical

or nuclear energy. The generation effects will be

included in the single term, q˙, which is the volumetric rate of thermal energy generation having units

W/m3 or Btu/h ft3. Thus, the ﬁrst term may be

expressed as

(6-10)

y

y

z

x

x

z

**Figure 16.1 A differential control
**

volume.

dQ

@T

@T

@T

@T

¼ k

k Dy Dz þ k

k Dx Dz

dt

@x xþDx

@x x

@y yþDy

@y y

@T

@T

k Dx Dy þ q_ Dx Dy Dz

þ k

@z zþDz

@z z

ð16-1Þ

The shaft work rate or power term will be taken as zero for our present purposes. This term is speciﬁcally related to work done by some effect within the

217

218

Chapter 16

Differential Equations of Heat Transfer

**control volume that, for the differential case, is not present. The power term is thus
**

evaluated as

dWs

¼0

dt

(16-2)

**The viscous work rate, occurring at the control surface, is formally evaluated by
**

integrating the dot product of the viscous stress and the velocity over the control surface.

As this operation is tedious, we shall express the viscous work rate as L Dx Dy Dz, where

L is the viscous work rate per unit volume. The third term in equation (6-10) is thus

written as

dWm

¼ L Dx Dy Dz

(16-3)

dt

The surface integral includes all energy transfer across the control surface due to ﬂuid

ﬂow. All terms associated with the surface integral have been deﬁned previously. The

surface integral is

ZZ

P

eþ

r(v: n)dA

r

c:s:

2

2

v

P

v

P

þ gy þ u þ

þ

gy

þ

u

þ

¼ rv x

rv

Dy Dz

x

r xþDx

r x

2

2

"

2

#

v2

P

v

P

þ gy þ u þ

þ gy þ u þ

rv y

Dx Dz

þ rv y

r yþDy

r y

2

2

2

2

v

P

v

P

þ gy þ u þ

þ gy þ u þ

rv z

Dx Dy

ð16-4Þ

þ rv z

r zþDz

r z

2

2

The energy accumulation term, relating the variation in total energy within the control

volume as a function of time, is

ZZZ

@

@ v2

þ gy þ u r Dx Dy Dz

(16-5)

er dV ¼

@t

@t 2

c:v:

Equations (16-1) through (16-5) may now be combined as indicated by the general ﬁrstlaw expression, equation (6-10). Performing this combination and dividing through by the

volume of the element, we have

k(@T/@x)jxþDx k(@T/@x)jx k(@T/@y)jyþDy k(@T/@y)jy

þ

Dx

Dy

k(@T/@z)jzþDz k(@T/@z)jz

þ q_ þ L

þ

Dz

frv x (v 2 /2) þ gy þ u þ (P/r) jxþDx rv x (v 2 /2) þ gy þ u þ (P/r) jx g

¼

Dx

frv y (v 2 /2) þ gy þ u þ (P/r) jyþDy rv y (v 2 /2) þ gy þ u þ (P/r) jy g

þ

Dy

frv z (v 2 /2) þ gy þ u þ (P/r) jzþDz rv z (v 2 /2) þ gy þ u þ (P/r) jz g

þ

Dz

2

@

v

þ r

þ gy þ u

@t

2

16.1

The General Differential Equation for Energy Transfer

Evaluated in the limit as Dx, Dy, and Dz approach zero, this equation becomes

@

@T

@

@T

@

@T

k

k

k

þ

þ

þ q_ þ L

@x @x

@y @y

@z @z

2

2

@

v

P

@

v

P

rv x

rv y

þ gy þ u þ

þ gy þ u þ

þ

¼

@x

r

@y

r

2

2

2

2

@

v

P

@

v

rv z

r

þ

þ gy þ u þ

þ gy þ u

þ

@z

r

@t

2

2

219

ð16-6Þ

**Equation (16-6) is completely general in application. Introducing the substantial
**

derivative, we may write equation (16-6) as

@

@T

@

@T

@

@T

k

k

k

þ

þ

þ q_ þ L

@x

@x

@y @y

@z @z

2

v

@r

r Dv 2

Du

D(gy)

:

:

þr

þ u þ gy = rv þ

þr

þ

¼ = (Pv) þ

@t

2 Dt

Dt

Dt

2

Utilizing the continuity equation, equation (9-2), we reduce this to

@

@T

@

@T

@

@T

k

k

k

þ

þ

þ q_ þ L

@x @x

@y @y

@z @z

¼ =: Pv þ

r Dv 2

Du

D(gy)

þr

þr

2 Dt

Dt

Dt

ð16-7Þ

**With the aid of equation (9-19), which is valid for incompressible flow of a fluid with
**

constant m, the second term on the right-hand side of equation (16-7) becomes

r Dv 2

¼ v: =P þ v: rg þ v: m =2 v

2 Dt

(16-8)

**Also, for incompressible flow, the first term on the right-hand side of equation (16-7)
**

becomes

=: Pv ¼ v: =P

(16-9)

**Substituting equations (16-8) and (16-9) into equation (16-7), and writing the conduction
**

terms as =: k=T, we have

=: k =T þ q_ þ L ¼ r

Du

D(gy)

þr

þ v: rg þ v: m =2 v

Dt

Dt

(16-10)

**It will be left as an exercise for the reader to verify that equation (16-10) reduces further
**

to the form

=: k=T þ q_ þ L ¼ rcv

DT

þ v: m = 2 v

Dt

(16-11)

**The function L may be expressed in terms of the viscous portion of the normal- and
**

shear-stress terms in equations (7-13) and (7-14). For the case of incompressible ﬂow, it is

written as

L ¼ v: m=2 v þ F

(16-12)

220

Chapter 16

Differential Equations of Heat Transfer

**where the ‘‘dissipation function,’’ F, is given by
**

" #

@v y 2

@v x 2

@v z 2

F ¼ 2m

þ

þ

@x

@y

@z

"

2

#

@v y @v z 2

@v x @v y

@v z @v x 2

þm

þ

þ

þ

þ

þ

@y

@x

@z

@y

@x

@z

Substituting for L in equation (16-11), we see that the energy equation becomes

DT

(16-13)

=: k=T þ q_ þ F ¼ rcv

Dt

From equation (16-12), F is seen to be a function of fluid viscosity and shear-strain rates, and

is positive-definite. The effect of viscous dissipation is always to increase internal energy at

the expense of potential energy or stagnation pressure. The dissipation function is negligible in

all cases that we will consider; its effect becomes significant in supersonic boundary layers.

**16.2 SPECIAL FORMS OF THE DIFFERENTIAL ENERGY EQUATION
**

The applicable forms of the energy equation for some commonly encountered situations

follow. In every case the dissipation term is considered negligibly small.

I.

**For an incompressible ﬂuid without energy sources and with constant k
**

rcv

II.

DT

¼ k =2 T

Dt

(16-14)

**For isobaric ﬂow without energy sources and with constant k, the energy equation is
**

rcv

DT

¼ k =2 T

Dt

(16-15)

**Note that equations (16-14) and (16-15) are identical yet apply to completely
**

different physical situations. The student may wish to satisfy himself at this point

as to the reasons behind the unexpected result.

III. In a situation where there is no ﬂuid motion, all heat transfer is by conduction. If

this situation exists, as it most certainly does in solids where cv ’ cp , the energy

equation becomes

@T

¼ =: k =T þ q_

rcp

(16-16)

@t

Equation (16-16) applies in general to heat conduction. No assumption has been made

concerning constant k. If the thermal conductivity is constant, the energy equation is

@T

q_

¼ a =2 T þ

(16-17)

@t

rcp

where the ratio k/rcp has been symbolized by a and is designated the thermal diffusivity. It is

easily seen that a has the units, L2/t; in the SI system a is expressed in m2/s, and as ft2/h in the

English system.

If the conducting medium contains no heat sources, equation (16-17) reduces to the

Fourier ﬁeld equation

@T

¼ a =2 T

(16-18)

@t

which is occasionally referred to as Fourier’s second law of heat conduction.

16.3

Commonly Encountered Boundary Conditions

221

**For a system in which heat sources are present but there is no time variation, equation
**

(16-17) reduces to the Poisson equation

=2 T þ

q_

¼0

k

(16-19)

**The ﬁnal form of the heat-conduction equation to be presented applies to a steady-state
**

situation without heat sources. For this case, the temperature distribution must satisfy the

Laplace equation

=2 T ¼ 0

(16-20)

**Each of equations (16-17) through (16-20) has been written in general form, thus each
**

applies to any orthogonal coordinate system. Writing the Laplacian operator, =2, in the

appropriate form will accomplish the transformation to the desired coordinate system. The

Fourier ﬁeld equation written in rectangular coordinates is

2

@T

@ T @2T @2T

þ

þ

¼a

@t

@x2 @y2 @z2

(16-21)

2

@T

@ T 1 @T 1 @ 2 T @ 2 T

þ

þ

¼a

þ

@t

@r 2 r @r r 2 @u2 @z2

(16-22)

in cylindrical coordinates

and in spherical coordinates

@T

1 @

@T

1

@

@T

1

@2T

¼a 2

r2

þ 2

sin u

þ 2

@t

r @r

@r

r sin u @u

@u

r sin2 u @f2

(16-23)

**The reader is referred to Appendix B for an illustration of the variables in cylindrical and
**

spherical coordinate systems.

**16.3 COMMONLY ENCOUNTERED BOUNDARY CONDITIONS
**

In solving one of the differential equations developed thus far, the existing physical situation

will dictate the appropriate initial or boundary conditions, or both, which the ﬁnal solutions

must satisfy.

Initial conditions refer speciﬁcally to the values of Tand v at the start of the time interval

of interest. Initial conditions may be as simply speciﬁed as stating that Tjt¼0 ¼ T0

(a constant), or more complex if the temperature distribution at the start of time measurement is some function of the space variables.

Boundary conditions refer to the values of T and v existing at speciﬁc positions on the

boundaries of a system, that is, for given values of the signiﬁcant space variables. Frequently

encountered boundary conditions for temperature are the case of isothermal boundaries,

along which the temperature is constant, and insulated boundaries, across which no heat

conduction occurs where, according to the Fourier rate equation, the temperature derivative

normal to the boundary is zero. More complicated temperature functions often exist at

system boundaries, and the surface temperature may also vary with time. Combinations of

heat-transfer mechanisms may dictate boundary conditions as well. One situation often

existing at a solid boundary is the equality between heat transfer to the surface by conduction

222

Chapter 16

**Differential Equations of Heat Transfer
**

Th

qx = hh ( Th – T1)

T1

qx = – k

dT

dx

T2

qx = hc ( T2 – Tc)

L

Tc

x

**Figure 16.2 Conduction and
**

convention at a system boundary.

**and that leaving the surface by convection. This condition is illustrated in Figure 16.2. At the
**

left-hand surface, the boundary condition is

@T

(16-24)

hk (Tk Tjx¼0 ) ¼ k

@x

x¼0

**and at the right-hand surface,
**

hc (Tjx¼L Tc ) ¼ k

@T

@x x¼L

(16-25)

**It is impossible at this time to foresee all the initial and boundary conditions that will be
**

needed. The student should be aware, however, that these conditions are dictated by the

physical situation. The differential equations of energy transfer are not numerous, and a

speciﬁc form applying to a given situation may be found easily. It remains for the user of

these equations to choose the appropriate initial and boundary conditions to make the

solution meaningful.

16.4 CLOSURE

The general differential equations of energy transfer have been developed in this chapter,

and some forms applying to more speciﬁc situations were presented. Some remarks

concerning initial and boundary conditions have been made as well.

In the chapters to follow, analyses of energy transfer will start with the applicable

differential equation. Numerous solutions will be presented and still more assigned as

student exercises. The tools for heat-transfer analysis have now been developed and

examined. Our remaining task is to develop a familiarity with and facility in their use.

PROBLEMS

The Fourier ﬁeld equation in cylindrical coordinates is

2

@T

@ T 1 @T 1 @ 2 T @ 2 T

¼a

þ

:

þ

@t

@r 2 r @r r 2 @u2 @z2

a. What form does this equation reduce to for the case of

steady-state, radial heat transfer?

b. Given the boundary conditions

16.1

T ¼ Ti

T ¼ To

at r ¼ ri

at r ¼ ro

**c. Generate an expression for the heat ﬂow rate, qr , using the
**

result from part (b).

16.2 Perform the same operations as in parts (a), (b), and (c) of

Problem 16.1 with respect to a spherical system.

16.3 Starting with the Fourier ﬁeld equation in cylindrical

coordinates,

a. Reduce this equation to the applicable form for steady-state

heat transfer in the u direction.

Problems

b. For the conditions depicted in the ﬁgure, that is, T ¼ To at

u ¼ 0, T ¼ Tz at u ¼ p, the radial surfaces insulated, solve

for the temperature proﬁle.

c. Generate an expression for the heat ﬂow rate, qu, using the

result of part (b).

d. What is the shape factor for this conﬁguration?

223

**b. the thermal conductivity varies with temperature according
**

to k ¼ ko aT, where a is a constant.

T1, A1

L

To

q

r1

T2, A2

Tp

16.4

ro

**Show that equation (16-10) reduces to the form
**

=: k =T þ q_ þ L ¼ rcv

DT

þ v: m=2 v

Dt

16.5 Starting with equation (16-7), show that, for a ﬂuid with

constant thermal conductivity and no energy sources, equations

(16-14) and (16-15) are obtained for incompressible and isobaric

conditions, respectively. (Neglect viscous dissipation.)

16.6 Solve equation (16-19) for the temperature distribution in

a plane wall if the internal heat generation per unit volume varies

according to q_ ¼ q_ 0 ebx/L . The boundary conditions that apply

are T ¼ T0 at x ¼ 0 and T ¼ TL at x ¼ L.

16.7 Solve Problem 16.6 for the same conditions, except that

the boundary condition at x ¼ L is dT/dx ¼ 0.

x

**16.13 Heat is generated in a radioactive plane wall according
**

to the relationship

X

q_ ¼ q_ max 1

L

where q_ is the volumetric heat generation rate, kW/m3 , L is the half

thickness of the plate, and x is measured from the plate center line.

X

L

**16.8 Solve Problem 16.6 for the same conditions, except that
**

at x ¼ L, dT/dx ¼ j (a constant).

16.9 Use the relation T ds ¼ dh dP/r to show that the effect

of the dissipation function, F, is to increase the entropy, S. Is the

effect of heat transfer the same as the dissipation function?

16.10 In a boundary layer where the velocity proﬁle is given by

vx

3 y 1 y 3

¼

v1 2 d 2 d

where d is the velocity boundary layer thickness, plot the

dimensionless dissipation function, F d2 /mv 21 , vs. y/d.

16.11 A spherical shell with inner and outer dimensions of ri

and ro, respectively, has surface temperatures Ti(ri) and To(ro).

Assuming constant properties and one-dimensional (radial)

conduction, sketch the temperature distribution, T(r). Give

reasons for the shape you have sketched.

16.12 Heat is transferred by conduction (assumed to be onedimensional) along the axial direction through the truncated

conical section shown in the ﬁgure. The two base surfaces are

maintained at constant temperatures: T1 at the top, and T2, at the

bottom, where T1 > T2 : Evaluate the heat transfer rate, qx, when

a. the thermal conductivity is constant.

**Develop the equation that expresses the temperature difference
**

between the plate center line and its surface.

16.14 Heat is generated in a cylindrical fuel rod in a nuclear

reactor according to the relationship

"

2 #

r

q_ ¼ q_ max 1

ro

where q_ is the volumetric heat generation rate, kW/m3 , and ro is

the outside cylinder radius. Develop the equation that expresses the

temperature difference between the rod center line and its surface.

16.15 Heat is generated in a spherical fuel element according

to the relationship

"

3 #

r

q_ ¼ q_ max 1

ro

where q_ is the volumetric heat generation rate, kW/m3 , and ro is the

radius of the sphere. Develop the equation that expresses the temperature difference between the center of the sphere and its surface.

Chapter

17

Steady-State Conduction

In most equipment used in transferring heat, energy ﬂows from one ﬂuid to another

through a solid wall. As the energy transfer through each medium is one step in the

overall process, a clear understanding of the conduction mechanism of energy

transfer through homogeneous solids is essential to the solutions of most heat-transfer

problems.

In this chapter, we shall direct our attention to steady-state heat conduction. Steady

state implies that the conditions, temperature, density, and the like at all points in the

conduction region are independent of time. Our analyses will parallel the approaches

used for analyzing a differential ﬂuid element in laminar ﬂow and those that will be

used in analyzing steady-state molecular diffusion. During our discussions, two types of

presentations will be used: (1) The governing differential equation will be generated by

means of the control-volume concept and (2) the governing differential equation will be

obtained by eliminating all irrelevant terms in the general differential equation for

energy transfer.

**17.1 ONE-DIMENSIONAL CONDUCTION
**

For steady-state conduction independent of any internal generation of energy, the general

differential equation reduces to the Laplace equation

=2 T ¼ 0

(16-20)

**Although this equation implies that more than one space coordinate is necessary to
**

describe the temperature field, many problems are simpler because of the geometry of

the conduction region or because of symmetries in the temperature distribution.

One-dimensional cases often arise.

The one-dimensional, steady-state transfer of energy by conduction is the simplest

process to describe as the condition imposed upon the temperature ﬁeld is an ordinary

differential equation. For one-dimensional conduction, equation (16-20) reduces to

d

dT

xi

¼0

(17-1)

dx

dx

where i ¼ 0 for rectangular coordinates, i ¼ 1 for cylindrical coordinates, and i ¼ 2 for

spherical coordinates.

One-dimensional processes occur in ﬂat planes, such as furnace walls; in cylindrical elements, such as steam pipes; and in spherical elements, such as nuclear-reactor

pressure vessels. In this section, we shall consider steady-state conduction through

simple systems in which the temperature and the energy ﬂux are functions of a single

space coordinate.

224

17.1

One-Dimensional Conduction

225

**Plane Wall. Consider the conduction of energy through a plane wall as illustrated in
**

Figure 17.1. The one-dimensional Laplace equation is easily solved, yielding

T ¼ C1 x þ C2

(17-2)

The two constants are obtained by applying the boundary conditions

L

T1

T

T2

at x ¼ 0 T ¼ T1

and

x

**Figure 17.1 Plane wall
**

with a one-dimensional

temperature distribution.

at x ¼ L

T ¼ T2

**These constants are
**

C2 ¼ T1

and

C1 ¼

T2 T1

L

**The temperature proﬁle becomes
**

T¼

T2 T1

x þ T1

L

or

T ¼ T1

T 1 T2

x

L

(17-3)

**and is linear, as illustrated in Figure 17.1.
**

The energy ﬂux is evaluated, using the Fourier rate equation

qx

dT

¼ k

dx

A

(15-1)

**The temperature gradient, dT/dx, is obtained by differentiating equation (17-3) yielding
**

dT

T1 T2

¼

dx

L

Substituting this term into the rate equation, we obtain for a ﬂat wall with constant thermal

conductivity

qx ¼

kA

(T1 T2 )

L

(17-4)

**The quantity kA/L is characteristic of a ﬂat wall or a ﬂat plate and is designated the thermal
**

conductance. The reciprocal of the thermal conductance, L/kA, is the thermal resistance.

Composite Walls. The steady ﬂow of energy

through several walls in series is often encounL1

L2

L3

tered. A typical furnace design might include

T1

T2

one wall for strength, an intermediate wall for

insulation, and the third outer wall for appearqx

qx

ance. This composite plane wall is illustrated in

T3

Figure 17.2.

T4

For a solution to the system shown in this

x

ﬁgure, the reader is referred to Section 5.

The following example illustrates the use of Figure 17.2 Temperature distribution for

the composite-wall energy-rate equation for pre- steady-state conduction of energy through a

dicting the temperature distribution in walls.

composite plane wall.

226

Chapter 17

EXAMPLE 1

**Steady-State Conduction
**

A furnace wall is composed of three layers, 10 cm of ﬁrebrick (k ¼ 1:560 W/mK), followed by 23 cm

of kaolin insulating brick (k ¼ 0:073 W/mK), and ﬁnally 5 cm of masonry brick (k ¼ 1:0 W/mK).

The temperature of the inner wall surface is 1370 K and the outer surface is at 360 K. What are the

temperatures at the contacting surfaces?

The individual material thermal resistances per m2 of area are

L1

0:10 m

¼ 0:0641 K/W

¼

R1 ; firebrick ¼

k1 A1 (1:560 W/m: K)(1 m2 )

L2

0:23

R2 ; kaolin ¼

¼ 3:15 K/W

¼

k2 A2 (0:073)(1)

L3

0:05

¼ 0:05 K/W

¼

k3 A3 (1:0)(1)

The total resistance of the composite wall is equal to 0:0641 þ 3:15 þ 0:05 ¼ 3:26 K/W. The total

temperature drop is equal to (T1 T4 ) ¼ 1370 360 ¼ 1010 K.

Using equation (15-16), the energy transfer rate is

R3 ; masonry ¼

T1 T4

1010 K

¼ 309:8 W

¼

3:26 K/W

åR

As this is a steady-state situation, the energy transfer rate is the same for each part of the transfer path

(i.e., through each wall section). The temperature at the ﬁrebrick–kaolin interface, T2, is given by

q¼

T1 T2 ¼ q(R1 )

¼ (309:8 W)(0:0641 K/W) ¼ 19:9 K

giving

T2 ¼ 1350:1

Similarly,

giving

T3 T4 ¼ q(R3 )

¼ (309:8 W)(0:05 K/W) ¼ 15:5 K

T3 ¼ 375:5 K

**There are numerous situations in which a composite
**

wall involves a combination of series and parallel

energy-ﬂow paths. An example of such a wall is illustrated in Figure 17.3, where steel is used as reinforcement for a concrete wall. The composite wall can be

divided into three sections of length L1, L2, and L3, and

the thermal resistance for each of these lengths may be

evaluated.

The intermediate layer between planes 2 and 3

consists of two separate thermal paths in parallel; the

effective thermal conductance is the sum of the conductances for the two materials. For the section of the

wall of height y1 þ y2 and unit depth, the resistance is

1

1

¼ L2

R2 ¼

k1 y1 k2 y2

k1 y1 þ k2 y2

þ

L2

L2

k2

y2

k1

k2

L1

L2

L3

**Figure 17.3 A series-parallel
**

composite wall.

y1

17.1

One-Dimensional Conduction

227

**The total resistance for this wall is
**

åRT ¼ R1 þ R2 þ R3

or

L1

1

L3

åRT ¼

þ

þ L2

k1 (y1 þ y2 )

k1 (y1 þ y2 )

k1 y1 þ k2 y2

The electrical circuit

R3

R1

is an analog to the composite wall.

R2

**The rate of energy transferred from plane 1 to plane 4 is obtained by a modiﬁed form of
**

equation (15-16).

q¼

T1 T4

¼

åRT

T T4

1

L1

1

L3

þ L2

þ

k1 (y1 þ y2 )

k1 (y1 þ y2 )

k1 y1 þ k2 y2

(17-5)

**It important to recognize that this equation is only an approximation. Actually, there is a
**

signiﬁcant temperature distribution in the y direction close to the material that has the higher

thermal conductivity.

In our discussions of composite walls, no allowance was made for a temperature drop

at the contact face between two different solids. This assumption is not always valid, has

there will often be vapor spaces caused by rough surfaces, or even oxide ﬁlms on the surfaces

of metals. These additional contact resistances must be accounted for in a precise energytransfer equation.

Long, Hollow Cylinder. Radial energy ﬂow by conduction through a long, hollow

cylinder is another example of one-dimensional conduction. The radial heat ﬂow for this

conﬁguration is evaluated in Example 1 of chapter 15 as

qr

2pk

¼

(Ti To )

L ln(ro /ri )

(17-6)

**where ri is the inside radius, ro is the outside radius, Ti is the temperature on the inside
**

surface, and To is the temperature on the outside surface. The resistance concept may

again be used; the thermal resistance of the hollow cylinder is

ln(ro /ri )

(17-7)

R¼

2pkL

The radial temperature distribution in a long, hollow cylinder may be evaluated by

using equation (17-1) in cylindrical form

d

dT

r

¼0

(17-8)

dr

dr

Solving this equation subject to the boundary conditions

at r ¼ ri T ¼ Ti

and

at r ¼ ro

we see that temperature proﬁle is

T(r) ¼ Ti

T ¼ To

Ti To

r

ln

ln(ro /ri )

ri

(17-9)

T2 . are R2 ¼ ln(r3 /r2 ) 2pk2 L for the insulation. and R3 ¼ 1 1 ¼ hA h2pr3 L for the air ﬁlm. we obtain qr ¼ 2pL(T2 T1 ) ½ln(r3 /r2 )/k2 þ 1/hr3 (17-11) The dual effect of increasing the resistance to energy transfer by conduction and simultaneously increasing the surface area as r3 is increased suggests that. the relative importance of each resistance term will change as the insulation thickness is varied. and the term 1/r3 decreases as r3 increases. under what conditions will hollow cylinder. In Example 3 of Chapter 15. whereas for the plane wall the temperature distribution is linear. Differentiating equation (17-11) with respect to r3. T1 . the total difference in temperature is T2 T1 and the two resistances. for a pipe of given size. the thermal resistance of a hollow cylindrical element was shown to be R¼ ln(ro /ri ) 2pkL (17-10) In the present example. hollow cylinder is a logarithmic function of radius r.228 Chapter 17 Steady-State Conduction Thus. the temperature in a long. L. Substituting these terms into the radial heat ﬂow equation and rearranging. this situation arise? Figure 17. EXAMPLE 2 A long steam pipe of outside radius r2 is covered with thermal insulation having an outside radius of r3. As the ratio r3/r2 increases logarithmically. hollow cylinder. and the temperature of the surrounding air.k2 h and r2 are considered constant. due to the insulation and the surrounding air ﬁlm.4 may be used to illustrate this composite cylinder.4 A series composite ness of insulation? If possible. The energy loss per unit area of outside surface of the insulation is described by the Newton rate equation qr ¼ h(T3 T1 ) (15-11) A r3 r1 r2 Can the energy loss increase with an increase in the thickFigure 17. a particular outer radius exists for which the heat loss is maximum. The following example illustrates the analysis of radial energy conduction through a long.T1 . The temperature of the outer surface of the pipe. we obtain dqr ¼ dr3 1 1 2 k2 r3 hr3 " #2 1 r3 1 þ ln r2 k2 hr3 2pL(T2 T1 ) (17-12) . In this example. T2. are ﬁxed.

17.1

One-Dimensional Conduction

229

**The radius of insulation associated with the maximum energy transfer, the critical radius, found
**

by setting dqr /dr3 ¼ 0; equation (17-12) reduces to

(r3 )critical ¼

k2

h

(17-13)

**In the case of 85% magnesia insulation (k ¼ 0:0692 W/mK) and a typical value for the
**

heat transfer coefﬁcient in natural convection (h ¼ 34 W/m2 K), the critical radius is

calculated as

rcrit ¼

k 0:0692 W/mK

¼ 0:0020 m

¼

h

34 W/m2 K

¼ 0:20 cm

(0:0067 ft)

(0:0787 in:)

**These very small numbers indicate that the critical radius will be exceeded in any practical
**

problem. The question then is whether the critical radius given by equation (17-13)

represents a maximum or a minimum condition for q. The evaluation of the second

derivative, d2 qr /dr32 , when r3 ¼ k/h yields a negative result, thus rcrit is a maximum

condition. It now follows that qr will be decreased for any value of r3 greater than

0.0020 m.

Hollow Sphere. Radial heat ﬂow through a hollow sphere is another example of

one-dimensional conduction. For constant thermal conductivity, the modiﬁed Fourier rate

equation

qr ¼ k

dT

A

dr

**applies, where A ¼ area of a sphere ¼ 4pr 2, giving
**

qr ¼ 4pkr 2

dT

dr

(17-14)

**This relation, when integrated between the boundary conditions
**

at T ¼ Ti

r ¼ ri

at T ¼ To

r ¼ ro

and

yields

q¼

4pk(Ti To )

1 1

r i r0

The hyperbolic temperature distribution

Ti To

1 1

T ¼ Ti

1/ri 1/ro

ri r

(17-15)

(17-16)

is obtained by using the same procedure that was followed to obtain equation (17-9).

230

Chapter 17

Steady-State Conduction

**Variable Thermal Conductivity. If the thermal conductivity of the medium through
**

which the energy is transferred varies signiﬁcantly, the preceding equations in this section

do not apply. As Laplace’s equation involves the assumption of constant thermal conductivity, a new differential equation must be determined from the general equation for heat

transfer. For steady-state conduction in the x direction without internal generation of energy,

the equation that applies is

d

dT

k

¼0

(17-17)

dx

dx

where k may be a function of T.

In many cases the thermal conductivity may be a linear function of temperature

over a considerable range. The equation of such a straight-line function may be

expressed by

k ¼ ko (1 þ bT)

where ko and b are constants for a particular material. In general, for materials satisfying

this relation, b is negative for good conductors and positive for good insulators. Other

relations for varying k have been experimentally determined for speciﬁc materials. The

evaluation of the rate of energy transfer when the material has a varying thermal

conductivity is illustrated in Example 2 of chapter 15.

**17.2 ONE-DIMENSIONAL CONDUCTION WITH INTERNAL
**

GENERATION OF ENERGY

In certain systems, such as electric resistance heaters and nuclear fuel rods, heat is

generated within the conducting medium. As one might expect, the generation of energy

within the conducting medium produces temperature proﬁles different than those for

simple conduction.

In this section, we shall consider two simple example cases: steady-state

conduction in a circular cylinder with uniform or homogeneous energy generation,

and steady-state conduction in a plane wall with variable energy generation. Carslaw

and Jaeger 1 and Jakob 2 have written excellent treatises dealing with more complicated problems.

Cylindrical Solid with Homogeneous

Energy Generation. Consider a cylindrical solid with internal energy generation

as shown in Figure 17.5. The cylinder

will be considered long enough so that

only radial conduction occurs. The density, r, the heat capacity, cp, and the

thermal conductivity of the material will

r

r

r

**Figure 17.5 Annular element in a long, circular
**

cylinder with internal heat generation.

1

H. S. Carslaw and J. C. Jaeger, Conduction of Heat in Solids, 2nd Edition, Oxford Univ. Press, New York,

1959.

2

M. Jakob, Heat Transfer, Vol. I, Wiley, New York, 1949.

17.2

One-Dimensional Conduction with Internal Generation of Energy

**be considered constant. The energy balance for the element shown is
**

9

9 8

9 8

8

>

=

= >

< rate of energy >

= >

< rate of energy >

< rate of energy >

conduction into þ generation within conduction out

>

>

>

>

;

; >

:

; >

:

:

of the element

the element

the element

9

8

>

=

< rate of accumulation >

¼

231

(17-18)

of energy

>

;

within the element

Applying the Fourier rate equation and letting q_ represent the rate of energy generated per

unit volume, we may express equation (17-18) by the algebraic expression

@T

@T

@T

_

k(2prL) þq(2prL

(2prL Dr)

Dr) k(2prL)

¼ rcp

@r

@r

@t

>

:

r

rþDr

**Dividing each term by 2prL Dr, we obtain
**

q_ þ

k½r(@T/@r)jrþDr r(@T/@r)jr

@T

¼ rcp

@t

r Dr

In the limit as Dr approaches zero, the following differential equation is generated:

k @

@T

@T

r

(17-19)

q_ þ

¼ rcp

r @r

@r

@t

For steady-state conditions, the accumulation term is zero; when we eliminate this term

from the above expression, the differential equation for a solid cylinder with homogeneous

energy generation becomes

k d

dT

r

q_ þ

¼0

(17-20)

r dr

dr

The variables in this equation may be separated and integrated to yield

rk

dT

r2

þ q_ ¼ C1

dr

2

k

dT

r C1

þ q_ ¼

dr

2

r

or

**Because of the symmetry of the solid cylinder, a boundary condition that must be satisﬁed
**

stipulates that the temperature gradient must be ﬁnite at the center of the cylinder, where

r ¼ 0. This can be true only if C1 ¼ 0. Accordingly, the above relation reduces to

k

dT

r

þ q_ ¼ 0

dr

2

(17-21)

_ 2

qr

þ C2

4k

(17-22)

**A second integration will now yield
**

T¼

**If the temperature T is known at any radial value, such as a surface, the second
**

constant, C2, may be evaluated. This, of course, provides the completed expression

for the temperature profile. The energy flux in the radial direction may be obtained

232

Chapter 17

Steady-State Conduction

from

qr

dT

¼ k

dr

A

by substituting equation (17-21), yielding

qr

r

¼ q_

2

A

or

r

qr ¼ (2prL)q_ ¼ pr 2 Lq_

2

2L

(17-23)

**Plane Wall with Variable Energy Generation. The second case associated with energy
**

generation involves a temperature-dependent, energy-generating process. This situation

develops when an electric current is passed through a conducting medium possessing an

electrical resistivity that varies with temperature. In our discussion, we shall assume that

the energy-generation term varies linearly with temperature, and that the conducting

medium is a ﬂat plate with temperature TL at both surfaces. The internal energy generation

is described by

(17-24)

q_ ¼ q_ L ½1 þ b(T TL )

**where q_ L is the generation rate at the surface and b is a constant.
**

With this model for the generation function, and as both surfaces have the same

temperature, the temperature distribution within the ﬂat plate is symmetric about the

x

midplane. The plane wall and its coordinate system are illustrated in Figure 17.6. The

–L

+L symmetry of the temperature distribution requires a zero temperature gradient at x ¼ 0.

With steady-state conditions, the differential equation may be obtained by eliminating the

Figure 17.6 Flat plate

irrelevant terms in the general differential equation for heat transfer. Equation (16-19) for

with temperaturethe case of steady-state conduction in the x direction in a stationary solid with constant

dependent energy

thermal conductivity becomes

generation.

d2 T q_ L

þ ½1 þ b(T TL ) ¼ 0

dx2

k

The boundary conditions are

at x ¼ 0

dT

¼0

dx

and

at x ¼ L T ¼ TL

These relations may be expressed in terms of a new variable, u ¼ T TL , by

d2 u q_ L

þ (1 þ bu) ¼ 0

dx2

k

or

d2 u

þ C þ su ¼ 0

dx2

where C ¼ q_ L /k and s ¼ bq_ L /k. The boundary conditions are

at x ¼ 0

du

¼0

dx

17.3

Heat Transfer from Extended Surfaces

233

and

at x ¼ L u ¼ 0

The integration of this differential equation is simpliﬁed by a second change in variables;

inserting f for C þ su into the differential equation and the boundary conditions, we obtain

d2 f

þ sf ¼ 0

dx2

for

x¼0

df

¼0

dx

and

x ¼ L

The solution is

f¼C

pﬃﬃ

pﬃﬃ

f ¼ C þ su ¼ A cos(x s) þ B sin(x s)

or

pﬃﬃ

pﬃﬃ

C

u ¼ A1 cos(x s) þ A2 sin(x s)

s

The temperature distribution becomes

pﬃﬃ

1 cos(x s)

pﬃﬃ 1

T TL ¼

b cos(L s)

(17-25)

**where s ¼ bq_ L /k is obtained by applying the two boundary conditions.
**

The cylindrical and spherical examples of one-dimensional temperature-dependent

generation are more complex; solutions to these may be found in the technical literature.

**17.3 HEAT TRANSFER FROM EXTENDED SURFACES
**

A very useful application of one-dimensional heat-conduction analysis is that of describing

the effect of extended surfaces. It is possible to increase the energy transfer between a

surface and an adjacent ﬂuid by increasing the amount of surface area in contact with the

ﬂuid. This increase in area is accomplished by adding extended surfaces that may be in the

forms of ﬁns or spines of various cross sections.

The one-dimensional analysis of extenq3

ded surfaces may be formulated in general

terms by considering the situation depicted

A(x)

in Figure 17.7.

The shaded area represents a portion of

the extended surface that has variable crossq1

q2

sectional area, A(x), and surface area, S(x),

which are functions of x alone. For steadystate conditions, the ﬁrst law of thermodynamics, equation (6-10), reduces to the simple

S(x)

x

expression

x

dQ

¼0

dt

**Figure 17.7 An extended surface of general
**

conﬁguration.

234

Chapter 17

Steady-State Conduction

**Thus, in terms of the heat ﬂow rates designated in the ﬁgure, we may write
**

q1 ¼ q2 þ q3

(17-26)

**The quantities q1 and q2 are conduction terms, while q3 is a convective heat-flow rate.
**

Evaluating each of these in the appropriate way and substituting into equation (17-26), we

obtain

dT

dT

kA hS(T T1 ) ¼ 0

(17-27)

kA

dx

dx

xþDx

x

**where T1 is the ﬂuid temperature. Expressing the surface area, S(x), in terms of the width,
**

Dx, times the perimeter, P(x), and dividing through by Dx, we obtain

kA(dT/dx)jxþDx kA(dT/dx)jx

hP(T T1 ) ¼ 0

Dx

Evaluating this equation in the limit as Dx ! 0, we obtain the differential equation

d

dT

kA

(17-28)

hPðT T1 Þ ¼ 0

dx

dx

One should note, at this point, that the temperature gradient, dT/dx, and the surface

temperature, T, are expressed such that T is a function of x only. This treatment assumes the

temperature to be ‘‘lumped’’ in the transverse direction. This is physically realistic when

the cross section is thin or when the material thermal conductivity is large. Both of these

conditions apply in the case of ﬁns. More will be said about the ‘‘lumped parameter’’

approach in Chapter 18. This approximation in the present case leads to equation (17-28),

an ordinary differential equation. If we did not make this simplifying analysis, we would

have a distributed parameter problem that would require solving a partial differential

equation.

A wide range of possible forms exist when equation (17-28) is applied to speciﬁc

geometries. Three possible applications and the resulting equations are described in the

following paragraphs.

(1) Fins or Spines of Uniform Cross Section. For either of the cases shown in Figure 17.8,

the following are true: A(x) ¼ A, and P(x) ¼ P, both constants. If, additionally, both k and h

are taken to be constant, equation (17-28) reduces to

d 2 T hP

(T T1 ) ¼ 0

dx2 kA

x

x

Figure 17.8 Two examples of extended surfaces with constant cross section.

(17-29)

17.3

x

Heat Transfer from Extended Surfaces

235

2tL

x

t0

t0

L

L

(a)

(b)

Figure 17.9 Two examples of straight extended surfaces with variable cross section.

**(2) Straight Surfaces with Linearly Varying Cross Section. Two conﬁgurations for
**

which A and P are not constant are shown in Figure 17.9. If the area and perimeter both vary

in a linear manner from the primary surface, x ¼ 0, to some lesser value at the end, x ¼ L,

both A and P may be expressed as

x

(17-30)

A ¼ A0 (A0 AL )

L

and

P ¼ P0 (P0 PL )

x

L

(17-31)

**In the case of the rectangular fin shown in Figure 17.9(b), the appropriate values of A and
**

P are

A0 ¼ 2t0 W

AL ¼ 2tL W

P0 ¼ 2½2t0 þ W PL ¼ 2½2tL þ W

where t0 and tL represent the semithickness of the ﬁn evaluated at x ¼ 0 and x ¼ L,

respectively, and W is the total depth of the ﬁn.

For constant h and k, equation (17-28) applied to extended surfaces with cross-sectional

area varying linearly becomes

h

x i d2 T A0 AL dT h h

xi

P

A0 (A0 AL )

(P

P

)

(T T1 ) ¼ 0 (17-32)

L

0

0

L dx2

dx k

L

L

(3) Curved Surfaces of Uniform Thickness. A common type of extended surface is that of the circular ﬁn

of constant thickness as depicted in Figure 17.10. The

appropriate expressions for A and P, in this case, are

)

A ¼ 4prt

and

r0

r

rL

P ¼ 4pr

When these expressions are substituted into equation

(17-28), the applicable differential equation, considering k and h constant, is

d2 T 1 dT h

(T T1 ) ¼ 0

þ

dr 2 r dr kt

(17-33)

r

r0

rL

Figure 17.10 A curved ﬁn of

constant thickness.

2t

236

Chapter 17

Steady-State Conduction

**Equation (17-33) is a form of Bessel’s equation of zero order. The solution is in terms of
**

Bessel functions of the ﬁrst kind. The description and use of these functions are beyond the

mathematical scope of this text. The interested reader may consult the work of Kraus et al.3

for a complete discussion of Bessel functions and their use.

In each of the cases considered, the thermal conductivity and convective heat-transfer

coefﬁcient were assumed constant. When the variable nature of these quantities is considered,

the resulting differential equations become still more complex than those developed thus far.

Solutions for the temperature proﬁle in the case of the straight ﬁn of constant cross

section will now be considered; equation (17-29) applies.

The general solution to equation (17-29) may be written

u ¼ c1 emx þ c2 emx

(17-34)

or

u ¼ A cosh mx þ B sinh mx

(17-35)

where m2 ¼ hP/kA and u ¼ T T1 . The evaluation of the constants of integration requires

that two boundary conditions be known. The three sets of boundary conditions that we shall

consider are as follows:

(a) T ¼ T0

at x ¼ 0

T ¼ TL

at x ¼ L

at x ¼ 0

(b)

T ¼ T0

dT

¼0

at x ¼ L

dx

and

(c)

T ¼ T0

dT

¼ h(T T1 )

k

dx

at

x¼0

at

x¼L

The ﬁrst boundary condition of each set is the same and stipulates that the temperature at the

base of the extended surface is equal to that of the primary surface. The second boundary

condition relates the situation at a distance L from the base. In set (a) the condition is that of a

known temperature at x ¼ L. In set (b) the temperature gradient is zero at x ¼ L. In set (c) the

requirement is that heat ﬂow to the end of an extended surface by conduction be equal to that

leaving this position by convection.

The temperature proﬁle, associated with the ﬁrst set of boundary conditions, is

mx

u

T T1

uL

e emx

¼

¼

emL

(17-36)

þ emx

u0 T0 T1

u0

emL emL

A special case of this solution applies when L becomes very large, that is, L ! 1, for

which equation (17-36) reduces to

u

T T1

¼

¼ emx

u0 T0 T1

(17-37)

**The constants, c1 and c2, obtained by applying set (b), yield, for the temperature proﬁle,
**

u

T T1

emx

emx

¼

¼

þ

u0 T0 T1 1 þ e2mL 1 þ e2mL

3

(17-38)

**A. D. Kraus, A. Aziz, and J. R. Welty, Extended Surface Heat Transfer, Wiley-Interscience, New York,
**

2001.

17.3

Heat Transfer from Extended Surfaces

237

**An equivalent expression to equation (17-38) but in a more compact form is
**

u

T T1

cosh½m(L x)

¼

¼

u 0 T 0 T1

cosh mL

(17-39)

**Note that, in either equation (17-38) or (17-39), as L ! 1 the temperature proﬁle
**

approaches that expressed in equation (17-37).

The application of set (c) of the boundary conditions yields, for the temperature proﬁle,

u

T T1

cosh½m(L x) þ (h/mk)sinh½m(L x)

¼

¼

u0 T0 T1

cosh mL þ (h/mk)sinh mL

(17-40)

**It may be noted that this expression reduces to equation (17-39) if du/dx ¼ 0 at x ¼ L and
**

to equation (17-37) if T ¼ T1 at L ¼ 1:

The expressions for T(x) that have been obtained are particularly useful in evaluating

the total heat transfer from an extended surface. This total heat transfer may be determined

by either of two approaches. The ﬁrst is to integrate the convective heat-transfer expression

over the surface according to

Z

Z

q¼

h½T(x) T1 dS ¼

hu dS

(17-41)

S

S

**The second method involves evaluating the energy conducted into the extended surface at
**

the base as expressed by

dT

q ¼ kA

(17-42)

dx x¼0

The latter of these two expressions is easier to evaluate; accordingly, we will use this

equation in the following development.

Using equation (17-36), we ﬁnd that the heat transfer rate, when set (a) of the boundary

conditions applies, is

uL /u0 emL

q ¼ kAmu0 1 2 mL

(17-43)

e emL

If the length L is very long, this expression becomes

q ¼ kAmu0 ¼ kAm(T0 T1 )

(17-44)

**Substituting equation (17-39) [obtained by using set (b) of the boundary conditions]
**

into equation (17-42), we obtain

q ¼ kAmu0 tanh mL

(17-45)

**Equation (17-40), utilized in equation (17-42), yields for q the expression
**

q ¼ kAmu0

sinh mL þ (h/mk)cosh mL

cosh mL þ (h/mk)sinh mL

(17-46)

**The equations for the temperature proﬁle and total heat transfer for extended surfaces of
**

more involved conﬁguration have not been considered. Certain of these cases will be left as

exercises for the reader.

A question that is logically asked at this point is, ‘‘What beneﬁt is accrued by the addition of

extended surfaces?’’ A term that aids in answering this question is the ﬁn efﬁciency, symbolized

as hf , deﬁned as the ratio of the actual heat transfer from an extended surface to the maximum

possible heat transfer from the surface. The maximum heat transfer would occur if the

temperature of the extended surface were equal to the base temperature, T0, at all points.

238

Chapter 17

**Steady-State Conduction
**

1.0

0.9

r0

rL

0.8

2t

3.0

0.7

1.6

0.6

nf

**For a straight fin rL – r0 = L
**

1.4

2.0

0.5

rL /r0 = 1.0 (straight fin)

0.4

1.8

4.0

0.3

0.2

0.1

0

0

1.0

2.0

(rL /r0 )

3.0

4.0

5.0

h/kt

Figure 17.11 Fin efﬁciency for straight and circular ﬁns of constant thickness.

**Figure 17.11 is a plot of hf as a function of a signiﬁcant parameter for both straight and
**

circular ﬁns of constant thickness (when ﬁn thickness is small, t rL r0 ).

The total heat transfer from a ﬁnned surface is

qtotal ¼ qprimary surface þ qfin

¼ A0 h(T0 T1 ) þ Af h(T T1 )

(17-47)

**The second term in equation (17-47) is the actual heat transfer from the fin surface in
**

terms of the variable surface temperature. This may be written in terms of the fin

efficiency, yielding

qtotal ¼ A0 h(T0 T1 ) þ Af hhf (T0 T1 )

or

qtotal ¼ h(A0 þ Af hf )(T0 T1 )

(17-48)

**In this expression A0 represents the exposed area of the primary surface, Af is the total fin
**

surface area, and the heat transfer coefficient, h, is assumed constant.

The application of equation (17-48) as well as an idea of the effectiveness of ﬁns is

illustrated in Example 3.

EXAMPLE 3

Water and air are separated by a mild-steel plane wall. It is proposed to increase the heat-transfer rate

between these ﬂuids by adding straight rectangular ﬁns of 1.27-mm thickness and 2.5-cm length,

spaced 1.27 cm apart. The air-side and water-side heat-transfer coefﬁcients may be assumed constant

with values of 11.4 and 256 W/m2 K respectively. Determine the percent change in total heat transfer

when ﬁns are placed on (a) the water side, (b) the air side, and (c) both sides.

17.3

Heat Transfer from Extended Surfaces

239

For a 1 m2 section of the wall, the areas of the primary surface and of the ﬁns are

0:00127 m

Ao ¼ 1 m2 79 fins (1 m)

fin

¼ 0:90 m2

A f ¼ 79 fins (1 m)½(2)(0:025 m) þ 0:10 m2

¼ 4:05 m2

Values of ﬁn efﬁciency can now be determined from Figure 17.11. For the air side

1=2

pﬃﬃﬃﬃﬃﬃﬃﬃ

11:4 W/m2 K

L h/kt ¼ 0:025 m

(42:9 W/mK)(0:00127 m)

¼ 0:362

and for the water side

pﬃﬃﬃﬃﬃﬃﬃﬃﬃ

L h/kT ¼ 0:025 m

256 W/m2 K

(42:9 W/mK)(0:00127 m)

1=2

¼ 1:71

The ﬁn efﬁciencies are then read from the ﬁgure as

hair ﬃ 0:95

hwater ﬃ 0:55

The total heat transfer rates can now be evaluated. For ﬁns on the air side

q ¼ ha D Ta ½Ao þ hfa Af

¼ 11:4 DTa ½0:90 þ 0:95(4:05)

¼ 54:1 DTa

and on the water side

q ¼ hw DTw ½Ao þ hfw Af

¼ 256DTw ½0:90 þ 0:55(4:05)

¼ 801 DTw

The quantities DTa and DTw represent the temperature differences between the steel surface at

temperature To and the ﬂuids.

The reciprocals of the coefﬁcients are the thermal resistances of the ﬁnned surfaces.

Without ﬁns the heat-transfer rate in terms of the overall temperature difference,

DT ¼ Tw Ta , neglecting the conductive resistance of the steel wall, is

q¼

DT

¼ 10:91 DT

1

1

þ

11:4 256

**With ﬁns on the air side alone
**

q¼

DT

¼ 44:67 DT

1

1

þ

54:1 256

**an increase of 310% compared with the bare-wall case.
**

With ﬁns on the water side alone

q¼

an increase of 3.0%.

DT

¼ 11:24 DT

1

1

þ

11:4 801

240

Chapter 17

**Steady-State Conduction
**

With ﬁns on both sides the heat-ﬂow rate is

q¼

DT

¼ 50:68 DT

1

1

þ

54:1 801

**an increase of 365%.
**

This result indicates that adding ﬁns is particularly beneﬁcial where the convection coefﬁcient

has a relatively small value.

**17.4 TWO- AND THREE-DIMENSIONAL SYSTEMS
**

In Sections 17.2 and 17.3, we discussed systems in which the temperature and the energy

transfer were functions of a single-space variable. Although many problems fall into this

category, there are many other systems involving complicated geometry or temperature

boundary conditions, or both, for which two or even three spatial coordinates are necessary

to describe the temperature ﬁeld.

In this section, we shall review some of the methods for analyzing heat transfer by

conduction in two- and three-dimensional systems. The problems will mainly involve twodimensional systems, as they are less cumbersome to solve yet illustrate the techniques of

analysis.

y

L

T=0

T=0

T = T1

x

**Figure 17.12 Model for
**

two-dimensional conduction

analysis.

**Analytical Solution. An analytical solution to any transfer problem must satisfy the
**

differential equation describing the process as well as the prescribed boundary conditions. Many mathematical techniques have been used to obtain solutions for particular

energy conduction situations in which a partial differential equation describes the

temperature ﬁeld. Carslaw and Jaeger4 and Boelter et al.5 have written excellent treatises

that deal with the mathematical solutions for many of the more complex conduction

problems. As most of this material is too specialized for an introductory course, a

solution will be obtained to one of the ﬁrst cases analyzed by Fourier6 in the classical

treatise that established the theory of energy transfer by conduction. This solution of a

two-dimensional conduction medium employs the mathematical method of separation of

variables.

Consider a thin, inﬁnitely long rectangular plate that is free of heat sources, as illustrated

in Figure 17.12. For a thin plate @T/@z is negligible, and the temperature is a function of x

and y only. The solution will be obtained for the case in which the two edges of the plate

are maintained at zero temperature and the bottom is maintained at T1 as shown. The steadystate temperature distribution in the plate of constant thermal conductivity must satisfy the

differential equation

@2T @2T

þ

¼0

@x2 @y2

(17-49)

4

H. S. Carslaw and J. C. Jaeger, Conduction of Heat in Solids, 2nd Edition, Oxford Univ. Press, New York,

1959.

5

L. M. K. Boelter, V. H. Cherry, H. A. Johnson, and R. C. Martinelli, Heat Transfer Notes, McGraw-Hill

Book Company, New York, 1965.

6

J. B. J. Fourier, Theorie Analytique de la Chaleur, Gauthier-Villars, Paris, 1822.

4 Two.17.and Three-Dimensional Systems 241 and the boundary conditions T¼0 T¼0 T ¼ T1 at at at x¼0 x¼L y¼0 T¼0 at y¼1 for all values of y for all values of y for 0 .

x .

L and for 0.

x.

y) ¼ B sin Cenpy/L þ Denpy/L (17-55) L The requirement that T ¼ 0 at y ¼ 1 stipulates that C must be zero.L Equation (17-49) is a linear. it follows that both must be independent of x and y. accordingly. yielding X ¼ A cos lx þ B sin lx and Y ¼ Cely þ Dely According to equation (17-50). y) ¼ XY ¼ (A cos lx þ B sin lx)(Cely þ Dely ) (17-54) where A. we obtain an expression in which the variables are separated 1 d2 X 1 d2 Y ¼ X dx2 Y dy2 (17-51) As the left-hand side of equation (17-51) is independent of y and the equivalent righthand side is independent of x. the temperature distribution is deﬁned by the relation T(x. y) ¼ X(x)Y(y) (17-50) where X(x) is a function of x only and Y(y) is a function of y only. lL must be an integral multiple of p or l ¼ np/L: Equation (17-54) is now reduced to npx T(x. T(x. sin lx must be zero at x ¼ L. y) ¼ Eenpy/L sin L . If we designate this constant l2 . The condition that T ¼ 0 at x ¼ 0 requires that A ¼ 0: Similarly. and hence must be equal to a constant. two ordinary differential equations result d2 X þ l2 X ¼ 0 dx2 (17-52) d2 Y l2 Y ¼ 0 dy2 (17-53) and These differential equations may be integrated. homogeneous partial differential equation. This type of equation usually can be integrated by assuming that the temperature distribution. A combination of B and D into the single constant E reduces equation (17-55) to npx T(x. and D are constants to be evaluated from the four boundary conditions. is of the form T(x. C. y). Substituting this equation into equation (17-49). B.

is used to evaluate En according to the expression npx 1 T1 ¼ å En sin for 0 . giving npx 1 (17-56) T ¼ å En enpy/L sin L n¼1 The last boundary condition. The general solution is obtained by summing all possible solutions.242 Chapter 17 Steady-State Conduction This expression satisﬁes the differential equation for any integer n greater than or equal to zero. T ¼ T1 at y ¼ 0.

x .

3. Shape Factors for Common Conﬁgurations The shape factor. As an alternative approach. 6. the dimension normal to the plane shown in very large. L L n¼1 The constants En are the Fourier coefﬁcients for such an expansion and are given by En ¼ 4T1 np for n ¼ 1. which may be integrated subject to the given boundary conditions.1 lists expressions for shape factors of ﬁve conﬁgurations. both with isothermal boundaries. In every case depicted. practical problems with complicated geometry and boundary conditions. that is. . it is presumed that the heat transfer problem is two dimensional. one must turn to numerical methods. is deﬁned and discussed brieﬂy in Chapter 15. Analytical solutions require relatively simple functions and geometries. which cannot be solved analytically. The isotherms are shown in the figure as solid lines. are energy-flow lines. three secondorder ordinary differential equations are obtained. When the variables are separated. however. and En ¼ 0 T=0 T=0 The solution to this two-dimensional conduction problem is T¼ T = T1 Figure 17. S. . Numerical Solutions Each of the solution techniques discussed thus far for multidimensional conduction has considerable utility when conditions permit its use. . the use of shape factors requires isothermal boundaries. a knowledge of the shape factor makes the determination of heat ﬂow a simple calculation. 4T1 1 e½(2nþ1)py/L (2n þ 1)px å sin L p n¼0 2n þ 1 (17-57) The isotherms and energy flow lines are plotted in Figure 17.13.12. There are. for n ¼ 2. and the dotted lines. which are orthogonal to the isotherms. .13 Isotherms and energy ﬂow lines for the rectangular plate in Figure 17. 4. The separation of variables method can be extended to three-dimensional cases by assuming T to be equal to the product X(x)Y(y)Z(z) and substituting this expression for T into the applicable differential equation. Note the similarity to the lines of constant velocity potential and stream function as discussed in momentum transfer. 5. . When a geometric case of interest involves conduction between a source and a sink. When the situation of interest becomes sufﬁciently complex or when . Analytical solutions are useful when they can be obtained. . Table 17.

New York. O. H.and Three-Dimensional Systems Table 17. j i. Wilkes. A more complete and detailed discussion of numerical solutions to heat conduction problems may be found in Carnahan et al. R. With the presence of digital computers to accomplish the large number of manipulations inherent in numerical solutions rapidly and accurately. 1974. j – 1 Figure 17. Luther. Shape factor.14 is a twodimensional representation of an element within a conducting medium.1 Conduction shape factors. B. J.4 Two. Wiley. this approach is now very common.7 and in Welty. New York. 1969. A. one must turn to numerical solutions. e e/ro Tb e Eccentric circular cylinders Tb Ti rb 2p ln(ro /ri ) 0:10669 ri Circular cylinder in a hexagonal cylinder Tb Ti 2p ln(ro /ri ) 0:27079 ri rb Circular cylinder in a square cylinder Tb r 2p cosh1 (r/r) Ti r Inﬁnite cylinder buried in semi-inﬁnite medium boundary conditions preclude the use of simple solution techniques. Welty. In this section we shall introduce the concepts of numerical problem formulation and solution.17. and J. j + 1 x i – 1. Wiley. S q/L ¼ kS(Ti To ) Shape 2p ln(ro /ri ) Tb Ti ri rb Concentric circular cylinders ri Ti rb 2p 1 þ r2 e2 cosh1 2r r ri /ro . Applied Numerical Methods.14 Two-dimensional volume element in a conducting medium. 243 .8 Shown in Figure 17. j y i. The element 7 8 i. Engineering Heat Transfer. Carnahan. j i + 1.

A direct application of equation (6-10) to node i. The designation. j þ q_ Dx2 ¼0 k (17-62) In the absence of internal generation. The wall material has a thermal conductivity of 1:21 W/mK: We may take advantage of the eightfold symmetry of this ﬁgure to lay out the simple square grid shown below (right). that is. j1 þ Ti. j1 þ Ti. A simple example showing the use of equation (17-63) in solving a two-dimensional heat conduction problem follows. j. heat transfer is assumed positive. j is centered in the ﬁgure along with its adjacent nodes. j jt þ q_ Dx Dy ¼ Dx Dy ð17-61Þ Dt This expression has been considered in a more complete form in the next chapter. and the last is the generation term. implies a general location in a two-dimensional system where i is a general index in the x direction and j is the y index. For the present we will not consider time-variant terms. we obtain dQ Dy Dy ¼ k (Ti1. but for now we will allow these dimensions to be different. . j ) þ q_ Dx Dy Dy Dy The ﬁrst two terms in this expression relate conduction in the x direction. Setting these expressions equal to each other and simplifying.’’ that is. j þ Tiþ1. j is the arithmetic mean of the temperatures of its adjacent nodes. j ) þ k (Tiþ1. jþ1 2Ti. and constant height. jþ1 Ti. The rate of energy increase within node i. j ¼ 4 or. moreover. Dx ¼ Dy: With these simplifications equation (17-61) becomes Ti1. j1 þ Ti. j 2Ti. may be evaluated allowing for conduction into node i. we have Dy Dx k ½Ti1. j Ti. j þ Ti. j may be written simply as ZZZ rcTjtþDt rcTjt @ er dV ¼ Dx Dy (17-60) @t Dt c:v: Equation (17-58) indicates that the expressions given by equations (17-59) and (17-60) may be equated. EXAMPLE 4 A hollow square duct of the conﬁguration shown (left) has its surfaces maintained at 200 and 100 K.244 Chapter 17 Steady-State Conduction or ‘‘node’’ i. Dx ¼ Dy. j þ Ti. j yields ZZZ dQ @ ¼ er dV (17-58) dt @t c:v: The heat input term. (17-62) may be solved for Tij to yield Ti1. j jtþDt rcTi. i. Evaluating dQ/dt in this manner. All of these terms are positive. we will consider the nodes to be square. the temperature of node i. j Dx Dy rcTi. jþ1 (17-63) Ti. The grid is set up with constant node width. j þ Tiþ1. Determine the steady-state heat transfer rate between the hot and cold surfaces of this duct. jþ1 4Ti.14. the third and fourth express y-directional conduction. j þ Tiþ1. Dy: It may be convenient to make the grid ‘‘square. j þ k ½Ti. Adjacent node indices are shown in Figure 17.j1 Ti. j ) dt Dx Dx Dx Dx ð17-59Þ þ k (Ti. dQ/dt. j ) þ k (Ti. j from the adjacent nodes and by energy generation within the medium. equation. j Ti. Dx. respectively.

heat transfer occurs to the cooler surface from nodes 1. their temperatures may be determined by proper application of equation (17-63). we have speciﬁed is the assumption that heat ﬂows in the x and y directions between nodes. We now solve for the heat transfer rate from the hotter surface. 2. T2 ¼ 141:67 K. and 3 to the cooler surface is written as k(T1 100) þ k(T2 100) þ k(T3 100) 2 145:83 100 þ (141:67 100) þ (120:83 100) ¼k 2 q¼ ¼ 85:415k (q in W/m. T2. and write T1 ¼ k(200 T1 ) þ k(200 T2 ) 2 200 145:83 þ (200 141:67) ¼k 2 q¼ ¼ 85:415k (q in W/m. Writing the proper expressions for T1. This is obviously a requirement of the analysis and serves as a check on the formulation and numerical work. we have 200 þ 100 þ 2T2 4 200 þ 100 þ T1 þ T3 T2 ¼ 4 100 þ 100 þ 2T2 T3 ¼ 4 This set of three equations and three unknowns may be solved quite easily to yield the following: T1 ¼ 145:83 K. k in W/mK) Observe that these two different means of solving for q yield identical results. k in W/mK) A similar accounting for the heat ﬂow from nodes 1. Implicit in the procedure of laying out a grid of the sort.17. of the heat transfer to and from node 1. The total heat transfer per meter of duct is calculated as q ¼ 8(8:415 K)(1:21 W/mK) ¼ 826:8 W/m . 2. and 3. only one half should be properly considered as part of the element analyzed.4 3m Two. We should also recall that the section of duct that has been analyzed is one-eighth of the total thus. and T3 using equation (17-68) as a guide. The example may now be concluded. On this basis heat transfer occurs from the hot surface to the interior only to nodes 1 and 2. T3 ¼ 120:83 K: The temperatures just obtained may now be used to ﬁnd heat transfer.and Three-Dimensional Systems 200 1 200 2 1m 1m 245 100 100 3m 1' 2' 3 100 100 The grid chosen is square with Dx ¼ Dy ¼ 1/2m: Three interior node points are thus identiﬁed.

the numerical approach to solving two-dimensional steady-state conduction problems. calculate the resulting percent error for values of ro /ri of 1. an a. which terms are relevant.2 The steady-state expression for heat conduction through a plane wall is q ¼ (kA/L)DT as given by equation (17-4). For steady-state heat conduction through a hollow cylinder. It is apparent that any added complexity in the form of more involved geometry. Numerical techniques may be used to solve complex problems involving nonuniform boundary conditions and variable physical properties.246 Chapter 17 Steady-State Conduction Example 4 has illustrated. occurs across a planewall having a constant thermal conductivity of 30 W/mK: The material is 30 cm thick. among others. If the arithmetic mean area. The deﬁning differential equations were frequently established by generating the equation through the use of the control-volume expression for the conservation of energy as well as by using the general differential equation for energy transfer. A 1. Show that A as deﬁned above satisﬁes the equations for steady-state radial heat transfer in a hollow cylindrical element.4 It is desired to transport liquid metal through a pipe embedded in a wall at a point where the temperature is 650 K. It is hoped that this approach will provide the student with an insight into the various terms contained in the general differential equation and thus enable one to decide.2-m-thick wall constructed of a material having a thermal conductivity varying with temperature according to . One-dimensional systems with and without internal generation of energy were considered. for each solution. Show a sketch of the temperature distribution for each case.3 Evaluate the appropriate ‘‘mean’’ area for steady-state heat conduction in a hollow sphere that satisﬁes an equation of the form q¼ kA DT ro ri Repeat part (b) of Problem 17. and 5. radiation. The analytical solution is recommended for problems of simple geometrical shapes and simple boundary conditions. Case T1 T2 1 2 3 4 350 K 300 K 275 K dT/dx (K/m) qx (W/m2 ) 2000 350 K 250 K 300 200 A ¼ 2p r o ri ln(ro /ri ) 17. determine the unknown quantities. PROBLEMS 17. 17. where A is the ‘‘log-mean’’ area deﬁned as 17. In this section.5 CLOSURE In this chapter. is used rather than the logarithmic mean. p(ro ri ). speciﬁed heat ﬂux. or simply a greater number of interior nodes. with no internal heat generation. Techniques for formulating such problems and some solution techniques are described by Welty. For each case listed in the table below. we have considered solutions to steady-state conduction problems.and three-dimensional steady-state conduction problems. 3. b. other types of boundary conditions such as convection. expression similar to equation (17-4) is q¼ kA DT r o ri 17.5.2 for the spherical case. we have considered techniques for solving two. in simple fashion. will render a problem too complex for hand calculation. Each of these approaches has certain requirements that limit their use.1 One-dimensional steady-state conduction.

8-cm layer of air. where T0 > TL : The wall material has a thermal conductivity that varies linearly according to k ¼ k0 (1 þ bT). The bus bar is supported by two plastic pedestals to which it is attached by an adhesive. respectively.5 for the case of a hollow cylinder with boundary conditions T ¼ T0 at r ¼ R0 and T ¼ TL at r ¼ R0 þ L: 17. is adjacent to the inside glass surface. of plastic (k ¼ 1:5 Btu/h ft F): Determine the thermal resistance of this wall if it is bolted together by 1/2-in.8. The outside surface is exposed to air at 300 K with a convective heat-transfer coefﬁcient of 23 W/m2 K: How far from the hot surface should the pipe be located? What is the heat ﬂux for the wall? 17. Toutside ho Tinside hi kfiberglass kplaster kwood 6 cm ¼ 10 C ¼ 20 W/m2 K ¼ 25 C ¼ 10 W/m2 K ¼ 0:035 W/m2 K ¼ 0:814 W/m2 K ¼ 0:15 W/m2 K Pine studs 30 cm 15 cm 2 cm Plaster Loose fiberglass insulation 17. there are two 3/4-in.6 Solve Problem 17.44 Btu/h ft 8F and a maximum usable temperature of 15008F and other with a k of 0.16 A copper bus bar measuring 5 cm by 10 cm by 2. Given the properties listed for the materials of construction.9 A furnace wall consisting of 0. The inside and outside wall temperatures are to be 2000 and 3008F. if they are made from two materials.14 A composite wall is to be constructed of 1/4 in. Bricks made of each material cost the same amount and may be laid in any manner. 0.94 Btu/h ft 8F and a maximum usable temperature of 2200 8F.9. At what point will the actual temperature proﬁle differ most from that which would exist in the case of constant thermal conductivity? 17. in addition to the conditions speciﬁed in Problem 17.8 A furnace wall is to be designed to transmit a maximum heat ﬂux of 200 Btu/h ft2 of wall area. one with a k of 0.15 A cross section of a typical home ceiling is depicted below. thickness of asbestos (k ¼ 0:15 Btu/h ft F) and then with a 1/ -in. what will be its equilibrium . k0 and b being constants. Under conditions where the extreme outside temperature of the glass is at 10 C and air. except that the outside temperature of the masonry brick cannot exceed 325 K. If 1 kW of energy is dissipated in the copper bar.10-m outer layer of masonry brick is exposed to furnace gas at 1370 K with air at 300 K adjacent to the outside wall. Determine the heat loss per square foot of wall and the temperature of the outside wall surface under these conditions. by how much must the thickness of kaolin be adjusted to satisfy this requirement? 17. b. stainless steel.-diameter bolts on 6-in. 17. 17. 3 in. The inside and outside convective heat transfer coefﬁcients are 115 and 23 W/m2 K. The convective heat-transfer coefﬁcient on the outside surfaces of both the plastic and aluminum is 12 W/m2 K. thickness of stainless steel (k ¼ 10 Btu/h ft F): If the 8 center temperature of this sandwich construction is considered constant at 10008F and the outside convective heat-transfer 247 coefﬁcient is 3 Btu/h ft2 F. how much energy must be supplied in W/ft2 to the heater? What will be the outside temperature of the stainless steel? 17. of stainless steel (k ¼ 10 Btu/h ft F). The pedestals are mounted on a wall whose temperature is 300 K.5-cm-thick sheet of plastic (k ¼ 2:42 W/mK) is to be bonded to a 5-cm-thick aluminum plate. respectively.25 m of ﬁre clay brick. respectively.5 m long is in a room in which the air is maintained at 300 K. and the heat to accomplish this bonding is to be provided by a radiant source. What is the required heat ﬂux if it is applied to the surface of (a) the plastic? (b) the aluminum? 17.Problems k ¼ 0:0073(1 þ 0:0054 T). determine how much heat is transferred through the insulation and through the studs.13 A 2.5 The temperature at the inner and outer surfaces of a plane wall of thickness L are held at the constant values T0 and TL. and a 0. the rate of heat transfer through the window unit. measuring 8 cm on a side. of corkboard (k ¼ 0:025 Btu/h ft F). b. the inside glass surface temperature. centers made of a. aluminum (k ¼ 120 Btu/h ft F): 17. 17.12 Determine the percent in heat ﬂux if. where T is in K and k is in W/mK.11 A heater composed of Nichrome wire wound back and forth and closely spaced is covered on both sides with 1/8-in. at 27 C. has its inside surface maintained at 925 K. The pedestals are square in cross section.20 m of kaolin. and the surrounding air is at 295 K.10 Given the conditions of Problem 17. Determine the most economical arrangement of bricks measuring 9 by 4 1/2 by 3 in.diameter steel bolts extending through the wall per square foot of wall area (k for steel ¼ 22 But/h ft F): 17. and 1/2 in. with hi ¼ 12 W/m2 K determine a. The unit measures 4 m in width and is 3 m wide. The glue that will accomplish the bonding is to be held at a temperature of 325 K to achieve the best adherence.7 A double-pane insulated window unit consists of two 1cm-thick pieces of glass separated by a 1. The air gap between glass panes may be treated as a purely conductive layer with k ¼ 0:0262 W/mK: Thermal radiation is to be neglected.

b. The maximum fuel temperature. Where. The fuel generates heat uniformly at a rate of 51:7 103 kJ/sm3 : The fuel is placed in an environment having a temperature of 360 K with a surface coefﬁcient of 4540 W/m2 K. End effects may be neglected. a.25 Radioactive waste (k ¼ 20 W/mK) is stored in a cylindrical stainless steel (k ¼ 15 W/mK) container with inner and outer diameters of 1. The fuel material has k ¼ 33:9 W/mK. Determine an expression for T(x) in terms of X. 11/2-in. while the surface at x ¼ 0 is maintained at a constant temperature of 320 K.24 A thin slab of material is subjected to microwave radiation that causes volumetric heating to vary according to 17. of 85% magnesia that costs $0.18 A 2-in. The heat loss per 10 ft of bare pipe. The resistivity of aluminum is 2:83 106 ohm-cm.20 A 10-kW heater using Nichrome wire is to be designed. The boundary at x ¼ L is perfectly insulated. respectively. 17.75 per foot. The convective heat-transfer coefﬁcient by condensing steam on the inside surface may be taken as 1500 Btu/h ft2 F: The surrounding air is at 80 F.21 if the ﬂuid surrounding the insulated wire was maintained at 70 F with a convective heat-transfer coefﬁcient between the insulation and the ﬂuid of 4 Btu/h ft2 F? What would be the surface temperature of the insulation under these conditions? 17.248 Chapter 17 Steady-State Conduction temperature? The convective heat-transfer coefﬁcient for all surfaces may be taken as 23 W/m2 K: The thermal conductivity of the plastic is 2:6 W/mK: Neglect thermal radiation. b. What size wire is required if the heater is to be in one piece 0. the steady-state temperature at the center of the waste material. 15 cm 17. The pipe is insulated with 1.6 m long? b. is 0. For this situation determine 17. The mass of steam condensed in 10 ft of bare pipe. What length of 14-gage wire is necessary to satisfy these design criteria? c. steel pipe. b. The temperature proﬁle as a function of radial position. will the maximum temperature occur? c.23 Work Problem 17. L. layer of material having a thermal conductivity of 0:14 Btu/h ft F: The outer surface of the insulation is maintained at 70 F: How much current may pass through the wire if the insulation temperature is limited to a maximum of 120 F? The resistivity of copper is 1:72 106 ohm-cm.19 Saturated steam at 40 psia ﬂows at 5 fps through a schedule-40.17 Solve Problem 17. is insulated with a 4-in.21 Copper wire having a diameter of 3/16 in. The heat loss per 10 ft of pipe insulated with 2 in. in the slab. 17. L. schedule-40 steel pipe carries saturated steam at 60 psi through a laboratory that is 60 ft long. c.2 m. The surface of the Nichrome is to be limited to a maximum temperature of 1650 K. What is the value of Tmax ? a.06 m.5 in. Thermal energy is generated uniformly within the waste material at a volumetric rate of 2 105 W/m3 : The outer container surface is exposed to water at 25 C. the steady-state temperatures at the inner and outer surfaces of the stainless steel. minimum convective heat-transfer coefﬁcient: 850 W/m2 K minimum temperature of the surrounding medium (air): 370 K The resistivity of Nichrome is 110 mV-cm and the power to the heater is available at 12 V. Other design criteria for the heater are a.9-cm steel bolt running through the center.16 if each plastic pedestal has a 1. and the outside surface coefﬁcient is 3 Btu/ft2 F: Determine the following: _ q(x) ¼ q_ o ½1 (x/L) where q_ o has a constant value of 180 kW/m3 and the slab thickness. k.6 W/m K. The surface temperature. c. and To : b. of 85% magnesia. q_ o . 17. with a surface coefﬁcient of 1000 W/m2 K: The ends of the cylindrical assembly are insulated so that all heat transfer occurs in the radial direction. How long must the steam line be in service to justify the insulation cost if the heating cost for the steam is $0:68 per 105 Btu? The outsidesurface convective heat-transfer coefﬁcient may be taken as 5 Btu/h ft2 F: 17. 17.77 cm in diameter.0 and 1. The thermal conductivity of the slab material is 0.26 A cylindrical nuclear fuel element is 10. How will the answers to parts (a) and (b) change if h ¼ 1150 W/m2 K? a.22 What would be the result of Problem 17. For the situation described evaluate the following at steady state: a.21 for the case of aluminum rather than copper. . 17.16 cm long and 10.

17.29 A 1-in. The latent heat of nitrogen is 200 kJ/kg.5 m. what will be its midpoint temperature? What will be the rate of heat transfer through both ends? 13 cm 17. . c.9 cm in diameter and 45 cm long. apart to the aluminum surface to aid in transferring heat.59 cm 17. which are at 60 F. which is at 300 K. which are at 480 K. its density.27? All values and dimensions in Problem 17. Plot the temperature proﬁle in the web if the convective heattransfer coefﬁcient between the steel surface and the surrounding air is 57 W/m2 K. 17. What conclusions may be reached regarding this result? 17. 17. respectively.30 Solve Problem 17. Evaluate the percent increase in heat transfer if these ﬁns are added to (a) the air side. If the convective heattransfer coefﬁcient is 17 W/m2 K. Assuming a negligible temperature change through both ﬂanges. a. (c) and both sides. 25 mm of insulation (k ¼ 0:002 W/mK) covers its outside surface. thickness and 3/4 in.33 A semiconductor material with k ¼ 2 W/mK and electrical resistivity. The surrounding air is 17.36 A 13 cm by 13 cm steel angle with the dimensions shown is attached to a wall with a surface temperature of 600 K. Plot the temperature proﬁle in the angle. long to the outside tube surface. and the convective heat-transfer coefﬁcient is 6 Btu/h ft2 F. a. It is proposed to add rectangular ﬁns 0. what will be the rate of liquid nitrogen boil-off? 17. The surrounding air is at 300 K. in the liquid phase. assuming a negligible temperature drop through the side of the angle attached to the wall.-OD steel tube has its outside wall surface maintained at 250 F.05 in. The air temperature is 300 K.27 Liquid nitrogen at 77 K is stored in an insulated spherical container that is vented to the atmosphere.34 An iron bar used for a chimney support is exposed to hot gases at 625 K with the associated convective heat-transfer coefﬁcient of 740 W/m2 K. b. is used to fabricate a cylindrical rod 40 mm long with a diameter of 10 mm. Determine the maximum temperature in the bar. The surrounding air is at 80 F.76 cm 28 cm 17.31 A cylindrical rod 3 cm in diameter is partially inserted into a furnace with one end exposed to the surrounding air. The container is made of a thin-walled material with an outside diameter of 0. Assuming the electrical resistivity of copper to be constant at 1:72 106 ohm-cm. respectively. Thickness = 1. r ¼ 2 105 V-m. 17. The temperatures measured at two points 7.37 A steel I-beam with a cross-sectional area as shown has its lower and upper surfaces maintained at 700 and 370 K.6 cm apart are 399 and 365 K.35 A copper rod 1/4 in. The bar is attached to two opposing chimney walls.38 Repeat Problem 17. develop an expression for the temperature variation in the web as a function of the distance from the upper ﬂange. b. respectively.29 if the convective heat-transfer coefﬁcient is increased to 60 Btu/h ft2 F by forcing air past the tube surface. Determine the heat loss from the sides of the angle projecting out from the wall. and the convective heat-transfer coefﬁcient is Btu/h ft2 F. The longitudinal surface of the rod is well insulated and may be considered adiabatic while the ends are maintained at temperatures of 100 and 0 C.27 apply.37 for the case of an aluminum beam. Compare the increase in heat transfer achieved by adding 12 longitudinal straight ﬁns or circular ﬁns with the same total surface area as the 12 longitudinal ﬁns.28 What additional thickness of insulation will be necessary to reduce the boil-off rate of liquid nitrogen to one-half of the rate corresponding to Problem 17. determine the thermal conductivity of the rod material. determine the maximum current the copper may carry if its temperature is to remain below 150 F. 249 at 60 F. It is proposed to increase the rate of heat transfer by adding ﬁns of 3/32-in. 17. The heat-transfer coefﬁcients on the air and water sides are 3 and 25 Btu/h ft2 F. respectively. If the rod carries a current of 10 amps.08 in. thick and 3/4 in. in diameter and 3 ft long runs between two bus bars.Problems 17. is 804 kg/m3. and the convective heat-transfer coefﬁcient between the angle surface and the air is 45 W/m2 K. What is the net heat transfer at the upper and lower ends of the web? y 0.32 Heat is to be transferred from water to air through an aluminum wall. The bar is 1. long spaced 0. For surroundings at 25 C and with a convective coefﬁcient of 18 W/m2 K at the outside surface of the insulation. (b) the water side.

they are 0. For these conditions. from the outside surface of the tube.46 Determine the heat ﬂow per foot for the conﬁguration shown. using the numerical procedure for a grid size of 1 1/2 ft. Using numerical methods. 50. The following speciﬁcations apply: h Tb (base) T1 (air) Fin base thickness. and 100 Btu/h ft2 F: Plot the percent increase in q vs. how much heat is transferred at each end of each rib? How far from the lower plate 17. of constant thickness. The surface temperatures at the inside and outside surfaces are 400 and 100 F. 17. at 280 K. The ﬁns are 1/16 in. L ¼ 60 W/m2 K ¼ 120 C ¼ 20 C ¼ 6 mm ¼ 20 mm Determine the ﬁn efﬁciency and heat loss per unit width for the ﬁnned surface.42 Repeat Problem 17. The space between the cylindrical surfaces is ﬁlled with rock wool (k ¼ 0:023 Btu/h ft F). The outside diameter of the engine cylinder is 0. How many ﬁns are required to cool a 3-kW engine.45 A cylindrical tunnel with a diameter of 2 m is dug in permafrost (k ¼ 0:341 W/m2 K) with its axis parallel to the permafrost surface at the depth of 2.5 m. If the outside surface of the tube wall is at 250 F. respectively.-OD pipe placed eccentrically inside a 6-in.44 Find the rate of heat transfer from a 3-in. What conclusions can be reached concerning this plot? is the rib temperature a minimum? What is this minimum value? T = 160°C T∞ = 25°C 8 mm 100 cm T = 400°C 17. from the axis of the large cylinder.-ID cylinder with the axis of the smaller pipe displaced 1 in. by 4. Determine the same information as in part (a) for values of h of 2. determine the percent increase in heat transfer for the ﬁnned pipe over that for the unﬁnned pipe. 17. section of building bring whose mean thermal conductivity may be taken as 0:38 Btu/h ft F: The convective heattransfer coefﬁcient between all surfaces and the surrounding air is 8 Btu/h ft2 F: The air temperature is 808F. respectively. The effective surface coefﬁcient between all surfaces and water is 300 W/m2 K. the surrounding air is at 80 F. 17. 5. and extend 2 cm from base to tip. to the permafrost surface at 220 K. determine a.41 A 2-in.41 for the case of an aluminum pipeand-ﬁn arrangement.-OD stainless-steel tube has 16 longitudinal ﬁns spaced around its outside surface as shown. The ribs that form the channels are also made of aluminum and are 8 mm thick. b.3 m thick.48 A 5-in.40 Heat from a ﬂat wall is to be enhanced by adding straight ﬁns. 17. 15. The material has a thermal conductivity of 0:15 Btu/h ft F: The inside and outside temperatures are at the uniform values of 200 and 0 F.250 Chapter 17 Steady-State Conduction 17.39 Circular ﬁns are employed around the cylinder of a lawn mower engine to dissipate heat. if 50% of the total heat given off is transferred by the ﬁns? 17. and the convective heat-transfer coefﬁcient is 8 Btu/h ft2 F. the total heat loss to the surrounding air. a. t Fin length. 17. using a grid size of 1 ft.47 Repeat the previous problem. . thick and extend 1 in. The ﬁns are made of aluminum. 9 ft 3 ft 3 ft 9 ft 17.3 m. h. operating at 30% thermal efﬁciency.375-in. Design operating conditions are T1 ¼ 30 C and h ¼ 12 W/m2 K: The maximum allowable cylinder temperature is 300 C: Estimate the amount of heat transfer from a single ﬁn.375-in. standard-steel angle is attached to a wall with a surface temperature of 600 F: The angle supports a 4. Determine the rate of heat loss from the cylinder walls. made of stainless steel.43 Water ﬂows in the channels between two aluminum plates as shown in the sketch.

17. which may be assumed to be at the steam temperature.Problems b. the location and value of the minimum temperature in the brick.2 m below the surface of the ground. The ground surface is at 280 K and the mean thermal conductivity of the soil is 0:66 W/mK.4-cm-OD pipe. how much steam will condense in a 50-ft length of pipe? 5 in. whose surface is at 100 F: If the space between the pipe and duct 1 ft 2 ft 17. 2 ft 5 in. The pipe is centered in the 2-ft-square duct. what is the heat loss per day from the pipe? . If the pipe surface is at 370 K. 145 cm long.49 Saturated steam at 400 F is transported through the 1-ft pipe shown in the ﬁgure.50 A 32. 251 is ﬁlled with powdered 85% magnesia insulation. is buried with its centerline 1.

in the derivation of this expression. in which the temperature at a given point varies with time. Examples of this second category would include metal stock or ingots undergoing heat treatment. In addition. In heating or cooling a conducting medium.1 ANALYTICAL SOLUTIONS The solution of an unsteady-state conduction problem is. these processes involve an unsteady-state ﬂux of energy. In this chapter. we shall consider problems and their solutions that deal with unsteady-state heat transfer within systems both with and without internal energy sources. will be considered in this chapter. The solution is approached by establishing the deﬁning differential equation and the boundary conditions. the rate of energy transfer is dependent upon both the internal and surface resistances. These design problems generally fall into two categories: the process that ultimately reaches steady-state conditions. or the thermal response of a thin laminate being bonded using a laser source. 252 . and the ﬂux of energy at a speciﬁc time can then be evaluated. Transient conduction processes are commonly encountered in engineering design. more difﬁcult than that for a steady-state problem because of the dependence of temperature on both time and position.Chapter 18 Unsteady-State Conduction Transient processes. the variation in the temperature distribution with time is established. the rate of internal generation. in general. It is repeated below for reference. and the process that is operated a relatively short time in a continually changing temperature environment. Lumped Parameter Analysis—Systems with Negligible Internal Resistance Equation (16-17) will be the starting point for transient conduction analysis. as well as the more general case in which both resistances are important. independent of position and time. By ﬁnding the solution to the partial differential equation that satisﬁes the initial and boundary conditions. however. 18. missile components during reentry into Earth’s atmosphere. Both of these cases will be considered. the limiting cases being represented either by negligible internal resistance or by negligible surface resistance. @T q_ ¼ a =2 T þ @t rcp (16-17) Recall that. thermal properties were taken to be _ can vary in both. q. As the transfer of energy is directly related to the temperature gradient. the initial temperature distribution in the conducting medium must be known.

may be expressed as ZZZ @ dT (18-3) er dV ¼ rVcp @t dt c:v: Equating these expressions as indicated by equation (18-1) we have. will show a temperature variation in the axial and radial directions as well as time. is not a signiﬁcant function of position. If a metallic specimen. The first of . The ratio. Because of this we will consider. initially with uniform temperature.1 Analytical Solutions 253 It is frequently the case that temperature within a medium varies signiﬁcantly in fewer than all three space variables. initially at uniform temperature T0 after it T∞ has been immersed in a hot oil bath at temperature T1 for a period of time t. V/A. A T(t) = T ﬁrst-law analysis using equation (6-10). with slight rearrangement dT hA(T1 T) ¼ (18-4) dt rVcp We may now obtain a solution for the temperature variation with time by solving equation (18-4) subject to the initial condition. is also seen to be a part of each of these new parametric forms. where the temperature of a body varies only with time. T(0) ¼ T0 . shape. is due to convection from the oil and is written as dQ ¼ hA(T1 T) (18-2) dt RRR The rate of energy increase within the specimen.18. applied to a spherical control volume coinciding with the specimen Figure 18. we have a spherical metallic specimen specimen. this case is the easiest of all to analyze. Spherical Shown in Figure 18. dQ/dt. is suddenly exposed to surroundings at a different temperature. If the cylinder has a length that is large compared to its diameter or. heated at one end with a ﬁxed boundary condition. @/@t c:v: er dV. having units of length. A circular cylinder. and thermal conductivity may combine in such a way that the temperature within the material varies with time only. with constant properties. that of a completely lumped-parameter system. as our ﬁrst transient conduction case.1 in question will reduce to ZZZ dQ @ ¼ er dV (18-1) dt @t c:v: The rate of heat addition to the control volume. it may be that size. if it is composed of a material with high thermal conductivity. A rearrangement of terms in the exponent may be accomplished as follows: 2 hAt hV A k hV/A at ¼ t ¼ (18-6) rcp V kA rV 2 c p k (V/A)2 Each of the bracketed terms in equation (18-6) is dimensionless. temperature will vary with axial position and time only.1. that is. These conditions are characteristic of a ‘‘lumped’’ system. and obtain T T1 ¼ ehAt/rcp V T0 T1 (18-5) The exponent is observed to be dimensionless. It is presumed that the temperature T(0) = T0 (uniform) of the metallic sphere is uniform at any given time.

01 0.4 0. 1. The lumped-parameter solution for transient conduction may now be written as T T1 ¼ eBiFo (18-9) T0 T1 Equation (18-9) is portrayed graphically in Figure 18. the conductive (internal) resistance to heat transfer.2. A large value of Bi indicates that the conductive resistance controls.2 T – T∞ T0 – T ∞ 254 0. to 1/h.02 0. The use of equation (18-9) is illustrated in the following example. A small value for Bi represents the case where internal resistance is negligibly small and there is more capacity to transfer heat by conduction than there is by convection.006 0.08 0. discussed at length earlier.008 0. lumpedparameter case.001 0 1 2 3 4 BiFo 5 6 7 Figure 18. there is more capacity for heat to leave the surface by convection than to reach it by conduction.2 Time-temperature history of a body at initial temperature. the Biot modulus is seen to be the ratio of (V/A)/k. that is. A commonly used rule of thumb is that the error inherent in a lumped-parameter analysis will be less than 5% for a value of Bi less than 0. In this latter case.06 0. The other bracketed term in equation (18-6) is the Fourier modulus. T0. The evaluation of the Biot modulus should thus be the ﬁrst thing done when analyzing an unsteady-state conduction situation. exposed to an environment at T1. the controlling heat transfer phenomenon is convection. and temperature gradients within the medium are quite small. An extremely small internal temperature gradient is the basic assumption in a lumpedparameter analysis.1. A natural conclusion to the foregoing discussion is that the magnitude of the Biot modulus is a reasonable measure of the likely accuracy of a lumped-parameter analysis. the convective (external) resistance to heat transfer.8 0. abbreviated Bi hV/A (18-7) Bi ¼ k By analogy with the concepts of thermal resistance.004 0. .0 0. The magnitude of Bi thus has some physical signiﬁcance in relating where the greater resistance to heat transfer occurs.6 0. abbreviated Fo.002 0.04 0.Chapter 18 Unsteady-State Conduction the new nondimensional parameters formed is the Biot modulus. where at Fo ¼ (18-8) (V/A)2 The Fourier modulus is frequently used as a nondimensional time parameter.1 0.

and its value is essentially equal to the ambient temperature. Ts is constant for all time.635 cm in diameter. is exposed to an air stream at a temperature. consider a large ﬂat plate of uniform thickness L. of 310 K. we can apply equation () to yield h¼ rc p V T0 T1 ln tA T T1 (8890 kg/m3 )(385 J/kg K) ¼ pD2 L 4pDL (30 s) ln 280 310 297 310 ¼ 51 W/m2 K This result is much less than the limiting value of h indicating that a lumped-parameter solution is probably very accurate. The initial temperature distribution through the plate will be assumed to be an arbitrary function of x. which is the limiting value of Bi for a lumped-parameter analysis to be valid. t > 0.1 EXAMPLE 1 Analytical Solutions 255 A long copper wire. the temperature of the surface. After 30 s the average temperature of the wire increased from 280 to 297 K. h. In order to determine if equation () is valid for this problem. The Biot number is expressed as pD2 L 0:00635 m h hV/A 4pDL 4 ¼ Bi ¼ ¼ 386 W/m K 386 W/m K k ¼ 4:11 106 h h Setting Bi ¼ 0. which is a near certainty.1. we obtain h ¼ 0:1/4:11 106 ¼ 24300 W/m2 K We may conclude that a lumped-parameter solution is valid if h < 24. that is. 300 W/m2 K. For this limiting case. T1 . estimate the average surface conductance. The solution for the temperature history must satisfy the Fourier ﬁeld equation @T ¼ a=2 T @t (16-18) @T @2T ¼a 2 @t @x (18-10) For one-directional energy flow with initial and boundary conditions T ¼ T0 (x) T ¼ Ts at t ¼ 0 at x ¼ 0 for 0 . To illustrate the analytical method of solving this class of transient heat-conduction problems. and solving for h. 0. T1 . A second class of time-dependent energy-transfer processes is encountered when the surface resistance is small relative to the overall resistance. Bi is 0:1. the value of Bi must be determined.18. Heating a Body Under Conditions of Negligible Surface Resistance. Proceeding. Using this information.

x .

L for t > 0 .

let Y ¼ (T Ts )/(T0 Ts ). the partial differential equation may be rewritten in terms of the new temperature variable as @Y @2Y ¼a 2 @t @x (18-11) and the initial and boundary conditions become Y ¼ Y0 (x) at t ¼ 0 Y ¼0 at x ¼ 0 for 0 .256 Chapter 18 Unsteady-State Conduction and T ¼ Ts at x ¼ L for t > 0 For convenience. where T0 is an arbitrarily chosen reference temperature.

x .

. as well as the central temperature history in other solids. at any plane in the conducting medium may be evaluated by qx ¼ kA @T @x (18-14) In the case of the infinite flat plate with an initial uniform temperature distribution of T0. . the heat rate at any time t is np 1 2 kA (Ts T0 ) å cos x e(np/2) Fo n ¼ 1. The central temperature history for the plane wall. T0 Ts p n¼1 n L The temperature history at the center of the infinite plane. . (18-15) qx ¼ 4 L L n¼1 . 3. in ‘‘Heissler charts. Y0 (x). Y0 (x) ¼ Y0 . The heat rate.3. is illustrated in Figure 18. L for t > 0 and Y ¼ 0 at x ¼ L for t > 0 Solving equation (18-11) by the method of separation of variables leads to product solutions of the form Y ¼ (C1 cos lx þ C2 sin lx)eal 2 t The constants C1 and C2 and the parameter l are obtained by applying the initial and boundary conditions. 5. q. and sphere is presented in Appendix F. Equation (18-12) points out the necessity for knowing the initial temperature distribution in the conducting medium. The complete solution is Z L np 2 2 1 np (np=2) Fo Y0 (x) sin e x dx Y ¼ å sin L n¼1 L L 0 (18-12) where Fo ¼ at/(L/2)2 . With this temperature distribution. Consider the special case in which the conducting body has a uniform initial temperature. infinite cylinder.’’ These charts cover a much greater range in the Fourier modulus than Figure 18. equation (18-12) reduces to np 2 T Ts 4 1 1 (18-13) ¼ å sin x eðnp/2Þ Fo n ¼ 1. 5. .3. . before the complete temperature history may be evaluated. . 3.

the Biot number is evaluated as 2 pD L h h(V/A) h(DL/4) 4 Bi ¼ ¼ ¼ k k(L þ D/2) pD2 k pDL þ 2 ¼ (8500 W/m2 K)(0:1 m)(0:1 m)/4 1:21 W/m K(0:1 þ 0:1/2) m ¼ 117 . Reading Mass. Ts (From P.4 0.. AddisonWesley Publishing Co. the ﬁnite cylinder.1 Analytical Solutions 257 1. 249. Schneider. T0.01 x1 x1 0.) In the following example. of 8500 W/m2 K.1 m in length and 0.0 0..6 at/x21 0.0 x1 x1 (T – Ts)/(T0 – Ts) 0.2 Figure 18. Conduction Heat Transfer.. Determine the time required for the center of this stubby cylinder to reach 310 K. Inc.8 1.3 Central temperature history of various solids with initial uniform temperature. If the cylinder were sufﬁciently long so that it could be considered inﬁnite.1 x1 x1 2x1 0. is initially at room temperature.001 0. how long would it take? For the ﬁrst case. It is suspended in a steam environment where water vapor at 373 K condenses on all surfaces with an effective ﬁlm coefﬁcient. and constant surface temperature. 0. p. 1955. h. J.2 0.18. EXAMPLE 2 A concrete cylinder. the use of the central temperature-history figure will be illustrated. 292 K. By permission of the publishers.0 1.1 m in diameter.

and convective boundary conditions T ¼ T0 @T ¼0 @x at t ¼ 0 at the centerline of the body and @T h ¼ (T T1 ) at the surface @x k . The second line from the bottom in this ﬁgure applies to a cylinder with height equal to diameter. the temperature history must satisfy the initial.11. The ordinate value of 0. as in this case. which may be expressed for one-dimensional heat ﬂow by @T @2T ¼a 2 @t @x (18-7) A case of considerable practical interest is one in which a body having a uniform temperature is placed in a new ﬂuid environment with its surfaces suddenly and simultaneously exposed to the ﬂuid at temperature T1. symmetry. The most general cases of transient heat-conduction processes involve signiﬁcant values of internal and surface resistances.3 may be used. is t¼ 0:13(0:05 m)2 ¼ 546 s 5:95 107 m2 /s ¼ 9:1 min Heating a Body with Finite Surface and Internal Resistances.3 will again be used. t ¼ 0:11 (0:05 m)2 ¼ 462 s 5:95 107 m2 /s ¼ 7:7 min In the case of an inﬁnitely long cylinder. for the abscissa.258 Chapter 18 Unsteady-State Conduction For this large value. The Biot number in this case is 2 pD L D h h h(V/A) 4 Bi ¼ ¼ 4 ¼ k(pDL) k k 2 (8500 W/m K)(0:1 m)/4 ¼ 176 ¼ 1:21 W/m K which is even larger than the ﬁnite cylinder case. The solution for the temperature history without internal generation must satisfy the Fourier ﬁeld equation. The required time. the fourth line from the bottom applies. in this case.778 yields. The ordinate is T Ts 310 373 ¼ ¼ 0:778 T0 Ts 292 373 and the corresponding abscissa value is approximately 0. a value of approximately 0. In this case. Figure 18.13. The time required may now be determined as at ¼ 0:11 x21 Thus. Figure 18.

Ingersoll. Heat Transfer to a Semi-Inﬁnite Wall. Carslaw and J. O. among which are the Laplace transformation and the Fourier transformation. Consider the situation illustrated in Figure 18.4 Temperature distribution in a semi-inﬁnite wall at time t. If we reconsider the inﬁnite ﬂat plate of thickness. J. The complex nature of equation (18-17) has led to a number of graphical solutions for the case of one-dimensional transient conduction. New York. Solutions to this case of time-dependent energy-transfer processes have been obtained for many geometries. hx1 /k. We shall use an alternative procedure. h. S. T1. R. where Ts > T0. C. L. when inserted into a medium at constant temperature. Zobel. The variables in equation (18-10) 1 H. The resulting plots. 2 . x/x1. (18-10) and the initial and boundary conditions are T ¼ T0 at t ¼ 0 for all x T ¼ Ts at x ¼ 0 for all t and T ! T0 as x ! 1 for all t The solution to this problem may be accomplished in a variety of ways. A large plane wall initially at a constant temperature T0 is subjected to a surface temperature Ts. C. and the relative distance. and A. and Ingersoll2. which is less involved mathematically. An analytical solution to the one-dimensional heat-conduction equation for the case of the semi-inﬁnite wall has some utility as a limiting case in engineering computations.18. are discussed in Section 18. which results in product solutions as previously encountered when only the internal resistance was involved. 2x1.2. 1948. The differential equation to be solved is @T @2T ¼a 2 @t @x Ts T0 x Figure 18. Conduction of Heat in Solids. Excellent treatises discussing these solutions have been written by Carslaw and Jaeger1 and by Ingersoll. Ingersoll. 1947. but now include a constant surface conductance. Oxford University Press. the following solution is obtained 1 sin d cos (d x/x ) T T1 n n 1 d2n Fo ¼2 å e T0 T1 n¼1 dn þ sin dn cos dn (18-16) where dn is deﬁned by the relation dn tan dn ¼ hx1 k (18-17) The temperature history for this relatively simple geometrical shape is a function of three dimensionless quantities: at/x21 . Zobel. with dimensionless temperature as a function of other dimensionless parameters as listed above.4. Heat Conduction (With Engineering and Geological Applications).1 Analytical Solutions 259 One method of solution for this class of problems involves separation of variables. McGraw-Hill Book Company. Jaeger.

Thus. we obtain x Y ¼ 1 erf pﬃﬃﬃﬃﬃ 2 at or or T T0 x ¼ 1 erf pﬃﬃﬃﬃﬃ Ts T0 2 at Ts T x ¼ erf pﬃﬃﬃﬃﬃ T s T0 2 at (18-20) . we may write T T0 x at ¼ f . designated ‘‘erf. in this problem there is no ﬁnite characteristic dimension. or with equal validity.’’ where Z f 2 2 erf f pﬃﬃﬃ eh dh p 0 and erf (0) ¼ 0. x1. the error function. Equation (18-18) may be integrated once to yield ln dY ¼ c1 h 2 dh or 2 dY ¼ c2 eh dh and integrated once more to yield Z Y ¼ c3 þ c2 eh dh 2 (18-19) The integral is related to a frequently encountered form. x1 x21 T s T0 However. Applying the boundary conditions to equation (18-19). If h ¼ x/2 at is selected as the independent variable and the dependent variable Y ¼ (T T0 )/ (Ts T0 ) is used. and thus pﬃﬃﬃﬃﬃ (T T0 )/ (Ts p ﬃﬃﬃﬃTﬃ 0 ) ¼ f (at/x2 ). A short table of erf f is given in Appendix L. and the boundary condition T ! T0 as x ! 1. (T T0 )/(Ts T0 ) ¼ f (x/ at). erf (1) ¼ 1. substitution into equation (18-10) yields the ordinary differential equation d2 Y/dh2 þ 2h dY/dh ¼ 0 (18-18) with the transformed boundary and initial conditions Y !0 as h ! 1 and Y ¼ 1 at h ¼ 0 The ﬁrst condition above is the same as the initial condition T ¼ T0 at t ¼ 0.260 Chapter 18 Unsteady-State Conduction may be expressed in dimensionless form by analogy with the previous case.

T1 . Solutions are presented in Appendix F for the ﬂat plate. . the equations describing temperature proﬁles have been solved3 and the results have been presented in a wide variety of charts to facilitate their use. Until the temperature change at x ¼ L exceeds some nominal amount.. relative position m. The surface temperature is particularly easy to obtain from the above equation. For the case of ﬁnite surface resistance. sphere. constant thermal diffusivity and no internal heat source. The value of L/(2 ﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ at) corresponding to a 0.5% change in (T T0 )/(Ts T0 ) is L/(2 at) ’ 2. 3 Equation (18-16) pertains to a plane wall of thickness. (b) The conducting medium has a uniform initial temperature. and boundary conditions T(x. so for L/(2 at) > 2.2 Temperature–Time Charts for Simple Geometric Shapes 261 This equation is extremely simple to use and quite valuable. if we let x ¼ 0. the solution for the ﬁnite and inﬁnite walls will be thepsame.2 TEMPERATURE–TIME CHARTS FOR SIMPLE GEOMETRIC SHAPES For unsteady-state energy transfer in several simple shapes with certain restrictive boundary conditions. and the heat transfer rate may be determined from q ¼ h(Ts T1 ) A 18. equation (18-20) may be used for ﬁnite geometry with little or no error.e. say (T T0 )/(Ts T0 ) equal pﬃﬃﬃﬃto ﬃ 0. unaccomplished temperature change ¼ X.5%. relative resistance where x1 is the radius or semithickness of the conducting medium. Consider a ﬁnite wall of thickness L subject to the surface temperature Ts. and long cylinder in terms of four dimensionless ratios: T1 T T1 T0 at ¼ 2 x1 x ¼ x1 k ¼ hx1 Y. the solution to equation (18-10) for a semiinﬁnite wall is pﬃﬃﬃﬃﬃ T1 T x hx h2 at h at x þ 2 þ pﬃﬃﬃﬃﬃ ¼ erf pﬃﬃﬃﬃﬃ þ exp (18-21) 1 erf T1 T0 k k k 2 at 2 at This equation may be used to determine the temperature distribution in finite bodies for small times in the same manner as equation (18-20). Two forms of these charts are available in Appendix F. T0.18. for t 0. These charts may be used to evaluate temperature proﬁles for cases involving transport of energy into or out of the conducting medium if the following conditions are met: (a) Fourier’s ﬁeld equation describes the process. (c) The temperature of the boundary or the adjacent ﬂuid is changed to a new value. relative time and n. 0) ¼ T0 and dT/dx(0. i. L. t) ¼ 0.

we ﬁnd the dimensionless temperature. This is equivalent to heat transfer from a 1-m-thick wall. 0. is 0. The relative position.1 in Appendix F. is equal to (1200 600)/(1200 200). k/hx1. will be 0. the desired temperature can be evaluated by Ts T 1200 T ¼ ¼ 0:74 Ts T0 1200 200 or T ¼ 460 K (368 F) . the relative time.305. the abscissa. Y ¼ (T1 T)/(T1 T0 ). has one of its faces suddenly exposed to a hot gas at 1200 K. Y. EXAMPLE 3 A ﬂat wall of ﬁre-clay brick.35 under these conditions. For transport in a cylinder. From Figure F. is 1.125/[(7. The relative position.7. is 1/2. the insulated face. position.6. x/x1. and Yb is evaluated with thickness x1 ¼ b.5)] or 0.262 Chapter 18 Unsteady-State Conduction For ﬂat plates where the transport takes place from only one of the faces. Using these values and Figure F. 2. and thickness x1 ¼ a. The dimensionless temperature. where x is then measured from the line of symmetry. x/x1. For transport in a rectangular parallelepiped Yparallelepiped ¼ Ya Yb Yc (18-23) where Ya is evaluated with width x1 ¼ a. including both ends Y cylinder ¼ Ycylinder Ya (18-24) plus ends where Ya is evaluated by using the flat-plate chart. or 0. 3. they may be combined to yield solutions for two. From the table of physical properties given in Appendix H. Yb is evaluated with thickness x1 ¼ b. The following is a summary of these combined solutions: 1. (b) the temperature of the insulated wall face at the time evaluated in (a). and Yc is evaluated with depth x1 ¼ c. Using this value. The relative resistance. The time required to raise the centerline to 6008F is t¼ 0:35x21 0:35(0:5)2 ¼ 1:651 105 s ¼ a 5:30 107 or 45:9 h The relative resistance and the relative time for (b) will be the same as in part (a). determine (a) the time necessary to raise the center of the wall to 600 K.5 m thick and originally at 200 K. at/x21 . If the heat-transfer coefﬁcient on the hot side is 7:38 W/m2 K and the other face of the wall is insulated so that no heat passes out of that face. to be 0. The use of temperature–time charts is demonstrated in the following examples.74. Although the charts were drawn for one-dimensional transport.38)(0. the following values are listed: k ¼ 1:125 W/m K c p ¼ 919 J/kg K r ¼ 2310 kg/m3 and a ¼ 5:30 107 m2 /s The insulated face limits the energy transfer into the conducting medium to only one direction. and resistance are evaluated as if the thickness were twice the true value. For transport in a rectangular bar with insulated ends Ybar ¼ Ya Yb (18-22) where Ya is evaluated with width x1 ¼ a. in Appendix F.and three-dimensional problems.

An initial temperature distribution may be nonuniform in nature.3 EXAMPLE 4 Numerical Methods for Transient Conduction Analysis 263 A billet of steel 30.18. actual initial and/or boundary conditions do not correspond to those mentioned earlier with regard to analytical solutions. if the surface conductance is 34 W/m2 K determine the center temperature of the billet after 1 h.3 NUMERICAL METHODS FOR TRANSIENT CONDUCTION ANALYSIS In many time-dependent or unsteady-state conduction processes. the following average physical properties will be used: k ¼ 49:2 W/m K c p ¼ 473 J/kg K r ¼ 7820 kg/m3 a ¼ 1:16 105 m2 /s Equation (18-22) applies. or system geometry may be variable or quite irregular.8 in Appendix F. For the cylindrical surface the appropriate values are X¼ at (1:16 105 m2 /s)(3600 s) ¼ x21 (0:1525 m)2 ¼ 1:80 n¼ x ¼0 x1 m ¼ k/hx1 ¼ 42:9/(34)(0:1525) ¼ 8:27 and. ambient temperature. for heat transfer across the cylindrical surface and both ends TCL T1 Yjtotal ¼ ¼ Ya YCL ¼ (0:95)(0:7) ¼ 0:665 T0 T1 The desired center temperature is now calculated as TCL ¼ T1 þ 0:665(T0 T1 ) ¼ 310 K þ 0:665(645 310) K ¼ 533 K (499 F) 18. numerical techniques offer the best means to achieve solutions. from Figure F. surface conductance. initially at 645 K.5 cm in diameter. the corresponding values of dimensionless temperature. For such complex cases.7 in Appendix F. is immersed in an oil bath that is maintained at 310 K. From Appendix H. we obtain YCL T T1 ¼ ﬃ 0:7 T0 T1 cyl Now. To evaluate Ya the following dimensionless parameters apply X¼ at (1:16 105 m2 /s)(3600 s) ¼ ¼ 0:449 x21 (0:305 m)2 n ¼ x/x1 ¼ 0 m ¼ k/hx1 ¼ 42:9 W/m K ¼ 4:14 (34 W/m2 K)(0:305 m) Using these values with Figure F. 61 cm long. Ya.95. is approximately 0. .

however. have an effect on the solution. The absence of variation in the y direction allows several terms to be deleted. quite obviously. to the space increment. numerical solutions are being obtained for heat transfer problems of all types. Digital computers obviously provide the only feasible way to accomplish solutions. For large numbers of nodes. it can be evaluated quite easily.264 Chapter 18 Unsteady-State Conduction More recently. The time-dependent term on the right of equation (18-25) is written such that the temperature at node i. a form resembling the Fourier modulus. The magnitude of this grouping will. Dx. This means of evaluating Tij at the end of a time increment is designated an ‘‘explicit’’ technique. For a space increment Dx. it is clear that a great number of calculations are needed and that much information must be stored for use in subsequent computation. Some numerical work is introduced in Chapter 17. we will consider variation in time as well as position. we obtain ! a Dt 2a Dt (Tiþ1 jt þ Ti1 jt ) þ 1 (18-27) Ti jt Ti jtþDt ¼ (Dx)2 ðDxÞ2 The ratio. it can be extended easily to three dimensions. equation (17-61) reduces to k Dy Dx (Ti1. A more thorough discussion of explicit solutions is given by Carnahan.4 Equation (18-25) may be solved to evaluate the temperature at node i. In this section. steady-state conduction. j1 þ Ti. We next consider properties to be constant and represent the ratio k=rc p as a. j þ Tiþ1. We will next consider the one-dimensional form of equation (18-25). Solving for Ti jtþDt . and this trend will doubtlessly continue. the reader is referred to equation (17-61) and the development leading up to it. j jtþDt rc p Ti. j 2Ti. this equation can then be solved to ﬁnd Tij at the end of time interval Dt. To begin our discussion. Dt. As Ti jjtþDt appears only once in this equation. It is likely that many users of this book will be involved in code development for such analysis. j ) þ k (Ti. j ) Dx Dy rc p Ti. jþ1 2Ti. For the case of no internal generation of energy. It has been determined that equation (18-27) is numerically ‘‘stable’’ when a Dt 1 . the simpliﬁed expression becomes rc p Ti jtþDt rc p Ti jt k Dx (18-26) (Ti1 jt þ Tiþ1 jt 2Ti jt ) ¼ Dt Dx where the j notation has been dropped. j for all values of i. is seen to arise naturally in this development. This grouping relates the time step. j that comprise the region of interest. j is presumed known at time t. dealing with two-dimensional. a Dt/(Dx)2 . with sophisticated computing codes available. j jt Dx Dy ¼ Dt (18-25) This expression applies to two dimensions.

Carnahan. Wilkes. Luther. Applied Numerical Methods. A. . 1969. H.4 4 B. and J. Wiley. New York. O. (18-28) 2 2 (Dx) For a discussion of numerical stability the reader is referred to Carnahan et al.

Either analytical or numerical methods must therefore be employed. i.5. having width Dx. If the limiting value is chosen. The illustration below depicts the wall divided into 10 increments.. equation (18-27) becomes TijtþDt ¼ EXAMPLE 5 Tiþ1jt þ Ti1jt 2 (18-29) A brick wall (a ¼ 4:72 107 m2 /s) with a thickness of 0.tþ1 ¼ 2 2 The problem solution now proceeds as equations (18-29) and (18-30) and are solved at succeeding times to update nodal temperatures until the desired result. The simplest approach is thus numerical and we will proceed with the ideas introduced in this section.5 m is initially at a uniform temperature of 300 K.t 425 þ T2. This relationship includes the dimensionless ratio. to the space increment Dx. the location of node 4.t ¼ 2 2 (18-30) T8. can have any value equal to or less than 0. however. This same idea prevails for all 11 nodes. the solution is in terms of inﬁnite series involving eigenvalues and the determination of a ﬁnal answer is cumbersome. However. this one-dimensional problem is not amenable to a solution using the charts because there is no axis of symmetry. a small time step will likely be used without major difﬁculty.tþ1 ¼ Ti1. aDt/Dx2 . T1.t þ 600 ¼ T9. this includes the surface nodes. The quantity. the algorithms for nodes 1 and 9 can be written as T0. ∆x 0 1 2 3 4 5 6 7 8 9 10 An energy balance for any internal node. Dx. equation (18-27) reduces to a simple algorithmic form Ti. An analytical solution using Laplace transform or separation-of-variables methodology is relatively straightforward. and computation time-a solution will be achieved more rapidly for larger values of Dt. When computing is done by machine.e. aDt/Dx2 . 0 and 10.tþ1 ¼ . Dt. For the case with a Dt/(Dx2 ) ¼ 1/2. and maintained at 425 and 600 K. An examination of equation (18-27) indicates considerable simpliﬁcation to be achieved if the equality in equation (18-28) is used.t þ T10. will yield equation (18-27) as a result. Ts ¼ 425 K.05 m. Each of the nodes within the wall is at the center of a subvolume. Determine the length of time required for its center temperature to reach 425 K if its surfaces are raised to.18. having a width.3 Numerical Methods for Transient Conduction Analysis 265 The choice of a time step involves a trade-off between solution accuracy-a smaller time step will produce greater accuracy. is achieved. Although relatively simple. as nodes 0 and 10 are at constant temperature for all time.t þ Tiþ1. which relates the time increment. respectively. In this example we have speciﬁed Dx ¼ 0. The shaded subvolume at node 4 is considered to have uniform properties averaged at its center. which is the limit for a stable solution.t 2 (18-29) This expression is valid for i ¼ 1 to 9.t þ T2.t T8.

5 402.5 A portion of a semi-inﬁnite wall used in integral analysis.1 405.1 526.8 408. The table below summarizes the form of the results for Ti.8 t ¼ 23 425 416.. originally at uniform temperature T0.4 AN INTEGRAL METHOD FOR ONE-DIMENSIONAL UNSTEADY CONDUCTION The von Ka´rma´n momentum integral approach to the hydrodynamic boundary layer has a counterpart in conduction... exposed to a ﬂuid at temperature T1 .0 419. 425 300 300 300 300 300 300 t ¼ 10 . . In this case.3 t ¼ 22 425 414.1 506. an interpolated value is n ¼ 22. As discussed earlier.8 512.0 428.t.0 600 411. a spreadsheet approach could also be used.3 445. with the surface of the wall at any time at temperature Ts.3 600 411.4 403. 425 394.2 600 The desired center temperature is reached between time increments 22 and 23.3 478.5 shows a portion of a semi-inﬁnite wall.8 372.2 445.6 time increments.5 550. .4 347.1 349.3 402.7 466.2 463. T∞ L Ts d x T0 Figure 18.2 408. the increment Dt is related to a and Dx according to the ratio a Dt ¼ 1/2 Dx2 or Dt ¼ 1 Dx2 1 (0:05 m)2 ¼ 2 a 2 4:72 107 m2 /s ¼ 2648 s ¼ 0:736 h The answer for total time elapsed is thus t ¼ 22:6(0:736) ¼ 16:6 h 18.3 556.4 600 409.5 409.4 t ¼ 20 . T0 T1 T2 T3 T4 T5 T6 t¼0 .4 436.4 512. 425 411.3 553.266 Chapter 18 Unsteady-State Conduction Equations (18-29) and (18-30) are quite simple and easily programed to achieve a solution. .3 395. Figure 18.9 346. .8 T7 T8 T9 T10 300 300 300 600 367.3 471.

giving qx d ¼ A dt Z d Z rc p T dx þ 0 L rcp T0 dx d and. since T0 is constant. we have ZZ ZZZ dQ dWs dWm P @ ¼ r(v n) dA þ eþ er dV (6-10) dt dt dt r @t c:s: c:v: with dWs dWm ¼ ¼ dt dt ZZ eþ c:s: P r(v n) dA ¼ 0 r The applicable form of the ﬁrst law is now ZZZ dQ @ ¼ er dV dt @t c:v: Considering all variables to be functions of x alone. Assuming the temperature proﬁle to be parabolic of the form T ¼ A þ Bx þ Cx2 and requiring that the following boundary conditions: T ¼ Ts T ¼ T0 at x ¼ 0 at x ¼ d . d) is assumed.’’ designated as d. At distance d. heat transfer from the ﬂuid to the wall affects the temperature proﬁle within the wall. Case 1.18.6.4 An Integral Method for One-Dimensional Unsteady Conduction 267 At any time t. The solution of equation (18-33) is subject to three different boundary conditions at the wall. is taken as zero. has its surface maintained at temperature Ts for t > 0. to a control volume extending from x ¼ 0 to x ¼ L. initially at uniform temperature T0. x ¼ 0. The temperature proﬁle at two different times is illustrated in Figure 18. in the sections to follow. this becomes Z d qx d ¼ rc p T dx þ rc p T0 (L d) A dt 0 The integral equation to be solved is now Z qx d d dd ¼ rc p T dx rc p T0 A dt 0 dt (18-33) If a temperature proﬁle of the form T ¼ T(x. and one may use this result to express the temperature proﬁle as T(x. t). equation (6-10). Constant wall temperature The wall. which may be solved. equation (18-33) will produce a differential equation in d(t). the temperature gradient. Applying the ﬁrst law of thermodynamics. where L > d. @T/@x. we may express the heat ﬂux as Z Z qx d L d L ¼ ru dx ¼ rc p T dx (18-32) A dt 0 dt 0 The interval from 0 to L will now be divided into two increments. is the distance from the surface wherein this effect is manifested. The ‘‘penetration distance.

yielding Z k d d x2 dd 2 (Ts T0 ) ¼ rc p T0 þ (Ts T0 ) 1 dx rc p T0 d dt 0 d dt and.268 Chapter 18 Unsteady-State Conduction T Ts t2 t1 d (t1) d (t2) x Figure 18. we see that the expression for T(x) becomes T T0 x2 ¼ 1 d Ts T0 The heat flux at the wall may now be evaluated as qx @T k ¼ k ¼ 2 (Ts T0 ) @x x¼0 d A (18-34) (18-35) which may be substituted into the integral expression along with equation (18-33). we have Z a d d x2 dd (18-36) 2 (Ts T0 ) ¼ T0 þ (Ts T0 ) 1 dx T0 d dt 0 d dt After integration. and @T ¼0 @x at x ¼ d be satisﬁed. after dividing through by rcp. equation (18-36) becomes 2a d d (Ts T0 ) ¼ (Ts T0 ) d dt 3 and cancelling (Ts T0 ). we obtain 6a ¼ d and thus the penetration depth becomes d¼ dd dt pﬃﬃﬃﬃﬃﬃﬃﬃﬃ 12at (18-37) (18-38) The corresponding temperature profile may be obtained from equation (18-34) as 2 T T0 x ¼ 1 pﬃﬃﬃ pﬃﬃﬃﬃﬃ (18-39) T s T0 3(2 at) . both quantities being considered constant.6 Temperature proﬁles at two times after the surface temperature is raised to Ts.

18.2 Exact 0 0 0.4 An Integral Method for One-Dimensional Unsteady Conduction 269 0. which compares reasonably well with the exact result T T0 x ¼ 1 erf pﬃﬃﬃﬃﬃ Ts T0 2 at Figure 18.0 1. when substituted into equation (18-38). (18-40) Case 2.0 2.6 0.5 1.7 A comparison of exact and approximate results for one-dimensional conduction with a constant wall temperature. yields d F(t)d2 aF(t) ¼ dt k 6k and 1/2 pﬃﬃﬃﬃﬃﬃ 1 Z t d(t) ¼ 6a F(t) dt F(t) 0 T T0 ¼ For a constant heat flux of magnitude q0/A the resulting expression for Ts is rﬃﬃﬃﬃﬃﬃﬃﬃ q0 3 at Ts T0 ¼ Ak 2 which differs by approximately 8% from the exact expression 1:13q0 pﬃﬃﬃﬃﬃ Ts T0 ¼ at Ak (18-41) (18-42) (18-43) (18-44) (18-45) . A speciﬁed heat ﬂux at the wall In this case the appropriate boundary conditions are and T ¼ T0 @T ¼0 @x at x ¼ d at x ¼ d @T F(t) ¼ at x ¼ 0 @x k where the heat ﬂux at the wall is expressed as the general function F(t). If the parabolic temperature proﬁle is used. the above boundary conditions yield ½F(t)(d x)2 2kd which.4 Approximate 0.5 x/2 at Figure 18.8 (T – T0)/(Ts – T0) 0.5 2.7 shows a comparison of these two results.

Convection at the surface The wall temperature is a variable in this case. whose surfaces are suddenly exposed to surroundings at a different temperature. some of the techniques for solving transient or unsteady-state heatconduction problems have been presented and discussed. with a uniform initial temperature.270 Chapter 18 Unsteady-State Conduction Case 3. however. The similarity between the penetration depth and the boundary-layer thickness from the integral analysis of Chapter 12 should also be noted. and those for which both resistances were signiﬁcant. . upon substituting equation (18-47) Ts T0 ¼ hd (T1 T0 ) Nk (18-48) or T0 þ (hd/Nk)T1 1 þ hd/Nk (18-49) T s T0 hd/Nk ¼ T1 T0 1 þ hd/Nk (18-50) T1 Ts 1 ¼ T1 T0 1 þ hd/Nk (18-51) Ts ¼ We may now write and The appropriate substitutions into the integral equation and subsequent solution follow the same procedures as in cases (a) and (b). Numerical and integral methods were also introduced. and spheres.5 CLOSURE In this chapter. additional boundary conditions are needed in such cases to evaluate the constants. charts are available for evaluating the temperature at any position and time. cylinders. At the surface we may write q @T ¼ k ¼ h(T1 Ts ) A x¼0 @x x¼0 which becomes. Situations considered included cases of negligible internal resistance. Temperature proﬁle expressions more complex than a parabolic form may be assumed. it may be easily determined. If the temperature variation within the medium is expressed generally as x T T0 ¼f (18-46) d Ts T0 we note that the temperature gradient at the surface becomes @T T s T0 N ¼ @x x¼0 d (18-47) where N is a constant depending upon the form of f(x/d). For ﬂat slabs. However. 18. The student should recognize the marked utility of the integral solution for solving onedimensional unsteady-state conduction problems. negligible surface resistance. the details of this solution are left as a student exercise.

At what minimum velocity must the billet travel through the furnace to satisfy these conditions? 18. The fuse material melts at 9008C.5 Cast-iron cannonballs used in the War of 1812 were occasionally heated for some extended time so that. with associated convective heattransfer coefﬁcient of 4 Btu/h ft2 8F. k ¼ 0:087 Btu/h ft F? 18.9 Water. and it must be raised to a minimum temperature of 15008F before working. The heat-transfer coefﬁcient between the furnace gases and the billet surface is 15 Btu/h ft2 8F.? (c) 12 in. Pertinent properties of the fuse material are Resistance ¼ 0:2 V k ¼ 20 W/m K a ¼ 5 105 m2 /s 18. is immersed in an oil bath that is at 258C.45 m wide. how long will it take for the rubber surface to reach 320 K? The dimensions of the block are 0. If the surroundings are at a temperature of 808F. at 308C. having a diameter of 0. d. 18. As a representative case.? (d) 24 in.10-m-diameter orange. 18.6 It is known that oranges can be exposed to freezing temperatures for short periods of time without sustaining serious damage.10 A short aluminum cylinder 0.1 mm. determine the steady state temperature of the wire.1 A household iron has a stainless-steel sole plate that weights 3 lb and has a surface area of 0. how long will it take for the surface of the orange to reach 08C? Properties of the orange are the following: r ¼ 940 kg/m3 k ¼ 0:47 W/m K c p ¼ 3:8 kJ/kg K: 18. The initial billet temperature is 2008F. The iron is rated at 500 W.Problems 271 PROBLEMS 18. is initially at a uniform temperature of 708F. in diameter. k ¼ 60 Btu/h ft F? mild steel.5 cm and diameters of 0. surrounds a fuse with a surface coefﬁcient of 10 W/m2 K. consider a 0. is initially at the uniform temperature of 708F. k ¼ 130 Btu/h ft F? zinc.? (e) 5 ft? 18. For conditions where an electric current of 100 A is ﬂowing through the wire and the surface coefﬁcient between the wire and oil bath is 550 W/ m2 K. The block sits on one of the 0.6 m high by 0. is contained within a thinwalled cylindrical vessel having a diameter of 18 in. Assume that the water is well stirred and that the convective heat-transfer coefﬁcient between the oil and cylindrical surface is 40 Btu/h ft2 8F. estimate the time it will take for the fuse to blow after a current of 3 A ﬂows through it.11 A type-304 stainless-steel billet.0 W/m2 K.? (b) 6 in. of 15 W/m2 K.794 mm. if the height of the cylinder is (a) 3 in.4 If a rectangular block of rubber (see Problem 18. The effective heat-transfer coefﬁcient at all exposed surface is 6. is passing through a 20-ft-long heat-treating furnace. How long. will it take for the wire to reach a temperature within 58C of its steady-state value? 18. For a surface coefﬁcient. The cylinder is immersed to a depth of 2 ft. It is suddenly exposed to a convective environment at 345 K with h ¼ 85 W/m2 K. after the current is supplied. k ¼ 212 Btu/h ft F? aluminum.3 m long by 0.0572 V/m. Aluminum wire of this size has an electrical resistance of 0. Plot the temperature of the water vs. how long after being exposed to air at 08F with an outside convective heat-transfer coefﬁcient of 16 But/h ft2 8F. between the air and orange surface. originally at a uniform temperature of 58C. initially at 408F. Air. If one of these the socalled ‘‘hot shot’’ were at a uniform temperature of 20008F. c. the ‘‘vulcanization’’ process requires that a tire carcass. k ¼ 10:5 Btu/h ft F? asbestos.2 An electrical system employs fuses that are cylindrical in shape and have lengths of 0.6 m long is initially at 475 K.5 ft2.7 A copper cylinder with a diameter of 3 in. k ¼ 25 Btu/h ft F? stainless steel.12 In the curing of rubber tires. would be required for the surface temperature to drop to 6008F? What would be the center temperature at this time? The ball diameter is 6 in. and the convective heat-transfer coefﬁcient between the sole plate and surroundings is 3 Btu/h ft2 8F. Assuming all heat transfer to be from the fuse surface. the adjacent surface may be considered an insulator.6 m in diameter and 0. be . Determine the temperature in the cylinder at a radial position of 10 cm and a distance of 10 cm from one end of the cylinder after being exposed to this environment for 1 h. when ﬁred at houses or ships. b. originally at 295 K.12 for properties) is set out in air at 297 K to cool after being heated to a uniform temperature of 420 K.3-m by 0. copper. e. How long after the cylinder is placed in a medium at 10008F.8 A cylinder 2 ft high with a diameter of 3 in. and the furnace gases are at 23008F. The following properties of cast iron may be used: k ¼ 23 Btu/h ft F c p ¼ 0:10 Btu/lbm F r ¼ 460 lbm /ft3 : 18. 6 in. How long after being placed in a medium at 10008F with an associated convective heat-transfer coefﬁcient of 4 Btu/h ft2 8F will the temperature at the center of the cylinder reach 5008F.45-m bases. they would set them aﬁre. will the center temperature reach 5008F if the cylinder is made from a. What will the maximum temperature within the rubber block be at this time? 18. time up to 1 h if the water and container are immersed in an oil bath at a constant temperature of 3008F.3 Aluminum wire. f. suddenly exposed to surrounding air at 58C. how long will it take for the iron to reach 2408F after it is plugged in? 18.

14 It is common practice to treat wooden telephone poles with tar-like materials to prevent damage by water and insects. placed in a pressurized oven. For an oven temperature of 3808C and h ¼ 140 W/m2 K.5 ft and is 10 ft long. exposed to boiling water at 1008C. Determine the temperature at a depth of 0. c p ¼ 3:35 kJ/kg K. how long after exposure to the ﬁre will the wood surface reach its combustion temperature of 4008C? Pertinent data are the following: h ¼ 30 W/m2 K Ti (initial) ¼ 21 C For oak: r ¼ 545 kg/m3 k ¼ 0:17 W/m K cp ¼ 2:385 kJ/kg K 18. having a thermal diffusivity of 5:16 107 m2 /s. If the meat is initially at 58C and the oven temperature is 1908C. The convective heat-transfer coefﬁcient between the nozzle and the gases. from the surface at this time? The nozzle is initially at 08F. The bar measures 0. 18. having its length and diameter equal to each other. every portion of the meat should attain a minimum temperature of 958C.26 The convective heat-transfer coefﬁcient between a large brick wall and air at 1008F is expressed as h ¼ 0:44 . is 200 Btu/h ft2 8F. with a surface coefﬁcient of 15 W/m2 K.25 m after a period of 5 h has elapsed at this surface condition.2 in. 18. for a 3-cm-thick tire carcass to reach the speciﬁed central temperature condition. and k ¼ 0:5 W/m K. For the hot dog initially at 58C. It will be removed when the tar has penetrated to a depth of 10 cm.3-m-diameter pole. If the interior wall surface is made of oak.13 Buckshot.8107 years for the Earth’s age: a ¼ 0:0456 ft2 /h T2 ¼ 0 F @T ¼ 0:02 F/ft. This heating is accomplished by introducing steam at 435 K to both sides. Properties of the wooden pole are k ¼ 0:20 W/m K a ¼ 1:1 107 m2 /s 18. which are at 30008F.21 A rocket-engine nozzle is coated with a ceramic material having the following properties: k ¼ 1:73 Btu/h ft F. Properties of rubber that may be used are the following: k ¼ 0:151 W/m K. with properties being those of water.25 kg. If the edges are suddenly all reduced to 1008F. Properly cooked. is quenched in 908F oil from an initial temperature of 4008F. (measured) @y y¼0 Comment on Lord Kelvin’s result by considering the exact expression for unsteady-state conduction in one dimension T Ts x ¼ erf pﬃﬃﬃﬃﬃ T0 Ts 2 at 18. 18. with a surface coefﬁcient of 90 W/m2 K.24 Determine an expression for the depth below the surface of a semi-inﬁnite solid at which the rate of cooling is maximum. 0. Lord Kelvin obtained an estimate of 9. respectively. 18.15. Substitute the information given in Problem 18. determine the time required for the pole to remain in the oven.4 for the case when air is blown by the surfaces of the rubber block with an effective surface coefﬁcient of 230 W/m2 K.22 to estimate how far below Earth’s surface this maximum cooling rate is achieved. The roast is to be modeled as a cylinder. determine the time required for the center of the cylinder to reach 530 K if end effects are neglected.2 ft by 0. 18.18 Consider a hot dog to have the following dimensions and properties: diameter ¼ 20 mm. Determine the time required.8 W/m2 K. 18. 18. in diameter. where H and D are the height and diameter of the cylinder. construct a plot of the time for the midpoint temperature to reach 530 K as a function of H/D. If the convective heat-transfer coefﬁcient between the lead and oil is 40 Btu/h ft2 8F.25 Soil.272 Chapter 18 Unsteady-State Conduction heated so that its central layer reaches a minimum temperature of 410 K. r ¼ 1201 kg/m3 . what will be the temperature of the shot as it reaches the bottom of the bath? 18. a ¼ 6:19 108 m2 /s. How long after startup will it take for the temperature at the ceramic surface to reach 27008F? What will be the temperature at a point 1/2 in. The roast weights 2. These tars are cured into the wood at elevated temperatures and pressures.15 For an asbestos cylinder with both height and diameter of 13 cm initially at a uniform temperature of 295 K placed in a medium at 810 K with an associated convective heat-transfer coefﬁcient of 22. cp ¼ 200 J/kg K.16 A copper bus bar is initially at 4008F. The buckshot is made of lead and takes 15 s to fall from the oil surface to the bottom of the quenching bath. what is the minimum cooking time required? 18.17 Rework Problem 18. a ¼ 0:35 ft2 /h. has its surface temperature suddenly raised and maintained at 1100 K from its initial uniform value of 280 K. It is known that a 10-cm depth of penetration will occur when a temperature of 1008C is achieved. how long will it take for the center to reach a temperature of 2508F? 18. after introducing steam. r ¼ 880 kg/m3 . what will be the cooking time if the required condition is for the center temperature to reach 808C? 18.19 This problem involves using heat transfer principles as a guide for cooking a pork roast. Using this value and the following properties for Earth’s crust. originally at 258C.23 After a ﬁre starts in a room the walls are exposed to combustion products at 9508C. Consider the case of a 0.22 One estimate of the original temperature of Earth is 70008F.20 Given the cylinder in Problem 18.

If the bar is considered semi-inﬁnite.34 A brick wall (a ¼ 0. 273 18. 6. and the convective heattransfer coefﬁcient between the soil and the surrounding air is 1. If the glass is initially at 308F.02 ft2/h. its thermal conductivity is 0. Its upper surface is suddenly exposed to an air stream at 2008C.35 A masonry brick wall 0. is suddenly exposed to combustion exhaust at 8008C. show that the penetration depth d for a semi-inﬁnite solid is of the form "R #1/2 pﬃﬃﬃ 0t F(t)dt d ¼ (constant) a F(t) 18. 18. how long will it take for a pipe buried 10 ft below the surface to reach 328F if the outside air temperature is suddenly dropped to 08F.32 If the heat ﬂux into a solid is given as F(t). If the wall is initially at a uniform temperature of 10008F. when the surface coefﬁcient between the wall and combustion gas is 20 W/m2 K. and 24 h.28 Air at 658F is blown against a pane of glass 1/8 in. 18. How long. the surface. 18. and has frost on the outside.5 Btu/h ft2 8F.33 If the temperature proﬁle through the ground is linear.31 A thick plate made of stainless steel is initially at a uniform temperature of 3008C. respectively.29 How long will a 1-ft-thick concrete wall subject to a surface temperature of 15008F on one side maintain the other side below 1308F? The wall is initially at 708F. after the wall surfaces are raised to 300 and 6008F. how long will it take for the temperature at a distance of 50 mm from the surface to reach 1008C? 18.Problems (T T1 )1/3 Btu/h ft2 F. estimate the temperature of the surface after 1. Determine the time of exposure required for the surface to reach its ignition temperature of 4008C. What will the surface temperature be at this time? . Evaluate the temperature after 3 min of elapsed time at a. initially at a uniform temperature of 258C. with a corresponding convective coefﬁcient of 22 W/m2 K.27 A thick wall of oak. estimate the length of time required for the frost to begin to melt. b. 18. increasing from 358F at the surface by 0. thick.58F per foot of depth.8 Btu/h ft 8F. Work this problem both analytically and numerically.45 m thick has a temperature distribution at time. t ¼ 0 which may be approximated by the expression T(K) ¼ 520 þ 330 sin p(x/L) where L is the wall width and x is the distance from either surface. The thermal diffusivity of soil may be taken as 0.016 ft2/h) with a thickness of 1½ ft is initially at a uniform temperature of 808F. 18.30 A stainless-steel bar is initially at a temperature of 258C. a depth of 50 mm. The surface is suddenly exposed to a coolant at 208C with a convective surface coefﬁcient of 110 W/m2 K. How long after both surfaces of this wall are exposed to air at 280 K will the center temperature of the wall be 360 K? The convective coefﬁcient at both surface of the wall may be taken as 14 W/m2 K. will it take for the temperature at the center of the wall to reach 3008F? 18.

Chapter 19 Convective Heat Transfer Heat transfer by convection is associated with energy exchange between a surface and an adjacent ﬂuid. and the heat transfer rate is increased. Thus. The distinction between laminar and turbulent ﬂow will thus be a major consideration in any convective situation. molecular energy exchange or conduction effects will always be present. on the contrary. then all energy transfer between a surface and contacting ﬂuid or between adjacent ﬂuid layers is by molecular means. and the geometry of the speciﬁc system of interest. It is related to the mechanism of ﬂuid ﬂow. and a thin layer of ﬂuid close to the surface will be in laminar ﬂow regardless of the nature of the free stream. 19. If. then there is bulk mixing of ﬂuid particles between regions at different temperatures. the associated density change and buoyant effect produce a natural circulation in which the affected ﬂuid moves of its own accord past the solid surface. we may expect many of the considerations from the momentum transfer to be of interest. however. The rate equation for convection has been expressed previously as q ¼ h DT (15-11) A where the heat ﬂux. the properties of the ﬂuid. the ﬂuid that replaces it is similarly affected by the energy transfer. ﬂow is turbulent. If ﬂuid ﬂow is laminar. In the analyses to follow. not at all a simple undertaking. q/A. Forced convection is the classiﬁcation used to describe those convection situations in which ﬂuid circulation is produced by an external agency such as a fan or a pump. and the process is repeated. but will now be considered in some depth. This simple equation is the deﬁning relation for h. The determination of the coefﬁcient h is. the convective heat-transfer coefﬁcient. and play a major role in any convection process. the ﬂuid particles immediately adjacent to a solid boundary are stationary. much use will be made of the developments and concepts of Chapters 4 through 14. When a ﬂuid is heated or cooled. There are very few energy-transfer situations of practical importance in which ﬂuid motion is not in some way involved. 274 . These have to do with the driving force causing ﬂuid to ﬂow. There are two main classiﬁcations of convective heat transfer.1 FUNDAMENTAL CONSIDERATIONS IN CONVECTIVE HEAT TRANSFER As mentioned in Chapter 12. This effect has been eliminated as much as possible in the preceding chapters. Natural or free convection designates the type of process wherein ﬂuid motion results from the heat transfer. occurs by virtue of a temperature difference. In light of the intimate involvement between the convective heat-transfer coefﬁcient and ﬂuid motion.

The molecular diffusivities of momentum and energy have been deﬁned previously as m momentum diffusivity : n r and k thermal diffusivity : a rcp That these two are designated similarly would indicate that they must also play similar roles in their speciﬁc transfer modes. as we shall see several times in the developments to follow. thus their ratio must be dimensionless. This ratio. Several of the new parameters to be encountered in energy transfer will arise in such a manner that their physical meaning is unclear. These are as follows: (a) dimensional analysis. 19. The Prandtl number is primarily a function of temperature and is tabulated in Appendix I. This is indeed the case. n mcp (19-1) Pr ¼ a k The Prandtl number is observed to be a combination of fluid properties. plays a major role in convective heat transfer.1. There are four methods of evaluating the convective heat-transfer coefﬁcient that will be discussed in this book. (c) approximate integral analysis of the boundary layer. Some parameters of this type have been encountered earlier. thus Pr itself may be thought of as a property.2 SIGNIFICANT PARAMETERS IN CONVECTIVE HEAT TRANSFER Certain parameters will be found useful in the correlation of convective data and in the functional relations for the convective heat-transfer coefﬁcients. which to be useful requires experimental results.19. is designated the Prandtl number. that of the molecular diffusivity of momentum to the molecular diffusivity of heat. For such a case the heat-transfer rate between the surface and the ﬂuid may be written as qy ¼ hA(Ts T1 ) (19-2) and. as one would expect. Additionally. these include the Reynolds and the Euler numbers.2 Signiﬁcant Parameters in Convective Heat Transfer 275 The hydrodynamic boundary layer. (b) exact analysis of the boundary layer. because heat transfer at the surface is by conduction qy ¼ kA @ (T Ts )jy¼0 @y (19-3) . For this reason. and (d) analogy between energy and momentum transfer. The temperature proﬁle that exists is due to the energy exchange resulting from this temperature difference. In the ﬁgure. the surface is at a higher temperature than the ﬂuid. we shall deﬁne and analyze the thermal boundary layer. which will also be vital to the analysis of a convective energytransfer process. The temperature proﬁle for a ﬂuid ﬂowing past a surface is depicted in Figure 19. we shall devote a short section to the physical interpretation of two such terms. analyzed in Chapter 12. those of L2 /t. For the moment we should note that both have the same dimensions. at various temperatures for each fluid listed.

Q. It is necessary to include two more dimensions—Q. with a temperature difference existing between the ﬂuid and the tube wall. will be encountered many times in the work to follow. h. thus all variables must be expressed dimensionally as some combination of M. The speciﬁc forced-convection situation. which was the case in the evaluation of the Biot modulus. and T. L. and dimensional representations are listed below. It may be considered a ratio of conductive thermal resistance to the convective thermal resistance of the fluid. The left-hand side of this equation is written in a manner similar to that for the Biot modulus encountered in Chapter 18. which we shall now consider. heat. Multiplying both sides by a representative length. thermal and ﬂow properties of the ﬂuid. thus h(Ts T1 ) ¼ k @ (T Ts )jy¼0 @y which may be rearranged to give h @(Ts T)/@yjy¼0 ¼ k T s T1 (19-4) Equation (19-4) may be made dimensionless if a length parameter is introduced. The important variables.1 Temperature and velocity proﬁles for a ﬂuid ﬂowing past a heated plate. v. These two parameters. is that of ﬂuid ﬂowing in a closed conduit at some average velocity. L. 19.276 Chapter 19 Convective Heat Transfer v∞ Ts – T∞ vx Ts – T y Figure 19. t. Ts These two terms must be equal. The above variables include terms descriptive of the system geometry. we have hL @(Ts T)/@yjy¼0 ¼ k (Ts T1 )/L (19-5) The right-hand side of equation (19-5) is now the ratio of the temperature gradient at the surface to an overall or reference temperature gradient. their symbols. Pr and Nu.3 DIMENSIONAL ANALYSIS OF CONVECTIVE ENERGY TRANSFER Forced Convection. This ratio is referred to as the Nusselt number Nu hL k (19-6) where the thermal conductivity is that of the fluid as opposed to that of the solid. and T. and the quantity of primary interest. temperature—to the fundamental group considered in Chapter 11. .

3 Variable Tube diameter Fluid density Fluid viscosity Fluid heat capacity Fluid thermal conductivity Velocity Heat-transfer coefﬁcient Dimensional Analysis of Convective Energy Transfer Symbol Dimensions D r m cp k v h L M/L3 M/Lt Q/MT Q/tLT L/t Q/tL2T 277 Utilizing the Buckingham method of grouping the variables as presented in Chapter 11. Solving for p2 and p3 in the same way will give p2 ¼ mc p ¼ Pr and k p3 ¼ hD ¼ Nu k . Note that the rank of the dimensional matrix is 4. and v as the four variables comprising the core.19. k. we have for L: 0 ¼ a b c þ d 3 Q: 0 ¼ b t : 0 ¼ b c d T: 0 ¼ b and M: 0 ¼ c þ 1 Solving these equations for the four unknowns yields a¼1 b¼0 c ¼ 1 d¼1 and p1 becomes p1 ¼ Dvr m which is the Reynolds number. Choosing D. the required number of dimensionless groups is found to be 3. we ﬁnd that the three p groups to be formed are p1 ¼ Da kb mc v d r p2 ¼ De kfmg v h cp and p3 ¼ Di k j mk v l h Writing p1 in dimensional form 1 ¼ ðLÞ a Q LtT b c d M L M Lt t L3 and equating the exponents of the fundamental dimensions on both sides of this equation. one more than the total number of fundamental dimensions. m.

The coefﬁcient of thermal expansion. They may be found by considering the relation for buoyant force in terms of the density difference due to the energy exchange. and h/rvc p : The first two of these we recognize as Re and Pr. r is the fluid density inside the heated layer. An alternative correlating relation for forced convection in a closed conduit is thus St ¼ f2 (Re. cp.278 Chapter 19 Convective Heat Transfer The result of a dimensional analysis of forced-convection heat transfer in a circular conduit indicates that a possible relation correlating the important variables is of the form Nu ¼ f1 (Re. g. and DT is the temperature difference between the heated fluid and the bulk value. The list of variables for the problem under consideration is given below. St h rvcp (19-8) This parameter could also have been formed by taking the ratio Nu/ðRe PrÞ. and DT into the list of those important to the natural convection situation. The buoyant force per unit volume. is given by r ¼ r0 (1 b DT ) (19-10) where r0 is the bulk fluid density. mc p /k. Fbuoyant. The velocity no longer belongs in the group of variables. the core group had been chosen to include r. is Fbuoyant ¼ (r0 r)g which becomes. the analysis would have yielded the groups Dvr/m. In the case of natural-convection heat transfer from a vertical plane wall to an adjacent ﬂuid. Pr) (19-9) Natural Convection. Pr) (19-7) If. the variables will differ signiﬁcantly from those used in the preceding case. The third is the Stanton number. as it is a result of other effects associated with the energy transfer. upon substituting equation (19-10) Fbuoyant ¼ bgr0 DT (19-11) Equation (19-11) suggests the inclusion of the variables b. New variables to be included in the analysis are those accounting for ﬂuid circulation. b. Variable Signiﬁcant length Fluid density Fluid viscosity Fluid heat capacity Fluid thermal conductivity Fluid coefﬁcient of thermal expansion Gravitational acceleration Temperature difference Heat-transfer coefﬁcient Symbol Dimensions L r m cp k b g DT h L M/L3 M/Lt Q/MT Q/LtT 1/T L/t2 T Q/L2tT . m. and v. in the preceding case.

incompressible continuity equation @v x @v y þ ¼0 @x @y (12-10) . 19. Pr) (19-13) or (b) Natural convection The similarity between the correlations of equations (19-7) and (19-13) is apparent. used in correlating natural-convection data. Gr has replaced Re in the correlation indicated by equation (19-7). is (bgr2 L3 DT )/m2 . m. This parameter. This becomes obvious when we observe the velocity. g. Choosing L. The boundary-layer equations considered previously include the two-dimensional. Pr) (19-9) Nu ¼ f3 (Gr.4 Exact Analysis of the Laminar Boundary Layer 279 The Buckingham p theorem indicates that the number of independent dimensionless parameters applicable to this problem is 9 5 ¼ 4. k. In equation (19-13).5. contained in the expression for St. we see that the p groups to be formed are p1 ¼ La mb kc b d ge cp p2 ¼ Lf mg kh bi g j r p3 ¼ Lk ml km bn go DT and p4 ¼ Lp mq kr bs gt h Solving for the exponents in the usual way. we obtain mc p ¼ Pr p3 ¼ b DT p1 ¼ k L3 gr2 hL ¼ Nu and p4 ¼ p2 ¼ k m2 The product of p2 and p3 . v. which must be dimensionless. Blasius’s solution for the laminar boundary layer on a ﬂat plate may be extended to include the convective heat-transfer problem for the same geometry and laminar ﬂow. Pr) (19-7) St ¼ f2 (Re. Gr bgr2 L3 DT m2 (19-12) From the preceding brief dimensional-analyses considerations.19.4 EXACT ANALYSIS OF THE LAMINAR BOUNDARY LAYER An exact solution for a special case of the hydrodynamic boundary layer is discussed in Section 12. is the Grashof number. we have obtained the following possible forms for correlating convection data: (a) Forced convection Nu ¼ f1 (Re. and b as the core group. It should be noted that the Stanton number can be used only in correlating forced-convection data.

isobaric ﬂow the energy equation that applies is now vx @T @T @2T þ vy ¼a 2 @x @y @y (19-15) From Chapter 12. Ts In steady. two-dimensional. @ 2 T/@x2 is much smaller in magnitude than @ 2 T/@y2 : T∞ Edge of thermal boundary layer x T = T(y) y Figure 19. The proper form of the energy equation will thus be equation (16-14). The solution was based upon the boundary conditions vy vx ¼ ¼0 v1 v1 at y ¼ 0 and vx ¼1 v1 at y ¼ 1 . written in two-dimensional form as 2 @T @T @T @ T @2T þ vx þ vy ¼a þ (19-14) @t @x @y @x2 @y2 With respect to the thermal boundary layer depicted in Figure 19. the applicable equation of motion with uniform free-stream velocity is vx @v x @v x @2vx þ vy ¼n 2 @x @y @y (12-11a) and the continuity equation @v x @v y þ ¼0 @x @y (12-11b) The latter two of the above equations were originally solved by Blasius to give the results discussed in Chapter 12. for isobaric ﬂow. incompressible.2.2 The thermal boundary layer for laminar ﬂow past a ﬂat surface.280 Chapter 19 Convective Heat Transfer and the equation of motion in the x direction @v x @v x @v x dv 1 @2vx þ vx þ vy ¼ v1 þn 2 @t @x @y dx @y (12-9) Recall that the y-directional equation of motion gave the result of constant pressure through the boundary layer.

4 Exact Analysis of the Laminar Boundary Layer 281 The similarity in form between equations (19-15) and (12-11a) is obvious. Using the nomenclature of Chapter 12. This is a consequence of having Pr ¼ 1: A logical consequence of this situation is that the hydrodynamic and thermal boundary layers are of equal thickness. we may now write the results obtained by Blasius for the energy-transfer case. (2) The boundary conditions for temperature must be compatible with those for the velocity. This may be accomplished by changing the dependent variable from T to (T Ts )/(T1 Ts ). In order that this be possible. the following conditions must be satisﬁed: (1) The coefﬁcients of the second-order terms must be equal.19. It is signiﬁcant that the Prandtl numbers for most gases are sufﬁciently close to unity that the hydrodynamic and thermal boundary layers are of similar extent. the dimensionless velocity proﬁle in the laminar boundary layer is identical with the dimensionless temperature proﬁle. This situation suggests the possibility of applying the Blasius solution to the energy equation. vx T Ts f0 ¼ 2 ¼2 (19-16) v1 T1 Ts rﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃﬃ y v1 y xv 1 y pﬃﬃﬃﬃﬃﬃﬃﬃ h¼ Rex (19-17) ¼ ¼ 2 nx 2x 2x n and applying the Blasius result. This requires that n ¼ a or that Pr ¼ 1. We may now obtain the temperature gradient at the surface @T 0:332 1/2 Re ¼ (T T ) (19-19) 1 s y¼0 x @y x Application of the Newton and Fourier rate equations now yields qy @T ¼ hx (Ts T1 ) ¼ k @y y¼0 A from which k @T hx ¼ Ts T1 @y ¼ y¼0 0:332k 1/2 Rex x (19-20) . we obtain df 0 d½2(v x /v 1 ) 00 pﬃﬃﬃﬃﬃﬃﬃﬃ ¼ f (0) ¼ dh y¼0 d½(y/2x) Rex y¼0 df2½(T Ts )/(T1 Ts )g pﬃﬃﬃﬃﬃﬃﬃﬃ ¼ d½(y/2x) Rex ¼ 1:328 ð19-18Þ y¼0 It should be noted that according to equation (19-16). The boundary conditions now are vy vx T Ts ¼ ¼ ¼0 v 1 v 1 T1 Ts vx T Ts ¼ ¼1 v 1 T1 Ts at y ¼ 0 at y ¼ 1 Imposing these conditions upon the set of equations (19-15) and (12-11a).

ZAMM.3. when used with the Fourier and Newton rate equations. 1 E. 115 (1921). A plot of the dimensionless temperature vs.328 (y/2x) Rex Figure 19. h Pr1/3 is shown in Figure 19.3 Temperature variation for laminar ﬂow over a ﬂat plate.Chapter 19 Convective Heat Transfer or hx x ¼ Nux ¼ 0:332 Re1x /2 k (19-21) Pohlhausen1 considered the same problem with the additional effect of a Prandtl number other than unity. . The temperature variation given in this form leads to an expression for the convective heat-transfer coefficient similar to equation (19-20). yields k hx ¼ 0:332 Rex1/2 Pr1/3 x (19-24) hx x ¼ Nux ¼ 0:332 Rex1/2 Pr1/3 k (19-25) or The inclusion of the factor Pr1/3 in these equations extends the range of application of equations (19-20) and (19-21) to situations in which the Prandtl number differs considerably from 1. 1. Pr1/3 which. Pohlhausen. He was able to show the relation between the thermal and hydrodynamic boundary layers in laminar ﬂow to be approximately given by d ¼ Pr1/3 dt (19-22) The additional factor of Pr1/3 multiplied by h allows the solution to the thermal boundary layer to be extended to Pr values other than unity. the gradient is @T 0:332 1/2 1/3 Rex Pr ¼ (T1 Ts ) @y y¼0 x f' = 2 (T – Ts) / (T∞ – Ts) 282 (19-23) Slope = 1. At y ¼ 0.

4. Consider the control volume designated by the dashed lines in Figure 19.19. For a plate of these dimensions Z qy ¼ hA(Ts T1 ) ¼ hx (Ts T1 ) dA A 1/ 3 h(wL)(Ts T1 ) ¼ 0:332kw Pr Z L (Ts T1 ) 1/2 Z 0 Rex1/2 dx x L v1r x1/2 dx m 0 1/2 1/3 v 1 r L1/2 ¼ 0:664k Pr m ¼ 0:664k Pr1/3 Re1/2 hL ¼ 0:332k Pr1/3 L The mean Nusselt number becomes NuL ¼ hL ¼ 0:664 Pr1/3 Re1L/2 k (19-26) and it is seen that NuL ¼ 2 Nux at x ¼ L (19-27) In applying the results of the foregoing analysis it is customary to evaluate all ﬂuid properties at the ﬁlm temperature.5 APPROXIMATE INTEGRAL ANALYSIS OF THE THERMAL BOUNDARY LAYER The application of the Blasius solution to the thermal boundary layer in Section 19. For ﬂow other than laminar or for a conﬁguration other than a ﬂat surface. . another method must be utilized to estimate the convective heattransfer coefﬁcient.4 was convenient although very limited in scope. applying to ﬂow parallel to a ﬂat surface with no pressure gradient. having width Dx.4 Control volume for integral energy analysis.5 Approximate Integral Analysis of the Thermal Boundary Layer 283 The mean heat-transfer coefﬁcient applying over a plate of width w and length L may be obtained by integration. An approximate method for analysis of the thermal boundary layer employs the integral analysis as used by von Ka´rma´n for the hydrodynamic boundary layer. 19. a height equal to the q3 q1 q2 ∆x y x q4 δth Figure 19. which is deﬁned as Tf ¼ Ts þ T1 2 (19-28) the arithmetic mean between the wall and bulk fluid temperatures. This approach is discussed in Chapter 12.

and an average value of cp is used. This contrasts slightly with the momentum integral solution in which the velocity proﬁle alone was assumed. An application of the ﬁrst law of thermodynamics in integral form ZZZ ZZ dQ dWs dWm @ ¼ (e þ P/r)r(v: n) dA þ er dV (6-10) dt @t dt dt c:s: c:v: yields the following under steady-state conditions: dQ @T ¼ k Dx dt @y y¼0 dWs dWm ¼ ¼0 dt dt ZZ Z dt 2 vx P : þ gy þ u þ (e þ P/r)r(v n) dA ¼ rv x dy r 2 c:s: 0 xþDx Z dt 2 vx P þ gy þ u þ rv x dy r 2 0 x Z dt 2 d v P rv x x þ gy þ u þ dy Dx dx 0 r dt 2 and @ @t ZZZ er dV ¼ 0 c:v: In the absence of signiﬁcant gravitational effects. In equation (19-29). the convective-energy-ﬂux terms become v 2x P þ u þ ¼ h0 ’ cp T0 r 2 where h0 is the stagnation enthalpy and cp is the constant-pressure heat capacity. for the energy equation both the variation in v x and in T with y must be assumed.4. and a unit depth. we obtain Z k @T d dt ¼ v x (T1 T) dy (19-30) rcp @y dx 0 y¼0 Equation (19-30) is analogous to the momentum integral relation. with the momentum terms replaced by their appropriate energy counterparts. T1 represents the free-stream stagnation temperature. The stagnation temperature will now be written merely as T (without subscript) to avoid confusion. The complete energy expression is now Z Z dt Z dt @T d dt k Dx ¼ rv c T dy rv c T dy rc Dx v x T1 dy x p x p p @y y¼0 dx 0 0 0 xþDx x ð19-29Þ Equation (19-29) can also be written as q4 ¼ q2 q1 q3 where these quantities are shown in Figure 19.284 Chapter 19 Convective Heat Transfer thickness of the thermal boundary layer. . Dividing both sides of equation (19-29) by Dx and evaluating the result in the limit as by Dx approaches zero. dt . equation (12-37). the product rcp may be taken outside the integral terms in this equation. If flow is incompressible. This equation may be solved if both velocity and temperature proﬁle are known. Thus.

. Problem 19.6 ENERGY. has proved effective in generating a modiﬁed solution for this situation. is related to the distance from the leading edge. is " #1 3 Pr Nux ﬃ 0:33 Rex (19-33) 1 (X/x)3/4 Note that this expression reduces to equations (19-25) for X = 0 19. according to and Ts ¼ T1 Ts > T1 for for 0<x<X X<x The integral technique. length X.and Momentum-Transfer Analogies 285 An assumed temperature proﬁle must satisfy the boundary conditions (1) T Ts ¼ 0 at y ¼ 0 (2) T Ts ¼ T1 Ts @ (T Ts ) ¼ 0 (3) @y @2 (4) 2 (T Ts ) ¼ 0 @y at y ¼ dt at y ¼ dt at y ¼ 0 ½see equation (19-15) If a power-series expression for the temperature variation is assumed in the form T Ts ¼ a þ by þ cy2 þ dy3 the application of the boundary conditions will result in the expression for T Ts T Ts 3 y 1 y 3 ¼ (19-31) 2 dt T1 Ts 2 dt If the velocity profile is assumed in the same form. although inexact. This section will deal with these analogies and use them to develop relations to describe energy transfer. then the resulting expression. is v 3 y 1 y3 ¼ (12-40) v1 2 d 2 d Substituting equations (19-31) and (12-40) into the integral expression and solving. It is interesting to note that equation (19-32) again involves the parameters predicted from dimensional analysis. is sufﬁciently close to the known value to indicate that the integral method may be used with conﬁdence in situations in which an exact solution is not known.19. and the unheated starting. x. Ts. and assuming both the hydrodynamic and temperature proﬁles to be cubic. as presented in this section. we obtain the result Nux ¼ 0:36 Rex1/2 Pr1/3 (19-32) which is approximately 8% larger than the exact result expressed in equation (19-25). This result. as obtained in Chapter 12.6 Energy.AND MOMENTUM-TRANSFER ANALOGIES Many times in our consideration of heat transfer thus far we have noted the similarities to momentum transfer both in the transfer mechanism itself and in the manner of its quantitative description. A condition of considerable importance is that of an unheated starting length.17 at the end of the chapter deals with this situation where the wall temperature. The result for Ts ¼ constant.

Proc. Trans.2 In 1883. Reynolds. a knowledge of the coefficient of frictional drag will enable the convective heat-transfer coefficient to be readily evaluated. Soc. . Soc. the manner of expressing them quantitatively must 2 3 O. 935 (1883). the dimensionless velocity and temperature gradients are related as follows: d v x d T Ts ¼ (19-34) dy v y¼0 dy T T 1 1 s y¼0 For Pr ¼ mcp /k ¼ 1. For those situations satisfying the basis for the development of equation (19-37). The restrictions on the use of the Reynolds analogy should be kept in mind. The latter is sensible when one considers that. giving mcp dv x h¼ (19-36) v 1 dy y¼0 Introducing next the coefficient of skin friction t0 2m dv x Cf ﬃ 2 ¼ 2 rv 1 /2 rv 1 dy y¼0 we may write equation (19-36) as h¼ Cf (rv 1 cp ) 2 which. The former of these was the starting point in the preceding development and obviously must be satisﬁed. Reynolds.286 Chapter 19 Convective Heat Transfer Osborne Reynolds ﬁrst noted the similarities in mechanism between energy and momentum transfer in 1874. 174A. for ﬂow past a solid surface with a Prandtl number of unity. As we have noted in the previous sections. 14:7 (1874). thus making possible the quantitative analogy between the two transport phenomena. we have mcp /k and we may write equation (19-34) as d v x d T Ts ¼k mc p dy v 1 y¼0 dy T1 Ts y¼0 which may be transformed to the form mcp dv x k d (T Ts ) ¼ Ts T1 dy v 1 dy y¼0 y¼0 Recalling a previous relation for the convective heat-transfer coefficient h d (Ts T) ¼ k dy (Ts T1 ) y¼0 (19-35) (19-4) it is seen that the entire right-hand side of equation (19-34) may be replaced by h. becomes h rv 1 cp St ¼ Cf 2 (19-37) Equation (19-37) is the Reynolds analogy and is an excellent example of the similar nature of energy and momentum transfer. in dimensionless form.. (London. they are (1) Pr ¼ 1 and (2) no form drag. Phil. Roy. in relating two transfer mechanisms. O. he presented3 the results of his work on frictional resistance to ﬂuid ﬂow in conduits. Manchester Lit.

A. Colburn4 has suggested a simple variation of the Reynolds analogy form that allows its application to situations where the Prandtl number is other than unity. A mass-transfer j factor. Thus. Trans. The instantaneous 4 A.7 TURBULENT FLOW CONSIDERATIONS The effect of the turbulent ﬂow on energy transfer is directly analogous to the similar effects on momentum transfer as discussed in Chapter 12.E. v 0x . and upon reaching its destination. Obviously the description of drag in terms of the coefﬁcient of skin friction requires that the drag be wholly viscous in nature. High and low Prandtl number ﬂuids falling outside this range would be heavy oils at one extreme and liquid metals at the other.19. Note that for Pr ¼ 1. described in Chapter 12. Consider the temperature proﬁle variation in Figure 19.7 Turbulent Flow Considerations 287 remain consistent.Ch. . 174 (1933). It can be easily shown that the exact expression for a laminar boundary layer on a ﬂat plate reduces to equation (19-38).1. Some possible areas of application would be ﬂow parallel to plane surfaces or ﬂow in conduits. The Prandtl number range is extended to include gases. The packet of ﬂuid moving through the distance L retains the mean temperature from its point of origin. The Colburn analogy is particularly helpful for evaluating heat transfer in internal forced ﬂows. The Colburn analogy expression is St Pr2/3 ¼ Cf 2 (19-38) which obviously reduces to the Reynolds analogy when Pr ¼ 1.5 to exist in turbulent ﬂow. the packet will differ in temperature from that of the adjacent ﬂuid by an amount TjyL Tjy : The mixing length is assumed small enough to permit the temperature difference to be written as dt (19-41) TjyL Tjy ¼ L dy y We now deﬁne the quantity T 0 as the ﬂuctuating temperature.5 – 50 as speciﬁed above. The coefﬁcient of skin friction for conduit ﬂow has already been shown to be equivalent to the Fanning ﬁction factor. The distance moved by a ﬂuid ‘‘packet’’ in the y direction. is denoted by L. is discussed in Chapter 28. and (2) 0:5 < Pr < 50. which may be evaluated by using Figure 14. P. the Prandtl mixing length. The restriction that Pr ¼ 1 makes the Reynolds analogy of limited use. which is normal to the direction of bulk ﬂow. Equation (19-38) is thus an extension of the Reynolds analogy for ﬂuids having Prandtl numbers other than unity. synonymous with the ﬂuctuating velocity component. Colburn applied this expression to a wide range of data for ﬂow and geometries of different types and found it to be quite accurate for conditions where (1) no form drag exists. and several other liquids of interest. Colburn.I. water. 29. 19. within the range 0. The Colburn analogy is often written as jH ¼ Cf 2 (19-39) where jH ¼ St Pr2/3 (19-40) is designated the Colburn j factor for heat transfer.. equation (19-37) is applicable only for those situations in which form drag is not present. the Colburn and Reynolds analogies are the same.

as defined in equation (12-52). Substituting for T its equivalent. or.5(b). it is apparent from equations (19-41) and (19-42) that T 0 ¼ L dT dy The energy ﬂux in the y direction may now be written as qy ¼ rc p Tv 0y A y (19-43) (19-44) where v 0y may be either positive or negative. with T 0 in terms of the mixing length qy dT ¼ rcp v 0y L A turb dy (19-46) The total energy ﬂux due to both microscopic and turbulent contributions may be written as qy dT ¼ rc p ½a þ jv 0y Lj (19-47) A dy As a is the molecular diffusivity of heat. for bulk flow occurring in the x direction. is accomplished because of the fluctuating temperature. . eH a for all fluids except liquid metals. for the y-directional energy ﬂux due to turbulent effects qy ¼ rcp (v 0y T 0 ) (19-45) A turb or. in equation form T ¼ T þ T0 (19-42) Any significant amount of energy transfer in the y direction.5 Turbulent-ﬂow temperature variation.288 Chapter 19 Convective Heat Transfer y T T L T' T t T (a) (b) Figure 19. we obtain. designated as eH . In a region of turbulent flow. This quantity is exactly analogous to the eddy diffusivity of momentum. eM . according to equation (19-42) qy ¼ rcp v 0y (T þ T 0 ) A y and taking the time average. the quantity jv 0y Lj is the eddy diffusivity of heat. T 0 . thus. as indicated in Figure 19. temperature is the sum of the mean and ﬂuctuating values.

we have. In his analysis solutions were obtained in each region and then joined at y ¼ j. we have Prturb ¼ eM L2 jdv x /dyj L2 jdv x /dyj ¼ 2 ¼1 ¼ eH jLv 0y j L jdv x /dyj (19-48) Thus. Zeit. can be formed by the ratio e M / e H . It is in this latter region that the Reynolds analogy applies. In terms of the eddy diffusivity of heat. .19. Physik.. in a region of fully turbulent flow the effective Prandtl number is unity. including both molecular and turbulent contributions. for which a e H . Within the laminar sublayer the momentum and heat ﬂux equations reduce to t ¼ rv dv x (a constant) dy and qy dT ¼ rcp a A dy Separating variables and integrating between y ¼ 0 and y ¼ j. thus becomes qy dT ¼ rc p (a þ eH ) A dy (19-50) Equation (19-50) applies both to the region wherein flow is laminar. the turbulent Prandtl number. and to that for which flow is turbulent and eH a. Prandtl. the hypothetical distance from the wall that is assumed to be the boundary separating the two regions. Utilizing equations (19-47) and (12-55). an analogous term.7 Turbulent Flow Considerations 289 As the Prandtl number is the ratio of the molecular diffusivities of momentum and heat. for the momentum expression Z Z vx j j t j dv x ¼ dy rv 0 0 and for the heat ﬂux Z Tj Ts dT ¼ qy Arc p a Z j dy 0 Solving for the velocity and temperature proﬁles in the laminar sublayer yields v x jj ¼ tj rv (19-51) qy j Arc p a (19-52) and Ts T j ¼ 5 L. 1072 (1910). 11. Prandtl5 achieved a solution that includes the influences of both the laminar sublayer and the turbulent core. and the Reynolds analogy applies in the absence of form drag. the heat ﬂux can be expressed as qy DT ¼ rc p eH (19-49) turb A dy The total heat flux.

Thus for the case at hand pﬃﬃﬃﬃﬃﬃﬃ v þ ¼ v x jj /( t/r) ¼ 5 . which has been defined previously as the Prandtl number. we have v i rAc rh p v 1 þ v x jj 1 ¼ (Ts T1 ) (19-55) t a qy Introducing the coefficient of skin friction Cf ¼ t rv 21 /2 and the convective heat-transfer coefﬁcient h¼ qy A(Ts T1 ) we may reduce equation (19-54) to v 1 þ v x jj (v/a 1) rcp ¼ v 21 Cf /2 h Inverting both sides of this expression and making it dimensionless.290 Chapter 19 Convective Heat Transfer Eliminating the distance j between these two expressions gives rvv x jj rAc p a ¼ (Ts Tj ) qy t (19-53) Directing our attention now to the turbulent core where the Reynolds analogy applies. this may be accomplished by recalling some results from Chapter 12. For a value of Pr ¼ 1. the Stanton number is a function of Cf. For Pr ¼ 1. we obtain Cf /2 h St ¼ rcp v 1 1 þ (v x jj /v 1 )½(v/a) 1 (19-56) This equation involves the ratio v/a. expressing h and Cf in terms of their deﬁning relations. we obtain qy /A t ¼ rcp (v 1 v x jj )(Tj T1 ) r(v 1 v x jj )2 Simplifying and rearranging this expression. we may write equation (19-37) Cf h ¼ (19-37) rcp (v 1 v x jj ) 2 and. It would be convenient to eliminate the velocity ratio. and the ratio v x jj /v 1 . equation (19-56) reduces to the Reynolds analogy. Pr. At the edge of the laminar sublayer v þ ¼ yþ ¼ 5 pﬃﬃﬃﬃﬃﬃﬃ and by deﬁnition v þ ¼ v x /( t/r). we have r(v 1 v x jj ) (Tj T1 ) ¼ rAcp t qy (19-54) which is a modiﬁed form of the Reynolds analogy applying from y ¼ j to y ¼ ymax : Eliminating Tj between equations (19-53) and (19-54).

estimate the exit temperature of the water using (a) the Reynolds analogy. 705 (1939).19. by fluid flow by convection by fluid flow 6 T. quite logically. His result. and a length of 10 ft. 61. For a constant wall temperature of 210 F. The water ﬂows at 20 gal/min. Entrance effects are to be neglected. (c) the Prandtl analogy. and (d) the von Ka´rma´n analogy. we see that an application of the ﬁrst law of thermodynamics to the control volume indicated will yield the result that 8 9 8 9 8 9 rate of heat < rate of heat = < rate of heat = < = transfer into c:v: þ transfer into c:v: ¼ transfer out of c:v: : . (b) the Colburn analogy. Considering a portion of the heat-exchanger tube shown in Figure 19. just as for the Prandtl analogy. The application of the Prandtl and von Ka´rma´n analogies is. Trans. These equations yield the most accurate results for Prandtl numbers greater than unity. restricted to those cases in which there is negligible form drag. .7 Turbulent Flow Considerations 291 Again introducing the coefﬁcient of skin friction in the form Cf ¼ we may write t rv 21 /2 rﬃﬃﬃﬃﬃ rﬃﬃﬃ Cf t ¼ v1 r 2 which. von Ka´rma´n6 extended Prandtl’s work to include the effect of the transition or buffer layer in addition to the laminar sublayer and turbulent core. This equation is written entirely in terms of measurable quantities. and the properties of water may be evaluated at the arithmetic-mean bulk temperature. : . equation (19-59) reduces to the Reynolds analogy for a Prandtl number of unity. gives rﬃﬃﬃﬃﬃ v x jj Cf ¼5 (19-57) v1 2 Substitution of equation (19-57) into (19-56) gives St ¼ Cf /2 pﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ 5 Cf /2(Pr 1) (19-58) which is known as the Prandtl analogy. An illustration of the use of the four relations developed in this section is given in the example below. the von Ka´rma´n analogy. when combined with the previous expression given for the velocity ratio. is expressed as St ¼ Cf /2 pﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ 5 Cf /2fPr 1 þ ln½1 þ 56(Pr 1)g (19-59) Note that. EXAMPLE 1 Water at 50 F enters a heat-exchanger tube having an inside diameter of 1 in.6. von Ka´rma´n. : . ASME.

Observe that the coefﬁcient of the right-hand term. assuming smooth tubing. at which n ¼ 0:474 105 ft2 /s. is the Stanton number. q2.000 v 0:474 105 ft2 /s At this value of Re. we will assume the mean bulk temperature to be 908F. the friction factor. h/rv x c p . The coefﬁcient of skin friction may be evaluated with the aid of Figure 14.0042. we have dT h 4 dx ¼ 0 þ T Ts rv x cp D and integrating between the limits indicated. is 0. This parameter has been achieved quite naturally from our analysis. The Reynolds number is Re ¼ Dv x (1/12 ft)(8:17 ft/s) ¼ ¼ 144. ff.6 Analog analysis of water ﬂowing in a circular tube. equation (19-59) reduces to dT h 4 þ (T Ts ) ¼ 0 dx rv x cp D (19-61) Separating the variables. the Stanton number is evaluated as follows: . The ﬁlm temperature will then be 1508F.1.292 Chapter 19 Convective Heat Transfer D q3 q2 q1 ∆x Flow Figure 19. we obtain Z Z TL dT h 4 L þ dx ¼ 0 rv x c p D 0 T0 T Ts TL Ts h 4L ¼0 þ ln T0 Ts rv x cp D (19-62) Equation (19-62) may now be solved for the exit temperature TL. Ts If these heat-transfer rates are designated as q1. The velocity is calculated as v x ¼ 20 gal/min (ft3 /7:48 gal)½144/(p/4)(12 )ft2 (min/60 s) ¼ 8:17 fps Initially. and q3. they may be evaluated as follows: pD2 v x cp Tjx 4 q2 ¼ hpD Dx (Ts T) q1 ¼ r and q3 ¼ r pD2 v x c p TjxþDx 4 The substitution of these quantities into the energy balance expression gives r pD2 v x c p ½TjxþDx Tjx hpD Dx(Ts T) ¼ 0 4 which may be simpliﬁed and rearranged into the form D TjxþDx Tjx h þ (T Ts ) ¼ 0 4 rv x cp Dx (19-60) Evaluated in the limit as Dx ! 0. For each of the four analogies.

19. the following results: (a) TL ¼ 152 F (b) TL ¼ 115 F (c) TL ¼ 132 F (d) TL ¼ 125 F Some ﬁne tuning of these results may be necessary to adjust the physical property values for the calculated ﬁlm temperatures. Stanton. we obtain.8 Closure 293 (a) Reynolds analogy St ¼ Cf ¼ 0:0021 2 (b) Colburn analogy St ¼ Cf 2/3 Pr ¼ 0:0021(2:72)2/3 ¼ 0:00108 2 (c) Prandtl analogy St ¼ ¼ Cf /2 pﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ 5 Cf /2(Pr 1) 0:0021 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ 0:00151 1 þ 5 0:0021(1:72) (d) von Ka´rma´n analogy St ¼ ¼ Cf /2 pﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ 5 Cf /2 Pr 1 þ ln 1 þ 56 (Pr 1) 0:0021 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 1 þ 5 0:0021 2:72 1 þ ln 1 þ 56 (2:72 1) ¼ 0:00131 Substituting these results into equation (19-62). New parameters pertinent to convection are the Prandtl. as the Prandtl number was considerably above a value of one. . The last three analogies yielded quite consistent results.8 CLOSURE The fundamental concepts of convection heat transfer have been introduced in this chapter.19. The Reynolds analogy value is much different from the other results obtained. Nusselt. and Grashof numbers. for TL. The Colburn analogy is the simplest to use and is preferable from that standpoint. In none of these cases is the assumed ﬁlm temperature different than the calculated one by more than 68F. This is not surprising. so the results are not going to change much.

the mean air temperature vs. and v from Problem 19. 1. The distance x may be taken as 0. gap between them. If air at 120 F. water. the magnitude of hx at 15 cm.294 Chapter 19 Convective Heat Transfer Four methods of analyzing a convection heat-transfer process have been discussed. prepare plots showing a.22 m long.6.21 for velocity and temperature proﬁles of the form v ¼ a þ by þ cy2 T Ts ¼ a þ by þ gy2 q px ¼ a þ b sin A L 19. and glycerin. and plate surface temperature as functions of x. determine. (3) integral analysis of the boundary layer. using the values for x. These are as follows: (1) dimensional analysis coupled with experiment. for the case of air at 310 K adjacent to a vertical wall with its surface at 420 K. is used to cool the plates.8 In a thermal heat sink the heat ﬂux variation along the axis of a cooling passage is approximated as px q ¼ a þ b sin A L where x is measured along the passage axis and L is its total length. 30 cm.5 m. demonstrate that the parameters T T1 T0 T1 x L at L2 and hL k are possible combinations of the appropriate variables in describing unsteady-state conduction in a plane wall. 508.26 for the case of a wall heat ﬂux varying according to where a ¼ 250 Btu/h ft2 . the heat ﬂux vs.1 Using dimensional analysis. x. 80 psi. x. hx/k.16. Where does the maximum surface temperature occur and what is its value? 19. 19. determine the total heat transferred for a stack of plates with a combined surface area of 640 ft2.9 Determine the total heat transfer from the vertical wall described in Problem 19. 30 cm.2 Dimensional analysis has shown the following parameters to be signiﬁcant for forced convection: xv 1 r m mc p k hx k h rcp v 1 Evaluate each of these parameters at 340 K. benzene. and h ¼ 34 W/m2 K. The ﬂow passages are 1. each plate being 4 ft wide. x is the distance from the leading edge of the plates. and glycerin. (2) exact analysis of the boundary layer.2. 19. for air.5 m high. and h/rcp v 1 vs. construct a plot of local heat-transfer coefﬁcient vs. 19. temperature for air. and (4) analogy between momentum and energy transfer. and the heat ﬂux in the plates varies according to the above equation where a ¼ 900 W/m2 and v ¼ 2500 W/m2 : Air enters at 100 C with a mass velocity (the product of rV) of 7:5 kg/s m2 : The surface coefﬁcient along the ﬂow passage can be considered constant with a value of 56 W/m2 K: Generate a plot of heat ﬂux. and 808F. ﬂowing at a mass velocity of 6000 lbm /h ft2 . b.4 Using the relations from Problem 19. 19.4 to the surrounding air per meter of width if the wall is 2. h. water. the thickness of the boundary layer at x ¼ 15 cm. v 1 ¼ 15 m/s. mcp /k. 19.5 Given the conditions speciﬁed in Problem 19.10 Repeat Problem 19. a. 19. 19.3 m.7 Given the information in Problem 19. mean air temperature. The heat ﬂux along the plate surfaces varies sinusoidally according to the equation length. Several empirical equations for the prediction of convective heat-transfer coefﬁcients will be given in the chapters to follow. b ¼ 1500 Btu/h ft2 .3 Plot the parameters xv 1 r/m.11 Solve Problem 19. position along the plate for glycerin temperatures of 308. 19. 1. A large installation involves a stack of plates with a 3-mm air space between them.6 The fuel plates in a nuclear reactor are 4 ft long and stacked with a 1/2-in.5 m. b. PROBLEMS 19. and L is the total plate q px ¼ a þ b sin A L .19. mercury.

25 for the case in which the ﬂowing ﬂuid is air at 15 fps. and hx at a distance of 40 cm from the leading edge of a ﬂat plane.8. is aligned parallel to the direction of ﬂow. Btu/h ft2 F: Determine the values for a and b for the plane vertical wall. What heat ﬂux will result.17 Shown in the ﬁgure is the case of a ﬂuid ﬂowing parallel to a ﬂat plate. c. Atmospheric air. (h) total heat transfer. in each case. and L is the tube length. x is the distance from the entrance. Assuming a cubic proﬁle for both the hydrodynamic and the thermal boundary layers. 508. L is a signiﬁcant length. (e) hx.16 Glycerin ﬂows parallel to a ﬂat plate measuring 2 ft by 2 ft with a velocity of 10 fps. (g) total drag force. wide. where for a distance X from the leading edge. For values of x>X. where Ts > T1 . Using the results of boundary-layer analysis. at 25 C. has its temperature maintained at 80 C. DT is Ts T1 . 19. 19.18 Simpliﬁed relations for natural convection in air are of the form h ¼ aðDT/LÞb where a. if the plate temperature is 508F above that of the glycerin? 19. A ﬂat plate 6 in. b are constants. determine the following (a) d. the total heat transfer rate from the plate to the air stream.8 m/s. (d) Cﬂ. and h is the convective heat-transfer coefﬁcient. b ¼ 1500 Btu/h ft2 . Ts. j. (c) Cfx. in ft. show that the ratio of the thickness. Also show that the local Nusselt number can be expressed as !1/3 Pr Re1x /2 Nux ﬃ 0:33 1 (X/x)3/4 19. using the equation from Problem 19. C fx . 19. at a temperature of 200 F. the plate is maintained at a constant temperature. the mean coefﬁcient of skin friction. and 1808F. the plate and ﬂuid are at the same temperature. b. the total drag force exerted on the plate by the air ﬂow. CfL.13 Nitrogen at 100 F and 1 atm ﬂows at a velocity of 100 fps. show that the solution in terms of d and v x from each integral equation reduce to 2a d dv x ¼ d dx 30 and vv x d d dv 2x þ bg DT ¼ d 3 dt 105 Next.19 Using the integral relations from Problem 19. At a position 4 ft from the leading edge. (f ) h. for the case of natural convection adjacent to a plane vertical wall. Air with a free stream . is expressed as " 3/4 #1/3 dt 1 X 1 j¼ ﬃ 1 / 3 d x Pr dt X 0 Z d1 Z dv x d d 2 þ bg ðT T1 Þ dy ¼ v dy v @y y¼0 dx 0 x 0 295 where d is the thickness of both the hydrodynamic and thermal boundary layers. (b) dt .Problems where a ¼ 250 Btu/h ft2 .14. and assuming the velocity and temperature proﬁles of the form y v y2 1 ¼ vx d d and T T1 y2 ¼ 1 Ts T1 d and 19. 19.20 Use the results of Problem along with those of Chapter 12 to determine d. ﬂows parallel to the surface with a velocity of 2. d1 .14 A plane surface. 25 cm wide. Determine values for the mean convective heat-transfer coefﬁcient and the associated drag force imposed on the plate for glycerin temperatures of 3508. in 8F.12 Work Problem 19. assuming that both d and v x vary with x according to d ¼ Axa and v x ¼ Bxb show that the resulting expression for d becomes d/x ¼ 3:94 Pr1/2 ðPr þ 0:953Þ1/4 Grx1/4 and that the local Nusselt number is Nux ¼ 0:508 Pr1=2 ðPr þ 0:953Þ1=4 Gr1=4 x 19.15 Show that. determine the following for a 1-m long plate: a. the appropriate integral equations for the hydrodynamic and thermal boundary layers are @T a @y ¼ y¼0 d dx Z d1 v x ðT1 T Þdy d x 19.

enters a 1-in.25 for the case in which the ﬂowing ﬂuid is sodium entering the tube at 2008F 19. . Awind.29 Work Problem 19. that is 1/7 T Ts y ¼ d1 T1 Ts and assuming that d ¼ d1 .296 Chapter 19 Convective Heat Transfer velocity of 5 m/s and T1 ¼ 300 K ﬂows parallel to the plate surface. The ﬁrst 20 cm of the plate is unheated.28 Work Problem 19. 19.27 Work Problem 19. develop expressions for the local Nusselt number in terms of Rex and Pr for velocity and temperature proﬁles of the form v ¼ a þ by.-ID tube that is used to cool a nuclear reactor. the velocity proﬁle has been shown to follow closely the form y1/7 v ¼ v1 d Assuming a temperature proﬁle of the same form.22 Repeat Problem 19. The temperature gradient at the surface may be considered similar to the velocity gradient at y ¼ 0 given by equation (13-26). at 608F. Determine the wind velocity that will cause the road surface to be at 308 K if all energy not reradiated to the sky is removed by convection. 19. The water ﬂow rate is 30 gal/min.25 Water.23 For the case of a turbulent boundary layer on a ﬂat plate.21 Using the appropriate integral formulas for ﬂow parallel to a ﬂat surface with a constant free-stream velocity. Compare the answer obtained. the exiting water temperature. 19. ﬂows across the road. Determine the total heat transfer and the exiting water temperature for a 15-ft-long tube if the tube surface temperature is a constant value of 3008F. 19. Determine the total heat transfer. T Ts ¼ a sin bu 19. T Ts ¼ a þ by 19. The water ﬂow rate is 30 gal/min.-ID tube that is used to cool a nuclear reactor.24 A blacktop road surface 18. and the wall temperature at the exit of a 15-ft long tube if the tube wall condition is one of uniform heat ﬂux of 500 Btu/hr ft2.21 for velocity and temperature proﬁles of the form v ¼ a sin by. using the Reynolds and Colburn analogies.26 for the case in which the ﬂowing ﬂuid is air at 15 fps.26 for the case in which the ﬂowing ﬂuid is sodium entering the tube at 2008F. 19. use the integral relation for the boundary layer to solve for hx and Nux.26 Water at 608F enters a 1-in. at 300 K. 19.3 m wide receives solar radiation at the rate of 284 W/m2 at noon and 95 W/m2 are lost by reradiation to the atmosphere. the surface temperature is maintained at 400 K beyond that point.

it may not offer a practical solution to every problem. The natural convection system most amenable to analytical treatment is that of a ﬂuid adjacent to a vertical wall.. 1. Beckmann. those analytical relations that are available are presented along with the most satisfactory empirical correlations for a particular geometry and ﬂow condition. respectively. Gregg. Sparrow and J.5-cm-high vertical plate are shown in Figures 20. 1953. There are many situations for which no mathematical models have as yet been successfully applied. M.M.3 1 2 3 E. Schmidt and Beckmann1 measured the temperature and velocity of air at different locations near a vertical plate and found a signiﬁcant variation in both quantities along the direction parallel to the plate. Although the analytic approach is very meaningful. and forced convection for external ﬂow. Because of this. with y measured normal to the plane surface. Schmidt and W. 341 and 391 (1930).3 for the conditions Ts ¼ 65 C. L. The former of these cases have been solved by Ostrach2 and the latter by Sparrow and Gregg. S.S. E. Most correlations are in the forms indicated by dimensional analysis. In this chapter. we shall present some of the most useful correlations of experimental heat-transfer data available. The sections to follow include discussion and correlations for natural convection. 297 . Trans. it is quite natural that the heat-transfer coefﬁcients and their correlating equations will vary with the geometry of a given system. A. Tech. U. The x direction is commonly taken along the wall. 435 (1956).Chapter 20 Convective Heat-Transfer Correlations Convective heat transfer was treated from an analytical point of view in Chapter 19. T1 ¼ 15 C: The two limiting cases for vertical plane walls are those with constant surface temperature and with constant wall heat ﬂux.2 and 20. Vertical Plates. Even in those cases for which an analytical solution is possible.E. Mech. Standard nomenclature for a two-dimensional consideration of natural convection adjacent to a vertical plane surface is indicated in Figure 20. Ostrach.1 NATURAL CONVECTION The mechanism of energy transfer by natural convection involves the motion of a ﬂuid past a solid boundary. forced convection for internal ﬂow. NACA Report 1111. The variations of velocity and temperature for a 12. In each case. which is the result of the density differences resulting from the energy exchange. Thermodynamik. it is necessary to verify the results by experiment. 20.1. 78.

3 0.1. We usually ﬁnd the mean Nusselt number.1 Coordinate system for the analysis of natural convection adjacent to a heated vertical wall.6 0. in cm/sec 60 x= 20 2 cm 10 30 1 8 2 cm 4 cm cm 7 cm 6 1 3 20 cm 0.0 y.9 1. in cm Figure 20. an expression for NuL may be determined 26 70 24 22 11 18 16 cm 7c 50 m 40 cm T. to be of more value than Nux. in °C 11 m 12 x= 14 4c vx.2 Velocity distribution in the vicinity of a vertical heated plate in air.5 0. y Ostrach. in cm 0.8 0. 10 0 0. NuL.8 1.4 0. obtained an expression for local Nusselt number of the form 1/4 Grx Nux ¼ f (Pr) (20-1) 4 The coefficient.298 Chapter 20 Convective Heat-Transfer Correlations Heated wall Fluid T∞ x Ts Figure 20. as discussed earlier. 4 cm 2 T∞ 0 0 0.0 Figure 20.1 0.3 Temperature distribution in the vicinity of a vertical heated plate in air.2 0. with values given in Table 20.2 0. and energy in a free-convection boundary layer. Using an integration procedure. motion.6 y. employing a similarity transformation with governing equations of mass conservation.7 0. . f(Pr). varies with Prandtl number.4 0.

This powerful equation is 92 8 > > > > = < 0:387 RaL1/6 NuL ¼ 0:825 þ h (20-4) i 8/27 > > > > 9 / 16 . may thus be used. Equations (20-1) and (20-2).716 10 1. Trans.191 1000 3.1 Natural Convection 299 Table 20. is related to the local value according to Z 1 L hL ¼ hx dx L 0 Inserting equation (20-1) appropriately. L. Chu. : 1 þ (0:492/Pr) 4 5 E. Churchill and H. Gregg. When turbulence is present. S. and we must rely heavily on correlations of experimental data. M.01 0. to evaluate any vertical plane surface regardless of wall conditions. Transition from laminar to turbulent ﬂow in natural convection boundary layers adjacent to vertical plane surfaces has been determined to occur at. . The mean heat transfer coefﬁcient for a vertical surface of height. provided boundary layer flow is laminar. 1323 (1975). W.S. 78. is often referred to as Ra. in dimensionless form.1. an analytical approach is quite difﬁcult.1 Values of the coefﬁcient f(Pr) for use in equation (20-1) Pr f (Pr) 0. the Rayleigh number. They propose a single equation for NuL that applies to all ﬂuids.. Gr Pr. H.169 100 2. indicates transition. 435 (1956). Heat & Mass Tr. with reasonable accuracy. L. we proceed to Z k bgDT 1/4 L 1/4 hL ¼ f (Pr) x dx L 4v2 0 4 k bgL3 DT 1/4 ¼ f (Pr) 3 L 4v2 and. being temperature dependent.081 0. t. turbulent ﬂow will also occur in free convection boundary layers. J. therefore. 18. we have 4 GrL 1/4 NuL ¼ f (Pr) 3 4 (20-2) Sparrow and Gregg’s4 results for the constant wall heat flux case compared within 5% to those of Ostrach for like values of Pr. Fluid properties.M.567 2 0.E. along with coefficients from Table 20. Ts þ T1 Tf ¼ 2 As with forced convection.20. S. It is important.966 using equation (20-1). that properties involved in equations (20-1) and (20-2) be evaluated at the ﬁlm temperature. or near Grt Pr ¼ Rat ﬃ 109 (20-3) where the subscript. Churchill and Chu5 have correlated a large amount of experimental data for natural convection adjacent to vertical planes over 13 orders of magnitude of Ra. The product. A. Sparrow and J. Int. will have some effect on calculated results.505 1 0.072 0..

300 Chapter 20 Convective Heat-Transfer Correlations Churchill and Chu show this expression to provide accurate results for both .

18. L. New York. Churchill and H. With cylinders of sufﬁcient length that end effects are insigniﬁcant. The correlations suggested by McAdams6 are well accepted for this geometry. should be used for fluid property evaluation. Tf . Int. The length scale. For the case of cylinders with their axes vertical. Chapter 7. H. Heat Transmission. S. S. Heat & Mass Tr. Churchill and Chu7 suggest replacing g by g cos u in equation (20-5) when boundary layer ﬂow is laminar. speciﬁcally. A distinction is made regarding whether the ﬂuid is hot or cool. Chu. For plane surfaces inclined at an angle. Heat & Mass Tr. and whether the surface faces up or down. with modiﬁcation. Churchill and Chu8 suggest the following correlation 92 8 > > > > = < 1/ 6 0:387 RaD (20-10) NuD ¼ 0:60 þ h i8/27 > > > > 9 / 16 . J. the film temperature. is the ratio of the plate-surface area to perimeter. 18. 1323 (1975). with the vertical. McAdams’s correlations are. It is clear that the induced buoyancy will be much different for a hot surface facing up than down. for a hot surface facing up or cold surface facing down 105 < RaL < 2 107 NuL ¼ 0:54 RaL1/4 (20-7) 2 107 < RaL < 3 1010 NuL ¼ 0:14 Ra1L/3 (20-8) and for a hot surface facing down or cold surface facing up 3 105 < RaL < 1010 NuL ¼ 0:27 Ra1L/4 (20-9) In each of these correlating equations. u. D. two correlations are recommended. the expressions presented for plane surfaces can be used provided the curvature effect is not too great.. : 1 þ (0:559/Pr) over the Rayleigh number range 105 < RaD < 1012 : 6 W. a vertical cylinder can be evaluated using correlations for vertical plane walls when D 35 L GrL1/4 (20-6) Physically. W. Churchill and H. 1049 (1975). 1957.. laminar and turbulent flows. equations (20-4) and (20-5) may be used. With turbulent ﬂow. McGraw-Hill Book Company. Horizontal Plates. this represents the limit where boundary layer thickness is small relative to cylinder diameter. H. 7 S. The criterion for this is expressed in equation (20-6). J. Some improvement was found for the laminar range RaL < 109 by using the following equation: 0:670 RaL1/4 (20-5) NuL ¼ 0:68 þ h i4/9 1 þ (0:492/Pr)9/16 Vertical Cylinders. Third Edition. McAdams. W. relative to the adjacent surface. 8 S. Horizontal Cylinders. for values of u up to 608. Int. H. equation (20-4) may be used without modiﬁcation. Chu. .

Heat Transfer. Eds. New York.125 0. Advances in Heat Transfer. 82.. 9 V. J.9 in terms of variable coefﬁcients NuD ¼ C RanD (20-11) where values of c and n are specified as functions of RaD in Table 20. Rectangular Enclosures. F.250 0. T.058 0. Clearly.20.333 The ﬁlm temperature should be used in evaluating ﬂuid properties in the above equations.148 0.2. Irvine and J. Table 20.4 is the conﬁguration and nomenclature pertinent to rectangular enclosures. and 1 < RaD < 105 : 1/ 4 NuD ¼ 2 þ 0:43 RaD (20-12) We may notice that.850 0.480 0. Shown in Figure 20. These cases have become much more important in recent years due to their application in solar collectors. for the sphere. pp. P. 214 (1960). 10 T.188 0. Hartnett. Vol. The correlation suggested by Yuge10 is recommended for the case with Pr ’ 1. Yuge. Morgan. Academic Press. This problem may be solved to yield a limiting value for NuD equal to 2. Spheres. 199–264. heat transfer will be affected L g H H Heated surface q W L Figure 20.675 1. .4 The rectangle enclosure.02 0. II.2 Values of constants C and n in equation (20-11) 1010 < RaD < 102 102 < RaD < 102 102 < RaD < 104 104 < RaD < 107 107 < RaD < 1012 C n 0. T. heat transfer from the surface to the surrounding medium is by conduction. This result is obviously compatible with equation (20-12). as Ra approaches zero. 1975.1 Natural Convection 301 A simpler equation has been suggested by Morgan.

1 < Pr < 2 104 . Horizontal enclosures. Catton12 suggests the use of the following correlations: 0:29 Pr RaL NuL ¼ 0:18 (20-15) 0:2 þ Pr when 1 < H/L < 2. A correlation for this case has been proposed by Globe and Dropkin11 in the form NuL ¼ 0:069 RaL1/3 Pr0:074 (20-14) for the range 3 105 < RaL < 7 109 : When u ¼ 180 . Dropkin. Heat Transfer. Globe and D. and by the usual dimensionless parameters. P. heat transfer is by conduction. 13 (1978). 103 < Pr < 105 . MacGregor and A. 91. Tf ¼ (T1 /T2 )/2: Convective heat ﬂux is expressed as q ¼ h(T1 T2 ) (20-13) A Case 1. Proc. K. These are NuL ¼ 0:42 RaL1/4 Pr0:012 (H/L)0:3 (20-17) for 10 < H/L < 40. the temperature of the hotter of the two large surfaces is designated as T1. u. Toronto. Heat Tr. by the aspect ratio. u = 0 With the bottom surface heated. For higher values of H/L. (20-18) . RaL < 1010 . In each of the correlations to follow. and the cooler surface is at temperature T2. 103 < RaL Pr/(0:2 þ Pr) and Pr NuL ¼ 0:22 RaL 0:22 þ Pr 0:28 1/4 H L (20-16) when 2 < H/L < 10. 6. J. NuL ¼ 1: Case 2. or when RaL < 1700. 24 (1959). Pr and RaL. Heat Transfer. Fluid properties are evaluated at the ﬁlm temperature. the correlations of MacGregor and Emery13 are recommended. Emery. Vertical enclosures. I. a critical Rayleigh member has been determined by several investigators to exist. 391 (1969).302 Chapter 20 Convective Heat-Transfer Correlations by the angle of tilt. Conference. Pr < 105 . Catton. thus. Canada. that is. P. the upper surface is heated. 81C. For cases where RaL ¼ bgL3 (T1 T2 ) > 1700 av conditions within an enclosure are thermally unstable and natural convection will occur. u = 90 For aspect ratios less than 10. J. H/L. 6th Int. 104 < RaL < 107 and NuL ¼ 0:046 RaL1/3 11 12 13 S.

Catton. Equation (20-19) was suggested by Hollands et al. are the following: 1708 1708(sin1:8u)1:6 1 RaL cos u RaL cos u 1/3 RaL cos u þ 1 5830 NuL ¼ 1 þ 1:44 1 (20-19) when H/L 12. 1 < Pr < 20. K. and s (sides) apply to the surfaces in question. 0 < u < 70 : In applying this relationship any bracketed term with a negative value should be set equal to zero. b (bottom). J. Raithby. For smaller aspect ratios the critical angle of tilt is also smaller. S. Hollands. Buchberg. S.2 m high. we have qtotal 14 15 16 k k k ¼ Nut (0:442) þ Nub (0:442) þ Nus (2:83) (T 295) 0:75 0:75 1:2 K.14 With enclosures nearing the vertical. T. G. Tilted vertical enclosures. 543 (1973). Surface areas that apply are p (0:75 m)2 ¼ 0:442 m2 4 ¼ p(1:2 m)(0:75 m) ¼ 2:83 m2 Atop ¼ Abottom ¼ Aside The total heat transfer is the sum of the contributions from the three surfaces. The heat dissipation rate from the transformer is constant at 1. Ayyaswamy and I.75 m in diameter and 1.5 KW. 0 < u < 90 Numerous publications have dealt with this conﬁguration. Rewriting this expression in terms of Nu.20. The tank contains a transformer immersed in an oil bath that produces a uniform surface temperature condition. Konicek. . P. A recommended review article on the subject of inclined rectangular cavities is that of Buchberg. and Edwards. J. Ayyaswamy and Catton15 suggest the relationship NuL ¼ NuLV ðsinuÞ1/4 (20-20) for all aspect ratios. E. Edwards. 98. All heat loss from the surface may be assumed due to natural convection to surrounding air at 295 K. The value u ¼ 70 is termed the ‘‘critical’’ tilt angle for vertical enclosures with H/L > 12. and 70 < u < 90 . This may be written as qtotal ¼ ½ht (0:442) þ hb (0:442) þ hs (2:83)(T 295) where subscripts t (top). D. Catton and D. Heat Transfer. 98. 182 (1976). and L. J. I. 106 < RaL < 109 Case 3. Heat Transfer. Heat Transfer.16 EXAMPLE 1 Determine the surface temperature of a cylindrical tank measuring 0. Catton. 189 (1976). Unny. H.1 Natural Convection 303 for 10 < H/L < 40. 95. G. Correlations for this case. when the aspect ratio is large ðH/L > 12Þ.

k ¼ 0:0293 W/mK. a surface temperature value is assumed for property evaluation. (20-9). the effect of curvature may be neglected if D 35 ¼ 0:106 L (11:7 109 )1/4 In the present case D/L ¼ 0:75/1:2 ¼ 0:625. and Nus. v ¼ 1:955 105 m2 /s.304 Chapter 20 Convective Heat-Transfer Correlations or qtotal ¼ ½0:589 Nut þ 0:589 Nub þ 2:36 Nus k(T 295) A complication exists in solving this equation because the unknown quantity is surface temperature. and (20-4) will be employed. we assume that Tsurface ¼ 38:5 K. and then solved for T. Properties will thus be evaluated at Tf ¼ 340 K: For air at 340 K. thus the vertical surface will be treated using equations for a plane wall. The procedure to be used is trial and error where. Pr ¼ 0:699. Values must now be determined for Nut. and bg/v2 ¼ 0:750 108 1/K m3 . To begin the problem. a ¼ 2:80 105 m2 /s. For the vertical surface bg Gr ¼ 2 L3 DT v ¼ (0:750 108 1/K m3 )(1:2 m)3 (90 K) ¼ 11:7 109 According to equation (20-6). Nub. Equations (20-8). This new value for surface temperature will then be used and the procedure continued until the resulting temperature agrees with the value used in ﬁnding ﬂuid properties. 0:699 ¼ 236 The solution for T is now T ¼ 295 þ 1500 W ½0:589(44) þ 0:589(20:2) þ 2:36(236)(0:0293) ¼ 381:1 K . initially. The three values for Nu are determined as follows: Nut: pD2 /4 D ¼ ¼ 0:1875 m pD 4 Nut ¼ 0:14½(0:750 108 )(0:1875)3 (90)(0:699)1/3 ¼ 44:0 L ¼ A/ p ¼ Nub: Nub ¼ 0:27½(0:750 108 )(0:1875)3 (90)(0:699)1/4 ¼ 20:2 Nus: 8 > > > > > > < 92 > > > > > 1/6 > 3 8 0:387½(0:750 10 )(1:2) (90)(0:699) = Nus ¼ 0:825 þ " > > #8/27 > > > > 0:492 9/16 > > > > 1 þ > > : .

25. we may write the velocity proﬁle as r 2 v x ¼ v max 1 (8-7) R 17 L. Ann.2 Forced Convection for Internal Flow 305 Using this as the new estimate for Tsurface. Considering the system as depicted in Figure 20. The velocity proﬁle is parabolic and fully developed before any energy exchange between the tube wall and the ﬂuid occurs.20. The momentum transfer associated with this type of ﬂow is studied in Chapter 13. Graetz. . The ﬁrst analytical solution for laminar ﬂow forced convection inside tubes was formulated by Graetz17 in 1885. Tf ﬃ 338 K. Many of the concepts and terminology of that chapter will be used in this section without further discussion. Laminar Flow. 2. we have a ﬁlm temperature. Energy transfer associated with forced convection inside closed conduits will be considered separately for laminar and turbulent ﬂow. The surface temperature of the tube is constant at a value Ts during the energy transfer. Chem. 1 þ (0:492/0:699)9/16 ¼ 235 The revised value for Ts is now 1500/0:0293 ½0:589(43:8) þ 0:589(20:1) þ 2:36(235) ¼ 381:4 K Ts ¼ 295 þ This result is obviously close enough and the desired result for surface temperature is Tsurface ﬃ 381 K 20. The reader will recall that the critical Reynolds number for conduit ﬂow is approximately 2300. Air properties at this temperature are v ¼ 1:936 105 m2 /s a ¼ 2:77 105 m2 /s k ¼ 0:0291 W/mK Pr ¼ 0:699 bg/v2 ¼ 0:775 108 1/K m3 The new values for Nu become 1/3 ¼ 43:8 Nut ¼ 0:14 (0:775 108 )(0:1875)3 (86)(0:699) 1 / 4 ¼ 20:10 Nut ¼ 0:27 (0:775 108 )(0:1875)3 (86)(0:699) 8 92 > 1/6 > > > < = 3 8 0:387 0:775 10 (1:2) (86)(0:699) Nus ¼ 0:825 þ h i8/27 > > > > : .. u. 337 (1885).2 FORCED CONVECTION FOR INTERNAL FLOW Undoubtedly the most important convective heat-transfer process from an industrial point of view is that of heating or cooling a ﬂuid that is ﬂowing inside a closed conduit. The assumptions basic to the Graetz solution are as follows: 1. 3.5. All properties of the ﬂuid are constant. Phys.

x or. recalling that v max ¼ 2v avg . assuming radial symmetry.306 Chapter 20 Convective Heat-Transfer Correlations T = Te for x < 0 T = Ts for x > 0 Fully developed velocity profile Temperature Te Figure 20. for 0. we may write r 2 v x ¼ 2v avg 1 R (20-21) The applicable form of the energy equation written in cylindrical coordinates. and neglecting @ 2 T/@x2 (axial conduction) in comparison to the radial variation in temperature is @T 1 @ @T ¼a r vx (20-22) @x r @r @r Substituting equation (20-21) for yx into equation (20-22) gives r 2 @T 1 @ @T ¼a r 2v avg 1 R @x r @r @r (20-23) which is the equation to be solved subject to the boundary conditions T ¼ Te at x ¼ 0 T ¼ Ts at x > 0.5 Boundary and ﬂow conditions for the Graetz solution.

r.

in terms of dimensionless parameters already introduced. The argument of the exponential. r ¼ 0 The solution to equation (20-23) takes the form 1 r X T Te x 2 a ¼ cn f exp bn R Rv avg R Ts Te n¼0 (20-24) The terms cn. f(r/R). that is. may be rewritten as 4 4 ¼ (2Rv avg /a)(2R/x) (Dv avg r/m)(c p m/k)(D/x) or. (a/Rv avg )(x/R). and bn are all coefficients to be evaluated by using appropriate boundary conditions. this becomes 4 4x/D ¼ Re Pr D/x Pe .R r¼R and @T ¼0 @r at x > 0. exclusive of bn.

Sieder and G. Katz.6. in Figure 20. we must resort to correlations of experimental data as suggested by 18 J. the ratio of the fluid viscosity at the arithmetic-mean bulk temperature to that at the temperature of the wall. Another parameter encountered in laminar forced convection is the Graetz number. Gz. . Tate. Eng. 1429 (1936). 370.6 presents the results of the Graetz solution graphically for two different boundary conditions at the wall. N. 1958. Chem. Turbulent Flow.6 along with the two Graetz results. McGraw-Hill Book Company. Pe. Ind. 19 F. takes into account the significant effect that variable fluid viscosity has on the heat-transfer rate. Knudsen and D. When considering energy exchange between a conduit surface and a ﬂuid in turbulent ﬂow. New York. these being (1) a constant wall temperature and (2) uniform heat input at the wall. 28. the analytical results approach constant limiting values for large values of x.6 Variation in the local Nusselt number for laminar ﬂow in tubes. and Knudsen and Katz18 summarize these quite well. Figure 20.. Note that. Fluid Dynamics and Heat Transfer.2 Forced Convection for Internal Flow 307 The product of Re and Pr is often referred to as the Peclet number. L. The last part of equation (20-27).20. These limits are Nux ¼ 3:658 for Twall ¼ constant (20-25) Nux ¼ 4:364 for q/Awall ¼ constant (20-26) Experimental data for laminar ﬂow in tubes have been correlated by Sieder and Tate19 by the equation D 1/3 mb 0:14 (20-27) NuD ¼ 1:86 Pe L mw The Sieder–Tate relation is also shown in Figure 20. Nux = hD k x 102 Sieder and Tate 10 Graetz 102 Pe (D/x) 10 Constant wall heat flux Constant wall temperature 103 104 Figure 20. All properties other than mw are evaluated at the bulk fluid temperature. E. deﬁned as Gz pD Pe 4 x Detailed solutions of equation (20-24) are found in the literature. p. These results cannot be compared directly because the Graetz results yield local values of hx and the Sieder–Tate equation gives mean values of the heat-transfer coefficient. G.

the ﬁrst two are most often used for those ﬂuids whose Prandtl numbers are within the speciﬁed range. Eng. All ﬂuid properties except mw are evaluated at bulk temperature. all ﬂuid properties are evaluated at the arithmetic-mean bulk temperature. Dittus and Boelter20 proposed the following equation of the type suggested earlier by dimensional analysis. W. n ¼ 0:3 if the ﬂuid is being cooled.. His equation is 2/3 St ¼ 0:023 Re0:2 D Pr (20-29) where 1. and 5. ReD > 104 . Tate. Boelter.308 Chapter 20 Convective Heat-Transfer Correlations dimensional analysis. A. Trans. 2. Ind.. 3.. 2. the value of ReD should be > 104. N. K. 0:7 < Pr < 17 000. Dittus and L. Eng. 443 (1930). E. ReD. L/D > 60. The Dittus–Boelter equation is simpler to use than the Colburn equation because of the ﬂuid property evaluation at the bulk temperature. 2. The three most-used equations of this nature and the restrictions on their use are as follows. 2. Sieder and Tate22 proposed the equation 0:14 2/3 mb Pr (20-30) St ¼ 0:023 Re0:2 D mw where 1. and L/D should have values within the following limits: ReD > 104 0:7 < Pr < 160 and L/D > 60 To account for high Prandtl number ﬂuids. 29. Sieder and G.I. Pr is in the range 0:7 < Pr < 100. Of the three equations presented. University of California. P. Colburn21 proposed an equation using the Stanton number. 1429 (1936). A. 28. Chem. equation (19-7): n NuD ¼ 0:023 Re0:8 D Pr (20-28) where 1. such as oils. L/D > 60. in place of NuD as related in equation (19-9). and St is evaluated at the bulk temperature. 4. E. and 4. 20 21 22 F. ReD and Pr are evaluated at the ﬁlm temperature.E. . St. M. 174 (1933). Colburn. n ¼ 0:4 if the ﬂuid is being heated. Publ. Pr.Ch. 3.

we need to know the exiting oil temperature. at this temperature. To evaluate oil properties at either the ﬁlm temperature or the mean bulk temperature. Equation (19-61) applies in this case ln TL Ts L h þ4 ¼0 D rvc p To Ts (19-62) If the thermal resistance of the copper tube wall is negligible. 400 W/m2 K. in fully developed ﬂow. EXAMPLE 2 Hydraulic ﬂuid (MIL-M-5606). is ReD ¼ (0:025 m)(0:05 m/s) ¼ 126 9:94 106 m2 /s and the ﬂow is clearly laminar. ﬁnd the rate of heat transfer to the oil. Using ﬂuid properties at these temperatures. the mean bulk temperature of the oil is Tb ¼ 295 þ 304 ¼ 299:5 K 2 which is sufﬁciently close to the initial assumption that there is no need to iterate further.05 m/s. we have 0:14 (0:123 W/mK)(1:86) 0:025 0:33 1:036 104 hi ¼ (126)(155) 3:72 103 0:025 m 0:61 ¼ 98:1 W/m2 K Substituting into equation (19-61). we get Ts TL 0:61 m 98:1 W/m2 K ln ¼ 4 3 0:025 m (843 kg/m )(0:05 m/s)(1897 J/kg K) Ts To ¼ 0:120 Ts TL ¼ e0:120 ¼ 0:887 Ts To TL ¼ 372 0:887(372 295) ¼ 304 K With this value of TL. steam condenses on the outside tube surface with an effective heat-transfer coefﬁcient of 11. The oil enters at 295 K.5 cm diameter copper tube that is 0. we will assume a bulk oil temperature of 300 K. ﬂows through a 2. the bulk temperature of the oil and the wall temperature will be assumed to be 300 and 372 K. The Reynolds number.20.61 m long. respectively.2 Forced Convection for Internal Flow 309 The following examples illustrate the use of some of the expressions presented in this section. The heat-transfer coefﬁcient on the oil side can then be determined using equation (20-27) k k D 0:33 mb 0:14 hi ¼ NuD ¼ 1:86 Pe D D L mw Initially. with a mean velocity of 0. the heat transfer rate can be expressed as q¼ Asurf (Tstm Toil ) ¼ rAvc p (TL To ) 1/hi þ 1/ho To get an indication of whether the ﬂow is laminar or turbulent. .

27-cm-ID tube at a velocity of 24 m/s.310 Chapter 20 Convective Heat-Transfer Correlations With an exiting temperature of 304 K the heat-transfer rate to the oil is q ¼ rAvc p (TL To ) p (0:025 m)2 (0:05 m/s)(1897 J/kgK)(9 K) ¼ (843 kg/m ) 4 ¼ 353 W 3 EXAMPLE 3 Air at 1 atmosphere and a temperature of 290 K enters a 1.. G. A. or (20-30) may be used. Deissler23 has 23 R. Equation (20-29) will be used.E. Re ¼ Dv (0:0127 m)(24 m/s) ¼ ¼ 20 600 v 1:478 105 Pa s Flow is clearly turbulent and Re is sufﬁciently large that equation (20-28). We now have h ¼ 0:023 Re0:2 Pr2/3 rvc p (0:0127)(24) 0:2 ¼ 0:023 (0:697)2/3 2:05 105 ¼ 0:00428 St ¼ Substituting into equation (19-61) we have TL Ts 1:52 m (0:00428) ¼ exp 4 0:0127 m To Ts ¼ 0:129 and the calculated value of TL is TL ¼ 372 (0:129)(372 290) ¼ 361 K This value agrees closely with the initially assumed value for TL. Deissler.M. Trans. Evaluate the convective heat-transfer coefﬁcient for this situation if the tube is 1. 1221 (1955). The wall temperature is maintained at 372 K by condensing steam. An exit temperature of 360 K will be assumed. The heat-transfer coefﬁcient is now evaluated as h ¼ rvc p St ¼ (1:012 kg/m3 )(24 m/s)(1009 J/kgK)(0:00428) ¼ 105 W/m2 K For ﬂow in short passages the correlations presented thus far must be modiﬁed to account for variable velocity and temperature proﬁles along the axis of ﬂow. we ﬁrst evaluate the Reynolds number at the tube entrance. . so there is no need to perform a second calculation. (20-29). the corresponding mean bulk temperature is 325 K and Tf ¼ 349. As in the previous example. it will be necessary to evaluate the exiting air temperature by ln TL Ts L h þ4 ¼0 To Ts D rvc p (19-62) To determine the type of ﬂow.S.52 m long. 77.

an application of the Colburn analogy Cfx (19-37) Stx Pr2/3 ¼ 2 along with equation (12-73) yields 1/ 3 Nux ¼ 0:0288 Re4=5 (20-33) x Pr A mean Nusselt number can be calculated using this expression for Nux.E.20. 74. This condition is amenable to analysis and has already been discussed in Chapter 19. The following equations may be used to modify the heat-transfer coefﬁcients in passages for which L/D < 60: for 2 < L/D < 20 hL ¼ 1 þ (D/L)0:7 h1 and for 20 < L/D < 60 hL ¼ 1 þ 6 D/L h1 (20-31) (20-32) Both of these expressions are approximations relating the appropriate coefﬁcient. The analysis of such ﬂow and of heat transfer in these situations is complicated when the phenomenon of boundary-layer separation is encountered. 20.S. Cylinders in Crossﬂow. the boundary layer ﬂow regimes are laminar for Rex < 2 105 and turbulent for 3 106 < Rex. The resulting expression is 4=5 (20-34) NuL ¼ 0:036 ReL Pr1/3 Fluid properties should be evaluated at the film temperature when using these equations.3 Forced Convection for External Flow 311 analyzed this region extensively for the case of turbulent ﬂow. For the laminar range and Nux ¼ 0:332 Rex1/2 Pr1/3 (19-25) NuL ¼ 0:664 Re1L/2 Pr1/3 (19-26) With turbulent flow in the boundary layer. hL.. We recall that. G. 343 (1952). The signiﬁcant results are repeated here for completeness. The reader will recall the nature of momentum-transfer phenomena discussed in Chapter 12 relative to external ﬂow.3 FORCED CONVECTION FOR EXTERNAL FLOW Numerous situations exist in practice in which one is interested in analyzing or describing heat transfer associated with the ﬂow of a ﬂuid past the exterior surface of a solid. A. The sphere and cylinder are the shapes of greatest engineering interest.M. such a condition will exist for most situations of engineering interest. Eckert and Soehngen24 evaluated local Nusselt numbers at various positions on a cylindrical surface past which ﬂowed an air stream with a range 24 E. in this case. where h1 is the value calculated for L/D > 60. Flow Parallel to Plane Surfaces. Soehngen. with heat transfer between these surfaces and a ﬂuid in crossﬂow frequently encountered. . Trans. R. Eckert and E. in terms of h1. Separation will occur in those cases in which an adverse pressure gradient exists.

7 and 20. 1949. as illustrated in Figure 20. which is greater than the stagnation-point value. . A. in practice. A second effect is that the Nusselt number reaches a value that is higher than the stagnation-point value. In the bottom curves of Figure 20. 378 (1949). McAdams26 has plotted the data of 13 separate 25 26 W. At low Reynolds numbers. Eckert and E. A much higher Reynolds number range was investigated by Giedt25 whose results are shown in Figure 20. Trans. the ﬁlm coefﬁcient reaches a second maximum. New York.7. thus less of the cylinder is engulfed in the wake. H. H.E.7.M. Third Edition. 71.8. the laminar boundary layer separates from the cylinder near 808 from the stagnation point. The increase is due to the greater conductance of the turbulent boundary layer. At higher Reynolds numbers.E. complex manner in external ﬂow about a cylinder. the ﬁlm coefﬁcient decreases almost continuously from the stagnation point. Trans. It is likely.) Figures 20. W. that an average h for the entire cylinder is desired. Soehngen.. 346 (1952). McGraw-Hill Book Company.M.8 show a smooth variation in the Nusselt number near the stagnation point. It is quite apparent from the ﬁgures that the convective heat-transfer coefﬁcient varies in an irregular. 74. By permission of the publishers. Heat Transmission. First.7 Local numbers for crossﬂow about a circular cylinder at low Reynolds numbers. Giedt.S. (From E. R. 100 80 Top 60 120 140 40 160 20 Direction of flow Nu 0 20 16 12 8 180 4 23 85 172 218 340 200 289 ReD 597 220 320 Bottom 300 280 240 260 Figure 20. and no large change in the Nusselt number occurs. the only departure being a slight rise in the separatedwake region of the cylinder. G. McAdams. Their results are shown in Figure 20. The second peak in the Nusselt number at high Reynolds numbers is due to the fact that the boundary layer undergoes transition from laminar to turbulent ﬂow.312 Chapter 20 Convective Heat-Transfer Correlations in Reynolds numbers from 20 to 600.S.. A.7. The effect of higher Reynolds number is twofold. the separation point moves past 908 as the boundary layer becomes turbulent.

This correlation is expressed in equation (20-36). 71. Note that values of NuD are for Pr ¼ 1. Pr1/3. " 5/8 #4/5 1/2 Pr1/3 0:62 ReD ReD NuD ¼ 0:3 þ h i1/4 1 þ 282.3 Forced Convection for External Flow 313 800 700 600 ReD = 219. should be employed. ReD.300 300 70.M. A widely used correlation for these data is of the form NuD ¼ B Ren Pr1/3 (20-35) where the constants B and n are functions of the Reynolds number. Churchill and M. A. His plot is reproduced in Figure 20.000 170. W. Trans. Giedt. Bernstein.20.000 500 186. 378 (1949).000 Nuq 400 101. 99. 300 (1977). For other ﬂuids a correction factor. The film temperature is appropriate for physical property evaluation.9. NuD ¼ NuD(figure) Pr1/3 . Values for these constants are given in Table 20.800 200 100 0 0 40 80 120 q-Degrees from stagnation point 160 Figure 20.S.3.000 140. 000 1 þ (0:4/Pr)2/3 27 S. By permission of the publishers.. J. (20-36) . that is.H.8 Local Nusselt numbers for crossﬂow about a circular cylinder at high Reynolds numbers. Heat Transfer. (From W.E.) investigations for the ﬂow of air normal to single cylinders and found excellent agreement when plotted as NuD vs. Churchill and Bernstein27 have recommended a single correlating equation covering conditions for which ReD Pr > 0:2.

Re for ﬂow normal to single cylinders.1 1.314 Chapter 20 Convective Heat-Transfer Correlations 600 400 300 200 Kennely et al. McAdams.618 0. Small Griffiths and Awbery Benke 100 80 60 40 20 NuD 1. Kennely and Sanborn L.84 1. Trans..4 0.193 0. following the work of Cary.0 3 4 6 8 10 2 3 4 6 8 102 2 ReD 3 4 6 8 103 2 3 4 6 8 104 0. A.I. McAdams. 1954. A recent correlation proposed by Whitaker30 is recommended for the following conditions: 0:71 < Pr < 380.Ch.45 0.28 McAdams29 has plotted the data of several investigators relating NuD vs. W. 361 (1972).) Table 20. ReD for air ﬂowing past spheres. Note: These values are strictly valid when Pr ﬃ 1.11. 75. All properties are 28 29 30 J. V. Third Edition.. New York.8 0.5 245 520 2 1 0. Local convective heat-transfer coefﬁcients at various positions relative to the forward stagnation point for ﬂow past a sphere are plotted in Figure 20. A.000 250.J. 3:5 < ReD < 7:6 104 . H.1 r v D hD Re = ∞ m∞ hf f 2 3 4 6 8 2 1. His plot is duplicated in Figure 20. 259. King Huges Gibson Reiher Paltz and Starr Vornehm Hilpert Gouhkman et al. Heat Transmission.000 100.805 Single Spheres.1 15.330 0. cit.3 0. H.989 0.10.3 Values of B and n for use in equation (20-35) ReD 0. Whitaker.0 8 6 4 Coordinates of recommended curve Nu = 0.E.9 Nu vs.000 40. Cary.E.385 0. McGraw-Hill Book Company. (From W.000 2 3 4 6 8 105 2 3 Figure 20. op.83 5. .027 0. R.0 10 100 1000 10. p.6 0. 1:0 < m1 /ms < 3:2.911 0.7 56.S.M. S.4/–4 4–40 40–4000 4000–40.000–400.000 B n 0.683 0.466 0. 483 (1953). 18.

) Angle from stagnation point.3 Forced Convection for External Flow 315 16 14 Re = 150. A. 485 (1953).000 58.000 89.000 h.26 1.0 10 102 103 rv D ∞ ReD = m 104 105 Figure 20.E.0–5. in atm Bider and Lahmeyer Loyzansky and Schwab Johnstone.D.0 7.10 Local heat-transfer coefﬁcients for ﬂow past a sphere. 75. (From McAdams.0 5. in cm 7.000 120.7–12. deg Key 1000 Observer Ds.11 NuD vs. ReD for air-ﬂow past single spheres.0 0. Pigford.0 Dorno 100 P.5 4. New York.055 1. and Chapin Theoretical line Recommended approx.S.5 1.M.0 Vyroubov 1–2 1. Trans. Third Edition. Cary. 0 0 30 60 90 120 150 180 (From J. McGraw-Hill Book Company.) . p.0 Johnstone. By permission of the publishers.033–0.31 ReD0.2 Buttner Meissner and Buttner 0. Heat Transmission. By permission of the publishers.8 1-15 0.0 Schmidt 7..0 7–15 1.6 1 1. 266.0 Kramers 0.24–15 1.5 1 V. R.000 44. and Chapin Nu = hD/kf 1. 1954.20. in Btu/ft2 hr °F 12 10 8 6 4 2 Figure 20. line Vyroubov 10 NuD = 0.9 1. Borne 5.71–1. Pigford.

Ranz and W. Several investigators have made signiﬁcant contributions to the analysis of these conﬁgurations. Delaware. O. 74. Engr. In addition to the greater Reynolds number range. Bulletin No. 2 (1950). This temperature indicates how cold the wind actually makes one feel. Bergelin. was used. In that ﬁgure all ﬂuid properties except mw are evaluated at the average bulk temperature. in which Deq. are presented in Figure 20. Wind–Chill Equivalent Temperature. . Colburn.. A. in addition to those factors already considered for ﬂow past single cylinders. Whitaker’s correlation is 1/ 2 þ 0:06 Re2D/3 )Pr0:4 (m1 /ms )1/4 NuD ¼ 2 þ (0:4 ReD (20-37) An important case is that of falling liquid drops.S. and Doberstein33 extended the work just mentioned for ﬁve of the tube arrangements to include values of Re up to 104. deﬁned by equation (20-37). as might be encountered in a heat exchanger. the tube diameter. Doberstein. G. ST is the center-to-center distance between tubes normal to the flow direction. One such term is the equivalent diameter of a tube bundle Deq. the colder the air feels to the human body. P. 48. P. Bergelin. Brown.12. For a complete explanation of the modeling used in determining this quantity. an indication of how cold one actually feels is expressed as the wind–chill equivalent temperature.13 involves Re calculated by using the tube diameter. Marshall. plotted as St Pr2/3 (mw /mb )0:14 vs.12. Trans. some investigators have chosen signiﬁcant lengths other than D. Re for various conﬁgurations.316 Chapter 20 Convective Heat-Transfer Correlations evaluated at T1 except for ms. P.E.M. O. Sta. D. which is the value at the surface temperature.13. Brown. deﬁned as Deq ¼ 4(SL ST pD2/4) pD (20-39) where SL is the center-to-center distance between tubes along the direction of flow. is 1/ 2 NuD ¼ 2 þ 0:6 ReD Re1/3 (20-38) Tube Banks in Crossﬂow. and H. 141 (1952). the stronger the wind blows. 953 (1952). Figure 20. as opposed to Figure 20. Bergelin. Re in Figure 20. C. to use in calculating Reynolds numbers. the effective heat-transfer coefﬁcient is affected by the tube arrangement and spacing. A. Eng. The determination of wind–chill equivalent temperature is an interesting example of the combined effects of convective and conductive heat transfer between the body and the adjacent air. When a number of tubes are placed together in a bank or bundle. Univ. and Hull32 studied the ﬂow of liquids past tube bundles in the region of laminar ﬂow with 1 < Re < 1000: Their results. 31 32 33 W. in addition to measured air temperatures. L. Hull. Their results are presented both for energy transfer and friction factor vs.31 for this case. Bergelin. and D is the OD of a tube. Colburn. For liquids in transition ﬂow across tube bundles. As ﬂuid ﬂow through and past tube bundles involves an irregular ﬂow path. and S. Expt. The correlation of Ranz and Marshall. A. modeled as spheres. The reader is undoubtedly familiar with weather reports where.. Chem. Progr.

50 Transition zone Model No.. A. By permission of the publishers. Hull.12 Convective heat-transfer exchange between liquids in laminar ﬂow and tube bundles.1 0.M. 6) 0. 1958 (1952). 7) (4) (1) (5) 0.S. P.14 1. Engr. 3 Oil Model No.01 1 10 100 1000 Re = GD m Figure 20. 5 flow 4 Oil Oil flow flow 2 Heat transfer 10–2 1 8 6 4 1 101 2 4 6 8 1 102 2 4 6 8 1 103 Re = Dt Gm m 2 4 6 8 1 104 Figure 20. and S. Brown. Bergelin. Bergelin.25 D.50 1. A. C. Station Bulletin 10. Dept. (From O. G.06 j' = k cpm 0. Colburn.08 0. By permission of the publishers.) 2 f mw mb 0.50 1.25 1.0 Tube arrangement equilateral triangle (4) equilateral triangle (1) staggered square (3) staggered square (7) In-line square (5) in-line square (2) in-line square (6) mw mb 0.25 1.0 0. and H.4 Rows 10 10 14 14 10 10 10 317 Pitch/D 1.2 0.25 1.02 (2. Doberstein.) . 2 Oil 2 flow 2/3 cpm h cpGm Rows 1 2 3 4 5 Oil 4 j' = Model 10–1 1 8 6 Model No.25 1.2. 4 Model No.25 1. A. (From O. p. L.13 Energy transfer and frictional loss for liquids in laminar and transition ﬂow past tube bundles. 74. 1950. 8. Univ.25 1.6 2/3 Forced Convection for External Flow 0. of Delaware.14 mw mb k Dt in. P.04 (3. P.8 0.3 2. Trans.25 1.E.14 4 100 1 8 6 0.50 1. in in..20. Pitch/Dt 10 10 14 10 10 3/8 3/8 3/8 3/8 3/8 1. 3/8 3/8 3/8 3/4 3/8 3/8 3/4 h cpGm 0. 1 flow Friction Model No.

App.5. 674–683 (1971). Steadman. Natural convection past vertical and horizontal surfaces.34 Tables 20. in any case. 34 R. Meteorol. The information included should. and tube bundles being the types of surfaces so considered. 20. allow most of the more common convection heattransfer coefﬁcients to be predicted with some conﬁdence.4 Wind–chill equivalent temperature—english units 5 10 15 20 25 30 35 40 35 30 25 20 15 32 22 16 12 8 6 4 3 27 16 9 4 1 2 4 5 22 10 2 3 7 10 12 13 16 3 5 10 15 18 20 21 11 3 11 17 22 25 27 29 Air temperature (8F) 10 5 6 9 18 24 29 33 35 37 0 15 25 31 36 41 43 45 0 5 10 15 20 25 5 22 31 39 44 49 52 53 10 27 38 46 51 56 58 60 15 34 45 53 59 64 67 69 21 40 51 60 66 71 74 76 26 46 58 67 74 79 82 84 31 52 65 74 81 86 89 92 20 24 28 32 36 40 20 26 36 43 47 49 51 24 31 42 48 53 56 58 28 35 47 54 59 62 64 32 40 52 60 65 68 70 36 44 57 65 71 74 77 40 49 63 71 77 80 83 Wind speed (km/hr) Table 20. . provide values of the wind–chill equivalent temperature as a function of air temperature and wind speed. Those graphs and equations presented are a small part of the information of this type available in the literature. 2. plus some useful simpliﬁed expressions for air. J. Indices of windchill of clothed persons. spheres.4 CLOSURE Many of the more useful experimentally developed correlations for predicting convective heat-transfer coefﬁcients have been presented in this chapter.318 Chapter 20 Convective Heat-Transfer Correlations Wind speed (mi/hr) Table 20. in English and SI units. respectively. including laminar and turbulent ﬂow correlations. The convection phenomena considered have included the following: 1.4 and 20.5 Wind–chill equivalent temperature—SI units 0 10 20 30 40 50 60 8 4 0 4 8 5 0 3 5 6 7 4 0 5 8 11 12 13 0 4 10 14 17 18 19 4 8 15 20 23 25 26 Air temperature (8C) 8 12 16 8 13 21 25 29 31 32 12 17 26 31 35 37 39 16 22 31 37 41 43 45 the reader is referred to a 1971 paper by Steadman. Forced convection for internal ﬂow. 10. Forced convection for external ﬂow with cylinders. G. and 3.

vertical.58C and then plunged into a 18C liquid bath.13 A ‘‘swimming-pool’’ nuclear reactor. 20.14. and the collector surface temperature is 1508F.8 Rubber balls are molded into spheres and cured at 360 K. apart. 20. What fraction of incident solar energy is lost by convection to the stagnant surrounding air at a temperature of 508F? What effect on the convective losses would result if the collector were crisscrossed with ridges spaced 1 ft apart? 20. What will be the surface temperature when the center temperature reaches 320 K? a.1 if the stagnant liquid is 20. and 1. 20.12 A 0. How long will it take for the center of the cylinder to reach 1008F? c. 20. 20. what is the maximum power level at which the reactor may operate? 20.10 if the medium surrounding the tube is stagnant water at 608F. is plunged vertically into a large tank of water at 508F. copper cylinder. vertically? The tube length is 10 ft. Determine the rate of heat transfer from the bulb by natural convection. What is the surface temperature when the center temperature is 1008F? Heat transfer from ends of the cylinder may be neglected.3 cm long with a diameter of 2. Under these conditions the surface temperature of the glass is 1408C. The tank is constructed of stainless steel 0.. is immersed in water at 808F. a. horizontal.14 A solar energy collector measuring 20 20 ft is installed on a roof in a horizontal position. When heated to a uniform temperature of 32. in length is placed in 958F stagnant water. which is suspended horizontally in still air in which heat is dissipated by the cable at a rate of 27 W per meter of length. 20.1 m/s. Determine the rate of heat gain if the tank is surrounded by air at 278 K. at a uniform temperature of 2008F. 25 cm in diameter. r ¼ 1120 kg/m3 . Note that your result is not equal to 100 W. is illuminated in air at 258C and atmospheric pressure. the center temperature of the cylinder reaches a value of 4.1 A 750-W immersion heater in the form of a cylinder with 3/4-in. spaced 21/2 in. Properties of rubber that may be used are k ¼ 0:24 W/m K. and what is the allowable Prandtl and Reynolds number range for a given set of data. Following this operation they are allowed to cool in room air. 20.7 A copper cylinder 20. and is oriented horizontally. a. The bulb is cylindrical. The incident solar energy ﬂux is 200 Btu/h ft2 . How long will it take for the outside surface of the cylinder to reach 1008F? b. 5. bismuth at 7008F. what heat ﬂux will be achieved if the tube is oriented a.Problems 319 The reader is reminded to observe any special considerations relative to the equations and plots in this chapter.6-m-diameter spherical tank contains liquid oxygen at 78 K. the 16-cm dimension vertical. Calculate the surface temperature of the heater if it is oriented with its axis Assuming the heat exchange between the cylinder and water bath to be purely by convection. Where does the remaining energy go? 20.15 Given the conditions for Problem 20.5 cm. This tank is covered with 5 cm of glass wool. cp ¼ 1020 J/kgK: 20.5. If this tube is located in still air at 608F. What will be the elapsed time for the surface temperature of a solid rubber ball to reach 320 K when the surrounding air temperature is 295 K? Consider balls with diameters of 7.11 Solve Problem 20.8 m. what value for the surface coefﬁcient is indicated? a. having a diameter of 35 mm and a length of 0. the 10-cm dimension vertical. horizontally? b. If 2008F is the maximum allowable plate temperature. The air temperature is 308C. rated at 1000 W. 20. what signiﬁcant length is used in a given correlation.6 Determine the steady-state surface temperature of an electric cable.32 cm thick.54 cm is being used to evaluate the surface coefﬁcient in a laboratory experiment. Such considerations include whether to evaluate ﬂuid properties at the bulk or ﬁlm temperature. diameter and 6 in. in length.10 A 1-in. 6 in. b.3 An immersion heater. 16-BWG copper tube has its outside surface maintained at 2408F. b. consisting of 30 rectangular plates measuring 1 ft in width and 3 ft in height.88C in 3 min. is in the form of a rectangular solid with dimensions 16 cm by 10 cm by 1 cm. b. determine the fraction of incident solar energy lost by convection to the surrounding air at 283 K ﬂowing parallel to the collector surface at a velocity of 6.2 Repeat Problem 20.4 A 2-in. PROBLEMS 20. hydraulic ﬂuid at 08F. . rated at 100 W.8 to reach the condition such that the center temperature is 320 K.9 Determine the required time for the rubber balls described in Problem 20. Determine the surface temperature of the heater if it is oriented in 295 K water with 20.5 A ﬂuorescent light bulb.

41 if. What will be the top surface temperature? 20.22 must be added to the steam pipe of Problem 20. a. The tubes are 5 ft long. construct a plot of convective heat-transfer coefﬁcient vs. 20.20 if the bare pipe is located so that 295 K air ﬂows normal to the pipe axis at a velocity of 6.1 bar condenses on the outside of a copper tube having inner and outer diameters of 16.-OD tubes in a staggered array with tubes arranged in equilateral triangle fashion having a pitchto-diameter ratio of 1. estimate the rate of steam condensed per meter of tube length.31 Solve Problem 20. If water at 1608F ﬂows at 20 ft/s past the tubes with constant surface temperature of 2128F. a.20 If the steam line described in Problem 20.65-cm layer of insulating material having a thermal conductivity of 0:242 W/mK: The air adjacent to the insulation is at 290 K. respectively.17 Given the information in Problem 20.32.27 Solve Problem 20. What is the temperature at the outside surface of the insulation under these conditions? 20. If the wire carries a current of 400 A.5 cm long. 16-BWG copper tube.5 cm is covered with a 0.21 Solve Problem 20.33 For the heater consisting of the tube bank described in Problem 20. respectively. room air is blown across the top with a velocity of 20 m/s.5 and 19 mm. What will be the outside surface temperature of the insulation? 20. The heater has a diameter of 1.29 A 1-in. evaluate the heat transferred to the water if the tube array consists of six rows of tubes in the ﬂow direction with eight tubes per row. 20. b. All heat transfer is from the lateral surface. when ﬁred at houses and ships.40 in order that the outside temperature of the insulation does not exceed 2508F? 20. normal to the tube axis. temperature for values of Tsurface between 420 and 1300 K. Air at 608F and atmospheric pressure is forced past this tube with a velocity of 40 fps. 20. k ¼ 39:8 W/mK cp ¼ 4:8 J/kgK r ¼ 7370 kg/m3 20. 20. Find the temperature of the air after passing through 20 ft of tubing if its entering velocity is 40 fps. determine a.320 Chapter 20 Convective Heat-Transfer Correlations 20. What will be the rate of heat loss under this new operating condition? b. and heat is transferred from the top surface at 40 W. 20. b. Neglect radiation from the insulation. Determine the heat ﬂux from the tube to the air if the ﬂow of air is a. 20..29 if the medium ﬂowing past the tube in forced convection is MIL-M-5606 hydraulic ﬂuid. has its outside surface maintained at 2408F.5 m/s. in W.16. The surface of the heater is at 380 K and the water is at 295 K. At this condition the inside oven temperature and room air temperature are 180 and 208C.20 if 3 in.5. 20. of insulation having a thermal conductivity of 0:060 Btu/h ft F is applied to the outside of the pipe.26 Saturated steam at 0. instead of six tubes. Conditions inside the oven may be considered unchanged. How long would it take for the surface temperature of a cannonball to reach 600 K? What would be its center temperature at this time? 20. the temperatures at the insulation-copper interface and at the outside surface of the insulation.25 A cooking oven has a top surface temperature of 458C when exposed to still air.32 An industrial heater is composed of a tube bundle consisting of horizontal 3/8-in. If one of these the socalled ‘‘hot shot’’ with a 15-cm diameter were at a uniform temperature of 1300 K.29 if the medium ﬂowing past the tube in forced convection is water at 608F.23 What thickness of insulation having a thermal conductivity as given in Problem 20. energy exchange is by natural convection with the heater axis oriented horizontally.16 Cast-iron cannonballs used in the War of 1812 were occasionally heated for some extended time so that.22 Solve Problem 20. 16-BWG copper tube whose surface is maintained at 2408F by condensing steam. what thickness will be necessary if radiation losses from the outside surface of the insulation account for no more than 15% of the total? The surroundings may be considered black at 708F. they would set them aﬁre..34 Heat transfer between an electrically heated circular cylinder and water is to be examined for three different conditions. as required by safety regulations. respectively.24 If insulation having a thermal conductivity of 0:060 Btu/h ft F is added to the outside of the steam pipe described in Problem 20. . parallel to the tube. Evaluate the heat transfer. what total heat transfer would be predicted from a 20-ft length of bare pipe? Consider the bare pipe to be a black surface and the surroundings black at 708F. three tubes each 5 m in length are used to heat the oil.19 Work Problem for an aluminum conductor of the same size ðresistivity of aluminum ¼ 2:83 106 ohm-cmÞ: 20. When the mean water temperature is 28 K.26 cm and is 7.18 Copper wire with a diameter of 0. b. what will be the effective heat-transfer coefﬁcient? 20. the convective heat-transfer coefﬁcient between the insulation surface and the surrounding air. The latent heat of condensation of steam may be taken as 2:390 kJ/kg: 20. energy exchange is by natural convection with heater axis oriented vertically. 20. what heat ﬂux value would exist if it were suddenly placed in still air at 270 K? The following properties of cast iron may be used: To reduce the surface temperature.40.40 is bare and surrounded by still air at 708F. 10 ft long. The surface coefﬁcients on the inner (water) surface and outer (steam) surface are 5200 and 6800 W/m2 K.30 Solve Problem 20. under conditions where 20.28 Air at 608F and atmospheric pressure ﬂows inside a 1-in.

Water at 290 K is circulated through the tube. J/kgK k.0080 0.18 if a fan provided an air ﬂow normal to the conductor axis at a velocity of 9 m/s? 20.-ID tube whose surface is held at a constant temperature of 1208F. Using this apparatus. and the inside wall of the 1/2-in. 321 20. Inside and outside surface coefﬁcients are 1700 and 8500 W/m2 K.133 0.45 A system for heating water with an inlet temperature of 258C to an exiting temperature of 708C involves passing the water through a thick-walled tube with inner and outer diameters of 25 and 45 mm. . W/mK v.394 2. blood. 20.133 0. The following information is pertinent: tube OD ¼ 1:905 cm tube ID ¼ 1:656 cm steam saturation temperature ¼ 319:5 K steam latent heat. respectively. the water velocity is 35 fps. What is the heat loss per 5 ft of water line in this case if the water and pipe temperatures are the same as speciﬁed in Problem 20.47 Air is transported through a rectangular duct measuring 2 ft by 4 ft. consists of a coiled tube that is immersed in an ice bath. What is the total heat loss through 5 ft of water line under these conditions? What is the exit water temperature? 20.43 An apparatus. The bulb can be modeled as a sphere with a diameter of 7.0203 0. 20. The outer tube surface is well insulated. six tubes. The tubes are 25.06 fps. K kg/m3 c p. The inside diameter of the tube is 2.38 A valve on a hot-water line is opened just enough to allow a ﬂow of 0. The air enters at 1208F and ﬂows with a mass velocity of 6 lbm /sft2 : If the duct walls are at a temperature of 808F. The tubes are 5 m long and the water enters at 290 K. m2 /s 103 Pr 373 393 413 433 842 831 817 808 2.5 mm.5 cm. and the surface coefﬁcient between the ice bath and outer tube surface is 500 W/m2 K: The thermal resistance of the tube wall may be neglected. If the oil enters at 1608C.130 0. W/mK m.42 Engine oil with properties given below ﬂows at a rate of 136 kg per hour through a 7. Determine the heat transfer from the bulb by the mechanism of forced convection.35 Solve Problem 20. 20.37 An electric light bulb. hfg ¼ 2393 kJ=kg: 20. schedule140 steel pipe at a rate of 10.0228 7:89 103 3:72 103 Determine the rate of heat transfer to the oil.44 A 1.4 mm in diameter. each 2.5-cm-ID pipe whose inside surface is maintained at 1008C.128 0. has a surface temperature of 1458C when cooled by atmospheric air at 258C. 20. K r. used in an operating room to cool blood.131 0.5 m/s.39 When the valve on the water line in Problem 20.13 kPa pressure.38? 20.47 kg/s. what will its temperature be at the exit of a 15-m-long pipe? T. what is the local convective coefﬁcient at this location? 20. how long must a. is to be cooled from 40 to 308C. For a mass ﬂow rate of water.5 m/s. OD ¼ 2:67 cm: The oil ﬂow rate is 1. If the inner surface of the tube at the outlet is Ts ¼ 110 C. The properties of oil to be used are as follows: T.16 with all conditions as given except that the ‘‘hot shot’’ is traveling through the air at 270 K with a velocity of 150 m/s.137 0. respectively. What is the length of the tube required for an air velocity of 25 fps? At 15 fps? 20. Find the rate of steam condensed per hour per meter of the tube length under these conditions. ﬂowing at 0:006 m3 /h.0124 0. m the tube be to achieve the desired outlet temperature? b. rated at 60 W.0414 0.38 is opened wide. The air ﬂows past the bulb with a velocity of 0.135 0.48 Cooling water ﬂows through thin-walled tubes in a condenser with a velocity of 1. schedule-r0 water line is at 808F. J/kgK k. 000 lbm /h: Estimate the value of h that applies at the inside pipe surface.41 Oil at 300 K is heated by steam condensing at 372 K on the outside of steel pipes with ID ¼ 2:09 cm.905-cm-diameter brass tube is used to condense steam on its outer surface at 10. Properties of blood are the following: r ¼ 1000 kg/m2 k ¼ 0:5 W/mK c p ¼ 4:0 kJ/kgK v ¼ 7 107 m2 /s Determine the total heat transfer to the oil and its temperature at the heater exit.5 m long. kg/m3 c p . The tube-wall temperature is maintained constant at 370 K by condensing steam on the outer surface.132 0. 3/4-in. 8008F ﬂows through a 1-in.482 0.40 Steam at 400 psi. and the electrical heating within the tube wall provides for a uniform generation of q˙ ¼ 1:5 106 W/m3 : ˙ ¼ 0:12 kg/s.46 Air at 25 psia is to be heated from 60 to 1008F in a smooth.306 2. Pas 300 310 340 370 910 897 870 865 1:84 103 1:92 103 2:00 103 2:13 103 0.36 What would be the results of Problem 20. energy exchange is by forced convection with water ﬂowing across the lateral surface at 1:5 m/s: 20. are used.0056 276 175 116 84 r.219 2. Determine the required length of tubing to accomplish the desired cooling. how much heat is lost by the air per foot of duct length? What is the corresponding temperature decrease of the air per foot? 20. 20. The water is maintained at 1808F.Problems c.

and the tubes are 1.49 Air.22 m and its walls are at 300 K. Estimate the heattransfer coefﬁcient.30 cm in outside diameter at ST ¼ SL ¼ 1:625 cm.61 m by 1. 20. Determine the rate of heat loss by the air per meter of duct length and the corresponding decrease in air temperature per meter.51 for a staggered arrangement. enters a rectangular duct with a mass velocity of 29:4 kg/sm2 : The duct measures 0.322 Chapter 20 Convective Heat-Transfer Correlations Estimate the exiting water temperature and the heattransfer rate per tube. and the tubes are 1. The tube bank is eight rows deep.51 A tube bank employs an in-line arrangement with ST ¼ SL ¼ 3:2 cm and tubes that are 1. . There are eight rows of tubes.5 and 15 cm in cross section. ﬂows normal to the tubes with a free stream velocity of 1. which are held at a surface temperature of 858C. There 20. what is the required ﬂow rate of the air? 20. at 322 K. Air. All other conditions remain the same.8 m long. at atmospheric pressure and a bulk temperature 278C. 20. which are held at a surface temperature of 908C.54 Rework Problem 20. 20. If the exiting air temperature is to be 308C.8 cm in outside diameter. are 10 rows of tubes.25 m/s. Determine the amount of heat transferred.52 Rework Problem 20.8 m long. The tube bank is 10 rows deep.53 for a staggered arrangement. 20.53 A tube bank employs tubes that are 1. The surfaces are maintained at 708C by solar irradiation. Air at atmospheric pressure and 208C ﬂows normal to the tubes with a free stream velocity of 6 m/s.50 Air at atmospheric pressure and 108C enters a rectangular duct that is 6 m long having dimensions of 7.

Natural convection currents circulate the superheated liquid. These phenomena. and the boiling surface may itself be a portion of the ﬂow passage. Extremely high heat ﬂuxes may be achieved in conjunction with boiling phenomena. These bubbles form at certain surface sites. An increase in wire temperature is accompanied by the formation of vapor bubbles on the wire surface.Chapter 21 Boiling and Condensation Energy-transfer processes associated with the phenomena of boiling and condensation may achieve relatively high heat-transfer rates. associated with the change in phase between a liquid and a vapor. whereas the accompanying temperature differences may be quite small. Another is the cooling of electronic devices where space is very critical. This is due to the additional considerations of surface tension. latent heat of vaporization. Pool boiling occurs on a heated surface submerged in a liquid pool that is not agitated. and other properties of two-phase systems that were not involved in the earlier considerations. One such application is the cooling of nuclear reactors. where vapor bubble nuclei are 323 .1. In regime I. making the application particularly valuable where a small amount of space is available to accomplish a relatively large energy transfer. There are six different regimes of boiling associated with the behavior exhibited in this ﬁgure. The advent of these applications has spurred the interest in boiling. These phenomena will be considered separately in the following sections. The processes of boiling and condensation deal with opposite effects relative to the change in phase between a liquid and its vapor. the wire surface temperature is only a few degrees higher than that of the surrounding saturated liquid. the temperature difference between the heated surface and saturated water is depicted in Figure 21. The ﬂow of liquid and vapor associated with ﬂow boiling is an important type of two-phase ﬂow. Flow boiling occurs in a ﬂowing stream. and concentrated research in this area in recent years has shed much light on the mechanism and behavior of the boiling phenomenon There are two basic types of boiling: pool boiling and ﬂow boiling. 21. surface characteristics.1 BOILING Boiling heat transfer is associated with a change in phase from liquid to vapor. An electrically heated horizontal wire submerged in a pool of water at its saturation temperature is a convenient system to illustrate the regimes of boiling heat transfer. and evaporation occurs at the free liquid surface as the superheated liquid reaches it. are more involved and thus more difﬁcult to describe than the convective heat-transfer processes discussed in the preceding chapters. A plot of the heat ﬂux associated with such a system as the ordinate vs. Regimes of Boiling.

1 Pool boiling in water on a horizontal wire at atmospheric pressure. and the heat ﬂux curve rises once more. Regimes II and III are associated with nucleate boiling. This is region IV on the curve. This condition continues until point b is reached. electrical heating is used. present. DT increases through regions I. break away from the wire surface. If. the heat ﬂux decreases. Note the somewhat anomalous behavior exhibited by the heat ﬂux associated with boiling. At a still higher wire surface temperature. Point a on the curve is often referred to as the ‘‘burnout point’’ for these reasons. then regime IV will probably not be obtained because of wire ‘‘burnout. however.1 can be achieved if the energy source is a condensing vapor. the heat ﬂux might be expected to increase continuously as the temperature difference between the . the stable ﬁlm-boiling regime. and III. a vapor ﬁlm forms around the wire. and reach the free surface. This ﬁlm collapse and reformation and this unstable nature of the ﬁlm is characteristic of the transition regime. II. and condense before reaching the free liquid surface. rise. larger and more numerous bubbles form. Other than normal gravitational effects are encountered in space vehicles. as in regime III.Boiling and Condensation IV VI Radiation coming into play Stable film boiling a V Partial nucleate boiling and unstable film Nucleate-boiling bubbles rise to interface III II Nucleate-boiling bubbles I Free convection Chapter 21 Log (q/A) 324 Boiling curve 0. the transition boiling regime is entered. the vapor ﬁlm around the wire becomes stable. °F 1000 10. When the peak value of q/A is exceeded slightly. The result is an increase in DT accompanied by a further decrease in the possible q/A. In this regime. This is the process occurring in regime II.000 Figure 21. thus. As DT at point b is extremely high. When present.1 1.1 The curve in Figure 21. One normally considers a ﬂux to be proportional to the driving force. thus. the wire will long since have reached its melting point. brieﬂy exposing a portion of the wire surface. Beyond the peak of this curve. the magnitude of the body-force intensity will affect both the mechanism and the magnitude of boiling heat transfer. the vapor ﬁlm provides a considerable resistance to heat transfer.’’ As the energy ﬂux is increased. When the surface temperature reaches a value of approximately 4008F above the saturated liquid. This is designated as region VI in Figure 21. This is region V. For surface temperatures of 10008F or greater above that of the saturated liquid. and portions of this ﬁlm break off and rise. break off. radiant energy transfer comes into play. the required amount of energy cannot be transferred by boiling. As the mechanism of energy removal is intimately associated with buoyant forces. rise.0 10 100 Tw – Tsat.

may be determined from Gb ¼ q/A hfg (21-4) where hfg is the latent heat of vaporization Rohsenow1 has used equation (21-1) to correlate Addoms’s2 pool-boiling data for a 0. Massachusetts Institute of Technology. PrL. the regime of partial nucleate boiling and partial natural convection.024-in. Gb. Chemical Engineering Department. the very high heat ﬂuxes associated with moderate temperature differences in the nucleate-boiling regime are much higher than the heat ﬂuxes resulting from much higher temperature differences in the ﬁlm-boiling regime. Db is the maximum bubble diameter as it leaves the surface.1.M. The reason for this is the presence of the vapor ﬁlm.S. The bubble Reynolds number. is of great engineering importance because of the very high heat ﬂuxes possible with moderate temperature differences. the most useful of these follow. Trans. 74.-diameter platinum wire immersed in water. and the results for each of these two regimes may be superposed to describe a process in regime II. the correlations presented in Chapter 20 for natural convection may be used. M. is the Prandtl number for the liquid. N. of course. 2 . Thesis. Ts Tsat is the excess temperature or the difference between the surface and saturatedliquid temperatures. Rohsenow..21. That data for this regime are correlated by equations of the form Nub ¼ f(Reb . which covers and insulates the heating surface in the latter case Correlations of Boiling Heat-Transfer Data. Reb. 1 W. In the natural convection regime.E. The nucleate-boiling regime.2 and is expressed in equation form as q g(rL rv ) 1/2 cpL (Ts Tsat ) 3 ¼ mL hfg (21-5) A s Csf hfg Pr1:7 L where cpL is the heat capacity for the liquid and the other terms have their usual meanings. 969 (1952). is defined as Reb Db Gb mL (21-3) where Gb is the average mass velocity of the vapor leaving the surface and mL is the liquid viscosity The mass velocity. Addoms. A. and kL is the thermal conductivity of the liquid. As the ﬂuid behavior in a boiling situation is very difﬁcult to describe. June 1948. The quantity. regime III. is not the case. there is no adequate analytical solution available for boiling transfer. Various correlations of experimental data have been achieved for the different boiling regimes. is a combination of regimes I and III. Regime II. D. PrL ) (21-1) The parameter Nub in equation (21-1) is a Nusselt number defined as Nub (q/A)Db (Ts Tsat )kL (21-2) where q/A is the total heat flux.1 325 Boiling heated surface and the saturated liquid increases. This. regime I of Figure 21. J. This correlation is shown in Figure 21.Sc.

J.006 0.0027 0.0030 3 W.1. and Momentum Transfer.015 0.1 0. From earlier discussion it is clear that the burnout point has considerable importance.013 0. By permission of the publishers. A table of Csf for various combinations of ﬂuid and surface is presented by Rohsenow and Choi3 and duplicated here as Table 21. 1961. Choi. M. Heat.7 10 Figure 21.1 cL 1 (Tw – Tsat) hfg PrL1.024 in.0054 0.. p. Table 21.010 0.J. Mass. N. Choi.) The coefﬁcients Csf in equation (21-5) vary with the surface–ﬂuid combination.7 100 10 1. 224. Rohsenow and H.013 0. cL (T – Tsat) = 0.Boiling and Condensation s g(rL – rv) 0.2 Correlation of poolboiling data.013 hfg w s g(rL – rv) Chapter 21 q/A mL hfg 326 14.1. and Momentum Transfer. Heat. . Inc.. diam...01 0. N. (From W.33 cL mL kL 1. Englewood Cliffs.7 psia 383 psia 770 psia 1205 psia 1602 psia 2465 psia 0.1 Values of Csf for equation (21-5) Surface/ﬂuid combination Csf Water/nickel Water/platinum Water/copper Water/brass CCl4/copper Benzene/chromium n-Pentane/chromium Ethyl alcohol/chromium Isopropyl alcohol/copper 35% K2CO3/copper 50% K2CO3/copper n-Butyl alcohol/copper 0. Inc.013 0. The curve drawn in Figure 21.2 is for Csf ¼ 0:013. Mass.0 q/A mL hfg Data of Addoms pool boiling Platinum wire–water 0.0027 0. Englewood Cliffs.006 0. M. Rohsenow and H.0025 0. Prentice-Hall. The ‘‘critical heat ﬂux’’ is the value of q/A represented by point a in Figure 21. Y. 1961. Prentice-Hall.

and no satisfactory correlation has been found for this region as yet. et al. Hsu9 states that heat-transfer rates for film boiling are higher for vertical tubes than for horizontal tubes when all other conditions remain the same..J. Trans. 45. few experimental data have been reported for this region. Buffalo. Eng Prog. however. Y. Do. . A. Engineering Heat Transfer. Bromley obtained the expression 3 kv rv (rL rv )g(hfg þ 0:4 c pL DT) 1/4 h ¼ 0:62 (21-7) Do mv (Ts Tsat ) where all terms are self-explanatory except Do. is not of great engineering interest. Zuber. Their test results were correlated by the equation 1/3 m2v h ¼ 0:0020 Re0:6 grv (rL rv )kv3 where Re ¼ 4m_ pDv mv (21-9) (21-10) m_ being the flow rate of vapor in lbm/h at the upper end of the tube and the other terms being identical to those in equation (21-7). In Berenson’s correlation. August 14–17. the tube diameter. Ind..I. Hsu.S. A. and the recommended expression is " 3 # kv f rvf (rL rv )g(hfg þ 0:4c pL DT) 1/4 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ h ¼ 0:425 (21-8) mvf (Ts Tsat ) s/g(rL rv ) where kvf. L. W. Van Nostrand.I. rvf. that of unstable ﬁlm boiling. A. Eng Chem. L. 1963. the superimposed contribution of radiation is appreciable. Without any appreciable ﬂow of liquid. Y.. 221 (1950). N.Y. 4.Ch. Paper No. Westwater. Princeton. 80. S.M. A modiﬁcation in equation (21-7) was proposed by Berenson7 to provide a similar correlation for stable ﬁlm boiling on a horizontal surface. 960. Bromley. In regime VI. requires high surface temperatures. 2639 (1953). regime V. A. Chem. A. N. Berenson. 711 (1958). 18. Heat Transfer Conference. (5). Bromley.. thus. P. becoming dominant at extremely high values of DT.21. 46. the correlations for ﬁlm boiling still apply. T. Stable ﬁlm boiling on the surface of horizontal tubes and vertical plates has been studied both analytically and experimentally by Bromley. Hsu and J. the two contributions may be combined.E. 4 5 6 7 8 9 N. which is the outside diameter of the tube.E.. Regime IV. as indicated by equation (21-11)..Ch.5.1 Boiling 327 An analysis of conditions at burnout modiﬁed by experimental results is expressed in equation 21-6 as sg(rL rv ) 1/4 q/Ajcritical ¼ 0:18hfg rv (21-6) r2v The interested reader is referred to the work of Zuber4 for a discussion of this subject..6 Considering conduction alone through the ﬁlm on a horizontal tube.E. J. Hsu and Westwater8 considered ﬁlm boiling for the case of a vertical tube. The stable-ﬁlm-boiling region. is replaced by the term ½s/g(rL rv )1/2 . and mvf are to be evaluated at the film temperature as indicated. 59 (1958).

Nusselt10 achieved an analytical result for the problem of ﬁlmwise condensation of a pure vapor on a vertical wall. . on the contrary. associated with the higher heat-transfer rates of the two types of condensation phenomena. and forms a ﬁlm. 21. After a condensate ﬁlm has been developed in ﬁlmwise condensation. When the liquid condensate forms on the surface. In this ﬁgure. Dropwise condensation is. then droplets form and run down the surface. It is evident that for vertical surfaces or large-diameter horizontal tubes the density difference between liquid and vapor will produce signiﬁcant local velocities. This process is designated dropwise condensation. Dropwise condensation is very difﬁcult to achieve or maintain commercially. always has some surface present as the condensate drop forms and runs off. Ver. z y x 1 d ∆x 10 Figure 21. 514 (1916).2 CONDENSATION Condensation occurs when a vapor contacts a surface that is at a temperature below the saturation temperature of the vapor. therefore. The interested reader is referred to the recent literature for pertinent discussion of these phenomena. In 1916. Zeitschr. If the surface is not wetted by the liquid. This term is discussed in Chapter 23. The meanings of the various terms in this analysis will be made clear by referring to Figure 21. coalescing as they contact other condensate droplets.3. and the associated energy transfer must occur by conduction through the condensate ﬁlm.. Any correlation that neglects ﬂow contributions should. Ing. Nusselt. Normally the liquid wets the surface. When there is appreciable ﬂow of either the liquid or the vapor. deutsch. The description of ﬂow boiling or two-phase ﬂow will not be discussed here. additional condensation will occur at the liquid–vapor interface. all equipment is designed on the basis of ﬁlmwise condensation.328 Chapter 21 Boiling and Condensation The contribution of radiation to the total heat-transfer coefﬁcient may be expressed as 1/3 hc þ hr (21-11) h ¼ hc h where h is the total heat-transfer coefficient. Dropwise condensation. therefore. and hr is an effective radiant heat-transfer coefficient considering exchange between two parallel planes with the liquid between assigned a value of unity for its emissivity. Film Condensation: The Nusselt Model. it will ﬂow under the inﬂuence of gravity. Such a process is called ﬁlm condensation. therefore. W. be used with caution. spreads out. d. hc is the coefficient for the boiling phenomenon. the foregoing correlations are unsatisfactory. 60.3 Filmwise condensation on a vertical plane wall.

the process may occur at a sufficiently high pressure that the vapor density.21. rv . Tsat. Relating these two effects. On this basis. dG. to the wall–liquid boundary at temperature. x ¼ 0. and to increase with increasing values of x. in the flow rate is evaluated from this expression to be dG ¼ (rL rv )gd2 dd d (21-14) This result has been obtained from momentum considerations alone. rL . d. the density of the gas or vapor at the liquid surface was neglected. we may write Z d qy (Tsat Tw ) 1 dG ¼ rL hfg þ ¼k rL v x cpL (Tsat T) dy d rL G 0 dx A which.2 Condensation 329 the ﬁlm thickness. To account for this possibility. sin u ¼ 1 and L ¼ d. It is also necessary to modify the density for the present case. the temperature proﬁle is linear and the heat ﬂux to the wall is qy (Tsat Tw ) ¼k d A (21-15) This same amount of energy must be transferred from the vapor as it condenses and then cools to the average liquid temperature. to be purely by conduction. the density function to be used in the present case is rL rv instead of simply rL. look at the related energy transfer. however. as Nusselt did originally. The resulting expression for the velocity profile in the condensate film at a particular distance x from the top of the wall becomes (r rv )gd2 y 1 y 2 vx ¼ L (21-12) d 2 d m The flow rate per unit width. is significant in comparison to that of the liquid. it is not unreasonable to consider energy transfer through the ﬁlm from the temperature at the vapor–liquid interface. at any value x > 0 is Z d v x dy G¼ 0 (r rv )gd3 ¼ L 3m (21-13) A differential change. the velocity proﬁle may be easily obtained from equation 8-12 rgL2 sin u y 1 y 2 vx ¼ (8-12) m L 2 L For the present application. We shall now. Under these conditions. Tw . is seen to be zero at the top of the vertical wall. As the ﬂow of condensate is assumed to be laminar. G. This may be true in many cases of a condensation process. becomes qy k(T sat Tw ) dG ¼ rL [hfg þ 38cpL (Tsat Tw )] ¼ d dx A (21-16) . The initial assumption made by Nusselt was that of wholly laminar ﬂow in the condensate ﬁlm. In the derivation of equation (8-12). if a linear temperature variation in y is utilized.

Example 1 illustrates such a case. Trans. Rohsenow11 performed a modiﬁed integral analysis of this same problem. M. An expression similar to equation (21-20) may be achieved for a surface inclined at an angle u from the horizontal if sin u is introduced into the bracketed term. 1645 (1956). in equation (21-20) and those preceding it. 78.E. . we obtain " # 1/4 4km(Tsat Tw )x d¼ rL g(rL rv )½hfg þ 38 cpL (Tsat Tw ) (21-18) We may now solve for the heat-transfer coefficient. from the expression h¼ qy /A k ¼ Tsat Tw d The substitution of equation (21-18) into this expression yields ( ) rL gk3 (rL rv )½hfg þ 38 c pL (Tsat Tw ) 1/4 hx ¼ 4m(Tsat Tw )x (21-19) The average heat-transfer coefficient for a surface of length L is determined from Z 1 L h¼ hx dx L 0 which. when equation (21-19) is substituted. Liquid properties should all be taken at the ﬁlm temperature. becomes ( ) rL gk3 (rL rv )½hfg þ 38 c pL (Tsat Tw ) 1/4 h ¼ 0:943 Lm(Tsat Tw ) (21-20) The latent heat term. giving (rL rv )g 2 k(T sat Tw ) dx d dd ¼ m rL d½hfg þ 38 cpL (Tsat Tw ) Simplifying this result and solving for d.S. This extension obviously has a limit and should not be used when u is small.. so the condensate that forms cannot ﬂow away. the analysis is quite simple. A. obtaining a result that differs only in that the term ½hfg þ 38 cpL (Tsat Tw ) is replaced by [hfg þ 0:68cpL (Tsat Tw )]. that is.M. Rohsenow’s results agree well with experimental data achieved for values of Pr > 0:5 and c pL (Tsat Tw )/hfg < 1:0: EXAMPLE A square pan with its bottom surface maintained at 350 K is exposed to water vapor at 1 atm pressure and 373 K. How deep will the condensate ﬁlm be after 10 min have elapsed at this condition? 11 W.330 Chapter 21 Boiling and Condensation Solving equation () for dG. when the surface is near horizontal. h. The pan has a lip all around. Rohsenow. should be evaluated at the saturation temperature. For such a condition. we have dG ¼ k(T sat Tw ) dx rL d½hfg þ 38 cpL (Tsat Tw ) (21-17) which may now be equated to the result in equation (21-20). hfg.

The criterion for turbulent ﬂow is. we have rhfg ¼ dd kL ¼ (Tsat Ts ) d dt and. we are assuming d to be constant. If d is not rapidly varying. may be expressed as follows: ˙ cond ¼ rV˙ cond ¼ rA m dd dt where dd/dt is the rate at which the condensate ﬁlm thickness. the condensate ﬁlm thickness is seen to vary with time according to dd kL (Tsat Ts ) ¼ dt rhfg Z d Z t kL d dd ¼ (Tsat Ts ) dt rhfg 0 0 1/2 2kL (Tsat Ts ) t1/2 d¼ rhfg d A quantitative answer to our example problem now yields the result 2(0:674 W/m K)(23 K)(600 s) 1/2 d¼ (966 kg/m3 )(2250 kJ/kg) ¼ 2:93 mm Film Condensation: Turbulent-Flow Analysis. Now. are related as ˙ cond hfg m q ¼ A A in The condensation rate.2 Condensation 331 We will employ a ‘‘pseudo-steady-state’’ approach to solve this problem. that is. m_ cond . the applicable Reynolds number is Re ¼ 4A rL v P mf (21-21) . The heat ﬂux at the interface may now be expressed as q dd ¼ rhfg A in dt This heat ﬂux is now equated to that which must be conducted through the ﬁlm to the cool pan surface. An energy balance at the vapor–liquid interface will indicate that the heat ﬂux and rate of mass condensed. m_ cond . It is logical to expect the ﬂow of the condensate ﬁlm to become turbulent for relatively long surfaces or for high condensation rates.21. a Reynolds number for the condensate ﬁlm. as we should expect. this ‘‘pseudo-steady-state’’ approximation will give satisfactory results. In terms of an equivalent diameter. progressing. grows. equating the two heat ﬂuxes. d. The heat ﬂux expression that applies is q kL ¼ (Tsat Ts ) d A out This is a steady-state expression.

Colburn formulated the plot shown in Figure 21. McAdams. deutsch. d being ﬁlm thickness and v avg the average velocity. Nusselt. Gc ¼ rL v avg d. Heat Transmission..0 0. The critical value of Re in this case is approximately 2000. 432 (1934). Ind. d. as h ¼ 0:0077 rL g(rL rv )kL3 1/3 0:4 Red m2L (21-24) Film Condensation: Analysis of the Horizontal Cylinder. H. Ind. 4 (1930). C. On the basis partly of analysis and partly of experiment.12 who used the same j factor determined for internal pipe ﬂow. New York. G.8 Pr = 1 Pr = 5 0. that is. Eng. P is the wetted perimeter. Red > 2000. W. Ing. 3rd edition.4 Film condensation including the regions of both laminar and turbulent ﬂows.. The term 4Gc /mf is thus a Reynolds number for a condensate ﬁlm on a plane vertical wall. McAdams14 recommends a simpler expression for the turbulent range. Gc is the mass ﬂow rate per unit width of surface. Zeitschr.4 m2 p2g Chapter 21 h k 332 0. Chem.Boiling and Condensation 1/3 1. where A is the condensate flow area. W. Chem. 1954.4. The ﬁrst attempt to analyze the case of turbulent ﬂow of a condensate ﬁlm was that of Colburn. 60. . The data points shown are those of Kirkbride. Eng. and v is the velocity of condensate. McGraw-Hill Book Company.1 100 2 4 8 1000 2 4 8 10000 2 4 8 100000 Re Figure 21.2 0.. P. 569 (1916). An analysis by Nusselt15 produced the following expression for the mean heat-transfer coefﬁcient for a horizontal cylinder: ( ) rL g(rL rv )k3 hfg þ 38 c pL (Tsat Tw ) 1/4 havg ¼ 0:725 (21-25) mD(T sat Tw ) 12 13 14 15 A. 26.13 The correlating equations for the two regions shown are for 4Gc /mf < 2000 k3 r2 g 1/3 4Gc 1/3 m2 f mf (21-22) (k3 r2 g/m2 )f1/3 (4Gc /mf )Pr1/3 i (4Gc /mf )4/5 364 þ 576 Pr1/3 (21-23) havg ¼ 1:51 and for 4Gc /mf > 2000 havg ¼ 0:045 h In these expressions. Colburn. Ver. Kirkbride. 26.

naturally. Film Condensation: Banks of Horizontal Tubes.21. as the condensate ﬁlm from one tube will drop on the next tube below it in the line. Nusselt also considered this situation analytically and achieved. a different value of h for each tube.86. the expression ( ) rL g(rL rv )k3 hfg þ 38 cpL (Tsat Tw ) 1/4 havg ¼ 0:725 (21-28) n Dm(Tsat Tw ) This equation yields a mean heat-transfer coefficient averaged over all n tubes. the horizontal position has the greater heat-transfer capability. for a vertical bank of n tubes in line.E.S.) Series C. Observing that experimental data exceeded thosevalues predicted from equation (21-28). Chen’s equation agrees reasonably well with experimental data for condensation on vertical banks of horizontal tubes. 83. For a bank of horizontal tubes there is. His resulting equation is cpL (Tsat Tw ) (n 1) havg ¼ 0:725 1 þ 0:02 hfg ( ) rL g(rL rv )k3 hfg þ 38 cpL (Tsat Tw ) 1/4 (21-29) n Dm(Tsat Tw ) which is valid for values of cpL (Tsat Tw )(n 1)/hfg > 2. A. Chen. 16 M. 48 (1961). For L/D values greater than 2. This process is depicted in Figure 21. M.5 Condensation on a horizontal tube bank.86. we obtain the result that 1/4 hvert 0:943 D 1/4 D ¼ ¼ 1:3 (21-26) L hhoriz 0:725 L For the case of equal heat-transfer coefficients. Chen16 modiﬁed this expression to include the effect of condensation on the liquid layer between the tubes. the relation between D and L is L ¼ 2:86 (21-27) D or. Condensate film Figure 21. (Trans. . Combining these expressions and canceling similar terms.M.5.2 Condensation 333 The similarity between equation (21-25) for a horizontal tube and equation (21-20) for a vertical tube is marked. equal amounts of energy can be transferred from the same tube in either the vertical or the horizontal position if the ratio L/D is 2.

Analytical solutions have been presented. copper? b. a needed quantity in several of the following problems. however.2 Plot values of the heat-transfer coefﬁcient for the case of pool boiling of water on horizontal metal surfaces at 1 atm total pressure and surface temperatures varying from 390 to 450 K. Boiling is normally described as nucleate type. or a combination of the two. its surface temperature rises above that of the adjacent saturated water. 21. The system pressure is maintained at 1 atm and the tube surface is held at 2808F.3 were initially heated to 5008F. Thus. the surface must not be ‘‘wetted’’ by the condensate. for ﬁlmwise condensation on vertical and horizontal plates and cylinders and for banks of horizontal cylinders. is related to temperature according to the expression s ¼ 0:1232[1 0:00146 T]. it is difﬁcult to maintain this condition for several reasons. ﬁlm type. where s is in N/m and T is in K. Very high heat-transfer rates are possible in the nucleate-boiling regime with relatively low temperature differences between the primary surface and the saturation temperature of the liquid. how long would it take for the center of the cylinder to cool to 2408F if it were constructed of a. Because of its uncertain nature and the conservative approach of a design based on lower heat-transfer coefﬁcients. then becoming a nucleate-boiling phenomenon at higher DTs. nickel? 21. PROBLEMS The surface tension of water. (c) platinum.5 Four immersion heaters in the shape of cylinders 15 cm long and 2 cm in diameter are immersed in a water bath at 1 atm total pressure. 21. Normally this requires that metal surfaces be specially treated. (b) copper. estimate the temperature of the heater . Condensation is categorized as either ﬁlmwise or dropwise. 21.3 CLOSURE The phenomena of boiling and condensation have been examined in this chapter. q/Ajboiling . Each heater is rated at 500 W.1 An electrically heated square plate measuring 20 cm on a side is immersed vertically in water at atmospheric pressure. in diameter is immersed in water. along with empirical results. In the English system with s given in lbf /ft and T in 8F. If the heaters operate at rated capacity. Several empirical correlations for these phenomena for various surfaces oriented in different ways have been presented. Each condition has a prominent place in engineering practice and both are difﬁcult to describe analytically. At low power levels the heat-transfer mechanism is natural convection. is associated with higher heat-transfer coefﬁcients than the ﬁlmwise condensation phenomenon. ﬁlmwise condensation is of primary interest. At present. Dropwise condensation is associated with much higher heat-transfer coefﬁcients than ﬁlmwise. At what value of DT are the heat ﬂuxes due to boiling and natural convection the same? Plot q/Ajconvection . For dropwise condensation to occur. as mentioned earlier.4 If the cylinder described in Problem 21. the surface tension may be calculated from s ¼ (8:44 103 )[1 0:00082 T]: 21.334 Chapter 21 Boiling and Condensation Drop Condensation. 21. Dropwise condensation. This anomalous behavior is peculiar to the boiling phenomenon. Determine the nucleate-boiling heat-transfer coefﬁcient and the rate of heat dissipation for this system. it is difﬁcult both to achieve and to maintain. ﬁlmwise condensation is the type predominantly used in design.3 A cylindrical copper heating element 2 ft long and 12 in. (d) brass. Dropwise condensation is an attractive phenomenon for applications where extremely large heat-transfer rates are desired. Consider the following metals: (a) nickel. brass? c. As the electrical energy supplied to the plate is increased. and q/Ajtotal versus DT values from 250 to 300 K. Film boiling is associated with a higher temperature difference yet a lower rate of heat transfer.

time up to 1 h for this situation. time.7 Estimate the heat-transfer rate per foot of length from a 0. ﬂowing in a pipe.5-mmID tube enters at 208C. 21. in diameter has its outside surface at a temperature of 12008F.15 A circular pan has its bottom surface maintained at 2008F and is situated in saturated steam at 2128F.16 are arranged in a vertical bank and the ﬂow is assumed laminar. c. Determine the rate of condensation and the heat transfer coefﬁcient for the case of a 1. the average heat-transfer coefﬁcient for the bank.14 Saturated steam at atmospheric pressure ﬂows at a rate of 0.042 kg/s/m between two vertical surfaces maintained at 340 K that are separated by a distance of 1 cm. If this pan is situated in saturated steam at 372 K. what heattransfer coefﬁcient will apply to the condensation process? What total heat transfer will occur? 21.02-in.16 Saturated steam at 365 K condenses on a 2-cm tube whose surface is maintained at 340 K.16. Evaluate the condensation rate if the pipe is oriented a. 21.-diameter pipe whose surface is at 1608F. 21.18 Determine the heat-transfer coefﬁcient for a horizontal 5-in. Construct a plot of condensate depth in the pan vs.20 Given the conditions of Problem 21. How many plates would you recommend? Substantiate all of the design criteria used. above the base is oriented with its base at an angle of 208 from the horizontal. The pan surface is kept at 1808F and it is situated in an atmosphere of 2108F steam.9 A steel plate is removed from a heat-treating operation at 600 K and is immediately immersed into a water bath at 373 K. The wire temperature is 22008F. 21.21 A vertical ﬂat surface 2 ft high is maintained at 608F. 21. and is 4 ft long. evaluate a.6 A horizontal circular cylinder 1 in. How tall may this conﬁguration be if the steam velocity is not to exceed 15 m/s? 21. inclined at 108 to the horizontal? c. The pipe has an inside diameter of 34 in.22 A square pan measuring 40 cm on a side and having a 2-cm-high lip on all sides has its surface maintained at 350 K. At 40 psi.8 Two thousand watts of electrical energy are to be dissipated through copper plates measuring 5 cm by 10 cm by 0. c. 335 21.23 A square pan with sides measuring 1 ft and a perpendicular lip extending 1 in. horizontally.14 are oriented horizontally in a vertical bank. 21. The sides of the pan may be considered nonconducting. b. b. third. 21. 21. The surface temperature is maintained at 918C. what heat-transfer rate will occur? 21. How long will it be before condensate spills over the lip of the pan? . 21. distance from the top of the pipe. vertically. Saturated atmospheric steam condenses on the outside of the tube. The tube outside diameter is 19 mm. vertically.19 If eight tubes of the size designated in Problem 21. inclined at 308 to the horizontal? 21.6 cm immersed in water at 390 K. horizontal? b. b. convective coefﬁcient on the condensate side.13 Water ﬂowing at a rate of 4000 kg/h through a 16. This tube is immersed in saturated water at a pressure of 40 psi. If the water is to be at 2128F throughout its residence in the pipe. the the the the convective coefﬁcient on the water side. For a horizontal brass tube 2 m long. the temperature of saturated water is 2678F. is to receive heat at a rate of 3 104 Btu/hft2 of pipe surface. diameter nichrome wire immersed in water at 2408F.5-m-long tube oriented a.10 Water. plate temperature for this system.12 Saturated steam at atmospheric pressure condenses on the outside surface of a 1-m-long tube with 150 mm diameter. What is the convective heat-transfer coefﬁcient in this case? 21. b. Construct a plot of convective heat-transfer coefﬁcient vs. a. exit water temperature. determine a. Estimate the heat ﬂux due to ﬁlm boiling that may be achieved with this conﬁguration. For a mild-steel plate 3 cm thick and 30 cm square plot the plate temperature vs. b.11 Saturated steam at atmospheric pressure is enclosed within a vertical 12-in. 21.Problems surface. Construct a plot of heat ﬂux vs. the heat-transfer coefﬁcient for the ﬁrst. what rate of water ﬂow would you suggest for safe operation? Support your results with all design criteria used. plate temperature. what height of vertical wall will cause the ﬁlm at the bottom of the tube to be turbulent? 21. Construct a plot for the thickness of condensate ﬁlm vs. d. 21. If saturated ammonia at 858F is adjacent to the surface. how long will it be before condensate spills over the lip if the pan is a.17 If eight tubes of the size designated in Problem 21. and eighth tubes. condensation rate.-OD tube with its surface maintained at 1008F surrounded by 8 steam at 2008F. horizontally.

1 TYPES OF HEAT EXCHANGERS In addition to being considered a closed-type exchanger. Hot and cold ﬂuids enter open-type heat exchangers and leave as a single stream. devices wherein physical mixing of the two ﬂuid streams actually occurs. In the recuperator. is the one of primary importance and the one to which we shall direct most of our attention. the hot and cold ﬂuid streams do not come into direct contact with each other but are separated by a tube wall or a surface that may be ﬂat or curved in some manner. using techniques from Chapter 13. Open-type heat exchangers are. The third type of heat exchanger. resulting in a continuous change in the temperature of at least one of the ﬂuid streams involved. investigate the conditions under which these three energy-transfer processes act in series with one another.Chapter 22 Heat-Transfer Equipment A device whose primary purpose is the transfer of energy between two ﬂuids is called a heat exchanger. regenerators. 2. We shall be concerned with a thermal analysis of these exchangers. 22. in the following sections. open-type exchangers. through the wall or plate by conduction. The required analytical tools for handling this type of heat exchanger have been developed in the preceding chapters. and then by convection from the surface to the second ﬂuid. closed-type exchangers or recuperators. the recuperator. A complete design of such equipment involves an analysis of pressure drop. 336 . No rate equations are necessary for the analysis of this type of exchanger. The nature of the exit stream is predicted by continuity and the ﬁrst law of thermodynamics. as well as material and structural considerations that are not within the scope of this text. and 3. Each of these energy-transfer processes has been considered separately in the preceding chapters. We shall. a recuperator is classiﬁed according to its conﬁguration and the number of passes made by each ﬂuid stream as it traverses the heat exchanger. Energy exchange is thus accomplished from one ﬂuid to an intermediate surface by convection. as implied in their designation. Heat exchangers are usually classiﬁed into three categories: 1. Regenerators are exchangers in which the hot and cold ﬂuids ﬂow alternately through the same space with as little physical mixing between the two streams as possible. The amount of energy transfer is dependent upon the ﬂuid and ﬂow properties of the ﬂuid stream as well as the geometry and thermal properties of the surface.

the ﬂuid streams would be unseparated or mixed. A common single-pass conﬁguration is the double-pipe arrangement shown in Figure 22.2 A crossﬂow heat exchanger. countercurrent ﬂow or simply counterﬂow if the ﬂuids ﬂow in opposite directions. In a condition such as that depicted in the ﬁgure. In order to accomplish as much transfer of energy in as little space as possible. A crossﬂow arrangement is shown in Figure 22. seldom are more than two shellside passes used. the tube-side ﬂuid makes two passes. and less-than-optimum performance is achieved. An additional descriptive term identiﬁes the relative directions of the two streams.1 Types of Heat Exchangers 337 A single-pass heat exchanger is one in which each ﬂuid ﬂows through the exchanger only once. In this ﬁgure.1 A double-pipe heat exchanger. Without these bafﬂes the ﬂuid becomes stagnant in certain parts of the shell.2. . Variations on the number of tubeand-shell passes are encountered in numerous applications. Figure 22. and crossﬂow if the two ﬂuids ﬂow at right angles to one another. The arrangement shown in Figure 22.3. it is desirable to utilize multiple passes of one or both ﬂuids. or both ﬂuids are mixed. Variations on the crossﬂow conﬁguration occur when one or the other. TH in Tc in Tc out TH out Figure 22.1. whereas the shell-side ﬂuid makes one pass. the terms used being parallel ﬂow or cocurrent ﬂow if the ﬂuids ﬂow in the same direction.2 is one in which neither ﬂuid is mixed. If the bafﬂes or corrugations were not present. A popular conﬁguration is the shelland-tube arrangement shown in Figure 22. as each section contacts an adjacent ﬂuid stream at a different temperature. the ﬂuid leaving at one end of the sandwich arrangement will have a nonuniform temperature variation from one side to the other. the ﬂow is partially channeled past these stagnant or ‘‘dead’’ regions.22. It is normally desirable to have one or both ﬂuids unmixed. Good mixing of the shell-side ﬂuid is accomplished with the bafﬂes shown.

1 W. M.338 Chapter 22 Heat-Transfer Equipment Figure 22. As each is a composite of several single-pass arrangements. McGraw-Hill Book Company.3 Shell-and-tube heat exchanger. compact. (a) (b) (c) (d) Figure 22. London. A number of more recent heat-transfer applications require more compact conﬁgurations than that afforded by the shell-and-tube arrangement. 1964. Kays and A. .4 Compact heat-exchanger conﬁgurations. The analysis of shell-and-tube. The subject of ‘‘compact heat exchangers’’ has been investigated and reported both carefully and quite thoroughly by Kays and London. 2nd edition. Compact Heat Exchangers. or any multiple-pass heat exchanger is quite involved. we shall initially focus our attention on the single-pass heat exchanger.1 Typical compact arrangements are shown in Figure 22.4. L.

this process may be considered the superposition of a condenser and a counterﬂow exchanger.5 Temperature proﬁles for single-pass. it is possible for the hot ﬂuid to leave the exchanger at a temperature below that at which the cold ﬂuid leaves.2 SINGLE-PASS HEAT-EXCHANGER ANALYSIS: THE LOG-MEAN TEMPERATURE DIFFERENCE It is useful.5(c) and (d). when considering parallel or counterﬂow single-pass heat exchangers. This situation occurs when energy transfer results in a change of phase rather than of temperature as in the cases of evaporation and condensation shown. In such a case.5. Also quite noticeable from Figure 22. the direction of ﬂow of the condensate stream is important. TH in TH out Tc out TH in TH TH out Tc out Tc Tc in Composite = Condenser + Tc in Counterflow exchanger Figure 22. to draw a simple sketch depicting the general temperature variation experienced by each ﬂuid stream. double-pipe heat exchangers.6.2 Single-Pass Heat-Exchanger Analysis: The Log-Mean Temperature Difference 339 22. In the counterﬂow arrangement. If the situation occurs where the complete phase change such as condensation occurs within the exchanger along with some subcooling. all of which are shown and labeled in Figure 22. It is a simple exercise to show that this temperature is the one resulting if the two ﬂuids are mixed in an open-type heat exchanger. For purposes of analysis.6 Temperature proﬁle in a condenser with subcooling. one of the two ﬂuids remains at constant temperature while exchanging heat with the other ﬂuid whose temperature is changing. as depicted in the diagram. Each of these may be found in a double-pipe arrangement. There are four such proﬁles in this category. then the diagram will appear as in Figure 22. In Figure 22. TH in TH in TH out TH out Tc in Tc out Tc out Tc in (a) Parallel flow (b) Counterflow TH in TH TH TH out Tc Tc Tc out Tc in (c) Evaporator (d) Condenser Figure 22.22. It is apparent that the exit temperatures of the hot and cold ﬂuids in the parallel-ﬂow case approach the same value.5(a) and (b) is the signiﬁcant difference in temperature proﬁle exhibited by the parallel and counterﬂow arrangements. This situation obviously corresponds . as it is of no consequence to the analysis. The direction of ﬂow of the ﬂuid undergoing a change in phase is not depicted in the ﬁgure.

we have dq ¼ U dA(TH Tc ) (22-3) which utilizes the overall heat-transfer coefficient. introduced in Chapter 15. the heat-transfer area varies linearly with distance from one end of the exchanger. U. we have d(DT) ¼ dTH dTc (22-4) . The abscissa of this ﬁgure is area. For a double-pipe arrangement. 1 2 TH 2 ∆T2 Tc 2 TH 1 ∆T1 ∆A Tc 1 A Figure 22. between the ends of this unit. in the case shown.7 Diagram of temperature vs. It is thus the single-pass counterﬂow arrangement to which we shall direct our primary attention. mc Writing equation (15-17) for the energy transfer between the two ﬂuids at this location. a ﬁrstlaw-of-thermodynamics analysis of the two ﬂuid streams will yield _ p )c DTc Dq ¼ (mc and _ p )H DTH Dq ¼ (mc As the incremental area approaches differential size. DA. The obvious conclusion to this discussion is that the counterﬂow conﬁguration is the most desirable of the single-pass arrangements. we may write _ p )c dTc ¼ Cc dTc dq ¼ (mc (22-1) _ p )H dTH ¼ CH dTH dq ¼ (mc (22-2) and where the capacity coefficient. is introduced in place of the more cumbersome product. _ p. C.7. the zero reference is the end of the exchanger at which the cold ﬂuid enters. With reference to a general increment of area. The detailed analysis of a single-pass counterﬂow heat exchanger that follows is referred to the diagram and nomenclature of Figure 22. Designating TH Tc as DT.340 Chapter 22 Heat-Transfer Equipment to a case of greater total energy transfer per unit area of heat exchanger surface than would be obtained if the same ﬂuids entered a parallel-ﬂow conﬁguration. contact area for single-pass counterﬂow analysis.

it is equally valid for any of the single-pass operations depicted in Figure 22. (DT2 DT1 )/ln (DT2 /DT1 ). then a step-by-step numerical integration is necessary. This coefﬁcient will not. we obtain 1 1 d(DT ) ¼ dq CH Cc dq CH 1 ¼ CH Cc (22-5) We should also note that dq is the same in each of these expressions. calculations based upon a value of U taken midway between the ends of the exchanger are usually accurate enough. U.5. yielding. upon integration. becomes ln DT2 UA (DT2 DT1 ) ¼ q DT1 This result is normally written as q ¼ UA DT2 DT1 DT2 ln DT1 (22-9) The driving force. equations (22-1)–(22-3) being evaluated repeatedly over a number of small-area increments. however. . is seen to be a particular sort of mean temperature difference between the two fluid streams. on the right-hand side of equation (22-9). equations (22-1) and (22-2) may be equated and integrated from one end of the exchanger to the other. but bears repeating. is designated DTlm . Z DT2 Z A d(DT ) U ¼ (DT2 DT1 ) dA (22-8) DT q 0 DT1 which. remain constant. for constant U. the logarithmic-mean temperature difference. and noting that CH (TH2 TH1 ) ¼ q.22. for the ratio CH/Cc CH Tc2 Tc1 ¼ Cc TH2 TH1 (22-6) which may be substituted into equation (22-5) and rearranged as follows: dq Tc2 Tc1 dq TH2 TH1 Tc2 þ Tc1 d(DTÞ ¼ 1 ¼ CH CH TH2 TH1 TH2 TH1 dq DT2 DT1 ¼ CH TH2 TH1 (22-7) Combining equations (22-3) and (22-7). that equation (22-10) is based upon a constant value of the overall heat-transfer coefﬁcient.2 Single-Pass Heat-Exchanger Analysis: The Log-Mean Temperature Difference 341 and substituting for dTH and dTc from equations (22-1) and (22-2). in general. This ratio. we have. and the expression for q is written simply as q ¼ UA DTlm (22-10) Even though equation (22-10) was developed for the specific case of counterflow. It was mentioned earlier. If there is considerable variation in U from one end of the exchanger to the other. thus.

Water at 280 K is available in sufﬁcient quantity to allow 0. In such a case. and ﬂows at a rate of 0.8 Single-pass temperature proﬁles for counterﬂow and parallel ﬂow. The overall heat-transfer coefﬁcient may be taken as 250 W/m2 K. respectively.5 kg/s.8). DT2 DT1 0 ¼ . Determine the required heat-transfer area for (a) counterﬂow and (b) parallel-ﬂow operations (see Figure 22. EXAMPLE 1 Light lubricating oil (cp ¼ 2090 J/kg K) is cooled by allowing it to exchange energy with water in a small heat exchanger.342 Chapter 22 Heat-Transfer Equipment It is also possible that the temperature differences in equation (22-9).201 kg/s to be used for cooling purposes. L’Hoˆpital’s rule may be applied as follows: DT2 DT1 DT1 f(DT2 /DT1 ) 1g ¼ lim lim ln (DT2 /DT1 ) DT2 !DT1 ln(DT2 /DT1 ) DT2 /DT1 !1 when the ratio DT2 /DT1 is designated by the symbol F. . It turns out that a simple arithmetic mean is within 1% of the logarithmic-mean temperature difference for values of (DT2 /DT1 ) < 1:5. evaluated at either end of a counterﬂow exchanger. that is. The outlet water temperature is determined by applying equations (22-1) and (22-2) q ¼ (0:5 kg/s)(2090 J/kg K)(25 K) ¼ 26 125 W ¼ (0:201 kg/s)(4177 J/kg K)(Tw out 280 K) 375 K 375 K 350 K 350 K Tw out 280 K Tw out 280 K (a) Counterflow (b) Parallel flow Figure 22. we may write F1 ¼ lim DT F!1 ln F Differentiating numerator and denominator with respect to F yields the result that lim DT2 DT1 DT2 !DT1 ln(DT2 /DT1 ) ¼ DT or that equation (22-10) may be used in the simple form q ¼ UA DT (22-11) From the foregoing simple analysis. ln(DT2 /DT1 ) 0 if DT1 ¼ DT2 In such a case. the log-mean temperature difference is indeterminate. it should be apparent that equation (22-11) may be used and achieve reasonable accuracy so long as DT1 and DT2 are not vastly different. are equal. The oil enters and leaves the heat exchanger at 375 and 350 K.

125 W is seen to be lower for the counterﬂow arrangement by approximately 7%. we see that the area required to accomplish this energy transfer is A¼ 26 125 W ¼ 1:562 m2 (250 W/m2 K)(66:9 K) (16:81 ft2 ) Performing similar calculations for the parallel-ﬂow situation. 1952. Mueller. . Correction factors to be used with equation (22-10) have been presented in chart form by Bowman. M. 283 (1940). The ﬁrst three are for different shell-and-tube conﬁgurations and the latter three are for different crossﬂow conditions.3 Figures 22. respectively. and Nagle2 and by the Tubular Exchanger Manufacturers Association. Trans.3 CROSSFLOW AND SHELL-AND-TUBE HEAT-EXCHANGER ANALYSIS More complicated ﬂow arrangements than the ones considered in the previous sections are much more difﬁcult to treat analytically.E. A.3 Crossﬂow and Shell-and-Tube Heat-Exchanger Analysis 343 from which we obtain Tw out ¼ 280 þ (0:5)(2090)(25) ¼ 311:1 K (0:201)(4177) (100 F) This result applies to both parallel ﬂow and counterﬂow. Tubular Exchanger Manufacturers Association.10 present correction factors for six types of heat-exchanger conﬁgurations. Nagle. The quantity read on the ordinate of each plot. we obtain DTlm ¼ A¼ 95 38:9 ¼ 62:8 K 95 ln 38:9 (113 F) 26 125 W ¼ 1:66 m2 (250 W/m2 K)(62:8 K) (17:9 ft2 ) The area required to transfer 26. New York. The parameters in Figures 22. 3rd edition.10 are evaluated as follows: Y¼ Tt out Tt in Ts in Tt in (22-12) Z¼ _ p )tube Ct Ts in Ts out (mc ¼ ¼ _ p )shell Cs Tt out Tt in (mc (22-13) where the subscripts s and t refer to the shell-side and tube-side ﬂuids. Bowman. 22. C. A.M.9 and 22. For the counterﬂow conﬁguration.9 and 22. and W. 62. A. DTlm is calculated as DTlm ¼ 70 63:9 ¼ 66:9 K 70 ln 63:9 (120:4 F) and applying equation (22-10).S. Standards.22. is F. Mueller. TEMA. for given values of Y and Z. the correction factor to be applied to equation (22-10). and thus these more complicated conﬁgurations 2 3 R.

4 0. four.7 0. F. and W.4 TH1 – TH 2 Tc2 – Tc1 Y Shell fluid Tube fluid (b) Correction factor.0 0.9 1.5 0 TH1 – TH2 Tc – Tc 2 1 0. M. based on counterﬂow LMTD.5 0 TH1 – TH2 Tc 2 – Tc1 0.8 0.5 1.6 0.M.4 0.0 3.2 0.8 0. By permission of the publishers.9 0. Trans. F 1.5 0. . or any multiple of four tube passes (c) Figure 22.1 0. 284.62.S. (a) One shell pass and two or a multiple of two tube passes.5 Y 1.0 0.3 TH1 Tc2 Tc1 TH2 0.0 3.E.8 0.0 0.) Correction factors.0 TH1 Tc2 Tc1 TH2 Correction factor plot for exchanger with one shell pass and two.8 z = 4.2 0.8 0.9 1. Bowman.0 Correction factor plot for exchanger with two shell passes and four..9 Correction factors for three shell-andtube heat-exchanger conﬁgurations.7 0.6 1.0 3.5 0. (b) One shell pass and three or a multiple of three tube passes.6 z = 0. A.5 Y 0. 285 (1940).7 0. or any multiple of tube passes (a) 1.2 0.1.0 0.8 0.8 0.4 0.0 1.0 Correction factor F A 0. (From R.6 0.0 1. Nagle.4 0.3 0. C.6 0. eight. A.0 2.6 0.9 0.7 0.6 z= 0. (c) Two shell passes and two or a multiple of two tube passes.0 1.2 0.8 z = 4.6 0. Mueller.1 0. A.0 2.0 Correction factor F B z = 4.0 2.2 0.

8 0.2 0. ﬂuid ﬂows over ﬁrst and second passes in series.6 Tc2 – Tc1 0. 62. one ﬂuid unmixed.7 0. Mueller.8 z = 4. single-pass.0 0. By permission of the publishers.E.6 0.0 1.0 0. A.8 z = 4.5 0.8 0. Bowman.3 Crossﬂow and Shell-and-Tube Heat-Exchanger Analysis 345 Correction factor. 288–289 (1940).0 TH1 – Tc1 TH1 Tc1 Tc2 TH2 (a) Correction factor.6 0. (b) Crossﬂow.5 1.6 z = 0.9 0.0 0.0 2.2 0.5 1.2 0.5 0 TH1 – TH2 Tc2 – Tc1 0. A.0 3.4 0.9 0.10 Correction factors for three crossﬂow heat-exchanger conﬁgurations. both ﬂuids unmixed.7 0.7 0.1 0. F 1. single-pass.0 1. C.9 1.22. and W.3 0. tube passes mixed.6 z = TH1 – TH2 Tc – Tc 2 0.3 0.4 Y= 0.5 0.1 0.0 0. Nagle.5 0 1 0.0 0.0 3. Trans.) . (a) Crossﬂow.2 0. (c) Crossﬂow. (From R.0 2.8 0.9 1.S. A.4 0. F 1.6 Tc2 – Tc1 0. M..M.8 TH1 – Tc1 TH1 Tc1 Tc2 TH2 (b) Figure 22.4 Y= 0.7 0.

The manner of using Figures 22.2 0.6 0.6 0.10(b) must be used.5 3.7 TH1 – TH2 0. water-mixed.8 1.6 0.0 Y TH1 Tc2 Tc1 TH2 (c) Figure 22.5 Tc2 – Tc1 0 0.0 1. For part (a).0 0. The reader is cautioned to apply equation (22-10).346 Chapter 22 Heat-Transfer Equipment Correction factor. using the factor F as in equation (22-14). F 1. oil being the tube-side ﬂuid.8 1.2 0. EXAMPLE 2 In the oil–water energy transfer described in Example 1. q ¼ UA(F DTlm ) (22-14) with the logarithmic-mean temperature difference calculated on the basis of counterﬂow. compare the result obtained with the result that would be obtained if the heat exchanger were (a) crossﬂow.4 0. The parameters needed to use this ﬁgure are Y¼ Tt out Tt in 25 ¼ 0:263 ¼ 95 Ts in Tt in . Figure 22.9 2. (b) shell-and-tube with four tube-side passes.0 0.9 and 22.4 0.10 may be illustrated by referring to the following example.10 Continued may be treated in much the same way as the single-pass double-pipe case.8 0.0 0.0 z = 4.

It is apparent that the ﬂuid experiencing the larger change in temperature is the one having the smaller capacity coefﬁcient. Tech. the expression for E becomes Cc (Tc out Tc in ) Cmax (Tc out Tc in ) E¼ (22-16) ¼ CH (TH in TH out )max Cmin (TH in Tc in ) 4 W.11. The required area for part (a) is thus equal to (1:562)/(0:96) ¼ 1:63 m2 . The book ‘‘Compact Heat Exchangers.22. as in Figure 22.4 THE NUMBER-OF-TRANSFER-UNITS (NTU) METHOD OF HEAT-EXCHANGER ANALYSIS AND DESIGN Earlier mention was made of the work of Kays and London1 with particular reference to compact heat exchangers. TH in TH in TH out TC out TC out TH out TC in (a) CH > CC. also presents charts useful for heat-exchanger design on a different basis than discussed thus far. Mechan. CH = Cmin TC in Figure 22. Thermodyn.. proposed the method of analysis based upon the heat-exchanger effectiveness E. as in Figure 22. in general. .11(a). The area for part (b) becomes (1:562)/(0:97) ¼ 1:61 m2 . which we designate Cmin. the exit temperature of the cold ﬂuid will equal the inlet temperature of the hot ﬂuid.11 Temperature proﬁles for counterﬂow heat exchangers. 12 (1930). Nusselt. This term is deﬁned as the ratio of the actual heat transfer in a heat exchanger to the maximum possible heat transfer that would take place if inﬁnite surface area were available. one ﬂuid undergoes a greater total temperature change than the other. Nusselt. as in Figure 22.4 in 1930.4 The Number-of-Transfer-Units (NTU) Method of Heat-Exchanger Analysis and Design 347 and Z¼ Ts in Ts out 31:1 ¼ ¼ 1:244 Tt out Tt in 25 and from the ﬁgure we read F ¼ 0:96. By referring to a temperature proﬁle diagram for counterﬂow operation. it is seen that. and if there is inﬁnite area available for energy transfer.’’ by Kays and London.11(b). yielding a value of F equal to 0. If Cc ¼ Cmin . According to the deﬁnition of effectiveness. The values of Y and Z determined above are the same in part (b). we may write E¼ CH (TH in TH out ) Cmax (TH in TH out ) ¼ Cc (Tc out Tc in )max Cmin (TH in Tc in ) (22-15) If the hot fluid is the minimum fluid. CC = Cmin (b) CC > CH.97. 22.

348 Chapter 22 Heat-Transfer Equipment Notice that the denominators in both equations (22-15) and (22-16) are the same and that. in all of its useful forms as far as heat-exchanger analysis and design are concerned. the numerator represents the actual heat transfer. Equation (22-17) is conspicuous among these others. we initially write equation (22-17) in the form CH (TH in TH out ) Cc (Tc out Tc in ) E¼ ¼ (22-18) Cmin (TH in Tc in ) Cmin (TH in Tc in ) The appropriate form for equation (22-18) depends on which of the two fluids has the smaller value of C. as the temperature difference appearing is that between the inlet streams alone. and equation (22-17) is obviously the easiest means of attaining this knowledge if one can determine the value of E. expresses q. TH2. equation (22-10) may be written as follows (numerical subscripts correspond to the situation shown in Figure 22.7): q ¼ Cc (Tc2 Tc1 ) ¼ UA (TH1 Tc1 ) (TH2 Tc2 ) ln½(TH1 Tc1 )/(TH2 Tc2 ) (22-19) The entering temperature of the hot fluid. may be written in terms of E by use of equation (22-18). along with the integrated forms of equations (22-1) and (22-2). the rate of heat transfer. To determine E for a single-pass case. It is thus possible to write a fifth expression for q as q ¼ ECmin (TH in Tc in ) (22-17) which. we have Cc TH2 TH1 ¼ CH Tc2 Tc1 which may be rearranged to the form TH1 ¼ TH2 Cmin (Tc2 Tc1 ) Cmax or TH1 Tc1 ¼ TH2 Tc1 Cmin (Tc2 Tc1 ) Cmax (22-22) . This is a definite advantage when a given heat exchanger is to be used under conditions other than those for which it was designed. We shall consider the cold fluid to be the minimum fluid and consider the case of counterflow. For these conditions. as well as equations (22-10) and (22-14). The exit temperatures of the two streams are then needed quantities. in each case. yielding TH2 ¼ Tc1 þ 1 (Tc2 Tc1 ) E (22-20) and also 1 TH2 Tc2 ¼ Tc1 Tc2 þ (Tc2 Tc1 ) E 1 1 (Tc2 Tc1 ) ¼ E (22-21) From the integrated forms of equations (22-1) and (22-2).

22.12 and 22. an analogous development to the preceding will yield Cmin 1 exp NTU 1 þ Cmax E¼ 1 þ Cmin /Cmax (22-26) Kays and London1 have put equations (22-25) and (22-26) into chart form. the same result would have been achieved. we obtain 1 Cmin (Tc2 Tc1 ) (Tc2 Tc1 ) E Cmax 1 Cmin ¼ ðTc2 Tc1 ) E Cmax TH1 Tc1 ¼ (22-23) Now substituting equations (22-21) and (22-23) into equation (22-19) and rearranging. Equation (22-24) was derived on the basis that Cc ¼ Cmin . UA Cmin 1 exp 1 Cmin Cmax E¼ UA Cmin 1 ðCmin /Cmax Þ exp 1 Cmin Cmax (22-24) The ratio UA/Cmin is designated the number of transfer units. ﬁnally. The utility of the NTU approach is illustrated in the following example.4 The Number-of-Transfer-Units (NTU) Method of Heat-Exchanger Analysis and Design 349 Combining this expression with equation (22-20). we have 1/E Cmin /Cmax UA Cmin ln ¼ 1 Cmin 1/E 1 Cmax Taking the antilog of both sides of this expression and solving for j. along with comparable expressions for the effectiveness of several shell-and-tube and crossflow arrangements. abbreviated NTU. equation (22-17) may be used both as an original design equation and as a means of evaluating existing equipment when it operates at other than design conditions. .13 are charts for E as functions of NTU for various values of the parameter Cmin /Cmax. if we had initially considered the hot ﬂuid to be minimum. equation (22-25) Cmin 1 exp NTU 1 Cmax E¼ Cmin 1 (Cmin /Cmax )exp NTU 1 Cmax (22-25) is valid for counterflow operation in general. Thus. For parallel flow. With the aid of these ﬁgures. we have. Figures 22.

00 60 40 20 0 0 1 2 3 4 Number of transfer units. NTU = AU/C min (b) 5 Figure 22.Heat-Transfer Equipment Heat-transfer surface 100 Cmin/Cmax = 0 0.50 60 0.25 Effectiveness E . (a) Counterﬂow. . NTU = AU/C min (a) 5 Heat-transfer surface 100 Cmin/Cmax = 0 80 0. in % Chapter 22 1.50 0. (c) One shell pass and two or a multiple of two tube passes.75 80 Effectiveness E .25 0.12 Heat-exchanger effectiveness for three shell-and-tube conﬁgurations.75 1. (b) Parallel ﬂow. in % 350 0.00 40 20 0 0 1 2 3 4 Number of transfer units.

75 60 1.4 351 The Number-of-Transfer-Units (NTU) Method of Heat-Exchanger Analysis and Design Shell fluid Tube fluid 100 Cmin/Cmax = 0 Effectiveness E . (c) Crossﬂow.25 0. in % 80 0. (b) Crossﬂow. . (a) Crossﬂow.12 Continued 100 Cmin/Cmax = 0 Effectiveness E .22. NTU = AU/C min 5 (a) Figure 22.00 40 20 0 0 1 2 3 4 Number of transfer units.00 40 20 0 0 1 2 3 4 Number of transfer units.50 0. one ﬂuid mixed.13 Heat-exchanger effectiveness for three crossﬂow conﬁgurations.75 60 1. both ﬂuids unmixed. in % 80 0.25 0. multiple pass.50 0. NTU = AU/C min (c) 5 Figure 22.

in % Chapter 22 0.33 Cmixed = 0.00 Cunmixed 60 Cmixed =1 Cunmixed 40 20 0 0 1 2 3 4 Number of transfer units.5 2 0. NTU = AU/C min (c) Figure 22. NTU = AU/C min (b) Two-pass arrangement 100 Four pass 80 Effectiveness E .13 Continued 5 .25 4 0.Heat-Transfer Equipment 100 80 Mixed fluid Unmixed fluid Effectiveness E .75 1. in % 352 Counterflow (n = ∞) Three pass 70 Two pass One pass 60 50 40 0 1 2 3 4 5 Number of transfer units.

the required area may be evaluated for each heat-exchanger conﬁguration.4 EXAMPLE 3 The Number-of-Transfer-Units (NTU) Method of Heat-Exchanger Analysis and Design 353 Repeat the calculations for Examples 1 and 2 to determine the required heat-transfer area for the speciﬁed conditions if the conﬁgurations are (a) counterﬂow. (a) Counterﬂow NTU ¼ 0:47 A¼ (0:47)(841:2) ¼ 1:581 m2 250 (b) Parallel ﬂow NTU ¼ 0:50 A¼ (0:50)(841:2) ¼ 1:682 m2 250 (c) Crossﬂow. the effectiveness is evaluated as E¼ 26 125 W ¼ 0:327 (841:2 J/kg s)(95 K) By using the appropriate chart in Figures 22. From equation (22-16).22. (b) parallel ﬂow. (c) crossﬂow. . water-mixed NTU ¼ 0:48 A¼ (0:48)(841:2) ¼ 1:615 m2 250 (d) Shell-and-tube. four tube-side passes NTU ¼ 0:49 A¼ (0:49)(841:2) ¼ 1:649 m2 250 These results are comparable to those obtained earlier. with some possible inaccuracies involved in reading the chart.12 and 22.13. the appropriate NTU values and. water-mixed. in turn. and (d) shell-and-tube with four tube-side passes. It is ﬁrst necessary to determine the capacity coefﬁcients for the oil and water _ p )oil ¼ (0:5 kg/s)(2090 J/kg K) ¼ 1045 J/s K Coil ¼ (mc and _ p )w ¼ (0:201 kg/s)(4177 J/kg K) ¼ 841:2 J/s K Cwater ¼ (mc thus the water is the minimum ﬂuid.

the NTU approach is clearly superior. the surface is said to be ‘‘fouled. Some .354 Chapter 22 Heat-Transfer Equipment The NTU method offers no distinct advantage over the procedure introduced earlier.5 ADDITIONAL CONSIDERATIONS IN HEAT-EXCHANGER DESIGN After a heat exchanger has been in service for some time. In Example 4. a crossﬂow heat exchanger with the shell-side ﬂuid (water) mixed is constructed with a heat-transfer area of 1:53 m2 . a bit simpler. EXAMPLE 4 In the energy exchange between water and lubricating oil as considered in the preceding examples. using the log-mean temperature difference. and F are all dependent on one or both exit stream temperatures.’’ When a fouling resistance exists.13. It is extremely difﬁcult to predict the rate of scale buildup or the effect such buildup will have upon heat transfer. When the nature of the surface is altered in some way as to affect the heat-transfer capability. from Figure 22. the thermal resistance is increased and a heat exchanger will transfer less energy than the design value. however. thus. when performing calculations of the type involved in the preceding examples. With Coil ¼ Cmin we have NTU ¼ UA (250 W/m2 K)(1:53 m2 ) ¼ Cmin 1045 J/s K ¼ 0:366 and. A new pump is attached to the water supply line enabling the water ﬂow rate to be increased to 1000 kg/h. This value may now be used in equations (22-1) and (22-2) to yield the required answers. The NTU method is. the effectiveness is E ﬃ 0:29 Using equation (22-17) we may evaluate the heat-transfer rate as q ¼ (0:29)(1045 J/s K)(95 K) ¼ 28:8 K an increase of over 10%. Y. its performance may change as a result of the buildup of scale on a heat-transfer surface or of the deterioration of the surface by a corrosive ﬂuid. Using the NTU method. Toil out ¼ 375 28:8 kW/(1045 W/K) ¼ 347:4 K Tw out ¼ 280 þ 28:8 kW/(1160 W/K) ¼ 304:8 K 22. a trial-and-error method would be necessary as DTlm . What will be the exit temperatures of the water and oil for the new operating conditions? If the DTlm method were used in this problem. it is ﬁrst necessary to calculate the capacity coefﬁcients Coil ¼ (0:5 kg/s)(2090 J/kg K) ¼ 1045 J/s K Cw ¼ (1000 kg/h)(h/3600 s)(4177 J/kg K) ¼ 1160 J/kg K Oil is now the ‘‘minimum’’ ﬂuid.

below 325 K above 325 K Boiler feed water.6 35.2.1 70. treated City or well water.5 Some useful values are given in Table 22. Fouling resistances that have been obtained from experiments may be used to roughly predict the overall heat-transfer coefﬁcient by incorporation into an expression similar to equation (15-19).8 35. 3rd edition.6 17. Uf is the overall heattransfer coefﬁcient of the fouled exchanger.5 8. The thermal resistance of the scale is determined by Rsc ¼ 1 1 Uf U0 (22-27) where U0 is the overall heat-transfer coefﬁcient of the clean exchanger. . C.5 Additional Considerations in Heat-Exchanger Design 355 evaluation can be done after a heat exchanger has been in service for some time by comparing its performance with that when the surfaces were clean. Sta.2 8. ﬂow rates.6 17. The most difﬁcult quantity to estimate quickly is the overall heat-transfer coefﬁcient. below 325 K above 325 K Refrigerating liquids Refrigerating vapors Liquid gasoline. organic vapors Fuel oil Quenching oil Steam. The following equation includes the fouling resistances.1 Heat-exchanger fouling resistances Fluid Distilled water Sea water. U. and Rsc is the thermal resistance of the scale. A.2 17. Eng.6 35. Mueller6 has prepared the very useful table of approximate U values which is reproduced here as Table 22. Table 22. New York. Bull. Ser. 121 (1954).1. and the like. Eng.8 88.8 17. Mueller.2 It is often useful to have ‘‘ball-park’’ ﬁgures on heat-exchanger size.22. non-oil-bearing Industrial air Fouling resistances (m2 K / W 105 ) 8. TEMA Standards. Ri on the inside tube surface and Ro on the outside tube surface: U0 ¼ 1 A0 /Ai hi þ Ri þ ½A0 ln(ro /ri )=2pk=L þ Ro þ 1=ho (22-28) Fouling resistances to be used in equation (22-28) have been compiled by the Tubular Exchanger Manufacturers Association. Exp. 5 6 Tubular Exchanger Manufacturers Association. Res.8 8. 1952. Purdue Univ.

Either is reasonably rapid and straightforward for designing an exchanger. Equation (22-17) is a simpler and more direct approach when analyzing an exchanger that operates at other than design conditions. instantaneous heater storage-tank heater Steam to oil. alcohol Water to condensing alcohol Water to lubricating oil Water to condensing oil vapors Water to condensing or boiling Freon-12 Water to condensing ammonia Steam to water. The two methods for heat-exchanger design utilize either equation (22-10) or (22-17).356 Chapter 22 Heat-Transfer Equipment Table 22. All heat-exchanger design and analysis involve one or more of the following equations: dq ¼ Cc dTc (22-1) dq ¼ CH dTH (22-2) dq ¼ U dAðTH Tc ) (22-3) q ¼ UA DTlm (22-10) and q ¼ ECmin (TH in Tc in ) (22-17) Charts were presented by which single-pass techniques could be extended to include the design and analysis of crossflow and shell-and-tube configurations. . heavy fuel light fuel light petroleum distillate Steam to aqueous solutions Steam to gases Light organics to light organics Medium organics to medium organics Heavy organics to heavy organics Heavy organics to light organics Crude oil to gas oil 55–165 850–1560 570–1140 340–480 200–340 280–850 250–680 110–340 220–570 280–850 850–1350 2280–3400 990–1700 55–165 165–340 280–1140 570–3400 28–280 220–425 110–340 55–220 55–340 165–310 22. jacket water coolers Water to brine Water to gasoline Water to gas oil or distillate Water to organic solvents.6 CLOSURE The basic equations and procedures for heat-exchanger design are presented and developed in this chapter.2 Approximate values for overall heat-transfer coefﬁcients Fluid combination U(W/m2 K) Water to compressed air Water to water.

for the exchanger of part (a). The coolant (Cc ¼ 4:20 kJ/kg K) enters the shell at 208C and leaves the shell at 428C. The overall heat-transfer coefﬁcient is 12 Btu/h ft2 8F. determine a. enters the exchanger at 2508F.9 Consider the exchanger in Problem 22.7 Water at 508F is available for cooling at a rate of 400 lbm/h. The kerosene enters at a rate of 2500 lbm/h. 22. ﬁnd the maximum ﬂow of oil that may be cooled with this unit. with cp ¼ 0.4 remain the same. The mass ﬂow rate of the water is 2. b. respectively.12 A shell-and-tube exchanger having one shell pass and eight tube passes is to heat kerosene from 80 to 1308F. for the water and 305 and 350 K. The overall heat-transfer coefﬁcient is 160 W/m2 K. If 100.13 If the overall heat-transfer coefﬁcient. Water entering at 2008F and at a rate of 900 lbm/h is to ﬂow on the shell side. Determine the required heat-transfer area. respectively.2 One hundred thousand pounds per hour of water are to pass through a heat exchanger. The water is to pass through the smooth horizontal tubes in turbulent ﬂow. For an overall tube-side heat-transfer coefﬁcient of 1200 W/m2 K and an oil ﬂow of 11 kg/s. the exchanger surface area. and the available tube-side pressure drop (neglecting entrance and exit losses) are all speciﬁed. double-pipe heat exchanger at a rate of 150 lbm/min and is heated from 60 to 1408F by an oil with a speciﬁc heat of 0. It is to be cooled to 350 K. b. the mass ﬂow rate of the heat stream. the corrected logarithmic-mean temperature difference. for the oil. the water ﬂow rate were decreased to 120 lbm/min? 22. Determine the surface area required if the overall heattransfer coefﬁcient is 230 W/m2 K. the initial and ﬁnal water temperatures. The overall heat-transfer coefﬁcient if 260 Btu/h ft2 8F. Determine the fouling resistance on the oil side of the exchanger. the required surface area in the exchanger.1 A single tube-pass heat exchanger is to be designed to heat water by condensing steam in the shell. 22. which is to raise the water temperature from 140 to 2008F. 22. 22. and the steam is to be condensed dropwise in the shell.11 A shell-and-tube heat exchanger is used to cool oil (CH ¼ 2:2 kJ/kg K) from 110 to 658C. A 2.24 Btu/lbm 8F are available at 8008F.6 A water-to-oil heat exchanger has entering and exiting temperatures of 255 and 340 K. both ﬂuids unmixed. Combustion products having a speciﬁc heat of 0.5 Water enters a counterﬂow. 22. What heat-transfer area is required? b. The water ﬂow rate. and the oil must leave the exchanger at no more than 1608F. 22. initial ﬂuid temperature. the exit temperature of the ﬂue gas. Water is available to cool the oil at a rate of 2 kg/s and a temperature of 280 K. 22. the required surface area in the exchanger. crossﬂow. and the average exit air temperature from this section is 300 K. the outlet of the oil reaches 308C instead of 388C with all other conditions remaining the same. It enters a double-pipe heat exchanger with a total area of 18 ft2. . Calculate the air ﬂow rate and the total heat transfer.4 Air at 203 kPa and 290 K ﬂows in a long rectangular duct with dimensions 10 cm by 20 cm.7 kg/s. What is the effectiveness of this heat exchanger? 22. b.8. The oil enters at 2408F and leaves at 808F.5-m length of this duct is maintained at 395 K. The exiting water temperature is limited at 2128F. Hot water (CH ¼ 4:18 kJ/kg K) enters the shell at 758C and leaves the shell at 558C. 22. What exit water temperature would result if. and total heat-transfer area determined in Problem 22. The heat exchanger has two shell passes and four tube passes. the required heat-transfer area for a counterﬂow exchanger. determine the effect of tube diameter on the total area required in the exchanger. b.000 lbm/h of the combustion products are available. b. After 4 years of operation. In order to determine the optimum exchanger design.8 A shell-and-tube heat exchanger is used for the heating of oil from 20 to 308C. a. The overall heat-transfer coefﬁcient is 150 Btu/h ft2 8F. Given the value of U ¼ 60 Btu/h ft2 8F. Determine a. The overall heat-transfer coefﬁcient based on the outside surface of the tubes is estimated to be 1080 W/m2 K.45 Btu/lbm 8F. The hot stream (CH ¼ 1:2 kJ/kg K) enters the heat exchanger at 2808C and leaves at 1208C. the condensation temperature of the steam.Problems 357 PROBLEMS 22. Oil. it is desirable to know how the total required area of the exchanger varies with the tube diameter selected. Determine a. shell-and-tube with two tube passes and one shell pass. What area is required if all conditions remain the same except that a shell-and-tube heat exchanger is used with the water making one shell pass and the oil making two tube passes? c. the coolant mass ﬂow rate.10 A ﬁnned-tube crossﬂow heat exchanger with both ﬂuids unmixed is used to heat water (Cc ¼ 4:2 kJ/kg K) from 20 to 758C.45 Btu/lbm 8F.3 An oil having a speciﬁc heat of 1880 J/kg K enters a single-pass counterﬂow heat exchanger at a rate of 2 kg/s and a temperature of 400 K. ﬁnd the exit oil temperature if the conﬁguration is changed to a. 22. the oil ﬂow rate is 12 kg/s (Cc ¼ 2:2 kJ/kg K). The heat exchanger has one shell pass and two tube passes. determine a. Assuming that the water ﬂow remains turbulent and that the thermal resistance of the tube wall and the steam–condensate ﬁlm is negligible. 22.

22. 22.25 in this service for a circulating water ﬂow of 0. The hot stream ﬂows at 2400 lbm/h and enters at 4008F. Surface coefﬁcients on the inside and outside tube surfaces are 470 and 210 W/m2 K. The water leaves the unit at a temperature of 388C. . what will be the approximate maximum ﬂow rate of steam in kg/s that can be condensed? What will be the leaving temperature of the circulating water under these conditions? Under these conditions. If the NTU rating for the condenser is given by the manufacturer as 1.19 Water ﬂowing at a rate of 10 kg/s through 50 tubes in a double-pass shell-and-tube heat exchanger heats air that ﬂows on the shell side. (b). The shell side is one-pass with water. Determine the following: a. the heat vaporization is 2257 kJ/kg and cp is 4:18 kJ/kg K. If the overall heat-transfer coefﬁcient is 300 W/m2 K. respectively. and (c).17 A shell-and-tube heat exchanger with two shell passes and four tube passes is used to exchange energy between two pressurized water streams. The coefﬁcient on the condensate side is 10. b.7 m long. heat-exchanger effectiveness. determine the required surface area. counterﬂow. ﬂowing at 6. shell-side ﬂuid mixed. above. This is for the same ﬂow rates and entering temperatures of both streams. The average water velocity in the 1. Determine the required number of tube passes. after a long period of operation. Because of space limitations. after a long period of operation.9 kg/s entering at 948C.15 Determine the required heat-transfer surface area for a heat exchanger constructed from 10-cm OD tubes. c. and water enters and leaves the exchanger at 90 and 1258F.16 Compressed air is used in a heat-pump system to heat water.20 Water ﬂowing at a rate of 3. The heat transfer rate is 0:2 106 kW. single pass.07 kg/s. that the cold stream leaves at 1848F instead of at the design value of 2208F. A counterﬂow surface with U ¼ 30 Btu/h ft2 8F and a surface-to-volume ratio of 130 ft2/ft3. parallel ﬂow.93 kg/s is cooled from 340 to 312 K by 6.6 cm and are 6. which is subsequently used to warm a house.14 A condenser unit is of a shell-and-tube conﬁguration with steam condensing at 858C in the shell.000 Btu/h. Three different exchanger conﬁgurations are of interest: a. 22. Evaluate the fouling factor that exists at the new conditions. the number of tubes per pass. exiting temperatures of the water and air streams.30 kg/s of water that is available at 283 K.21 Saturated steam at 373 K is to be condensed in a shelland-tube exchanger. 22.44 m in length. b. The entering water temperature is 350 K. A 95% ethanol solution (cp ¼ 3:810 kJ/kg K). Air enters the exchanger at 2008F and leaves at 1208F.18 For the heat exchanger described in Problem 22. a scale has been built up inside the tubes resulting in an added fouling resistance of 0:0021 m2 K/W. An overall heattransfer coefﬁcient of 4600 W/m2 K may be assumed to apply. which make two passes through the single-shell unit. What must be the required length of tubes for this case? 22. The house demand is 95. Air enters the unit at 158C with a ﬂow rate of 16 kg/s.366 m/s. The overall heat-transfer coefﬁcient is 1420 W/m2 K. ﬂowing at 1. determine the new results for parts (a). A crossﬂow unit with both ﬂuids unmixed with U ¼ 50 Btu/ h ft2 8F and surface-to-volume ratio of 90 ft2/ft3.358 Chapter 22 Heat-Transfer Equipment 22. The overall heat-transfer coefﬁcient based on outside tube area is 568 W/m2 K. the tubes may not exceed 2. b. 22. heat-transfer rate to the air. A crossﬂow conﬁguration with the water unmixed and air mixed having U ¼ 40 Btu/h ft2 8F and a surface-to-volume ratio of 100 ft2/ft3. 22. a.600 W/m2 K.8 kg/s is heated from 38 to 558C in the tubes of a shell-and-tube heat exchanger. respectively. and the length of tubes consistent with this restriction. The tubes are made of brass with outside diameters of 2. Water at 208C enters the tubes. it is to enter as steam at 373 K and leave as condensate at approximately 373 K. c. c. single pass. Choose from the following alternative units the one that is most compact.17 it is observed. If. One stream ﬂowing at 5000 lbm/h is heated from 75 to 2208F.905-cm-ID tubes is 0. and circulating water is available at 280 K. crossﬂow with one tube pass and one shell pass.

The electromagnetic spectrum shown in Figure 23. having both wave properties and particle-like properties. In the range from l ¼ 0:38 to 0.Chapter 23 Radiation Heat Transfer The mechanism of radiation heat transfer has no analogy in either momentum or mass transfer. Gamma rays and x-rays also have great penetrating ability. by lv ¼ c. matter is not required for radiant heat transfer. plastics. v. To produce radiation of this type we must disturb the nucleus or the inner-shell electrons of an atom.76 microns. we will concern ourselves ﬁrst with understanding the nature of thermal radiation. and spacecraft. of radiation is related to the wavelength l. radiation is sensed by the optical nerve of the eye and is what we call light. 23. radiation from a hot object will be different in quality than radiation from a body at a lower temperature. Radiation travels at the speed of light. symbolized m. First. the amount of heat transfer depends upon both the temperature difference between two bodies and their absolute temperatures. home heating. The color of incandescent objects is observed to change as the temperature is changed. Radiation heat transfer is extremely important in many phases of engineering design such as boilers. the amount of heat transfer was found to depend upon the temperature difference. In conduction and convection. Finally. The unit of wavelength which we shall use in discussing radiation is the micron. The radiation between wavelengths of 0. and glasses. In this chapter. the energy associated with these waves is much less than that for short-wavelength radiation.1 NATURE OF RADIATION The transfer of energy by radiation has several unique characteristics when contrasted with conduction or convection. we will illustrate some techniques for solving relatively simple problems where surfaces and some gases participate in radiant energy exchange. Next. indeed the presence of a medium will impede radiation transfer between surfaces. Very-long-wavelength radiation.1 and 100 microns is 359 . Radiation in the visible range is observed to have little penetrating power except in some liquids. One micron is 106 m or 3:94(10)5 in: The frequency. Cloud cover is observed to reduce maximum daytime temperatures and to increase minimum evening temperatures.1 illustrates the tremendous range of frequency and wavelength over which radiation occurs. where c is the speed of light. In addition. A second unique aspect of radiation is that both the amount of radiation and the quality of the radiation depend upon temperature. The changing optical properties of radiation with temperature are of paramount importance in determining the radiant-energy exchange between bodies. in radiation. surfaces that are opaque to visible radiation are easily traversed by gamma and x-rays. Short-wavelength radiation such as gamma rays and x-rays is associated with very high energies. we will discuss properties of surfaces and consider how system geometry inﬂuences radiant heat transfer. such as radio waves. also may pass through solids. both of which are dependent upon radiant energy transfer between earth and space. however.

There are two types of reﬂection that can occur. 23. Incident radiation Reflected radiation Absorbed radiation Transmitted radiation Figure 23.3 0. and t is called transmissivity. If r. in microns 0.1 The electromagnetic spectrum. .7 0.2 may be either absorbed. the angle of incidence of the radiation is equal to the angle of reﬂection. and t are the fractions of the incident radiation that are reﬂected. respectively.2 is specular reﬂection. The reﬂection shown in Figure 23. absorbed and transmitted. then rþaþt ¼1 (23-1) where r is called the reflectivity.8 Figure 23. the difference being caused by the different population of energy states in electrical conductors.2 Fate of radiation incident upon a surface. specular reﬂection and diffuse reﬂection. they reﬂect radiation in all directions.05 in.2 THERMAL RADIATION Thermal radiation incident upon a surface as shown in Figure 23. termed as thermal radiation. Diffuse reﬂection is sometimes likened to a situation in which the incident thermal radiation is absorbed and then reemitted from the surface. Most bodies do not reﬂect in a specular manner. a.4 Yellow Red 0. Absorption of thermal radiation in solids takes place in a very short distance.5 0. reﬂected.360 Chapter 23 Radiation Heat Transfer λ 10–6 10–5 Cosmic rays 10–4 Gamma rays 10–3 10–2 10–1 x-rays 100 Ultraviolet 101 102 103 104 105 106 Radio waves Infrared Visible spectrum Blue. still retaining its initial wavelength. The thermal band of the spectrum includes a portion of the ultraviolet and all of the infrared regions. In specular reﬂection. green 0. in electrical nonconductors. a is called the absorptivity.6 Wavelength. on the order of 1 mm in electrical conductors and about 0. or transmitted. which can absorb energy at thermal radiation frequencies.

3 The Intensity of Radiation 361 For most solids. A monochromatic absorptivity. whereas equilibrium refers to the equality of temperatures. or E ¼ El dl 0 The monochromatic emissivity. Kirchhoff’s law states that. The amount of energy traveling in a given direction is determined from I. Black bodies may also be made of bright objects. As the total emissive power includes radiant-energy contributions from all wavelengths. regardless of the nature of the interior surface. is called a black body. of a surface is deﬁned as the total rate of thermal energy emitted via radiation from a surface in all directions and at all wavelengths per unit surface area. so we may write e¼ E Eb (23-2) where Eb is the total emissive power of a black body. as can be shown by looking at a stack of razor blades. for which a ¼ 1. The radiant energy El contained between wavelengths l and l þ dl is the monochromatic emissive power.23. e. The monochromatic absorptivity is deﬁned as the ratio of the incident radiation of wavelength l that is absorbed by a surface to the incident radiation absorbed by a black surface. sharp edge forward. The emissivity.3. may be deﬁned in the same manner as the monochromatic emissivity. The total emissive power is also referred to elsewhere as the emittance or the total hemispheric intensity. al . A small hole in a large cavity closely approaches a black body. the monochromatic emissive power.3 THE INTENSITY OF RADIATION In order to characterize the quantity of radiation that travels from a surface along a speciﬁed path. the following equality holds for each surface: el ¼ a l (23-3) Thermodynamic equilibrium requires that all surfaces be at the same temperature so that there is no net heat transfer. equation (23-1) becomes. El . Steady state means that time derivatives are zero. Radiation incident to the hole has very little opportunity to be reﬂected back out of the hole. the transmissivity is zero. 23. is simply el ¼ El /El. the intensity of radiation. The total emissive power. where El. Kirchhoff’s law does not apply. r þ a ¼ 1: The ideally absorbing body. With reference to Figure 23. a so-called black body will appear black. thus Z 1 dE ¼ El dl. we . is deﬁned as the ratio of the total emissive power of a surface to the total emissive power of an ideally radiating surface at the same temperature. For radiation between bodies at greatly different temperatures. A frequent error in using Kirchhoff’s law arises from confusing thermal equilibrium with steady-state conditions. As we see reﬂected light (radiation). the concept of a single ray is not adequate. for an opaque body. In such situations the emissivity and the absorptivity may be assumed to be equal. A black body neither reﬂects nor transmits any thermal radiation. may also be defined.b. A relation between the absorptivity and the emissivity is given by Kirchhoff’s law. E. such as between Earth and the sun. The utility of Kirchhoff’s law lies in its use for situations in which the departure from equilibrium is small. el . The ideal radiating surface is also called a black body. Closely related to the total emissive power is the emissivity. no light being reﬂected from it.b is the monochromatic emissive power of a black body at wavelength l at the same temperature. for a system in thermodynamic equilibrium. and thus they may be called opaque to thermal radiation.

which will be discussed shortly. cos u sin u du df 0 0 or simply E ¼ pI If the surface does not radiate diffusely. Rearranging equation (23-4).3. A r0 sin q solid angle is defined by V ¼ A/r 2 or dV ¼ dA/r 2 . we see that the relation between the total emissive power. these being r. E ¼ dq/dA. where dV is a differential solid angle. a portion of space. in Figure 23. and f. the apparent size of the emitting area is dA cos u. and the intensity. Consiq der an imaginary hemisphere of radius r covering the plane surface on which dA is located. dA.3 The intensity of radiation. Our perspective will be that of an observer at point P looking at dA. I. the radial coordinate. is Z Z dq ¼ E ¼ I cos u dV ¼ I cos u dV (23-5) dA The relation is seen to be purely geometric for a diffusely radiating (I 6¼ I(u)) surface. and thus (r sin u df)(r du) ¼ sin u du df r2 The total emissive power per unit area becomes Z E ¼ I cos u dV r0 dV ¼ 2p Z p/2 Z ¼I dA f df Figure 23. Note that with the eye located at point P. If a unit area of surface. dA. It is important to remember that the intensity of radiation is independent of direction for a diffusely radiating surface. that is. Standard spherical coordinates will be used.3. emits a total energy dq. the azimuthal angle.4. the zenith angle shown in Figure 23. then the intensity of radiation is given by I d2 q dA dV cos u I dΩ q (23-4) dA Figure 23. dq The solid angle dV intersects the shaded area on the hemisphere as shown in Figure 23. then Z 2p Z p/2 E¼ I cos u sin u du df 0 0 (23-6) . u.362 Chapter 23 Radiation Heat Transfer P are interested in knowing the rate at which radiant energy is emitted from a representative portion.4 Integration of intensity over solid angles. of the surface shown in a prescribed direction.

Wien obtained this result in 1893.l vs. 2.4 Planck’s Law of Radiation 363 The relation between the intensity of radiation. c is the speed of light.5 the area under the curve of Eb. Planck. Eb. This functional relationship is plotted in Figure 23. The peak energy is also observed to occur at shorter and shorter wavelengths as the temperature is increased.6.1. For a black body at 5800 K (the effective temperature of solar radiation). is Ib. will be discussed in the next section.4 PLANCK’S LAW OF RADIATION Planck1 introduced the quantum concept in 1900 and with it the idea that radiation is emitted not in a continuous energy state but in discrete amounts or quanta.l ¼ 2pc2 hl5 ch exp 1 klT (23-7) Figure 23. functionally. we get Ebl ¼ T5 2p2 h(lT)5 ch exp 1 klT (23-8) where the quantity Ebl /T 5 is expressed as a function of the lT product. Weare ofteninterestedinknowing how much emissionoccursin aspeciﬁcportion ofthetotal wavelength spectrum.. derived by Planck. The peak energy is observed to be emitted at lT ¼ 2897:6 mm K(5215:6 mm R). Verh. We will now consider the means of such a description. as can be determined by maximizing equation (23-8). and discrete values of Ebl /sT 5 are given in Table 23. physik.l as a function of wavelength and temperature.l is the intensity of radiation from a black body between wavelengths l and l þ dl. lmax T ¼ 2897 m K. as already discussed. Equation (23-7) expresses.l ¼ 2c2 hl5 ch exp 1 klT where Ib. The relation. This is conveniently expressed as a fraction of the total emissive power. s. The fraction between wavelengths l1 and l2 is designated Fl1 l2 and may be expressed as R l2 R l2 l1 Ebl dl l Ebl dl ¼ 1 4 Fl1 l2 ¼ R 1 (23-9) sT 0 Ebl dl 1 M. Equation (23-6) relates intensity to emissive power that. potentially. h is Planck’s constant. .5 illustrates the spectral energy distribution of energy of a black body as given by equation (23-7). In Figure 23. is called Wien’s displacement law. 23. d. and T is the temperature. The total emissive power between wavelengths l and l þ dl is then E b . Radiation intensity is fundamental in formulating a quantitative description of radiant heat transfer but its deﬁnition. I. l (the total emitted energy) is seen to increase rapidly with temperature. Dividing both sides of this equation by T 5. is much easier to describe.23. Gesell. and the total emissive power is an important step in determining the total emissive power. 237 (1900). The intensity of radiation emitted by a black body. is cumbersome. k is the Boltzmann constant. a large part of the emitted energy is in the visible region. which can be treated as a single independent variable. 7 years prior to Planck’s development. deut. The constant.

555 Locus of maximum Values. Equation (23-10) .7mm) 10 12 Figure 23. the fraction of emission between any two wavelengths can be determined by subtraction. at a given temperature. By permission of the publishers. Third Edition. Howell. 5555 104 5000. (From R. mm Visible region (0. 000.Radiation Heat Transfer 108 108 107 107 106 105 104 103 102 Hemispherical spectral emissive power Eλb(l. This may be accomplished by using the fraction Ebl /sT 5 . T). T ) 101 101 100 0 Violet 2 4 5 8 Red Wavelength l. as discussed. Siegel and J. 1992.4–0. Thermal Radiation Heat Transfer. R. W/(m2 • µm) Chapter 23 Hemispherical spectral emissive power Elb(λ. T. 1667 103 1500. 2778 3000. K 10 . T).5 Spectral emissive power for a black body for several temperatures. This process can be simpliﬁed if the temperature is eliminated as a separate variable. e lb( lmax.) Equation (23-9) is conveniently broken into two integrals as follows: Z l2 Z l1 1 Ebl dl Ebl dl Fl1 l2 ¼ 4 sT 0 0 (23-10) ¼ F0l2 F0l1 So. Hemisphere Publishers. 833 102 1000. °R. W/m 2 • mm) 364 106 105 Blackbody temperature. Washington.

s.l dl ¼ ¼ sT 4 (23-12) Eb ¼ 15c2 h3 0 where s is called the Stefan–Boltzmann constant and has the value s ¼ 5:676 108 W/m2 K4 (0:1714 108 Btu/h ft2 R4 ). 1992.000 8000 6000 365 Planck’s law Wien’s distribution Rayleigh–Jeans'. distribution 240 × 10–15 Elb(l. This constant is observed to be a combination of other physical constants. Hemisphere Publishers.) may be modiﬁed in this manner to yield Z l2 T Z l1 T Ebl Ebl Fl1 Tl2 T ¼ d(lT) d(lT) 5 5 sT sT 0 0 ¼ F0l2 T F0l1 T (23-11) Values of F0lT are given as functions of the product. By permission of the publishers. and its relation to other physical constants were obtained after the presentation of Planck’s law in 1900. lT. The exact value of the Stefan–Boltzmann constant. The Stefan–Boltzmann relation.000 × 10–15 Stefan–Boltzmann Law lT. F0 – λT 5216 7398 11. The result is Z 1 2p5 k4 T 4 Eb. Thermal Radiation Heat Transfer.890 99 160 120 80 4000 2000 0 40 0 2 4 6 8 1 2 4 6 3 104 Wavelength–temperature product lT. R. (From R.6 Spectral energy distribution for a black body as a function of lT.23. Btu/(h • ft 2 • 8R 5 • mm) 14. T)/ T5.5 STEFAN–BOLTZMANN LAW Planck’s law of radiation may be integrated over wavelengths from zero to inﬁnity to determine the total emissive power. Third Edition. Eb ¼ sT 4 .1. µm • °R 1 6 8 1 2 4 6 8 10 20 Wavelength–temperature product lT. Washington. (µm)(°R) 2606 lT. T)/ T5. via experiment by Stefan in 1879 and via a thermodynamic derivation by Boltzmann in 1884. W/(m2 • K 5 • mm) 12. . was obtained prior to Planck’s law.5 Elb(l.000 10. Siegel and J.200 22.067 2898 4107 6148 25 50 75 41. Howell. (µm)(K ) 1448 Percent 200 emissive power below lT. 23. in Table 23. µm • K 40 3 103 Figure 23.

366 Chapter 23 Radiation Heat Transfer Table 23.7543 0.4679 1.1797 1.8871 0.7961 0.8081 0.0393 0.5327 0.0429 1.8485 0.4449 0.2175 2.2688 7000 7100 7200 7300 7400 0.1462 2.7825 0.6803 1.3472 4000 4100 4200 4300 4400 0.6461 0.0021 0.3028 0.2739 0.9552 0.8562 0.1 Planck radiation functions Eb 1 sT 5 cm K Eb 1 sT 5 cm K lT(mm K) F0lT 0.7291 1.9054 0.9099 0.4222 5500 5600 5700 5800 5900 0.0128 0.0667 0.2117 lT(mm K) F0lT 1000 1100 1200 1300 1400 .0197 0.7111 1.5552 0.8157 1.7692 0.3089 9000 9100 9200 9300 9400 0.2255 0.5807 8500 8600 8700 8800 8900 0.1402 1.2650 0.0339 0.0939 2.2447 2.6337 0.4624 2.5488 1.9324 1.6387 1.1200 0.2404 0.7010 0.5348 0.2107 1.7461 0.3594 0.9888 1.2958 0.1646 0.1190 1.8295 0.0285 0.2774 0.4291 0.9181 0.9938 0.2565 0.0372 0.3995 0.2328 0.2623 2.7108 0.8244 0.6220 0.5643 0.1831 0.4987 0.9030 0.8746 0.8928 0.8192 0.0003 0.0521 0.6075 0.7378 0.5238 1.6462 0.2505 2.3713 6000 6100 6200 6300 6400 0.8712 0.3856 0.8602 0.5499 1.9717 2.4036 0.8137 0.3132 0.9809 1.8344 0.8521 1.4614 0.2185 0.4135 1.3604 1.8745 8000 8100 8200 8300 8400 0.7255 2000 2100 2200 2300 2400 0.8981 0.4140 3500 3600 3700 3800 3900 0.2279 0.7541 0.8810 0.0830 0.7565 1.5160 0.6979 0.3241 0.4434 0.2832 0.7844 0.6974 1.6909 0.7763 0.8676 0.8900 0.2409 2.8022 0.8480 0.9077 0.3722 0.0078 0.7897 0.4965 3000 3100 3200 3300 3400 0.8158 0.5765 0.1009 0.4238 0.2028 2.6209 1.6694 0.8640 0.0756 9500 9600 9700 9800 9900 0.8522 0.3401 0.9121 0.7201 0.3617 2.8391 0.8779 0.0009 0.2928 4500 4600 4700 4800 4900 0.4786 0.6715 0.5933 0.3829 0.1640 1.7832 0.8436 0.2483 5000 5100 5200 5300 5400 0.5987 2500 2600 2700 2800 2900 0.5937 0.8841 0.0695 6500 6600 6700 6800 6900 0.3354 0.4809 0.2732 0.2590 1.0855 0.2624 2.5793 0.1613 0.1408 7500 7600 7700 7800 7900 0.6579 0.3181 0.0043 0.7618 0.5152 0.9006 0.8826 1500 1600 1700 1800 1900 0.1824 2.8955 0.2053 0.

000 41.9978 0.9972 0.000 21.000 36. Eb ¼ sT 4 . Consequently.9961 0.0087 0.9901 0.0285 0.000 14.000 11.1145 0.0432 0.6 EMISSIVITY AND ABSORPTIVITY OF SOLID SURFACES Whereas thermal conductivity.9985 0.9953 0.0050 45.000 23. speciﬁc heat.9689 0.0075 0.0015 0.0065 0.0009 0.0540 0.0039 0.9984 0.000 31.9834 0.9142 0.0031 0. Figure 23. Q.9976 0.9922 0. By deﬁnition. Brewster.9974 0.0008 0.000 44.000 48.0009 0.9930 0.000 Eb 1 sT 5 cm K (From M.9987 0.0684 30.9318 0. New York.9982 0. . density.000 38.0235 35.0010 25.9889 0.0014 0.000 18. E ¼ eEb .0118 0.0057 0. Thermal Radiative Transfer and Properties. For actual surfaces.9912 0.9943 0. Monochromatic Emissivity.000 0.000 0.9738 0.9628 0.0349 0.9979 0. is a gross factor.0025 0.0022 0. The emissivity of the surface.000 0.000 33.0139 0.000 49.000 47.9551 0. From preceding sections it is seen that.0007 lT(mm K) F0lT 10.9986 0.000 32.000 28.000 17.0020 0. as radiant energy is being sent out from a body not only in all directions but also over various wavelengths.9808 0. John Wiley & Sons.9777 0.0012 0.1518 0.9451 0.0007 0. For actual surfaces.9988 0.9970 0.000 13.9873 0.000 29.2052 0. 23.000 16. following the deﬁnition of emissivity.000 39. for black-body radiation.6 Emissivity and Absorptivity of Solid Surfaces 367 Eb 1 sT 5 cm K lT(mm K) F0lT 0. so deﬁned.9988 0.000 34. 1992.000 12.000 0.7 Emissivity at various wavelengths.0101 40.000 24.23.000 43.000 0.9937 0.0018 0.9957 0.9967 0.000 42.9856 0.9964 0.0878 0.b P Q 0 λ1 λ Figure 23.000 19.9983 0. and viscosity are the important physical properties of matter in heat conduction and convection.000 0.000 0.9981 0.7 represents a typical distribution of the intensity of radiation of two Eλ.0016 20.000 22.0028 15.9948 0.0196 0.0164 0.0044 0.000 27.0011 0. the monochromatic emissivity of an actual surface is the ratio of its monochromatic emissive power to that of a black surface at the same temperature.000 37.0035 0.000 26. the emissivity may vary with wavelength as well as the direction of emission. emissivity and absorptivity are the controlling properties in heat exchange by radiation.000 46. By permission of the publishers). we have to differentiate the monochromatic emissivity el and the directional emissivity eu from the total emissivity e.

The emissivity stays fairly constant in the neighborhood of the normal direction of emission. (a) Wet ice. This is the direct consequence of Kirchhoff’s law.2 0 0.04 0. Most nonconductors have much smaller emissivities for emission angles in the neighborhood of 908 (see Figure 23.14 εq Figure 23. Deviation from the cosine law is even greater for many conductors (see Figure 23.8 Emissivity variation with direction for nonconductors.04 0. equation (23-5). (g) Aluminum oxide.10 0. The cosine variation discussed previously. l1 .9).0 εq 0. If the cosine law is fulﬁlled. The total emissivity of the surface is given by the ratio of the shaded area shown in Figure 23.08 0.6 0. This is due to the fact that the emissivity (averaged over all wavelengths) of actual surfaces is not a constant in all directions.4 0. the distribution curves should take the form of semicircles.14 0. 60° 40° 20° 0° 20° 40° 60° Mn Cr Ni polished Al Ni dull 0.8). The variation of emissivity of materials with the direction of emission can be conveniently represented by polar diagrams.06 0.02 0. (b) Wood.6 0.2 0. .4 0.02 εq 0 0.06 0. it ﬁrst increases and then decreases as the former approaches 908. such surfaces at the same temperature over various wavelengths. as the emission angle is increased.0 Figure 23. (c) Glass. (f) Copper oxide. The monochromatic emissivity at a certain wavelength. is strictly applicable to radiation from a black surface but fulﬁlled only approximately by materials present in nature. (e) Clay. Directional Emissivity.368 Chapter 23 Radiation Heat Transfer 0° 20° 20° (e) 40° (g) (d) 40° (b) (c) 60° 60° (f ) (a) 80° 1.9 Emissivity variation with direction for conductors.12 0. is seen to be the ratio of two ordinates such as OQ and OP.12 0.10 0. (d) Paper. That is el1 ¼ OQ OP which is equal to the monochromatic absorptivity al1 from radiation of a body at the same temperature.08 0.8 0.7 to that under the curve for the black-body radiation.8 80° εq 1.

2 The ratio e/en for bright metallic surfaces 0:049 ¼ 1:25 0:039 0:046 ¼ 1:12 0:041 0:053 ¼ 1:18 0:045 0:057 ¼ 1:19 0:048 0:071 ¼ 1:22 0:058 0:158 ¼ 1:23 0:128 0:340 ¼ 1:08 0:336 Aluminum. emissivity depends on surface conditions.3 The ratio e/en for nonmetallic and other surfaces Copper oxide (3008F) Fire clay (1838F) Paper (2008F) Plywood (1588F) Glass (2008F) Ice (328F) 0.97 0. is.2 lists the ratio of e/en for a few representative bright metallic surfaces. in general.95 A few generalizations may be made concerning the emissivity of surfaces: (a) In general. It has been found that for most bright metallic surfaces. polished (423 K) Iron.3). e. the normal emissivity values can be used. (f) The emissivities of nonmetallic surfaces are much higher than for metallic surfaces and show a decrease as temperature increases. (d) The formation of a thick oxide layer and roughening of the surface increase the emissivity appreciably.2 can be taken as a good average. bright rolled (392 K) Chromium. Table 23. The value 1. (b) The emissivity of highly polished metallic surfaces is very low. bright rolled (443 K) Nickel. For nonmetallic or other surfaces.99 0.97 0. bright (353 K) The average total emissivity may be determined by using the following expression: Z p/2 e¼ eu sin 2u du 0 The emissivity. different from the normal emissivity. en (emissivity in the normal direction). Table 23. Because of the inconsistency that can often be found among various sources. bright etched (423 K) Bismuth. (c) The emissivity of all metallic surfaces increases with temperature.93 0. (e) The ratio e/en is always greater than unity for bright metallic surfaces.23. bright matte (373 K) Nickel. the ratio e/en is slightly less than unity.6 Emissivity and Absorptivity of Solid Surfaces 369 Table 23. without appreciable error.96 0. for total emissivity (see Table 23. polished (373 K) Manganin. the total emissivity is approximately 20% higher than en . .

For a gray surface. the emissivity and absorptivity of ordinary surfaces may differ widely. Absorptivity.05. the total average absorptivity will be independent of the spectralenergy distribution of the incident radiation. When exposed to solar radiation. the surface is called gray. may be employed. Like emissivity. say solar radiation ( 5800 K). for several materials. if the incident radiation is that from a very-high-temperature source.4. may be used in place of a. If al is a constant and thus independent of l. in particular. and dark linoleum. However. That is. at various temperatures.2 or even 0. Mg. al . the ﬁrst subscript referring to the temperature of the receiving surface and the second subscript to the temperature of the incident radiation. freshly polished metallic surfaces have an emissivity value (and absorptivity under equilibrium conditions) of about 0. even though the temperatures of the incident radiation and the receiver are not the same. 23. in addition. It may be remarked once again that Kirchhoff’s law holds strictly true under thermal equilibrium. on the quality of the incident radiation. Good approximations of a gray surface are slate.2 . The absorptivity of a surface depends on the factors affecting the emissivity and. Fe. Contrary to the above.370 Chapter 23 Radiation Heat Transfer (g) The emissivities of colored oxides of heavy metals like Zn. but their absorptivity drops sharply to 0.10 Radiant heat transfer between two surfaces. a1. Consequently. the monochromatic absorptivity. dA2 A2 .95 at ordinary temperature. The radiant energy emitted from a black surface at dA1 and received at dA2 is dq1!2 ¼ Ib1 cos u1 dV12 dA1 q2 r q1 dA1 A1 Figure 23.15 if these oxides are exposed to solar radiation. and Al. For most materials. Gray surfaces.4 lists emissivities. Table 23.10. their absorptivity increases to 0. White metal oxides usually exhibit an emissivity (and absorptivity) value of about 0. the emissivity. then a ¼ e. if a body at temperature T1 is receiving radiation from a black body also at temperature T1 .7 RADIANT HEAT TRANSFER BETWEEN BLACK BODIES The exchange of energy between black bodies is dependent upon the temperature difference and the geometry with the geometry. tar board. of a surface may vary with wavelength. Consider the two surfaces illustrated in Figure 23. playing a dominant role. Under these latter circumstances a double-subscript notation. e. and Cr are much larger than emissivities of white oxides of light metals like Ca. in the usual range of temperature encountered in practice (from room temperature up to about 1370 K) the simple relationship a ¼ e holds with good accuracy.

polished plate Strip Filament Wire Silver: Polished.020–0.28 67 100–700 0.28 260–440 75 0.12–0. highly polished Iron and steel (not including stainless): Metallic surfaces (or very thin oxide layer) Iron.052 0. Hottel)y T.96%). pure Polished 440–1070 212 390–1110 200–940 0.20–0.31 100–600 390–1110 0.69 0.192 0.66–0.032 0. unoxidized Gray oxidized Nickel alloys: Chromnickel Copper–nickel. 8Fz Surface Emissivity A. polished Nichrome wire.104 0.57 0.21 0. oxidized Platinum: Pure.08–0.65–0.13 440–1160 0.79 0.057 0.073–0.64–0.4 Normal total emissivity of various surfaces (Compiled by H. Metals and their oxides Aluminum: Highly polished plate.018–0.95–0.182 440–1160 100–700 0.059 0.039–0.3% pure Commercial sheet Oxidized at 11108F Heavily oxidized Brass: Polished Oxidized by heating at 11108F Chromium (see nickel alloys for Ni–Cr steels): Polished Copper Polished Plate heated at 11108F Cuprous oxide Molten copper Gold: Pure.09 0.14–0.075 0. completely rusted Steel plate.19 0. C. highly polished Oxidized surfaces Iron plate.28 125–1894 212 120–1830 120–930 0.38 0.61–0.97 2370–2550 2910–3270 0. 98.76 0.17 0.031 (Continued ) .59 100–2000 0.94–0.29 0.7 Radiant Heat Transfer Between Black Bodies 371 Table 23. bright Nichrome wire.98 440–1160 1700–2960 80–2240 440–2510 0.36 212 390–1110 1470–2010 1970–2330 0.057–0.10 0.022–0. rough Molten surfaces Cast iron Mild steel Lead: Pure (99.054–0.16–0.036–0.54 0.23. polished Wrought iron.11–0.035 800–1880 392 100–480 0. polished Cast iron.

Metals and their oxides Stainless steels: Polished Type 310 (25 Cr. polished 72 Oak.1% pure.35 . splotched. and miscellaneous Asbestos: Board 74 Paper 100–700 Brick Red.90 0. and soda 500–1000 Gypsum.96 0.91 0.06 0.85 0. lead. sprayed on iron 76 Black shiny shellac on tinned iron sheet 70 Black matte shellac 170–295 Black or white lacquer 100–200 Flat black lacquer 100–200 Oil paints.94 0. 8Fz Emissivity 212 0.97 0.875 0. thick on smooth or blackened plate 70 Magnesite refractory brick 1832 Marble. 20 Ni) Brown.066 0.064 0.372 Chapter 23 Radiation Heat Transfer Table 23.93 0.032–0.4 Normal total emissivity of various surfaces (Compiled by H. light gray. 0.80–0.07.053 0.98 0.35 0.526 0.38 0.906 0.11 0.93–0. planed 70 Paints. C.95–0.96–0.39 0.96–0.75 0.043 and 0.02 in. building materials. glazed 2012 Building 1832 Fireclay 1832 Carbon: Filament 1900–2560 Lampblack-waterglass coating 209–440 Thin layer of same on iron plate 69 Glass: Smooth 72 Pyrex.95 0. but no gross irregularities 70 Brick.045–0.074 A.903 0.96 0.93 0. rough. Hottel)y Surface T. 0. all colors 212 A1 paint. oxidized from furnace service Tin: Bright tinned iron Bright Commercial tin-plated sheet iron Tungsten: Filament.75 0. aged Filament Polished coat Zinc: Commercial 99.821 0. varnishes: Snow-white enamel varnish on rough iron plate 73 Black shiny lacquer. Refractories.45 0. 16 different.94 0. paints.90–0.95 0. after heating to 6208F 300–600 0.92–0. polished Oxidized by heating at 7508F 420–980 76 122 212 80–6000 6000 212 440–620 750 B. lacquers.927 0.08 0.

Refractories.). and miscellaneous Plaster. Table of normal total emissivity compiled by H. as has been used throughout the text thus far. rough lime Rooﬁng paper Rubber: Hard. building materials. C.91 0.963 y By permission from W. Hottel. Table 23. Thus dV12 ¼ cos u2 dA2 r2 and as Ib1 ¼ Eb1 /p.95–0. H. may be determined. McAdams (ed. in contrast to K.86 0. This is n cos u cos u dA o 2 1 1 dq2!1 ¼ Eb2 dA2 pr 2 The net heat transfer between surfaces dA1 and dA2 is then simply dq12 net ¼ dq1Ð2 ¼ dq1!2 dq2!1 or dq1Ð2 ¼ (Eb1 Eb2 ) cos u1 cos u2 dA1 dA2 pr 2 Integrating over surfaces 1 and 2. they correspond and linear interpolation is permissible. In exactly the same manner the energy emitted by dA2 and captured by dA1. Heat Transmission. z When temperatures and emissivities appear in pairs separated by dashes. glossy plate Soft.7 Radiant Heat Transfer Between Black Bodies 373 Table 23.4 (Continued) T. rough (reclaimed) Water 50–190 69 0. 1954. The reader should note that the units of temperatures in Table 23.23. where dV12 is the solid angle subtended by dA2 as seen from dA1 . paints. McGraw-Hill Book Company.4 is presented as originally published by McAdams. gray. we obtain Z Z cos u1 cos u2 dA2 dA1 q1Ð2 ¼ (Eb1 Eb2 ) pr 2 A1 A2 the insertion of A1 /A1 yields q1Ð2 ¼ (Eb1 Eb2 )A1 1 A1 Z Z A1 cos u1 cos u2 dA2 dA1 pr 2 A2 (23-13) .4 are F.94 0. the heat transfer from 1 to 2 is n cos u cos u dA o 1 2 2 dq1!2 ¼ Eb1 dA1 pr 2 The bracketed term is seen to depend solely upon geometry.91 74 76 32–212 0. Third Edition. 8Fz Surface Emissivity B.

It is purely geometric. and A1 ¼ 2pr02 . A physical interpretation of the view factor may be obtained from the following argument. the net heat transfer is not affected by these operations. The view factor is independent of temperature. 1. A2 ¼ pr02 . Determine the view factors F11. For surface 1 we may write F11 þ F12 ¼ 1 and A1 F12 ¼ A2 F21 . then the view factor would be F21. Clearly. A1 F12 ¼ A2 F21 . and thus A1 F12 ¼ A2 F21 . F12. For an enclosure. This simple but extremely important expression is called the reciprocity relationship. The net rate of heat transfer between A1 and A2 is the difference or Eb1 A1 F12 Eb2 A2 F21. be determined by integration. Before some speciﬁc view factors are examined. If we had used A2 as a reference.374 Chapter 23 Radiation Heat Transfer The bracketed term in the above equation is called the view factor F12. EXAMPLE 1 The view factor F21 is unity. An example of such a case follows. In many cases the view factor may be determined without integration. there are several generalizations worthy of note concerning view factors. The reciprocity relation. As 1 F12 A1 Z Z A1 cos u1 cos u2 dA2 dA1 pr 2 A2 (23-14) this integration process becomes quite tedious. as surface 2 sees only surface 1. and the view factor for a complex geometry is seen to require numerical methods. the amount of heat that surface A2 receives is Eb1 A1 F12 . F11 þ F12 þ F13 þ ¼ 1. the view factor F12 can be interpreted as the fraction of black-body energy leaving A1 which reaches A2. Clearly the view factor cannot exceed unity. whereas the amount that reaches A1 is Eb2 A2 F21 . the above relations give Surface 1 F12 ¼ F21 r0 pr02 A2 ¼ (1) A1 2pr02 ! ¼ 1 2 and Surface 2 F11 ¼ 1 F12 ¼ 1 2 The view factor F12 can. In order to illustrate the analytical evaluation of view factors. As the total energy leaving surface A1 is Eb A1 . This may be arranged to yield (Eb1 Eb2 )A1 F12 . 2. consider the view factor between the differential area dA1 and the parallel plane A2 shown in Figure 23.11. The amount of heat lost by surface A2 is Eb2 A2 . The view factor FdA1 A2 is given by FdA1 A2 1 ¼ dA1 Z dA1 Z cos u1 cos u2 dA2 dA1 pr 2 A2 . and F21. in general. As F21 ¼ 1. Consider the view factor between a hemisphere and a plane as shown in the ﬁgure. Thus. 3. is always valid.

0 4. where r 2 ¼ D2 þ x2 þ y2 .03 L2 D 0.5 distance from dA to rectangle 2.12. 3.11 Differential area and parallel-ﬁnite area.0 5.0 L1 0. Engrg.16 0. The resulting integral becomes Z Z 1 L1 L2 D2 dx dy FdA1 A2 ¼ 2 2 2 2 p 0 0 (D þ x þ y ) or FdA1 A2 ( ) 1 L1 L2 L2 L1 1 1 qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ tan qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ þ qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ tan qﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ ¼ (23-15) 2p D2 þ L21 D2 þ L21 D2 þ L22 D2 þ L22 The view factor given by equation (23-15) is shown graphically in Figure 23.04 dA 0.15 also illustrate some view factors for simple geometries. hence Z 1 cos u1 cos u2 FdA1 A2 ¼ dA2 p A2 r2 Also.’’ Mech. By permission of the publishers.10 0.14 1. Figures 23. C.12 0.5 D/L1. D is 2.5 3.0 1.20 0.12 View factor for a surface element and a rectangular surface parallel to it.5 5. dA1 and as A2 dA1 the view of dA2 from dA1 is independent of the position on dA1.18 0.23. Hottel.5 0.7 y Radiant Heat Transfer Between Black Bodies 375 L2 A2 dA2 L1 D x q1 q2 r Figure 23. (From H.5 FdA1 – A2 = 0. 52 (1930).22 0. Dimension ratio 4. ‘‘Radiant Heat Transmission.13–23. it may be noted that u1 ¼ u2 and cos u ¼ D/r.5 Figure 23.02 D/L2.5 2.08 0.06 0 0 0.0 3.24 0.5 0. Dimension ratio 3.) .05 1.5 1.0 L1 and L2 are sides of rectangle.0 2.0 0..

0 6. C.0 0. 5-6-7-8 Planes connected by nonconducting but reradiating – wall.0 Asymptotes Scale changes here 8.6 0.0 – View factor F or F 376 0.0 2. 4. ‘‘Radiant Heat Transmission.50 Y = 0. 7. F.4 Radiation between parallel planes. C. The curves labeled 5. narrow rectangles 2 1 0. Engrg. 2. Hottel. smaller side or diameter distance between planes 6 7 Figure 23. Hottel. rectangles. By permission of the publishers. Engrg.0 3.0 3.2 0.2 0 0 1 2 3 4 5 Ratio. 52 (1930).0 Dimension ratio. ‘‘Radiant Heat Transmission.. (From H.) .40 0.Radiation Heat Transfer 0.. 52 (1930).5 2. directly opposed 1-2-3-4 Direct radiation between the planes. (From H.30 0.0 1.’’ Mech. 6.’’ Mech.10 0 0 4. Y = 0. By permission of the publishers.3 0. 1. 6 Squares. Z 6 8 10 A2 z x y A1 A1 = Area on which heattransfer equation is based Y = y/x Z = z /x Figure 23.20 1.13 View factor for adjacent rectangles in perpendicular planes.8 0. 3.8 8 7 6 0. and disks. 8 long.) 1.0 1.0 4.14 View factors for equal and parallel squares. 7 2:1 Rectangle.1 Dimension ratio. F. and 8 allow for continuous variation in the sidewall temperatures from top to bottom. 5 Disks.1 0.4 View factor F12 Chapter 23 0.6 5 4 3 0.

4 0.and ﬁnite-size areas have been expressed in equation form thus far. EXAMPLE 2 Determine the view factor from a 1 m square to a parallel rectangular plane 10 m by 12 m centered 8 m above the 1 m square. The total view factor is the sum of the view factors or 0. using the symbol Gij.09.4 r1 1. at ﬁrst glance.4 0. will be incident on the other surfaces that it can ‘‘see.3 0.36.2 0. designated i.2 r2 0.’’ If there are n surfaces in total. and Figure 23.5 0. Using D ¼ 8.6 0.12 may be used.2 0 0.5 D 0. L2 ¼ 5. The smaller area may be considered a differential area. View-Factor Algebra View factors between combinations of differential.8 6 5 FA1 – A2 0.23. with j designating any surface that receives energy from i. A simpliﬁed notation will be introduced.12 is 0. L1 ¼ 6.6 0.15 View factors for parallel opposed circular disks of unequal size.0 1.8 2 0. the total view factor is the sum of the view factors to each subdivided rectangle. Some generalizations can be made that will be useful in evaluating radiant energy exchange in cases that.7 377 Radiant Heat Transfer Between Black Bodies 1.0 10 8 0.8 1 2 D/r1 4 6 8 10 20 Figure 23.1 0. we ﬁnd that the view factor from Figure 23. In an enclosure all energy leaving one surface.6 1 r2/O = 1. The 10 m by 12 m area may be divided into four 5 m by 6 m rectangles directly over the smaller area. seem quite difﬁcult.2 4 3 2 0. Thus. we may write n å Fij ¼ 1 (23-16) j¼1 A general form of the reciprocity relationship may be written as Ai Fij ¼ Aj Fji (23-17) these two expressions form the basis of a technique designated view-factor algebra. deﬁned as Gij Ai Fij .

can be obtained using view-factor algebra in the following steps. 2 2 2m 4 5m 3 3m 3 1 2m 3m 2m 1 2m 3m 2m Inspection indicates that.13. we may write G1(2þ3) ¼ G12 þ G13 (23-20) This relation says simply that the energy leaving surface 1 and striking both surfaces 2 and 3 is the total of that striking each separately. G21 ¼ G2(1þ3) G23 . in case (a). Some special symbolism will now be explained.378 Chapter 23 Radiation Heat Transfer Equations (23-16) and (23-17) may now be written as åGij ¼ Ai Gij ¼ Gji (23-18) (23-19) The quantity Gij is designated the geometric flux. If surface 1 ‘‘sees’’ two surfaces. EXAMPLE 3 Determine the view factors. view factors F23 and F2ð1þ3Þ can be read directly from Figure 23. for the ﬁnite areas shown. is reduced to G(1þ2)(3þ4) ¼ G1(3þ4) þ G2(3þ4) which decomposes further to the form G(1þ2)(3þ4) ¼ G13 þ G14 þ G23 þ G24 Examples of how view-factor algebra can be used follow. designated 2 and 3. The desired view factor. involving four surfaces. Relations involving geometric fluxes are dictated by energy conservation principles. F12 . F12 . Equation (23-20) can be reduced further to A1 F1(2þ3) ¼ A1 F12 þ A1 F13 or F1(2þ3) ¼ F12 þ F13 A second expression. G2(1þ3) ¼ G21 þ G23 Thus.

a surface that views n other surfaces may be described according to F11 þ F12 þ þ F1i þ þ F1n ¼ 1 or n å F1i ¼ 1 i¼1 (23-21) . we have 5 (0:15 0:10) ¼ 0:125 2 3 ¼ (0:22 0:165) ¼ 0:0825 2 F1(2þ4) ¼ F14 The solution to case (b) now becomes F12 ¼ 0:125 0:0825 ¼ 0:0425 23.13 we read F2(1þ3) ¼ 0:15 F23 ¼ 0:10 Thus. the appropriate values are F(2þ4)(1þ3) ¼ 0:15 F4(1þ3) ¼ 0:22 F43 ¼ 0:165 F(2þ4)3 ¼ 0:10 Making these substitutions.23. for conﬁguration (a). by reciprocity. we may solve for F12 according to G12 ¼ G21 ¼ G2(1þ3) G23 A1 F12 ¼ A2 F2(1þ3) A2 F23 A2 F23 ½F A1 2(1þ3) F12 ¼ From Figure 23.13. the solution steps are G12 ¼ G1(2þ4) G14 which may be written as F12 ¼ F1(2þ4) F14 The result from part (a) can now be utilized to write A2 þ A4 ½F(2þ4)(1þ3) F(2þ4)3 A1 A4 F43 ¼ ½F A1 4(1þ3) F1(2þ4) ¼ F14 Each of the view factors on the right side of these two expressions may be evaluated from Figure 23.8 RADIANT EXCHANGE IN BLACK ENCLOSURES As pointed out earlier. for case (b).8 Radiant Exchange in Black Enclosures 379 Finally. we obtain F12 ¼ 5 (0:15 0:10) ¼ 0:125 2 Now.

respectively. 23.16 show a path to ground at each of the junctions. have electrical counterparts I. Eb4 1 A1F1 – 3 Eb3 1 A1F1 – 3 Eb1 1 A2 F2 – 3 1 A1F1 – 2 1 A3F3 – 4 Eb3 1 A2F2 – 3 1 A1F1 – 4 1 A2F2 – 4 Eb2 Eb1 1 A1F1 – 2 Eb2 Figure 23.380 Chapter 23 Radiation Heat Transfer Obviously. and the thermal resistance. n n i¼1 i¼1 q1others ¼ å q1i ¼ å A1 F1i (Eb1 Ebi ) (23-25) Equation (23-25) can be thought of as an analog to Ohm’s law where the quantity of transfer. In such situations one would resort to numerical methods. q23. the potential driving force. respectively. designated i. When analyzing enclosures with four or more surfaces. an analytical solution becomes impractical. we may write for the net heat transfer with 1. Eb1 Ebi . Figure 23. The thermal analog is a surface that has some external inﬂuence whereby its temperature is maintained at a certain level by the addition or rejection of energy. 1/A1 F1i . in a black enclosure the radiant exchange is given as q1i ¼ A1 F1i (Eb1 Ebi ) (23-24) For an enclosure where surface 1 views n other surfaces. although somewhat tedious.16 depicts the analogous electrical circuits for enclosures with three and four surfaces.16 Radiation analogs. and R. that is. . can be accomplished in reasonable time. q.9 RADIANT EXCHANGE WITH RERADIATING SURFACES PRESENT The circuit diagrams shown in Figure 23. q13. The solution to a three-surface problem. Such a surface is in contact with its surroundings and will conduct heat by virtue of an imposed temperature difference across it. the inclusion of A1 with equation (23-12) yields n å A1 F1i ¼ A1 (23-22) i¼1 Between any two black surfaces the radiant heat exchange rate is given by q12 ¼ A1 F12 (Eb1 Eb2 ) ¼ A2 F21 (Eb1 Eb2 ) (23-23) For surface 1 and any other surface. DV. to ﬁnd q12.

For gray bodies. G. irradiation.18 Heat transfer at a surface. determination of heat transfer becomes Incident radiation. the reradiating view factor. The irradiation. J The net heat transfer from the surface shown in Figure 23. G more involved.10 RADIANT HEAT TRANSFER BETWEEN GRAY SURFACES In the case of surfaces that are not black. Figure 23. the electrical analog may be used by the simple modiﬁcation that no path to ground exists at the reradiating surface. is seen equivalent to the square-bracketed term in the previous expression. we encounter surfaces that effectively are insulated from the surroundings. we have Eb1 Eb2 Requiv ¼ A1 F12 þ 381 3 1 A1F1 – 3 Eb1 1 A2F2 – 3 1 A1F1 – 2 Eb2 Figure 23. F12. F 12 .17 q12 ¼ 1 (Eb1 Eb2 ) 1/A1 F13 þ 1/A2 F23 1 ¼ A1 F12 þ (Eb1 Eb2 ) 1/F13 þ A1 /A2 F23 (23-26) ¼ A1 F 12 (Eb1 Eb2 ) The resulting expression. equation (23-62). deﬁned as the rate at which radiation leaves a given surface per unit area. contains a new term. F 12 . is deﬁned as the rate at which radiation is incident on a surface per unit area. Evaluating the net heat transfer between the two black surfaces. 23. The radiosity. and the total emissive power are related by J ¼ rG þ eEb (23-27) .18 is determined by the difference between the radiation leaving the surface and the radiation incident upon the surface. This new factor. is Figure 23. the radiosity. Figure 23. In other situations where curves such as in this ﬁgure are not available. Such a surface will reemit all radiant energy that is absorbed— usually in a diffuse fashion. J. q12 . which includes direct exchange between surfaces 1 and 2. For a gray body.23.10 Radiant Heat Transfer Between Gray Surfaces In radiation applications. These surfaces thus act as reﬂectors and their temperatures ‘‘ﬂoat’’ at some value that is required for the system to be in equilibrium. plus terms that account for the energy that is exchanged between these surfaces via the intervening reradiating surface. surface.17 shows a physical situation and the corresponding electric analog for a three-surface enclosure with one being a nonabsorbing reradiating surface. surfaces for which the absorptivity and emissivity are independent of wavelength.14 allows reradiating view factors to be read directly for some simple geometries. that is. Radiation leaving considerable simpliﬁcations can be made. It is apparent that F 12 will always be greater than F12.

. Such an expression is q¼ Eb 1 Eb 2 r1 /A1 e1 þ 1/A1 F12 þ r2 /A2 e2 (23-31) where the terms in the denominator are the equivalent resistances due to the characteristics of surface 1. where the net heat leaving a surface can r /ε A be thought of in terms of a current.19 illustrates this analogy. Setting a ¼ e. From equation (23-17) we may write q1Ð2 ¼ A1 F12 (J1 J 2 ) ¼ A2 F21 (J1 J2 ) We may now write the net heat exchange in terms of the different ‘‘resistances’’ offered by each part of the heat transfer path as follows: A1 e1 (Eb1 J1 ) r1 Rate of heat leaving surface 1: q¼ Rate of heat exchange between surfaces 1 and 2: q ¼ A1 F12 (J1 J2 ) Rate of heat received at surface 2: q¼ A2 e2 (J2 Eb2 ) r2 If surfaces 1 and 2 view each other and no others then each of the qs in the previous equations is equivalent. In such a case an additional expression for q can be written in terms of the overall driving force. Now the net exchange of heat via radiation between two surfaces will depend upon their radiosities and their relative ‘‘views’’ of each other. geometry. an important simplification may be made in equation (23-29). and the quotient r/eA may be termed a resistance. The net heat transfer from a surface is qnet ¼ J G ¼ eEb þ rG G ¼ eEb (1 r)G (23-28) A In most cases it is useful to eliminate G from equation (23-28).382 Chapter 23 Radiation Heat Transfer where r is the reflectivity and e is the emissivity.19 Electrical analogy for Eb J may be likened to a potential difference. we obtain qnet ¼ Ae (Eb J) r (23-30) Eb J (Eb – J) which suggests an analogy with Ohm’s law. the difference Figure 23. Eb1 Eb2 .20. This yields qnet (J eEb ) ¼ eEb (1 r) r A As a þ r ¼ 1 for an opaque surface qnet eEb aJ ¼ r A r (23-29) When the emissivity and absorptivity can be considered equal. The electrical analog to this equation is portrayed in Figure 23. Figure 23. qnet = r /ε A V ¼ IR. and the characteristics of surface 2. radiation from a surface. respectively.

Examples 4 and 5. 2. Each surface must be gray. Kirchhoff’s law must apply. A simple series electrical circuit is useful in solving this problem. which follow. There is no heat absorbing medium between the participating surfaces.10 Eb1 J1 R= r1 383 Eb2 J2 R= e1A1 Radiant Heat Transfer Between Gray Surfaces r2 A1F12 R= r2 e2 A2 Figure 23. that is. 3. 4. R1 = r1/A1e1 Eb1 = sT 14 R3 = r2/A2e2 R2 = 1/A1F12 J1 Eb2 = sT 24 J2 Utilizing Ohm’s law.20 Equivalent network for gray-body relations between two surfaces. EXAMPLE 4 Two parallel gray surfaces maintained at temperatures T1 and T2 view each other. Generate an expression for the net heat transfer between these surfaces. a ¼ e. we obtain the expression q12 . The circuit and important quantities are shown here. Each surface is sufﬁciently large that they may be considered inﬁnite. illustrate features of gray-body problem solutions.23. Each surface must be isothermal. The assumptions required to use the electrical analog approach to solve radiation problems are the following: 1.

Determine the net heat transfer from the high temperature surface under the following conditions: (a) the plates are black and the surroundings are at 0 K and totally absorbing. . s T14 T24 Eb1 Eb2 ¼ ¼ r1 1 r åR þ þ 2 A1 e1 A1 F12 A2 e2 Now. we obtain the result AsðT14 T24 Þ 1 e1 1 e2 þ1 þ e1 e2 AsðT14 T24 Þ ¼ 1 1 þ 1 e1 e2 q12 ¼ EXAMPLE 5 Two parallel planes measuring 2 m by 2 m are situated 2 m apart. noting that for inﬁnite parallel planes A1 ¼ A2 ¼ A and F12 ¼ F21 ¼ 1 and writing r1 ¼ 1 e1 and r2 ¼ 1 e2 . (b) the plates are black and the walls connecting the plates are reradiating. Plate 1 is maintained at a temperature of 1100 K and plate 2 is maintained at 550 K.

8. F1R . and F 12 . Analog electrical circuits for parts (a).4 and 0. with black surroundings at 0 K.14 F 12 ¼ 0:54 from Figure 23.21.80 EbR = 0 R= 1 A1F1R R= ER 1 A2 F2R R= 1 = 1 A1F1R AR FR1 Eb1 Eb2 1 1 = R= A1F12 A2F21 R= 1 = 1 ARFR2 A2F2R Eb1 Eb2 1 = 1 R= A1F12 A2F21 (a) (b) R= 1 A1F1R R= J1 J1 1 A2F2R Eb1 Eb2 R= r1 A1e1 R= R= r2 A2e2 1 = 1 A1F12 A2F21 (c) Figure 23. (b).384 Chapter 23 Radiation Heat Transfer (c) the plates are gray with emissivities of 0. 300 W þ 266. and (c) are shown in Figure 23.21 Equivalent circuits for example 3. F1R = 1 – F12 = 0.14 and F1R ¼ 1 F12 ¼ 0:80 Part (a). 000 W ¼ 328:3 kW . respectively. Heat ﬂux evaluations will require evaluating the quantities F12 . The net rate of heat leaving plate 1 is q1 net ¼ q12 þ q1R ¼ A1 F12 (Eb1 Eb2 ) þ A1 F1R Eb1 ¼ (4 m2 )(0:2)(5:676 108 W/m2 : K4 )(11004 5504 )K4 þ (4 m2 )(0:8)(5:676 108 W=m2 : K4 )(1100 K)4 ¼ 62. The appropriate values are F12 ¼ 0:20 from Figure 23.

When reradiating walls are present the heat ﬂux becomes 2 3 6 q12 ¼ (Eb1 Eb2 )6 4A1 F12 þ 1 1 1 þ A1 F1R A2 F2R 7 7 5 and. from Figure 23. ﬁnally. For a representative surface having area. Gi. since A1 ¼ A2 and F1R = F2R F1R q12 ¼ (Eb1 Eb2 )A1 F12 þ 2 Since F12 þ F1R ¼ 1. Such an assumption. equations (20–28) and (20–30) can be written as qi ¼ Ebi Ji ¼ Ai (Ji Gi ) ri /Ai ei (23-32) where qi is the net rate of heat transfer leaving surface i.10 Radiant Heat Transfer Between Gray Surfaces 385 Part (b). radiosity. The formalism to be developed in this section is directly applicable for solution by numerical methods.23.14 allow for the temperatures along the reradiating walls to vary from T1 to T2 . in this example. leads to an error of approximately 11%. The use of the analog circuit considers the radiating surface to be a constant temperature. in an enclosure bounded by n surfaces. Ai . using reciprocity. can be expressed as n Ai Gi ¼ å Jj Aj Fji (23-33) j¼1 or. and irradiation are particularly useful in generalizing the analysis of radiant heat exchange in an enclosure containing any number of surfaces.14.out ¼ 131:3 kW. An evaluation of the circuit shown in Figure 23. using the value F 12 ¼ 0:54. the heat ﬂux is q12 ¼ (4 m4 )(5:678 104 W/m2 : K4 )(11004 5504 )K4 (0:6) ¼ 187 kW We should note that an equivalent expression for the heat ﬂux is q12 ¼ A1 F 12 (Eb1 Eb2 ) and. the bracketed term is evaluated as F12 þ F1R 0:8 ¼ 0:2 þ ¼ 0:6 2 2 and. the result would be q12 ¼ 168:3 kW This alternate result is the more accurate in that the values of F 12 plotted in Figure 23. The concepts related to the quantities. The irradiation. Part (c).2(C) yields q1. as n Ai Gi ¼ Ai å Jj Fij j¼1 (23-34) .

Ji. have been separated out of the summation.. Fii . In most cases Fii will be 0. This set of equations can be represented in matrix form as [A][J] ¼ [B] (23-39) where [A] is the coefficient matrix. itself. ﬁnally n ei ei 1 Fii þ Ji å Fij Jj ¼ r Ebi j¼1 ri i (23-38) j 6¼ i Equations (23-37) and (23-38) comprise the algorithm for evaluating quantities of interest in a many-surface enclosure. .386 Chapter 23 Radiation Heat Transfer Combining equations (23-32) and (23-34) we obtain n qi ¼ Ai Ji å Fij Jj (23-35) j¼1 Ai e i Ai e i ¼ r Ebi r Ji i i (23-36) We can now write the two basic expressions for a general surface in an enclosure. i. the latter is written when the surface temperature is specified. [B] is a column matrix involving the right-hand sides of equations (23-37) and (23-38). equation (23-35) can be expressed in the form n Ji å Fij Jj ¼ j¼1 qi Ai or n Ji (1 Fii ) å Fij Jj ¼ j¼1 j 6¼ i qi Ai (23-37) and. equations (23-35) and (23-36) yield n Ai e i å Fij Jj ri (Ebi Ji ) ¼ Ai Ji j¼1 2 3 n ¼ Ai 4Ji (1 Fii ) å Fij Jj 5 j¼1 j 6¼ i and. This quantity. The solution for the Ji then proceeds according to [J] ¼ [C][B] (23-40) [C] ¼ [A]1 (23-41) where is the inverse of the coefficient matrix. if the temperature at surface i is known. and [J] is a column matrix of the unknowns. Fii will have a nonzero value in those cases when surface i ‘‘sees’’. involving the unknowns Ji . In these two equations the terms involving the view factor. Example 6 illustrates the application of this approach. it is concave. The former applies to a surface of known heat flux.e. If the surface heat ﬂux is known. When writing equation (23-37) or (23-38) for each surface in an enclosure a series of n simultaneous equations is generated.

The following conditions are known: T1 ¼ 1100 K F11 ¼ 0 T2 ¼ 550 K F21 ¼ 0:2 T3 ¼ 0 K F31 ¼ 0:2 F12 ¼ 0:2 F13 ¼ 0:8 e1 ¼ 1 F23 ¼ 0 F23 ¼ 0:8 e2 ¼ 1 F32 ¼ 0:2 F33 ¼ 0:6 e3 ¼ 1 We can write the following: e1 e1 J1 ½F12 J2 þ F13 J3 ¼ r Eb1 r1 1 e2 e2 J2 ½F21 J1 þ F23 J3 ¼ r Eb2 1þ r2 2 e3 e3 J3 ½F31 J1 þ F32 J2 ¼ r Eb3 1 F33 þ r3 3 1þ which. The only change from part (a) is that e3 ¼ 0. The set of equations. Each of the surfaces is at a known temperature in this case.10 EXAMPLE 6 Radiant Heat Transfer Between Gray Surfaces 387 Solve the problem posed in example 5 using the methods developed in this section. For this case n ¼ 3 and the problem formulation will involve 3 equations—one for each surface. reduce to J1 ¼ Eb1 ¼ sT14 J2 ¼ Eb2 ¼ sT24 J3 ¼ 0 The net heat leaving plate 1 is thus.23. applying to the three surfaces are again e1 e1 J1 ½F12 J2 þ F13 J3 ¼ r Eb1 r1 1 e2 e2 J2 ½F21 J1 þ F23 J3 ¼ r Eb2 1þ r2 2 e3 e3 J3 ½F31 J1 þ F32 J2 ¼ r Eb3 1 F33 þ r3 3 1þ as before. and ei . equal to q1 ¼ A1 [J1 F12 J2 ] ¼ A1 [sT14 0:2sT24 ] ¼ 4 m2 (5:676 108 W/m K4 )[11004 0:2(550)4 ]K4 ¼ 328:3 kW Part (b). Fij . Part (a). Values of Ti and Fij remain the same. Substituting values for Ti. for the given conditions. according to equation (23-37). thus equation (23-38) applies. we have J1 ¼ Eb1 ¼ sT14 J2 ¼ Eb2 ¼ sT24 (1 F33 )J3 F31 J1 F32 J2 ¼ 0 .

Gases emit and absorb radiation in discrete energy bands dictated by the allowed energy states within the molecule. Emissivities are e1 ¼ 0:4 e2 ¼ 0:8 e3 ¼ 1 Equations for the three surfaces are. with numerical values inserted.388 Chapter 23 Radiation Heat Transfer The expression for qi is F13 (F31 J1 þ F32 J2 ) q1 ¼ A1 [J1 F12 J2 F13 J3 ] ¼ A1 J1 F12 J2 1 F33 F13 F31 F13 F32 J2 F12 þ ¼ A1 J1 1 1 F33 1 F33 and. it follows that the amount of energy emitted or absorbed by a molecule will have a .11 RADIATION FROM GASES So far. the vibrational or rotational motion of a molecule may have only certain values. again e1 e1 1þ J1 [F12 J2 þ F13 J3 ] ¼ Eb1 r1 r1 e2 e2 J2 [F21 J1 þ F23 J3 ] ¼ Eb2 1þ r2 r2 e3 e3 J3 [F31 J1 þ F32 J2 ] ¼ Eb3 1 F33 þ r3 r3 which become 0:4 0:4 J1 (F12 J2 þ F13 J3 ) ¼ Eb1 1þ 0:6 0:6 0:8 0:8 J2 (F21 J1 þ F23 J3 ) ¼ Eb2 1þ 0:2 0:2 J3 ¼ 0 We now have 1:67J1 0:2J2 ¼ 0:67Eb1 5J2 0:2J1 ¼ 4Eb2 Solving these two equations simultaneously for J1 and J2 we get J1 ¼ 33 900 W/m2 J2 ¼ 5510 W/m2 and the value for qi is evaluated as 5510 4 q1 ¼ 33 900 5 ¼ 131:2 kW 23. say. we obtain (0:8)(0:2) (0:8)(0:2) q1 ¼ 4(5:676 108 ) (1100)4 1 (550)4 0:2 þ ¼ 187:0 kW 1 0:6 1 0:6 Part (c). As the energy associated with. the interaction of radiation with gases has been neglected. Values of Ti and Fij remain the same.

in microns 14 16 18 Figure 23. frequency. while the energy emitted by a solid will comprise a continuous spectrum. The gas is in thermodynamic equilibrium. These gases are also associated with the products of combustion of hydrocarbons. composition.26 give the corresponding data for CO2. For nonluminous gases. the radiation emitted and absorbed by a gas will be restricted to bands. as it involves the temperature. and geometry of the gas. the gray gas emissivities of H2O and CO2 may be obtained from the results of Hottel. Figures 23. The state of the gas may therefore be characterized locally by a single temperature. density. Figure 23.25 and 23. The emission of radiation for these gases is seen to occur in the infrared region of the spectrum. the inert gases and diatomic gases of symmetrical composition such as O2. . and H2 may be considered transparent to thermal radiation. For geometries not covered in the table. the total emissivity may be determined from the relation etotal ¼ eH2 O þ eCO2 De where De is given in Figure 23. While the graphs apply strictly only to a hemispherical gas mass of radius L. Thus. The gas may be considered gray. Figure 23. n ¼ DE/h.24 gives the correction factor.22. other shapes can be treated by consideration of a mean beam length L as given in Table 23.27. The determination of the absorption and emission of radiation is very difﬁcult. A hemispherical mass of gas at 1 atm pressure was used by Hottel to evaluate the emissivity. In the range of temperatures associated with the products of hydrocarbon combustion. which is the ratio of the emissivity at total pressure P to the emissivity at a total pressure of 1 atm. For pressures other than atmospheric. the mean beam length may be approximated by the relation L ¼ 3:4 (volume)/(surface area). N2. This simpliﬁcation allows the absorption and emission of radiation to be characterized by one parameter as a ¼ e for a gray body.23 gives the emissivity of a hemispherical mass of water vapor at 1 atm total pressure and near-zero partial pressure as a function of temperature and the product pw L. There are several simpliﬁcations that allow estimation of radiation in gases to be made in a straightforward manner. Figure 23. where pw is the partial pressure of the water vapor.11 Water vapor Intensity 2 389 Carbon dioxide 2000 °R 0 Radiation From Gases 4 6 12 8 10 Wavelength. Important types of media that absorb and emit radiations are polyatomic gases such as CO2 and H2O and unsymmetrical molecules such as CO. it may be seen that the emission bands of CO2 and H2O overlap.22 illustrates the emission bands of carbon dioxide and water vapor relative to black-body radiation at 15008F. corresponding to the difference in energy DE between allowed states. From Figure 23. Cw . 2. These idealizations are as follows: 1.23. When both carbon dioxide and water vapor are present.22 Emission bands of CO2 and H2O.5.

02 0.2 0.03 0. New York.2 0 0.3 0. 5 1500 0. 01 06 0 2000 2500 3000 3500 Absolute temperature. By permission of the publishers.7 0.009 0.5 0.1 0.015 0.02 03 0.1 0. McAdams (ed. L.09 0.05 1. 08 0. 0.6 0. 0. 1964. H. in °R 4000 4500 5000 Figure 23.04 5 0.1 0.3 10 0. 2 0. 5 2 07 03 02 01 1000 04 5 0.08 0.12 0. tube diameter equals clearance Same as preceding except tube diameter equals one-half clearance 2:8 clearance 3:8 clearance yFrom H. IV in W.07 0.5 0. 01 00 7 0.06 0. 5 0. Heat Transmission.’’ Chap. ‘‘Radiation. . Third Edition.007 500 3 05 0.4 Pw L = 20 atm ft 0.15 Cas emissivity. McGraw-Hill Book Company.5 2 1.8 5 0. C. 0.2 0. 0.012 0. for various geometriesy Shape L Sphere 2 diameter 3 1 diameter 1:8 distance between planes 2 side 3 Inﬁnite cylinder Space between inﬁnite parallel planes Cube Space outside inﬁnite bank of tubes with centers on equilateral triangles. 390 5 1.5 Mean beam length.10 0.6 0.4 0.). 0.2 0.23 Emissivity of water vapor at one atmosphere total pressure and near-zero partial pressure. 00 0.010 0.008 0. 0. Hottel.0 0.Chapter 23 Radiation Heat Transfer Table 23.

0 15 0.08 0.0 03 0.2 2.0 1.09 0.009 0.11 Radiation From Gases 391 1.0 Cw 0.0 6 0.8 1.24 Correction factor for converting emissivity of H2O at one atmosphere total pressure to emissivity at P atmospheres total pressure.005 0.5 0.5 5. .04 0.10 0.6 0.2 (P + pw).0 0.02 0.2 0 0 0.0 1.07 0.50 2. in °R Figure 23. 0.008 0.15 0.6 0.2 pw L = 10.007 0.0 0.0 1.8 0.25 0.03 0.0 0.0 3 06 0.001 0.25 Emissivity of CO2 at one atmosphere total pressure and near-zero partial pressure.2 0.1 0 0.4 0.0 2 0.4 0.0 atm ft 1. in atm Figure 23.0 4 0.06 0.4 0.6 1.0 10 05 0.004 0.3 0.8 0.0 1.8 pw L = 10.4 0.0 atm ft 1.015 0.0 atm ft 3.6 1 2 0.002 1000 1500 2000 2500 3000 3500 4000 4500 5000 Absolute temperature.3 pC L = 5.2 8 0.010 0.05 0.003 500 0.23.0 1.0 0.0 08 0.006 0.0 04 0.1 5 0.

05 0-0.01 0 0 0.0 2. Wiley and Sons.3 0. 23.0 0. J. 4 M.4 0.2 0.8 1.5 0.4 0.75 0. R. Thermal Radiation Heat Transfer. Howell.5 0. For a more complete treatment.03 2.5 1.8 1.20 0. Thermal Radiative Transfer and Properties. McGraw-Hill.75 0.20 0.1 0.2 0.0 0.08 0.0 1.30 0.5 m ft pc L = 2. F.05 0.2 0. Siegel and J.3 0.04 pcL + pwL = 5 atm ft 3. in atm 2.0 3.5 0.0 Total pressure P. The results presented here for the gray gas are gross simpliﬁcations. Modest.27 Correction to gas emissivity due to spectral overlap of H2O and CO2.0 5.5 at Cc 0.0 0.4 0.5 0. textbooks by Siegel and Howell..25 0.05 0. 1992. 1992.0 pw pc + pw Figure 23.0 0. 3rd Edition.2 Modest.05 1. 3 M. Hemisphere Publishing Corp.2 0.5 1. New York.12 THE RADIATION HEAT-TRANSFER COEFFICIENT Frequently in engineering analysis.6 0.3 0.8 0.75 0.8 1.3 and Brewster4 present the fundamentals of nongray-gas radiation.0 Figure 23.25 1. Q.5 pcL 2 -0.06 pcL + pwL = 5 atm ft 0.07 1700°F and above pcL + pwL = 5 atm ft 3 2 1.5 0. . New York.0 1000°F 260°F 0.0 0.0 pw pc + pw 0.8 1.0 pw pc + pw 0 3.2 0. 1993.0 1.392 Chapter 23 Radiation Heat Transfer 2.0 1. along with extensive bibliographies.0 =0 1.0 0. Radiative Heat Transfer.30 0.6 0. Brewster.26 Correction factor for converting emissivity of CO2 at one atmosphere total pressure to emissivity at P atmospheres total pressure.4 0.6 0.5 2.12 0. Washington. convection and radiation occur simultaneously rather than as isolated phenomena.50 0.50 0 0. 0. An important approximation in such cases is the linearization of the radiation contribution so that htotal ¼ hconvection þ hradiation 2 (23-42) R.02 1.6 0.02 0.12 0.

determine the percent of solar radiation that will pass through the glass.35 and 2. PROBLEMS 23.28 Tangent approximation for hr .Problems qr A1 where qr /A1 (T TR ) s(T 4 T24 ) ¼ F12 T TR 393 hr (23-43) Here TR is a reference temperature. the following relations are obtained for hr and TR : Ᏺ12s (T 4 – T 42) hr (T – TR) T – T2 T1 – T2 Figure 23. Considering the sun to emit as a black body at 5800 K. In effect. The presence of absorbing and emitting gases between surfaces was also examined. .1 and 100 mm.13 CLOSURE Radiation heat transfer has been considered in this chapter. introduced in Chapter 15. The gases of principle interest in this regard are water vapor and carbon dioxide. Modiﬁcations to this relationship were made for nonblack surfaces and for geometric relationships between multiple surfaces in view of each other. and s is the Stefan–Boltzmann constant. which is generally referred to as the thermal band. estimate the sun’s effective surface temperature.7 mm. 23. accounts for geometry and surface condition of the radiating and absorbing surface. and its diameter is 860.1 The sun is approximately 93 million miles distant from Earth. F. having units of W/m2 K4 in the SI system. On a clear day solar irradiation at Earth’s surface has been measured at 360 Btu/h ft2 and an additional 90 Btu/h ft2 are absorbed by Earth’s atmosphere. is designated the Stefan–Boltzmann equation.28. The factor. and T1 and T2 are the respective surface temperatures. By constructing a tangent to the relation curve at T ¼ T1 .2 A greenhouse is constructed of silica glass that is known to transmit 92% of incident radiant energy between wavelengths of 0. The glass may be considered opaque for wavelengths above and below these limits. T is the absolute temperature. hr ¼ 4sT 31 F12 (23-44) and TR ¼ T1 T 41 T 42 4T 31 (23-45) 23. Radiant energy transfer is associated with the portion of the electromagnetic spectrum between 0. The fundamental rate equation for thermal radiation. it is expressed as Eb ¼ sT 4 (23-12) where Eb is the black body emissive power. With this information.000 miles. equation (23-43) represents a straight-line approximation to the radiant heat transfer as illustrated in Figure 23.

10 cm2 and is located 1 m from the furnace opening.4 A radiation detector. Determine the amount of radiant energy reaching the detector under two conditions: a.3 to 0. is used to estimate heat loss through an opening in a furnace wall.3 A tungsten ﬁlament. the opening is covered by a semitransparent material with spectral transmissivity given by t l ¼ 0:8 tl ¼ 0 for 0 .394 Chapter 23 Radiation Heat Transfer If the plants on the inside of the greenhouse have an average temperature of 300 K. The opening in this case is circular with a diameter of 2. At what wavelength is the emissive power maximum? What portion of the total emission lies within the visible-light range. oriented as shown in the sketch. what fraction of their emitted energy will be transmitted through the glass? 23. The detector has a surface area of 0. is heated to a temperature of 40008R.75 mm? 23. and emit as a black body.5 cm. radiating as a gray body. the detector has a clear view of the opening. 0. b.

l .

. are being considered for use in windows. If the plate temperature is 700 K and the surroundings are at 310 K. Solar irradiation may be taken as 450 Btu/h ft2 of satellite disc area.13 A sheet-metal box in the shape of a 0.7 The sun’s temperature is approximately 5800 K and the visible light range is taken to be between 0. what will be the net heat ﬂux to or from a 308C surface? 23. determine the energy loss through the hole. plain and tinted. Taking the satellite diameter as 50 in. Incident solar energy reaches the collector with a ﬂux of 800 W/m2 .7 to 100 mm.5 The distribution of solar energy. estimate the temperature of the satellite skin. Compare the fraction of visible radiant energy transmitted through each. incident on Earth. with a surface area of 60 m2 . The surroundings are considered black with an effective temperature of 308C. Neglecting any conductive loss from the collector. hollow.5-cm-diameter hole in Problem 23. is placed between these two. determine: a. is placed on the roof of a house. For an effective convective heat-transfer coefﬁcient of 12 W/m2 K applying. and 4500 K. what heat loss would result? 23. The spectral transmissivity for these two glasses is approximated as plain glass : tinted glass : tl ¼ 0 0:9 0 for for for 0 < l < 0:3 mm 0:3 < l < 2:5 m 2:5 mm < l tl ¼ 0 0:9 0 for for for 0 < l < 0:5 mm 0:5 < l < 1:5 m 1:5 mm < l Compare the fraction of incident solar energy transmitted through each material. 23. Determine (a) the wavelength of maximum emission and (b) the fraction of emission in the visible region of the spectrum. The box encloses electronic equipment that dissipates 1200 W of energy.12 If the 7. Its orbit may be considered circular at a height of 500 miles above Earth. (b) the number of watts emitted in the visible range from 0.5-cm-diameter hole is drilled in a 10-cm-thick iron plate.9 An opaque gray surface with e ¼ 0:3 is irradiated with 1000 W/cm.16 A small circular hole is to be drilled in the surface of a large. is 35 W/m2 K. and air at 208C adjacent to the plate. 2000. spherical enclosure maintained at 2000 K. 2 mm for 2 mm < l < 1 Opening diameter = 25 cm Furnace 30° T = 1500 K Detector 23. the equilibrium temperature of the collector. Determine the fractional change in radiant exchange between the two plane surfaces due to the intervening plane and evaluate the temperature of this intervening plane.70-m cube has a surface emissivity of 0. What fraction of solar emission is visible? What fraction of solar emission lies in the ultraviolet range? The infrared range? At what wavelength is solar emissive power a maximum? 23.4 mm. determine (a) the hole diameter. 23. what will be the temperature of the box surface? 23. respectively.15 The ﬁlament of an ordinary 100 W light bulb is at 2910 K and it is presumed to be a black body. . and (d) the infrared range from 0.14 Two very large black plane surfaces are maintained at 900 and 580 K. and the emissivity of Earth may be taken as 0.95. The convective heat-transfer coefﬁcient between the collector and the surrounding air. 23.4 and 0. having e ¼ 0:8.7 mm. which lies in the wavelength band between 0.11 A 7.6 Determine the fraction of total energy emitted by a black body. b. (c) the ultraviolet range between 0 and 0.10 A black solar collector. 23. 23. If 100 W of radiant energy exits through the hole. the net radiant exchange between the collector and its surroundings. 0. A third large plane surface. 3000. emits 8 W.7.8 and 5. 23.4 and 0.0025 m2 in area. Earth may be considered to be at a uniform temperature of 508F. Two kinds of glass.0 mm for surface temperatures of 500.17 A large cavity with a small opening. 23. Determine the wall temperature of the cavity. at 308C. If the surroundings are taken to be black at 280 K.8 A satellite may be considered spherical with its surface properties roughly those of aluminum. 23.11 were drilled to a depth of 5 cm. The hole sides may be considered to be black. and the top and sides of the box are considered to radiate uniformly. can be approximated as being from a black body at 5800 K.7 mm.

The rod diameter is 2 in. and that of the reﬂector is 18 in. The intervening space between the two tubes is evacuated. What is the radiant energy loss from the heater per foot of length? How does this compare to the loss from the heater without the reﬂector present? 23.2 mm and 0. with e ¼ 0:05 and its surface temperature is 300 K. What is the net rate of radiant transfer to the surroundings? 6 ft 12 ft 23.8. is concentric with the smaller one. 23.2 mm and 1. The glass in the peephole has a transmissivity of 0. The inner sphere has an outside diameter of 1 m and the outer sphere has an inside diameter of 1. passes horizontally through a 12 14 9 ft room whose walls are maintained at 708F and have an emissivity of 0. is made of two concentric spheres separated by an evacuated space. Nitrogen. Determine the heat lost through the peephole.20 A cryogenic ﬂuid ﬂows in a 20-mm-diameter tube with an outer surface temperature of 75 K and an emissivity of 0.08 between 3.19 A furnace that has black interior walls maintained at 1500 K contains a peephole with a diameter of 10 cm. at 1 atmosphere.24 The circular base of the cylindrical enclosure shown may be considered a reradiating surface.3 m. black? b. This larger tube is gray. 23. 23.3 in. in watts per meter of tube length.18 Determine the wavelength of maximum emission for (a) the sun with an assumed temperature of 5790 K.27 A dewar ﬂask. The pipe surface is at a temperature of 2058F.2.-OD iron pipe e ¼ 0:7.80 and are maintained at 5408F. and the thermocouple indicates 3108F. Determine the heat gain by the cryogenic ﬂuid. estimate the actual temperature of the gas. 23. The emissivity of the heater surface is 0.28 A cylindrical cavity is closed at the bottom and has an opening centered in the top surface. used to contain liquid nitrogen. having a diameter of 50 mm.8. determine the net energy exchange between the ﬂoor and ceiling. A cross section of this conﬁguration is shown in the sketch.21 A circular duct 2 ft long with a diameter of 3 in.78 between 0 and 3. For the conditions stated below. gray with an emissivity of 0. Assuming the convective heat-transfer coefﬁcient between the thermocouple and gas in the duct to be 30 Btu/h ft2 8F. The cylindrical walls have an effective emissivity of 0. The shield surfaces may be considered gray and diffuse with an emissivity of 0. determine the rate of radiant energy passing through the . 23. estimate the boil-off rate of nitrogen. These surfaces are both graydiffuse with e ¼ 0:2.22 A heating element in the shape of a cylinder is maintained at 20008F and placed at the center of a half-cylindrical reﬂector as shown. 23.23 A 12-ft-long. A larger tube. (b) a light bulk ﬁlament at 2910 K. How much energy will be lost through the hole if the cavity is a. and (d) human skin at 308 K. and the entire assembly is placed in a room maintained at 708F.2.7? What will be the temperature of the refractory under each condition? 15-cm radius 23. Under conditions when the inner sphere is full of liquid nitrogen and the outer sphere is at a temperature of 300 K. which are maintained at 408F. 23.04 on both sides. Assuming the walls to be reradiating and all surfaces to have an emissivity of 0. has a thermocouple in its center with a surface area of 0.6. The top of the enclosure is open to the surroundings. The emissivity of the duct walls may be taken as 0.23. Evaluate the heat gain per meter of length if there is a thin walled radiation shield placed midway between the two tubes. respectively.25 The hemispherical cavity shown in the ﬁgure has an inside surface temperature of 700 K.8 and that of the thermocouple as 0. 3-in. A plate of refractory material is placed over the cavity with a circular hole of 5 cm diameter in the center. Compare the radiant energy loss from the pipe with that due to convection to the surrounding air at 708F. The duct walls are at 2008F. (c) a surface at 1550 K. has a saturation temperature of 78 K and a latent heat of vaporization of 200 kJ/kg.8.26 A room measuring 12 ft by 20 ft by 8 ft high has its ﬂoor and ceiling temperatures maintained at 85 and 658F.

The plate having e ¼ 0:6 is maintained at 1000 K and the other is at 420 K.8 is placed between the two plates described in Problem 23. How much energy will be lost through the hole? 23. the surroundings are black at 278C. The two surfaces are maintained at 200 and 100 K. determine a. they are gray e1 ¼ 0:6. b. The side wall is made of polished stainless steel. The surroundings (the room) can be considered black and the heater surface emissivity is 0.30 Two parallel black rectangular surfaces. the net radiant heat transfer between the two surfaces.5 in 4 in 1 40 mm 3 10 in 2 30 mm 23.. having the same diameter. . reradiating surface. b.31. the radiant energy interchange between the two plates.5 and 4 in. This heater–receiver assembly is placed in a large room at a temperature of 275 K.8. What will be effective emissivity of the opening? respectively. how will the answer to part (a) of Problem 23. They measure 5 m by 10 m. to 395 K. Determine the following: a. The bottom of a tank.36 A heavily oxidized aluminum surface at 755 K is the source of energy in an enclosure. The surroundings may be considered to absorb all energy that escapes the two-plate system. the space between the two cylindrical surfaces is enclosed by an adiabatic surface. the heater surface temperature. 2 1 2. 23. b. whose back sides are insulated. The bottom surface is diffuse-gray with e ¼ 0:6. Determine the radiant energy reaching the bottom of the tank if a. respectively. In this case the two disks comprise the bases of a cylinder with side wall at constant temperature of 3508F. b.6 and 0. assume that all three surfaces have uniform temperatures and that they are diffuse and gray.33 if they are bases of a cylinder with the side wall considered a nonconducting. and has a temperature of 600 K. diffuse receiver with a spacing of 7.33. with both surfaces having an emissivity of 0.5 between them. which radiantly heats the side walls of a circular cylinder surface as shown. Find the heat exchange between these disks if a. b. For purposes of calculation. 23. The top of the enclosure is made of ﬁre clay brick and is adiabatic. b.396 Chapter 23 Radiation Heat Transfer 5-mm-diameter cavity opening. are oriented parallel to each other with a spacing of 5 m. diffuse circular heater with a diameter of 15 cm is placed parallel to a second gray.35 Evaluate the heat transfer leaving disk 1 for the geometry shown in Problem 23. respectively.37 A gray. is oriented parallel to the heater with a separation distance of 10 cm. Evaluate for the case where: a.31 be affected? Draw the thermal circuit for this case. they are black. the side wall is black.2 m wide and 2. All interior surfaces are black at 600 K. the total energy lost from the hot plate. e2 ¼ 0:3. All interior surfaces are diffuse-gray with e ¼ 0:6 and are at a uniform temperature of 600 K. the net heat transfer between each surface and the surroundings. When the power input to the heater is 300 W. b. c. the receiver surface temperature. The disk surface temperature is 2108F. 23.29 A circular heater. c. The backs of both surfaces are insulated and convective effects are to be neglected. Determine a. These rectangles are 1. the net heat supplied to each surface. Fire clay brick 2m Stainless steel Aluminum 2m 23.32 If a third rectangular plate. The heater surface is gray (e ¼ 0:6) and the tank surface is also gray (e ¼ 0:7). The disk to the left has an inner ring cut out such that it is annular in shape with inner and outer diameters of 2.4 m high and are 0.6 m apart. a. Evaluate the heat transfer to the stainless steel surface. in diameter and is at a temperature of 5008F.34 Evaluate the net heat transfer between the disks described in Problem 23..9. as shown in the accompanying ﬁgure.33 Two disks are oriented on parallel planes separated by a distance of 10 in. 23. The surroundings are black at 0 K. 23.31 Two parallel rectangles have emissivities of 0. measuring 20 cm in diameter. All other surfaces are reradiating. has its surface temperature maintained at 10008C. 23. The disk to the right is 4 in. the side wall is gray with e ¼ 0:2: Determine the rate of heat loss through the hole in each case.

41 A gas consisting of 20% CO2 and 80% oxygen and nitrogen leaves a lime kiln at 20008F and enters a square duct measuring 6 in. by 6 in. The inside surfaces of the steel tube are newly galvanized. diameter 1 in. 23.40 A gas of mixture at 1000 K and a pressure of 5 atm is introduced into an evacuated spherical cavity with a diameter of 3 m. The water-cooled walls and ends of the tube are maintained at 508F. an approximation for the radiant energy exchange between the enclosure and gas contained within an arbitrary control volume is given by Aw Fwg sew (eg Tg4 ag Tw4 ): . d.8. b. The other three walls may be considered refractory surfaces. the net radiant exchange to the surroundings.-diameter) silica glass viewing port. and whose walls have an emissivity of 0. The cavity walls are black and initially at a temperature of 600 K. in cross section. 397 23. whose inside surface is maintained at 8008F. c. the energy radiated through the viewing port. The metal is maintained at a temperature of 25008F. At what temperature would the gas leave the duct if the length of the duct were twice the value determined in part (a)? 12 in 4 in (Courtesy of the American Institute of Chemical Engineers.9. the net radiant exchange between the heater and receiver.Problems c. In the upper end is a small (1/4-in. long) metal test specimen is suspended by very ﬁne wires in a large evacuated tube. The speciﬁc heat of the gas is 0:28 Btu/lbm F. One wall of the duct is held at 420 K and has an emissivity of 0. at which temperature it has an emissivity of approximately 0. As the response of the gas to emission and absorption of radiant energy differs. the total net heat-transfer rate by radiation from the test specimen. Determine the required length of duct to cool the gas to 10008F. 23.) Hint. Determine the rate of radiant energy transfer to the cold wall from the water vapor. Test specimen a. the view factor from the specimen to the window. and it is to be cooled to 10008F in the duct.38 A small (1/4 in.39 A duct with square cross section measuring 20 cm by 20 cm has water vapor at 1 atmosphere and 600 K ﬂowing through it.2. b. c. Determine the ratio of radiant energy transfer to that by convection. What initial rate of heat transfer will occur between the gas and spherical walls if the gas contains 15% CO2 with the remainder of the gas being nonradiating? Viewing port 23. Room temperature is 708F. Estimate a. The mass velocity of the kiln gas is 0:4 lbm /ft2 : s and the convective heat-transfer coefﬁcient between the gas and duct walls is 1:5 Btu/h ft2 F.

minimizing the concentration differences within the system. we should immediately realize that the two mechanisms often act simultaneously. However. whereas in gas absorption and liquid–liquid extraction processes. the doping of a silicon wafer to form a semiconducting thin ﬁlm. When a system contains two or more components whose concentrations vary from point to point. As in the case of heat transfer. If we consider the lump of sugar added to the cup of black coffee. Each of these modes of mass transfer will be described and analyzed. in adsorption or crystallization processes. The mechanism of mass transfer. there is a natural tendency for mass to be transferred. Perfume presents a pleasant fragrance that is imparted throughout the surrounding atmosphere. and the puriﬁcation of ores and isotopes. These two distinct modes of transport. the aeration of wastewater. Mass can be transferred by random molecular motion in quiescent ﬂuids. in the conﬂuence of the two modes of mass transfer. aided by the dynamic characteristics of the ﬂow. one mechanism can dominate quantitatively so that approximate solutions involving only the dominant mode need be used. experience teaches us that the length of time required to distribute the sugar will depend upon whether the liquid is quiescent or whether it is mechanically agitated by a spoon. Biological processes include the oxygenation of blood and the transport of ions across membranes within the kidney. the components remain at the interface. The transport of one constituent from a region of higher concentration to that of a lower concentration is called mass transfer. Chemical processes include the chemical vapor deposition (CVD) of silane (SiH4) onto a silicon wafer. molecular mass transfer and convective mass transfer. Mass transfer underlies the various chemical separation processes where one or more components migrate from one phase to the interface between the two phases in contact. or it can be transferred from a surface into a moving ﬂuid. as we have also observed in heat transfer. depends upon the dynamics of the system in which it occurs. Mass transfer is the basis for many biological and chemical processes. For example. Many of our day-to-day experiences involve mass transfer. A lump of sugar added to a cup of black coffee eventually dissolves and then diffuses uniformly throughout the coffee. 398 . Water evaporates from ponds to increase the humidity of the passing air stream. the components penetrate the interface and then transfer into the bulk of the second phase. are analogous to conduction heat transfer and convective heat transfer.Chapter 24 Fundamentals of Mass Transfer The previous chapters dealing with the transport phenomena of momentum and heat transfer have dealt with one-component phases that possessed a natural tendency to reach equilibrium conditions.

a deﬁnite number of the molecules in the lower element of the volume will cross the hypothetical section from below. For example. is deﬁned as molecular diffusion. sometimes toward a region of higher concentration. there are more solute molecules in one of the elements of volume than in the other. As each component may possess a different mobility. As pointed out in Chapters 7 and 15. or diffusion.1 Molecular Mass Transfer 399 24. The Fick Rate Equation The laws of mass transfer show the relation between the ﬂux of the diffusing substance and the concentration gradient responsible for this mass transfer. In the speciﬁc case of gaseous mixtures. In a multicomponent mixture. This macroscopic transport of mass. each solute molecule behaves independently of the other solute molecules. At temperatures above absolute zero. and that the theoretical analyses of all three phenomena will have much in common. independent of any convection within the system.1 shows an elemental volume dV that contains a . an overall net transfer from a region of higher concentration to one of lower concentration will result. Although it is not possible to state which way any particular molecule will travel in a given interval of time. The two thin. one should expect that the three transport phenomena will depend upon many of the same characteristic properties. and the same number of molecules will leave the upper element and cross the section from above. sometimes toward a lower concentration. With the existence of a concentration gradient.1 MOLECULAR MASS TRANSFER As early as 1815. its evaluation must involve an examination of the effect of each component. as it is also called. Within dilute gas mixtures. Figure 24. as stipulated by Avogadro’s law. Let us consider a hypothetical section passing normal to the concentration gradient within an isothermal. As mass transfer. the species A mixture velocity must be evaluated by averaging the velocities of all of the components present.1 Elemental volume containing a multicomponent mixture. we will often desire to know the diffusion rate of a speciﬁc component relative to the velocity of Molecule of the mixture in which it is moving. In order to establish a common basis for future discussions. the solute molecules move along a zigzag path. The net ﬂow of each molecular species occurs in the direction of a negative concentration gradient. As a result of the collisions. Figure 24. individual molecules are in a state of continual yet random motion. Unfortunately. Parrot observed qualitatively that whenever a gas mixture contains two or more molecular species. the concentration of a molecular species can be expressed in many ways. Concentrations. whose relative concentrations vary from point to point. which tends to diminish any inequalities of composition. the quantitative description of molecular diffusion is considerably more complex than the analogous descriptions for the molecular transfer of momentum and energy that occur in a onecomponent phase. Collisions between the solute and the solvent molecules are continually occurring. as it seldom encounters them. isobaric gaseous mixture containing solute and solvent molecules. such as mean free path. Accordingly. a logical explanation of this transport phenomenon can be deduced from the kinetic theory of gases. the molecular transport of momentum and the transport of energy by conduction are also due to random molecular motion. occurs only in mixtures. accordingly. an apparently natural process results. equal elements of volume above and below the section will contain the same number of molecules. let us ﬁrst consider deﬁnitions and relations that are often used to explain the role of components within a mixture.24.

Under conditions in which the ideal gas law. As each molecule of each species has a mass. yA. c ¼ ntotal /V ¼ P/RT. pA V ¼ nA RT. When dealing with a gas phase. and R is the gas constant. r. rA . concentrations are often expressed in terms of partial pressures. For species A.400 Chapter 24 Fundamentals of Mass Transfer mixture of components. is deﬁned as the mass of A per unit volume of the mixture. The mole fraction for liquid or solid mixtures. applies. that is. V is the gas volume. mass concentration. where P is the total pressure. one mole of any species contains a mass equivalent to its molecular weight. that is. xA. must be 1: n å vi ¼ 1 (24-3) i¼1 The molecular concentration of species A. is deﬁned as the number of moles of A present per unit volume of the mixture. is the mass concentration of species A divided by the total mass density vA ¼ rA n ¼ åri rA r (24-2) i The sum of the mass fractions. by deﬁnition. are the molar concentrations of species A divided by the total molar density cA (liquids and solids) c cA (gases) yA ¼ c xA ¼ (24-7) For a gaseous mixture that obeys the ideal gas law. the molar concentration is nA pA ¼ (24-5) cA ¼ V RT where pA is the partial pressure of the species A in the mixture. a mass concentration for each species. yA. is the total moles of the mixture contained in the unit volume. T is the absolute temperature. is the total mass of the mixture contained in the unit volume. can be deﬁned. The total mass concentration or density. as well as for the mixture. n r ¼ å ri (24-1) i¼1 where n is the number of species in the mixture. the mole fraction. and for gaseous mixtures. nA is the number of moles of species A. c. including species A. can be written in terms of pressures yA ¼ cA pA /RT pA ¼ ¼ P/RT c P (24-8) . vA . n c ¼ å ci (24-6) i¼1 or for a gaseous mixture that obeys the ideal gas law. By deﬁnition. cA. the mass concentration and molar concentration terms are related by the following relation: cA ¼ rA MA (24-4) where MA is the molecular weight of species A. The total molar concentration. The mass fraction.

028 kg/mol.032 kg/mol and of nitrogen is 0. Table 24. nitrogen.24. The molecular weight of oxygen is 0. The sum of the mole fractions.1.1 Concentrations in a binary mixture of A and B Mass concentrations r ¼ total mass density of the mixture rA ¼ mass density of species A rB ¼ mass density of species B vA ¼ mass fraction of species A ¼ rA /r vB ¼ mass fraction of species B ¼ rB /r r ¼ rA þ rB 1 ¼ vA þ vB Molar concentrations Liquid or solid mixture Gas mixture c ¼ molar density of mixture ¼ n=V cA ¼ molar density of species A ¼ nA /V cB ¼ molar density of species B ¼ nB /V xA ¼ mole fraction of species A ¼ cA /c ¼ nA /n xB ¼ mole fraction of species B ¼ cB /c ¼ nB /n c ¼ cA þ cB c ¼ n/V ¼ P/RT cA ¼ nA /V ¼ pA /RT cB ¼ nB /V ¼ pB /RT yA ¼ cA /c ¼ nA /n ¼ pA /p yB ¼ cB /c ¼ nB /n ¼ pB /p pA pB P c ¼ cA þ cB ¼ þ ¼ RT RT RT 1 ¼ yA þ yB 1 ¼ xA þ xB Interrelations xA EXAMPLE 1 or rA ¼ cA MA vA /MA yA ¼ vA /MA þ vB /MB xA MA vA ¼ xA MA þ xA MA (24-10) or yA MA yA MA þ yB MB (24-11) The composition of air is often given in terms of only the two principal species in the gas mixture oxygen. . must be 1: n å xi ¼ 1 i¼1 n (24-9) å yi ¼ 1 i¼1 A summary of the various concentration terms and of the interrelations for a binary system containing species A and B is given in Table 24. yo2 ¼ 0:21 yN2 ¼ 0:79 Determine the mass fraction of both oxygen and nitrogen and the mean molecular weight of the air when it is maintained at 258C (298 K) and 1 atm (1.013 105 Pa). O2 . N2 . by definition.1 Molecular Mass Transfer 401 Equation (24-8) is an algebraic representation of Dalton’s law for gas mixtures.

r. is Pa m3 (298 K) (1 mol) 8:314 nRT mol K ¼ V¼ P 1:013 105 Pa ¼ 0:0245 m3 The concentrations are 0:21 mol mol O2 ¼ 8:57 0:0245 m3 m3 0:79 mol mol N2 ¼ ¼ 32:3 0:0245 m3 m3 cO2 ¼ cN2 n c ¼ å ci ¼ 8:57 þ 32:3 ¼ 40:9 mol/m3 i¼1 The total density. accordingly. at 298 K.0288. At ideal conditions. an evaluation of a velocity for the gas mixture requires the averaging of the velocities of each species present. This problem could also be solved using the ideal gas law. When one takes into account the other constituents that are present in air.402 Chapter 24 Fundamentals of Mass Transfer As a basis for our calculations. In a multicomponent system the various species will normally move at different velocities.029 kg/mol.0288 kg. the mean molecular weight of the air must be 0. the mean molecular weight of air is often rounded off to 0. PV ¼ nRT. . consider 1 mol of the gas mixture oxygen present ¼ (1 mol)(0:21) ¼ 0:21 mol (0:032 kg) ¼ 0:00672 kg mol nitrogen present ¼ (1 mol)(0:79) ¼ 0:79 mol ¼ (0:21 mol) (0:028 kg) ¼ 0:0221 kg mol total mass present ¼ 0:00672 þ 0:0221 ¼ 0:0288 kg ¼ (0:79 mol) 0:00672 kg ¼ 0:23 0:0288 kg 0:0221 kg ¼ ¼ 0:77 0:0288 kg vO2 ¼ vN2 As 1 mol of the gas mixture has a mass of 0. is r¼ 0:0288 kg ¼ 1:180 kg/m3 0:0245 m3 and the mean molecular weight of the mixture is M¼ r 1:180 kg/m3 ¼ ¼ 0:0288 kg/mol c 40:9 mol/m3 Velocities. 0 C or 273 K and 1 atm of 1:013 105 Pa pressure. the gas constant is evaluated to be R¼ PV (1:013 105 Pa)(22:4 m3 ) Pa m3 ¼ ¼ 8:314 nT (1 kg mol)(273 K) mol K (24-12) The volume of the gas mixture.

Physik. 1951. and DAB.z is the molar flux in the z direction relative to the molar-average velocity. a species can have a velocity relative to the mass. The mass (or molar) ﬂux of a given species is a vector quantity denoting the amount of the particular species. 59 (1855). deﬁnes the diffusion of component A in an isothermal. JA.z ¼ DAB 1 2 A. This is the velocity that would be measured by a pitot tube and is the velocity that was previously encountered in the equations of momentum transfer. Ann. S. An empirical relation for this molar ﬂux.24. accordingly. the proportionality factor. North-Holland. The molar-average velocity for a multicomponent mixture is defined in terms of the molar concentrations of all components by n å c i vi n å c i vi V ¼ i¼1n ¼ i¼1 c å ci (24-14) i¼1 The velocity of a particular species relative to the mass-average or molar-average velocity is termed a diffusion velocity. that passes per given increment of time through a unit area normal to the vector. in either mass or molar units. Fick. dcA/dz is the concentration gradient in the z direction. R. the Fick rate equation is dcA (24-15) dz where JA. the diffusion velocity of species i relative to the molar-velocity average According to Fick’s law. 94. A more general ﬂux relation that is not restricted to isothermal. . The ﬂux may be deﬁned with reference to coordinates that are ﬁxed in space. We can deﬁne two different diffusion velocities vi v. isobaric system: JA ¼ DAB rcA For diffusion in only the z direction. the diffusion velocity of species i relative to the mass-average velocity and vi V. The basic relation for molecular diffusion deﬁnes the molar ﬂux relative to the molaraverage velocity. coordinates that are moving with the massaverage velocity. often referred to as Fick’s ﬁrst law.or molar-average velocity only if gradients in the concentration exist.1 Molecular Mass Transfer 403 The mass-average velocity for a multicomponent mixture is deﬁned in terms of the mass densities and velocities of all components by n å r i vi n å ri vi v ¼ i¼1n ¼ i¼1 r å ri (24-13) i¼1 where vi denotes the absolute velocity of species i relative to stationary coordinate axes. Fluxes. ﬁrst postulated by Fick1 and. or coordinates that are moving with the molar-average velocity. is the mass diffusivity or diffusion coefficient for component A diffusing through component B.. Thermodynamics of Irreversible Processes. isobaric systems was proposed by de Groot2 who chose to write concentration diffusion overall flux ¼ gradient coefficient density JA. Amsterdam. de Groot.

z.z ¼ cDAB dvA (24-17) dz where dvA /dz is the concentration gradient in terms of the mass fraction.z. we obtain cA v A.z ¼ cDAB dyA þ c A Vz dz For this binary system. is JA.z þ cB v A.404 Chapter 24 Fundamentals of Mass Transfer or dyA (24-16) dz As the total concentration c is constant under isothermal. we obtain JA. we symbolize this new type of ﬂux that is relative to a set of stationary axes by N A ¼ c A vA .z ¼ cA (v A. For a binary system with a constant average velocity in the z direction. the quantities cAv A.z þ cB v B. accordingly. v A. Steffan (1872) and Maxwell (1877).z ) dz (24-19) As the component velocities. using the kinetic theory of gases.z ¼ DAB drA dz Initial experimental investigations of molecular diffusion were unable to verify Fick’s law of diffusion. This was apparently due to the fact that mass is often transferred simultaneously by two possible means: (1) as a result of the concentration differences as postulated by Fick and (2) by convection differences induced by the density differences that resulted from the concentration variation. An equivalent expression for jA. are velocities relative to the ﬁxed z axis.z Vz ) (24-18) Equating expressions (24-16) and (24-18). proved that the mass ﬂux relative to a ﬁxed coordinate was a result of two contributions: the concentration gradient contribution and the bulk motion contribution.z ¼ cA (v A. When the density is constant. this relation simplifies to jA.z ) c cA Vz ¼ yA (cA v A.z and cBv B.z are ﬂuxes of components A and B relative to a ﬁxed z coordinate. upon rearrangement.z Vz ) ¼ cDAB dyA dz which. isobaric conditions. Vz can be evaluated by equation (24-14) as or 1 Vz ¼ (cA v A.z ¼ rDAB jA. the molar ﬂux in the z direction relative to the molar-average velocity may also be expressed by JA.z and v B. the mass ﬂux in the z direction relative to the mass-average velocity.z ) Substituting this expression into our relation.z þ cB v B. equation (24-15) is a special form of the more general relation (24-16). yields cA v A.z ¼ cDAB dyB þ yA (cA v A.

z þ NB. is deﬁned for a binary system in terms of mass density and mass fraction by nA ¼ rDAB =vA þ vA (nA þ nB ) (24-22) where nA ¼ rA vA and nB ¼ rB vB Under isothermal. JA . the concentration gradient contribution: vA (nA þ nB ) ¼ rA v. this relation simpliﬁes to nA ¼ DAB =rA þ vA (nA þ nB ) As previously noted. and yA (NA þ NB ) ¼ cA V the molar flux resulting as component A is carried in the bulk flow of the fluid: This flux term is designated the bulk motion contribution: Either or both quantities can be a signiﬁcant part of the total molar ﬂux. the direction of each of two vector quantities must be evaluated. must be considered and then. we obtain a relation for the ﬂux of component A relative to the z axis NA.24. If species A were diffusing in a multicomponent mixture. the mass flux.1 Molecular Mass Transfer 405 and NB ¼ cB vB Substituting these symbols into equation (24-19). relative to a ﬁxed spatial coordinate system. the expression equivalent to equation (24-21) would be n NA ¼ cDAM =yA þ yA å Ni i¼1 where DAM is the diffusion coefﬁcient of A in the mixture. is a resultant of the two vector quantities: cDAB =yA the molar flux. NA. in turn. resulting from a concentration gradient. the ﬂux is a resultant of two vector quantities: DAB =rA . NA. the bulk motion contribution: .z ) dz (24-20) This relation may be generalized and written in vector form as NA ¼ cDAB = yA þ yA (NA þ NB ) (24-21) It is important to note that the molar flux. Whenever equation (24-21) is applied to describe molar diffusion. resulting from the concentration gradient: This term is referred to as the concentration gradient contribution. nA. the mass flux resulting as component A is carried in the bulk flow of the fluid. isobaric conditions. the vector nature of the individual ﬂuxes. The mass ﬂux. jA . NA and NB.z ¼ cDAB dyA þ yA (NA.

A generalized driving force in chemical thermodynamic terms is dmc /dz where mc is the chemical potential. The ﬂuxes JA and jA are used to describe the mass transfer in diffusion cells used for measuring the diffusion coefﬁcient. JA and NA. If the dye-ﬁlled balloon were dropped into the moving stream. are often used to describe engineering operations within process equipment.2 summarizes the equivalent forms of the Fick rate equation. or the resultant velocity of the molecule while under the influence of a unit driving force.2 Equivalent forms of the mass ﬂux equation for binary system A and B Flux Gradient Fick rate equation Restrictions nA =vA =rA nA ¼ rDAB =vA þ vA (nA þ nB ) nA ¼ DAB =rA þ vA (nA þ nB ) Constant r NA =yA =cA NA ¼ cDAB =yA þ yA (NA þ NB ) NA ¼ DAB =cA þ yA (NA þ NB ) Constant c jA =vA =rA jA ¼ rDAB =vA jA ¼ DAB =rA Constant r jA =yA =cA JA ¼ cDAB =yA JA ¼ DAB =cA Constant c If a balloon. As an example. Related Types of Molecular Mass Transfer According to the second law of thermodynamics. Since chemical reactions are described in terms of moles of the participating reactants. the dye would diffuse radially while being carried downstream.406 Chapter 24 Fundamentals of Mass Transfer Table 24. are used when the Navier– Stokes equations are also required to describe the process. and nA are equivalent statements of the Fick rate equation.z Vz ) ¼ cA DAB dmc RT dz (24-24) Equation (24-24) may be used to deﬁne all molecular mass-transfer phenomena. Any one of these equations is adequate to describe molecular diffusion. The molar diffusion velocity of component A is deﬁned in terms of the chemical potential by v A . The mass ﬂuxes.z V z ¼ u A dmc DAB dmc ¼ dz RT dz (24-23) where uA is the ‘‘mobility’’ of component A. jA. Table 24. thus both contributions participate simultaneously in the mass transfer. systems not in equilibrium will tend to move toward equilibrium with time. the molar ﬂuxes. NA. The diffusion coefﬁcient. the dye will diffuse radially as a concentration gradient contribution. is identical in all four equations. the chemical potential of a component in a homogeneous ideal solution at constant temperature and pressure is deﬁned by mc ¼ m0 þ RT ln cA (24-25) . When a stick is dropped into a moving stream. nA and jA. however. is dropped into a large lake. certain ﬂuxes are easier to use for speciﬁc cases. The ﬂuxes relative to coordinates ﬁxes in space. nA and NA. consider the conditions speciﬁed for equation (24-15). The molar flux of A becomes JA.z ¼ cA (v A. DAB. are used to describe mass-transfer operations in which chemical reactions are involved. The four equations deﬁning the ﬂuxes. Equation (24-23) is known as the Nernst–Einstein relation. it will ﬂoat downstream by the bulk motion contribution. JA. ﬁlled with a color dye.

2 THE DIFFUSION COEFFICIENT Fick’s law of proportionality. and differences in the forces created by external ﬁelds. After the fast neutrons are slowed down through elasticscattering collisions between the neutrons and the nuclei of the reactor’s moderator. and their migration is described by Fick’s law of diffusion. The mean free path of fast neutrons is approximately one million times greater than the free paths of gases at ordinary pressures. a. n. We can. magnetic. electrolytic precipitation due to an electrostatic force ﬁeld. these slower moving neutrons.2 The Diffusion Coefﬁcient 407 where m0 is a constant. results from the random molecular motion over small mean free paths. for example. and electrical ﬁelds. which may be obtained from equation (24-15) JA.24. initially possess high kinetic energies and are termed fast neutrons because of their high velocities. When we substitute this relation into equation (24-24). k/rc p . such as gravity. which is a very small target compared to the volume of most atoms and molecules. the fast neutrons must collide with a nucleus. which is a factor 104 smaller. thermal neutrons. Its fundamental dimensions. the SI units are m2/s. is used successfully in the separation of isotopes. migrate from positions of higher concentration to positions of lower concentration. produced in a nuclear ﬁssion process.z M 1 L2 DAB ¼ ¼ ¼ L2 t M/L3 1/L dcA /dz t are identical to the fundamental dimensions of the other transport properties: kinematic viscosity. Neutrons. that is. DAB. they are very speciﬁc processes. 24. is known as the diffusion coefﬁcient. obtain mass transfer by applying a temperature gradient to a multicomponent system. At these high velocities. the Fick rate equation for a homogeneous phase is obtained JA. the Soret effect or thermal diffusion. The diffusion of fast neutrons and molecules in extremely small pores or at very low gas density cannot be described by this relationship. To be deﬂected. up to 15 million meters per second. This transport phenomenon. which will produce a chemical potential gradient: temperature differences. In the English system ft2/h is commonly used. the chemical potential of the standard state. and thermal diffusivity. in addition to differences in concentration. neutrons pass through the electronic shells of other atoms or molecules with little hindrance. The molecular mass transfer. There are many well-known examples of mass ﬂuxes being induced in a mixture subjected to an external force ﬁeld: separation by sedimentation under the inﬂuence of gravity. or its equivalent ratio. and magnetic separation of mineral mixtures through the action of a magnetic force ﬁeld. Although these mass-transfer phenomena are important. The mass diffusivity has been reported in cm2/s. pressure differences. Components in a liquid mixture can be separated with a centrifuge by pressure diffusion. Conversion between these systems involves the simple relations DAB (cm2 /s) ¼ 104 DAB (m2 /s) DAB (ft2 /h) ¼ 3:87 DAB (cm2 /s) (24-26) . although normally small relative to other diffusion effects. independent of any containment walls.z ¼ DAB dcA dz (24-15) There are a number of other physical conditions. resulting from concentration differences and described by Fick’s law.

1 (1894). Jeans. 38. given by kT (24-28) l ¼ pﬃﬃﬃ 2 2ps A P where u is the mean speed of species A with respect to the molar-average velocity rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ 8kNT (24-29) u¼ pMA Insertion of equations (24-28) and (24-29) into equation (24-27) results in 3 1=2 2T 3/2 k N DAA ¼ (24-30) 2 3 / 2 3p s A P MA where MA is the molecular weight of the diffusing species A. the diffusion coefficients are generally higher for gases (in the range of 5 106 to 1 105 m2 /s). the molecules are regarded as rigid spheres that exert no intermolecular forces.2. Mathematical Theory of Non-Uniform Gases. respectively. Collisions between these rigid molecules are considered to be completely elastic. (g/mol). G. As one might expect from the consideration of the mobility of the molecules. Cowling.4 and Chapman and Cowling. 36. Chapman and T. In the absence of experimental data. 1921. 1959. Experimental values for the diffusivities of gases. K is the Boltzmann constant (1:38 1016 ergs/K). Cambridge University Press. T is the absolute temperature (K). Using a similar kinetic theory of gases approach for a binary mixture of species A and B composed of rigid spheres of unequal diameters. Mag. P is the system pressure. J. 507. and solids are tabulated in Appendix Tables J. and s AB is the Lennard–Jones diameter of the spherical molecules. liquids. Gas Mass Diffusivity Theoretical expressions for the diffusion coefﬁcient in low-density gaseous mixtures as a function of the system’s molecular properties were derived by Sutherland.3. Phil. deﬁned as 1 (24-27) DAA ¼ lu 3 and l is the mean free path of length of species A. London. N is Avogadro’s number (6:022 1023 molecules/mol).. and composition of the system. sometimes as valid as experimental values due to the difﬁculties encountered in their measurement. In the simplest model of gas dynamics. a simpliﬁed model for an ideal gas mixture of species A diffusing through its isotope A yields an equation for the self-diffusion coefﬁcient. semitheoretical expressions have been developed which give approximations. J. London. than for liquids (in the range of 1010 to 109 m2 /s).5 based upon the kinetic theory of gases. the gas-phase diffusion coefﬁcient is shown to be 1 1 1/2 þ 2 K 3/2 1/2 3/2 2MA 2MB N T (24-31) DAB ¼ s þ s 2 3 p B A P 2 3 W. Cambridge University Press. which are higher than the values reported for solids (in the range of 1014 to 1010 m2 /s). Dynamical Theory of Gases.3 Jeans.408 Chapter 24 Fundamentals of Mass Transfer The diffusion coefficient depends upon the pressure. Sutherland. 5 S. 4 . and J.1. With these assumptions. temperature.

Figure 24. the gas-phase diffusion coefficient is dependent on the pressure and the temperature. Rev. the gas-phase diffusion coefficient is an inverse function of total system pressure 1 P a 3/2 power-law function of the absolute temperature DAB / DAB / T3/2 As equation (24-31) reveals.24. Unlike the other two molecular transport coefficients. in atmospheres. Modern versions of the kinetic theory have been attempted to account for forces of attraction and repulsion between the molecules. This is not the case for liquid diffusion coefﬁcients. MB are the molecular weights of A and B. on the dimensionless temperature. B.. s AB is the ‘‘collision diameter. we did not find any composition dependency in equation (24-30) or in the similar equations for viscosity and thermal conductivity. this information is available for only a very few pure gases. a dimensionless function of A the temperature and of the intermolecular potential field for one molecule of A and one molecule of B.2 presents the graphical dependency of the ‘‘collision integral. nonreacting molecules: 1 1 1/2 3=2 0:001858T þ MA MB DAB ¼ (24-33) Ps 2AB VD where DAB is the mass diffusivity of A through B. Bird. In the absence of experimental data.’’ VD . respectively. MA.2 tabulates these values. k is the Boltzmann constant. the diffusion coefficient is dependent on pressure as well as on a higher order of the absolute temperature. and as one of the problems at the end of this chapter points out. When the transport process in a single component phase was examined. which is 1:38 1016 ergs/K. (24-34) (24-35) (24-36) (24-37) . see equation (24-31). 205 (1949). the diffusion coefﬁcients for gases DAB ¼ DBA .44. r and eD . Appendix Table K. in cm2/s. L. Chem. Hirschfelder et al. R. presented an equation for the diffusion coefﬁcient for gas pairs of nonpolar. (1949). and VD is the ‘‘collision integral’’ for molecular diffusion. O. in K. Hirschfelder. Appendix Table K. and E. Specifically. are usually obtained from viscosity data.1 lists VD as a function of kT/eAB . Spotz. and eAB is the energy of molecular interaction for the binary system A and B. the values for pure components may be estimated from the following empirical relations: s ¼ 1:18 Vb1/3 s ¼ 0:841 Vc1/3 1/3 Tc s ¼ 2:44 Pc eA /k ¼ 0:77 Tc 6 J. Unfortunately. T is the absolute temperature. The Lennard–Jones parameters. a Lennard–Jones parameter. viscosity and thermal conductivity.6 using the Lennard–Jones potential to evaluate the inﬂuence of the molecular forces.2 The Diffusion Coefﬁcient 409 Unlike the other two molecular transport coefficients for gases. in ergs. the viscosity and thermal conductivity. P is the absolute pressure. in ˚ .’’ a Lennard–Jones parameter. kT/eAB .

00 10.5 1. v H2 D2 He N2 O2 Air 7.69 Cl S Aromatic ring Heterocyclic ring 19.9 H2O CCIF2 SF6 Cl2 Br2 SO2 12.7 67.50 2.00 0.00 Dimensionless temperature.98 5.6 20.’’ and eA /k ¼ 1:15 Tb (24-38) 3 where Vb is the molecular volume at the normal boiling point. v C H O N 16.50 1.70 2.9 35.00 1.00 0.2 20.1 22.50 0. in K.07 6. in atmospheres.00 4. in (cm) /g mol (this is evaluated by using Table 24.Chapter 24 Fundamentals of Mass Transfer 3.7 37.8 18. WD 410 2.9 14.2 Binary gasphase Lennard–Jones ‘‘collision integral.9 16. Tb is the normal boiling temperature.1 Ar Kr CO CO2 N2O NH3 16.88 17. kT/eΑΒ Figure 24.0 20.7 114.1 For a binary system composed of nonpolar molecular pairs.5 17.3). Tc is the critical temperature.00 Collision integral for diffusion. in K.00 2.3 Atomic diffusion volumes for use in estimating DAB by method of Fuller. Vc is the critical molecular volume.00 8. and Pc is the critical pressure.9 26. Table 24. in (cm)3/g mol.2 41. Schettler. and Giddings Atomic and structure diffusion-volume increments.8 69.2 Diffusion volumes for simple molecules.48 5. the Lennard–Jones parameters of the pure component may be combined empirically by the following relations: s AB ¼ sA þ sB 2 (24-39) pﬃﬃﬃﬃﬃﬃﬃﬃﬃ eA eB (24-40) and eAB ¼ .00 6.

O. John Wiley & Sons. the values of s and e/k are obtained 8 eA /k. P1 ¼ DABT1 . P1 P T V j 2 1 D T2 In Appendix Table J. . we obtain DAB ¼ ¼ 7 0:001858 T 3/2 (1/MA þ 1/MB )1/2 Ps 2AB VD (0:001858)(293)3/2 (1/44 þ 1/29)1/2 (1)(3:806)2 (1:047) ¼ 0:147 cm2 /s J.. unfortunately. Hirschfelder. up to 25 atm. and R.24. Compare this value with the experimental value reported in appendix table J. in A Carbon dioxide 3:996 190 Air 3:617 97 The various parameters for equation (24-33) may be evaluated as follows: s A þ s B 3:996 þ 3:617 ¼ ¼ 3:806 A8 2 2 pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ eAB /k ¼ (eA /k)(eB /k) ¼ (190)(97) ¼ 136 T ¼ 20 þ 273 ¼ 293 K P ¼ 1 atm eAB 136 ¼ ¼ 0:463 kT 293 kT ¼ 2:16 eAB VD (Table K:1) ¼ 1:047 MCO2 ¼ 44 s AB ¼ and MAir ¼ 29 Substituting these values into equation (24-33). we may extend these values to other temperatures. Higher pressures apparently require dense gas corrections. From Appendix Table K. EXAMPLE 2 Evaluate the diffusion coefﬁcient of carbon dioxide in air at 20 C and atmospheric pressure. the proposed modifications are discussed by Hirschfelder. F.1. experimental values of the product DABP are listed for several gas pairs at a particular temperature. Equation (24-33) also states that the diffusion coefﬁcient varies with the temperature as T 3/2 /VD varies. in K s. Curtiss. Bird.2 The Diffusion Coefﬁcient 411 These relations must be modified for polar–polar and polar–nonpolar molecular pairs. no satisfactory correlation is available for high pressures. and Bird. the diffusion coefﬁcient varies inversely with the pressure. B. For moderate ranges of pressure. New York.1. we can predict the diffusion coefﬁcient at any temperature and at any pressure below 25 atm from a known experimental value by 3/2 VD jT1 P1 T2 (24-41) DABT2 .7 The Hirschfelder equation (24-33) is often used to extrapolate experimental data. C. Using equation (24-41). 1954.2. Simplifying equation (24-33). Inc. Curtiss. Molecular Theory of Gases and Liquids.

The Properties of Gases and Liquids. H.T2 T2 VD jT1 Values for VD may be evaluated as follows: at T2 ¼ 273 eAB /kT ¼ at T1 ¼ 293 136 ¼ 0:498 273 VD jT1 ¼ 1:074 VD jT2 ¼ 1:074 (previous calculations) The corrected value for the diffusion coefﬁcient at 208C is 3/2 293 1:074 (0:136) ¼ 0:155 cm2 /s DAB.. 1.I. s i and ei . F. Slattery and R. J.. A. Manual for Predicting Chemical Process Design Data. J.412 Chapter 24 Fundamentals of Mass Transfer From Appendix Table J. Gilliland. J.31 and the diffusion volumes for H2 to be corrected to 6.. Giddings. 681 (1934). Schettler. Prausnitz. T. P. C. and Giddings permits the evaluation of the diffusivity when reliable Lennard–Jones parameters. Ind. Accordingly. Bird. and J.T1 ¼ 273 1:047 (1:55 105 m2 /s) We readily see that the temperature dependency of the ‘‘collision integral’’ is very small. 26. 3937 (1930). 18 (1966).Ch. Danner. M. E. J. 58(5). Chem. . (1983). 86. 82(6). 1977. this equation gives good results for most nonpolar. New York. Arnold.8 Other empirical equations have been proposed9 for estimating the diffusion coefﬁcient for nonpolar. we have DAB ¼ 0:136 cm2 /s Equation (24-41) will be used to correct for the differences in temperature 3/2 DAB. G. Equation (24-33) was developed for dilute gases consisting of nonpolar. Third Edition. Schettler. Othmer and H. Ind. Chem.1 for CO2 in air at 273 K. 10 R. 9 J. Bailey. binary gas systems at low pressures.E. Chapter 11. K. Chem. E. A. R. Eng.7. binary gas systems over a wide range of temperatures. Soc. 8 R. Ind. Eng. E.. (1975). most scaling of diffusivities relative to temperature include only the ratio ðT1 /T2 Þ3/2 . B. Danner and Daubert10 have recommended the atomic and structure diffusion-volume increments for C to be corrected to 15. D. Sherwood. are unavailable. However. T is in K. the authors recommend the addition of the atomic and structural diffusion-volume increments v reported in Table 24.E. spherical monatomic molecules. 52. and P is in atmospheres. P. 249 (1962). C. 4. Eng. The Fuller correlation is 1 1 1/2 103 T 1:75 þ MA MB DAB ¼ (24-42) 2 1 / 3 P (Sv)A þ (Sv)B1/3 where DAB is in cm2/s.9 and for H to 2. C. R. Am. Process Des. 1 atm. Dev.I.12 and for air to 19. To determine the v terms. Reid. Engr. 137 (1958). Chen.3. The empirical correlation recommended by Fuller. McGraw-Hill Book Company. Fuller.. and T. D. Chem.Ch.. Chem. N.T1 T1 VD jT2 ¼ DAB. and T. Daubert.

Brokaw. 1 1 1/2 103 T 1:75 þ MA MB DAB ¼ h i 1=3 2 1 / 3 P (åv)A þ (åv)B 1 1 1/2 þ 44 29 ¼ (1)[(26:9)1/3 þ (20:1)1/3 ]2 103 (293)1:75 ¼ 0:152 cm2 /s This value compares very favorably to the value evaluated with Hirschfelder equation. S. Dev. (24-46) . cm3 /g mol Tb ¼ normal boiling point. The Hirschfelder equation () is still used. Chem. and its determination was easily accomplished. Engr.155 cm2/s. and Giddings equation and compare the new value with the one reported in example 2. Brokaw11 has suggested a method for estimating diffusion coefﬁcient for binary gas mixtures containing polar compounds.24. 0. And VD0 ¼ 11 A C E G þ þ þ (T )B exp(DT ) exp(FT ) exp(HT ) R.2 EXAMPLE 3 The Diffusion Coefﬁcient 413 Reevaluate the diffusion coefﬁcient of carbon dioxide in air at 208C and atmospheric pressure using the Fuller. however. Debye Vb ¼ liquid molar volume of the specific compound at its boiling point. 240 (1969). the collision integral is evaluated by VD ¼ VD0 þ 0:196d2AB T (24-43) where dAB ¼ (dA dB )1/2 d¼ 1:94 103 m2p (24-44) VD TD m p ¼ dipole moment. Schettler. Process Des.. 8. Ind. K and T ¼ kT/eAB where eAB eA eB 1/2 ¼ k k k e/k ¼ 1:18(1 þ 1:3 d2 )Tb (24-45) d is evaluated with (24–44).

B. component 1 diffusing through component n. and y0n is the mole fraction of component n in the gas mixture evaluated on a component-1-free basis. Reid. and Bird13 present an expression in its most general form. O. a process gas stream rich in an inert nitrogen (N2) carrier gas has the following composition: ySIH4 ¼ 0:0075. 13 J. J.. F. Engr. permitting the evaluation of the diffusion coefficients for gases involving polar compounds with errors less than 15%. 14 C. C. Hirschfelder. and R. Wiley. New York. K. Bird. Wilke. D1n is the mass diffusivity for the binary pair. Hirschfelder. The Lennard–Jones constants for silane are eA /k ¼ 207:6 K and ˚. 95–104 (1950). 1977.35 E F G H ¼ 1:035. Molecular Theory of Gases and Liquids. McGraw-Hill Book Company. equation (24-49) is developed by using Wilke’s approach for extending the Stefan and Maxwell theory in order to explain the diffusion of species A through a gas mixture of several components. that is y2 y2 ¼ y02 ¼ y2 þ y3 þ y n 1 y1 In Problem 24. and T. Prausnitz.36 B ¼ 0:156. Curtiss. . Sherwood. s A ¼ 4:08 A 12 R. Mass transfer in gas mixtures of several components can be described by theoretical equations involving the diffusion coefﬁcients for the various binary pairs involved in the mixture. Prog.414 Chapter 24 Fundamentals of Mass Transfer with A ¼ 1:060. R.74 ¼ 3:894. 46. EXAMPLE 4 In the chemical vapor deposition of silane (SiH4) on a silicon wafer. Chapter 11. M. p. yH2 ¼ 0:015. is evaluated with s AB ¼ (s A s B )1/2 (24-47) with each component’s characteristic length evaluated by 1:585 VD 1/3 s¼ 1 þ 1:3 d2 (24-48) Reid. Curtiss.7 at the end of this chapter. and Sherwood12 noted that the Brokaw equation is fairly reliable. The Properties of Gases and Liquids.96 ¼ 1:764. Wilke14 has simpliﬁed the theory and has shown that a close approximation to the correct form is given by the relation D1mixture ¼ 1 y02 /D12 þ y03 /D13 þ þ y0n /D1n (24-49) where D1mixture is the mass diffusivity for component 1 in the gas mixture. s AB . Determine the diffusivity of silane through the gas mixture.87 ¼ 1:529. 718. Chem. Third Edition. New York.00 D ¼ 0:476.10 C ¼ 0:193. yN2 ¼ 0:9775 The gas mixture is maintained at 900 K and 100 Pa total system pressure. C. Prausnitz.11 The collision diameter.

we obtain DSiH4 mixture ¼ 1 y0N2 DSiH4 N2 þ y0H2 DSiH4 H2 ¼ 1 cm2 ¼ 1:10 103 0:9849 0:0151 s þ 1:09 103 4:06 103 This example veriﬁes that for a dilute multicomponent gas mixture. the diffusion coefﬁcient of the diffusing species in the gas mixture is approximated by the binary diffusion coefﬁcient of the diffusing species in the carrier gas. theories of the structure of liquids and their transport characteristics are still inadequate to permit a rigorous treatment. The composition of nitrogen and hydrogen on a silane-free basis are y0N 2 ¼ 0:9775 ¼ 0:9849 1 0:0075 and y0H 2 ¼ 0:015 ¼ 0:0151 1 0:0075 Upon substituting these values into the Wilke equation (24-49). New York. NaCl. separate correlations for predicting the relation between the liquid mass diffusivities and the properties of the liquid solution will be required for electrolytes and nonelectrolytes. Laidler. have been postulated as possible explanations for diffusion of nonelectrolyte solutes in lowconcentration solutions. Two theories. The laws of hydrodynamics provide 15 S.15 The hydrodynamical theory states that the liquid diffusion coefﬁcient is related to the solute molecule’s mobility. Theory of Rate Processes. Glasstone. while others that are designated as electrolytes ionize in solutions and diffuse as ions.24. Inspection of published experimental values for liquid diffusion coefﬁcients in Appendix J. that is.2 The Diffusion Coefﬁcient 415 The binary diffusion coefﬁcients at 900 K and 100 Pa total system pressure estimated by the Hirschfelder equation (24-33) are DSiH4 N2 ¼ 1:09 103 cm2 /s and DSiH4 H2 ¼ 4:06 103 cm2 /s The binary diffusion coefﬁcients are relatively high because the temperature is high and the total system pressure is low. the ideal liquid is treated as a quasicrystalline lattice model interspersed with holes. Eyring. Though each ion has a different mobility. These jumps are empirically related to Eyring’s theory of reaction rate. and molecular diffusion coefﬁcients have no meaning. J. accordingly. . Liquid-Mass Diffusivity In contrast to the case for gases. IX. Certain molecules diffuse as molecules. In the Eyring concept. to the net velocity of the molecule while under the inﬂuence of a unit driving force. 1941. The transport phenomenon is then described by a unimolecular rate process involving the jumping of solute molecules into the holes within the lattice model. Chap. and H.2 reveals that they are several orders of magnitude smaller than gas diffusion coefﬁcients and that they depend on concentration due to the changes in viscosity with concentration and changes in the degree of ideality of the solution. sodium chloride. Needless to say. where we have available an advanced kinetic theory for explaining molecular motion. the electrical neutrality of the solution indicates that the ions must diffuse at the same rate. the diffusion rates of the individual cations and anions must be considered. However. K. McGraw-Hill Book Company. it is possible to speak of a diffusion coefﬁcient for molecular electrolytes such as NaCl. if several ions are present. For example. the Eyring ‘‘hole’’ theory and the hydrodynamical theory. diffuses in water as the ions Naþ and Cl .

5 44. T is the absolute temperature. Cl2 Iodine. the following Table 24. T is absolute temperature. When certain ring structures are involved. COS Sulfur dioxide.Ch.8 18.I. I2 23.6 36. Molecular volumes at normal boiling points.9 30. in cm3 /g mol 14.4.0 51. in cm3 /g mol.5 lists the contributions for each of the constituent atoms. N2 Air Carbon monoxide. R.4 25. A. N2O Ammonia. This equation has been fairly successful in describing the diffusion of colloidal particles or large round molecules through a solvent that behaves as a continuum relative to the diffusing species. Wilke and P.416 Chapter 24 Fundamentals of Mass Transfer relations between the force and the velocity. An equation that has been developed from the hydrodynamical theory is the Stokes–Einstein equation DAB ¼ kT 6prmB (24-50) where DAB is the diffusivity of A in dilute solution in D.. CO2 Carbonyl sulﬁde. MB is the molecular weight of the solvent. the atomic volumes of each element present are added together as per the molecular formulas.E. VA is the molal volume of solute at normal boiling point.4 Molecular volumes at normal boiling point for some commonly encountered compounds Compound Hydrogen.J. mB is the viscosity of the solution. k is the Boltzmann constant. corrections must be made to account for the speciﬁc ring conﬁguration. NH3 Water.6 31.9 32. NO Nitrous oxide. Br2 Chlorine. Empirical correlations. O2 Nitrogen. 264 (1955). in cm3 /g mol Compound Molecular volume. Wilke and Chang16 have proposed the following correlation for nonelectrolytes in an infinitely dilute solution: DAB mB 7:4 108 (FB MB )1/2 ¼ T VA0:6 (24-52) where DAB is the mass diffusivity of A diffusing through liquid solvent B. which attempt to predict the liquid diffusion coefficient in terms of the solute and solvent properties.2 29. H2 Oxygen.9 53. have been developed. H2O Hydrogen sulﬁde.7 34. for some commonly encountered compounds. and mB is the solvent viscosity. 1. Chang. using the general form of equation (24-51). in cm2 /s.3 25. . CO Carbon dioxide. The results of the two theories can be rearranged into the general form DAB mB ¼ f(V) kT (24-51) in which f(V) is a function of the molecular volume of the diffusing solute.4 71. For other compounds. and FB is the ‘‘association’’ parameter for solvent B. in centipoises. are tabulated in Table 24.2 48. in K. VA .5 C.8 Nitric oxide. SO2 16 Molecular volume. H2S Bromine. r is the solute particle radius. Table 24.

Third Edition. Prausnitz.24.6 3. M. in cm3 /g mol 7. VA.0 15.2 The Diffusion Coefﬁcient 417 Table 24. Appendix A.19 17 The correction of FB is recommended by R. 1977. The Properties of Gases and Liquids. (4): 16 (1975). as cyclobutane ﬁve-membered ring.7 37.5 11.0 If data for computing the molar volume of solute at its normal boiling point. . Sherwood. in acids Sulfur Atomic volume.5 12. FB .9 11.K. London..6 y G.8 21. in methyl esters Oxygen.4 9. Third Edition.F. New York. in higher ethers and other esters Oxygen. are given below for a few common solvents.M. in secondary amines 27.5 Recommended values of the association parameter. 1915. are not available. Prausnitz and. New York. The Molecular Volumes of Liquid Chemical Compounds. double bond Nitrogen. in primary amines Nitrogen. Processing.0 25.2617 1. and other unassociated solvents 2. M.5 Atomic volumes for complex molecular volumes for simple substancesy Element Atomic volume.6 10. Ltd. in cm3 /g mol Bromine Carbon Chlorine Hydrogen Iodine Nitrogen. Tyn and Calus18 recommend the correlation VA ¼ 0:285Vc1:048 where Vc is the critical volume of species A in cm3/g. mol. Longmans. Green & Company. 1977. 578.0 12. ether. and Sherwood. as furan pyridine benzene ring naphthalene ring anthracene ring deduct deduct deduct deduct deduct deduct deduct 6 8.C. Reid. 18 Tyn.9 1. Prausnitz. McGraw-Hill Book Company.5 15 15 30 47. K. The Properties of Gases and Liquids. and T. J.5 1. Le Bas. 21. McGraw-Hill Book Company. 19 R. Solvent FB Water Methanol Ethanol Benzene. Sherwood. Values of Vc are tabulated in Reid.0 14. and W. as ethylene oxide four-membered ring. T. J.1 9. C. Calus. Reid.T. p. corrections are recommended: for for for for for for for three-membered ring. in methyl ethers Oxygen. except as noted below Oxygen.0 Element Oxygen. heptane.

I. 20. with the value of the gas diffusivity of ethanol in air at 10 C and 1 atm pressure. 7:96 106 cm2 /s. we would obtain a diffusion coefficient for ethanol in a dilute water solution of 7:85 106 cm2 /s. Scheibel. J.E. G. (24-54) . Eng. Laudie. and VA is the molal volume of the solute at normal boiling point. Hayduk and Laudie20 have proposed a much simpler equation for evaluating inﬁnite dilution diffusion coefﬁcients of nonelectrolytes in water DAB ¼ 13:26 105 m1:14 VA0:589 B (24-53) where DAB is the mass diffusivity of A through liquid B. 611 (1974).05 mol of alcohol/liter of water is 1. Performing a similar calculation.. the viscosity of a solution containing 0. yielding DAB mB K ¼ T VA1/3 20 21 W. mB is the viscosity of water. Ind. Let us compare this value of the liquid diffusivity of ethanol in a dilute solution of water at 10 C. E. in centipoises.45 centipoises. C2 H5 OH. the remaining parameters to be used are T ¼ 283 K FB for water ¼ 2:26 and MB for water ¼ 18 Substituting these values into equation (24-52). Scheibel21 has proposed that the Wilke–Chang relation be modiﬁed to eliminate the association factor.5 as follows: VC2 H5 OH ¼ 2VC þ 6VH þ VO VC2 H5 OH ¼ 2(14:8) þ 6(3:7) þ 7:4 ¼ 59:2 cm3 /mol At 10 C. the liquid diffusion coefﬁcient of water in an inﬁnite dilute solution of ethanol at the same 10 C temperature predicts that the diffusion coefﬁcient DBA is equal to 1:18 105 cm2 /s.418 Chapter 24 EXAMPLE 5 Fundamentals of Mass Transfer Estimate the liquid diffusion coefﬁcient of ethanol. Chem. It is important to note that liquid diffusivities DABL and DBAL are not equal as were the gas diffusivities at the same temperature and pressure. this value is essentially the same value obtained using the Wilke–Chang equation. FB . This relation is much simpler to use and gives similar results to the Wilke–Chang equation.. we obtain DC2 H5 OHH2 O ¼ 7:4 108 (2:26 18)1/2 (59:2)0:6 ¼ 7:96 106 cm2 /s ! 283 1:45 (7:96 1010 m2 /s) This value is in good agreement with the experimental value of 8:3 1010 m2 /s reported in Appendix J. This emphasizes the order of magnitude difference between the values of the liquid and gas diffusivities.Ch. 2007 (1954). in cm3 /g mol. in cm2 /s. 46. If we substitute the values used in example 4 into the Hayduk and Laudie relationship. in a dilute solution of water at 108C. The molecular volume of ethanol may be evaluated by using values from Table 24. 0:118 cm2 /s. Hayduk and H. A.

84 (1970).. Trans. K. Vignes23 recommended the following relationship: DAB ¼ (DAB )xB (DBA )xA where DAB is the inﬁnitely dilute diffusion coefﬁcient of A in solvent B. Chem. DHv . the known relations between electrical conductance and the liquid diffusion coefﬁcient are valid only for dilute solutions of salts in water. Tyne. use K ¼ 17:5 108 . and Sherwood22 recommend this equation for solutes diffusing into organic solvents.000 30. such as an alcohol. R. Lefﬂer and H. Reid. at its normal boiling point temperature.900–30. Most methods for predicting the liquid diffusion coefﬁcients in concentration solutions have combined the inﬁnite dilution coefﬁcients. J. Cullinan. 25 M. Prausnitz. McGraw-Hill Book Company. 24 J.000 46. 23 A.000–39. and xA and xB are the molar fraction composition of A and B.2 where K is determined by The Diffusion Coefﬁcient 419 " K ¼ (8:2 10 8 # 3VB 2/3 ) 1þ VA except 1. if VA < 2:5VB . (kJ/kmol) n 7. 25 M. Chem. Reid. Chem. C. For benzene as a solvent. 9.000 39. Chapter 11. New York. 112 (1981). use K ¼ 18:9 108 . if VA < 2VB . 1977. This exponent may be evaluated from the following table: DHv .. they noted that this equation might evaluate values that had errors up to 20%.. This Vignes equation has been less successful for mixtures containing an associating compound. DAB and DBA . Tc is the critical temperature of solvent B in K. 189 (1966). E. The Properties of Gases and Liquids. DBA is the inﬁnitely dilute diffusion coefﬁcient of B in solvent A. A modiﬁcation for this type of concentrated solution has been proposed by Lefﬂer and Cullinan24 DAB m ¼ (DAB mB )xB (DBA mA )xA As the values of liquid diffusion coefﬁcients reported in the literature were obtained in the neighborhood of the ambient temperature. Tyne25 recommended the following equation for extrapolating to higher temperatures (DAB T1 ) Tc T2 n ¼ (24-55) (DAB T2 ) Tc T1 where T1 and T2 are in K. E. 112 (1981). and n is the exponent related to the latent heat of vaporization of solvent. T. 9. Even so. however.000 >50.24.000–46. in a simple function of composition. J. Vignes. I.. Trans. M. Eng. 2. Fundam. and T. Ind. Eng. Prausnitz. Chem. 22 . 9. Third Edition.000 3 4 6 8 10 The properties of electrically conducting solutions have been studied intensively for more than 75 years.000–50. Sherwood. J. 5. Tyne. For other organic solvents. I. Ind.

Many adsorbent materials are porous to provide a high internal surface area for solute adsorption. Consequently. If the Kn number is much greater than one. This process is known as Knudsen ﬂow or Knudsen diffusion. the diffusing molecule can interact with the wall of the pore. l0þ . typically near the molecular diameter of the molecule itself. respectively. the Kn number goes up as the total system pressure P decreases and absolute temperature T increases. l0 are the limiting (zero concentration) ionic conductances in (amp/cm2 ) (volt/cm) (g equivalent/cm3 ). The Knudsen number. 96. then Knudsen diffusion can be important. The separation of solutes from dilute solution by the process of adsorption is another example. as species A is now more likely to collide with the pore wall as opposed to another molecule. the gas molecules will collide with the pore walls more frequently than with each other. If the pore diameter is smaller than the mean free path of the diffusing gas molecules and the density of the gas is low. The diffusivity for Knudsen diffusion is obtained from the self-diffusion coefﬁcient derived from the kinetic theory of gases rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ lu l 8kNT DAA ¼ ¼ (24-57) 3 3 pMA For Knudsen diffusion. The porous catalyst possesses a high internal surface area to promote chemical reactions at the catalytic surface. The gas ﬂux is reduced by the wall collisions.316 joules/(K)(g mol). In an adsorption process. 8. Pore Diffusivity There are many instances where molecular diffusion occurs inside the pores of porous solids. where nþ and n are the valences of the cation and anion. As the pore diameter approaches the diameter of the diffusing molecule. In both examples. R is the gas constant. we replace path length l with pore diameter dpore. Below. the molecules must diffuse through a gas or liquid phase residing inside the pores. in cm2 /s. Knudsen diffusion applies only to gases because the mean free path for molecules in the liquid state is very small.420 Chapter 24 Fundamentals of Mass Transfer The diffusion coefﬁcient of a univalent salt in dilute solution is given by the Nernst equation DAB ¼ 2RT (1/l0þ þ 1/l0 )F (24-56) where DAB is the diffusion coefficient based on the molecular concentration of A. T is absolute temperature. many catalysts are porous solid pellets containing catalytically active sites on the pore walls. given by Kn ¼ l mean free path length of the diffusing species ¼ dpore pore diameter is a good measure of the relative importance of Knudsen diffusion. At a given pore diameter. Knudsen diffusion. Kn for liquids is very small.500 coulombs/g equivalent. For example. in K. we describe two types of pore diffusion: the Knudsen diffusion of gases in cylindrical pores and the hindered diffusion of solutes in solvent-ﬁlled cylindrical pores. Consider the diffusion of gas molecules through very small capillary pores. This equation has been extended to polyvalent ions by replacing the numerical constant 2 by (1/nþ þ 1/n ). Kn. the solute sticks to a feature on the solid surface that is attractive to the solute. In this . In practice. and F is Faraday’s constant.

Second. and the path for diffusion of the gas molecule within the pores is ‘‘tortuous. there are instances where both Knudsen diffusion and molecular diffusion (DAB ) can be important.’’ For these materials. is determined by 1 1 ayA 1 ¼ þ DAe DKA DAB (24-59) with a¼1þ NB NA For cases where a ¼ 0 (NA ¼ NB ). if an average pore diameter is assumed.2 The Diffusion Coefﬁcient instance. We can make two comparisons of DKA to the binary gas phase diffusivity. or where yA is close to zero. cylindrical pores aligned in a parallel array. a reasonable approximation for the effective diffusion coefficient in random pores is D0Ae ¼ e2 DAe where e ¼ (24-61) the volume occupied by pores within the porous solid total volume of porous solid ðsolid þ poresÞ e is the volume void fraction of the porous volume within the porous material. in most porous materials. and temperature T has units of K. DKA . and temperature. DAB . species A molecular weight. is dependent on the pore diameter. the Knudsen diffusivity for diffusing species A. pure molecular diffusion. or the presence of species B in the binary gas mixture. The four possible types of pore diffusion are illustrated in Figure 24. If we consider that Knudsen diffusion and molecular diffusion compete with one another by a ‘‘resistances in series’’ approach. The Knudsen diffusivity. then the effective diffusivity of species A in a binary mixture of A and B. However. the Knudsen process is signiﬁcant only at low pressure and small pore diameter. it is not a function of absolute pressure P. pure Knudsen . is rﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ dpore dpore 8kNT DKA ¼ u¼ pMA 3 3 sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃ dpore 8 g cm molecules T 16 23 1:38 10 DKA ¼ 6:023 10 p s2 K mol MA 3 rﬃﬃﬃﬃﬃﬃﬃ T ¼ 4850dpore MA 421 (24-58) This simpliﬁed equation requires that dpore has units of cm. vs.3. The ﬁrst three. DAe . Generally. DAB / T 3=2 for the binary gas phase diffusivity. each with their respective diffusivity correlation.24. This ‘‘void fraction’’ is usually experimentally determined for a speciﬁc material. DKA . First. equation (24-59) reduces to 1 1 1 ¼ þ DAe DAB DKA (24-60) The above relationships for the effective diffusion coefficient are based on diffusion within straight. However. MA has units of g/mol. pores of various diameters are twisted and interconnected with one another. the temperature dependence for the Knudsen diffusivity is DKA / T 1=2 .

Silane is diluted to 1.1 kPa) total system pressure. The gas-phase molecular diffusivity of SiH4 –He. EXAMPLE 6 One step in the manufacture of optical ﬁbers is the chemical vapor deposition of silane (SiH4 ) on the inside surface of a hollow glass ﬁber to form a very thin cladding of solid silicon by the reaction SiH4 (g) ! Si(s) þ 2H2 (g) Optical fiber as shown in Figure 24. assess the importance of Hollow glass fiber Knudsen diffusion for SiH4 inside the ﬁber Figure 24. with s SiH 4 ¼ 4:08 A eSiH 4 /k ¼ 207:6 K.001858T3/2 D AB = Knudsen + molecular diffusion 1/2 D KA = d pore 8kNT 3 pMA Random porous material Pore wall A B 1 D Ae @ 1 D AB + 1 D KA D' Ae = e 2 D Ae Figure 24.4 Optical ﬁber. 100 Pa for high bandwidth data transmission have dpore = 10 µm very small inner pore diameters. Shaded areas represent nonporous solids. If the inner diameter of the Si-coated hollow glass ﬁber is 10 mm. The gas-phase . The molecular weight of silane is 32 g/mol. are based on diffusion within straight. the process is carried out at high temperature and SiH4 gas very low total system pressure. The binary gas phase diffusivity of silane in helium at 25 C ˚ and (298 K) and 1.3 kPa) total system pressure is 0:571 cm2 /s. The fourth involves diffusion via ‘‘tortuous paths’’ that exist within the compacted solid. Typically. diffusion. cylindrical pores that are aligned in parallel array. and Knudsen and molecular combined diffusion.0 mol % in the inert carrier gas helium (He).3 Types of porous diffusion. Optical ﬁbers 900 K.4. typically Si thin film H2 gas less than 20 mm (1 mm ¼ 1 106 m).0 atm (101.422 Chapter 24 Fundamentals of Mass Transfer Pure molecular diffusion Pure knudsen diffusion Pore wall Pore wall A A dpore B 1 + 1 MA MB P s2AB WD 0. and effective diffusivity for SiH4 at 900 K and 100 Pa total system pressure must be calculated. lumen at 900 K and 100 Pa (0. Knudsen diffusivity for SiH4.

F1 (w).e ¼ 1 1 DSiH4 He þ 1 ¼ DK.24. General models for diffusion coefﬁcients describing the ‘‘hindered diffusion’’ of solutes in solvent-ﬁlled pores assume the form of DAe ¼ DAB F1 (w)F2 (w) (24-62) The molecular diffusion coefficient of solute A in the solvent B at infinite dilution. then Knudsen diffusion controls the silane transport inside the optical ﬁber if no external bulk transport is supplied.668 at 900 K for gaseous SiH4 –He mixtures. both of which are theoretically bounded by 0 and 1.SiH4 ¼ 4850 dpore ¼ 25:7 ¼ 4850(1 10 ) s MSiH4 32 As the SiH4 is signiﬁcantly diluted in He. Note that the gas phase molecular diffusivity is high due to high temperature and very low system pressure. DA e . reﬂecting the resistance in series approach. As the molecular diameter of the solute approaches the diameter of the pore. Finally. This phenomena is known as solute exclusion. with dpore ¼ 1 103 cm (10 mm) sﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ rﬃﬃﬃﬃﬃﬃﬃﬃ T 900 cm2 3 DK . Hindered solute diffusion in solvent-ﬁlled pores. The Knudsen diffusivity of SiH4 inside the optical ﬁber is calculated using equation (24-58).802 at 298 K and 0. and F2 (w). is reduced by two correction factors. and is used to separate large biomolecules such as proteins from dilute aqueous . Consider the diffusion of a solute molecule through a tiny capillary pore ﬁlled with liquid solvent. the diffusive transport of the solute through the solvent is hindered by the presence of the pore and the pore wall. then the solute is too large to enter the pore.SiH4 1 cm2 ¼ 25:5 1 1 s þ 3:32 103 25:7 The effective diffusivity for SiH4 is smaller than its Knudsen diffusivity. we calculate the Knudsen number for SiH4 erg 1 N m 1:38 1016 900 K K 107 erg kT l ¼ pﬃﬃﬃ 2 ¼ ¼ 1:68 104 m ¼ 168 mm 2 2ps A P pﬃﬃﬃ 1m N 100 2 2p 0:408 nm 9 10 nm m l 168 mm ¼ 16:8 ¼ Kn ¼ dpore 10 mm As Kn 1 and the effective diffusivity is close to the Knudsen diffusivity. the process is dilute with respect to SiH4 and so equation (24-60) can be used to estimate the effective diffusivity DSiH4 .2 The Diffusion Coefﬁcient 423 molecular diffusivity of silane in helium is scaled to process temperature and pressure using the Hirschfelder extrapolation. both correction factors are functions of the reduced pore diameter w w¼ ds dpore ¼ solute molecular diameter pore diameter (24-63) If w > 1. Furthermore. equation (24-41) 900 K cm2 900 K 1:5 101:3 kPa 0:802 cm2 ¼ 3:32 103 DSiH4 He ¼ 0:571 0:1 kPa 298 K 0:1 kPa 0:668 s s It is left to the reader to show that the collision integral VD is equal to 0.

As w approaches 1. that is.424 Chapter 24 Fundamentals of Mass Transfer mixtures containing solutes of much smaller diameter. the effective diffusion coefficient is zero. both F1 (w) and F2 (w) decrease asymptotically toward zero so at w ¼ 1. F1 (w) ¼ flux area available to solute p(dpore ds )2 ¼ ¼ (1 w)2 2 total flux area pdpore (24-64) and holds for 0 . is based on simple geometric arguments for stearic exclusion. The correction factor F1 (w). the stearic partition coefﬁcient.

F1 (w) .

developed by Renkin. assuming diffusion of a rigid spherical solute in a straight cylindrical pore. have been developed. The most common equation. Equations for F2 (w). and ignore electrostatic or other energetic solute–solvent-pore wall intereactions. the hydrodynamic hindrance factor. polydisperity of solute diameters. is based on the complicated hydrodynamic calculations involving the hindered Brownian motion of the solute within the solvent-ﬁlled pore. The analytical models are generally asymptotic solutions over a limited range of w.26 is reasonable for 0 . 1:0 The correction factor F2 ðwÞ. and noncircular pore cross sections.

w .

Renkin. Tanford. B = 10. A mesoporous membrane with cylindrical pores of 30 nm diameter is available (Figure 24. Gen. 0:6 F2 (w) ¼ 1 2:104w þ 2:09w3 0:95w5 EXAMPLE 7 (24-65) It is desired to separate a mixture of two industrial enzymes.44 nm ds. in a dilute. 225 (1954). 38. The properties of each enzyme as reported by Tanford27 are given below. C. J.5). John Wiley & Sons.. 26 27 E. lysozyme and catalase. A = 4.12 nm DAe DBe Bulk solvent dpore = 30 nm Figure 24. .5 Hindered diffusion of solutes in solvent-ﬁlled pores. Physical Chemistry of Macromolecules. M. aqueous solution by a gel ﬁltration membrane. New York. 1961. The following separation factor (a) for the process is proposed a¼ ds. Physiol. Determine the separation factor for this process.

F1 ðwA Þ by equation (24-64) and F2 ðwA Þ by the Renkin equation (24-65) are F1 (wA ) ¼ (1 wA )2 ¼ (1 0:137)2 ¼ 0:744 F2 (wA ) ¼ 1 2:104wA þ 2:09w3A 0:95w5A ¼ 1 2:104(0:137) þ 2:09(0:137)3 0:95(0:137)5 ¼ 0:716 The effective diffusivity of lysozyme in the pore. This mechanism has been mathematically described by assuming a unimolecular rate process and applying Eyring’s ‘‘activated state’’ concept. and the energy of activation associated with the jump. This normally requires a distortion of the lattice.B ¼ 10:44 nm DoBH2 O ¼ 4:10 107 cm2 /s The transport of large enzyme molecules through pores ﬁlled with liquid water represents a hindered diffusion process. ‘‘impurity atoms. the separation factor is DAe 5:54 107 cm2 /s a¼ ¼ 9:06 ¼ DBe 6:12 108 cm2 /s It is interesting to compare the value above with a0 . the length of the jump path. are introduced into solid silicon to control the conductivity in a semiconductor device.2 Lysozyme (species A) MA ¼ 14 100 g/g mol ds.A ¼ 4:12 nm DoAH2 O ¼ 1:04 106 cm2 /s The Diffusion Coefﬁcient 425 Catalase (species B) MB ¼ 250 000 g/g mol ds. The resulting equation is a complex equation relating the diffusivity in terms of the geometric relations between the lattice positions. The hardening of steel results from the diffusion of carbon and other elements through iron. DAe is estimated by equation (24-62) DAe ¼ DAH2 O F1 (wA )F2 (wA ) ¼ 1:04 106 cm2 cm2 (0:744)(0:716) ¼ 5:54 107 s s Likewise. as discussed in the ‘‘hole’’ theory for liquid diffusion.’’ commonly called dopants. the ratio of molecular diffusivities at inﬁnite dilution DAH O 1:04 106 cm2 /s ¼ 1:75 a0 ¼ 2 ¼ 4:1 107 cm2 /s DBH2 O The small pore diameter enhances the value for a because the diffusion of the large catalase molecule is signiﬁcantly hindered inside the pore relative to the smaller lysozyme molecule. The reduced pore diameters for lysozyme and catalase are wA ¼ ds.6. for catalase F1 (wB ) ¼ 0:425. F2 (wB ) ¼ 0:351. In vacancy diffusion. as illustrated in Figure 24. the transported atom ‘‘jumps’’ from a lattice position of the solid into a neighboring unoccupied lattice site or vacancy. and DBe ¼ 6:12 108 cm2 /s: Finally.A 4:12 nm ¼ ¼ 0:137 dpore 30:0 nm and wB ¼ ds. The atom continues to diffuse through the solid by a series of jumps into other neighboring vacancies that appear to it from time to time. Vacancy diffusion and interstitial diffusion are the two most frequently encountered solid diffusion mechanisms.B 10:44 nm ¼ ¼ 0:348 dpore 30:0 nm For lysozyme. In semiconductor manufacturing processes.24. . Solid Mass Diffusivity The diffusion of atoms within solids underlies the synthesis of many engineering materials.

G.28 Appendix Table J. Figure 24. Diffusion of Solids..6 Solid-state vacancy diffusion.. 1941. 1993. Process Engineering Analysis in Semiconductor Device Fabrication. Shewmon. S. Excellent references are available for a more detailed discussion on the diffusion characteristics of atoms in solids (Barrer.7. S. Kou. 1996. Diffusion In and Through Solids. McGraw-Hill Inc. New York. Cambridge University Press.3 lists a few values of binary diffusivities in solids.8 illustrates the dependence of solid-phase diffusion coefﬁcients on temperature. R is the thermodynamic constant (8. Barrer. New York. . New York. speciﬁcally for the diffusion of common dopants in solid silicon. Q is the activation energy (J/mol). as illustrated in Figure 24. London. M. P.7 Solid-state interstitial diffusion.Fundamentals of Mass Transfer State 1 Energy Chapter 24 State 2 State 3 2 Activation energy 1 3 Figure 24. K. John Wiley & Sons Inc.314 J/mol K). 1963. Shewmon. This mechanism is also mathematically described by Eyring’s unimolecular rate theory. This normally involves a dilation or distortion of the lattice.. Do is a proportionality constant of units consistent with DAB . Hochberg. R. The solid-phase diffusion coefﬁcient has been observed to increase with increasing temperature according to an Arrhenius equation of the form DAB ¼ Do eQ/RT (24-66) or ln(DAB ) ¼ Q 1 þ ln(Do ) RT (24-67) where DAB is solid diffusion coefficient for the diffusing species A within solid B. McGraw-Hill Inc. Middleman and A. Kou. State An atom moves in interstitial diffusion by jumping from one interstitial site to a neighboring one. Middleman and Hochberg. Transport Phenomena and Materials Processing. State 1 Energy 426 State 2 2 Activation energy 1 3 State 28 State 3 Figure 24. and T is the absolute temperature (K).

Kramer.9 Figure 24.29 and by Crank and Park. Spalding. Diffusion in Polymers. Chem.. J. Data from Figure 24. Engineering Design for Plastics. S. 1964. L. Table 24.9 200.8 1000/ T ( K –1) 0. 1311 (1930). B = Si (cm2/s) 10–10 10–11 10–12 10–13 Ga Sb 10–14 0. New York.7 238. Chem.7 284.1 239.8 Diffusion coefﬁcients of substitutional dopants in crystalline silicon.6 Data for self-diffusion in pure metals Structure Metal Do (mm2 /s) Q (kJ/mole) fcc fcc fcc fcc bcc bcc Au Cu Ni FeðgÞ FeðaÞ FeðdÞ 10.7 31 190 49 200 1980 176. Rogers. These tables point out the significant energy barrier that must be surpassed when an atom jumps between two lattice sites by vacancy diffusion (Table 24. O.32 29 30 31 32 C.8 can be used to estimate Q for a given dopant in silicon using equation (24-67).6 In Al B. Academic Press. Phys.3 279.7 provide the diffusion data needed to evaluate DAB by equation (24-66) for self-diffusion in pure metals and interstitial solutes in iron. Reinhold Press. E.7). G.30 Diffusivities of solutes in dilute biological gels are reported by Friedman and Kramer31 and by Spalding. Soc. Park. Tables 24. Crank and G. E.. Am. 1968. J. J.6) and a significantly smaller energy barrier encountered in interstitial diffusion (Table 24.2 The Diffusion Coefﬁcient 427 10–9 DAB. 52.24. P 0.6 and 24. .7 0. New York. Friedman and E. 3380 (1969).5 Diffusion coefﬁcients and solubilities of solutes in polymers are reported by Rogers.

DcA is the concentration difference between the boundary surface concentration and the average concentration of the fluid stream of the diffusing species A. kc.3 CONVECTIVE MASS TRANSFER Mass transfer between a moving ﬂuid and a surface or between immiscible moving ﬂuids separated by a mobile interface (as in a gas/liquid or liquid/liquid contactor) is often aided by the dynamic characteristics of the moving ﬂuid. the process is called free or natural convection. As this is always true. a distinction must be made between two types of ﬂow. It is important for the student to recognize the close similarity between the convective mass-transfer coefﬁcient and the convective heat-transfer coefﬁcient.4 144. a function of system geometry. The reciprocal of the coefﬁcient. in general. the mechanism of mass transfer between a surface and a ﬂuid must involve molecular mass transfer through the stagnant and laminar ﬂowing ﬂuid layers. represents the resistance to the transfer through the moving ﬂuid. ﬂuid and ﬂow properties. is accordingly referred to as the ﬁlm mass-transfer coefﬁcient. This mode of transfer is called convective mass transfer. and kc is the convective mass-transfer coefficient. with the transfer always going from a higher to a lower concentration of the species being transferred.2 24. we can recall that there is always a layer. equation 15. When an external pump or similar device causes the ﬂuid motion. close to the surface where the ﬂuid is laminar. the process is called forced convection. The rate equation for convective mass transfer.11 is NA ¼ kc DcA (24-68) where NA is the molar mass transfer of species A measured relative to fixed spatial coordinates.1 2.1 76. Equation (24-68) deﬁnes the coefﬁcient kc in terms of the mass ﬂux and the concentration difference from the beginning to the end of the masstransfer path. This immediately suggests that the techniques developed for evaluating the convective heat-transfer coefﬁcient may be repeated for convective mass transfer. A complete discussion of convective mass-transfer coefﬁcients and their evaluation is given in Chapters 28 and 30. As in the case of convective heat transfer.0 0. It is. sometimes extremely thin. 1/kc. If the ﬂuid motion is due to a density difference.3 0. The controlling resistance to convective mass transfer is often the result of this ‘‘ﬁlm’’ of ﬂuid and the coefﬁcient.428 Chapter 24 Fundamentals of Mass Transfer Table 24. .5 84.7 Diffusion parameters for interstitial solutes in iron Structure bcc bcc bcc fcc Solute Do (mm2 /s) Q (kJ/mole) C N H C 2.1 13. As in the case of molecular mass transfer. and that ﬂuid particles next to the solid boundary are at rest. and the concentration difference DcA : From our experiences in dealing with a ﬂuid ﬂowing past a surface. Convective transfer depends on both the transport properties and the dynamic characteristics of the ﬂowing ﬂuid. generalized in a manner analogous to Newton’s ‘‘law’’ of cooling. Chapters 28 and 30 consider the methods of determining this coefﬁcient. convective mass transfer occurs in the direction of a decreasing concentration.

fundamental relations were presented for concentrations and velocities of the individual species as well as for the mixture. liquid. acetone exerts a vapor pressure of 161 mmHg or 2:148 104 Pa: Therefore. . The molar composition of the commercial LNG is methane. C2H6 4. The acetone temperature is maintained at 290 K.4 105 Pa.6% Propane. CH4 93. As diffusion of mass involves a multicomponent mixture. is to be shipped from the Alaskan Kenai Peninsula by an ocean carrier to processing plant on Yaquina Bay. CO2 0. the partial pressure of methane when the total pressure is 1:4 105 Pa.7% determine a. b. The molecular transport property. the mass fraction of carbon dioxide in parts per million by weight.mol/s. Thus m kg:mol kg:mol 0 ¼ 0:1732 WA ¼ kc A(cAs cA1 ) ¼ 0:0324 (0:6 m2 ) 8:91 s m3 s 24. for the mass transfer of acetone into the nitrogen stream is 0. the concentration of acetone in the gas phase at the acetone surface is cAs ¼ PA 2:148 104 Pa kg mol ¼ ¼ 8:91 RT m3 Pa m3 (290 K) 8:314 kg mol K and the concentration of acetone in the nitrogen carrier gas is near zero because the molar ﬂowrate of the carrier gas is in a large excess relative to the rate of acetone transfer.1 Liquiﬁed natural gas. molecular and convective mass transfer. has been discussed and correlating equations presented. LNG. C5H8 1. the density of the gas mixture when heated to 207 K and at 1. and solid systems. c. The rate equations for the mass transfer of species A in a binary mixture are as follows: molecular mass transfer: JA ¼ cDAB =yA jA ¼ rDAB =vA NA ¼ cDAB =yA þ yA (NA þ NB ) nA ¼ rDAB =vA þ vA (nA þ nB ) molar ﬂux relative to the molar-average velocity mass ﬂux relative to the mass-average velocity molar ﬂux relative to ﬁxed spatial coordinates mass ﬂux relative to ﬁxed spatial coordinates convective mass transfer: NA ¼ kc DcA PROBLEMS 24. The total molar rate of acetone transfer from the liquid to the gas phase can be evaluated by WA ¼ NA A ¼ kc A(cAs cA1 ) The mass transfer area is speciﬁed as 0:6 m2 : At 290 K. e. If the average mass-transfer coefﬁcient.5 mol % ethane. kc. the average molecular weight of the LNG mixture. determine the total rate of acetone release in units of kg.2% Carbon dioxide.4 CLOSURE In this chapter. the weight fraction of ethane. d. have been introduced. DAB .0324 m/s. Oregon. the diffusion coefﬁcient or mass diffusivity in gas. the two modes of mass transport.Problems EXAMPLE 8 429 A pure nitrogen carrier gas ﬂows parallel to the 0:6 m2 surface of a liquid acetone in an open tank.

6 Starting with Fick’s equation for the diffusion of A through a binary mixture of A and B. prove a. Estimate the diffusivity of ammonia in air at 1:013 105 Pa and 373 K using the Brokaw equation (24-43). The air is at 373 K and 1:5 105 Pa. Prog. jA þ jB ¼ 0: 24. mole fraction of O2.13 An absorption tower has been proposed to remove selectively two pollutants.2 K and the critical volume (VC) of H2S is 98. mass density of N2. JA ¼ DAB =cA 24. a thin ﬁlm of solid silicon (Si) is uniformly deposited on a wafer surface by the chemical decomposition of silane (SiH4) in the presence of H2 gas. b. weight of the mixture. isobaric gaseous system: dcA ¼ b rA rB ðvAz vBz Þdz MA MB Wilke33 extended this theory to a multicomponent gas mixture. the weight fraction of these species.4 Starting with Fick’s equation for the diffusion of A through a binary mixture of species A and B as given by NAz ¼ cDAB dydzA þyA ðNAz þ NBz Þ and Fick’s equation for the diffusion of B through the same binary mixture given by NBz ¼ cDBA dydzB þyB (NBz þ NAz ).5 cm3/mol. Compare the evaluated value with the experimental value reported in Appendix Table J.9 The isomerization of n-butane to iso-butane is carried out on a catalyst surface at 2. The appropriate form of the Maxwell-type equation was assumed to be dcA r r r r ¼ bAB A B ðvAz vBz Þ þ bAC A C ðvAz vCz Þ dz MA MB MA MC r r þbAD A D ðvAz v0z Þ þ MA MD Using this relation. and 20 mol % HCl. and a proportionality constant. The following equation expresses mathematically the resistances for an isothermal. d. the resistances that must overcome the molecular mass transfer. verify equation (24-49).2 In the manufacture of microelectronic devices. f. The Lennard–Jones parameters for SiCl4 ˚. Wilke. b. nA þ nB ¼ rv c. c. prove the two gas diffusivities. average molecular weight of the gas mixture. . The dipole moment for ammonia is 1.0 atm and 4008C. b. 24.8 Determine the value of the following gas diffusivites using the Hirschfelder equation: a. 24. carbon dioxide/air at 310 K and 1:5 105 Pa ethanol/air at 325 K and 2:0 105 Pa carbon monoxide/air at 310 K and 1:5 105 Pa carbon tetrachloride/air at 298 K and 1:913 105 Pa 24. the average molecular weight of the gas mixture.11 An absorption tower is proposed to remove selectively ammonia from an exhaust gas stream. stating the assumptions made in the derivations: a.. DAB and DBA . are equal. 33 C. Eng. of SiH4 if the feed gas is maintained at 900 K and a system pressure of 60 torr. 24. e. Chem.430 Chapter 24 Fundamentals of Mass Transfer 24. s A ¼ 5:08 A 24. determine a. d. c. NA þ NB ¼ cV. b. 40 mol % H2.5 Starting with the Fick’s equation for the diffusion of A through a binary mixture of components A and B NA ¼ cDAB =yA þ yA (NA þ NB ) derive the following relations. What is the gas-phase molecular diffusion coefﬁcient of n-butane in iso-butane? Compare values obtained from both the Hirschfelder and Fuller– Schettler–Giddings equations. nA ¼ DAB =rA þ wA ðnA þ nB Þ b. Estimate the molecular diffusion coefﬁcient for (a) SiCl4 in H2 and (b) SiCl4 in a gas phase mixture containing 40 mol % SiCl4. g. the molar concentration.7 Stefan and Maxwell explained the diffusion of A through B in terms of the driving force dcA . 95 (1950). 46. hydrogen sulﬁde (H2S) and sulfur dioxide (SO2). Does the Hirshfelder equation for gas evaluating gas diffusivities verify this same equality? 24. 24. b. 24.12 Highly puriﬁed tetrachlorosilane (SiCl4) gas is reacted with hydrogen gas (H2) to produce electronic-grade polycrystalline silicon at 8008C and 1:5 105 Pa according to the equction: SiCl4 ðgÞ þ 2H2 ðgÞ ! SiðsÞ þ 4HClðgÞ: There are concerns that the reaction experiences diffusional limitations at the growing Si solid surface. 24. If the gas composition is maintained at 40 mol % SiH4 and 60 mol % H2. volume fraction of O2. mass density of the air. from an exhaust gas stream containing H2S SO2 N2 3 vol % 5 vol % 92 vol % Estimate the diffusivity of hydrogen sulﬁde in the gas mixture at 350 K and 1:013 105 Pa: The critical temperature (TC) of H2S is 373. h. mass density of O2. mass density of the air. c.10 Determine the diffusivity of methane in air using (a) the Hirschfelder equation and (b) the Wilke equation for a gas mixture.46 debye. cA. (species A) are eA /k ¼ 358 K.3 Air is contained in a 30 m3 container at 400 K and 1:013 105 Pa: Determine the following properties of the gas mixture: a.1.

3. Learn about the structures of fcc and bcc iron in a materials science textbook. Benzene is very sparingly soluble in water and has a molecular diameter of 0. n-butanol in water at 288 K. Determine the liquid diffusion coefﬁcient of chlorine in an inﬁnitely dilute solution of water at 289 K using (a) the Wilke–Chang equation and (b) the Hayduk–Laudie equation. species B) at 1008C and 1. Compare this value with experimental value reported in Appendix J. gas diffusion processes may play a role in the operation of this device. Hence. From the measurements.4. which is very small.5 atm total system pressure. What is the effective diffusion coefﬁcent of O2 (DAez ) within the microchamber? 24. the measured effective diffusion coefﬁcient of hydrogen was 0. Random pores 100 A of 0.22 Researchers are proposing the development of a ‘‘nanochannel reactor’’ for steam reforming of methane (CH4) to fuelcell hydrogen gas to power microscale devices. spherical. mol.2. species A within the water-ﬁlled pores of the particle. 24. a.18 The aeration of water is an important industrial operation. MB ¼ 18 g/g: mol) at 3008C and 0. Estimate the liquid-phase diffusion of benzene in ethanol and ethanol in benzene at 288 K by (a) the Wilke–Chang equation and (b) the Scheibel equation. The process is isothermal at 298 K. From Figure 24. The benzene solute does not adsorb onto the intersurfaces of the pores. the gas ﬂow is likely to be very small within a given channel. Estimate the mean diameter of a serum albumin molecule. Determine the liquid diffusion coefﬁcient of oxygen in an inﬁnitely dilute solution of water at 288 K using (a) the Wilke– Chang equation and (b) the Hayduk–Laudie equation.0 atm total system pressure and 808C in a porous material of monodispersed pore size distribution and void volume fraction of 0. the nominal value of the diffusion coefﬁcient for phosphorus in silicon at 1316 K is 1 1013 cm2 /s and at 1408 K is 1 1012 cm2 /s: Determine the value of the diffusion coefﬁcient at 1373 K. Random pores 1000 A of 0. 24.0 mol % O2 (species A) in a helium carrler gas (species B) enters the microscale chamber. ˚ in diameter with void fraction c. d. d.8.15 Estimate the liquid diffusivity of the following solutes that are transferred through dilute solutions: a.20 The case-hardening of mild steel involves the diffusion of carbon into iron. Estimate the diffusion coefﬁcient for carbon diffusing into fcc iron and bcc iron at 1000 K.0 atm within the following materials: ˚ pores in parallel array. b. methanol in water at 283 K. CAc . particularly during the mixing and heating steps. and then explain why the diffusion coefﬁcients are different. What is the molar concentration of oxygen gas at the entrance to the microchamber? b. inert mineral particle. Straight 20.23 Diffusion experiments were conducted with a binary mixture of synthesistic gas containing H2 (species A) diluted in a large excess of CO (species B) at 2.24 A mixture 1.15 nm.22 nm. The total system pressure within the chamber is 300 Pa. 24. Gas phase diffusion in nanochannel A = CH4. The molecular weight of oxygen is 32 g/g.4.40.0 g/g.25 Nanochannel 300°C. water in methanol at 288 K.25 Consider a single.16 Water supplies are often treated with chlorine as one of the processing steps in treating wastewater. mol and helium is 4.5 atm 200 nm As each channel diameter is so small. 0. . What is effective diffusion coefﬁcient of CH4 in the nanochannel at the feed gas conditions? Is Knudsen diffusion important? 24. The diameter of the channel is 200 nm ð1 109 nm ¼ 10 mÞ: A feed gas containing 20 mol % CH4 in water vapor is fed to the nanochannel with a ﬂux ratio NA /NB ¼ 0:25. The average diameter of the pores is 150 nm and the void fraction is 0. B= H2O A+B 20 mol% CH 4 NA/NB = 0.14 The Stokes–Einstein equation is often used to estimate the molecular diameter of large spherical molecules from the molecular diffusion coefﬁcient. Compare the results with the experimental value reported in Appendix J. The measured molecular diffusion coefﬁcient of the serum albumin (an important blood protein) in water at inﬁnite dilution is 5:94 107 cm2 =s at 293 K. porous.036 cm2/s. ˚ pores in parallel array.Problems 24. oxygen in ethanol at 293 K. species A) diffusing into nitrogen gas (N2. Straight 100 A ˚ in diameter with void fraction b.2.17 Benzene (species A) is often added to ethanol to denature the ethanol (species B). The pores inside the particle are ﬁlled with liquid water (species B). Does DAB ¼ DBA ? 24. The chamber consists of a channel that is 5:0 mm (microns) in diameter.000 A 431 24. The chamber temperature is maintained at 1008C. c. The known value is 7. We are interested in analyzing the molecular diffusion of the contaminant benzene C6H6. What is the mean pore size ðdpore Þ of this material? 24. a. MA ¼ 16g/g: mol) in water vapor (species B. We are speciﬁcally interested in evaluating the effective diffusion coefﬁcient of methane gas (species A.21 Determine the effective diffusion coefﬁcient for hydrogen gas (H2. 24. 24. 24. The concentration of dissolved benzene in the water surrounding the particle.19 A silicon wafer is doped with phosphorus.

Estimate the mean pore diameter of the support. glucose (solute A) in aqueous solution is diffusing across a microporous membrane.0 nm diameter. . The average pore diameter of the porous membrane material is 15 nm. there is no dissolved benzene inside the water-ﬁlled pores.38 nm. The molecular diffusion coefﬁcient of ribonuclease in water is 1:19 106 cm2 /s at 298 K. The molecular diffusion coefﬁcient of urease in water at inﬁnite dilution is 3:46 107 cm2 /s at 2988 K. Diffusion of glucose across a microporus polymer membrane ds. DAz . Initially. Is Knudsen diffusion important? You may assume that the CO is dilute in the gas mixture.86 nm (nanometers). The system operates at 5.86 nm Glucose solution (30°C) Microporous membrane (2 mm thick) Glucose solution (30°C) 3.29 The diffusion rate of the enzyme ribonuclease into a porous chromatography support was measured at 298 K.26 Steam reforming of hydrocarbons is one way to make hydrogen gas (H2) for fuel cells.0 atm total system pressure and 4008C.27 As part of a bioseparation process. The critical volume (Vc) of benzene is 259 cm3 /g:mol: What is the effective diffusion coefﬁcient of benzene inside the porous particle? Is pore diffusion important? 24. of glucose through the membrane? 24. An important aspect of this separation process is the diffusion of the protein into the porous matrix of the chromatography support used to affect the separation. It is desired to separate a mixture of CO and H2 using a microporous ceramic membrane. as shown in the following ﬁgure. and the diameter of the molecule is 12.A = 0. What is the effective diffusion coefﬁcient.0 mm. The temperature is 308C.28 Protein mixtures in aqueous solution are commonly separated by molecular sieve chromatography.0 nm The thickness of the membrane is 2. 24. and the diameter of the molecule is 3.432 Chapter 24 Fundamentals of Mass Transfer is constant with time. Determine the effective diffusion coefﬁcient of CO in the gas mixture within the microporous membrane. the product gas contains H2 contaminated with carbon monoxide (CO). Estimate the effective diffusion coefﬁcient of the enzyme urease in a silica gel support with 100 nm diameter pores. which must be further enriched in H2 in order for the fuel cell to work better. The mean diameter of a single glucose molecule is 0. 24. and the void fraction e ¼ 0:10. and the pores running through the membrane consists of parallel channels of 3.6 nm. However. Assume that the molecular diffusion coefﬁcient of glucose in water is described by Stokes–Einstein relationship. and an effective diffusion coefﬁcient of 5:0 107 cm2 /s was backed out of the data.

Additional differential equations will be obtained when we insert.1. The control-volume expression for the conservation of mass is ZZ @ rðv : nÞdA þ @t c:s: ZZZ y r dV ¼ 0 (4-1) c:v: ∆y ∆z which may be stated in words as 8 9 8 9 <net rate of mass= <net rate of accumulation= efflux from þ of mass within control ¼ 0 : . : volume 433 . Once again. : . we shall establish the equation of continuity for a given species. Figure 25. into the continuity equation. : . mass ﬂux relationships developed in the previous chapter. control volume volume ∆x x z If we consider the conservation of a given species A. The general relation for a mass balance of species A for our control volume may be stated as 9 8 8 9 8 9 > rate of chemical > > > = < net rate of mass = < net rate of accum = < production of A ¼ 0 (25-1) efflux of A from þ ulation of A within within the control > : . 25. through which a mixture including component A is ﬂowing.1 THE DIFFERENTIAL EQUATION FOR MASS TRANSFER Consider the control volume. the production or disappearance of A by chemical reaction within the volume. By making a mass balance over a differential control volume. we shall use this approach to develop the differential equations for mass transfer. Dx Dy Dz. the general differential equations for momentum transfer are derived by the use of a differential control volume concept. > > > control volume control volume . By an analogous treatment. the general differential equations for heat transfer are generated in Chapter 16.Chapter 25 Differential Equations of Mass Transfer In Chapter 9. as shown in Figure 25.1 A differential control this relation should also include a term that accounts for volume.

the mass of A transferred across the area Dy Dz at x will be rA v Ax Dy Dzjx .z Dx Dyjz @r þ A Dx Dy Dz rA Dx Dy Dz ¼ 0 @t (25-2) Dividing through by the volume. Substituting each term in equation (25-1). and a discussion of their meanings will be given below.z are the rectangular components of the mass flux vector.z Dx DyjzþDz nA. and canceling terms. this yields @ @ @ @r nA. Dx Dy Dz.x jxþDx nA. and Dz approach zero. x þ nB. where rA has the units (mass of A produced)/(volume)(time).y þ nA. nA ¼ rA vA . The net rate of mass efﬂux from the control volume may be evaluated by considering the mass transferred across control surfaces.z jz @rA þ þ þ rA ¼ 0 Dx Dy Dz @t (25-3) Evaluated in the limit as Dx. Dy.y .x .x Dy DzjxþDx nA. nA. The differential equations are @ @ @ @r nB.y jy nA. the rate of production of A is rA Dx Dy Dz This production term is analogous to the energy generation term that appeared in the differential equation for energy transfer.x Dy Dzjx þ nA. For example. nA.z Dx Dyjz The rate of accumulation of A in the control volume is @rA Dx Dy Dz @t If A is produced within the control volume by a chemical reaction at a rate rA.434 Chapter 25 Differential Equations of Mass Transfer The individual terms will be evaluated for constituent A. The net rate of mass efﬂux of constituent A will be in the x direction: nA.x jx nA.y Dx Dzjy and in the z direction: nA. as discussed in Chapter 16.x þ nA. it would be nA. or in terms of the ﬂux vector.x Dy DzjxþDx nA. As nA.x Dy Dzjx in the y direction: nA. y þ nB. equation (25-4) may be written @r (25-5) =: nA þ A rA ¼ 0 @t A similar equation of continuity may be developed for a second constituent B in the same manner.y Dx Dzjy þ nA.y Dx DzjyþDy nA. z þ B rB ¼ 0 @x @y @z @t (25-6) .z jzþDz nA. and nA.z Dx DyjzþDz nA. we have nA. x Dy Dzjx .y Dx DzjyþDy nA. we obtain nA.z þ A rA ¼ 0 @x @y @z @t (25-4) Equation (25-4) is the equation of continuity for component A.y jyþDy nA.

This equation is rDvA þ=: jA rA ¼ 0 Dt (25-10) We could follow the same development in terms of molar units. the equation of continuity for species A in terms of the substantial derivative may be derived. Adding equations (25-5) and (25-7).1 The Differential Equation for Mass Transfer 435 and =: nB þ @rB rB ¼ 0 @t (25-7) where rB is the rate at which B will be produced within the control volume by a chemical reaction. Equation (25-9) is identical to the equation of continuity (9-2) for a homogeneous fluid. we obtain =: rv þ @r ¼0 @t (25-9) This is the equation of continuity for the mixture. If RA represents the rate of molar production of A per unit volume. Substituting these relations into (25-8). the molar-equivalent equations are for component A for component B =: NA þ @cA RA ¼ 0 @t (25-11) =: NB þ @cB RB ¼ 0 @t (25-12) and for the mixture =: ðNA þ NB Þ þ @ðcA þ cB Þ ðRA þ RB Þ ¼ 0 @t (25-13) . we obtain =: (nA þ nB ) þ @(rA þ rB ) ðrA þ rB Þ ¼ 0 @t (25-8) For a binary mixture of A and B. the continuity equation for the mixture can be rearranged and written Dr þr=: v ¼ 0 Dt (9-5) Through similar mathematical manipulations.25. we have nA þ nB ¼ rA vA þ rB vB ¼ rv rA þ rB ¼ r and r A ¼ r B by the law of conservation of mass. The equation of continuity for the mixture and for a given species can be written in terms of the substantial derivative. and RB represents the rate of molar production of B per unit volume. As shown in Chapter 9.

we obtain =: cDAB =yA þ =: cA V þ @cA RA ¼ 0 @t (25-16) Either equation (25-15) or (25-16) may be used to describe concentration profiles within a diffusing system.2 SPECIAL FORMS OF THE DIFFERENTIAL MASS-TRANSFER EQUATION Special forms of the equation of continuity applicable to commonly encountered situations follow. the equation of continuity for the mixture in molar units is =: cV þ @c ðRA þ RB Þ ¼ 0 @t (25-14) 25. with their qualifying assumptions. Important forms of the equation of continuity. we obtain =: rDAB =vA þ =: rA v þ @rA rA ¼ 0 @t (25-15) and substituting equation (24-21) into equation (25-11). and the diffusion coefﬁcient. Both equations are completely general. however.436 Chapter 25 Differential Equations of Mass Transfer For the binary mixture of A and B. DAB can be assumed constant. nA and NA . they are relatively unwieldy. These equations can be simplified by making restrictive assumptions. we replace the ﬂuxes. These expressions are NA ¼ cDAB =yA þ yA (NA þ NB ) (24-21) or its equivalent NA ¼ cDAB =yA þ cA V and nA ¼ rDAB =vA þ vA (nA þ nB ) (24-22) or its equivalent nA ¼ rDAB =vA þ rA v Substituting equation (24-22) into equation (25-5). by the appropriate expressions developed in Chapter 24. only when the stoichiometry of the reaction is Aﬁ B which stipulates that one molecule of B is produced for each mole of A disappearing. r. include: (i) If the density. we have NA þ NB ¼ cA v A þ cB v B ¼ cV and cA þ cB ¼ c However. can we stipulate that RA ¼ RB . In general. In order to use the equations for evaluating the concentration proﬁles. equation (25-15) becomes 0 DAB = rA þ rA =: v þ v: =rA þ 2 ! @rA rA ¼ 0 @t .

or stationary liquids. RA ¼ 0. The assumption of no fluid motion restricts its applicability to diffusion in solids. the case of equimolar counterdiffusion. (iii) In a situation in which there is no ﬂuid motion. we obtain v: =cA ¼ DAB r2 cA (25-22) If additionally. but acting in the opposite direction to N. and (25-20) may be simpliﬁed further when the process to be deﬁned is a steady-state process. we obtain v: =cA þ @cA ¼ DAB =2 cA þ RA @t (25-17) (ii) If there is no production term. rewriting the left-hand side of equation (25-18). and no chemical production. constant diffusivity. and if the density and diffusion coefﬁcient are assumed constant. the equation becomes v: =cA ¼ DAB =2 cA þ RA (25-21) For constant density. and no variation in the diffusivity or density. that is. (25-18). where NA is equal in magnitude. The similarity between these two equations is the basis for the analogies drawn between heat and mass transfer.2 Special Forms of the Differential Mass-Transfer Equation 437 Dividing each term by the molecular weight of A and rearranging.25. equation (25-18) reduces to @cA ¼ DAB =2 cA @t (25-20) Equation (25-20) is commonly referred to as Fick’s second ‘‘law’’ of diffusion. v ¼ 0 the equation reduces to =2 cA ¼ 0 Equation (25-23) is the Laplace equation in terms of molar concentration. equation (25-17) reduces to @cA þv: =cA ¼ DAB =2 cA @t (25-18) DcA ¼ DAB =2 cA Dt (25-19) We recognize that ð@cA =@tÞ þ v: =cA is the substantial derivative of cA. RA ¼ 0. @cA =@t ¼ 0. For constant density and a constant-diffusion coefﬁcient. we obtain which is analogous to equation (16-14) from heat transfer DT k 2 ¼ = T Dt rcP (16-14) or DT ¼ a=2 T Dt where a is the thermal diffusivity. and for binary systems of gases or liquids. Equation (25-20) is analogous to Fourier’s second ‘‘law’’ of heat conduction @T ¼ a=2 T @t (16-18) (iv) Equations (25-17). (25-23) . that is. v ¼ 0. RA ¼ 0. no production term.

r ) þ þ þ ¼ RA r @r r @u @t @z and in spherical coordinates is @cA 1 @ 1 u 1 @NA.x @NA.y @NA. By writing the Laplacian operator.z @cA þ þ þ (25-27) ¼ RA @t @x @y @z in cylindrical coordinates is @cA 1@ 1 @NA.u sin u) þ þ 2 (r 2 NA.3 for energy transfer.z (rNA. for example at t ¼ 0.3 COMMONLY ENCOUNTERED BOUNDARY CONDITIONS A mass-transfer process is fully described by the differential equations of mass transfer only if the initial boundary and initial conditions are speciﬁed. The initial condition in mass transfer processes is the concentration of the diffusing species at the start of the time interval of interest expressed in either mass or molar concentration units. or the equation of continuity of A. written in rectangular coordinates is @NA. Initial conditions are associated only with unsteady-state or pseudo-steady-state processes. Fick’s second ‘‘law’’ of diffusion written in rectangular coordinates is 2 @cA @ cA @ 2 cA @ 2 cA ¼ DAB þ þ (25-24) @t @x2 @y2 @z2 in cylindrical coordinates is 2 @cA @ cA 1 @cA 1 @ 2 cA @ 2 cA ¼ DAB þ þ þ 2 @t @r 2 r @r r 2 @u2 @z and in spherical coordinates is @cA 1 @ 1 @ @cA 1 @ 2 cA 2 @cA r sin u ¼ DAB 2 þ 2 þ 2 2 r @r r sin u @u @t @r @u r sin u @f2 (25-25) (25-26) The general differential equation for mass transfer of component A. thus each applies to any orthogonal coordinate system. Typically. cA ¼ cAo ðmolar unitsÞ at t ¼ 0. The initial and boundary conditions used for mass transfer are very similar to those used in Section 16.f (NA. .u @NA. initial and boundary conditions are used to specify limits of integration or to determine integration constants associated with the mathematical solution of the differential equations for mass transfer. The reader may wish to refer to that section for further discussion of initial and boundary conditions. the transformation of the equation to the desired coordinate system is accomplished. =2. in the appropriate form. rA ¼ rAo ðmass unitsÞ or may be more complex if the initial concentration distribution within the control volume for diffusion is speciﬁed.r ) þ ¼ RA r @r r sin u @u r sin u @f @t (25-28) (25-29) 25. The concentration may be simply equal to a constant.438 Chapter 25 Differential Equations of Mass Transfer Each of the equations (25-15) through (25-23) has been written in vector form.

600 76. Values of H in pressure units for selected gaseous solutes dissolved in aqueous solution are listed in Table 25.800 35. Surface concentration can assume a variety of units. for a gas mixture in contact with a pure volatile liquid A or pure volatile solid A. there are two ways to specify the concentration at the gas–liquid interface. c A so that cAs ¼ c A . (1) The concentration of the transferring species A at a boundary surface is speciﬁed. so that pAs ¼ PA . and PAs is the partial pressure of species A in the gas.600 45.800 58. A similar equation may also be used to determine the boundary conditions at a gas–solid interface cA. mass concentration rAs . PA.800 50. For a contacting gas and liquid where transferring species A is present in both phases. the partial pressure of species A in the gas at the surface is saturation vapor pressure. When the boundary surface is deﬁned by a pure component in one phase and a mixture in the second phase. liquid mole fraction xAs. First.500 66. for solutions where species A is only weakly soluble in the liquid.600 . molar concentration cAs.300 25.000 56. PA is the vapor pressure of species A evaluated at the temperature of the liquid.200 42.500 37. then the boundary condition at the gas–liquid surface is deﬁned for an ideal liquid mixture by Raoult’s law pAs ¼ xA PA where xA is the mole fraction in the liquid.1.700 52.solid ¼ S pA Table 25. Henry’s law may be used to relate the mole fraction of A in the liquid to the partial pressure of A in the gas p A ¼ H xA where coefﬁcient H is known as Henry’s constant. the concentration of species A in the liquid at the surface is the solubility limit of A in the liquid.25. gas mole fraction yAs. for example. then the concentration of transferring species A in the mixture at the interface is usually at thermodynamic saturation conditions.1 Henry’s constant for various gases in aqueous solutions (H in bars) T (K) 273 280 290 300 310 320 NH3 Cl2 H2S SO2 CO2 CH4 O2 H2 21 23 26 30 265 365 480 615 755 860 260 335 450 570 700 835 165 210 315 440 600 800 710 960 1300 1730 2175 2650 22.500 30. For a liquid mixture in contact with a pure solid A.800 27. if both of the species in the liquid phase are volatile.3 Commonly Encountered Boundary Conditions 439 Four types of boundary conditions are commonly encountered in mass transfer. The partial pressure of species A at the interface is related to surface mole fraction yAs by Dalton’s law pAs yAs ¼ P or to surface concentration cAs by the Ideal Gas law pAs cAs ¼ RT Second.000 78.500 56.000 61. Speciﬁcally.500 71. etc.

consider the generic chemical reaction at the boundary surface A þ 2B ! 3C. M. or at the centerline of symmetry of the control volume. and product C diffuses away from the surface. (3) The ﬂux of the transferring species is zero at a boundary or at a centerline of symmetry.2 Solubility constants for selected gas–solid combinations (1 bar ¼ 105 Pa) Gas Solid T (K) O2 N2 CO2 He H2 Natural rubber Natural rubber Natural rubber Silicon Ni 298 298 298 293 358 S ¼ cA. (4) The convective mass transfer ﬂux at the boundary surface is speciﬁed.3. First. There are three common situations. Second. if component A is consumed by a ﬁrst-order on a surface at z = 0. which in turn sets the ﬂux at the surface. This situation can arise at an impermeable boundary. For example. then ¼ kc cAs N A z¼0 where ks is a surface reaction rate constant with units of m/s. where the net ﬂux is equal to zero.solid /PA (kg mol/m3 bar) 3:12 103 1:56 103 40:15 103 0:45 103 9:01 103 where cA. the ﬂux of one species may be related to the ﬂux of another species by chemical reaction stoichiometry. also known as the solubility constant. the convective mass transfer ﬂux across the ﬂuid boundary layer is ¼ kc ðcAs cA1 Þ N A z¼0 where cA1 is the bulk concentration of A the ﬂowing ﬂuid. (2) A reacting surface boundary is speciﬁed. the reaction may be so rapid that cAs ¼ 0 if species A is the limiting reagent in the chemical reaction.440 Chapter 25 Differential Equations of Mass Transfer Table 25. Values of S for several gas–solid pairs reported by Barrer1 are listed in Table 25. Macmillan Press. Third. where reactants A and B diffuse to the surface. cAs is the surface concentration of A at z = 0 and kc is the convection mass-transfer coefﬁcient deﬁned in Section 24. For example. The ﬂuxes for A and B move in the opposite direction to the ﬂux for C. the ﬂux NA is related to the ﬂux of the other species by NB ¼ þ2 NA or NC ¼ 3 NA. For example. a ﬁnite rate of chemical reaction might exist at the surface. solid is the molar concentration of A within the solid at the interface in units of kg mol/m3 and pA is the partial pressure of gas phase species A over the solid in units of Pa. The partition coefﬁcient S. Diffusion In and Through Solids. for a one-dimensional ﬂux along z @cA @cA ¼ DAB ¼0 or ¼0 NA z¼0 @z z¼0 @z z¼0 where the impermeable boundary or the centerline of symmetry is located at z = 0. 1941. Consequently. 1 R. In either case. New York. the ﬂux can be deﬁned by convection. and the positive z direction is opposite to the direction of ﬂux of A along z. When a ﬂuid ﬂows over the boundary. at some surface located at z = 0.2. . all dealing with heterogeneous surface reactions. has units of kg mol/m3 : Pa. Barrer.

25. make a ‘‘list of nomenclature’’ and update the list as you add more terms to the model development. In the ﬁrst approach. In general. = NA ¼ 1 @(r2 NAr ) r2 @r In the second approach. most molecular diffusion problems involve working through the following ﬁve steps: Draw a picture of the physical system. z). For example: @cA ¼ 0. then yA (NAz þ NBz ) ¼ 0. @t (b) If no chemical reaction occurs uniformly within the control volume for diffusion. or spherical (r.4 STEPS FOR MODELING PROCESSES INVOLVING MOLECULAR DIFFUSION Processes involving molecular diffusion can be modeled by the appropriate simpliﬁcations to Fick’s equation and the general differential equation for mass transfer. making use of Fick’s law and the general differential equation for mass transfer. Fick’s law is simpliﬁed by establishing the relationship between the ﬂuxes in the bulk-contribution term. y. Step 2: Make a ‘‘list of assumptions’’ based on your consideration of the physical system. u. then RA ¼ 0. Next. Decide where the source and the sink of mass transfer are located. = NA ¼ @NAz @z by cylindrical geometry in the r and z directions. For example. perform a ‘‘shell balance’’ for the component of interest on a differential volume element of the process. cylindrical (r. including the system boundaries. u). simply reduce or eliminate the terms that do not apply to the physical system.4 Steps for Modeling Processes Involving Molecular Diffusion 441 25. Two approaches may be used to simplify the general differential equation for mass transfer. then (c) If the molecular mass transfer process of species A is one-dimensional in the z direction. Both of these approaches are discussed and illustrated in Chapter 26. Then formulate differential material balances to describe the mass transfer within a volume element of the process based on the geometry of the physical system and the assumptions proposed. recall the one-dimensional ﬂux of a binary mixture of components A and B NAz ¼ cDAB dyA þ yA (NAz þ NBz ) dz If NAz ¼ NBz . Label the important features. Step 1: Step 3: Pick the coordinate system that best describes the geometry of the physical system: rectilinear (x. As appropriate. f). (a) If the process is steady state. = NA ¼ @NAz 1 @(rNAr ) þ @z r @r for radial symmetry in spherical coordinates. If yA (NAz þ NBz ) does not equal 0. then NA is always equal to cA Vz and reduces to cA v z for low concentrations of A in the . z.

at z ¼ 0. NAz ¼ k0 cAs . (d) Known ﬂux of species A at a surface or interface.g. NAz jz¼0 ¼ 0 ¼ dcA =dz. consider asymptotic solutions or limiting cases to more difﬁcult problems ﬁrst.442 Chapter 25 Differential Equations of Mass Transfer Differential equation for mass transfer Fick's equation Assumptions Assumptions Simplified differential equation for mass transfer (NA) Fick's equation differential from (NA) Analytical integration Boundary conditions More assumptions Simplified differential equation for mass transfer (cA) Integral form flux (NA) or transfer rate (wA) Analytical integration Integral form flux (NA) or transfer rate (wA) Differentiation Boundary conditions Concentration profile (cA) Figure 25. the ﬂux. If a differential equation for the concentration proﬁle is desired. at z ¼ 0. Step 5: Solve the differential equations resulting from the differential material balances and the boundary/initial conditions described to get the concentration proﬁle. cA ¼ cAo .. For the rapid disappearance of species A at the surface or interface. e. (b) Symmetry condition at a centerline of the control volume for diffusion. (e) Known chemical reaction at a surface or interface. If appropriate. e. Step 4: Recognize and specify the boundary conditions and initial conditions.. or no net diffusive ﬂux of species A at a surface or interface at z ¼ 0.2 Model development pathways for processes involving molecular diffusion. Figure 25. then the simpliﬁed form of Fick’s law must be substituted into the simpliﬁed form of the general differential equation for mass transfer.g..g. NAz jz¼0 ¼ NAo .. or other parameters of engineering interest. cAs ¼ 0. mixture.g. e. The examples cover many of . (c) Convective ﬂux of species A at a surface or interface. The following examples illustrate how physical and chemical processes involving molecular diffusion can be modeled by the appropriate simpliﬁcations of Fick’s equation and the general differential equation for mass transfer. This concentration can be speciﬁed or known by equilibrium relationships such as Henry’s law. where k0 is a ﬁrst-order chemical reaction rate constant.g.. NA ¼ kc (cAs cA1 ).2 illustrates this process. e. For a slower chemical reaction at the surface or interface with ﬁnite cAs at z ¼ 0. For example (a) Known concentration of species A at a surface or interface at z ¼ 0 e.

(1) The reaction occurs only at the surface of growing Si thin ﬁlm. (3) The feed gas provides silane in high excess relative to that consumed by reaction. The general differential equation for mass transfer in terms of rectilinear coordinates is @NAx @NAy @NAz @cA þ þ þ RA ¼ @x @y @z @t . In this context. a thin ﬁlm of solid silicon (Si) serves as a semiconductor. A diffuser provides a quiescent gas space over the growing Si ﬁlm. The silane in the feed gas serves as the source for mass transfer. Furthermore. whereas the feed gas serves as the sink for H2 mass transfer. SiH4 vapor + H2 gas To vacuum Diffuser SiH4 Quiescent gas z=0 H2 Si thin film Heated plate Si(s) + 2 H2 (g) SiH4(g) z=δ Figure 25.3. Chapters 26 and 27 provide analytical solution techniques for steady-state and unsteady-state diffusion processes. In contrast. so that molecular diffusion dominates. (4) The ﬂux of silane is one-dimensional along z. Silicon thin ﬁlms are commonly formed by the chemical vapor deposition. or CVD. We have taken extra time at the beginning of each example to describe the interesting technology behind the process.3 Chemical vapor deposition of silicon hydride. (6) The mass transfer process within the diffusion zone is at steady state. Consider the very simpliﬁed CVD reactor shown in Figure 25. Develop a differential model for this process. The examples emphasize the ﬁrst four steps of model development outlined. including statements of assumptions and boundary conditions. the gas phase over the Si ﬁlm is not mixed. so that RA ¼ 0. Therefore. EXAMPLE 1 Microelectronic devices are fabricated by forming many layers of thin ﬁlms onto a silicon wafer. the formation of H2 at the Si ﬁlm surface serves as the source for H2 mass transfer.25. The assumptions are used to reduce the general forms of the differential equation for mass transfer and Fick’s equation. Co