Differential Quadrature and Differential Quadrature Based Element Methods: Theory and Applications

Differential Quadrature and Differential Quadrature Based Element Methods

Theory and Applications

Xinwei Wang

Table of Contents

Cover

Title page

Copyright

Preface

Acknowledgments

Chapter 1: Differential Quadrature Method

Abstract

1.1. Introduction

1.2. Integral quadrature

1.3. Differential quadrature method

1.4. Determination of weighting coefficients

1.5. Explicit formulation of weighting coefficients

1.6. Various grid points

1.7. Error analysis

1.8. Local adaptive differential quadrature method

1.9. Differential quadrature time integration scheme

1.10. Summary

Chapter 2: Differential Quadrature Element Method

Abstract

2.1. Introduction

2.2. Differential quadrature element method

2.3. DQEM with Hermite interpolation

2.4. DQEM with Lagrange interpolation

2.5. Assemblage procedures

2.6. Discussion

2.7. Summary

Chapter 3: Methods of Applying Boundary Conditions

Abstract

3.1. Introduction

3.2. Basic equations of a Bernoulli–Euler beam

3.3. Methods for applying multiple boundary conditions

3.4. Discussion

3.5. Numerical examples

3.6. Summary

Chapter 4: Quadrature Element Method

Abstract

4.1. Introduction

4.2. Quadrature element method

4.3. Quadrature bar element

4.4. Quadrature Timoshenko beam element

4.5. Quadrature plane stress (strain) element

4.6. Quadrature thick plate element

4.7. Quadrature thin beam element

4.8. Quadrature thin rectangular plate element

4.9. Extension to quadrilateral plate element with curved edges

4.10. Discussion

4.11. Summary

Chapter 5: In-plane Stress Analysis

Abstract

5.1. Introduction

5.2. Formulation-I

5.3. Formulation-II

5.4. Results and discussion

5.5. Equivalent boundary conditions

5.6. Summary

Chapter 6: Static Analysis of Thin Plate

Abstract

6.1. Introduction

6.2. Rectangular thin plate under general loading

6.2.3. Equivalent load

6.3. Applications

6.4. Summary

Chapter 7: Linear Buckling Analysis of Thin Plate

Abstract

7.1. Introduction

7.2. Buckling of rectangular thin plate

7.3. Applications

7.4. Summary

Chapter 8: Free Vibration Analysis of Thin Plate

Abstract

8.1. Introduction

8.2. Free vibration of rectangular thin plate

8.3. Examples

8.4. Summary

Chapter 9: Geometric Nonlinear Analysis

Abstract

9.1. Introduction

9.2. Nonlinear stability of thin doubly curved orthotropic shallow shell

9.3. Buckling of inclined circular cylinder-in-cylinder

9.4. Buckling and postbuckling analysis of extensible beam

9.5. Summary

Chapter 10: Elastoplastic Buckling Analysis of Plate

Abstract

10.1. Introduction

10.2. Elastoplastic buckling of thin rectangular plate

10.3. Elastoplastic buckling of thick rectangular plate

10.4. Summary

Chapter 11: Structural Analysis by the QEM

Abstract

11.1. Introduction

11.2. Dynamic analysis of a flexible rod Hit by rigid ball

11.3. Free vibration of thin plates

11.4. Buckling of thin rectangular plate

11.5. Summary

Appendix I

Appendix II

Appendix III

Appendix IV

Appendix V

Appendix VI

Appendix VII

Appendix VIII

Appendix IX

Index

Copyright

Butterworth-Heinemann is an imprint of Elsevier

The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK

225 Wyman Street, Waltham, MA 02451, USA

Copyright © 2015 Elsevier Inc. All rights reserved.

No part of this publication 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. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein).

Notices

Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary.

Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility.

To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein.

British Library Cataloguing-in-Publication Data

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

Library of Congress Cataloging-in-Publication Data

A catalog record for this book is available from the Library of Congress

ISBN: 978-0-12-803081-3

For information on all Butterworth-Heinemann publications visit our website at http://store.elsevier.com/

Typeset by Thomson Digital

Printed and bound in the United States of America

MATLAB® is a trademark of The MathWorks, Inc. and is used with permission. The MathWorks does not warrant the accuracy of the text or exercises in this book. This book’s use or discussion of MATLAB® software or related products does not constitute endorsement or sponsorship by The MathWorks of a particular pedagogical approach or particular use of the MATLAB® software.

Preface

Time is flying. More than 20 years have passed since the differential quadrature method (DQM) was introduced to the author by Professor Charles W. Bert in 1991. The author joined Dr. Bert’s differential quadrature (DQ) research group at the University of Oklahoma, Norman, Oklahoma, USA. A year later, he returned to his home country and became a professor at Nanjing University of Aeronautics and Astronautics, Nanjing, China. He has continued to do research on the engineering applications of the DQM and contributed to the development of the DQM.

The DQM can often yield numerically exact results as to the analytical solutions. Due to its attractive features of rapid convergence, high accuracy, and computational efficiency, the DQM is now a well-known method worldwide. The early developments of the DQM and its applications in general engineering up to the year of 1999 are well documented in a book written by Professor Shu (Chang Shu: Differential Quadrature and Its Application in Engineering, London: Springer-Verlag Limited, 2000). New developments on the DQ method and its applications to structural mechanics have been made since then. Although the progression of the development and application of the DQM in the area of structural mechanics is clear from the past researches, but these have been scattered over many papers. In addition, a variety of different quadrature formulations by varying the degree of the polynomials, treatment of boundary conditions and employing regular or irregular grid points exist in literature. This has often caused confusion for researchers and engineers and led to a difficulty to make a choice of a DQM or DQ-based element method for solving practical problems.

From time to time, the author receives emails domestically and internationally to request papers or ask questions related to the applications of the DQM as well as to the implementation. Therefore, there is a need to write a book to explain these new developments and their applications in detail as well as to provide FORTRAN programs and MATLAB files to overcome the difficulty in the implementation by using the well-developed methods.

The book is primarily written for scientists and engineers who are interested in applying the DQM and DQ-based element methods to problems in the area of structural mechanics, including static, buckling, vibration, and dynamic problems. The book is also suitable for graduate students majoring in engineering. With the help of the useful information provided in the book, they can use these methods effectively to solve a variety of problems in their research as well as in practice. The book can serve as a reference book for researchers in many fields such as biosciences, transport processes, fluid mechanics, static and dynamic structural mechanics, static aeroelasticity, and lubrication mechanics, where quantitative analysis is needed.

This book is focused to explore the applications of the DQ technique to problems in the area of structural mechanics, including static stress analysis, buckling analysis, and vibration and dynamic analysis. Linear, geometrically nonlinear, and material nonlinear problems are involved. Important aspects are emphasized and discussed in detail. The main body of the book contains two segments. One is theory and the other is applications. To ease the implementation effort, FORTRAN programs and MATLAB files are included in the appendices. The outline of this book is as follows.

Chapter 1 presents the basic principle of the DQM. For completeness considerations, the original method to determine the weighting coefficients is also included. The explicit formulas to compute the weighting coefficients are given, including the one based on the polynomials and the one based on the harmonic functions. The latter is called the harmonic differential quadrature method (HDQM). Various grid distributions are summarized. Error analysis of the DQM is briefly discussed. Local adaptive DQM is given. Special attention is paid to the DQ-based time integration scheme, an unconditional stable numerical integration method. Examples are given to demonstrate its capability over the existing time integration schemes.

Chapter 2 presents the basic principle of the differential quadrature element method (DQEM). Two different ways are described to formulate the weighting coefficients of the DQ beam element. One way uses the Hermite interpolation and the other employs the Lagrange interpolation. For DQ plate element, the approach of the mixed Hermite interpolation with Lagrange interpolation can also be used. Assemblage procedures are given and several examples are worked out in detail for illustrations.

Chapter 3 presents various approaches to apply the multiple boundary conditions. Although all methods work equally well in one-dimensional problem; however, some of them have difficulty in applying the multiple boundary conditions at the corners for two-dimensional problems, such as the rectangular plate problems, since these approaches have four degrees of freedom at corners but only three boundary conditions are available. Examples are given to show the importance of choosing an appropriate way to apply the multiple boundary conditions.

Chapter 4 presents the basic principle of the weak-form quadrature element method (QEM), one of the DQ-based element methods. Different from the strong-form DQEM presented in Chapter 2, the formulations of the QEM are essentially similar to the high-order finite element method (FEM) or the time-domain spectral element method (SEM). However, differences do exist and are demonstrated in details.

Chapters 5–11 present various applications of the DQM and the DQ-based element methods. The applications include in-plane stress analysis, static analysis, linear and nonlinear buckling analysis, and free vibration and dynamic analysis. The importance of using an accurate way to apply the multiple boundary conditions and of choosing the right nonuniform grid distribution is demonstrated.

In appendices, several FORTRAN programs and subroutines are provided to save the readers’ programming effort. With simple modifications, these programs can be used to handle different materials, different applied loads as well as different boundary conditions, and to solve a variety of problems in the area of structural mechanics, such as static, buckling, and free vibration analysis. These programs and subroutines have also been converted to the MATLAB files and functions.

Finally, a summary on our research work is given in Appendix XI. Our contribution to the development of the DQM as well as to its applications is briefly described. More results, which are not included in the book, can be found in the listed journal papers.

The book was completed in the State Key Laboratory of Mechanics and Control of Mechanical Structures, Nanjing University of Aeronautics and Astronautics, Nanjing, China.

Xinwei Wang

Nanjing, China

Acknowledgments

It is clearly impossible for me to acknowledge everyone who has made a contribution, in some manner, to publish this book. I owe a major debt to my colleagues and graduate students who made many contributions to the development of the DQM as well as applications of the DQM. They are Professor Renhuai Liu, Yongliang Wang, Beiqing He, Guangming Zhou, Chuwei Zhou, and Xinfeng Wang, Dr. Meilan Tan, Feng Liu, Lifei Gan, Jian Liu, Zhangxian Yuan, and Chunhua Jin, and Mr. Bin Liu, Xudong Shi, Linghai Jiang, Mengsheng Wang, Chengwei Dai, Zhongbin Zhou, Wei Zhang, Chunling Xu, Feng Wang, Guohui Duan, Luyao Ge, and Yu Wang, and Miss Zhe Wu. The materials presented in this book include their contributions.

I am particularly grateful to Professor Charles W. Bert at the University of Oklahoma, USA, who introduced the DQM to me and encouraged me to do research on the DQM continuously. Special thanks go to Dr. Zhangxian Yuan and Mr. Luyao Ge for helping me to convert the FORTRAN programs into MATLAB files, and also to Mr. Luyao Ge and Yu Wang for redrawing some plots.

My sincere thanks and appreciation also go to each of the reviewers for their helpful reviews and valuable comments to improve the book. They include Professor Alfred G. Striz at the University of Oklahoma, USA, Professor Wen Chen at Hohai University, China, and several anonymous reviewers.

The staff at Elsevier have been very helpful and cooperative. I was assisted in the manuscript preparation by Cari Owen, Editorial Project Manager, Science & Technology Books (Engineering). I express my heartfelt thanks to Cari who works with great care and in a friendly and considerate way; to Pauline Wilkinson, the Production Project Manager, who prepared the proofs with great accuracy, and to the other individuals at Elsevier who were involved in the editing and production aspects of the book.

Last but not least, I would like to thank my family for their understanding, encouragement, and support. I devote this book to my wife, Guoying, for her infinite patience and taking most of the household duties, and also to my little Grandson, Chenqi, to ease my (his Grandpa’s) guilt from spending too little time to play with him.

The researches were partially supported by the National Natural Science Foundation of China (10972105), the Aeronautical Science Foundation of China (2004ZB52006), Jiangsu Natural Science Foundation (BK99116), National Doctorial Foundation of China (20020287003), PAPD, and the Research Fund of State Key Laboratory of Mechanics and Control of Mechanical Structures (Nanjing University of Aeronautics and Astronautics) (Grant No. 0214G02).

Chapter 1

Differential Quadrature Method

Abstract

A variety of different quadrature (DQ) formulations exist in literatures. This has often caused confusion for researchers and engineers and led to a difficulty in making a choice of a different quadrature method (DQM) for solving practical problems. This chapter presents the basic principle of the DQM and summarizes various DQ formulations, including the original DQM, the modified DQM, harmonic differential quadrature method (HDQM), local adaptive differential quadrature method (LaDQM), and the DQ-based time integration scheme. Existing explicit formulas to compute the weighting coefficients are given. Various grid distributions are summarized and their discrete error is briefly discussed. Some recommendations are made. Although Grid III is the most widely used grid spacing in literature, however, research shows that Grid V is the most reliable grid spacing and thus recommended, especially for dynamic analysis by using the DQM. The LaDQM is recommended if the number of grid points is large.

Keywords

differential quadrature

weighting coefficients

harmonic differential quadrature

LaDQM

grid spacing

DQ-based time integration

explicit formulas

discrete error

1.1. Introduction

With the advance of computer technology, nowadays numerical simulations play an important role in science and engineering. Various numerical methods have been used in numerical analysis and are regarded as powerful tools for solving partial differential equations (PDEs). To name a few, finite element method (FEM) [1], finite difference method (FDM) [2], finite volume method (FVM) [3], and methods of weighted residuals (MWR) [4,5] such as Galerkin method and collocation method. Among all aforementioned methods, the most widely used methods are FEM and FDM. Perhaps due to its flexibility and ability in dealing with complex geometries and boundary conditions, FEM is more widely used in the area of structural mechanics [1].

It is known that none of the aforementioned numerical methods is versatile and can be used to solve all problems efficiently. Each method has its own merits and limitations. Even the most widely used FEM still suffers from difficulty in analyzing problems when phenomena such as singularities, steep changes, stress concentration, and large deformation exist. For example, the computational efficiency of the FEM is lost when the method is used to analyze problems of metal forming and elastoplasticity, high-velocity impact, dynamic crack propagation, explosion, and shock wave. Even with the modern computing machines, analysis for guided wave propagation in three-dimensional solids by the classical FEM would require impractical computational resources (the computational time and memory storage requirements) [6]. Therefore, along with the ever-growing advancement of faster computing machines, the research into the development of new efficient methods for numerical simulations is an ongoing parallel activity [7].

Motivated by the needs of modern science and technology, considerable efforts have been made in the development of new numerical methods, such as various meshless or mesh-free methods [8], wavelet-based numerical methods [9], the differential quadrature method (DQM) and the differential quadrature-based element method [7], the discrete singular convolution (DSC) algorithm [10], and the high-order FEM [11]. A few examples of the requirements of new methods are as follows: (1) simulations of many dynamic systems often require very fast numerical solution of the equations of the system mathematical models [7]; (2) the computer-aided design (CAD) process in which the database often requires large computer storage and the interpolative manipulations for the operating design parameters may be less accurate as well as quite timeconsuming [7]; and (3) structural health monitoring (SHM) applications call for both efficient and powerful numerical tools to predict the behavior of ultrasonic-guided waves since the existing well-known FEM would require impractical computational resources. The aforementioned methods try to resolve the limitations existing in the classical FEM and to fulfil the needs of modern science and technology.

To reduce the enormous computational costs in simulations of wave propagation in solids, higher- order FEMs with polynomial degrees p > 2 [11], time-domain spectral element methods (SEM) [12–14], and weak-form quadrature element method (QEM) [15] are proposed. Besides the merit of high rate of convergence, they also possess the advantages existing in the FEM and can be implemented into the commercial software such as ABAQUS, since these methods are essentially the FEMs. Obviously, these methods still have some limitations existing in the classical FEM. For example, they cannot be efficiently used in the large deformation analysis such as analysis of metal forming.

Aimed at resolving the limitations existing in the FEM for analyzing problems of metal forming and elastoplasticity, high-velocity impact, dynamic crack propagation, and explosion, various mesh-free methods [8] are proposed. Due to less mesh dependency, mesh-free methods can eliminate possible mesh distortion and entanglement encountered in FEM to analyze large deformation and explosion problems. For example, the method of smoothed particle hydrodynamic (SPH) [16], a mesh-free method, is proposed and implemented into the commercial finite element software LS-DYNA to complement the deficiency of the FEM in dealing with the explosion and high-velocity impact problems. Extended finite element method (XFEM) [17], another mesh-free method [8], is proposed to analyze the problem of dynamic crack propagations. Since the finite element mesh can remain unchanged during crack propagation with using the XFEM, the hybrid method of FEM together with XFEM is convenient to analyze such problems. It is seen that each method is proposed to overcome certain difficulty existing in FEM and to fulfil certain needs by the modern science and technology. Mesh-free methods have some disadvantages: their approximate functions are much complicated and larger computation effort is usually needed; and dealing with essential boundary conditions is far more complicated than the classical FEM [8].

Since wavelets possess multiresolution and localization properties, various wavelet-based methods are proposed to meet some needs of modern science and technology [9]. Similar to the mesh-free methods, wavelet-based methods can be more efficiently used to analyze problems with singularities, steep changes, and stress concentration. For example, the wavelet Galerkin method (WGM) is proposed and has been successfully used in solving a variety of PDEs in regular and irregular computational domains [18]. The drawback of the WGM is that the method can only handle simple boundary conditions and is complicated when nonhomogeneous boundary conditions are considered in two-dimensional cases [9].

The DSC algorithm [10] is efficient and robust for solving the Fokker–Planck equation describing various physical phenomena. The method employs compactly support wavelet interpolating functions. Banded differential matrices with well-behaved condition numbers are obtained thus the DSC is suitable for large-scale computations. Due to employing exterior grid points for treating boundary conditions, spurious eigenvalues are removed although the matrices are not symmetric. The unique advantage of the DSC is that it can also yield accurate high-order mode frequencies. Perhaps its drawback is not easy to apply the free boundary conditions, especially at free corners of anisotropic rectangular plates. Although with the aid of the iteratively matched boundary (IMB) method [19], this issue may be resolved to a certain degree but is very complicated and the existing problem has not completely solved yet.

Because of clear superiority of the FEM and FDM in engineering applications, some new or hybrid methods are proposed to improve the classical FEM and FDM: the aforementioned higher-order FEMs [11], SEM [12–14], QEM [15], mesh-free least-squares-based FDM [20], wavelet finite element method (WFEM) [21], and wavelet-optimized finite difference method (WOFD) [22]. These methods raise the computational efficiency and extend the application ranges of the classical FEM and FDM.

DQM was originated by Bellman and Casti in the early 1970s [23,24]. The method has a relatively recent origin in the later 1980s [25] and has been gradually emerging as a distinct numerical solution technique for the initial- and/or boundary-value problems of physical and engineering sciences since then [7]. In fact, the DQM can be formulated via the polynomial-based collocation method, one of the popular methods of MWR [26,27]. As a numerical method, the DQM can be applied in the fields of biosciences, transport processes, fluid mechanics, static and dynamic structural mechanics, static aeroelasticity, and lubrication mechanics. It has been shown that the DQM is simple and can yield highly accurate numerical solutions with minimal computational effort. The differential quadrature-based element methods are proposed to overcome some deficiency existing in the conventional DQM and extended the application range of the DQM in dealing with complex geometry and boundary conditions. The developed DQM has seemingly a high potential as an alternative to the classical finite difference and FEMs [7].

Summaries on the new development of the DQM as well as on its applications to structural mechanics problems up to the year of 2000 can be found in Refs. [7,27–30]. Since 2000, further developments on the DQM have been made and the DQM has been used successfully for solutions of many engineering problems. Although the progression on the development and application of the DQM in the area of structural mechanics is clear from the past researches, these have been scattered over many papers. Besides, a variety of different quadrature formulations by varying the degree of the polynomials, treating boundary conditions, and employing regular or irregular grid points exist in literatures. This has often caused confusion for researchers and engineers and led to a difficulty in making a choice between a DQM and a DQ-based element method for solving practical problems. Improper choice can be very dangerous since the DQM may admit spurious complex eigenvalues, namely, the stability problem whose occurrence is not well understood in general [31]. To resolve this issue, this book presents new developments on the DQM, the strong-form differential quadrature element method (DQEM) and the weak-form differential quadrature element method (QEM) systematically. A variety of applications are demonstrated. Suggestions are made, and some FORTRAN programs and converted MATLAB files are provided. It is hoped that the provided information can help the researchers and engineers to use the DQM for solving practical problems.

In this chapter, the basic principle of DQM is presented. Various existing DQ formulations are summarized, including the original DQM, the modified DQM, the HDQM, the local adaptive differential quadrature method (LaDQM), and the DQ-based time integration scheme. Existing explicit formulas to compute the weighting coefficients are given. Seven grid distributions are summarized and their discrete error is briefly discussed. Numerical examples are given and some recommendations are made.

1.2. Integral quadrature

There are many integral quadrature methods available, such as trapezoidal rule, Newton–Cotes quadrature, Gaussian quadrature, and Gauss–Lobatto–Legendre (GLL) quadrature. With the advent of computer, Gaussian quadrature has become a well-known numerical integration method. Since it is ideally suited for computers, Gaussian quadrature has been exclusively used in the FEM [1]. On the other hand, the GLL quadrature has become a well-known method in recent years and is widely used in the time-domain SEM [12–14].

The advantage of Gaussian quadrature is its high accuracy. The method is accurate up to a polynomial of degree (2N − 1) with only N abscissas. To achieve the same accuracy, Gaussian quadrature requires the least number of functional evaluations as compared to other quadrature methods. The advantages of GLL quadrature are that its accuracy is up to a polynomial of degree (2N − 3) and that it can result a diagonal mass matrix.

Consider one-dimensional integration of f(x) in the range of [−1, 1]. Numerical integration can be written by

(1.1)

where N is the total number of integral points, and Hi and xi are the weights and abscissas in the quadrature, respectively.

The key to success is the right choice of abscissas. For Gaussian quadrature, the abscissas are the roots of the Nth-order Legendre polynomial. For GLL quadrature, the abscissas are the roots of the first-order derivative of the Nth-order Legendre polynomial together with the two end points (xi = ±1). The abscissas are often called as the GLL points [13].

Once the abscissas are given, the weights can be obtained by

(1.2)

where lj(x) are the Lagrange interpolation function defined by

(1.3)

For GLL quadrature, Eq. (1.1) can be expressed by

(1.4)

The (N – 2) weights can be computed by

(1.5)

where PN−1(x) is the Legendre polynomial of degree (N − 1).

In the time-domain SEM, GLL quadrature plays an important role in formulations of the element stiffness matrix and mass matrix. Due to a diagonal mass matrix, a crucial reduction of the complexity and of the cost of the numerical time integration is achieved, since the dynamic equation can be explicitly integrated with the usage of the central FDM.

Appendix I lists the abscissas and weights (2 < N < 22) in GLL quadrature for readers’ reference. If larger N is required, the abscissas and weights can be obtained by using the software Maple (http://www.maplesoft.com). A simple Maple program is also given in Appendix I for readers’ reference.

1.3. Differential quadrature method

Consider one-dimensional function f(x) in the domain of [−1,1]. Assume f(x) is continuous and differentiable with respect to x. Analog to Eq. (1.1), one has

(1.6)

where N is the total number of grid points, and Aij and xi are called the weighting coefficients of the first-order derivative with respect to x and grid points in the DQM, respectively.

For higher-order derivatives, similar expressions to Eq. (1.6) exist. For example,

(1.7a)

(1.7b)

(1.7c)

where Bij, Cij, and Dij are called the weighting coefficients of the second-, third-, and fourth-order derivatives, respectively, with respect to x in the DQM.

It is seen that the essence of the DQM is that the derivative of a function f(x) with respect to x at a given discrete point is approximated as a weighted linear sum of the function values at all discrete points, including the two end points.

Similar to the integral quadrature method, the key to success