Textbook Computational Fluid Dynamics Incompressible Turbulent Flows 1St Edition Takeo Kajishima Ebook All Chapter PDF

Computational Fluid Dynamics:
Incompressible Turbulent Flows 1st

Edition Takeo Kajishima
Visit to download the full and correct content document:
https://textbookfull.com/product/computational-fluid-dynamics-incompressible-turbule
nt-flows-1st-edition-takeo-kajishima/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...
Computational fluid dynamics for incompressible flow

First Edition Roychowdhury
https://textbookfull.com/product/computational-fluid-dynamics-
for-incompressible-flow-first-edition-roychowdhury/
Computational Methods for Fluid Dynamics Joel H.

Ferziger
https://textbookfull.com/product/computational-methods-for-fluid-
dynamics-joel-h-ferziger/
Computational fluid dynamics in food processing Second

Edition. Edition Sun
https://textbookfull.com/product/computational-fluid-dynamics-in-
food-processing-second-edition-edition-sun/
Computational Fluid Dynamics A Practical Approach 3rd

Edition Jiyuan Tu
https://textbookfull.com/product/computational-fluid-dynamics-a-
practical-approach-3rd-edition-jiyuan-tu/
Uncertainty Quantification in Computational Fluid
Dynamics and Aircraft Engines 1st Edition Francesco
Montomoli
https://textbookfull.com/product/uncertainty-quantification-in-
computational-fluid-dynamics-and-aircraft-engines-1st-edition-
francesco-montomoli/
Computational Fluid Dynamics for Built and Natural

Environments Zhiqiang (John) Zhai
https://textbookfull.com/product/computational-fluid-dynamics-
for-built-and-natural-environments-zhiqiang-john-zhai/
Computational Fluid Dynamics in Food Processing 2nd

Edition Da-Wen Sun (Editor)
https://textbookfull.com/product/computational-fluid-dynamics-in-
food-processing-2nd-edition-da-wen-sun-editor/
Cellular Flows Topological Metamorphoses in Fluid

Mechanics 1st Edition Vladimir Shtern
https://textbookfull.com/product/cellular-flows-topological-
metamorphoses-in-fluid-mechanics-1st-edition-vladimir-shtern/
Non Newtonian Fluid Mechanics and Complex Flows Angiolo

Farina
https://textbookfull.com/product/non-newtonian-fluid-mechanics-
and-complex-flows-angiolo-farina/
Takeo Kajishima · Kunihiko Taira
Computational
Fluid Dynamics
Incompressible Turbulent Flows
Computational Fluid Dynamics
Takeo Kajishima Kunihiko Taira
•
Computational Fluid
Dynamics
Incompressible Turbulent Flows
123
Takeo Kajishima Kunihiko Taira
Department of Mechanical Engineering Florida State University
Osaka University Tallahassee, FL
Osaka USA
Japan
ISBN 978-3-319-45302-6 ISBN 978-3-319-45304-0 (eBook)

DOI 10.1007/978-3-319-45304-0
Library of Congress Control Number: 2016948598
© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made.
Printed on acid-free paper
This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
To our families:
Yoko, Ryohei, and Junpei
Yuki, Kai, and Rio
Preface
The majority of flows encountered in nature and engineering applications are tur-
bulent. Although turbulence has been studied as part of classical mechanics for over
a century, it is still one of the unsolved problems in physics and remains to be an
active area of research. The field of study that uses numerical simulation to examine
fluid dynamics is called Computational Fluids Dynamics (CFD). There is great
demand in utilizing CFD for analyzing problems ranging from turbulent flows
around aircraft and ground vehicles to larger scale problems related to weather
forecasting and environmental assessment. Such demands are very likely to grow in
the coming years as the engineering community at large pursues improvements in
energy efficiency and performance for various fluid-based systems.
Presently, there are several commercial CFD solvers released with turbulence
analysis capability. With software creating beautiful visualizations of turbulent
flows, it may appear that any type of flows can be numerically predicted. While
there may be some truth to such capability, it is still difficult to solve most turbulent
flow problems without relying on companion experiments. That raises a question of
why we still are not able to perfectly predict the behavior of turbulent flows. The
dynamics of turbulent flows obeys the Navier–Stokes equations, upon which CFD
solvers are based. However, turbulence exhibits flow structures over a wide range
of spatial and temporal scales that all interacts amongst them in a complex nonlinear
manner. That means that the spatial grid must be fine enough to resolve the smallest
scales in turbulent flows while ensuring that the computational domain is large
enough to encompass the largest flow structures. Such grid requirement becomes
increasingly costly as we tackle flows at a higher Reynolds number. Despite the
significant improvement in the computational capability with recent
high-performance computers, we still do not expect computers to be able to handle
these large grids for very high Reynolds number flows.
For this particular reason, turbulence is not likely to be completely solved in the
near future. Flow physics taking place at scales below the resolvable scales must be
represented with appropriate models, referred to as turbulence models. Presently,
there is not a universally accepted turbulence model or numerical algorithm that can
yield a solution unaffected by discretizations of the flow field. Thus, CFD should
vii
viii Preface
continue to be an active field of research with efforts focused towards predicting the
essential features of turbulent flows with turbulence models. As such, engineers and
scientists using CFD must understand how the governing equations are numerically
solved. We must also be equipped with the ability to correctly interpret the
numerical solution. With these points in mind, we should construct a necessary and
sufficient computer program appropriate to simulate the fluid flow of interest. For
commercial software, sufficient details on the solver technique should be provided
in the reference manual so that users can determine whether the solver can be
appropriately used for the problem at hand.
This book describes the fundamental numerical methods and approaches used to
perform numerical simulations of turbulent flows. The materials presented herein
are aimed to provide the basis to accurately analyze unsteady turbulent flows. This
textbook is intended for upper level undergraduate and graduate students who are
interested in learning CFD. This book can also serve as a reference when devel-
oping incompressible flow solvers for those already active in CFD research. It is
assumed that readers have some knowledge of fluid mechanics and partial differ-
ential equations. This textbook does not assume the readers to have advanced
knowledge of numerical analysis.
This textbook aims to enable readers to construct his or her own CFD code from
scratch. The present textbook covers the numerical methods required for CFD and
places emphasis on the incompressible flow solver with detailed discussions on
discretization techniques, boundary conditions, and turbulent flow physics. The
introduction to CFD and the governing equations are offered in Chap. 1, followed
by the coverage of basic numerical methods in Chap. 2. Incompressible flow sol-
vers are derived and discussed in detail in Chaps. 3 and 4. We also provide dis-
cussions on the immersed boundary methods in Chap. 5. A brief overview on
turbulent flows is given in Chap. 6 with details needed for analyzing turbulent flows
using Reynolds-Averaged Navier–Stokes equations (RANS) and Large-Eddy
Simulation (LES) provided in Chaps. 7 and 8, respectively. At the end of the
book, an appendix is attached to offer details on the generalized coordinate system,
Fourier analysis, and modal decomposition methods.
A large portion of the present book is based on the material taught over the years
by the first author for the course entitled “Computational Fluid Dynamics and
Turbulent Flows” at Osaka University and his textbook entitled, “Numerical
Simulations of Turbulent Flows” (1st and 2nd editions in Japanese) that has been
available in Japan since 1999. Chapters 1–4 and 6–8 as well as Appendices A and B
in the present book are founded heavily on the Japanese version by Kajishima. The
present textbook enjoys additions of stability analysis (Sect. 2.5), immersed
boundary methods (Chap. 5), and modal decomposition methods (Sect. 6.3.7 and
Appendix C) by Taira based on the courses taught at the Florida State University.
Furthermore, exercises have been added after each chapter to provide supplemental
materials for the readers.
The preparation of this book has benefited greatly from comments, feedback,
and encouragements from Takashi Ohta, Shintaro Takeuchi, Takeshi Omori, Yohei
Morinishi, Shinnosuke Obi, Hiromochi Kobayashi, Tim Colonius, Clarence
Preface ix
Rowley, Steven Brunton, Shervin Bagheri, Toshiyuki Arima, and Yousuff

Hussaini. The stimulating discussions with them on various topics of CFD over the
years have been invaluable in putting together the materials herein. We must also
thank our research group members and students for providing us with detailed
comments on the drafts of this book that helped improve the organization and
correctness of the text. Finally, we would like to acknowledge Michael Luby and
Brian Halm at Springer for working with us patiently and Nobuyuki Miura and
Kaoru Shimada at Yokendo for their support on the earlier Japanese versions.
Osaka, Japan Takeo Kajishima

Tallahassee, Florida, USA Kunihiko Taira
July 2016
Contents
1 Numerical Simulation of Fluid Flows. . . . . . . . . . . . . . . . . . . . . . . . . . 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Overview of Fluid Flow Simulations . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Governing Equations of Fluid Flows . . . . . . . . . . . . . . . . . . . . . . 4
1.3.1 Conservation Laws . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3.2 Closure of the Governing Equations . . . . . . . . . . . . . . . . 6
1.3.3 Divergence and Gradient Forms . . . . . . . . . . . . . . . . . . . 7
1.3.4 Indicial Notation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.3.5 Governing Equations of Incompressible Flow . . . . . . . . . 10
1.3.6 Properties of Partial Differential Equations . . . . . . . . . . . 13
1.4 Grids for Simulating Fluid Flows . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.5 Discretization Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
1.6 Verification and Validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
1.7 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
1.8 Exercises. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2 Finite-Difference Discretization of the Advection-Diffusion
Equation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.2 Advection-Diffusion Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
2.3 Finite-Difference Approximation . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3.1 Taylor Series Expansion . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.2 Polynomial Approximation . . . . . . . . . . . . . . . . . . . . . . . 32
2.3.3 Central Difference at Midpoints . . . . . . . . . . . . . . . . . . . . 35
2.3.4 Compatibility of Finite Differencing . . . . . . . . . . . . . . . . 36
2.3.5 Spatial Resolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
2.3.6 Behavior of Discretization Error . . . . . . . . . . . . . . . . . . . 42
2.4 Time Stepping Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.4.1 Single-Step Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2.4.2 Multi-Step Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
xi
xii Contents
2.5 Stability Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

2.5.1 Stability of Time Stepping Methods . . . . . . . . . . . . . . . . 52
2.5.2 von Neumann Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 54
2.5.3 Stability of the Discrete Advection Equation . . . . . . . . . . 55
2.5.4 Stability of the Discrete Diffusion Equation . . . . . . . . . . 57
2.5.5 Stability of the Discrete Advection-Diffusion
Equation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
2.5.6 Time Step Constraints for Advection and Diffusion . . . . 62
2.5.7 Amplitude and Phase Errors . . . . . . . . . . . . . . . . . . . . . . 64
2.6 Higher-Order Finite Difference . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
2.7 Consistency of Finite-Difference Methods . . . . . . . . . . . . . . . . . . 67
2.8 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
2.9 Exercises. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72
3 Numerical Simulation of Incompressible Flows . . . . . . . . . . . . . . . . . . 73
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
3.2 Time Stepping for Incompressible Flow Solvers. . . . . . . . . . . . . . 73
3.3 Incompressible Flow Solvers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 76
3.3.1 Fractional-Step (Projection) Method . . . . . . . . . . . . . . . . 77
3.3.2 Simplified MAC (SMAC) Method . . . . . . . . . . . . . . . . . 78
3.3.3 Highly Simplified MAC (HSMAC) Method
and Semi-Implicit Method for Pressure Linked
Equation (SIMPLE) . . . . . . . . . . . . . . . . . . . . . . . . . .... 79
3.3.4 Accuracy and Stability of Time Stepping . . . . . . . . .... 80
3.3.5 Summary of Time Stepping for Incompressible
Flow Solvers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82
3.4 Spatial Discretization of Pressure Gradient Term . . . . . . . . . . . . . 87
3.4.1 Pressure Poisson Equation . . . . . . . . . . . . . . . . . . . . . . . . 87
3.4.2 Iterative Method for the Pressure Poisson Equation. . . . . 92
3.4.3 Iterative Method for HSMAC Method . . . . . . . . . . . . . . . 99
3.5 Spatial Discretization of Advection Term . . . . . . . . . . . . . . . . . . . 100
3.5.1 Compatibility and Conservation . . . . . . . . . . . . . . . . . . . 101
3.5.2 Discretization on Nonuniform Grids . . . . . . . . . . . . . . . . 107
3.5.3 Upwinding Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.6 Spatial Discretization of Viscous Term . . . . . . . . . . . . . . . . . . . . . 114
3.7 Summary of the Staggered Grid Solver . . . . . . . . . . . . . . . . . . . . 118
3.8 Boundary and Initial Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . 120
3.8.1 Boundary Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
3.8.2 Solid Wall Boundary Condition . . . . . . . . . . . . . . . . . . . 123
3.8.3 Inflow and Outflow Boundary Conditions . . . . . . . . . . . . 128
3.8.4 Far-Field Boundary Condition . . . . . . . . . . . . . . . . . . . . . 133
3.8.5 Initial Condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
Contents xiii
3.9 High-Order Accurate Spatial Discretization . . . . . . . . . . . . . .... 136

3.9.1 High-Order Accurate Finite Difference . . . . . . . . . . .... 136
3.9.2 Compatibility of High-Order Finite Differencing
of Advective Term . . . . . . . . . . . . . . . . . . . . . . . . . .... 137
3.9.3 Boundary Conditions for High-Order Accurate
Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
3.10 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
3.11 Exercises. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
4 Incompressible Flow Solvers for Generalized Coordinate System . . . 147
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.2 Selection of Basic Variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
4.3 Strong Conservation Form of the Governing Equations . . . . . . . . 150
4.3.1 Strong Conservation Form . . . . . . . . . . . . . . . . . . . . . . . . 150
4.3.2 Mass Conservation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
4.3.3 Momentum Conservation. . . . . . . . . . . . . . . . . . . . . . . . . 151
4.4 Basic Variables and Coordinate System . . . . . . . . . . . . . . . . . . . . 153
4.5 Incompressible Flow Solvers Using Collocated Grids. . . . . . . . . . 157
4.6 Spatial Discretization of Pressure Gradient Term . . . . . . . . . . . . . 160
4.6.1 Pressure Gradient Term . . . . . . . . . . . . . . . . . . . . . . . . . . 160
4.6.2 Pressure Poisson Equation . . . . . . . . . . . . . . . . . . . . . . . . 163
4.6.3 Iterative Solver for the Pressure Poisson Equation. . . . . . 165
4.7 Spatial Discretization of Advection Term . . . . . . . . . . . . . . . . . . . 166
4.7.1 Compatibility and Conservation . . . . . . . . . . . . . . . . . . . 166
4.7.2 Upwinding Schemes . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
4.8 Spatial Discretization of Viscous Term . . . . . . . . . . . . . . . . . . . . . 170
4.9 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 170
4.10 High-Order Accurate Spatial Discretization . . . . . . . . . . . . . . . . . 172
4.11 Evaluation of Coordinate Transform Coefficients . . . . . . . . . . . . . 173
4.12 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
4.13 Exercises. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177
5 Immersed Boundary Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5.2 Continuous Forcing Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
5.2.1 Discrete Delta Functions . . . . . . . . . . . . . . . . . . . . . . . . . 182
5.2.2 Original Immersed Boundary Method . . . . . . . . . . . . . . . 188
5.2.3 Immersed Boundary Projection Method. . . . . . . . . . . . . . 190
5.3 Discrete Forcing Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193
5.3.1 Direct Forcing Method . . . . . . . . . . . . . . . . . . . . . . . . . . 193
5.3.2 Consistent Direct Forcing Method . . . . . . . . . . . . . . . . . . 194
5.3.3 Cut-Cell Immersed Boundary Method . . . . . . . . . . . . . . . 197
xiv Contents
5.4 Applications of Immersed Boundary Methods . . . . . . . . . . . . . . . 198

5.4.1 Flow Around a Circular Cylinder . . . . . . . . . . . . . . . . . . 199
5.4.2 Turbulent Flow Through a Nuclear Rod Bundle . . . . . . . 199
5.5 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
5.6 Exercises. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 204
6 Numerical Simulation of Turbulent Flows. . . . . . . . . . . . . . . . . . . . . . 207
6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
6.2 Direct Numerical Simulation of Turbulent Flows . . . . . . . . . . . . . 208
6.2.1 Reynolds Number . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 208
6.2.2 Full Turbulence Simulation . . . . . . . . . . . . . . . . . . . . . . . 210
6.2.3 Direct Numerical Simulation of Turbulence . . . . . . . . . . 212
6.2.4 Turbulence Simulation with Low Grid Resolution. . . . . . 213
6.3 Representation of Turbulent Flows . . . . . . . . . . . . . . . . . . . . . . . . 218
6.3.1 Turbulence Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
6.3.2 Governing Equations for Turbulent Flow . . . . . . . . . . . . 220
6.3.3 Turbulence Modeling Approaches . . . . . . . . . . . . . . . . . . 221
6.3.4 Visualization of Vortical Structures . . . . . . . . . . . . . . . . . 222
6.3.5 Coherent Structure Function . . . . . . . . . . . . . . . . . . . . . . 225
6.3.6 Rotational Invariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
6.3.7 Modal Decomposition of Turbulent Flows . . . . . . . . . . . 227
6.4 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
6.5 Exercises. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
7 Reynolds-Averaged Navier–Stokes Equations . . . . . . . . . . . . . . . . . . . 237
7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
7.2 Reynolds-Averaged Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
7.2.1 Reynolds Average . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
7.2.2 Reynolds Stress Equation . . . . . . . . . . . . . . . . . . . . . . . . 239
7.3 Modeling of Eddy Viscosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
7.4 k-ε Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 246
7.4.1 Treatment of Near-Wall Region . . . . . . . . . . . . . . . . . . . 249
7.4.2 Computational Details of the k-ε Model . . . . . . . . . . . . . 252
7.4.3 Features and Applications of the k-ε Model . . . . . . . . . . 254
7.5 Other Eddy-Viscosity Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . 256
7.6 Reynolds Stress Equation Model . . . . . . . . . . . . . . . . . . . . . . . . . 259
7.6.1 Basic Form of the Stress Equation . . . . . . . . . . . . . . . . . 259
7.6.2 Features of the Stress Equation Model . . . . . . . . . . . . . . 263
7.7 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
7.8 Exercises. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 267
Contents xv
8 Large-Eddy Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

8.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
8.2 Governing Equations for LES . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
8.2.1 Filtering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 270
8.2.2 Governing Equations for Large-Eddy Simulation . . . . . . 274
8.3 Smagorinsky Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
8.3.1 Local Equilibrium and Eddy-Viscosity Assumptions . . . . 276
8.3.2 Derivation of the Smagorinsky Model . . . . . . . . . . . . . . . 277
8.3.3 Properties of the Smagorinsky Model . . . . . . . . . . . . . . . 278
8.3.4 Modification in the Near-Wall Region . . . . . . . . . . . . . . . 279
8.4 Scale-Similarity Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
8.4.1 Bardina Model. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
8.4.2 Mixed Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 283
8.5 Dynamic Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 284
8.5.1 Dynamic Eddy-Viscosity Model . . . . . . . . . . . . . . . . . . . 285
8.5.2 Extensions of the Dynamic Model . . . . . . . . . . . . . . . . . 288
8.6 Other SGS Eddy-Viscosity Models . . . . . . . . . . . . . . . . . . . . . . . . 289
8.6.1 Structure Function Model . . . . . . . . . . . . . . . . . . . . . . . . 289
8.6.2 Coherent Structure Model . . . . . . . . . . . . . . . . . . . . . . . . 290
8.6.3 One-Equation SGS Model . . . . . . . . . . . . . . . . . . . . . . . . 292
8.7 Numerical Methods for Large-Eddy Simulation . . . . . . . . . . . . . . 294
8.7.1 Computation of SGS Eddy Viscosity . . . . . . . . . . . . . . . 294
8.7.2 Implementation of Filtering . . . . . . . . . . . . . . . . . . . . . . . 297
8.7.3 Boundary and Initial Conditions . . . . . . . . . . . . . . . . . . . 300
8.7.4 Influence of Numerical Accuracy . . . . . . . . . . . . . . . . . . 302
8.8 Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 303
8.9 Exercises. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 305
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 306
Appendix A: Generalized Coordinate System . . . . . . . . . . . . . . . . . . . . . . 309
Appendix B: Fourier Analysis of Flow Fields . . . . . . . . . . . . . . . . . . . . . . 325
Appendix C: Modal Decomposition Methods . . . . . . . . . . . . . . . . . . . . . . 339
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353
Chapter 1
Numerical Simulation of Fluid Flows
1.1 Introduction
Numerical simulations, along with experiments and theoretical analysis, are often
used as a tool to support research and development in science and engineering.
The use of simulations has been popularized by the development and wide-spread
availability of computers. Since numerical computations are advantageous to exper-
iments from the aspects of speed, safety, and cost in many cases, their uses have been
widely accepted in the industry. Simulations have also become a valuable tool in
fundamental research due to its ability to analyze complex phenomena that may be
difficult to study with experimental measurements or theoretical analysis. Reflecting
upon these trends, the adjective computational is now widely used to describe sub-
fields that utilize simulation in various disciplines, such as computational physics
and computational chemistry.
The field of study concerned with analyzing various types of fluid flows with
numerical simulations and developing suitable simulation algorithms is known as
computational fluid dynamics (CFD). Applications of CFD can be found in the analy-
sis of the following studies but not limited to
• Flows around aircraft, ships, trains, and automobiles;
• Flows in turbo-machineries;
• Biomedical and biological flows;
• Environmental flows, civil engineering, and architecture;
• Large-scale flows in astrodynamics, weather forecasting, and oceanography.
The flows in these settings usually do not have analytical expressions because of
the complex physics arising from boundary geometry, external forcing, and fluid
properties. In CFD, flow physics is analyzed and predicted by numerically solving
the governing equations and reproducing the flow field with the use of computers.
In general, a fluid can be viewed as a continuum, for which there are established
conservation laws for mass, momentum, and energy. For many fluids that are used in
engineering applications, there are well-accepted constitutive relations. The equation
© Springer International Publishing AG 2017 1

T. Kajishima and K. Taira, Computational Fluid Dynamics,
DOI 10.1007/978-3-319-45304-0_1
2 1 Numerical Simulation of Fluid Flows
of state for gases or equations for phase transformation or chemical reactions are also
included in the system of equations as needed. The objective of flow simulations is
to numerically solve such system of equations with appropriate initial and boundary
conditions to replicate the actual flow. In the simulation, the flow field is represented
by variables such as velocity, pressure, density, and temperature at discrete set of
points. The evolution of these variables over time is tracked to represent the flow
physics.
There is always the question of whether a simulation correctly reproduces the
flow physics, because we are representing the continuum on discrete points. We
can only obtain reliable numerical solutions when the simulation methodology is
validated against experimental measurements or theoretical solutions. The use of such
validated method must be limited to parameters that are within the applicable range.
While there has been numerous achievements with computational fluid dynamics in
the design fields and fundamental research, the development of accurate simulation
methods that are widely applicable and more robust will continue in the future.
1.2 Overview of Fluid Flow Simulations
Numerical simulation of fluid flow follows the steps laid out in Fig. 1.1.
1. Decide to solve the full Navier–Stokes equations, the inviscid approximation,

or any other approximation to reproduce the flow physics of interest. Choose a
turbulence model and a non-Newtonian constitutive equation if necessary. Based
on these choices, the governing partial differential equations to be solved in the
simulation are obtained.
2. Discretize the governing equations with the finite-difference, volume or element
method and choose the appropriate grid (spatial discretization). The correspond-
ing algebraic equations to be solved are derived. We can then decide on the
numerical algorithm to solve these equations and develop a computer program.
In some cases, the program can be written to specifically take advantage of the
characteristics of the available computer hardware.
3. Numerical simulation of fluid flow can output a large number of numerical values
as the solution. Comprehending such solution is difficult with just pure numbers.
Hence, graphs and visualizations with computer graphics and animations are used
to aid the analysis of the simulation results.
Thus, to perform numerical simulations of fluid flows, it is not sufficient to only have
the understanding of fluid mechanics. One must have knowledge of numerical analy-
sis for discretization schemes and numerical algorithms, as well as computer science
for programing and visualization. It goes without saying that the combination of
these fields is necessary for CFD. To ensure that the results obtained from numerical
simulation of fluid flow are reliable, we must also be concerned with verification and
validation. We will discuss this in further details in Sect. 1.6.
1.2 Overview of Fluid Flow Simulations 3
experiments
(dye visualization)
Physical flow phenomenon
Approximation
Physical model
Governing equations
(partial differential equations)
Discretization
Grid generation
Discretized governing equations

(algebraic equations)
Numerical methods
Code development
Simulated flow phenomenon

(data set)
Data visualization/analysis
simulation 3D printing
(vorticity contour)
Fig. 1.1 General process of simulating fluid flows. Inserted visualizations are for unsteady flow
over a pitching plate (Reprinted with permission from [8]; copyright 2014, AIP Publishing LLC)
With the invention of 3D printers, it is possible to print three-dimensional flow

structures that can be held and examined up close for a deeper understanding of
the flow physics. An example of a 3D print of the vortex behind a pitching plate is
illustrated in the bottom of Fig. 1.1 based on the data from direct numerical simulation
[8]. Such novel 3D printed prototypes can expand how we study flow fields and
convey findings to others.
The objective of this book is to explain how one can develop a program for
numerical simulations of incompressible turbulent flows. In particular, we focus on
the treatment of the motion of Newtonian fluid (e.g., air and water) that is described
by the Navier–Stokes equations. We also discuss methodologies to simulate turbulent
flow with high-order accurate methods.
1.3 Governing Equations of Fluid Flows
In this section, we present the governing equations of fluid flows. There are two
formulations one can use to describe a flow field: the Eulerian and Lagrangian rep-
resentations. The Eulerian representation describes the flow field with functions of
space and time. The Lagrangian representation on the other hand describes the flow
field following individual fluid elements in the flow. In this book, we discretize the
governing equations on a grid which is based on the Eulerian formulation. For that
reason, the discussions herein are based on the Eulerian representation of the flow
field.
For details on the two representations of a flow field and the derivation of the
governing equations starting from vector and tensor analysis, we list Currie [4],
Panton [15], and Aris [1] as references.
1.3.1 Conservation Laws
The governing equations for fluid flows consist of the conservation laws for mass,
momentum, and energy.
First, let us consider mass conservation. Denoting the density (mass per unit
volume) of the fluid by ρ, we perform a budget analysis for a control volume. We let
the volume and surface of the control volume be V and S, respectively, with the unit
normal vector n on the surface (directed outward) and the flow velocity u, as shown
in Fig. 1.2. The time rate of change of the mass within the control volume consists of
the mass flux going in and out of the volume (ρu) · n through the surface assuming
there is no mass source or sink:

∂ρ
dV = − (ρu) · ndS. (1.1)
V ∂t S
1.3 Governing Equations of Fluid Flows 5
Fig. 1.2 Control volume for S n

conservation laws u
Here, mass influx is negative (ρu · n < 0) and efflux is positive (ρu · n > 0).
Using Gauss’ theorem, the above equation can be written only with a volume
integral
∂ρ
+ ∇ · (ρu) dV = 0. (1.2)
V ∂t
Equations (1.1) and (1.2) are integral representations of mass conservation. Since
Eq. (1.2) must hold for any arbitrary control volume, the integrand should be zero.
Thus, we have
∂ρ
+ ∇ · (ρu) = 0, (1.3)
∂t
which is the differential form of mass conservation.
Following similar control volume analysis, the equations for momentum and
energy conservation can be derived. We can represent these mass, momentum, and
energy conservation laws all together in a single equation using tensors. The integral
representation of the three conservation laws for a control volume V becomes

∂Λ
dV = − Π · ndS + Γ dV (1.4)
∂t
V S
V
∂Λ
+ ∇ · Π − Γ dV = 0 (1.5)
V ∂t
with the corresponding differential form being
∂Λ
+∇ ·Π = Γ. (1.6)
∂t
Here, the vector Λ denotes the conserved quantities (per unit volume)
⎡ ⎤
ρ
Λ = ⎣ ρu ⎦ , (1.7)
ρE
where E is the total energy per unit mass, which is comprised of the internal energy
(per unit mass) e and the kinetic energy k
E = e + k. (1.8)
The kinetic energy is defined as

|u|2
k= . (1.9)
2
The term Π is the flux of Λ
⎡ ⎤
ρu
Π=⎣ ρuu − T ⎦. (1.10)
ρE u − T · u + q
The term ρu can be recognized as the momentum per unit volume in the second row
of Eq. (1.7) and also as the mass flux vector for the first row of Eq. (1.10). The term
ρuu − T is the momentum flux tensor in the momentum equation which consists
of ρuu that describes the flux of momentum ρu moving at a velocity u and T that
represents the momentum exchange due to the stress at the surface of the control
volume. In the energy conservation equation, ρE u is the flux of energy, T · u is the
work performed by stress, and q is the heat flux. By taking an inner product of Π
and the unit normal vector n, we obtain the physical quantity that passes through the
control surface per unit time and area. Since n represents the outward unit normal
vector on a control surface S in Eq. (1.4), positive and negative Π · n correspond to
efflux and influx, respectively, from the control surface.
In what follows, we assume that there is no sink, source, or heat generation within
the volume of interest. If a body force f acts on the fluid, there is production of
momentum and work performed by the force, making the right-hand side of Eq. (1.6)
become ⎡ ⎤
0
Γ = ⎣ ρ f ⎦. (1.11)
ρu · f
Note that Eqs. (1.4) and (1.5) represent the change in conserved variables that is
attributed to the flux across surface of the control volume and the source within the
volume.
1.3.2 Closure of the Governing Equations
In order to solve the governing equations for fluid flow, we need to match the number
of unknowns to the number of equations. This is referred to as the closure of the
system of equations. For the flow equations, we need to express the stress T and heat
flux q in the flux term Π by ρ, u, and E. These relations are called the constitutive
equations.
For a Newtonian fluid, the stress tensor T is expressed as the sum of pressure and
viscous stress in the following manner
1
T = − p I + 2μ D − I∇ · u , (1.12)
3
in which the Stokes relation has been used. Here, I is the identity tensor, p is the
static pressure, μ is the dynamic viscosity, and D is the rate-of-strain tensor given
by
1
D= (∇u)T + ∇u . (1.13)
2
For the heat flux q, we can use the Fourier’s law
q = −k∇T, (1.14)
where T is the absolute temperature and k is the thermal conductivity. Note that μ
and k can be expressed as functions of T for Newtonian fluids (i.e., μ(T ) and k(T )).
We have introduced T and p, which can be related to ρ by the equation of state.
Here, we make an assumption that the fluid is an ideal gas in thermodynamic equi-
librium. The ideal gas law states that
p = ρRT = (γ − 1)ρe, (1.15)
where γ = c p /cv is the specific heat ratio, cv is the heat capacity at constant volume,
c p (= cv + R) is the heat capacity at constant pressure, and R is the gas constant.
The internal energy e is
e = cv T (1.16)
(or de = cv dT ). With the above relations, the system of equations is in closure.

Equations (1.4)–(1.6) expressed only in terms of the unknowns ρ, u, and E. Hence,
the number of equations and the number of unknowns are equated.
1.3.3 Divergence and Gradient Forms
The mass conservation equation is also referred to as the continuity equation and is
∂ρ
+ ∇ · (ρu) = 0 (1.17)
∂t
for flows without sinks or sources. We can consider the time rate of change and
transport of some physical quantity φ by the velocity field u to be decomposed as

∂(ρφ) ∂φ ∂ρ
+ ∇ · (ρuφ) = ρ + u · ∇φ + φ + ∇ · (ρu) , (1.18)
∂t ∂t ∂t
where the second term on the right-hand side becomes zero due to continuity,
Eq. (1.17). Accordingly, we have

∂(ρφ) Dφ ∂φ
+ ∇ · (ρuφ) = ρ =ρ + u · ∇φ , (1.19)
∂t Dt ∂t
where D/Dt is defined as

D ∂
≡ +u·∇ (1.20)
Dt ∂t
and called the material derivative (substantial derivative).
The second term on the left-hand side of Eq. (1.19) represents advection by the
divergence of ρuφ. For that reason, this form is called the divergence form (conserv-
ative form). The second term on the right-hand side is expressed as the inner product
of the velocity u and the gradient ∇φ, and is thus referred to as the gradient form
or advective form1 (non-conservative form). The terminology of conservative and
non-conservative forms is widely used in the CFD literature.
Equation (1.19) is merely a statement of differentiation rule with the continuity
equation incorporated. Therefore, calculations based on either form should be iden-
tical. For correctly discretized equations, the identity should hold, which makes the
use of the term non-conservative form for the right-hand side of Eq. (1.19) some-
what misleading. Depending on the discretization schemes, this relation may not
hold. It would appear more appropriate to use the term non-conservative to describe
incompatible discretization schemes rather than the form of the right-hand side of
Eq. (1.19). The notion of compatible numerical differentiation will be discussed in
detail in Chap. 2.
The conservation of momentum provides us with the equation of motion, which
is
∂(ρu)
+ ∇ · (ρuu − T ) = ρ f (1.21)
∂t
in the divergence form and
Du
ρ =∇ ·T +ρf (1.22)
Dt
in the gradient form. When the constitutive equation (Eq. (1.12)) for a Newtonian
fluid is used for T , the equation of motion is referred to as the Navier–Stokes equation
of motion. Since
Du ∂u
= + u · ∇u (1.23)
Dt ∂t
1 Advective form is often called convective form. We however use the term advective form to be
consistent with the use of the term advection instead of convection (= advection + diffusion) for
preciseness.
corresponds to the acceleration of a fluid element, we can notice that Eq. (1.22)
describes Newton’s second law (mass × acceleration = force) per unit mass.
Taking an inner product of the momentum equation, Eq. (1.22), with the velocity
u, we arrive at the conservation of kinetic energy
Dk
ρ = u · (∇ · T ) + ρu · f . (1.24)
Dt
Subtracting the above equation from the conservation of total energy
DE
ρ = ∇ · (T · u) − ∇ · q + ρu · f , (1.25)
Dt
we obtain the conservation of internal energy
De
ρ = T : (∇u) − ∇ · q, (1.26)
Dt
where T : S denotes the contraction of tensors (i.e., Ti j S ji ). The set of equations
consisting of the mass, momentum, and energy equations for a Newtonian fluid is
called the Navier–Stokes equations.
1.3.4 Indicial Notation
Up to this point, we have used vector notation in the governing equations for fluid
flows. We can also utilize what is called indicial notation to represent the components
for a Cartesian coordinate system, in which we denote the coordinates with x1 = x,
x2 = y, and x3 = z and the corresponding velocity components with u 1 = u, u 2 = u,
and u 3 = w.
We then can express the mass conservation as
∂ρ ∂(ρu j )
+ = 0. (1.27)
∂t ∂x j
The momentum conservation equation can be written in the divergence form

(Eq. (1.21)) as
∂(ρu i ) ∂
+ (ρu i u j − Ti j ) = ρ f i (1.28)
∂t ∂x j
and in gradient form (Eq. (1.22)) as

∂u i ∂u i ∂Ti j
ρ + uj = + ρ fi . (1.29)
∂t ∂x j ∂x j
The conservation of total energy in indicial notation becomes
∂(ρE) ∂
+ (ρEu j − Ti j u i + q j ) = ρu i f i (1.30)
∂t ∂x j
for the divergence form and

∂E ∂E ∂
ρ + uj = (Ti j u i − q j ) + ρu i f i (1.31)
∂t ∂x j ∂x j
for the gradient form (Eq. (1.25)).

The constitutive relations can also be written in indicial notation. The stress tensor
(Eq. (1.12)) is represented as

1 ∂u k ∂u i ∂u j
Ti j = −δi j p + 2μ Di j − δi j , where Di j = + (1.32)
3 ∂xk ∂x j ∂xi
and the Fourier heat conduction law becomes

∂T
q j = −k . (1.33)
∂x j
When the same index appears twice in the same term, summation over that index is
implied (summation convention). That is, in two dimensions, a j b j = 2j=1 a j b j , and

in three dimensions, a j b j = 3j=1 a j b j . When summation is not to be performed, it
will be noted in the text. The symbol for the index can be different in the summation
but results in the same sum (i.e., a j b j = ak bk ). The symbol δi j appears often when
using indicial notation and is called the Kronecker delta, defined as

1 i= j
δi j = (1.34)
0 i = j
This is the component-wise representation of the basis tensor I for the Cartesian
coordinate system. Note that the trace in three dimensions is δkk = 3 (contraction of
δi j ).
1.3.5 Governing Equations of Incompressible Flow
Here, we summarize the governing equations for incompressible flow of a Newtonian

fluid, which is the focus of this book. For incompressible flow, the material derivative
of the density does not change
Dρ ∂ρ
= + u · ∇ρ = 0. (1.35)
Dt ∂t
Note that incompressibility does not necessarily mean that the density ρ is constant.
This continuity equation becomes much more complex for multispecies system with
diffusion, even if the flow can be treated as incompressible [9]. With Eq. (1.35), the
continuity equation, Eq. (1.17), turns into
∇ · u = 0, (1.36)
which is also referred to as the incompressibility constraint or the divergence-free

constraint. This relation implies that the volumetric flux budget is enforced instan-
taneously at each moment in time, since there is no term with time rate of change.
Next, let us consider the momentum equation. For a Newtonian fluid in incom-
pressible flow, the stress tensor is
T = − p I + 2μ D. (1.37)
Thus, the momentum equation for incompressible flow becomes

∂u
ρ + ∇ · (uu) = −∇ p + ∇ · (2μ D) + ρ f . (1.38)
∂t
If we can treat viscosity to be a constant, the above equation can be further simplified
to become
∂u ∇p
+ ∇ · (uu) = − + ν∇ 2 u + f , (1.39)
∂t ρ
where ν = μ/ρ is the kinematic viscosity.

Taking the divergence of Eq. (1.39) with the assumption of ρ and ν being constant
and utilize Eq. (1.36), we arrive at an equation for pressure
∇2 p
= −∇ · ∇ · (uu) + ∇ · f , (1.40)
ρ
where ∇ 2 is the Laplacian operator. Here, the flow field is constrained by incom-
pressibility, Eq. (1.36), at all times and the corresponding pressure field is determined
from the instantaneous flow field. Equation (1.40) is called the pressure Poisson equa-
tion and is an elliptic partial differential equation that is solved as a boundary value
problem. While pressure for compressible flow is provided thermodynamically by
the equation of state, Eq. (1.15), the pressure for incompressible flow is not deter-
mined using thermodynamics. This leads to adopting different numerical methods for
incompressible and compressible flows. While both incompressible and compressible
flows have attributes of wave propagation2 and viscous diffusion, incompressible

flow needs to enforce the divergence-free constraint, Eq. (1.36), which necessitates
an elliptic solver. On the other hand, compressible flow does not need an elliptic solver
since incompressibility need not be satisfied. Different types of partial differential
equations are briefly discussed in Sect. 1.3.6.
The conservation of internal energy for incompressible flow becomes
De
ρ = μ D : D − ∇ · q. (1.41)
Dt
With the use of Eqs. (1.14) and (1.16), the equation for the temperature field reads
DT
ρcv = μ D : D + k∇ 2 T. (1.42)
Dt
The first term on the right-hand side represents the heat generation due to fluid
friction (viscous effect). If the influence of friction on the temperature field is small,
the temperature equation simplifies to
DT
ρcv = k∇ 2 T. (1.43)
Dt
Under this assumption, we can consider the kinetic and internal energies separately
for incompressible flows. Since the conservation of kinetic energy depends only
passively on the conservation of mass and momentum, it is not necessary to explicitly
handle the kinetic energy in the fluid flow analysis. The conservation of internal
energy can be used as a governing equation for temperature. In the discussions to
follow, we do not consider the temperature equation.
Component-wise Representation
As a summary of the above discussion and for reference, let us list the governing
equations for incompressible flow with constant density and viscosity for a Cartesian
coordinate system using the component-wise representation. The continuity equation
(1.36) is
∂u i
=0 (1.44)
∂xi
and the momentum equation (1.39) is
∂u i ∂(u i u j ) ∂u i ∂u i 1 ∂p ∂2ui
+ = + uj =− +ν + fi . (1.45)
∂t ∂x j ∂t ∂x j ρ ∂xi ∂x j ∂x j
2 For incompressible flow, the Mach number M = u/a (the ratio of characteristic velocity u and
sonic speed a) is zero or very small. That means that the acoustic wave propagation is very fast
compared to the hydrodynamic wave propagation. In the limit of M → 0, the acoustic propagation
is considered to take place instantaneously over the whole domain, which leads to the appearance
of ellipticity in the governing equations.
The pressure Poisson equation derived from the above two equations is
1 ∂2 p ∂ 2 (u i u j ) ∂ f i ∂u i ∂u j ∂ fi
=− + =− + . (1.46)
ρ ∂xi ∂xi ∂xi ∂x j ∂xi ∂x j ∂xi ∂xi
Note that Eqs. (1.45) and (1.46) make use of Eq. (1.44).
1.3.6 Properties of Partial Differential Equations
As we have seen above, the conservation laws that describe the flow are represented
by partial differential equations (PDEs). Hence, it is important to understand the
characteristics of the governing PDEs to numerically solve for the flow field. It
is customary to discuss the classification of PDEs and the theory of characteristic
curves, but unless readers go beyond the scope of this book or become involved in
the development of advanced numerical algorithms, there is not a critical need to
dive deeply into the theory of PDEs.
Depending on how information travels, the second-order PDEs are classified into
elliptic, parabolic, and hyperbolic PDEs as shown in Table 1.1. These types are not
influenced by the choice of coordinate systems but can be affected by the location
in a flow field. For example, let us consider flow over a bluff body. Flow away from
the body and outside of the boundary layer (without any vorticity) can be treated as
potential flow which is described by an elliptic PDE. However, in the region right
next to the surface of the body (boundary layer), viscous diffusion becomes the
dominant physics describing the flow which can be captured by a parabolic PDE.
Hence, the classification of PDEs is in general performed locally. We will let other
textbooks [10, 13] on PDEs describe the classification of PDEs, characteristic curves,
and initial/boundary value problems. For the purpose of this book, we will focus on
presenting the PDE types of the governing equations for flow in a brief fashion. The
classification of PDEs is based on the existence of a unidirectional coordinate [16].
Table 1.1 Canonical second-order partial differential equations in fluid mechanics

Type Examples
Elliptic Poisson equation ∇ 2 p = −ρ∇ · ∇ · (uu) + ∇ · f
Laplace equation ∇2φ = 0
∂u
Parabolic Diffusion equation = ν∇ 2 u
∂t
∂T k 2
Heat equation = ∇ T
∂t ρcv
∂2 u
Hyperbolic Wave equation = U 2∇2 u
∂t 2
The temporal axis t is obviously unidirectional (past to future). On the other hand,
the spatial coordinate xi is bidirectional.
For viscous incompressible flows in general, the PDEs for unsteady flow (depen-
dent on t) are parabolic and the PDEs for steady flow (independent of t) are elliptic.
In some steady cases where the influence from downstream can be neglected, the flow
may be solved by marching from upstream to downstream. This is called parabolic
approximation (parabolization). One can also solve for steady flow by leaving the
temporal derivative term and reframing the problem as an unsteady one. Once the
solution is converged to the steady profile after sufficient time advancement, the time
derivative term would vanish and the flow field becomes the solution to the original
elliptic PDE. Such approach could be described as a parabolic technique for solving
an elliptic PDE.
1.4 Grids for Simulating Fluid Flows
For Eulerian methods, variables such as velocity, pressure, and density are determined
at a large number of discrete points to represent the motion of a liquid or a gas as
a continuum. A polygon (in two-dimensional and polyhedron in three-dimensional
space) made from local collection of discrete points (vertices) is called a cell and
the space filled by these cells is referred to as a grid (or mesh). Physical variables of
interest are positioned at various locations on the cells chosen to satisfy particular
numerical properties.
A few representative Eulerian grids are shown in Fig. 1.3. For a Cartesian grid,
the governing equations for the flow takes the simplest form and makes spatial
discretization effortless. However, it is not suitable for discretizing a flow field around
a body of complex geometry, as illustrated in Fig. 1.3a. Even a body with rather simple
geometry, such as a sphere or a circular cylinder, would requires a very fine mesh
near the body boundary to resolve the flow. One solution to this issue is to employ an
immersed boundary method that generates a body without regard to the underlying
grid. Chapter 5 is devoted to the immersed boundary method. Another remedy to
this problem is to use a curvilinear coordinate grid that fits around the boundary, as
shown in Fig. 1.3b. Such curvilinear grid is called a boundary-fitted coordinate grid
or a body-fitted coordinate grid (BFC). A transformation of such BFC in physical
(a) (b) (c)
Fig. 1.3 Grids for fluid flow simulations (two-dimensional). a Cartesian grid. b Curvilinear grid.
c Unstructured grid
1.4 Grids for Simulating Fluid Flows 15
(a) (b)
A D
η
ξ Δη = 1
C D
B A η
y Δξ = 1
Computational domain
x B
ξ
C
Fig. 1.4 Mapping between the physical space and computation space for a boundary-fitted grid. a
Physical space. b Computational space
domain to another domain (computation domain), becomes convenient, especially

if we make the grid size unity in the computational domain (see Fig. 1.4). Grids that
has the same topology as a Cartesian grid is referred to as structured grids, as shown
by Fig. 1.3a, b.
One can also discretize the spatial domain as shown in Fig. 1.3c with cells that
are arranged in an irregular manner to accommodate complex boundary geometries.
Boundary-fitted grids of this type are called unstructured grids. Unstructured grids
often utilize triangles, quadrilaterals, and hexagons in two dimensions and tetrahedra
and hexahedra in three dimensions.
Structured grids are not as versatile as unstructured grids for discretizing com-
plex geometry. However, structured grids are often preferred for computing highly
accurate flow data in fundamental research. Given the same number of grid points,
structured grids allow for more orderly access to memory on computers, compared
to unstructured grids, and achieve higher computational efficiency. The boundary-
fitted coordinate is a generalized curvilinear coordinate which has basis vectors that
are not always orthonormal and these vectors can vary in space. Thus, the govern-
ing equations for flow become complex with the use of the generalized coordinate
system. See Appendix A for details.
Let us present a few types of boundary-fitted grids. Depending on the boundary
shape, different types of boundary-fitted grids is chosen. In Fig. 1.5, we show canoni-
(a) (b) (c)
Fig. 1.5 Boundary-fitted grids. a O-grid. b C-grid. c H-grid

(a) (b) (c)
Fig. 1.6 Examples of hybrid grids. a Overset grid. b Patched grid. c Hybrid grid
cal boundary-fitted grids that are used to discretize the computational domain around
an airfoil. The O-grid places the grid around a body efficiently with low skewness
in general. For bodies with sharp corners (cusp), for example the trailing edge of an
airfoil, there can be highly skewed grids. The C-grid is aligned with the flow around
the body and is able to generate unskewed grids near the trailing edge. While such
arrangement of grids is suitable for viscous flows, there may be unnecessarily large
number of grids downstream of the wing. For airfoil cascade (arrangement of a series
of blades, such as in turbines) or flow over a body in a channel, multi-block approach
is often utilized with H-grids and L-grids (not shown).
It would be ideal to discretize the flow field with only one type of grid. With
the development of grid generation techniques, it has become possible to generate a
mesh of a single type even for somewhat complex geometries. However, generating
high-quality meshes remains a challenge (e.g., mesh with low skewness, mesh with
necessary and sufficient resolution). In many cases, one cannot know where the grid
should be refined a priori.
For flows with bodies of complex geometry, with multiple bodies, or with bodies
that move (relative to each other) or deform, we can consider the use of multiple
types of grids, as illustrated in Fig. 1.6. In most cases, information is transmitted
during calculation from one grid to the other at the overlap or along the interface of
the grids.
Concentrating grids in regions where the flow exhibits changes in its features
is effective for attaining accurate solutions. For example, we know a priori that the
boundary layer near the wall has large velocity gradients. Thus, it would be beneficial
to place a large number of grid points to resolve the flow there. One can also consider
adaptively generating additional grids at regions where the flow shows large gradients
in the variables of interest. For unstructured grids, this would be handled by simply
adding extra vertices in that region. Such approach is referred to as adaptive mesh
refinement and is often used to capture shock waves and flames. In case of flows with
time varying boundary shapes, such as the waves around a ship, a moving adaptive
mesh is used.
In what follows, we assume that the grid has been generated for the computational
domain. For details on grid generation, readers should consult with [6, 19].
Another random document with
no related content on Scribd:
marry? For him it would be all very well, but what about her? Mrs.
Burgess, a good sort and no chatterbox, in whom he had confided,
thought this absence of his in England, ostensibly to see lawyers
might serve to make Daisy reconsider, think what it meant. It was a
question of her position, Mrs. Burgess said; the social barrier; giving
up her children. She’d be a widow with a past one of these days,
draggling about in the suburbs, or more likely, indiscriminate (you
know, she said, what such women get like, with too much paint). But
Peter Walsh pooh-poohed all that. He didn’t mean to die yet.
Anyhow she must settle for herself; judge for herself, he thought,
padding about the room in his socks smoothing out his dress-shirt,
for he might go to Clarissa’s party, or he might go to one of the Halls,
or he might settle in and read an absorbing book written by a man he
used to know at Oxford. And if he did retire, that’s what he’d do—
write books. He would go to Oxford and poke about in the Bodleian.
Vainly the dark, adorably pretty girl ran to the end of the terrace;
vainly waved her hand; vainly cried she didn’t care a straw what
people said. There he was, the man she thought the world of, the
perfect gentleman, the fascinating, the distinguished (and his age
made not the least difference to her), padding about a room in an
hotel in Bloomsbury, shaving, washing, continuing, as he took up
cans, put down razors, to poke about in the Bodleian, and get at the
truth about one or two little matters that interested him. And he would
have a chat with whoever it might be, and so come to disregard
more and more precise hours for lunch, and miss engagements, and
when Daisy asked him, as she would, for a kiss, a scene, fail to
come up to the scratch (though he was genuinely devoted to her)—
in short it might be happier, as Mrs. Burgess said, that she should
forget him, or merely remember him as he was in August 1922, like a
figure standing at the cross roads at dusk, which grows more and
more remote as the dog-cart spins away, carrying her securely
fastened to the back seat, though her arms are outstretched, and as
she sees the figure dwindle and disappear still she cries out how she
would do anything in the world, anything, anything, anything....
He never knew what people thought. It became more and more
difficult for him to concentrate. He became absorbed; he became
busied with his own concerns; now surly, now gay; dependent on
women, absent-minded, moody, less and less able (so he thought as
he shaved) to understand why Clarissa couldn’t simply find them a
lodging and be nice to Daisy; introduce her. And then he could just—
just do what? just haunt and hover (he was at the moment actually
engaged in sorting out various keys, papers), swoop and taste, be
alone, in short, sufficient to himself; and yet nobody of course was
more dependent upon others (he buttoned his waistcoat); it had
been his undoing. He could not keep out of smoking-rooms, liked
colonels, liked golf, liked bridge, and above all women’s society, and
the fineness of their companionship, and their faithfulness and
audacity and greatness in loving which though it had its drawbacks
seemed to him (and the dark, adorably pretty face was on top of the
envelopes) so wholly admirable, so splendid a flower to grow on the
crest of human life, and yet he could not come up to the scratch,
being always apt to see round things (Clarissa had sapped
something in him permanently), and to tire very easily of mute
devotion and to want variety in love, though it would make him
furious if Daisy loved anybody else, furious! for he was jealous,
uncontrollably jealous by temperament. He suffered tortures! But
where was his knife; his watch; his seals, his note-case, and
Clarissa’s letter which he would not read again but liked to think of,
and Daisy’s photograph? And now for dinner.
They were eating.
Sitting at little tables round vases, dressed or not dressed, with their
shawls and bags laid beside them, with their air of false composure,
for they were not used to so many courses at dinner, and
confidence, for they were able to pay for it, and strain, for they had
been running about London all day shopping, sightseeing; and their
natural curiosity, for they looked round and up as the nice-looking
gentleman in horn-rimmed spectacles came in, and their good
nature, for they would have been glad to do any little service, such
as lend a time-table or impart useful information, and their desire,
pulsing in them, tugging at them subterraneously, somehow to
establish connections if it were only a birthplace (Liverpool, for
example) in common or friends of the same name; with their furtive
glances, odd silences, and sudden withdrawals into family jocularity
and isolation; there they sat eating dinner when Mr. Walsh came in
and took his seat at a little table by the curtain.
It was not that he said anything, for being solitary he could only
address himself to the waiter; it was his way of looking at the menu,
of pointing his forefinger to a particular wine, of hitching himself up to
the table, of addressing himself seriously, not gluttonously to dinner,
that won him their respect; which, having to remain unexpressed for
the greater part of the meal, flared up at the table where the Morrises
sat when Mr. Walsh was heard to say at the end of the meal, “Bartlett
pears.” Why he should have spoken so moderately yet firmly, with
the air of a disciplinarian well within his rights which are founded
upon justice, neither young Charles Morris, nor old Charles, neither
Miss Elaine nor Mrs. Morris knew. But when he said, “Bartlett pears,”
sitting alone at his table, they felt that he counted on their support in
some lawful demand; was champion of a cause which immediately
became their own, so that their eyes met his eyes sympathetically,
and when they all reached the smoking-room simultaneously, a little
talk between them became inevitable.
It was not very profound—only to the effect that London was
crowded; had changed in thirty years; that Mr. Morris preferred
Liverpool; that Mrs. Morris had been to the Westminster flower-show,
and that they had all seen the Prince of Wales. Yet, thought Peter
Walsh, no family in the world can compare with the Morrises; none
whatever; and their relations to each other are perfect, and they
don’t care a hang for the upper classes, and they like what they like,
and Elaine is training for the family business, and the boy has won a
scholarship at Leeds, and the old lady (who is about his own age)
has three more children at home; and they have two motor cars, but
Mr. Morris still mends the boots on Sunday: it is superb, it is
absolutely superb, thought Peter Walsh, swaying a little backwards
and forwards with his liqueur glass in his hand among the hairy red
chairs and ash-trays, feeling very well pleased with himself, for the
Morrises liked him. Yes, they liked a man who said, “Bartlett pears.”
They liked him, he felt.
He would go to Clarissa’s party. (The Morrises moved off; but they
would meet again.) He would go to Clarissa’s party, because he
wanted to ask Richard what they were doing in India—the
conservative duffers. And what’s being acted? And music.... Oh yes,
and mere gossip.
For this is the truth about our soul, he thought, our self, who fish-like
inhabits deep seas and plies among obscurities threading her way
between the boles of giant weeds, over sun-flickered spaces and on
and on into gloom, cold, deep, inscrutable; suddenly she shoots to
the surface and sports on the wind-wrinkled waves; that is, has a
positive need to brush, scrape, kindle herself, gossiping. What did
the Government mean—Richard Dalloway would know—to do about
India?
Since it was a very hot night and the paper boys went by with
placards proclaiming in huge red letters that there was a heat-wave,
wicker chairs were placed on the hotel steps and there, sipping,
smoking, detached gentlemen sat. Peter Walsh sat there. One might
fancy that day, the London day, was just beginning. Like a woman
who had slipped off her print dress and white apron to array herself
in blue and pearls, the day changed, put off stuff, took gauze,
changed to evening, and with the same sigh of exhilaration that a
woman breathes, tumbling petticoats on the floor, it too shed dust,
heat, colour; the traffic thinned; motor cars, tinkling, darting,
succeeded the lumber of vans; and here and there among the thick
foliage of the squares an intense light hung. I resign, the evening
seemed to say, as it paled and faded above the battlements and
prominences, moulded, pointed, of hotel, flat, and block of shops, I
fade, she was beginning, I disappear, but London would have none
of it, and rushed her bayonets into the sky, pinioned her, constrained
her to partnership in her revelry.
For the great revolution of Mr. Willett’s summer time had taken place
since Peter Walsh’s last visit to England. The prolonged evening was
new to him. It was inspiriting, rather. For as the young people went
by with their despatch-boxes, awfully glad to be free, proud too,
dumbly, of stepping this famous pavement, joy of a kind, cheap,
tinselly, if you like, but all the same rapture, flushed their faces. They
dressed well too; pink stockings; pretty shoes. They would now have
two hours at the pictures. It sharpened, it refined them, the yellow-
blue evening light; and on the leaves in the square shone lurid, livid
—they looked as if dipped in sea water—the foliage of a submerged
city. He was astonished by the beauty; it was encouraging too, for
where the returned Anglo-Indian sat by rights (he knew crowds of
them) in the Oriental Club biliously summing up the ruin of the world,
here was he, as young as ever; envying young people their summer
time and the rest of it, and more than suspecting from the words of a
girl, from a housemaid’s laughter—intangible things you couldn’t lay
your hands on—that shift in the whole pyramidal accumulation which
in his youth had seemed immovable. On top of them it had pressed;
weighed them down, the women especially, like those flowers
Clarissa’s Aunt Helena used to press between sheets of grey
blotting-paper with Littré’s dictionary on top, sitting under the lamp
after dinner. She was dead now. He had heard of her, from Clarissa,
losing the sight of one eye. It seemed so fitting—one of nature’s
masterpieces—that old Miss Parry should turn to glass. She would
die like some bird in a frost gripping her perch. She belonged to a
different age, but being so entire, so complete, would always stand
up on the horizon, stone-white, eminent, like a lighthouse marking
some past stage on this adventurous, long, long voyage, this
interminable (he felt for a copper to buy a paper and read about
Surrey and Yorkshire—he had held out that copper millions of times.
Surrey was all out once more)—this interminable life. But cricket was
no mere game. Cricket was important. He could never help reading
about cricket. He read the scores in the stop press first, then how it
was a hot day; then about a murder case. Having done things
millions of times enriched them, though it might be said to take the
surface off. The past enriched, and experience, and having cared for
one or two people, and so having acquired the power which the
young lack, of cutting short, doing what one likes, not caring a rap
what people say and coming and going without any very great
expectations (he left his paper on the table and moved off), which
however (and he looked for his hat and coat) was not altogether true
of him, not to-night, for here he was starting to go to a party, at his
age, with the belief upon him that he was about to have an
experience. But what?
Beauty anyhow. Not the crude beauty of the eye. It was not beauty
pure and simple—Bedford Place leading into Russell Square. It was
straightness and emptiness of course; the symmetry of a corridor;
but it was also windows lit up, a piano, a gramophone sounding; a
sense of pleasure-making hidden, but now and again emerging
when, through the uncurtained window, the window left open, one
saw parties sitting over tables, young people slowly circling,
conversations between men and women, maids idly looking out (a
strange comment theirs, when work was done), stockings drying on
top ledges, a parrot, a few plants. Absorbing, mysterious, of infinite
richness, this life. And in the large square where the cabs shot and
swerved so quick, there were loitering couples, dallying, embracing,
shrunk up under the shower of a tree; that was moving; so silent, so
absorbed, that one passed, discreetly, timidly, as if in the presence of
some sacred ceremony to interrupt which would have been impious.
That was interesting. And so on into the flare and glare.
His light overcoat blew open, he stepped with indescribable
idiosyncrasy, leant a little forward, tripped, with his hands behind his
back and his eyes still a little hawklike; he tripped through London,
towards Westminster, observing.
Was everybody dining out, then? Doors were being opened here by
a footman to let issue a high-stepping old dame, in buckled shoes,
with three purple ostrich feathers in her hair. Doors were being
opened for ladies wrapped like mummies in shawls with bright
flowers on them, ladies with bare heads. And in respectable quarters
with stucco pillars through small front gardens lightly swathed with
combs in their hair (having run up to see the children), women came;
men waited for them, with their coats blowing open, and the motor
started. Everybody was going out. What with these doors being
opened, and the descent and the start, it seemed as if the whole of
London were embarking in little boats moored to the bank, tossing
on the waters, as if the whole place were floating off in carnival. And
Whitehall was skated over, silver beaten as it was, skated over by
spiders, and there was a sense of midges round the arc lamps; it
was so hot that people stood about talking. And here in Westminster
was a retired Judge, presumably, sitting four square at his house
door dressed all in white. An Anglo-Indian presumably.
And here a shindy of brawling women, drunken women; here only a
policeman and looming houses, high houses, domed houses,
churches, parliaments, and the hoot of a steamer on the river, a
hollow misty cry. But it was her street, this, Clarissa’s; cabs were
rushing round the corner, like water round the piers of a bridge,
drawn together, it seemed to him because they bore people going to
her party, Clarissa’s party.
The cold stream of visual impressions failed him now as if the eye
were a cup that overflowed and let the rest run down its china walls
unrecorded. The brain must wake now. The body must contract now,
entering the house, the lighted house, where the door stood open,
where the motor cars were standing, and bright women descending:
the soul must brave itself to endure. He opened the big blade of his
pocket-knife.
Lucy came running full tilt downstairs, having just nipped in to the
drawing-room to smooth a cover, to straighten a chair, to pause a
moment and feel whoever came in must think how clean, how bright,
how beautifully cared for, when they saw the beautiful silver, the
brass fire-irons, the new chair-covers, and the curtains of yellow
chintz: she appraised each; heard a roar of voices; people already
coming up from dinner; she must fly!
The Prime Minister was coming, Agnes said: so she had heard them
say in the dining-room, she said, coming in with a tray of glasses.
Did it matter, did it matter in the least, one Prime Minister more or
less? It made no difference at this hour of the night to Mrs. Walker
among the plates, saucepans, cullenders, frying-pans, chicken in
aspic, ice-cream freezers, pared crusts of bread, lemons, soup
tureens, and pudding basins which, however hard they washed up in
the scullery seemed to be all on top of her, on the kitchen table, on
chairs, while the fire blared and roared, the electric lights glared, and
still supper had to be laid. All she felt was, one Prime Minister more
or less made not a scrap of difference to Mrs. Walker.
The ladies were going upstairs already, said Lucy; the ladies were
going up, one by one, Mrs. Dalloway walking last and almost always
sending back some message to the kitchen, “My love to Mrs.
Walker,” that was it one night. Next morning they would go over the
dishes—the soup, the salmon; the salmon, Mrs. Walker knew, as
usual underdone, for she always got nervous about the pudding and
left it to Jenny; so it happened, the salmon was always underdone.
But some lady with fair hair and silver ornaments had said, Lucy
said, about the entrée, was it really made at home? But it was the
salmon that bothered Mrs. Walker, as she spun the plates round and
round, and pulled in dampers and pulled out dampers; and there
came a burst of laughter from the dining-room; a voice speaking;
then another burst of laughter—the gentlemen enjoying themselves
when the ladies had gone. The tokay, said Lucy running in. Mr.
Dalloway had sent for the tokay, from the Emperor’s cellars, the
Imperial Tokay.
It was borne through the kitchen. Over her shoulder Lucy reported
how Miss Elizabeth looked quite lovely; she couldn’t take her eyes
off her; in her pink dress, wearing the necklace Mr. Dalloway had
given her. Jenny must remember the dog, Miss Elizabeth’s fox-
terrier, which, since it bit, had to be shut up and might, Elizabeth
thought, want something. Jenny must remember the dog. But Jenny
was not going upstairs with all those people about. There was a
motor at the door already! There was a ring at the bell—and the
gentlemen still in the dining-room, drinking tokay!
There, they were going upstairs; that was the first to come, and now
they would come faster and faster, so that Mrs. Parkinson (hired for
parties) would leave the hall door ajar, and the hall would be full of
gentlemen waiting (they stood waiting, sleeking down their hair)
while the ladies took their cloaks off in the room along the passage;
where Mrs. Barnet helped them, old Ellen Barnet, who had been with
the family for forty years, and came every summer to help the ladies,
and remembered mothers when they were girls, and though very
unassuming did shake hands; said “milady” very respectfully, yet had
a humorous way with her, looking at the young ladies, and ever so
tactfully helping Lady Lovejoy, who had some trouble with her
underbodice. And they could not help feeling, Lady Lovejoy and Miss
Alice, that some little privilege in the matter of brush and comb, was
awarded them having known Mrs. Barnet—“thirty years, milady,”
Mrs. Barnet supplied her. Young ladies did not use to rouge, said
Lady Lovejoy, when they stayed at Bourton in the old days. And Miss
Alice didn’t need rouge, said Mrs. Barnet, looking at her fondly.
There Mrs. Barnet would sit, in the cloakroom, patting down the furs,
smoothing out the Spanish shawls, tidying the dressing-table, and
knowing perfectly well, in spite of the furs and the embroideries,
which were nice ladies, which were not. The dear old body, said
Lady Lovejoy, mounting the stairs, Clarissa’s old nurse.
And then Lady Lovejoy stiffened. “Lady and Miss Lovejoy,” she said
to Mr. Wilkins (hired for parties). He had an admirable manner, as he
bent and straightened himself, bent and straightened himself and
announced with perfect impartiality “Lady and Miss Lovejoy ... Sir
John and Lady Needham ... Miss Weld ... Mr. Walsh.” His manner
was admirable; his family life must be irreproachable, except that it
seemed impossible that a being with greenish lips and shaven
cheeks could ever have blundered into the nuisance of children.
“How delightful to see you!” said Clarissa. She said it to every one.
How delightful to see you! She was at her worst—effusive, insincere.
It was a great mistake to have come. He should have stayed at
home and read his book, thought Peter Walsh; should have gone to
a music hall; he should have stayed at home, for he knew no one.
Oh dear, it was going to be a failure; a complete failure, Clarissa felt
it in her bones as dear old Lord Lexham stood there apologising for
his wife who had caught cold at the Buckingham Palace garden
party. She could see Peter out of the tail of her eye, criticising her,
there, in that corner. Why, after all, did she do these things? Why
seek pinnacles and stand drenched in fire? Might it consume her
anyhow! Burn her to cinders! Better anything, better brandish one’s
torch and hurl it to earth than taper and dwindle away like some Ellie
Henderson! It was extraordinary how Peter put her into these states
just by coming and standing in a corner. He made her see herself;
exaggerate. It was idiotic. But why did he come, then, merely to
criticise? Why always take, never give? Why not risk one’s one little
point of view? There he was wandering off, and she must speak to
him. But she would not get the chance. Life was that—humiliation,
renunciation. What Lord Lexham was saying was that his wife would
not wear her furs at the garden party because “my dear, you ladies
are all alike”—Lady Lexham being seventy-five at least! It was
delicious, how they petted each other, that old couple. She did like
old Lord Lexham. She did think it mattered, her party, and it made
her feel quite sick to know that it was all going wrong, all falling flat.
Anything, any explosion, any horror was better than people
wandering aimlessly, standing in a bunch at a corner like Ellie
Henderson, not even caring to hold themselves upright.
Gently the yellow curtain with all the birds of Paradise blew out and it
seemed as if there were a flight of wings into the room, right out,
then sucked back. (For the windows were open.) Was it draughty,
Ellie Henderson wondered? She was subject to chills. But it did not
matter that she should come down sneezing to-morrow; it was the
girls with their naked shoulders she thought of, being trained to think
of others by an old father, an invalid, late vicar of Bourton, but he
was dead now; and her chills never went to her chest, never. It was
the girls she thought of, the young girls with their bare shoulders, she
herself having always been a wisp of a creature, with her thin hair
and meagre profile; though now, past fifty, there was beginning to
shine through some mild beam, something purified into distinction by
years of self-abnegation but obscured again, perpetually, by her
distressing gentility, her panic fear, which arose from three hundred
pounds’ income, and her weaponless state (she could not earn a
penny) and it made her timid, and more and more disqualified year
by year to meet well-dressed people who did this sort of thing every
night of the season, merely telling their maids “I’ll wear so and so,”
whereas Ellie Henderson ran out nervously and bought cheap pink
flowers, half a dozen, and then threw a shawl over her old black
dress. For her invitation to Clarissa’s party had come at the last
moment. She was not quite happy about it. She had a sort of feeling
that Clarissa had not meant to ask her this year.
Why should she? There was no reason really, except that they had
always known each other. Indeed, they were cousins. But naturally
they had rather drifted apart, Clarissa being so sought after. It was
an event to her, going to a party. It was quite a treat just to see the
lovely clothes. Wasn’t that Elizabeth, grown up, with her hair done in
the fashionable way, in the pink dress? Yet she could not be more
than seventeen. She was very, very handsome. But girls when they
first came out didn’t seem to wear white as they used. (She must
remember everything to tell Edith.) Girls wore straight frocks,
perfectly tight, with skirts well above the ankles. It was not becoming,
she thought.
So, with her weak eyesight, Ellie Henderson craned rather forward,
and it wasn’t so much she who minded not having any one to talk to
(she hardly knew anybody there), for she felt that they were all such
interesting people to watch; politicians presumably; Richard
Dalloway’s friends; but it was Richard himself who felt that he could
not let the poor creature go on standing there all the evening by
herself.
“Well, Ellie, and how’s the world treating you?” he said in his genial
way, and Ellie Henderson, getting nervous and flushing and feeling
that it was extraordinarily nice of him to come and talk to her, said
that many people really felt the heat more than the cold.
“Yes, they do,” said Richard Dalloway. “Yes.”
But what more did one say?
“Hullo, Richard,” said somebody, taking him by the elbow, and, good
Lord, there was old Peter, old Peter Walsh. He was delighted to see
him—ever so pleased to see him! He hadn’t changed a bit. And off
they went together walking right across the room, giving each other
little pats, as if they hadn’t met for a long time, Ellie Henderson
thought, watching them go, certain she knew that man’s face. A tall
man, middle aged, rather fine eyes, dark, wearing spectacles, with a
look of John Burrows. Edith would be sure to know.
The curtain with its flight of birds of Paradise blew out again. And
Clarissa saw—she saw Ralph Lyon beat it back, and go on talking.
So it wasn’t a failure after all! it was going to be all right now—her
party. It had begun. It had started. But it was still touch and go. She
must stand there for the present. People seemed to come in a rush.
Colonel and Mrs. Garrod ... Mr. Hugh Whitbread ... Mr. Bowley ...
Mrs. Hilbery ... Lady Mary Maddox ... Mr. Quin ... intoned Wilkin. She
had six or seven words with each, and they went on, they went into
the rooms; into something now, not nothing, since Ralph Lyon had
beat back the curtain.
And yet for her own part, it was too much of an effort. She was not
enjoying it. It was too much like being—just anybody, standing there;
anybody could do it; yet this anybody she did a little admire, couldn’t
help feeling that she had, anyhow, made this happen, that it marked
a stage, this post that she felt herself to have become, for oddly
enough she had quite forgotten what she looked like, but felt herself
a stake driven in at the top of her stairs. Every time she gave a party
she had this feeling of being something not herself, and that every
one was unreal in one way; much more real in another. It was, she
thought, partly their clothes, partly being taken out of their ordinary
ways, partly the background, it was possible to say things you
couldn’t say anyhow else, things that needed an effort; possible to
go much deeper. But not for her; not yet anyhow.
“How delightful to see you!” she said. Dear old Sir Harry! He would
know every one.
And what was so odd about it was the sense one had as they came
up the stairs one after another, Mrs. Mount and Celia, Herbert Ainsty,
Mrs. Dakers—oh and Lady Bruton!
“How awfully good of you to come!” she said, and she meant it—it
was odd how standing there one felt them going on, going on, some
quite old, some....
What name? Lady Rosseter? But who on earth was Lady Rosseter?
“Clarissa!” That voice! It was Sally Seton! Sally Seton! after all these
years! She loomed through a mist. For she hadn’t looked like that,
Sally Seton, when Clarissa grasped the hot water can, to think of her
under this roof, under this roof! Not like that!
All on top of each other, embarrassed, laughing, words tumbled out
—passing through London; heard from Clara Haydon; what a chance
of seeing you! So I thrust myself in—without an invitation....
One might put down the hot water can quite composedly. The lustre
had gone out of her. Yet it was extraordinary to see her again, older,
happier, less lovely. They kissed each other, first this cheek then
that, by the drawing-room door, and Clarissa turned, with Sally’s
hand in hers, and saw her rooms full, heard the roar of voices, saw
the candlesticks, the blowing curtains, and the roses which Richard
had given her.
“I have five enormous boys,” said Sally.
She had the simplest egotism, the most open desire to be thought
first always, and Clarissa loved her for being still like that. “I can’t
believe it!” she cried, kindling all over with pleasure at the thought of
the past.
But alas, Wilkins; Wilkins wanted her; Wilkins was emitting in a voice
of commanding authority as if the whole company must be
admonished and the hostess reclaimed from frivolity, one name:
“The Prime Minister,” said Peter Walsh.
The Prime Minister? Was it really? Ellie Henderson marvelled. What
a thing to tell Edith!
One couldn’t laugh at him. He looked so ordinary. You might have
stood him behind a counter and bought biscuits—poor chap, all
rigged up in gold lace. And to be fair, as he went his rounds, first with
Clarissa then with Richard escorting him, he did it very well. He tried
to look somebody. It was amusing to watch. Nobody looked at him.
They just went on talking, yet it was perfectly plain that they all knew,
felt to the marrow of their bones, this majesty passing; this symbol of
what they all stood for, English society. Old Lady Bruton, and she
looked very fine too, very stalwart in her lace, swam up, and they
withdrew into a little room which at once became spied upon,
guarded, and a sort of stir and rustle rippled through every one,
openly: the Prime Minister!
Lord, lord, the snobbery of the English! thought Peter Walsh,
standing in the corner. How they loved dressing up in gold lace and
doing homage! There! That must be, by Jove it was, Hugh
Whitbread, snuffing round the precincts of the great, grown rather
fatter, rather whiter, the admirable Hugh!
He looked always as if he were on duty, thought Peter, a privileged,
but secretive being, hoarding secrets which he would die to defend,
though it was only some little piece of tittle-tattle dropped by a court
footman, which would be in all the papers to-morrow. Such were his
rattles, his baubles, in playing with which he had grown white, come
to the verge of old age, enjoying the respect and affection of all who
had the privilege of knowing this type of the English public school
man. Inevitably one made up things like that about Hugh; that was
his style; the style of those admirable letters which Peter had read
thousands of miles across the sea in the Times, and had thanked
God he was out of that pernicious hubble-bubble if it were only to
hear baboons chatter and coolies beat their wives. An olive-skinned
youth from one of the Universities stood obsequiously by. Him he
would patronise, initiate, teach how to get on. For he liked nothing
better than doing kindnesses, making the hearts of old ladies
palpitate with the joy of being thought of in their age, their affliction,
thinking themselves quite forgotten, yet here was dear Hugh driving
up and spending an hour talking of the past, remembering trifles,
praising the home-made cake, though Hugh might eat cake with a
Duchess any day of his life, and, to look at him, probably did spend a
good deal of time in that agreeable occupation. The All-judging, the
All-merciful, might excuse. Peter Walsh had no mercy. Villains there
must be, and God knows the rascals who get hanged for battering
the brains of a girl out in a train do less harm on the whole than
Hugh Whitbread and his kindness. Look at him now, on tiptoe,
dancing forward, bowing and scraping, as the Prime Minister and
Lady Bruton emerged, intimating for all the world to see that he was
privileged to say something, something private, to Lady Bruton as
she passed. She stopped. She wagged her fine old head. She was
thanking him presumably for some piece of servility. She had her
toadies, minor officials in Government offices who ran about putting
through little jobs on her behalf, in return for which she gave them
luncheon. But she derived from the eighteenth century. She was all
right.
And now Clarissa escorted her Prime Minister down the room,
prancing, sparkling, with the stateliness of her grey hair. She wore
ear-rings, and a silver-green mermaid’s dress. Lolloping on the
waves and braiding her tresses she seemed, having that gift still; to
be; to exist; to sum it all up in the moment as she passed; turned,
caught her scarf in some other woman’s dress, unhitched it, laughed,
all with the most perfect ease and air of a creature floating in its
element. But age had brushed her; even as a mermaid might behold
in her glass the setting sun on some very clear evening over the
waves. There was a breath of tenderness; her severity, her prudery,
her woodenness were all warmed through now, and she had about
her as she said good-bye to the thick gold-laced man who was doing
his best, and good luck to him, to look important, an inexpressible
dignity; an exquisite cordiality; as if she wished the whole world well,
and must now, being on the very verge and rim of things, take her
leave. So she made him think. (But he was not in love.)
Indeed, Clarissa felt, the Prime Minister had been good to come.
And, walking down the room with him, with Sally there and Peter
there and Richard very pleased, with all those people rather inclined,
perhaps, to envy, she had felt that intoxication of the moment, that
dilatation of the nerves of the heart itself till it seemed to quiver,
steeped, upright;—yes, but after all it was what other people felt,
that; for, though she loved it and felt it tingle and sting, still these
semblances, these triumphs (dear old Peter, for example, thinking
her so brilliant), had a hollowness; at arm’s length they were, not in
the heart; and it might be that she was growing old but they satisfied
her no longer as they used; and suddenly, as she saw the Prime
Minister go down the stairs, the gilt rim of the Sir Joshua picture of
the little girl with a muff brought back Kilman with a rush; Kilman her
enemy. That was satisfying; that was real. Ah, how she hated her—
hot, hypocritical, corrupt; with all that power; Elizabeth’s seducer; the
woman who had crept in to steal and defile (Richard would say,
What nonsense!). She hated her: she loved her. It was enemies one
wanted, not friends—not Mrs. Durrant and Clara, Sir William and
Lady Bradshaw, Miss Truelock and Eleanor Gibson (whom she saw
coming upstairs). They must find her if they wanted her. She was for
the party!
There was her old friend Sir Harry.
“Dear Sir Harry!” she said, going up to the fine old fellow who had
produced more bad pictures than any other two Academicians in the
whole of St. John’s Wood (they were always of cattle, standing in
sunset pools absorbing moisture, or signifying, for he had a certain
range of gesture, by the raising of one foreleg and the toss of the
antlers, “the Approach of the Stranger”—all his activities, dining out,
racing, were founded on cattle standing absorbing moisture in sunset
pools).
“What are you laughing at?” she asked him. For Willie Titcomb and
Sir Harry and Herbert Ainsty were all laughing. But no. Sir Harry
could not tell Clarissa Dalloway (much though he liked her; of her
type he thought her perfect, and threatened to paint her) his stories
of the music hall stage. He chaffed her about her party. He missed
his brandy. These circles, he said, were above him. But he liked her;
respected her, in spite of her damnable, difficult upper-class
refinement, which made it impossible to ask Clarissa Dalloway to sit
on his knee. And up came that wandering will-o’-the-wisp, that
vagulous phosphorescence, old Mrs. Hilbery, stretching her hands to
the blaze of his laughter (about the Duke and the Lady), which, as
she heard it across the room, seemed to reassure her on a point
which sometimes bothered her if she woke early in the morning and
did not like to call her maid for a cup of tea; how it is certain we must
die.
“They won’t tell us their stories,” said Clarissa.
“Dear Clarissa!” exclaimed Mrs. Hilbery. She looked to-night, she
said, so like her mother as she first saw her walking in a garden in a
grey hat.
And really Clarissa’s eyes filled with tears. Her mother, walking in a
garden! But alas, she must go.
For there was Professor Brierly, who lectured on Milton, talking to
little Jim Hutton (who was unable even for a party like this to
compass both tie and waistcoat or make his hair lie flat), and even at
this distance they were quarrelling, she could see. For Professor
Brierly was a very queer fish. With all those degrees, honours,
lectureships between him and the scribblers he suspected instantly
an atmosphere not favourable to his queer compound; his prodigious
learning and timidity; his wintry charm without cordiality; his
innocence blent with snobbery; he quivered if made conscious by a
lady’s unkempt hair, a youth’s boots, of an underworld, very
creditable doubtless, of rebels, of ardent young people; of would-be
geniuses, and intimated with a little toss of the head, with a sniff—
Humph!—the value of moderation; of some slight training in the
classics in order to appreciate Milton. Professor Brierly (Clarissa
could see) wasn’t hitting it off with little Jim Hutton (who wore red
socks, his black being at the laundry) about Milton. She interrupted.
She said she loved Bach. So did Hutton. That was the bond between
them, and Hutton (a very bad poet) always felt that Mrs. Dalloway
was far the best of the great ladies who took an interest in art. It was
odd how strict she was. About music she was purely impersonal.
She was rather a prig. But how charming to look at! She made her
house so nice if it weren’t for her Professors. Clarissa had half a
mind to snatch him off and set him down at the piano in the back
room. For he played divinely.
“But the noise!” she said. “The noise!”
“The sign of a successful party.” Nodding urbanely, the Professor
stepped delicately off.
“He knows everything in the whole world about Milton,” said Clarissa.
“Does he indeed?” said Hutton, who would imitate the Professor
throughout Hampstead; the Professor on Milton; the Professor on
moderation; the Professor stepping delicately off.
But she must speak to that couple, said Clarissa, Lord Gayton and
Nancy Blow.
Not that they added perceptibly to the noise of the party. They were
not talking (perceptibly) as they stood side by side by the yellow
curtains. They would soon be off elsewhere, together; and never had
very much to say in any circumstances. They looked; that was all.
That was enough. They looked so clean, so sound, she with an
apricot bloom of powder and paint, but he scrubbed, rinsed, with the
eyes of a bird, so that no ball could pass him or stroke surprise him.
He struck, he leapt, accurately, on the spot. Ponies’ mouths quivered
at the end of his reins. He had his honours, ancestral monuments,
banners hanging in the church at home. He had his duties; his
tenants; a mother and sisters; had been all day at Lords, and that
was what they were talking about—cricket, cousins, the movies—
when Mrs. Dalloway came up. Lord Gayton liked her most awfully.
So did Miss Blow. She had such charming manners.
“It is angelic—it is delicious of you to have come!” she said. She
loved Lords; she loved youth, and Nancy, dressed at enormous
expense by the greatest artists in Paris, stood there looking as if her
body had merely put forth, of its own accord, a green frill.
“I had meant to have dancing,” said Clarissa.
For the young people could not talk. And why should they? Shout,
embrace, swing, be up at dawn; carry sugar to ponies; kiss and
caress the snouts of adorable chows; and then all tingling and
streaming, plunge and swim. But the enormous resources of the
English language, the power it bestows, after all, of communicating
feelings (at their age, she and Peter would have been arguing all the
evening), was not for them. They would solidify young. They would
be good beyond measure to the people on the estate, but alone,
perhaps, rather dull.
“What a pity!” she said. “I had hoped to have dancing.”
It was so extraordinarily nice of them to have come! But talk of
dancing! The rooms were packed.
There was old Aunt Helena in her shawl. Alas, she must leave them
—Lord Gayton and Nancy Blow. There was old Miss Parry, her aunt.
For Miss Helena Parry was not dead: Miss Parry was alive. She was
past eighty. She ascended staircases slowly with a stick. She was
placed in a chair (Richard had seen to it). People who had known
Burma in the ’seventies were always led up to her. Where had Peter
got to? They used to be such friends. For at the mention of India, or
even Ceylon, her eyes (only one was glass) slowly deepened,
became blue, beheld, not human beings—she had no tender
memories, no proud illusions about Viceroys, Generals, Mutinies—it
was orchids she saw, and mountain passes and herself carried on
the backs of coolies in the ’sixties over solitary peaks; or descending
to uproot orchids (startling blossoms, never beheld before) which
she painted in water-colour; an indomitable Englishwoman, fretful if
disturbed by the War, say, which dropped a bomb at her very door,
from her deep meditation over orchids and her own figure journeying
in the ’sixties in India—but here was Peter.
“Come and talk to Aunt Helena about Burma,” said Clarissa.
And yet he had not had a word with her all the evening!
“We will talk later,” said Clarissa, leading him up to Aunt Helena, in
her white shawl, with her stick.
“Peter Walsh,” said Clarissa.
That meant nothing.
Clarissa had asked her. It was tiring; it was noisy; but Clarissa had
asked her. So she had come. It was a pity that they lived in London
—Richard and Clarissa. If only for Clarissa’s health it would have
been better to live in the country. But Clarissa had always been fond
of society.
“He has been in Burma,” said Clarissa.
Ah. She could not resist recalling what Charles Darwin had said
about her little book on the orchids of Burma.
(Clarissa must speak to Lady Bruton.)
No doubt it was forgotten now, her book on the orchids of Burma, but
it went into three editions before 1870, she told Peter. She
remembered him now. He had been at Bourton (and he had left her,
Peter Walsh remembered, without a word in the drawing-room that
night when Clarissa had asked him to come boating).
“Richard so much enjoyed his lunch party,” said Clarissa to Lady
Bruton.
“Richard was the greatest possible help,” Lady Bruton replied. “He
helped me to write a letter. And how are you?”
“Oh, perfectly well!” said Clarissa. (Lady Bruton detested illness in
the wives of politicians.)
“And there’s Peter Walsh!” said Lady Bruton (for she could never
think of anything to say to Clarissa; though she liked her. She had
lots of fine qualities; but they had nothing in common—she and
Clarissa. It might have been better if Richard had married a woman
with less charm, who would have helped him more in his work. He
had lost his chance of the Cabinet). “There’s Peter Walsh!” she said,
shaking hands with that agreeable sinner, that very able fellow who
should have made a name for himself but hadn’t (always in
difficulties with women), and, of course, old Miss Parry. Wonderful
old lady!
Lady Bruton stood by Miss Parry’s chair, a spectral grenadier,
draped in black, inviting Peter Walsh to lunch; cordial; but without
small talk, remembering nothing whatever about the flora or fauna of
India. She had been there, of course; had stayed with three
Viceroys; thought some of the Indian civilians uncommonly fine
fellows; but what a tragedy it was—the state of India! The Prime
Minister had just been telling her (old Miss Parry huddled up in her
shawl, did not care what the Prime Minister had just been telling
her), and Lady Bruton would like to have Peter Walsh’s opinion, he
being fresh from the centre, and she would get Sir Sampson to meet
him, for really it prevented her from sleeping at night, the folly of it,
the wickedness she might say, being a soldier’s daughter. She was
an old woman now, not good for much. But her house, her servants,
her good friend Milly Brush—did he remember her?—were all there
only asking to be used if—if they could be of help, in short. For she
never spoke of England, but this isle of men, this dear, dear land,
was in her blood (without reading Shakespeare), and if ever a

Textbook Computational Fluid Dynamics Incompressible Turbulent Flows 1St Edition Takeo Kajishima Ebook All Chapter PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Textbook Computational Fluid Dynamics Incompressible Turbulent Flows 1St Edition Takeo Kajishima Ebook All Chapter PDF

Uploaded by

Copyright:

Available Formats

Computational Fluid Dynamics:

Incompressible Turbulent Flows 1st

Computational fluid dynamics for incompressible flow

Computational Methods for Fluid Dynamics Joel H.

Computational fluid dynamics in food processing Second

Computational Fluid Dynamics A Practical Approach 3rd

Computational Fluid Dynamics for Built and Natural

Computational Fluid Dynamics in Food Processing 2nd

Cellular Flows Topological Metamorphoses in Fluid

Non Newtonian Fluid Mechanics and Complex Flows Angiolo

ISBN 978-3-319-45302-6 ISBN 978-3-319-45304-0 (eBook)

© Springer International Publishing AG 2017

Printed on acid-free paper

This Springer imprint is published by Springer Nature

Rowley, Steven Brunton, Shervin Bagheri, Toshiyuki Arima, and Yousuff

Osaka, Japan Takeo Kajishima

1 Numerical Simulation of Fluid Flows. . . . . . . . . . . . . . . . . . . . . . . . . . 1

2.5 Stability Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51

3.9 High-Order Accurate Spatial Discretization . . . . . . . . . . . . . .... 136

5.4 Applications of Immersed Boundary Methods . . . . . . . . . . . . . . . 198

8 Large-Eddy Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269

© Springer International Publishing AG 2017 1

1.2 Overview of Fluid Flow Simulations

1. Decide to solve the full Navier–Stokes equations, the inviscid approximation,

Physical ﬂow phenomenon

Discretized governing equations

Simulated ﬂow phenomenon

With the invention of 3D printers, it is possible to print three-dimensional flow

1.3 Governing Equations of Fluid Flows

1.3.1 Conservation Laws

Fig. 1.2 Control volume for S n

with the corresponding differential form being

The kinetic energy is defined as

1.3.2 Closure of the Governing Equations

p = ρRT = (γ − 1)ρe, (1.15)

(or de = cv dT ). With the above relations, the system of equations is in closure.

1.3.3 Divergence and Gradient Forms

where D/Dt is defined as

1.3.4 Indicial Notation

The momentum conservation equation can be written in the divergence form

and in gradient form (Eq. (1.22)) as

The conservation of total energy in indicial notation becomes

for the divergence form and

for the gradient form (Eq. (1.25)).

and the Fourier heat conduction law becomes

1.3.5 Governing Equations of Incompressible Flow

Here, we summarize the governing equations for incompressible flow of a Newtonian

which is also referred to as the incompressibility constraint or the divergence-free

Thus, the momentum equation for incompressible flow becomes

where ν = μ/ρ is the kinematic viscosity.

flows have attributes of wave propagation2 and viscous diffusion, incompressible

and the momentum equation (1.39) is

1.3.6 Properties of Partial Differential Equations

Table 1.1 Canonical second-order partial differential equations in fluid mechanics

1.4 Grids for Simulating Fluid Flows

(a) (b) (c)

domain to another domain (computation domain), becomes convenient, especially

(a) (b) (c)

Fig. 1.5 Boundary-fitted grids. a O-grid. b C-grid. c H-grid

(a) (b) (c)

You might also like