Structural Acoustics and Vibration: Mechanical Models, Variational Formulations and Discretization

Structural Acoustics and Vibration

Mechanical Models Variational Formulations and Discretization

First edition

Roger Ohayon

Conservatoire National des Arts et Métiers (CNAM) Paris, France

Christian Soize

Office National d'Etudes et de Recherches Aérospatiales (ONERA) Châtillon, France

ACADEMIC PRESS

San Diego  London  Boston

New York  Sydney  Tokyo  Toronto

Table of Contents

Cover image

Title page

Copyright page

Preface

Chapter I: A Strategy for Structural-Acoustic Problems

1 Introduction

2 Structural-Acoustic Master System

3 Concept of Fuzzy Structure and Structural-Acoustic Fuzzy System

4 LF, MF and HF Frequency Ranges

5 Strategy Used for the LF and MF Ranges

6 Excitations and Responses

7 Organization of the Book

Chapter II: Basic Notions on Variational Formulations

1 Introduction

2 Boundary Value Problem

3 Strong Solution of the Boundary Value Problem

4 Variational Formulation and Weak Solution of the Boundary Value Problem

5 Converse

6 Associated Linear Operator Equation

7 Ritz-Galerkin Approximation and Finite Element Method

8 Bibliographical Comments

Chapter III: Linearized Vibrations of Conservative Structures and Structural Modes

1 Introduction

2 Conservative Elastodynamic Boundary Value Problem in a Bounded Medium with Initial Cauchy Conditions

3 Associated Spectral Problem: Eigenfrequencies and Structural Modes

4 Variational Formulation of the Spectral Problem

5 Associated Linear Operators and Algebraic Properties

6 Basic Properties of the Eigenfrequencies and Structural Modes for a Structure Fixed on Γ0

7 Basic Properties of the Eigenfrequencies and Structural Modes for a Free Structure

8 Cases of Structures with Symmetry Properties

9 Finite Element Discretization and Generalized Symmetric Matrix Eigenvalue Problem

10 Dynamic Substructuring Procedures for Calculation of the Structural Modes

Chapter IV: Dissipative Constitutive Equation for the Master Structure

1 Introduction

2 Notation for the Fourier Transform

3 Damping Model with Frequency-Independent Coefficients

4 Model with Frequency-Dependent Coefficients Based on the Linear Theory of Viscoelasticity

5 Summary

Chapter V: Master Structure Frequency Response Function

1 Introduction

2 Equations in the Frequency Domain

3 Variational Formulation

4 Linear Operators and Algebraic Properties

5 Frequency Response Function

6 Finite Element Discretization

7 Boundary Impedance Operator

Chapter VI: Calculation of the Master Structure Frequency Response Function in the LF Range

1 Introduction

2 FRF Model in the LF Range

3 Projection of the FRF on the Elastic Structural Modes of the Associated Conservative System

4 Remark on a Nonviscoelastic Model Diagonalized by the Elastic Structural Modes

5 Introduction of Quasi-Static Correction Terms

6 Frequency-by-Frequency Construction of the FRF

Chapter VII: Calculation of the Master Structure Frequency Response Function in the MF Range

1 Introduction

2 FRF Model in the MF Range

3 Definition of an MF Narrow Band

4 Class of MF Narrow Band Excitation Force Fields

5 Approximation of the FRF on an MF Narrow Band Due to the Frequency-Dependent Damping and Stiffness Operators

6 Analytical Processing of the Short Time Scale in the Frequency Domain and Numerical Processing of the Long Time Scale in the Time Domain

7 Construction of the FRF on an MF broad band

8 Finite Element Discretization

Chapter VIII: Reduced Model in the MF Range

1 Introduction

2 Definition of a Narrow MF Band

3 Energy Properties of a Simple-Linear-Oscillator Response

4 Variational Formulation of the Master Structure in the MF Range

5 Construction of an Appropriate Functional Basis for the Reduced Model in the MF Range

6 Construction of a Reduced Model in the MF Range

7 Finite Element Discretization

8 Construction of the Dominant Eigensubspace Using the Subspace Iteration Method

Chapter IX: Response to Deterministic and Random Excitations

Publisher Summary

1 Introduction

2 Operator-Valued FRF and Associated Linear Filter in the Time Domain

3 LF and MF Deterministic Cases: Time-Periodic Excitation

4 LF and MF Deterministic Cases: Time Square Integrable Excitation

5 LF and MF Random Cases: Time Stationary Stochastic Excitation

6 Random Case: Nonstationary Stochastic Excitation

Chapter X: Linear Acoustic Equations

Publisher Summary

1 Introduction

2 Inviscid Acoustic Fluid

3 Dissipative Acoustic Fluid

Chapter XI: Internal Acoustic Fluid Formulation for the LF and MF Ranges

Publisher Summary

1 Introduction

2 Statement of the Internal Acoustic Problem

3 Boundary Value Problem

4 Variational Formulation and Linear Operator Equation

5 Frequency Response Function

6 Finite Element Discretization

7 Acoustic Modes of the Acoustic Cavity

8 Finite Element Discretization and Generalized Symmetric Matrix Eigenvalue Problem

9 FRF Calculation in the LF Range

10 FRF Calculation in the MF Range

11 Case of a Zero Pressure Condition on Part of the Boundary

12 Particular Case of an Axisymmetrie Acoustic Cavity

13 Response to Deterministic and Random Excitations

Chapter XII: External Acoustic Fluid. Boundary Integral Formulation for the LF and MF Ranges

Publisher Summary

1 Introduction

2 Statement of the External Acoustic Problem

3 Exterior Neumann Problem Related to the Helmholtz Equation

4 Acoustic Impedance Boundary Operator and Radiation Impedance Operator

5 Free-Space Green’s Function and Solution of the InhomogeneousHelmholtz Equation

6 Response to Prescribed Wall Displacement, Acoustic Source Density and Incident Plane Wave

7 Results of Potential Theory: Single- and Double-Layer Potentials

8 Helmholtz Integral Representations and their Normal Derivatives

9 Boundary Integral Equations for the Exterior Neumann Problem Related to the Helmholtz Equation

10 Acoustic Impedance Boundary Operator Construction: Appropriate Symmetric Boundary Integral Method Valid for all Real Values of the Frequency

11 Construction of the Radiation Impedance Operator and Asymptotic Formula for the Radiated Pressure Field

12 Symmetric Boundary Element Method Valid for all Real Values of the Frequency

13 Case of a Free Surface

Chapter XIII: Structural-Acoustic Master System in the LF Range

1 Introduction

2 Statement of the Structural-Acoustic Problem in the Frequency Domain

3 Boundary Value Problem of the Structural-Acoustic Master System

4 Variational Formulation of the Structural-Acoustic Problem

5 Finite Element Discretization

6 Symmetric Reduced Matrix Model in the LF Range

7 FRF Calculation in the LF Range

8 Location of the Resonant Frequencies of the Coupled System

9 Structural-Acoustic Modes of the Master Structure Coupled with an Internal Acoustic Fluid

10 Case of a Master Structure Coupled with an External Acoustic Fluid

11 Structure Coupled with an External and an Internal Acoustic Fluid. Case of a Zero Pressure Condition on Part of the Internal Fluid Boundary

12 Case of an Axisymmetric Structural-Acoustic Master System

13 Response to Deterministic and Random Excitations

Chapter XIV: Structural-Acoustic Master System in the MF Range

1 Introduction

2 Statement of the Structural-Acoustic Problem in the Frequency Domain

3 Boundary Value Problem of the Structural-Acoustic Master System

4 Variational Formulation of the Structural-Acoustic Problem

5 Finite Element Discretization

6 FRF Calculation in the MF Range

7 Case of a Master Structure Coupled with an External Acoustic Fluid

8 Structure Coupled with an External and an Internal Acoustic Fluid. Case of a Zero Pressure Condition on Part of the Internal Fluid Boundary

9 Case of an Axisymmetric Structural-Acoustic Master System

10 Response to Deterministic and Random Excitations

11 Bibliographical Comments

Chapter XV: Fuzzy Structure Theory

1 Introduction

2 Statement of the Problem

3 Fuzzy Structure Theory Stated as an Inverse Problem

4 Random Equation of the Master Structure Coupled With Fuzzy Substructures in Terms of the Displacement Field of the Master Structure

5 Homogeneous Model of a Fuzzy Substructure

6 Construction of a Homogeneous Fuzzy Impedance Law for a Fuzzy Substructure and for a Local Direction

7 Solving Method for the Random Equation of the Master Structure Coupled with Fuzzy Substructures

8 Ritz-Galerkin Approximation and Finite Element Discretization

9 Identification Method for the Parameters of Type I and Type II Fuzzy Impedance Laws

10 Case of a Real Structure Constituted by a Master Structure Coupled with a Very Large Number of Simple Oscillators. Application of the Fuzzy Structure Theory

Appendix: Mathematical Notations

m

2 Vector Product

3 Real and Complex Matrices

4 Linear and Bilinear Forms, Algebraic Duality and Algebraic Transposition on Complex Vector Spaces

5 Linear and Bilinear Forms and Linear Operators on Complex Hilbert Spaces

6 Antilinear and Sesquilinear Forms and Linear Operators on Complex Hilbert Spaces

7 Gradient, Divergence, Curl and Laplacian Operators

8 Fourier Transform of Functions 4

10 Order Symbols o and O

References

Subject Index

Symbol Index

Copyright

ACADEMIC PRESS LIMITED

24–28 Oval Road

LONDON NW1 7DX

U.S. Edition Published by

ACADEMIC PRESS

San Diego, CA 92101

This book is printed on acid free paper

Copyright © 1998 ACADEMIC PRESS LIMITED

All rights reserved

No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical including photocopying, recording, or any information storage and retrieval system without permission in writing from the publisher.

A catalogue record for this book is available from the British Library

ISBN 0–12–524945–4

Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall

Preface

Roger Ohayon; Christian Soize

This book is devoted to mechanical models, variational formulations and discretization for calculating linear vibrations in the frequency domain of complex structures with arbitrary shape, coupled or not with external and internal acoustic fluids at rest. Such coupled systems are encountered in the area of internal and external noise prediction, reduction and control problems. The excitations can arise from different mechanisms such as mechanical forces applied to the structure, internal acoustic sources, external acoustic sources and external incident acoustic plane waves. These excitations can be deterministic or random. We are interested not only in the low-frequency domain for which modal analysis is suitable, but also in the medium-frequency domain for which additional mechanical modeling and appropriate solving methods are necessary. The main objective of the book is to present appropriate theoretical formulations, constructed so as to be directly applicable for developing computer codes for the numerical simulation of complex systems.

Paris, May 1997

Chapter I

A Strategy for Structural-Acoustic Problems

1 Introduction

This book is devoted to mechanical models and appropriate numerical formulations for calculating the vibration of complex structures with arbitrary shape coupled with external and internal acoustic fluids. In practice, such coupled systems are encountered in the area of internal and external noise prediction, reduction and control problems. The external noise problem is also called the acoustic radiation of structures. Excitations can arise from different mechanisms such as mechanical forces applied to the structure, internal acoustic sources, external acoustic sources and external incident acoustic plane waves. These excitations can be deterministic or random. Concerning problems solved by analytical methods, such as rectangular thin plates, circular cylindrical thin shells, with particular acoustic and structural boundary conditions, we refer the reader to Junger and Feit, 1993; Bowman et al., 1969; Fahy, 1987; Jones, 1986.

In Section 2, we define a structural-acoustic master system.

In Section 3, we introduce the concept of fuzzy structure related to the structural complexity and we define a structural-acoustic fuzzy system.

Section 4 is devoted to the definition of the LF (low-frequency), MF (medium-frequency) and HF (high-frequency) ranges for structural-acoustic problems.

In Section 5, we present the theoretical and numerical strategy used for the LF and MF ranges.

In Section 6, we introduce the different types of prescribed deterministic and random excitations for a structural-acoustic system.

Finally, in Section 7, we indicate the general organization of the book.

2 Structural-Acoustic Master System

In general, a complex structure is composed of a main part called the master structure, defined as the primary structure accessible to conventional modeling, and a secondary part called the fuzzy substructure related to the structural complexity and including for example many equipment units attached to the master structure.

An internal (or an external) structural-acoustic master system is defined as a master structure coupled with an internal acoustic fluid (or an external acoustic fluid). The acoustic fluid can be a gas or a liquid. A master structure coupled with internal and external acoustic fluids is called a structural-acoustic master system.

In this book, mechanical modeling of a structural-acoustic master system is based on the use of the linearized equations of structural and fluid mechanics (acoustic fluids) around a state of equilibrium at rest. For a structural-acoustic problem, the fields of interest are the displacement, velocity and acceleration fields of the master structure, the pressure field in the internal fluid and the pressure field in the external fluid (near field and far field). Since the problem is linear, the structural-acoustic master system can be characterized by a Frequency Response Function (FRF) which allows the response to deterministic and random excitations to be calculated.

3 Concept of Fuzzy Structure and Structural-Acoustic Fuzzy System

A fuzzy structure is defined as a master structure coupled with a fuzzy substructure. A fuzzy substructure is defined as a structural complexity consisting of a large number of small secondary dynamical subsystems attached to the master structure. Each secondary subsystem is a discrete or continuous elastodynamic system, whose resonant frequencies lie in the frequency band of analysis considered (for instance, a large number of equipment units together can constitute a fuzzy substructure). More generally, a fuzzy substructure is a part of the structure that is not accessible to conventional modeling because the details of it are unknown, or are known with insufficient accuracy (this explains the choice of the word fuzzy). For this reason a statistical approach is proposed for modeling fuzzy substructures, by introducing a random boundary impedance operator related to the interfaces where the fuzzy substructure is connected to the master structure. This approach, called fuzzy structure theory, was introduced by Soize, 1986. It should be noted that the term fuzzy has nothing to do with the mathematical theory concerning fuzzy sets and fuzzy logic. In the definition introduced in Section 2, if the master structure is replaced by a fuzzy structure, then we obtain a structural-acoustic fuzzy system.

4 LF, MF and HF Frequency Ranges

The different types of dynamical responses of a weakly dissipative complex structure leads to the definition of three frequency ranges of analysis. Fig. 1 is a qualitative diagram showing the frequency response function (FRF) of a component of the displacement in a fixed point of the structure.

Fig. 1 Frequency ranges for a complex structure

The three frequency ranges are defined as follows.

- The low-frequency range (LF) is defined as the modal domain for which the associated conservative system has a small number of modes (low modal density).

- The high-frequency range (HF) is defined as the frequency band for which there is a uniform high modal density. Among the analysis methods for this band, we can mention the wave approach (see for instance Maidanik and Dickey, 1988 and 1994), the global statistical energy approach and the local energy approach. The most popular global statistical energy approach is the Statistical Energy Analysis. For its basic formulation, we refer the reader to Lyon, 1975; Maidanik, 1981; Lyon and DeJong, 1995; Soize, 1995a; Lesueur, 1988; Cremer et al., 1988; Crighton et al., 1992. For extensions to the case of a structure coupled with an external unbounded liquid, see David and Soize, 1994. Concerning the local energy approach, we can mention the power flow analysis based on continuous energy equations (Nefske and Sung, 1988).

- For complex systems, an intermediate frequency range called medium-frequency range (MF) appears. This MF range is defined as the intermediate frequency band for which the modal density exhibits large variations over the band. In addition, if there is a structural complexity related to the presence of a fuzzy substructure, this fuzzy substructure plays an important role in the dynamical behavior of the master structure. The fuzzy substructure induces an apparent strong dissipation in the master structure due to the power flow between the master structure and all the dynamical subsystems constituting the fuzzy substructure. Therefore, the frequency response functions of the master structure coupled with the fuzzy substructure have smooth variations which means that this system is not resonant in the MF range. The analysis is similar for internal structural-acoustic systems (master system or fuzzy system). The presence of an external acoustic fluid does not substantively modify the above qualitative analysis.

This book is devoted to LF and MF analyses for complex structures and complex structural-acoustic systems.

5 Strategy Used for the LF and MF Ranges

5.1 Structural-acoustic master systems in the LF and MF ranges

Concerning the internal structural-acoustic master system (master structure coupled with internal fluids), the LF and MF models are based on the symmetric variational formulation of the local equations including appropriate dissipative terms. The master structure is described by the displacement field and the internal fluids are described by scalar fields. An adapted formulation is presented for the LF range and the MF range. The internal structural-acoustic master system is discretized by the finite element method. The external fluid model is based on a symmetric variational formulation of an appropriate boundary integral representation which is valid for all values of the real frequencies and does not exhibit the nonphysical irregular frequencies (spurious frequencies) encountered in standard boundary integral representation. Discretization is carried out using the finite element method which leads to boundary elements. For the structural-acoustic master system, a linear symmetric matrix system is obtained whose dimension is the number of degrees of freedom of the discretized model of the internal structural-acoustic master system due to the use of a boundary integral formulation for the external acoustic fluid.

LF computation strategy

We construct a modal reduction procedure using the structural modes of the master structure in vacuo and the acoustic modes of the internal acoustic cavities. The reduced symmetric matrix equation obtained is solved frequency by frequency.

MF computation strategy

In the MF range, the dimension of the matrix equation resulting from finite element discretization is large because the mesh size must be adapted with respect to the wavelengths in the system. Consequently, a frequency-by-frequency solution is not practicable. The method proposed consists in solving the finite element matrix equation MF narrow band by MF narrow band.

5.2 Fuzzy structure in the MF range

In this case, the matrix equation obtained is the sum of the finite element matrix systems corresponding to the master structure with a frequency-dependent random symmetric matrix corresponding to the fuzzy substructure. An appropriate process is used for solving the linear random matrix equation.

5.3 Structural-acoustic systems in the MF range

In this case, the matrix equation obtained is the sum of the finite element matrix systems corresponding to the structural-acoustic master system with a frequency-dependent random symmetric matrix corresponding to the fuzzy substructure and, as above, an appropriate solver is used.

6 Excitations and Responses

For deterministic excitations such as mechanical forces applied to the master structure, internal acoustic sources, external acoustic sources and external incident acoustic plane waves, the responses are calculated using the frequency response functions constructed in the LF and MF ranges. Concerning random excitations, two cases are considered. The first case concerns discrete random excitations modeled by vector-valued stochastic processes. This is the case for random mechanical forces or for random aeroacoustic internal sources induced by machines. The second case concerns continuous random excitations corresponding to random wall pressure fields modeled by vector-valued random fields. This is the case for a turbulent boundary layer due to an external flow around a launcher, an aircraft, an automobile, a naval structure, a high-speed train or an external random field due to a spatial distribution of multipole sources such as complex aeroacoustic sources generated by a launcher engine. For time-stationary random excitations, spectral analysis of the linear filtering theory is used and, for nonstationary random excitations, appropriate methods are used. In all these random cases, the frequency response functions are used in the LF and MF ranges.

7 Organization of the Book

This book is organized in three main parts. Chapters III to IX are devoted to vibration of the master structures. Chapters X to XIV deal with vibration of structural-acoustic master systems and finally, in Chapter XV, fuzzy structures are introduced.

Chapter II is devoted to basic notions on variational formulations, the Ritz-Galerkin method and finite element discretization used throughout the book.

In Chapters III, IV and V, we introduce the structural modes, the damping model based on the use of the linear viscoelasticity theory and the frequency response function (FRF) for the master structure.

Chapter VI presents the master structure FRF calculation in the LF range using modal reduction and a frequency-by-frequency construction.

Chapters VII and VIII give the master structure FRF calculation in the MF range.

Chapter VII presents a narrow-frequency-band-by-narrow-frequency-band construction applied to the finite element discretization of the master structure and Chapter VIII introduces a reduced model in the MF range for which a frequency-by-frequency construction can be used.

Chapter IX gives the methods for constructing the response of the master structure submitted to deterministic and stationary or nonstationary random excitations in the LF and MF ranges.

Chapter X is devoted to the derivation of the linear acoustic equations introducing a damping model, a frequency-dependent wall acoustic impedance and appropriate terms for the internal fluid to ensure that the equations are correctly-stated at zero frequency.

Chapter XI introduces the acoustic modes of an internal cavity and presents the construction of the FRF for internal acoustic problems in the LF and MF ranges.

Chapter XII presents a symmetric boundary integral formulation for the external acoustic problem which is valid for all frequency values in the LF and MF ranges.

Chapter XIII describes the construction of the FRF calculation in the LF range for a structural-acoustic master system using a modal reduced symmetric matrix model and a frequency-by-frequency construction. The methods presented in Chapter IX can be applied directly for calculation of the response of a structural-acoustic master system submitted to deterministic or random excitations.

Chapter XIV deals with the construction of the FRF calculation in the MF range for a structural-acoustic master system using a narrow-frequency-band-by-narrow-frequency-band construction applied to the finite element symmetric matrix of the system. The methodology presented in Chapter VIII can be extended for constructing a reduced symmetric matrix model of the structural-acoustic master system in the MF range. The methods presented in Chapter IX can be applied directly for calculation of the response of a structural-acoustic master system submitted to deterministic or random excitations.

Chapter XV introduces the fuzzy structure theory in structural dynamics for the MF range. The basic tools introduced allow structural-acoustic fuzzy systems to be considered, i.e. structural-acoustic master systems coupled with fuzzy substructures corresponding to additional structural complexities.

Mathematical notations are given in the appendix, presenting the main mathematical tools used in the book and particularly elements concerning linear, antilinear, bilinear and sesquilinear forms on Hilbert space and Fourier transforms of functions.

A comprehensive list of references is given and is followed by a Subject Index.

It should be noted that the footnotes concern purely mathematical aspects which can be skipped in a first reading.

References

Bowman JJ, Senior TBA, Uslenghi PLE. In: Amsterdam: North-Holland; 1969. Electromagnetic and Acoustic Scattering by Simple Shapes.

Cremer L, Heckl M, Ungar EE. In: Berlin: Springer-Verlag; 1988. Structure-Born Sound.

Crighton DG, Dowling AP, Ffowcs-Williams JE, Heckl M, Leppington FG. In: Berlin: Springer-Verlag; 1992. Modern Methods in Analytical Acoustics.

David JM, Soize C. Prediction of the high-frequency behavior of coupled fluid-structure systems by the SEA method and applications. In: Crolet JM, Ohayon R, eds. Computational Methods for Fluid-Structure Interaction. Harlow: Longman Scientific and Technical; 1994:55–77.

Fahy F. In: London: Academic Press; 1987. Sound and Structural Vibration.

Jones DS. In: New York: Oxford University Press; 1986. Acoustic and Electromagnetic Waves.

(originally published in 1972, MIT Press, Cambridge) Junger MC, Feit D. In: Woodbury, NY, U.S.A. Acoust. Soc. Am. Publications on Acoustics; 1993. Sound, Structures and Their Interaction.

Lesueur C. In: Paris: Eyrolles; 1988. Rayonnement Acoustique des Structures.

Lyon RH. In: Cambridge, Massachusetts: MIT Press; 1975. Statistical Energy Analysis Systems: Theory and Applications.

Lyon RH, DeJong RG. In: Boston: Butterworth-Heinemann; 1995. Theory and Application of Statistical Energy Analysis.

Maidanik G. Extension and reformulation of statistical energy analysis with use of room acoustics concepts. Journal of Sound and Vibration. 1981;78(3):417–423.

Maidanik G, Dickey J. Response of coupled one-dimensional dynamic systems. Journal of Sound and Vibration. 1988;121:187–194.

Maidanik G, Dickey J. Flexural response matrix for ribbed panels. J. Acoust. Soc. Am. 1994;95(3):1245–1254.

Nefske DJ, Sung SH. Power flow finite element analysis of dynamic systems. ASMENCA, Vol. 3, New York. 1988.

(English edition) Soize C. Probabilistic structural modeling in linear dynamic analysis of complex mechanical systems. I—Theoretical elements. La Recherche Aérospatiale. 1986;5:23–48.

Soize C. Coupling between an undamped linear acoustic fluid and a damped nonlinear structure—Statistical Energy Analysis considerations. J. Acous. Soc. Am. 1995a;98(1):373–385.

Chapter II

Basic Notions on Variational Formulations

1 Introduction

The objective of this chapter is to review the classical tools for the construction of variational formulations of boundary value problems and the corresponding discretization, which will be used throughout the book. In this chapter, these tools will be illustrated using a classical basic problem.

In Sections 2 and 3, we introduce the basic boundary value problem in a bounded medium and its strong solution.

Sections 4 and 5 give the method for constructing a variational formulation of the boundary value problem using the test function (weighted function) method. For that, we define the admissible function space of the problem and we use Green’s formula. We introduce the weak solution of the boundary value problem which corresponds to the solution of the variational formulation. Finally the converse is established.

In Section 6, we introduce the linear operator equation associated with the variational formulation.

Section 7 reviews the Ritz-Galerkin and finite element methods which constitute the basic tools used in this book.

Section 8 is devoted to bibliographical comments.

2 Boundary Value Problem

³ with a smooth boundary Γ = ∂Ω. The external unit normal to Γ is denoted as n (see Fig. 1). We denote by x = (x1, x2, x3) any point in Ω. Let g be a given real-valued field defined on Ω and G be a given real-valued field defined on Γ. For k > 0, we consider the following boundary value problem

   (1)

with the Neumann boundary condition

   (2)

where u is a real-valued field defined on Ω. The Laplacian operator is defined by ∇² = ∇ · ∇ in which ∇ denotes the gradient operator with respect to x = (x1, x2, x3). The normal derivative ∂u/∂n is defined by ∇u · n.

Fig. 1 Geometrical configuration

3 Strong Solution of the Boundary Value Problem

For sufficiently regular functions g and G, the boundary value problem defined by Eqs. (1) and (2) has a unique solution u called the strong solution¹.

4 Variational Formulation and Weak Solution of the Boundary Value Problem

4.1 Variational formulation

We proceed by the test-function method (also called weighted-function method). In a first step, we introduce the admissible function space of the problem constituted by the sufficiently differentiable² real-valued functions δu defined on Ω. In a second step, we consider strong solution u and integrating over the domain Ω yields

   (3)

in which dx = dx1dx2dx3. We have Green’s formula³

   (4)

in which ds is the surface element. Taking into account the Neumann condition defined by Eq. (2) and using Eq. (4), Eq. (3) can be written as

   (5)

The variational formulation of the boundary value problem defined by Eqs. (1) and (2) is then stated as follows. For given g and G, find u such that Eq. (5) is satisfied, i.e.

   (6)

in which a(u, δudefined by

   (7)

and f(δudefined by

   (8)

4.2 Weak solution of the boundary value problem

A weak solution of the boundary value problem defined by satisfying Eq. (6). It should be noted that the strong solution of Eqs. (1) and (2) is a solution of Eq. (6), i.e. also a weak solution of Eqs. (1) and (2).

4.3 Existence and uniqueness of the weak solution of the boundary value problem

For sufficiently regular functions ⁴ g and G.

5 Converse

Let u be the solution of Eq. (6) (weak solution of Eqs. (1) and (2)) and let us assume that u is sufficiently regular⁶. Using Eq. (6) and applying the Green formula defined by Eq. (4) yields

   (9)

To recover Eqs. (1) and (2) from Eq. (9), we proceed in two steps. First, we choose a test-function δu in the set of infinitely differentiable functions vanishing in a neighborhood ⁷ of Γ. Then Eq. (9) yields

   (10)

From Eq. (10), we deduce that ⁸

   (11)

Substituting Eq. (11) into Eq. (9) yields

   (12)

We then deduce that ⁹

   (13)

6 Associated Linear Operator Equation

Bilinear form a(u, δu) given by Eq. (7) defines a linear operator A such that ¹⁰

   (14)

and the element f such that ¹¹

   (15)

Consequently, the linear operator equation corresponding to the variational formulation defined by Eq. (6) is written as ¹²

   (16)

7 Ritz-Galerkin Approximation and Finite Element Method

7.1 Ritz-Galerkin approximation

The Ritz-Galerkin method consists in constructing an approximate solution of the variational formulation defined by of finite dimension N, we consider a complete countable family {uα}αis spanned by {u1, …, uNcan then be written as

   (17)

in which q = (q1, …, qNN-vector of components {qα} called the generalized coordinates. A test function δu can be written as

   (18)

in which δq = (δq1, …, δqNN. The restriction of the variational formulation defined by is obtained by substituting Eqs. (17) and (18) into Eq. (6),

   (19)

From Eq. (19), we deduce the linear matrix equation of dimension N

   (20)

is an (N × N) matrix such that

   (21)

such that

   (22)

Since bilinear form a(u, δu) is