Matrix Structural Analysis Second Edition William McGuire Professor of Civil Engineering, Emeritus Comell University Richard 1. Gallagher Late Profesior and President Clarkson Universtiy Ronald D. Ziemian “Associate Professor of Civil Engineering ‘Bucknell University ® John Wiley & Sons, Inc. ‘New York/ Chichester / Weinheim Brisbane Singapore/ TorontoTA 642 H26 Boe @® wero umrary - 050547968 ACQUISITIONS EDITOR Wayne Anderon MARKETING MANAGER Katherine Hepburn PRODUCTION EDITOR Ken ator COVER DESIGNER Tyne Roan TLLUSTRATION COORDINATOR Sigmund MalinowakifBugene Aietlo ELECIRONIC ILLUSTRATIONS Precision Grophis, Inc 16-1 GGS Information Services, Inc, and poiated “This book wos eet in 10/12 Times Ten by UL rat hound by Hlamiton Fenting Company, Ine. The cover was printed by Phoeaix Color Coyorstion “Tis took sprinted on aie poet) ‘The paper in this book was manufactured by a mill whose forest management progam seen Sdstained yield Rorvesting oft inberlands. Sustained yield harvesting principles varie thatthe nonbere of res et each yet doesnot exceed the amount of new growth copyiht © 200 John Wiley & Son, ne. A sighs seseved No patof his pba mayb reper, store ia a ese pte ansied MEF Tm ot by any mean eetroni, echnieal photcopyng ren, easing ae reo entas permed under Sextions 107 or 108 ofthe 1976 United States Geant cy, witiowt iter the prior waten permission of he Pabst or Suoreaton troup payment ofthe appropiate pa coy foe to the Cyc dae Staten Rosoyood Div, Danvers, MA 01023, (08) 70840, fox Gayo tga Requests tbe Pablar for persion should ended to he (os Depart John Wiley & Sons Ine, Thad Avebin New York,NY Gro oun, Gl) SSD 013 fax (21) 650408, & Ml: PERMREQGWILEY.COM. Library of Cong Cataloging Poison Data MeGuie, Wiliam, 1920- Mat stettcel analy by Wiis MeCaire, Richt H. Cilagine, Ronld D. ieaian— 2nd ed Pe Ineiaes index ISBN 0-7-1286 (Cloth: lk pape) {Structural assis (Baginecrng) Matrix methods alapher Rind H. M-Zlogian, Rens D-H Tite ‘TAS2NDS 2000 arte, sassis7 cr Pate in the United States of Ameo. 098765432Preface to the Second Edition ‘The first edition ofthis book was written 20 years ago, but our aims remain the same: to place proper emphasis on the methods of matrix structural analysis used in eurzent practice and to provide the groundwork for forthcoming practice and allied, more ‘advanced subject mnatter. The present edition accounts for changes in practice that have taken place in the intervening years, and it incorporates advances in the art of analysis we regard as suitable for application now and of increasing importanc: to ‘practitioners in the years just ahead. ‘Among the major changes from the frst edition is the addition of chapters oa the nonlinesr analysis of framed structures, treatment ofthe solution of nonlinear prob- ‘ems, and peckaging of a compact disk containing the computer program, MASTAN2. ‘The major reductions include the deletion of Chapter 7 on the flexibility method, Seetion 10:7 on the transfer matrix method, and Chapter 12 on the finite element ‘method. The entire text and illustrative examples have also been edited exteasively, land several chapters have been repositioned. These changes deserve some explanation, ‘The reduction ia coverage of the flexibility method was presaged by a comme in the preface to the fist edition: “It is hardly used in practice.” Although the methcd is advantageous in certain cizeumstances and we believed it might receive greater aten- tion if east in a more efficient form, such has not been the case. Therefore, the de~ velopment of flexibility method equations of global analysis has been deleted. But extensive coverage of the flexibility approach to the determination of element force tisplacement relationships has been retained because of the value that it has in many instances. Elimination of the global flexibility method formulation also led to recon sideration of the ehapter on tho equations of statics and Kinematies (Chapter 6 of the list edition) which was in large measure preparation forthe identification of seduusent forces in the flexibility method. ‘That material has been deleted also. The other dele- tions material on the transfer matrix method and the general finite element concept— were made in the interest in retaining a reasonably wieldy volume. “The additions represent some departures from the spirit ofthe fist edition. In that edition’s preface it vas noted thet, in concentrating on the direct slifiess method, we ‘vere dealing with an accredited, widely practiced procedure, whereas an earlier gon tration of texts frequently dealt with subject matter that was not flly crystallized. Nonlinear structural analysis has a long history and it isthe subject of a numbe: of books and some highly sophisticated computer programs, but it can hardly be called “erystalized.” Many of its features aro ad hoe—in the good sense of the term—and are in need of reduction to more routine procedures through further research and zeneralization, Others may always require the expertise of specialists in analysis 2s, ‘opposed to the design engineer who has many additional concerns. Nevertheless, iis believed to be timely, and indeed necessary, to include an introduction to nonlinear analysis in a text on the analysis of framed strictures, particularly those of the civil fengineering variety. Techniques for its practical application are at hand, and there is, 4 growing awareness of ther place in design. Similarly in the first edition we decided to exclude computer programs, citing init preface their rapid obsolescence and the widespread availabilty of commercial pro- {rams such as STRUDL. Advances in all areas of computer technology over the pastvi Preface tothe Second Faitton twenty years have changed this, Particularly relevant to our subject isthe capacity for packaging structural analysis programs of significant power in an interactive graphic- upported medium suitable for running on readily available personal computers. "The Computer is nota subst for mastery ofthe subject mater, but in the ability for the (aor to contzol the terms ofthe analysis and to picture the results that it now provides, ican be an invaluable aid to understanding the theory and appreciating the physical ‘ignffcance of the outcome, We believe these pedagogical advantages outweigh the inevitability of obsolescence and are therefore packaging with every copy of the text MASTAN?, an interactive graphies program with provisions for geometric and ma- terial nonlinear analysis, as vell e¢ conventional linear elastic anlysis. MASTAN2 is toned on MATLAB?, a premier software packege for numeric computing and data analysis pero we used ST units excusvely nthe Bist edition, inthis one many of the examples and problems-—particulary those of a nonlinear natuce that have strong, implications for design—are in Imperial units, We do this because the SI system bas hot been fully assimilated in practice, and we believe that appreciation ofthe physical Significance of the results is essential to their understanding, ‘Our sims for the present edition are explained further inthe following discussion of the motivation and salient features of the respective chapters: “Three purposes are served by Chapter , the introduction. First, we present a concise sketch of the dovelopment of the subject. One intent ofthis history is to emphasize that computerized structural analysis methods are merely one part of a continuing process that extends back more than 150 years. Second, the role that computerized Etructural analysis bas played in the design of existing structures is outlined. Finally, the computer capabilities themselves are tied to the programs written for structural nalysis, Partienla: attention is given to the development and use of interactive com: puter graphics. ‘Chapters 25 represent closely allied subject matter related primarily tothe direct stiffness method. Chapter 2serves to define tenminology, coordinate systems, and the most fupdasental notions of structural behavior, bt it also contains two developments ‘Of gteat generality. The fist s the basic character of elemental relations in the form ‘Sf stitinees and Resibility and theie teansformability from one to another and even to Ghemnative fomnats. The second is the fundamental idea of direct stiftness analysis, ‘eseribed here by means ofthe simplest structural element. “A mote formal treatment is given to dizect stiffness analysis in Chapter 3 and, com- sequently, its possible to examine more closely the implications for large-scale pracy tical computation. The latter inchude considerations such as the characteristics of the algebraic equations that are to bé formed and solved. In Chapters 4 and 5 the emai 3 tools needed for the linear elastic stifness analysis of complete frames are estab~ fished, The siliness matrix ofa rather goneral space-frame clement is formulated and then applied, in illustrative examples, (0 a variety of specifc situations Two comments on Chapters 2-5 are appropriate. The fist i that, except for some ‘editing of the text and modification of the examples and problems, they are identical to the same chapicts ofthe first edition. The second is that this inital development of the subject features the basie physical conditions of structural analysis: equilibrium, compatibility, and the mechanical properties of the material. Reliance on work and tnorgy concepts is Kept to a minimum. Our feeling is that itis preferable to help the Sipe name MASTANDis an arog for Matix Structural Anaya, 2° Bain. 1'was developed by the tio foe edeatoal purposes aly, It bas no flagship 0 ay citing eowaercial rectal or Balt ‘emeatenltis peowran SRT Abe os fepaared radewuik of The MathWorks le, 3 Apple Hl Dine, Natick MA 01702008,Profoce tothe Second Edition yi student become attuned to the overall approach to computerized structural analysis in this way, before proceeding to the more powerful and versatile—but more mathe ‘matical—virtual work-based methods. We do introduce the formulation of matrix structural analysis on the basis of virtual ‘work concepts in Chapters 6 and 7. A treatment of this type is necessary to give ‘maximum scope to various aspects of practical design analysis of frameworks, such as tapered members and distributed loads, as well as to the study of nonlinear analysis, and the finite element approach. The theoretical groundwork of the virtual work prin- ciple is Jai in Chapter 6. Both virtual displacement and virtual force concepts are covered, but far greater attentions given to the former because of their roe in stiffness formulations. Chapter 7 examines the implementation of the virtual work principle in linear elastic matrix structural analysis, Chapters 8-10 comprise an introduction to the nontinear matrix analysis of frame- works, The approach is one of using selected structural systems and types of matezial a the basis for explaining how theory can be translated into procedures useful in the design office. Ia the interest of focusing on sources and end results in these chapters, wwe defer to Chapter 12 the treatment of methods and details of nonlinear equation solution and to appendices the coverage of some aspects ofthe basic structural me~ chanies and behavior ‘An overview of sources of structural nonlinearity, levels of analysis used in dealing with them, and a system for casting the required equations in matrix stiffness method forms presented in Chapter 8, Io the course of doing this the notions of the geometic stiffnest matrix and a plastic reduction matrix are introduced. Also, the range of con- ditions to be considered i illustrated by the solution of some elementary problemsby traditional methods. In Chapter 9, the groundwork prepared in the previous chapter is used as the base for the development and application of a matrix method for the nonlinear elastic (geometric nonlinear) analysis of both planar and space frames. The ‘aleulation of the elastic critical load of ideal systems is also treated. Chapter 10 ex: Plays the same base, plus tho concepts of plastic hinge theory, for the materiat and ‘geometric analysis of planar systems. Included is the determination of the inelastic ‘critical load of ideal systems by an adaptation of tangent modulus theory. Some of the more popular methods of solution oflinear algebraic equations are sted- {ed in Chapter 1, and the same is done for nonlinear equationsin Chapter 12. Equation solving is more in the realm of applied mathematics, numerical analysis, or computer science than in structural engineering. Nevertheless since primary responsibility for the entire analysis generally belongs fo structural engineers, they should have more ‘than superficial knowledge of the important aspects of equation solving. Operations peculia tothe handling of equatioas in MASTAN? are explained in detail. Otherwise, these chapters are introductions to the subject that summarize useful methods, men- ‘ion pitfalls that may be encountered, and give guidance to more specialized sources. In the interest of developing basic stiiness methods of analysis in the earlier chap- ters, we put off consideration of some procedures essential to the eflicent and realise solution of practical design analysis problems. Chapter 13 is devoted to the exposition fof certain of these practices, including condensation of analysis equations prior to solution by elimination of specified unknovins, substructuring, the imposition of con- straints, connection stiffness and Gnite joint siz, the exploitation of symmetry and antisymmetry, and procedures for the economical reanalysis of structures when in an iterative design sequeace. Attention is also given to various types of coordinate systeras that are alternative to global coordinates, such asthe local coordinate systems that are ‘convenient for sloping supports. Jn Appendix A, principles of structural mechanics and geometry fundamental ‘0 solving the problem of equilibrium of a deformed structure are reviewed and thenill Preface to the Second Baifon employed in the derivation of toms tat were incorporated without proof in he eo" cpio itiness matrices of Chapter 9. There are several easons for placing his roe saci eau appendix rather than im the main body of the text, "Throughout the text we ea io feandard matrix notation, Concise Weatment of the appended materia} safes we of some elementary concep of tensors and indciel noon. Mo reavies the use of some less obvious Jefnitions of tess and strain ich a econ Tea schol sreses and Green-Lagrange strains, and am excursion into the analyst aie te rotations. Tuese changes in syle and alteration of the analytical ase aro not of init Tut we believe ther introduction inthe earlier chapters would have been = aoereeSiay intrusion on the main agrument, Nevertheless, although iis materi) (rt aa ebial to an initial stody, itis fundamental too true understanding ofthe 5 Bol cos of aoplioear auaysis and to thee extension beyond the Timi explored fa the main text “rhe problem of noruniform torsion in tee-dimensional frameworks i 89 90 deened te this appendix but for a somewhat different reason. Practically, i» Key dressed the reistance of many systems to the destabilizing effects of combined clement Hi flowure Failure to treat the subject would be a serious omission, But co cor theoretical and experimental research is needed to reduce its analysis 1062 seeiaished peocedate. The approach described has e sound basis. Tt worthy of 3p Stlatou, but it must be viewed in the context of en area sill very much under ‘development. “Aprundin B contains «discussion of the problem of distinguishing beawoon Ue oho at tigi body motion and displacements resulting from the deformation of TeGy ea presoted in an appendix rather than in the main body of the fxtbecayse aeaes ee ef meitque rather than the presentation of methods, something that was done in Chapter 12. ore a tents outlined hore repreteat subject mater on the scale of two three- cred hot ono semester courses, wih considerable selectivity of coverage and lind it woergement available {0 the instructor. For example, the Gist seven chapters Pls cout portions of Chapters 11 and 13 might serve for an introductory juniow or ae ererercl course, At the same level, but fr classes of students with a previous, Haas rodustion tothe subject, the first ve chapters could serve for review and Te cnmont of understanding of basc concep, with formal instruction staring with a tel work approwch (Chapters 6 and 7). The seeond level eousse tthe elective aoe uductory graduate level, night feature the virtual work approach to batt Fane ord nonlinear elastic analysis, material nonlinear analysis, and stability studies (Chapters 6-10, Chapter 12, additional postions of Chapter 13, and reference to ‘he spendin). The program MASTAN2an be a valuable adjunct at all evel: for de sPpeMtions for review of the texts mumeTous examples, fr problem solving, as the cee analysis and design studies, and in support of individual programming, An ifctive application is to have students develo their own programs, either as ler, aes fo ine apalysis routines in MASTANGZ or as additions to the program's existing routines. Ourco-avthor, Professor R. H. Gallagher, died in September 1997, wile this edition was in preparation, The idea forthe original book was his, and he was active in he Funny of this edition aad in the resin of several chapters, His contributions are erat toghe work, The remaining authors prepared the material on noalinear analysis Jind we beat sole responsiblity for any ofits shortcomings. William McGuire Roneld D. ZiemianAcknowledgments ‘We are indebted to a number of people for their help in the prepatation of this edition, ‘Our work would not be complete without an acknowledgment of this debt and a par- ticular offering of thanks to the following: “To Professors J.P. Abel and G. G. Deierlein for using drafts ae classroom notesand for their wise advice throughout. To Professors F. C. Filippou, J. F. Haljar, E. M. Lai, and G. H.Pavlino for their thorough, constructive reviews of an carly draft and to Professor K. M. Mosalam for the benefit of his experience in using it in class. ‘Yo the men whose Comell graduate research contributed to the development of smsterial in the book: J. L. Gross, C. I. Pesquera, M. Gattass,J.G. Orbison, Y.B. Yang, S.1 Hilmy, J. L. Castaler, S. N. Sutharshana, D. W. White, C. 8. Chen, C. Chrysos tomov, aad M. R. Atala, ‘To Ms. N. Bulock and her colleagues at the MathWorks, Inc, and Professor J. Maneval of Bucknell University for their MATLAB. support in. developing MASTAN2. ‘To Ms. P. Welzel for her secretarial assistance in preparing the manuscript for publicationContents ‘Chapter 1 Chapter 2 ‘Chapter 3 Chapter 4 Symbols xvii Introduction 1 Li A Brief History of Structural Analysis 2 1.2 Computer Programs 3 12.1 Computational Flow and General Purpose Programs 4 12.2 The Program MASTAN2 6 References 8 Definitions and Concepts 9 24 Degrees of Freedom 9 22 Coordinate Systems and Conditions of Analysis 11 23° Structure Idealization 14 24 Axial Force Element-—Foree Displacement Relationships 16 244 Eloment Stiffness Equations 16 242. Element Flexibility Equations 18 25. Axial Foree Eloment~-Global Stiffness Equations 19 26 Examples 20 27 Problems 26 Formation of the Global Analysis Equations 31 34 Direct Stiffness Method—The Basic Equations 31 32 Direct Stiffness Method~—The General Procedure 39 33. Some Features of the Stiffness Bquations 46 34 Indeterminacy 47 35 Problems 49 Reforences 55 Stiffness Analysis of Frames—I 56 441 Stress-Strain Relationships 56 42. Work and Energy 58 43° Reciprocity 60 4.4 Flexibilty-Stifiness Transformations 62 444 Stiffness to Flexibility Transformation 62 442 Flexibility to Stiffaess Transformation 63sii Contents Chapter 5 Chapter 6 455. ‘The Framework Blement Stiffness Matrix 66 46 451 Axial Force Member 67 452 Pore Torsional Member 67 453. Beam Beat About Iisz Axis 68 454 Beam Bent About Its y Axis 72 455. Complete Element tifiness Matrix 73 ‘A Commentary on Deformations and Displacement Variables 74 4461 Neglected Deformations 74 462. Displacement Variables 75 47 Bxamples 76 48 Problems 88 References 9% Stiffness Analysis of Frame—II 93 54. Coordinate Transformations 93 52 33 34 5.441 ‘Transformation Matrices 94 512. Transformation of Degrees of Freedom 98 513 ‘Transformations and Energy 98 514 Rectangular Transformation Matrices 100 Loads Between Nodal Points 108 ‘Self-Straining—Initial and Thermal Strain Cone 120 Strain Problems 122 Problems 125 Problems 130 Virtual Work Principles 137 64 62 63 64 65 Principle of Virtual Displacements—Rigid Bodies 138 Principle of Virtual Displacements—Deformable Bodies 142 Virtual Displacement Analysis Procedure and Detailed Expressions 144 63:1 General Procedure 144 632 Internal Vistual Work 145 623 Histetwal Virtual Work 148 ‘Construction of Analytical Solutions by the Principal of Virtual Displacements 150 641 Fact Solutions 150 642 Approximate Solutions and the Significance of the Chosen Virtual Displacements 153 643. Bxamples 154 Principle of Visual Forces 158 65. Equations of Equilibrium 158 652. Characteristics of Virtual Forco Systems 161 'Contents xh 653 Formulation of the Virtual Forces Principle 162 654 Construction of Analytical Solutions by the Virtual Forces Principle 168 66 Problems 170 References 173 Chapter 7 Virtual Work Principles in Framework Analysis 174 74 Description ofthe Displaced State of Blements 175, 7A4 Definition of the Shape Funetion Mode of Description 175, 7412. Formulation of Shape Functions 176 713 Characteristics of Shape Functions 180 7.2. Virtual Displacements in the Formulation of Flement Stiffness Equations 181 724 Construction of Expressions for Real and Virtual Displacements 181 722 Virtwal Displacements Formula for an Element Stifiness Matrix 182 723. Application to Standard Axial, Torsional, and Flexural Elements 183 73° Nonuniform Elements 184 74 Nonuniform’Torsion 186 741 Au Element Sttiness Matix 187 742 Application and Examples 190 45° Loads Between Nodal Points and In Approach 194 7.6 Vietual Fores inthe Formulation of Element Force-Displacement Equations 199 7.61 Construction of Blement Equations by the Principle of Virtual Forces 199 762. Fusther Applications—Shearing Deformations and Curved Elements” 208, 7.7 Problems 209 References 214 Strain Bffect—A General Chapter 8 ‘Nonlinear Analysis of Frames—An Introduction 216 8.1 Nonlinear Behavior, Analysis, and Design 216 B11 Sources of Noalineerty 217 812 Levels of Analysis 218 81.3 Examples from Rstablished Theory 20 814 A Commentary on Stability 230 82 A Mateix Approach 234 83. ‘The Equations of Analysis and Their Solution 235 831 Equation Solution-—The Options 236 832 A Fundamental Problem 237 84 Problems 239 References 241dy. Contents Chapter 9 Chapter 10 Chapter H Geometric Nonlinear and Elastic Critical Load Analysis 242 9.4 Geometric Stiffness Matrices for Planar Blements 243 9414 Axial Force Member 243 912 Combined Bending and Axial Force 245 A 9.4.3 Examples of Plane Structure Analysis 246 92 Combined Torsion and Axial Force 254 93 ‘Three Dimensional Geonietric Nonlinear Analysis—An Overview 257 94 Examples of Three Dimensional Structure Analysis 259) 95 Problems 263 References 268 Material Nonlinear Analysis 269 401 Nonlinear Material Behavior 262 10.1 Plasticity Theory 270 3012 Plastic Analysis 272 1013 Further Considerations 273 102. A Plastic Hinge Method For Ductile Frames 274 1021. ‘The Yield Surface and a Plstic Reduction Matrix 274 4022 Defsiton ofthe Yield Susface 277 103 Inclastic Critical Load Theory 278 104 Examples 219 405 ‘The Yield Surface Concept—A Brief Survey of Farther Applications 290 105.1. Spread of Plasticity 290 3052 Moltple Yield Surfaces 282 1053 Reinforced Concrete Members 293 106 Problems 294 References 299 Solution of Linear Algebraic Equations 301 4UL1 The Basie Choice—Direct Inversion versus Elimination or Iteration 302 112 Direct Bimisation Methods 302 112.1 Gauss Elimination 304 3122 Cholesky Method 307 113 Iterative Methods 310 1131 Gauss-Seidel and Jacobi Iterations 310 1132 Conjugate Gradient Method 312 14 Sparseness and Bandedness 316 415 Frontal Solvers 317 ' 116 Errors 322 1161 The Condition Number 326 1162 Estimate of Condition Number 327 tChapter 12 Chapter 13 1163. Error Estimates and Preconditioning 329 N64 Detecting, Controlling, and Correcting Error 333 ALT Problems 325 References 337 Solution of Nonlinear Equilibrium Equations 339 124 Incremental Analysis 339 122 Incremental Single-Step Methods 340 122.1 Buler Method 341 122.2. Second-Order Runge-Kutta Methods 341 123 Tnoremental-Iterative Methods 245 123.1 Load Control Method 347 1232 Displacement Control Methods 348 123.3 Work Control Method 349 1234 Constant Are Length Methods 349 1235 Modified Iterative Technique 349 3236 Convergence Criteria 350 124 Automatic Load Inerementation 351 124. Change in Stiffness 351 1242 Number of Iterations 352 125 Element Result Calculations 352 125. Updating Deformed Geometry 352 3252 Farce Recovery 352 126 Plastic Hinge Consteaints 355 127 Limit Point and Post-Limit Analysis 358 128 Critical Load Analysis—An Figenproblem 360 3281 Reduction to Standard Form 361 1282 Polynomial Expansion 363 1283 Power Method 364 1284 Inverse Iteration 367 129 Problems 370 References 373 Special Analysis Procedures 375 131 Condensation 376 132 Substructuring 377 133 Constraints 385 134 Joint Coordinates 388 135 Connections and Joints 393 135.1 Flexible Connections 393 1352 Finite Joint Size 398 4136 Symmetry and Antisymmetry 403ast Contents| Appendix A Appendix B 1347 Reanalysis Techniques 408 138 Problems 413 References 418 Nonlinear Analysis—A Further Look 419 ‘AA. Vistual Displacement Principles in Lagrangian Formulations 420 ‘2 An Updated Lagrangian Formulation and Its Linearization 425 ‘a3. Application to the Framework Element 426 ‘tot Finite Rotations and Equilibrium in the Deformed Configuration ‘AAA The Finite Rotation Equation 432 ‘h2 Application ofthe Finite Rotation Equation to the Pramew Blement 433 AS Nonuniform Torsion 437 References 441 41 fork On Rigid Body Motion and Natural Deformation 444 B. ‘The Nature of the Problem 444 B2_ Distinguishing Between Rigid Body Motion a 3. Critique of Foree Recovery Methods 447 B31. The Natural Deformation Approach 447 B32 ‘The Blementary Approsch 448 References 450 nd Natural Deformation 445 Author Index 451 Subject Index 453Symbols In matrix structural anelysis, many physical quantities and mathematical operations _nust be represented symbolically. Preparation of the equations of analysis in a form suitable for computer solution requires that all symbols used be defined in a rigid fashion amenable to aumerical interpretation. On the other hand, the development of these equations, with stress upon their physical significance, is often accomplished through the use of simple, ess formal symbols-symbols that vary with the principle under discussion and have a clear physical connotation in the case at hand. In the interest of generality and uniformity, we shall use some basic symbols to denote cestain ‘quantities throughout the text, But the precise interpretation of any of these symbols ‘must be obtained from the local context in which it is used, and in which it wil be explained i general, we use the letter P to designate applied direct forces and P,, to designate applied moments, Similarly, R and Ry, ate used for direct and moment reactions. At their ends and withia elements, we generally denote direct forces by F and moments Dy AE All of these symbols may carry clarifying subscripts and superscripts to indicate sirection, point of application, or member to which the symbol applies. The symbols ‘may appear in either single component or vector form. The symbols 1, v, and » wil designate translational displacements in the x, y, and z dircetions and @,, 8, and 6, rotational displacements about these axes. Generally the etter k will refer toastifiness ‘quantity and d to a flexibility quantity. “Matrices ate denoted by a boldface lett within the symbols [ } (for a square or rectangular matrs), () (or a column veetor), | J (or a row vector), and [| for a Giagonal matrix, Matrix operations are represeated by { J" as the inverse, { J¥ atthe transpose, | | as @ determinant, and | || 48 a norm, ‘As a further guide, the following is alist of the principal symbols used in the text. {As indicated above, most of these may contain clarifying or modifying subscripts, or supplementary marks (overbars , hat symbols”, etc.) that wil be defined in son text. The same applies to the individual components of matrices and vectors lted below. A Area B Bimoment CG ‘Warping constant (a Element flexibility matrix @) Global flexibility matrix, diagonal matcix F,[E] Elastic modulus, matrix of elastic constants e Normal strain e Infinitesimal strain tensor F Normal or shearing force Vector of element nodal forcesx Symbols (6) (6) HBW Mp iN (, (0) e P R P (a) @ IR) fo 5 co) " tT T m u, Ue wi v Vol wwe EME ‘Shear modulus Gradient vector Constraint equation coefficients Halt-bandwidth ‘Moment of inertia Polar moment of inertia Identity matrix St. Venant torsion constant Spring stiffness, rotational spring stitiness ‘Element stiffness matric Globsl stiffness matrix Length Lower triangular matrix Direction cosines Ratio of bending moment to plastic moment Bending or twisting moment Plastic momeat ‘Vector of element shape functions ‘Number of degrees of freedom, number of nodes Nall matrix and vector ‘Vector of global nodal forces Axial force ‘Axial yield or squash load {Ratio of axial force to squash Joad, number of elements Vector of element force distribution functions Distributed load intensity ‘Vector of reaction forces, unbalanced load vector Residual vector Radius of gyration Current stiffness parameter ‘Basis vector ‘Temperature change from stress-free state, torsional moment (Cauchy stress tensor Second Piola-Kirchott stress tensor Force transformation matrix Strain energy and complementary strain energy Upper triangular matrix isplacoment components Potential energy of applied loads Volume ‘Work and complementary work Cartesian coordinates Plastic section modulusSymbols xix Greck Symbols a, p.3 6 bl y fa) Coetficient of thermal expansion, factor defined in context Direction angles ‘Rate of twist, relaxation factor ‘Transformation matrix ‘Shear strain ‘Vector of nodal point displacements Displacement ‘Virtual quantity, variation Green Lagrange strain tensor Absolute percent relative error Angular displacement Curvature, condition number 0, plastic strain factor, effective length factor ‘Nondimensional coordinate ‘Total potential energy 3416.. Radius of curvature, ‘Shear tees, fraction of load increment ‘Normal stress “Yield stress ‘igenvector “Yield surface function Static equilibrium mateix ‘Angle of measure “Acceplable percent error tolerance igenvalu In addition to the above literal aud matrix symbolism, we shall use the following sraphic symbols wherever iti desired to indicate or to stress some particular charac- teristic ofa force or structure, +> Direct force and moment components wr Direct force and moment resultants —-»> Reactive force and moment components, A Atoraivereresenaton of manentxx Symbols Manat Pinned joint Rigid joint Pinned connection Roller support Pinned support laste support Fixed supportChapter 1 Introduction (One of the responsibilities ofthe structural design engineer is to devise arrangements tnd proportions of members that can withstand, economically and efficiently, the con- ditions anticipated during the lfetime of a structure. A ceatral aspect ofthis function is the calculation of the distribution of forces within the structure and the displaced state of the system, Our objective is to describe modern methods for performing these calculations in the particular ease of framed structures, The number of structures that ‘are actually simple frazneworks represents only a part of those whose idealization in the form of a framework is acceptable for the purposes of analysis. Buildings of various types, portions of aerospace and ship structures, and radio telescopes and the like can often be idealized as frameworks. Tin design, both serviceability limit states and sirengts limit states should be considered, Acserviceability limit state is one in which the structure would become unfit for normal service because of excessive deformation or vibration, or problems of maintenance or durability, Generally, linear elastic analysis gives a good picture of the actual response to service loads, A strength limit stato is one in which the structure would beceme tunsafe. Except for sudden britle or fatigue crack propagation, the attainment of a strength limit is generally the result of increasingly nonlinear elastic or inelastic re- sponse that culminates in structural instability. Most often, the internal load distribu tion at a strength limit state is calculated by a Tinear method. In proportioning mem- bers, the analytical results are modified in an empirical or judgmental way to account {or the effects of nonlinearity. Whereas this practice is common, the availability and application of practicable methods of nonlinear analysis is steadily increasing. One of the sims of this book is to provide an introduction to typical methods of this type Fundamentally, the behavior of all types of structures-—frameworks, plates, shell, or solids—is described by means of differential equations. In practice, the writing of tillerential equations for framed structures is rarely necessary. It has been long eslab~ lished that such structures may be treated as assemblages of one~limensional mentbers. Exact or approximate solutions to the ordinary differential equations for each member are well-known. These solutions can be cast in the form of relationships between the forees and the displacements atthe ends of the member. Proper combinations ofthese relationships with the equations of equilibrium and compatibility at the joints and supports yields a system of algebraic equations that describes the behavior of the structure. ‘Structures consisting of two- or three-dimensional components—plates, membranes, shells, solids—are more complicated in that rarely do exact solutions exist for the applicable partial differential equatioas. One approach to obtaining practical, numer ical solutions is the finite element method. The basic concept of the method is that a 12 Chapter 1 ntodaction ‘continuum (the total structure) casi be modeled analytically by its subdivision into fezious (ihe finite clement), in each of which the behavior is described by a set of t seimed functions representing the stresses or displacements in that region. This per~ thits the problem formulation tobe altered to one of the establishment ofa system of algebraic equations, “The practical, aumerical solution of problems in structural analysis thus is seen t0 involve the formation and solutions of systems—sometimes very large systems —of algebraic equations. Also, it shouldbe fant clear that a member ofa framed structure | js simply one example of « moze broadly defined family of finite elements ‘Viewed in this way, structural analysis may be broken down into five parts: 1 Basle meckantes. ‘The fundamental relationships of stress and strain, compat t bility, and equibriam. : 2, Flite element mechanics. ‘equations ofthe element. 3 Eyuation formulation. ‘The establishment ofthe governing algebraic equations ' of the sysiem. 4 Equation solution. Counpttional methods and algocithms | & Salutton interpretation. ‘Toe presentation of resus in. form useful in design “his book deals hel with prt 3,4, and 5. Specially, to maniacal roan ee enc opoach these ors tt caret seems to bo mos Suitable aan aeeccaion fmulation poses sdf taking advantage o he eer os of compra solving lnegeordr syste of uation, A Fo a crema meshancs audbatematrixalgeraispresumed Ony ae ie lemeat method flating Co ned strstr Wil bein a a ats eng et to texts speciaizng a the andamentas of fin ele ca rest postures (og, Refs. 11 and 12), Computational methods and a ee pede in Chapt 11 and 1, bat ote comprehensivecoverage \ aan books on mame analysis (24, et 13) on eo ihe apronch to stetural ays we fe aking requires ome anos toy of hs and elated sujet, Tae following Bit view | may help, “The exact or approximate solution ofthe ditferential 1A. A BRIEF HISTORY OF STRUCTURAL ANALYSIS “Although it was immediately preceded by the great accomplishments of the schoo of Piney andcany such ap Navi and St. Venant, the eviod from 1850 to 1875 isa repeal stating ofa for our eeview. The concep of framework analysis emerged du- ee hs period, through the efforts of Maxwell (Ret. 14), Castiano (Ret. 15), and Mir (Rel 16), among others. At the sme tine, the concepts of matrices were being ‘Mtrodwoed and Joined by Sylvester, Hamilton, and Cayley (Ref, 1.7). These concepts ! ‘Rete foundations of matrix sructral analy, which didnot take fours wat nearly 80 year ater. eae eat chronicle of developments in sritural mechanics in the peiod 1875 to 19205 found in Timoshenko's History of Srength of Material (Ref 18). Ver tle (Bogess wes made in tis period in the development of theory and the analytical Peace subsidiary to matrix structural analysis. To a great exten, this was dc to i teeth Tintations on tho solvability of algebraic equations with more than a few Prt owas, For the structures of primary interest in thet petiod—trsses and flamer aa analysis approach based on member forces as unknowas was almost us versaly employe412 Computer Prograns 3 ‘Around 1920, esa result ofthe efforts of Maney (Ref. 19) inthe United States and Ostenfeld (Ref. 1.10) in Denmark, the basic ideas of a truss and framework analysis approach based on displacement parameters as unknowns took form. These eas reptesent the forerunners of matrix structural analysis concepts in vogue today. Severe Timitations on the size of the problems that could be handled by cither force or dis- placement unknowns nevertheless prevailed until 1952 when Hardy Cross introduced the method of moment distribution (Ref. 1.11). This method made feasible the solution of structural analysis problems that were an order of magnitude more complex than tthe most sophisticated problems treatable by the prior approaches. Moment distibu- tion became the staple of structural frame analysis practice for the next25 years. Today it remains of value as a ready method for solving small problems ancl as an aid in visualizing behavior, but itis priuitive in power compared to computer methods end has been superseded by them in the solution of large problems. "Association of the mathematical conczpts of matrix theory with the engineering concepts of structural analysis appeared in the 1930s, chiefly in the work of Frazer, Dunean, and Collar (Ret. 1.12). The liison developed erratically through the 1940s, but there was no motivation fora firm union until digital computers were born in the carly 1950s, Individuals who foresaw the impact of computers on both theory and practice then undertook the codification of the well-established framework auslysis procedures in the format best suited to the computer, the matrix format. Two note ‘worthy developments were tho publications of Argyris and Kelsey (Ref. 1.13) and ‘Turner, Clough, Martin, and Topp (Ref. 1.14). These papers wedded the concepts of framework analysis and coritinuum analysis and cast the resulting procedures in a matrix format. They were a strong influence on developments in subsequent year. An important factor in the acceptance of computerized siructural analysis in the gineering profession has been the series of ASCE conferences on electronic compu tation, especially the frst three (Refs. 115-117). “The desirability of accouatiag for nonlinear elastic and inelastic behavior vas rec- ognized from the earliest days of the development of structural analysis, and sigaicant tontcibutions were made in the nineteenth century. By 1888 a practical theory for including the effect of elastic displacements in the analysis of suspension bridges had been evolved by Melan (Ref. 1.18), and in 1899 Goodman presented a clear exp nation of the plastic hinge concept (Ref. 119). In the carly years of the twentieth century, Steiaman published an English translation of Melan’s deflection theory (Ref. 1120) and, as reviewed in Reference 1.21, contributions to plastic analysis and design ‘were made by Kazinery, Maier-Liebait, and others. But it was aot until the latter halt of this century that reasonably general, practicable methods for including both geo- metric and material nonlinearity in analysis became a realty ‘Among the seminal events of this period was Turner, Dill, Martin, andl Melosh's 41960 publication of a geometse stiffness matrix for the uniform axial force member (Ref, 1.22), On the material behavior side, a dramatic thrust ocourzed between the end fof World War IK and 1965, Most of the significant developments of that period are reported in References 1.21, 1.23, and 1.24 Since then, there has been a steadily in- Creasing flow of contributions to the theory and application of nonlinear methols of analysis and design. Most of them have been incremental, but their cumulative effect has been profound. 1.2. COMPUTER PROGRAMS ‘When the structural engineer leaves the university and caters the design office, itis probable that he or she will encounter computational hardware and software of con4 Chapter 1 Intoduction ‘ siderable sophistication. The availble programs may require only the simplest data in birder to bring about solutions for the problems at hand. With such tools availabe, it Wrnatural for the student of engineering to question the utlity, or even the necessity, ‘of what may appear to be a study of already-programmed procedures. The answer tO this question is in the responsibilities of the design engineer. The decisions that must i ‘be nude by this individual, extending from the analytical description of the structure, ' Thence to the evaluation of the correctness of the numerical analysis, and through 10 {he interpretation of the output, are not emenable to computerization, The structural Engineer must understand the principles of analysis, Computerization has relieved the pusdons of the rote operations in design analysis, but it cannot relieve the engineer of responsibility for the designed structure. “Proughout this book, therefore, illustrative numericl examples and problems are ‘used Mest of them ace simple. The aim is to avoid obscuring the point or principle vicjer stu in a elond of nimerical computations. But in the process the advantage bf matrix notation and the matsix approsch may not become immediately apparent \ Gihen itis clea thatthe algebraic equations for these elementary structures could have ‘been compiled in a less formal, and perhaps more physically obvious, way. Such i not he cave for most slzuetures encountered in practice, however. They are larger, more ‘complicated, and genertlly require more extensive calculations. Its the effiient ana Jets a such systems that we are smainly concerned with, Although adopting the matrix MRproach may aot induence the way we think about structural behavior, one of is aero trtues i that mentioned earlier: It provides the best format presently known for ‘analyzing structures by computer. 42 Computational Flow and General Paxpose Programs “The equations ofthe marine clement approach are of form so generally app Tie Rat iis posible in theary to write a single computer program that wil slve an \ sie ctimites variety of problems in structural mechanics. Many commercially av sin general-purpose programs attempt to obtain this objective alvkough usually on & ser ted tele The alvantage of general-purpose programs is not merely this capa: \ Tit bat the taity afforded in te instruction of prospective wsers, input and output ata interpretation procedures, and documentation. ‘The four components ia tbe Bowehart of Figure 1.1 are common to vrtally all goneal-purpow Hate element analysis programs. As a minimum, the input phase erpald require of dhe user n0 information beyond that relating to the material of cpotituction, geometric description ofthe finite element representation (including sup- out condition), andthe conditions of loading. The more sophisticated general pur Fete proggansfecitae ths inpst proces through suc features 5 pestored mattis reson arhedules and graphics! displays of the init elemat idealization so that rire in iu may be detected prior to performance of calculations Ta the phase vomprised of tho Library of finite elements is the coded formulative proces forthe individual elements. Most generl- purpose programs contin a varoty Prements, samples of which are Mlustzated in Figure 12. The only element we are ) ceaccrned ith Rere i the framework clement. Tents such as References 1.1 and 12 Seemtbe the ure of ether elements in analytically modeling a coatinsum. The element library phase ofthe general-purpose program receives the stored inpot data aud e8- { toutes the clement algebraic relationships by application ofthe relevant coded for- ‘aRtatve process. This phase of the program also includes all operations necessary to Poston the clement algebraerelationsbis for connection to the neighboring ele- 1 peeks as wel atthe connection process itself. The latter operations produce the al febreic equations that characterize the response of structure12 Computer Programs 5 ry meee peieatmaret eames’ -——] 21 brary of elements ‘Geaecton ofratematet - ‘Rodel for strvearasamens 2 Solution Sed applied loses ‘Constrtion pdsolaion of Siathenateat model tar fx ‘Shura syste. ‘oaipt Display of petite {Toners and forse 1 Stretralanaysie computational flow, Figur ‘The solution phase operates on the equations of the problem formed in the prior phase. In the case ofa linea static aualysis program this may mean no more that the ingle solution ofa set of linear algebraic equations fora known right-hand side. Non- ar static analysis normally requires a series of incremental linear analyses. Solutions, for dyaamie response may require very extensive coinputations over a time-history of applied loads, In ill other cases, particularly where the number of unknowns is very latge, it may be advantageous to divide the total structure into several substructeres. ‘Afinite element idealization of each substructure is made and analyzed, and the results fare properly recombined into a solution forthe full structure. Included in the solution phase aze the back-substitution operations needed fo obtain all the desired aspecs of the solution. © @ o ‘igure 1.2 Sample Suite elements. (0) Frameworks member. (b) Plan stress (2 Hla plate bending (4) Solié element. (e) Curve thin set.6 Chapter Intodveion “The ompus phase presents the analyst with a numerical or pictorial record of the solution upon which engineers can base decisions regarding the proportioning of the 1 Structure and other design questions, The numerical record is commonly a printed list ‘atforees or stresses at points or within the finite elements, of displacements ofspecisieg points, and of otber desired information, Pictoraly, data such as moment diagrams, ' eentour plots of the principal stresses, or plots of the displaced state ofthe structure fay be displayed on the computer screen and printed as permaneat records, “The phases ofthe general-purpose program have one feature in common: the mod- larity oftheir component aspects Insofar as possible, the procedures treated in this teat ove formulated in a manner and form consistent with’ those found in widely dis- ttibuted general purpose programs. But they should also be of use to the student in preparing more limited or specil-purpose programs. Often, programs for a particular Purpose can be designed to treat that problem more eficintly than the general- purpose programs can. i 12.2. The Program MASTAN2 ‘ACD containing a stroctural analysis computer program entitled MASTAN? is bun t dled with this textbook, As shown in Figures 13 and 1.4, MASTAN? is an interactive Graphics program that provides preprocessing, analysis, and postprocessing cap Piltes, Preprecossing options include definition of structural geometry, support com 1 Gitious, applied loads, and element properties. The analysis routines provide the ser the opportunity to perform fist- oF second-order elastio or inelastic anslyses Of two. or ibree-dimensional frames and trusses subjected to static loads. Post: \ ms igure 13 Preprocessing capabilities of MASTAN2,12 Computer Programs 7 (@)Detonmed shape and past ing cations & 2:2ed Order Blstie SuOete alate Cant onder neastic Applied Load Ratio as tert glace! (S007) ()Respontecares Figue 14 Posiprocessng expabiliies of MASTAN2. processing capabilities include the interpretation of structural behavior through de- Formation and force diagrams, printed output, and facilities for plotting response ‘in many ways, MASTAN2 is similar to today’s commercially available software in functionality, The number of pre- and post-processing options, however, has been Linted to misimize the amount of time needed for a user to become proficient wit jt. Many of the theoretical and numerical formulations presented inthe context of the following chepters provide a basis for the program's linear and nonlinear analysis routines la this regard, the reader i strongly encouraged to use this software asatool for demonstration, reviewing examples, solving problems, and perhaps performing ‘analysis and design studies, MASTAN2 has been written in modular format fore, the reader is also provided the opportunity to develop and implement additional to alternative analysis routines directly within the program. Finally it should be noted that MASTAN2 will erecute on any computing platform where MATLAB is available16 Chapter Definitions und Concepts May Fave ace - 2h aa te igure 2.7 Fromework element ‘within the purview of thi text, Io Chapters 2-5 we treat only cases in wi Component elements are prismatic, that i, straight and of uniform cross sectio. pode to node. Nonprismatie clements ate covered in Chapter 7 ‘he complete framework clement has 12 nodal degrees of freedom and 13 {force components, as indicated in Figure 2.72 Usually, however, members of ] ‘structure are joined in ways such that inthe idealization, a mumber of these ‘Components and degrees of freedom may be disregarded. For example, fox the pot of analysis, we may offen assume that trusses are pin jointed. In beams, tox moments may often be disregarded. Other reasons for neglecting certain effoee, ben cited in Section 21 * 24 AXIAL FORCE ELEMENE FORCE-DISPLA CEMENT RELATIONSHIPS Force-tisplacement relationships for the complete framework clement will be & ‘oped in Section 45. First, however, the simple aria! force element (also cafled the. element) will be used for further demonstration of eoncepts and definitions, Force-displacement equations, the relationships between joint forces and join, placements are most commonly written ia either of two forms: stiffness equatio, flexibility equations’ 24,1 Element Stiffness Equations Element stiffness equations are linear algebraic equations of the foram (= pata) : ¢ ‘The matrix fk] isthe element stiffness matrix, and (F) and (Aare element force displacement vectors, An individual term of the [k] matrix, ky, is an element stig coefficient. fa displacement A, of unit value is imposed and all other degrees of fy ‘dom are held fixed against movement (Ay = 0 for k # /) the force Fis equal in. va to ky. The spring constant of elementary mechanics isa stifiness coefficient, Consider a simple tuss member directed along a local axis x and having a kno Shs clement i “complain the sent haf is eapoble of accounting frail fer eae two ‘ons, nual bending, anion (St. Vena orion, Se Scion 74 or deusion ef nomenon toe nd the edition to dares of resdom need to sceount for warping son. Ex Gecipton of hed form, mined force diplaeaent equations, maybe founda he fret edition of text Bech fu iid, at 20 been cde in thi eiton "2A. Aulal Force Element: Force-Displacement Relationships. 17 — o ft eaewet 1 o Figue28 Axial forse element — rember on x ass (a) Anal force ——1, elemeat, (2) Unit dsplacement st @ | & (6) Unit displacement st 2 area A, and Young's modulus f (Figure 2.84). Since itis assumed to be pin-ended and therefore capable of resisting only an axial force, itis a two-degree-of-reedom mem- ber, In Figure 28D a unit displacement in degree-of freedom 1 is imposed with the other degree of freedom held fixed (A, 1; = 1y Ay = uy = 0). The column of joint forces for this ease is equal to the column of stifiness coefficients corresponding toy, or F=f} @ = 1,2) where (=n BP {kab = Less eal® Clearly, F, = ay ia tho force required to impoce a nit value of &,, and F, ~ ky ithe reactive force. Thus the column of stiffness coefficients {ky} represents a system of, forces in equilibrium. The same interpretation can be placed on the sacond columa of the element stiffness matsix (sco Figure 2.8). In Figures 2.8b and 28e the forces F, tnd Fare shown in the directions that correspond to the imposed positive displace- ments. In analysis, unknown forces are generally assumed to act in their positively ‘defined directions, and stiffness coctfcient signs are established accordingly. Thus the ‘correct direction of any foree can be determined from the sign ofthe analytical result. “To write the stiffness equations for this member, we have, from Figure 2.8 and the ‘basic mechanics of deformable bodies: = AL me cr since we wish an expression for the end forces pF and, by equilibrium Reha18 Chapter 2. Definifons and Concepts In the same way, from Figure 2.8¢ Const in matic for, tse equations ae fe) fn ka) 2Af tt] fon ffs ele)-20s TG} 2 Note that io pertinent egisim equation i BA, =O and hat this ein 610 ae ery coluna ofthe snes str Thematrx aso symmetie. aaa ae ene ofthe necessary epost of dsplcement the am of which Wilbe dmc in Seton 3. Ie ets equations for element, such 5 Eguatin 22, ssiates eee ie ae pt forts wih alo he degses of eedom, Maer ae ek tie sng. The reson tilly, the egress of a an nid bodp motion tras, For example, in efi te fst column eee ee oe mutex was sie tne 2 sppcesing the di itis a pl igre 280) Ui had not boc, aay attempt to pose 9 Mert et wol bave been unedited, The member cou respond as oe aid aut to ie ght an usted amount unde heaton of rare yale force Like the dplcenet at node | warsporested aaa eo column ofthe sess mate. Th eppesed deplcenets ae ey tach sup contin ne ss of plcements sociated th aa ton sich deers of eedom, when extracted, cable «more concve sear afthe element tines properties (ce Seton 44) 242. Blement Flexibility Equations ‘Sloment flexibility equations expres, for elements supported ina stable manner, the joint placements, (A) as a function of the joit forces, (Fy): (a) = LE) (aay ‘where [d] is the element flexibility matrix. An element flexibility coefficient, dy is the Value of the displacement A, caused by a unit value of the force J. The subscript fon The force and displacement vectors refers to the degrees of freedom that are free to Aispace. tis used to exmphasize the fact that these vectors exclude components related {fo the support condition, that is, that the constrained degrees of freedom are absent from Equation 24, “Flexibility elationships cam be waitten only for clements supported in a stable man not because rigid body motion of undetermined magnitude would otherwise result from application of applied foxces. They can be applied in as many ways as there are Stable sad statically determinate support conditions. In contrast to the stiffness matrix ‘of Equation 2, the Rexiblity mates not unique. The difference i that in compiling the clemont slfiness matrix, isplacoments are specified (symbolically), end therefore rigid body motion presents no problem of definition. ‘Flexbility relationships can also be derived for elements supported in a statically indeterminate manner but they would not be of general use sinee, with certain excep tioas they could not be combined with other elements ia the representation of complex structure (see Section 4.4.1).2S Asial Force Element—Gl For the axial force member, the flexibility relationships assume the simple forms associated either with a pinned support at node 2 (Figure 2.86): L w= [A] ora pinned support at node 1 (Figure 28): L wl [A] When the compete element framework element is considered in Seeion 4.5 more complex fixity equations wil be found. 28 AXIAL FORCE ELEMENT—GLOBAL STIFENESS EQUATIONS Normally, the simplest way to form the global stiffness equations for the analysis of a structure ig fo start with eloment stiffness equations in convenient local coordi- nates —relationships such as Equation .3—and then to transform them to global axes by methods to be developed in Section 5.1. Mathematical transformations are not ‘essential, however. Element stifiness equations in arbitrary coordinates—either local oF global_—caa be developed from basic principles. Consider the axial foree member in Figure 2.9, which is identical to the member in Figure 2.8. It has known properties ‘nd lies in an z-y plane, but in this case it makes an angle 6—which may be different from zer0—with the x axis. Defining the degrees of freedom as displacement com- ponents parallel to the coordinate axes, there are in this instance four relevant quaa- lies (uy, Uy, Ma v2) rather than the two (ji, us) that are sufficient for the member oriented as in Figure 28. Correspondingly, the resultant force acting on each end, 4 2h, o © [igure 29 Axial force element—member inlined tox axis. () Inclined axial soiment. (2) Small displaceinent, uy (e) Small displacement,20 Chapter? Detmiions and Concepts 2.6 EXAMPLES which must be directed along the bar, may be resolved into its components in the ceeinete directions, as shown in Figure 29a, The element stifiness matrix must now be of order 4 x 4. “fo determine the column of stfiuess coefficients relating the force components to tu, impose a small displacement in the x direction at node 2 and ld all the other gpreee of freedom fixed against displacement (Figure 2:96). Making the usual as Suniptions of small displacement theory, the lengthening of the bar, Ala, is Als rr eos ‘The resultant force in the bar, Fis EA Beavp Ale “Allofthe force components are related to F by equilibrium. Following our assumption. ‘hat the components ofthe force vector remain unchanged as node 2 moves from 2 to 2, we have EA =A cost bs Fas ~Fa = Froosb ao tans <2 sn gcanpom Likewise, for a small displacement in the y direction at node 2 (Figure 2,9¢),it follows, that EA Bg = Fy = By oon b= FA sing con 85 Ba 0 Fy = Basing 2A sit 605 Corresponding displacements can be imposed at node 1 and the results compiled in the element stifiness equations: Fe cost" singcoss vost ~sin cos #] fy Bl zA| sindcosp — sn? singe ~sat |I us Fa| L cos? —sin cos cos" 6 sin cos |) 2 Ba csingcoré sit singeass sit" en @s) ‘Equation 25s « more general form ofthe aial member stiffness equatih than Equa: ticn 23,1 yields the element stiffness equations in arbitrary coordinates directly. For fp = 0, Equation 25 reduces to Equation 23 after the deletion of null vectors, as of ‘course it should, “The following examples illustrate the application of the concepts discussed in this chapter, Although notin themselves examples of matrix structural analysis, they com tain previews of proceduces that are pert of formal matrix analysis methods. Treample 21 shows how member stiffness equations may be used to evaluate diser Aseati Consider ab 200 6% 10° Vera x ig ~ 1664 ENmm oA tao tnar(=f) = a6 = From theft pat of Equation 25 (with uy = 165A(e0? dy + sin ds £08 $05) 166.4(008234, + 0.4615%,) ® BAY ra ni determinate toss is subjected to the oad and resulting bar ces show. What is the displacement ofa? £ = 200,00 MPa. Sx 10" rom the fist part of Equation 25 (with uy © ve * Oy 3A0.0(C08 fact + 8 fi €08 fae.) 1667 = 3200(036004, ~ 0.48000.) o 26 Examples 2 placements in eases in which internal forces have already been calculated. The requize- ‘ment of compatibility of displacements s satisfied through the simultaneous solution of two equations. Note that superscripts are used for defining unambiguously the et- ‘ement on which the force component act. ‘Example 2.2 is an elementary application of the displacement method of analysis for forces and displacements, In Example 23 stiffness equations are used to determine directly the force needed to obtain a desired displacement. Example 2.4 is an elementary application of flexibility equations to the solutior of 4 thermal loading problem. Example 2.5 illustrates that structural equations may sometimes be used advanta- geously inthe solution of small displacement kinematic problems. In Example 26some of the many different waysin which element sifiness equations sre combined to form force-displacement relationships are illustrated. 200X810" _ 200 kNimm $) = ar22 Chapter? Definitions and Concepts Solve Equations and b simultaneously: 0.628, + 046150, = 2003 (0) 0.3600, ~ 048000, = 0.5209 (0) 4, = 241 am —> e072 mm t wy = 252 0m 7 ‘A statically indeterminate tas is subjected to the Jad shown, What fre the ber frees and the displacement ofa? E ~ 200,000 Mo. By symmotiy, joint a must displace horizontally. Therefore, there isoaly ne unkown egw of fcedom, ty Fiom the Sst past of Equation 25, using the nomenclature {ndieated aad denoting 9 as the pial support point (Bs cortar thus, forg = 2 HO X30 IS cote, aN ee BV X10 20 5 x 10°(cos 210° BEI" ue 200 x 5 x 10°08 150° a Bef, = 1000 ~ 2(90.71 + 1250)u, = 9 1 Solving for uy 55 0m26 Examples 23, Calculate the ar forces sing the member stil est equations: a Fepn BIDE AMAT Fg 18O6 AF =-255408 2504255 = 319388 Fog 3682 Feo assane ‘What isthe magnitude and dsecton of the fore P at point « required toSeploce that point vertically downward mun without “ny hosioetal displacement? E = 200,000 MPa, ‘All degrees of freedom, excepto, equal 210. ¢, = —5 mm. From the first and {econd pars of Equation 25, using the nomenclature indicated, with 9 typical oppor pois (Gi Pag 8 ba) “Ye = (B) oo For ab, ha 45% be OO % 1009 6 456 coe 45(-8) = ‘ at = OI in 5" cos 459(-5) = 1168 4N yf = TOT 457(-5) = —1768 WN For ac c= ta8"*Q/-A) = 153.43" pg = — BOX A000 Gin 159.4" one 158 45(-5) = Pin ARTO in 15849 one IDABY(-8) = STEEN FS = 17H A9(sa? 15343")(~5) = —1789 kN24 Chaptes 2 Defialions and Concepts ‘By equilibrium, the required force isthe vector sum ofthese components ‘Two bars ab and be are panes together as shown. Bar be is cooled arc: Determine the diplacement of & and the force in the bars. ‘Terma expansion costicient w= 117 X 10-° mmiun"C, = 200,000 Mo. ‘Assume the pio at bis disconnected and calculate the gap de to free therm contraction of bac be duc fo a temperature change, Taare, te ala? 27x 10-F 3 10 x 40-= 1404 mm ‘Apply equal senate frees in each bar to close the pa: tay ta Using fesibltyreltionships for ty and «Ae. (EL) 1x Dawe ot Hg eae ee Solving for Fand using the Mexibility relationships fr ty 884 fe F100 kN uy = 0501 mm ye = O18 um 10 ns ie Xpress.26 Kxamples 25 ‘A weightless ar thats Sma long rests agains a lope os shown I'the point is moved 10mm tothe sight as Indicated, how fa wll pot b ove up the slope? “Tr [No forces ave requscd the problem is ene of kinematics, From the fist pat of Equation 25, vith the above notation, ra Pot do in foe 4,— a f= sin stm) Bee ga20 asm ani ths 4c a ~ 3+ an "co AV seg = 897mm ‘Two straight bars of different propestes are pinned together as shown. De {elo the force displacement relationships for thee preseribed loading and J Fy PE support conitions. Rani isiaatiiaanmaina igus BP Ltd ® Capone (Lt ath sate BE a Set Cite " Kay * iy From (b), Propenial thee Bee B+ hy EP Bate26 Chapter? Detntions and Concepts “The two force displacement relationships ae: Case TE From (@), Pr From (b), fala Fue Oe ac $ 3 ‘Tae two foreo- displacement re own conn: ‘ali ees: P a aes Pa Case From (a From (6), == = FE okay Rounds canine = From equilibrium, ‘Eglbrsegueoct PPD FE kay + hats ‘he force-aiaplacement relationship i Pm (ban + Katy 2.7, PROBLEMS Jn the following problems all of the numerical calculations can be done with a pocket ‘calculator, but iti suggested that ifthe reader is not familiar with the use of computer programs, the resulls of sample problems be checked with @ program such as MASTAN2, 24 Ineach case shown, the vertical component of reaction at bis 273. Using stifiness ‘quations forthe support elements, calculate the displacement at b. Bach supporting Tink hes the same A, E, and L. Consider the beam itself undeformable, that i, capable of rigid body motion but not intemal straining, tere, . o ® 027 Problems 27 22. For the system shown, Z is constant. Calculate (a) the horizontal force at ze- uized to produce a displacement x, = QOOIL, (b) the displacement at c under this force, and (¢) the resctions at a and d. em 1 = a Problem 22 23. The system and the displacement uy are the same as in Problem 2.2, but the horizontal force is applied at c. Calculate ils magnitude, the displacement at, andthe reactions. 24 The truss shown isthe same as in Example 2.1 except that the area of bar a} is variable, E = 200,000 MPa. (a) For what value of Ags is 0, = 0? Is this a function of the magnitude of the applied load? (b) Is it possible to obtain zero u, by varying As? (© Suppose the applied load were vertical rather than horizontal, would it then be possible to obtain 2er0 u, by varying Aas? Ifso, for what value of Aj,? What would bbe the corresponding u,? 25 The two-bar axial system isthe same as in Example 2.6 except that external axial forces are shown acting at cach of the three nodes. Write the stifiness equations for the assembled system: (P] = (K]lA], where (P} = LP, Py PJ? and (A) = [us my uel Compare elements of (K] with values obtained in Example 2.6. of 26. Using the matsix equation assembled in Problem 2.5 and assuming P, = 0, eim- inate u, fom the equations by algebraic operations and develop condensed mati equations of the form [P,} = (la), where [F)}= LP, PeJ? and (A,} = Lue uel ‘Use the condensed equations to solve Case 1 of Example 2628 Chapter? Deflations and Concepts 27 Whatis the magnitude and disection of the force P required to obtain displace ‘ment components u, = 2 mm and v, = ~3 mm? Cross-sectional areas (mm* x 10°) V are indicated on each bar. £ = 200,000 MPa. ina fag ta Problem 27 ‘ 28 Using flexibility equations for the supporting links, caleulate the displacement a 1 for part of Problem 24. If the flexibility equations are used to ealeulate the dis- i placement at in part b of Problem 2.1, what additional condition must be invoked? Calculate the displacement at b in part b. As in Problem 2.1 consider the beam itself undeformable 29° Comparing the flexibility and stifiness approaches to part of Problem 2.3, at ‘hat points were the requirement of joint equilibrium and joint displacement com- patibility iavoked in the respective solutions? 2.40 Compaze the two systems shown: a structural system subjected to an applied oad P and a resistor network carrying a specified current I. Demonstrate that there is, 4 mathematical analogy between force and current, flexibility and resistance, and dis~ placemeat and voltage. Assume that the structural arrangement is such that the two paralle! bars must elongate equally te ¢ 5 \ & | 2 Problem 210 2cL Compare the structural system shown with @ capacitor network subjected to a baticry vollage V, Demonstrate that there is a mathematical analogy between force27 Probleme 29 and charge, stiffness and capacitance, and displacement and voltage, Make the seme ‘assumption regarding the elongation of the parallel bars asin Problem 2.10. cs cs oa Problem 24 2.12 _A stocl bar having a cross-sectional area of 200 min*is secured atits ends ton ‘aluminum cylinder having a cross-sectional area of 350 mun. The eylinder is heated SO°C while the temperature of the bar remains unchanged. Calculate the changes in Jengths and the internal forces in bar and cylinder. £, = 200,000 MP3, By = 70,000 MPa, a, = 1.17 X 10°F mmimm’C, aug = 2:31 X 10"? mmimnmC, [cee ‘yiacer Bar Problem 2.22 213. ‘The composite member of Problem 2:12 subjected to an external tension Tas sows calculate the changes in lengths and the internal forces in bar and eylinder. ‘Problem 2.13 244 Two stee! cylinders are clamped together by a steel bolt as shown. The boltis loaded to an intial tension of 200 kN by tightening the nut. Plot a diagram of the force in the bolt versus 7, an external load applied to the composite system after bolt igh ening. Asay = 500 mm, Agyindys ™ 4800 mm, E = 200,000 MPe. OYlinder fi, Cynder Ro, Bolt a Problem 2.1430. Chupter 2 Deflaltions und Concepts 215 Use flexibility relationships, geomettical considerations, and equilibrium equa tions to ealeulae the displacement at « and the bar forces. E ~ 200,000 MP. Problem 2.15Chapter 3 Formation of the Global Analys' Equations is In this chapter we discuss the fundamentals of the direct stiffness approach to the osmation ofthe equations of analysis, equations in which joint displacements play the tole of unknowns. This approach requires only the notion and the algebraic form of, the element stiffness matrix. And it merely involves the application ofthe conditions ‘of equilibrium and continuity of displacement at the joints of the analytical model. TThe immediate objective ofthe chapter is to furnish readers with a means of forming the equations of analysis and to help them see their purpose by illustrating their 2o- Ition and the interpretation of results in simple eases. Only axial force members ere considered, Novortheless, the text and the examples illustrate the essence of matrix steuctural analysis, and the ideas contained in them are part ofthe foundation on which the whole system is built Tn extending the scope and power of structural analysis, computer-oriented methids such as the dizect stiffness method require an order in the preparation of a problem for analysis that wat not always essential in the older, classical methods. The same ‘skill, understanding, and imagination are needed, but itis also necessary to establish tnd obey a formal routine in the statement of the problem and inits solution. 3. DIRECT STIFENESS METHOD—THE BASIC EQUATIONS A complete set of force-displacement relationships for a framework element ‘degrees of freedom is, from Equation 2.2, Fm Kady + ally 4000+ Ky 0 dh i + hay 61) Kags + Kala #000 + hyd ++ Fe oh + lg + egy ot Rae Jn woiting Equation 3.1 itis assumed thatthe degrees of freedom refer to the glotal axes of the structure. This is accomplished either by writing the equations directly in slobal coordinates a3 in Figure 2.9 and Equation 2.5, or by applying mathematical au32 Chapt 3. Formation ofthe Global Analysis Equations coordinate trasformations to foree-displacement relationships that have been written in local coordinates, as in Equation 2.3 “The numbers 1, in identity the degrees of freedom at the joints ofthe element, and inthis example they correspond to some convenient global numbering system for these joints, The global numbering system is completely independent ofthe local element MMumberingeyslem. There are no inherent restrictions on the ways of designating global jointe but in any problem, there will be prefered systems, The order of degree-of- Freedom numbering can have a strong effect on the efficiency of the equation solving process (see Section 11.4). ‘When the plobal anaiysis equations are formed using the direct stiffness method, all ‘of the degrees of freedom appear in cach of the rows of Equation 3.1: The element ‘Goes not have a defined support condition. Thus, as shown in Section 2.4.1, the equa tions include rigid-body-motion terms. ‘Once the element force-displacement relationships have been numerically evaluated for all members ofthe structure, application ofthe direct stiffness method consists of their combination in an algebraic form that satisfies the requirements of static equilib ‘um and joint compatiblity at all of the junction points of the assembled model. To illustrate the procedure, consider the formation of the force-displacement equations of the poiat q in the global x direction of the analytical model of the portion of & pin jointed tris shown in Figure 31. "The quantities in the x dicecton at this point are Tesignated by the subscript. The truss bars shown all ie in the xy plane. For eon- Vonience, we have identified the degrees of freedom at each joint, but the only force that has been identified is the externally applied foad, P, that acts in the x direction at point g Figue 34 Representative interior joint ofa plane ss ‘hve die formal Sevelopment ofthe tansormationeqeatins to Section 5.1. ort sal foree meses ‘ne etensstan mate rectangle Svton 51 anda Example 5 used orca testes IESE geton 2 fo tnt ef Equation 2 Even witout the bebe of te rathematie hat precedes Te pera oft exemple atti point wl pose lspht int the cordate waafornation proces.34. Direct Sitiress Method ~The Basie Equations 33 © w Ne ft a, - mA hyure 3.2 Study of equiviam in y (A) the Pry azections For jmstion point equilibrium, the applied load must be equal tthe sum of the internal forces ating on the bars mecting at that point? To emphasize this operation, in Figure 32 we show the bars separated from the join. From the condition of sedizetion equilibrium, FAA FE EP Da Re + PES FE + FP 2) where isthe global, x-direction internal free component on bar. A, and s0 on. The Tovce displacement equations forthe clements, each of the form of Equation 21, yeld cxpresions for FF." FP in terms of the comesponding element degiecs of eed A? ... AP, Substitution of such expressions into Equation 32 results in: Py = (kGMt + kaAd + kGAS + kag) + (KEM? + KEMP + KBAS + KBAR) + (KGAF + KEAE + KGAF + kGOS) ‘+ (kRAP + RRAP + KBAP + kBAP) Since by the condition of compatibility of displacement, As the sme for bats A, Gand D inceach degre of freedom, thai, f = AP = Af = AP = A, iolows that lef + REF AE + RE), + RO, a FUATAE SAE RB + kg te age, 08) 6a) Pos Kids + Kydhy + Keds + Keds #04 Kody 63) “This isthe final form of the desired equations. The capitalized terms Ka, Kay Ko Ko. ate global siffness coeficiens and Equation 35 isa global stifiess equation ‘Note that cach ofthe bars meeting atthe indicated junction points possesses stifiness coefficients with common subscripts, eg, ki, KF, Kf, kB. When the subscripts of ihe coefficients of two or more diferent elements are identical, the elements have a degre: of freedom in common, designated by the second subscript, and such coefficients are Added to form one coefficient of the stiffness equation forthe fores represented by te first subscript, Consider now what happens when an additional member, designated as H, frames into joint q Figute 33). In Equation 32 this merely means that a force Fis added feature ote nomencatre tht intra (or eset forces are entiegby the symbol Fandertemal sores bythe mba P.Appropistesubszits or supers are essigned inch xe34 Chapter Formation ofthe Global Analysis Equations ©. © a Figure 33 Eifect of adding a member toa joint. Figure 3.4) and Equation 3.4 is supplemented by stifiess terms kif, lf, Kllo fl. A. foree F would be added in the y-direction equlicium equation at point q, and sia- iy for Fores Fi and Ff atthe opposite end of the member. There are no other contributions of member H to joint equilibrium equatioas. Thus tis a simple matter to {eviee the stiffness equations to inclade another member, and the effects are quite local ‘Note, however, that after the equations have been solved forthe displacements as functions of the applied Toads, ao simple modifications can be made to the resulting equations to account precisely for the presence of the new member. When @ now seliber is added, the precise solution requires starting with stifiness equations that ‘Recount for this member, But there are simple procedures for obtaining approximate Solutions when structural changes are made. These will be taken up in Section 137. Tn summary, joint equilibrium is stistied through the formation of equations such as Pquation 32, and joint displacement compatibility is satisfied by equating degrees Of figedom, atin going from Equations 33 to 34, Therefore, if the element force- isplacemeat telationships of Equation 3.1 satisfy the proper laws of materials (the Ne Ls Ay 6 SS} ane Jj rr QF o ao w igure 34 Equilbrium in the PA, direction, member added34 Direct Stitiness Method —The Basie Equations. 35 constitutive relationships) and the requirements of equilibrium within the elemeats, nd if the proper supports (boundary conditions) are assigned to the assembled struc- ture, then all the requirements for an exact solution—within the limits of linear elastic theory-—ate satisfied, ‘These concepts are illustrated in Examples 3.1-3., In Example 3.1 the global stiffness equations for an unsupported truss are obtained from clement sliliness equations, and it is demonstrated that the resultig stiffness ‘matrix contains rigid-body-motion terms, Note thatthe axes used in forming the mem- ber stiffness equations are parallel to, but not coincident with, the overall global axes. ‘There is no need for coincidence; Equation 25 yields the correct results without further transformation, provided that the member and overall global axes are parallel. This point will be discussed in Section 5.1, Example 3.2 is an application of the equations of Example 3.1 to an adequately supported system. The ways in which various stiffness equations are used to determine displacements, reactions, and bar-foree components are illustrated. Note that multi pliers of known zero displacements have no effect on the results and thus may be climinated from consideration, Example 33 demonstrates the simplicity of modifying the stiffness matrixto account for strctural changes. Inthe determination ofthe dispiacoments, the invorsion ofthe stiffness matrix is indicated symbolically but, for brevity, the actual caleuletions are not reproduced. For the system shown: 1. Write the member force-dsplacement relationships in global coordinates 2 Assemble the global stifzess equations. 3 Show that the glob stitfaes equations contain rigk-body motion terms. © = 200000 MP. Aa TRAP aa 1098 dl Define the soocdintes, degrees of ficedom, ead exteraal forees as follows: pee 1. Member force-diplacement relationships (see Equation 25) Member a ex) 200 x20 x10 (EA) _ 200 20% 10 aon 34 eNom (2), Big = mani ‘0500-0500 -0500 0500] fay 300 0500-000 J, 05300-0500) a, 000 ls,36 Chapter3 Formation ofthe Global Analsss Equations eA) ro xx . @)- To.sm 1° 7 92949 KNIoun hee ae [Ps 1.000 0 1.000 0} [As fe tee Fl onal” oo” le : Bea rea 1.000 0) As| seer (28), 2928200 en tin Fe So coos eae : FS 7500) 0750 0433 |) As Lm wee 2 Cts quar aio tin 9) ] Un oe 386 “ia —asfe om Te a8 ht Hi Gon Salle ose 3. Rigi body motion. Adding rows 1 and 3 of the global sttiess matrix yields the vector: [2812 16% 3298 0 ~6107 ~10045 ‘which isthe negative of row 5. Therefore, thet is linear depen- ‘ence, the determinant is zero, and the matsbxf singular. This is ‘signal tat, ander an anbiteary load, the displacements are i ‘efnite tha ie, theze may be rigid Body motion, TED “The tras of lsxample 31 fs supported and loaded ss shown, 4, Caleulate the displacements at a and b. 2 Caleulate the reactions ‘3. Caleulate the bar forces, Use equations of Example 311343 A roller support is placed at joint b of the rigid frame of Example 4.13 as shows, below. Using the “joint coordinate” approach of Section 13.4 and the data of Exampe 413, solve the problem to account for the presence ofthis support. Probl 1344 Re-solve Example 5.7 after the assembled stiffness matrix at joint b has been revised s0 that the coordinate axes are aligned in the direction of member ab and perpendicular to it 43415 Example 5.10 has been revised so that the hinge at point c is replaced by a roller that is oriented as shown below. Calculate the displacements for this revised support condition. 1313 Prablen B15 . 13.16 Demonstrate that the use of Equations 13.34 and 13.35 result in the form of Equation, 13:32 13.17 Compute the displacements, reactions, and internal forces for the beam shown. , B= 200,000 MPa and J = 10 x 10° ma 0p 2 be |} 4-20 roblem B17 4348 For the beam shown, plot the relationship between a andthe vertical displace rent vp. For what value of a isthe beam essentially fixed at a? |, — Problem 13.18416 Chaptce 13. Specinl Analysis Procedures 13,19 Compute the displacements, reactions, and internal forces for the frame shown, {E = 29,000 ksi. Use the following connection stiffnesses: 12,000 in. kipsrad 120,000 in. kipslrad Allmsember: Wines —w. Problem 13.19 1320. Compute the displacements, reactions, and internal forces forthe frame shown, 29,000 ksi, Use the following connection siffnesses: (3) ky =k = M44 x 104i. Kipsrad (b) ky = 2ks = 1446 % 10" in. Kipslead. (6) ky =Oand k= © (OK, =Oand ky = 144 X10" in. ipsrad sigs a Altmenters | Wiad | {1s Problem 13.20 4821 Use Equation 13:38 to analyze the beam-column shown. Compare results with ‘an analysis that uses en equivalent axial force and moment acting at the ceatroid of the member, E = 29,000 ksi Problem 13.21138 PROBLEMS 138 Problems 413 proximate method is not always suficiently accurate. The design analyst must exercise judgment in identifying such cieumstances. To improve the accuracy ofthis method, Equation 1349 can be employed iteratively, that i, (8°) + (84°) can be used in ploce of [é"} to obtain a better estimato of the displacement changes. This approach addsto the computational cost, however, which might then exceed the cost of a completely new analysis. Indeed, the cost of feanalysis techniques versus completely new analysis. hhave been questioned in some sludies (Refs. 13.5, 13.6). Note also that a wide variety of reanalysis techniques has been proposed (e.,, Refs. 13.5~13.10),of which the above is merely a simple example. ‘Using the data of Bxample 18:12 and the the displacements in the chonged structure. Compare the results with the “exact” solution, ‘From Pquation 13.50: - (a) + (58) = CT = (2K IOC" P a) 1 fis © GIL Oo] fro ote a0 6 DLR i} =Afso as 9f[fo 1 of -2]o 0 offi as 9 |r uae Le 9 nifllo o 2, ooolls 9 nile 192 29 72 (a x0 m0 1927, nm im mile, 5] 0245 0153. 0092) (7, is 9153 0408 0245 | 7, ulopen Looe o2as o347|Lr, “Te exact solution ia Ueinel form, is us) F363 9227 0136075 to} =]oz27 asst 0273 |} wa) Loa36 ozrs oases, 431 Apply the condensation equations of Section 13.1 to the solution of Problem 34, 132 Condense the angular displacement 0, from Example 5.7 to produce stiffness equations in terms of uy and vy; then solve for the latter. 433 Form the stifiness matrix for the three-jointed axial member of Example 7.1, and then eliminate joint 2 by use of the condensation procedue of Section 13.1, Com pare the result with the axial member stifiness matrix. 134 Use substructuring to analyze the structures of Problem 4.9, 13S Use substructuring to analyze the structure of Problem 4.15. 136 Combine substructuring with, where appropriate, the methods of Section 5.2 to analyze the structures of Problem 58.414 Chapter 13 Special Analysls Procedures 13.7 Use substructoring to analyze the structure shown 1-10 x 108 mat sony a | 1 |e ts sola | a | | bf eae tt oo retin BF 138 Use substructuring in conjunction with a computer program to analyze the ‘russes of Problem 3.15. 139° (a) Compute the displacements, reactions, and bar forces for the truss shown ‘using substructuring, Cross-sectional areas (mom? x 10") are shown on each bar. 200,000 MP2. (Same as Problem 3.1). {(b) Solve the same problem but for a vertical load of 300 KN at d. soo Loe tel Problem 13.9 13,10 Remove the hinge from Example 414 by enforcement of the constraint con ition Oar ~ Sze, O and demonstrate thatthe tesulting stiffness equations are those ‘of a member of length L. 1341 Establish the matrix equations, similar to Equation 1325, to accommodate ‘constraint conditions ofthe type given by Equation 13.18 (ie, (M} ¥ 0). 13.12. Assume that the joints @ and b of Example 5.6 are so connected that Oy = =60,s. Enforce this constraint condition and solve for By.32 Direct Stitmese Method—The General Procedure 1 ports from the remainder, we partition Equation 3.6 in a conformable fashion to sie Pl _ [Ky| Kal far : a - [ee La eo” where all quantities pertinent to the supports are assigned the subscript sand those wisn to fe remaining degrees of featom have the subscript j (Note the separstion Shade fs Example 32 where, in addition to using the sbove parlonng the degrees dr feedom were convenicaiy numbered in advance. Assigning the lst sequence of STumbere to the support degrees of freedom made eubsequent renumbering uaneoes- sary Such a nombeting scheme may be advantageous, Dt i snot essential) Bepandiag Fguaton 3.7 and ating tha (4) = 0, we have (Pq) = (yt) (8a) {P= [Kyte] (8b) “Tho general solution to Equation 38a is obtained symbolically by (4) = B17) = (IEPs) es) where the matic [D] is the set of global exibilty (displacement) cocficients, ‘We have emphasized that the operation of matix inversion is symbolic. In practice, the process is notmally one of equation solving by @ method of the type described in Chapter 11, Also, the grouping of tems and partitioning ndiceted in Pquation 3.7 was adopted forthe sake of logic and clarity in preseutaton. In practice, itmay be nether necessary nor convenient. ‘When (Ay), the vector of displacements at all unsupported nodes, has been found from Fquation 39, the support seactions (Pare found by substitution ofthe result in Equation 38b, yielding, (P= OKIE) 10) ‘To obtain the internal force distribution in the th element, one may multiply the calculated deprees of freedom for that clement, designated by {A by the element ‘sifiness matrix}, resulting inthe aumerical evaluation of the joint force components (Thus ey = eta} eu “Those forces (direct forces and moments) wll be vector components in the directions ofthe global coordinate axes. Separate operations ae required to transform them into {@) cesltant frees, (2) components oriented with respec to the member axes (nonnal force, shea, bending moment, and torgue), or) ual stresses. These transformations ae often made part of the basic analysis, bat occasionally it may be preferable toleave them as separate tasks forthe analyst or designer. “The application of the general procedure is illustrated in Fxamples 34-36. The structures used are elemedtary pi-jointed trsss, but the examples contain mot of the festures found inthe analysis of complex systems. Example 24s similar to Example 3.1 exezpt thatthe calculations are arangedin a ‘nore formal feshion to follow the procedures outlined inthis section. The details of a basic system for forining the global stifiaess equations should be clear from this example, Haample 35s similar to Example 32 except thatthe resultant barfores calculations ‘now follow clear matrix formulation Tho five displacements Were calculated by >a puter, but the results may be verified easily by substitution inthe stiiness equations.42 Chapter’ Formatfon of the Global Analysis Equations In Fxaraple 3.6itis shown that an increase in the number of constrains reduces the labor of solving forthe displacements even though it increases the static redundancy of the system, Questions of indeterminacy are introduced in Section 3.4 For the sytem shown: 4. Waite the force displacement relationships in global coordinates 2 Assemble the global sifiaes equations |. Show thatthe stiffness equations cont E = 200,000 MPs. sigid-bodyanoton terms Define the coordinates, degrees of freedom, and external forces as feng Yoana pans follows wets |e la aby + z ls fe 1. Member foredsplacement relationships (ee Faation 25): " Member ob BA) 100 x 10 x10 . . (FB), - ARE - 0 Ft ra : Pe) kn kan hr be] (8 100 0 -1000 ofa.) ARF Tay re| |e tn A belle oo ofa, Fe [ie tee ky alfa = 1.000 0), re} [ko a ko bal Sym. olla. Member cd 400 Nin ) [is a) 10m 9 1000 oa, i Pel lke 4 oo ola FH | ke anf ~ 00% 1010 ofa, Ft) te se Sym olla, i22 Dicec Sess Method—The General Procedure 43 Member ae 200 15 3108 20 SIE = 300 Nim ag tas ie ( a has eal] Ba ea ss ee] Be a ay eae 0250 9433-0250 oss] (4 950-4333 ~0750 a 310.09 0250 033) Sym, 0750] {4, Member ba (a jac 4, i ‘oso 04s ~0.250 ~0.438] (4) 0750-0433 -0.750|} a. 3.9} 0250 0.433 | ay sym. 704, Member Be fr) Yn tox rb f6 Fe{ [eo ha ke Ralf FE[~ [ky ke hae Kee|] de Fe) [ko kee r (62) 20 ct 11000 0 1.000 |}. 34a] oe has sym. 1.000, 2, Global witness equations in matrix form (see Bquations 3.5 and 36) ey 4750 1299 4000, P 2250 0 Ps 4750 Ay sym, 0 0750-129 0 © 0 fA 0 1299-22590 00 0 || 129 0-0-0750 ~1.299] |, S74 0 ~3464 ~1299 ~2.250) | 6 4750, 1.299 ~400 0 | jas sm 00 [fe 4750, 4.299) |, 2.250) |e44 Chapter 3 Formation of the Global Analysis Equations 13. Righd body motion, Adding rows 1,5, and 7 ofthe global stitiness matrix yields the vector: [4000 0 4750-1299 0 0 0750 1299) Which isthe negative of row 3:Thetefore, the mati is singular. The displacements ate indeinite—there may be rigid body motion. ‘The truss of Example 3.4 Is supported and loaded as shown: 1. Calculate the displacements a 6c, nd 2 Calculate the seneions 4. Caleulate the br forces Use equations of Fxaunple 3.4 y= ama Tp tees ‘Boundary conditions: y = As = 1, Displacements, Remove columns and rows 1,2, and 4 from thestitfaes ‘equations, leaving. a ° 4750 0 0 “0750 ~1.2907)/8, Pe ° 4750 1299 4000 0 |}a, rebel oo \=s0 smd 0 0 Hae 2] | 2sa.sa| Sym 4750 1299) ee) (2s 2250 |(a,) Solving fr (} on computer yet the following resulis, which may be checked by substitution inthe above equations: LAL= LAs. 4) 86 yA) =L-0407 9.809 2.292 10926 ~7.801 mm32 Direct Stiiness Method —The General Procedure 45 2 Reactions, The remsining stiffness equations (ows 1,2, and 4) are sed as follows 0.407) at ete Ga Py) (Re 4000 0750 1259 0 0 5 pleln.t=10} 0 1299 -2250 0 0 |) ioe el Lie 190 3.464 ~1299 ~2250]} "oO 282.9) ~mol en 10362, 4 Bar forss, Develop a formula for caeulatog ba forces from ments: Rom equilibrium atthe 2end of geueral member 1-2 force Fis Fa = Fem Fagot + Fn cos ~ #) ‘In matric form this is A ss aft} seme nonin Fes an Kia Kay Ke | fe] fis as] Substiting the last two of Equations bin Equation a gives the desired formule Tx kin ye] 0 a Kez kas haul} ose sino © Momberas p= 0" Ao ar Ay Be (0 . revamas oft 23 Sf ef tases Membered $= 0 ‘hy Be By Ae, (9809 “io 1 o4]-220 Fag = 40001 of oo 0 of 10.925 f = +488 N 1401 Memberac = 60" a Me f 0 =0433 0250 oss] 0 =0750 0433 0.750]) 9809 ~22%7 +91446 Chopter3. Formation of the Global Analysis Equations Memberbd 9 6 fey (0a 0280 0435 0250 o433]} 0 r= s1n0}0300 0866 “pa33 0980 03 pad “nat = 3268 kN 8 OO “The truss show is the stme as in Example 3.5 except for the addition of horizontal constraints at 6 and c,Caleulate the dsplacements at cand d ‘Remove coluinns and rows 1 to 5 from the stiiness equations, en Pe) o Pye {rl aos Piss ed | rsa sot 907 nro sain co a a;b= 10%) 0250-0144 ae sym. os, ° ° aos} ={ 111} om -2s2a4) | -t.901, 33 SOME FEATURES OF THE STIFFNESS EQUATIONS ‘As has been shown, both element and global stiffness matrices are almost always ‘yinmetrical The conditions under which a set of stiffness coefiicients for a structure FHdeticaon of the ccurcece sd the estment of noseymctizal sites mates are beyond the hope of his tet Soe Releeace 31 for cwtion of pies es.34 Indeterminacy 47 possesses the property of symmetry will be discussed in Section 43. Practically, this fncans that only the main diagonal terms and terms to one side of the main diagonal need to be stored in 2 computer program (see Section 11.4). ‘Also, the stillness (equilibrium) equation fora given degree of freedom is influenced by the degrees of freedom associated with the elements connecting to that degree of freedom, The members shovma in Figure 3.1 or in Example 34, could comprise but a small region of what is actually a vory large structure. Bars such as £, F, and G of Figure 34, and any others that might exist beyond these, have no effect on Equation 35 In other words, the nonzero terms ina given row of a stiffness matrix consist only of the main diagonal and the terms corresponding to degrees of freedom at that joint fand at other joints on the elements mecting at that joint. All other terms in the row fate zero. When there are many degrees of freedom in the complete structure, the Stifiness matrix may contain relatively few nonzero terms, in which case itis cherac- terized as sparse ot weakly populated. Clearly, in the olution phase ofthe analysis its advantageous to cluster all nonzero terms as close to the main diagonal as possible (ee Figure 3.6), thereby isolating the zero terms and facilitating their removal in the solution process. This can be done by ‘bumbering the degrees of freedom in such a way that the columnar distance of the term most remote from the main diagoual term in each row is minimized, that is, by minimizing the bandwidih stint igure 3.6 Typical faa form of row of global stisiness matrix: Bandwidth minimization is but ono strategy for achieving efficiency in the equation- solving phase. Whatever the approach employed, in large-scale applications it is es- ential to the economy of the solution process that account be taken of the symmetry and sparseness of the sliffness matrix. Even the very elementary problem in Example 355 required the inversion of a 5 X 5 matrix. Real structures are often orders of mag nitude larger than this, Equation solving algorithms willbe discussed in Chapter Li. 34 INDETERMINACY Up to this point, the conventional concept of static indeterminacy has not been men- tioned. Cognizaice of static indeterminacy isin fact unnecessary inthe direct stifiaess approach. The structures in Examples 3.2 and 3.5 are statically determinate and those in Examples 33 and 3.6 are indeterminate, yet all were analyzed by identical proce:48 Chapter3 Formation ofthe Global Analysis Kauations dures, The displacement approach rests on a comparable concept thot may be ealled Kinematic indeterminacy. Both can be clarified by comparing definitios of the two. ‘In the early, classical methods of analysis, a flexibility approach was normally em- ployed. In that approach, one first determines the number of equations in addition to those of equilibrium, that are requized for analysis, The number of additional equations is the dogree of static indeterminacy, that is, the number of redundant forces. Ifthe ‘conventional procedure of cutting the steuctre to reduce it fo an equivalent statically cterminate one is used, tho additional equations are developed by employing the ‘clastic characteristics ofthe system and applying the requiremeat of restoring the con- tinuity that was destroyed when the system was cut. The unknowns in the analysis are the redundant forces at the cut sections or removed suppor. Kinematic indeterminacy refers to the number of displacements that are required t0 define the response ofthe structure. The degree of kinematic redundancy is equal t0 ‘the number of degrees of freeciom that must be conceptually constrained to reduce the system to one in which all joint displacements are zero or have predetermined ‘values! In the displacement approach the system i ist reduced to this kinematically determinate condition by considering all unspecified degrees of freedom equal to zero. “The number of such degrees of freedom isthe number of equations that are required {or analysis, that is, the number of rows of the [Kj] submatrix of Equation 3.7. These ‘equations are developed by employing the elastic characteristics of the system and applying the requirement of restoring the equilibrium that was impaired when the conceptual restraints were employed to reduce the system (o an equivalent kinemati- cally determinate one. Mathematically, the restoration of equilibrium was expressed jn Equations 3:2-3.5. The unkuowas in the analysis are the degrees of freedom. ‘From the preceding, two observations may be made: 1. The displacement approach appears more automatic since, in it, all of the un- known quantities are frst reduced to zef0 to produce the “determinate” counterpart, ‘whereas in the flexibility approach selectivity may be involved in designating “redundanis.”” 2, Since the labor of solution is lagely a function of the number of unknowns, the flexibility approach would appear to be advantageous in structures in which the num- ber of force reduundants is less than the number of unknown degrees of freedom, and the displacement approach in all other cases. This view, however, fails to account for differences in effort required to form the equations (0 be solved in the respective approaches.* "The comparative examples in Figure 3.7 indicate that, except in small structures, the number of reduadant forces is usually considerably greater than the degree of kine- ‘matic redundancy. In the igure, all relevant degrees of freedom are includ. In eval- ‘Suppor settlement problems in which ciated displacements ofthe supports are specedin he problema sus cles a which dplscements havo predetermined vals, “Proms or he ont sletion ends foxes athe formation fhe ities snail etods along peal! nes say be foud inthe st etion ofthis book. ‘Note that ie Peemple 35, the system is staaly determinate bt five tines Finally ieernint, ‘whereas in Exnple 36 salar ost tw ies Sate idctefnite and thre nes neatly Indeterminate. Aleit apposth to Example 35 wuld be very spl, whereasin Example 3.6 would ‘be relat time-consuming. Conversely, the displacement sltion of ample 3. smptes tan is in rample 15 But cemdlesofthe dieters in mer of ako within eseh approach, a amp (eas fea ani te plicermentietiod i a leat a eens excep! forthe Soterinton of ‘roe invatially deters sens38 PROBLEMS 35 Problems 4 Stoves Fores Tisedoss Sale Kinematic ® ep oa ndsarA yy datas. oP 41¢ pia : Figure 37 State and kinematic indeterminacy. ‘ating kinematic indeterminacy, we make the common ssstumption that axial defor- ‘mation may be neglected in analyzing flexurally Joaded frames. As an example of the ‘consequence of ths, one need consider only one translational degree of frcedom at teach floor ofthe multi-story frame of Figure 3.72? ‘tis suggested that, in'several ofthe following problems, particulary 3.6 and 3.15, a ‘computer program such as MASTAN? bo used, In all eases, the computed results should be verified by manually checking key equilibrium andlor compatibility conditions. “ina bangs the eal deformation of columns should be taken ito acount This woul ee adtonst degrees of teed ote ana50 Chapter’S Formation of the Global Analysis Equations 31 Compute the displacements, reactions, and bar forces for the trusses show, b 30 ¢_SO kN . PE Now ¢ hsm Lem ‘Problem 3.1 32 The ateas of bars ab and ae in mm X 10° are as indicated. A varies from 0 to 40 X10? mn? (a) Plo the variation in the force in each bar versus A. (b) Ifthe allowable ‘stress is 140 MPa, for approximately what value of A does the system have the highest ratio of load-carrying capacity to weight? Assume the same material is used for all three bars, Problem 32 33. Repeat Problem 3.2 but, in addition to varying A, vary the angle of inclination ‘of bar ad from 20 to 60° by 10° increments, What are the approximate values of A andAS Problems St the angle of iaclination of ad that yield the highest ratio of load-carrying capacity to weight? 34. (a) Astemble the global stifiness equations forthe truss shown, B and A ofall bats are the same, (b) Assuming there will bo no force epplied at e, eliminate the corresponding degrees of freedom from the global equations by proper algebraic op- trations and develop a condensed matrix equatioa containing only the forces and degrees of freedom at cand d, (c) Caleulate the reactions and bar forces for a 2004KN horizontal foree acting to the right at d opr KUSUE A Aes: ME TU. LERART an. Pie 35 (a) Assemble the global stiffness equations for the truss in patt a of the figare. (6) Modify the above to inclode bars db and bf asin part b. (c) Moly the above to inchide bars gh and ij a2 in part c. (4) Discuss the manner in which the changes may ‘or may not affect the labor of solution, E and A of al bars are the same. 4 de tole tut i I ‘CADE. ~ ADA, bowen bohetel 346 Use a computer program ta calculate the displacements, reactions, and bar forees ff the truss shown. A = 40 X 10" mm? for all chord members and 25 % 10° mas for all web members. E = 200,000 MPa, o0KN 100 “00 fe . a F on ee we . is —— Problem 3652 Chapter'3 Formation of the Global Analysis Equations 37 Assemble the [K,,] matrices for the two trusses shown, Test to see whether each Il ‘one is singular or not. What is the significance of your results? A and E are the same {forall bars. btm 4m o o Problem 37 38 The symmetrical truss shovin is mado of three rods whose compressive resistance | nay be considered negligible. In is externally ualoaded condition (part a of the figure) pretensioned so that the force in ab is 50 KN. £ = 200,000 MPa. (a) Calculate the i pretensioning force in ac and ad by statis. (b) Plot the vertical displacement ata versus, the applied load P. Assume linear elastic behavior. j © o t Problem 38AS Problems 53 39 The truss shown isthe same as in Example 3.4 but its now supported as shown. ‘Assume that the support at b sottles vertically 10 mm. (a) Are any forees developed in the members? (b) Caleulate the displacements of joint ¢ and d. (c) Sketch the dis placed structure, ‘ EA Le |—su—l Problem 39 340 A and E are the same for al bars.) Assemble the (Ky matrix (b) Sketch the tisplaced structure under the loads P. (¢) Use considerations of symmetry to reduce the order ofthe [K,7] matix. (A) Caleulae the displacements at b aed c hsm Problem 3.0 3411 Consider the structure of Problem 3.10, carrying—instead of the loads P—two horizontal loads 17 acting to the right at b and c. Sketch the displaced structure, use considerations of anisymmetry to reduce the order of [K,,], and calculate the displuce- ment at b and 342 Three bars of equal A and £ are arranged as shown, (a) Assemble the [Ky] Imatrx and test for singularity. (b) Can a useful solution be obtained when Py = #2? Demonstrate (e) What happens when P, > P,?54 Chapter’ Rormation of the Global Aualysis Equations 343. Compute the displacements, reactions, and bar forces for the system shown, ‘A= 2% 10" mm? and £ = 200,000 MPa for each member. 3.14 If the system shown were modeled as an analogous capacitor network, what ‘Would be the relative values ofthe xequited capacitances (see Problem 2.11)? Consider vertical displacemeat only. = 200,000 MPa, Problem 314 34S Use @ computer program to calculate the displacements, reactions, and bar35 Problems 55 forces ofthe trusses shown. A = 40 X 10* mn? for all chord members and 25 X 10° rm for all yeb members. E = 200,000 MPa, oy - , ee ES amet mebemtemd o eo CT \ 0 ‘hy - aT | teem ~ abel ) * * ZNNW1 | 1 7m Problem 3.15, ° REFERENCE, 3A KL Bathe, Finite Element Procedures, Prentice Hall, Englewood Cliffs, NJ. 1996Chapter 4 Stiffness Analysis of Frames—I Jn this and the following chapter we develop the remaining tools needed for the linear clastic stiffness analysis of complete frames. Here we are using the term frames in a [generic sense to include structures such as pin-jointed planar and space trusses, ‘gid jointed planar and space frames, and combinations ofthese formis;in short, frame ‘works of any shape, jointing scheme, and loading. By the end of Chapter 3 we ean ‘demonstrate the solution of comples, realistic problems. Nevertheless, there will till remain much more to be covered in later chapters—in fundamental areas of structural Inechanies aud behavior, ia nonlinear analysis, and in methods of application such as the techniques of equation solution. ‘Our main purpose in the present chapter isthe development and application of the stiffness matrix of the prismatic, bisymmetrical, 12) 1-0 los Me Syn. als ‘Member ab is beam oa an elastic foundation I = 1285 in’. Foundation ‘moss = 1.5 Kin = 29,00 kt Determine the deflection and the bending moment inthe beam atc by: 1. Blasi theor. 2. The siffnes method p 1 Assuming tinear clastic behavior of the beam ond the supporting foundation it may be sown that (see Ret. 43): Psi pl + sin Bt ‘Woosh pl + coxa © wast coin sam 8.2027 fon) Me 5 i votes npporet on ° ‘equally spsced axial members of 15. per support = 1.5 x 15 = 225 kia88 Chapter 4 Stitness Analysis of Hrames— Defected structure: Moment diagram: 40 ' TTI VATTTTTT u Mx 50. 48 PROBLEMS ‘As in Chapter 3, itis suggested that, in several of the following problems, a computer program such as MASTAN2? be used nd thatthe results be manually checked, 441 Assemble the global stifiness equations for the beam shown. Neglect axial de- formation. Compute the displacements, reactions, and internal forces for the loading and support conditions indicated. ! ps 2 i i; —+ =e i © ste (ppd st) | ae sen (oionlestaat at snd doy) _f . 8 Smmsettenent © — a. sie ‘ ° . in a a sor bowl G De oe o aa @ 42. Develop the stiffness matcx forthe beam in Figure 4.10, that is, verify Equation 433.48 Problems 89 43 One way to support a beam in a stable, statically determinate fashion is a indi- ccaed (rotation atthe left end is provented). (a) Verify that the Dexibility matrix is vi], Lf 2 30] on) GT |-se 6 |e (b) Verity thatthe complementary strain energy equals that of the cantilever and simply supported element (see Example 46). (€) Develop the beam stifiness matrix using this Mexibilty matrix and verify Equation 4.32, Problem 43 44 Compute the displacements, reactions, and internal forces for the beams shown, in the igure, Insert nodes as required. E and / are constant. yt ap til, it __] tt pt . Peablem 44 4S. Carry ovt all of the algebraic operations needed to obtain the final combined ‘member stiffness matrix of Example 4.14 46 (a) By algebraic operations on Equation 4.32 develop a condensed stiffness ma- trix for a beam element ia which the bending moment at one end is slvays zero (part a of the figure below). (b) Use this stiffness matrix in the analysis of the beam in part b. @ o Problem 46 41 Compare the stiffness method with the method of moment distribution.90 Chapter 4 Sitters Analysis of Frames 48 Compute the displacements, reactions, and internal forces for the systems shown, E~ 200,000 MPa, © o Problem 48 49 Compute the displacements, reactions, ad internal forces for the systeras shown. Forthe beams] ~ 700% 10° mat, A = 15 X 30° mm?.For the struts, A= 10 x 10° mai”, E = 200,000 MPa. 100 9 settee tie ft “ as ee Probiem 49 4.10 Compute the displacements, reactions, and intemal foress forthe system shown. For the beam J = 500 X 10° mm, A = 12 10? mn’. For the struts,A = 8 x 10" mm", E = 200,000 MPa. Poobem 410 4511. Same as Problem 4:10 but place an internal hinge halfway between @ and b. | 4132 Compute the displacements, reactions, and interna forces for the systems sown, = 200,000 Ma ° robles 41248 Problems 91 4.13. Plot the bending moment at a versus Zl, a8 [lly varies from 1 10 10. es [}—3—b 4-1} — Problem 4.13 4.4 Plot the ratio of the bending moment at ato the bending moment at ¢ versus Tilly 26 id vatis from 1 to 10. Assume A is infinite (neglect axial deformation ofthe strat). : a T a? ae —— va 445 Compute the displacements, reactions, snd internal forees forthe systsm shown, E = 200,000 MPa. an jt 9 ee — pe am ie > e a | t 416 Members ab and be only possess a torsional stiffness k = GJ/L. Calculate the tangles of twist 0, and 6, at node b due to the applied twisting moments Pras and Pp Problem 4169 Chapter 4 Stitness Analysis of Reames —t REFERENCES ‘Third Edition, MeGraw- 44 S. Timoshenko and J. N. Goodies, Theory of Elasticity Hill, New York, 1970, 42 C. FL Noms, 1B. Wilbur, and S. Utku, Elementary Structural Analyst, Third Béition, McGraw-Hill, New York, 1976. 43 S, Timoshenko, Strengih of Moterials, Part I, Advanced Theory and Problems, “Third Edition, Van Nostrand, Princeton, NJ, 1956. 4A J.P. Den Hartog, Advanced Strength of Materials, McGraw-Hill, New York, 1952.Chapter 5 is of Frames—IE Stiffness Analys' In this chapter, withthe introduction of methods of performing coordinate transfor ‘mations and for treating loads applied between nodal points, sel straining, and tem: perature change, we complete the development of the basic tools required for the linear elastic analysis ofa framework of any shape, connection type, and loading. ‘The transformation of degrees-of-freedom, forces, and force-displacement relation: ships from one coordinate system to another can be treated as an exercise in analytical ‘geometry. It is important, however, to keop in mind the physical problem being ad- ‘essed, tha i, to remember that mercly changing the frame of reference can in no ‘way alter the actual response of the structure. “The concept of fixed-ond forces will be introduced in dealing with distributed Inads, self.straining, and thermal effects. Understanding this notion requires that one visu, lize the application and subsequent removal of temporary, fictitious restraints oa the degrees-of freedom and grant to such constraints the ability to resist force. It is @ concept that is atthe heart ofthe displacement approach to analysis, 5.1 COORDINATE TRANSFORMATIONS We have already discussed the reasons for employing a number of coordinate systems in the analysis ofa particular structure and, asin Figure 24, we have indicated several Iogical choices. The problems can be illustrated further by Figure 5.1, a tripodal space frame with fixed supports and a rigid joint at a, The structure is sliown placed in convenient global coordinate system with axes x,y, and z. A certain natural action is texerted on the @ end of member ab as the system deforms under the applied loads. ‘This action can be portrayed graphically or its components can be listed in vector fashion. There are many ways to do this, but the action itself is immutable. ‘Tollustrate some possibilities we fist consider the ease in which the resultant ation on ab isa single force and not a couple, that itis the bound vector Fay with an crigin at point ¢ (Figure 5.1b). The position of q with respect to a can be stated ia terms of the radial distance , the azimuth angle a, and the elevation angle p. Similarly, we ‘could list the azimuth and elevation angles, y and 6, to the positive direction of the force, The six quantities a, 8, p, % 8 and the magnitude Fy, of the resultant fore= ‘completely define the natural action on the @ end of this member. This is physically cleat, and it would be a useful way to define the action if ab were an isolated member tnd Fy a load applied to a fixed external loading point. But it doesn’t lend itself to the representation of element stifines, and therefore it will not be considered in this connection ‘Two other optioas are illustrated in Figuses 5.Acand 5.14. In the former aconvesient 9394 Chapter §_ Sitess Analysis of Frames—Ul igure 51. Alternative representations of force. (a) Structure, (6) Resultant free (© Local coordinates. (2) Global coordinates. set of local coordinates is established and the action is described in (ems ofthe six i Aantites Fea, Fyn, Few Bens Mya abd Mey, tats, aiect force and tmomeat com } fonents acting petal fo the Toca aces. in the later we view the same ction sr the ‘ix quantities Fay Fs Peas Mans Myas Mces dteet force and moment components acting i parallel othe global axes “The following vectors are thus alternative representations ofthe same quantity WY =LFee Bye Fee Mea Myx Myaj” (BLE Be Fee Maa Me Maal™ Having either of thes, the other canbe obtained by a suitable transformation, We ! twe both repeatediy since we ae interested in transforming force and dpiacerent Yeors and, by logical extension, siiness matrices, rom local to global coordintes tnd vice vera d Sdc1 ‘Transformation Matrices \ I ‘Since, 9s show in Figure 2.7, we can portray displacements (translations and rotations) by line vectors in exactly the same way as we portray forees (direct forces and mo- iSA Coordinate Transformations 95 a) © igure 52 Diret force a a, (a) Local coordintes. (b) Global coordinates. ‘ments, the rules for transformation of displacements and forees are identical! Fur thermore, as showin below, the necessary trausformations may be accomplished ‘through rotation matrices that consist of direction cosines, This is tbe result of two things. First, the equations that comprise the set of global analysis equations invalve equilibrium aad compatiblity at particular joints (see Figure 32 and Equations 3.3-3,5) and, in these, we simply need to transform between vectors passing (hrough the joint: rotation, but not translation, is required. Second, we are using orthogonal coordinate systems. For the reasons given above, the direct force components in Figures Sle and S.1d have the same resultant, In the stme way, the resultants of the two sets of moment ‘components are equal, Consequently, in developing transformations linking Local co- ‘ordinates to global coordinates, one may treat direct forces and moments separately. Rules developed for direct forces also apply to moments, siace they are each repre- sented by sets of orthogonal vectors. Tn Figure 5:2, the direet force components in the two systems are compared, “The resultant direct foree (Sin the Figure), can be resolved into either its local com- ponents Fray Fyn and Fo, (Figure 5.2a) or its global components For, Fer tid Fay (Figore 5:28). 1 follows tliat each of the global components may in turn be resolved into components in the local dizections, and vice versa. The casiest way to express one set of components in terms of the other is through direction cosines. Using the no- Imenclature for direction cosines defined in Figure 5.3, we have, for the local ‘components Fieg = Fg €08 4. FB 608 By + Fog £08 By Brie = Fon 008 05: + By 608 By + Fey 608 3, Fog = Fon 08 047 + Foy 008 By + Fe, £05 Be ‘Ts satement ia valid for tinea ana However, sho in tet sec at Reference rotations of rgd boy Go ot obey ths ln ef vector addition. The fate rataton elect, wie toque spss onda a thes danesional pokes sly, died in Seton A of he Append96 Chapter Stiffaest Analysts of Frames Fy wa igure $3 Direction anges Using, respectively, the symbols J, m, and m, with appropriate subseripts, to designate the cosines of the ine direction angles between the global x, y, and z axes and the subscripted local axes, we have [eBay Mee + Me Pew Ba Foe + My Eon + My Fee 61) Fog = Leg + MF ya + MePes Ba Ba 62 Fra (Fe) = bth 63) “The matrix [y] of the nine direction cosines relating the local coordinate system to the slobal one is a rotation matrix. ‘Recalling that the sum of the squares of the direction cosines for any axis is unity, we have Jn mattix notation this becomes or, in short form, Bt mbt ob Bam te. Te mip + Ga) ‘The length of any vector ise square root ofthe sum of the squares ofits components ‘Thus cach row of fy] bas Uait length, Also, for any two exthogonal axes such 9.x’ and yi_x'and 2', and 2' and y', the sum of the products of the corresponding direction ‘cosines (the éealar product) is zero: Lely + monty + Mytty = 0 Iyde + mymy + yng = 0 65) o Tole + mei + Met “Together, Equations 5.4 and 5.5 state that the fy] matrix consists ofzows of orthogonal unit vectors, Such a matrix is called an orthogonal matrix. itis a square matrix having the distinguishing property of an inverse equal to its transpose: bP = oT (8.6)SA Coordinate Transformations 97 ‘Youngs modulus = Shear made = Figure 446 Bisymmetial framework element. “This property may be verified readily by the identity fy]fvI* = 08} = fvllyI*. teva bbe shov later (Equation 5.14) that use of the orthogonality ofthe [y] matrix simplifies considerably the labor of transforming coordinates. ‘Recognizing again that direct forces and moments transform independently, andthat all that has been said about forces applied to one end of a member applies equally well to the other end, we can use Equation 5.3 directly in compiling the force trans- formation equation for the 12-degree-of-freedom framework element. Referring to Figure 4.6 (repeated here) for general nomenclature, and using primes to desigaate local coordinates, this equation is Per Bn - en or, in shorthand He) 68) where Tea td w- i & oll (F Jsymboizes a diagonal matric in the text). Just sf] is orthogonsl,s0 100 i [X). we =r 10)98 Chapter S. Stituess Analysis of Frames—tt ‘Also, considering that the 12 degrees of freedom of the member in Figure 46 are portrayed by vectors just asthe forces are, we may use the same transformation for these, oF (a) = Oa) Gar) 5:2 Transformation of Degrees of Freedom Having the necessary transformation matsix, the transformation equations for the member stifiness matrix follows directly. If we write Equation 2.2 in local coordinates, ge) = (1A 12) From Bquation 5.11 ar) = TTC) (5:30) and, from Equation 58 (rips = OITA) (sas0) fe = TELAT (S113) ‘Using the property of orthogonality (Equation 5.10), (ey = OT PITH) oy ‘Thus () = (ta) (6.15) where = OFT (6.16) ‘This isthe equation most generally used for finding the member stiffness matrix in global coordinates once one bas formulated it in local eoordinates and knows the ‘geometrical relationship between the two sets of axes. 5.1.3 Transformations and Energy We have arrived at the necessary transformations in a dicect and rigorous way, but it js useful to reconsider the subject from an energy approach. It wil bring out certain principles that may not be apparent from the above and which will be used Iter. ‘The transformed degrees of freedom are not necessarily equal in number to the original degrees of freedom. For example, compare Figures 2.8 and 2.9. In the frst figure the behavior ofthe axial force member is defined by two degrees of freedom, ‘whereas in the second figure, in which the only difference is the coordinate system, our degrees of freedom are required. The relationships connecting the two displace- seat vectors may be written symbolically, as before (a) = era) ay where, now, [F] need not be the square matrix defined in Equation 5.7 an, iit isn't, the vere, [F}-" of Equation 5.130, Goes not exist. To teat this condition we agin suppose the equations to be rasformed are fe) = 8) (12)SA Coordinate Transformations. 99 [Aso, we assume that cack force component F ofthe vector {F) produees the work iat uring the displacement AY, with no work done under any other dsplacoment com. ponents in {A} This s another consequence of employing eomponeats that act along Drthogonal axes. Such foree and displacement vectors are ealled conjugate vecrs ‘Both sets, that is (A'], {H") and (A, {Fave taken to be conjugate sete of vectors. Changing the frat of reference cannot alter the amount of work done. Therefore, for the work to remaia invariant under the imposed transformation, the following equality saust be satisfied EIA) LENA and, from Equation 5.11 Lema) = Lea) ‘Thus (edn) = be os, by transpositos ie Since Equation 5.11 was invoked in this derivation, the transformation of displece- ments presented in Equation 5.11 implies the transformation of forces given by Fqua- tion 5.17, The force and displacement transformations are termed contragredient under the stipulated conditions of conjugacy. The conclusion of immediate interests that, if the force transformation is first defined as in Equation 5.17, then the displacement- transformation matrix of Equation 5.11 is given by the transpose of the force-trans- {formation matrix. The principe of contragredience is ofconsierable inoportance when ihe displacement (or foree) tranaformation is readily constructed from physical mean- Sng, but the formation of the conjugate vector is not readily perceived, This occurs, for teximple, when the condensation of degrees of freedom is accomplished by means of 4 trausformation process (sce Section 133). ‘To fellow this approach in developing element siifness transformations, itis con- venient to deal with strain energy and external work, quantities that were introduced jn Seetion 4.2. Again we require that work remain invariant under the imposed trans- formation, permitting direct substitution of Equation 5.1 into Equations 43 and 44a. ‘Thus, from Equation 43, ere) (617) we Ha'yey = 18") sierra} = 4S oe) and, rom Equation 4, w= aa pea = LAE UCC) 04) om) Hence the transformed stinss mtr again given by w= ere 619)100 Chapter Stiffness Analysis of Frames “The force vectors of cousse transformed by Equation 5.17. The transformation of [k'] into [k] by Equation 5.16 isa congruent transformation (see page 64). Therefore, if [RJ is symmetric, (wil also be symmetcic. S.A Rectangular Transformation Matrices "The main reason for introducing the energy approach at this juncture isto show that, under the condition of contragredience, its not necessary to uso the condition that the inverse of the transformation matrix equals its transpose. Transposition can be Jnvoked dieetly. Tis allows convenient definition of nonsquare cooxdinate axis trans Formation matrices. As noted earlier, the stiffness matrix ofthe axial member in Figure 28 features the two axial displacements. The same element disposed in an x-y plane. arin Figure 29 i described by four displacement componeats. Arranging the local 3 tis along the member of Figure 29 and defining the direction cosines as we have rship between the local and global degrees of freedom becomes : Gis. ak \ “The validity ofthis transformation may be verified by inspection, butts done formally F in Example 5.1 by deleting relevant rows and columns from the complete transfor | imation mattix of Equation 57. In this example the same transformation matrix is used {fo demonstrate the equivalence of Equations 2.3 and 2.5, which were developed in- [ ‘dependently from basic physical principles in Chapter : ‘Welitfonel demonstrations of basic transformations are contained in Examples 5.2 through 5.5. Example 52 is similer to Example 5.1 in that it illustrates the reduetion of the com- plete 12-degree-of fredom transformation mattix toa special situation—in this case {he planar framework member. The partitioning of the [matrix follows Equation 57. ‘in Example 53, the resulls of several previous examples are used to demonstrate the assembly of the global stiffness equations for 2 typical rigid frame. Axial defor- ‘pation effects are inchuded for completeness even though, in most practical civil en- Tineeringfeames of this type, they may be neglected in linear analysis, The example Entains merely a listing of representative matrices and not detsls of their manipula tion. The primed degrees of freedom refer to the local coordinates indicated. “The sual space truss analysis of Example 54 is illustrative of the application of the ‘same transformation techniques to space structures. Tn Example 5.5 the problem of determining the diection casines ofthe principal axes of an oblique member is solved by viewing the total transformation asa series of, Totations, Most general-puspose computer programs have some builtin method for ‘ransforming coordinates that require only simple descriptive input. Nevertheless, the ‘analyst should understand the geometrical and analytical problems involved ‘EXEMELESH “Verly Equation 25 wing Equation 23 andthe transformation Equations 57 and 5.16.© Fy) = Wye where e's a2 * 2 mates Tron ‘equation 23 was developed from a skotch similar ta sketch a of the above figure. It may be Equation 25 was developed from sketch b. It may be written (ey > aay ere {his a 4% 4 mati © ‘Transform to using Pquaton 5:16 (k= [FT W}(T}, as ftlows: ta Equation 5:7 deste Al columns excep those conresponding 10 Fy, J, Fy tnd Fy and al rou exept those or sole ca avs fa) fle me teh v2 ‘Tus Te me of 3 ‘which isthe transformation matrix of this section. rom Equations 23, 516 and the above figue, cod 0 FA/siag 0 ft 1 “ist sing ° eT a cose 0 sing. 14 oe FA WT cos sm. hic is the stiffness matrix of Equation 25, md neot g Hsing cos ¢ Fa) Bs Es Ba 0 o cose sing, soos sis sing cos sag ]102 Chepter Sstitness Amalyss of Frames—it ‘The meter shown is part o a fain located snd londed inthe plan. 1 Reduce the genetal Equation 5:7 tothe patieular transformation equation for this ease 2. Demonstate the orthogonality of the transformation matrix ‘Define the foal and global coordinates, degrees of freedom, and fares: Peay Fv mea, toda sytem Locate 1 Deleting a irxclevant rows aod solumas from Equation 5:7 leaves | i | | 2, Ohogonalty. essay be seen that (FEC) = t ‘The OF = "The rigid frame shown bas = 200/000 MPa andthe following member prop | | (nie for ab, ed, ed, A = 4 % 10° mn’, ~ $0 X 108 mun for be, = 6X $0’, ~ 200 % 10 mun Using the results of Example 48 and the trans i forthe, formation matrix of Equation 5.2, develop the global sifiness equat structure, eluding feeural and axial deformationsConsider member ab n det; how [k] oly for members be, ed, and ed Member ab Wet fe ah 0 0800 20 0 oaxie oo 0800 ‘oso o 008 o 0.048, -n.00 0 008 fen a 00 02 x10 o -r200 04 x 108 Sym Or: 00018 0 12.00 00 o 0800 bu =12200 o 04 108 0 0.900 ° ° 0300 ‘ooo sa Coordinate Transformations 108, =12.0 ° 02 x 10 200 a 04 x 108 ( H ~0.64s2 02591 70813 6452 6452. 02591 0.1095 Bs =70813 17381 xa 02591 0.1095 17381 02591 Sym, 0.1095 Member ty 6 a, 05137 03817 2911 7200 -0si37 03817 9600 03817 ~028i1 04 x 10° ~7200 -9.500 sist ~03817 sym o2sit a 1083 17381 05 x 10 79313 ~17381 1x10 Ou 7200 2.500 02x10 7200 9.600 04x 107i | I 104 Chapter Stitiess Analysis of Frames Member te um Mee Oe ‘oous «0 1200-0008 9 ~1200 050 0 0-080 0 4x1 1200 0 02x10 doo 0 12.00 Sym. 03000 04 x 10% ‘Assembled global stittness equation: ‘A planted space tre fs supported and loaded as chown. E ~ 200000 MPa, Bar Aug 20% 10" A= 30 10" mn? Aug = 4010 eon Ay = 30X10" mt 4, Calslate the dsplacement ata. 2. Galelate the reactions.5A Coordinate Transformations 10S ‘Develop the three-dimensionsl siness matrix for an axial force meater, Sting Equation 434 in the loeal coordinates siouin and llsnating all inelevant degrees of frcedown, es) BAL 1 AY fal) all fe) [4 Te} ote} Tn quation37 delet al ons excep hse coreaponting Uo Fa Byles and Fan owe ong sec tenant Fed's compare Exale © ying Ime mw 0 00 [5 FT ome ol ‘Using Equation 5.16, (k] = (CT TUT), Cr Manele me 7A a 1 Displacements, Record direction cosines foreach member. Locate lea origin for each mens Members ye L le me me -2~ 4 +8 9165 0am —o4G -08729 oe 6 4 "8 10770 Oss asa 0708 a 6 2 8 O19 Ose IS6L EAs a 22-8 aS 0357 ars? 0908 [Recon direction cosine produce multiplied by AIL, (Values of A/E are in mm.) Member (AID) XB meme ae ate 2 ae ‘0M OMS 1663 O20R 01568312 ae 28S 08643 05881 1537-05782 1459507683, ad 39m 1354 01508 2414 04s 1810 —0T38 ae 3.586 4964 0966 3143-0196 O78SR__—0.7858, = 25ui5 1.1469 8957 01s 1761202103 ‘Referring to the fk] matrix and adding stiftesses in the usual way, the global stitines equa- tions forte nonzero degrees offeedom are (a 2s ~01124 17612] (ue pep = 167 02103} v, Pod ( Sym. 8.957 Shwe, Solving fr the displacements, LAE= be oe jox783 2722 0.4863} mm‘ 106 Chapter $. Sttiess Analysis of Frames 2, Reactions. Using the propetis ofthe element stiiness matrices (fee above table) the global ‘components ofthe feartions are 1034 aise] 164 02078 os3i2 328 1.663 | 153 0768) 01964 0964-07858 07858 07858 3143 078m Psion Disgnced ‘oufguation , E—— Fp per t mg ts 13056 mat a ‘ind the direction cosines ofthe member shown. = find ace horizontal yi yertial. The x axis along the meer, The xy" plane males a dikedcl angle ‘Of30" with a vertical plane tough the member ax, ‘Demonstrate thatthe matrix of dicection cosines] fan be generated by three successive rotationsSA Coordinate Transformations 107 om the figure on the Hight OA = VIET eH = 11516 fy #10516 = 08639 ay = cose = 30288" ie = SIILSTG = 04319 By = cos my = GALI 31576 = 02592 8. = cos ny = 4ST Inthe figure atthe right, subseips indicate the postion of tach axe at the end of each rotation. Following our sign invention is a negative angle as pictured. 1 An angle p bout the y axis bringing the x” axe into the xy plane, p = ~tan"*(/10) = —16.699", ca aa a net ie tae ao claehee wth the ake GD" A) = 255. 3. An ang o about the x axe, bringing he yas into coincidence with the y ais an! the ¢ axis into coincidence with the 2a. = 30 ‘Symbolically we may write 4 = fol) = felines) ‘where {5} sot (a are coordinate vectors and fy] [4] and fy.) are transformation matices defining the thee rotations ‘eseibe! above, Using the definitions of Equation 52, 1 0 Of case siny O][cosp 9 —siny’ 0 covw sina |}-siny cosy ol] 0 1 0 0 -ciaw covoll 0° 0 Aillsing 0 cosp 1 0 0 PF oso19 04319 OFF 0.9578 0 oas7s =|0 8c aswoo|]-ossie aso olf 0 1 0 0-050 asso, 0 0 ijl -02873 0 o9s7a 08638 04319 02s91 =|-osm9 o7et ome 00420-04510 05015,108 Chapter Stiiness Analysis of Frames 52 LOADS BETWEEN NODAL POINTS ‘Genetally, strwctures must resist loads applied between joints or natural nodal points tthe system: Three possiblities ere shown in Figure SA. Loading may range from @ few consuntrated loads to an infinite variety of uniformly or nosuriformly distributed loads. o ° igure £4 Loads betveen nodal points ‘One way to eat intermediate concentrated loads is to insert artificial nodes, such as pp and gia Figuce Sa. Tho solution then proceeds asin many ofthe earlier examples ‘Tne degrees of freedom atthe artifiis] modes are added to the total inventory, andthe secessary additional equations are written by considering the requirements of equilib ‘Bom at these novles. The intemal element forces on each side of nodes such as p ancl Trust equilibrate the external loads applied at these points. Assigning the identieal de- free of ficedom designation to the corresponding unknown displacement components eis two sides of the node ensures the satisfaction of compatibility. Tis procedure teed in Exarnple 412 and its tse was implied in several problems of Chapter 4. : "A clitional nodes may also be used when the load is distributed as in Figures 5.4b and 5.4e, Following a process selected or devised by the analyst, the distributed loads ito “lumped” as concentrated loads at suitably selected arbitrary nodes, and the de- frees of freedom at there and the actual joins are treated asthe wnknowns of the Picblemn. ‘There are various ways of approaching or obtaining equivalence between Ne lumped and the actual loading. Ia all cases the lumped loads must be statically i Canivalent to the distsbuted loads they replace. For example, the transverse lond at {he entific joint must equal the total tansverse component of the distributed load rssociated with that joint ‘Use of artificial nodes may be the most expeditious soute to a solution that is in- herently approximate but satisfactory for all practical puzposes. Ttis not a very elegant i procedure, however, ad the addition of the displacements atthe arbitrary nodes tO Rho uaknown degrees of freedom may increase the Iabor of solution unduly. Por these \ ! reasons, iis desizable to have recourse to a rigorous method that eliminates the need for “artificial” joints. : “The approsch that is of most general use withthe displacement method is one that : employs the related concepts of fictitious jolt restraint, fixed end forces, and equi SJent nodal toads. It will be explained by example. But in preparation for doing this, Ge useful to recapitulate an approach we have already used by illustrating its appli- \ it fiom to the continuous beam of Figure 55a, Tbe solution ofthis problem of a system. onded only ata aatural node Would start withthe writing of Equation 3.6: h (ry = E18) eo | “The suppor and the remaining degrees of freedom would then be grouped asin Equt- tion 37 en52. Lond Retween Nodal Points 109 | i - we ® 4 afi fa te Bo se Mite 8 vimuttngen 10-92 AS = 1358 AN “ a7 1 Displacements, Record diection angles, dtection cosines, and products multiplied by AIL (Gee Example 5) AIL is in mm. See Figure 53 for definition ofdtestion anges, Member ar ee ML) GAIL) ik (AIL) Heme A1L) oO “110000100 ° ee 1S 086605000 0.625 a6 01563-02706, a4 o7o1_ ayn gow —_a4onn__ 0.400, 0.400 B68 15568 1294 Stiffness equations for dgplacement at a (see Equation 368): Ce} ()2n(ese S010} {oma Sia ssn) (= ER8g) = 2, Reastlons. Use negatives of member influence coetiiets in the sbove table. nw Rs o 8 a ° Re 0 1a -2s80] | 210] Aol gf tates vamos), 73] 2th Re oars ~01563|{-aoess} *) 200f ) 1002 Ru] | -naeco. 0000 uss] | 18 B,) 04000 0.4000 ass} | 118,53. SelfStraning—Iutal and Thermal Strain Conditions 129, 13. Nullifyeisplacemente Net, thecal, Sxed-end forces must be 2670. For uniform temperature change en fined geometry (@uss eoaiguration),fixed-en frees are a function of the bar reas alone (eee Pauation 523), Ths we have Ag 608 30-4 Agy 608 45 = Ace sin 30 + Aga sin 45 = 6000 ORG) A. + OFOT Any 05000 Ase + 0.7071 Ay = 6000 393 ma? Aus = 5380 ma Ae In Example 5:13 the differential heating of a continuous beam is studied. Although 4 point of inflection appears ia the beam, the bending moment is positive throughout. Sttessing may be thought of as occurring as the beam i strained from the positica it ‘would assume ifthe thermal deformation were permitted to occur in some unrestrained way (Gay, through removal of statically redundant supports). EXAMETE RS ‘The beam @ the same atin Example 56. The depth of span ab is 400 rin ad that of epan bei 200 ram. The bottom surface of both aa iy aaron Sos eP spans sheared 10'Cand the top surface 20°C, with a uniform gradient inbetween. a= 12% 10°? mmfastC. om. sa—| 1 Caleaite the displacements a and b 2, Caleuhte the seactions. Fsod-ead forces (Gee Equation 524) aw = y= AOAC «9510 =——=) Gg Mic pp 200 % 50 x 108 > 1.2 x 10-%(-10) Mees 20 1. Displacements. Use the tffess equations of Pxample 5.6: Ci} -Ghecrlas tN feo eoted 6.0 10° eNom Mi =110 Chapter 5 Stittess Analysis of Frames I Solving for the displacements, Le Cajl.0008609 00005217] red 2, Renettons. See Frample 56. fe Bas Re i875 185 ° Bo agg] 3875-625 | oovonseo) , | 0 Re 0-12.00 | {-0.0005217, 0 Bove 0 02x 1 5 x10. Lak 252 &N 1254N [3913 « 10° ee] ss kno, TO17 en Steed onigrntion _-—-Uasessed configura er ashen sau ‘2.00052 ae SA PROBLEMS ‘Asin the previous chapters, itis suggested tha, in several of the following problems, ‘a computer program sue as MASTAN2 be used and the results checked manually. SAL In the sketches below, the x and z axes are horizontal and the y axis is vertical a part aof the figure the x'y' plane makes a dihedral angle of 75° with the vertical plane through OA and in part b the x'y’ plane makes a dihedral angle of 90" with the ‘vertical plane, Deteraiine the direction cosines forthe two elements OA.54 Problems 181 5.2. ‘Two sets of parallel, orthogonal axes are shown. Six veetor components act at (the origin of one sot. Develop the transformation matrix that relates six statically equivalent parallel components (not shown on the sketch) acting at Oto those shcwn. Problem 52 53. The two sets of axes shown lic in the same plane. Using the rotation matrix [y] and the resits of Problem 5.2, develop a transformation matrix that makes P,?,,and Prag statically equivalent to Py, Py, and Png Show that it can be expressed a the product of a translation and a rotation matrix. 54 ‘The member A isthe some asin Example 5.5. Using the results ofthat example, plus anextension ofthe reasoning used in Problem 53, construct an equilibrium matrix [cP] that relates reactive components acting at O parallel to the globel axes (these152. Chapler$. Stffess Analysis of Frames i ‘components are not showa in the figure) to the forces acting at A that are expressed in local coordinates. Problem 54 i 55 Compute the displacements, reactions, and internal forces for the systems shown. Soom | [ew one some] | < je : . i 40m fed mooted maf 60—}-4n—-—# n+] ° ° 56 Assume E = constant and A = zero for all members; calculate u,v, and 0 for joint. Also calculate the disteibution of internal forces. Why isn’t the answer realistic? ‘What is the shortcoming of the solution and how can it be corrected? What answers ‘would one obtain to the same problem using moment distribution? What assumptions 8 to deformation are iroplicit in the moment distribution solution? | i Problem 56 87 Assume E = constant and A = zero for both members. Assemble the stiffness ‘equations required for the solution of u,v, and 0 at joints b and c. Show that this solution feils in this case, What isthe difficulty and how can it be corrected? Why wasS4 Problems 153 «stiffness solution (albeit an uarealistic one) obtainable in Problem 5.6 but not inthis case? rao 3 Problem 57 ‘58 Compute the displacements, reactions, and internal forces forthe systems shown. [Neglect axial deformation (i, agsime it to be 2er0) except inthe structures for which areas are presented. = 200,000 MPa. (Gee Problem 42s or properties) Sa wis i a ! ® Sina i ar * Es o Me Mm Problem 58 9 ‘The structure shown is the same as in Fxample 5.3. Using a computer, caleuate the displacements, zeactions, and internal forees. Solve for (a) axial deformation in- ‘cluded, and (6) axial deformation neglected. What difference does it make? Double134 Chapter S-Stitiness Analysis of Frames—IT the moment of inertia ofthe columns, solve, and compare results. Double the moment of inertia of the beams (with column's at original value) ancl do likewise. Problem 59 5.0 Compute the displacements, reactions, and internal forces for the systems shown. eame W636 inane W255 Baws suse 4 a Messe ame 4 Ieee amt <, ge co sectnte ta eoaware aie ues ileoa reson eames ° © Problem 5.1054. Problems 156 S41 Bars ac and ce are heated 30°C, Compute the nodal-point displacements aud the bar forces. A'~ 6000 mm? all bars, B= 200,000 MPa, a = 1.2 x 10-* mm/mm*C. 512 ‘The depth of spans ab and ed is 200 mm and that of span Be is 400 mm. ‘The bottom surface of span be is heated 40°C and the top surface 15°C, with a uniform temperature gradient in between; the other spans remain unheated, « = 1.2 X 10°? ‘mmimm'C, F = 200,000 MP2, Compute the displacements, reactions, and interaal forces. st a Lt [513 ‘The problem is the same as the provious one except for the struts at B and e ‘These aze Meated 15°C in adaition to the beam being heated ay presciled in Problem 52. * » . 4 se ° L Problem 53, 514 ‘The beam shown is of constant section throughout its length and it is subjected to-a depthwise thermal gradient A all along its length. Calculate the displacemen's, reactions, and internal forces for (a) La = Zs, and (6) La = 214. Compare results Problem 5415 (a) Il is desired to pretension the system in part a of the figure so that, when the geometrical configuration shown is obtained, the tension in ab is 50 KN, To what Jengihs should the individual bars be fabricated if this is to be attained? B126 Chapter 5 Stittuss Analysts of Pranes—It 200,000 MPa, (b) The problem is the same but, in this case, there is an additional bar fe, as shown in past . Problem 515, 546. ‘Two steel cylinders are clamped together by a hot driven rive. ‘The heads are oumed on the rivet at a temperature L000°C higher than the eylinder. The rivet coo's tains the tesistance of the eylinder. Assume that the sivet is elastic throughout the fooling process. What are the forces inthe system after the rivet has cooled 1000°C? Bove 500 mt?, Acyinee = 4800 mm?. H = 200,000 MPa, a = 1.2 x 10°* mmm’ lade Gvinder a Problem 546 1547 Given the axial force member stiffness equations in orthogonal coordinates = tad y (Equation 25), develop the transformations necessary to transform these equa tions into the eblique coordinate system defined by the axes 7yand £and the included angle y Compare the transformed equations. Problem 517 REFERENCES 54 FP, Beer and E, R, Johnson, Vector Mechanics for Engineers: Statics, Sixth Eai- ton, McGraw-Hill, New York, 1996. 52 A.B.U Love, A Treatise on the Mathematical Theory of Elasticity, Fourth Edition, ‘Dover Publications, New York, 1944,Chapter 6 Virtual Work Principles In prior chapters we established the relationships of framework analysis by dealing directly with the basic conditions of equilibrium and displacement continuity fr eestic Structures. That approach has considerable appeal because ofits straightforward na- ture, but another method virtual work—can be exploited for the same purpose as well as for the development of important analytical tools that are beyond the react of the simple physical approach, Virtual work isa general term that includes two distinct principles, those of viral displacements and virtual forces. ‘Virtual work concepts, ia one form or another, wll have been encountered by any student who has been exposed to elementary structural analysis. Because our purpose “dealing with such principle is to establish a basis for general framework analsis, "a necessary (0 take a closer look at theoretical questions than is usually the case in iroductory structural analysis. Ia Chapter 7 we demonstrate the utility of these dhe- ories in mattx structural analysis. (References 6.1~63 contain further discussions of the theories and their application.) "This chapter begins with an examination of the concept of work in structural me- chanics, a topic treated brielly in Section 4.2. The notion of a virtual, or imaginary, ‘eplacement is ntrxhiced and, by stdy of the response of a vigid body to this ype of displacement, the most clemsentary form of the principle of virtual-displacements is established. The usefulness ofthis form of the principle is limited to the calculation of support reactions and member forces in statically determinate structures. ‘The principle of virtual displacements is of value principally in the formulation of stifinese equations, which imply deformable behavior. Therefore, the concepts devel- ‘oped for rigid bodies are established for deformable bodies in Section 62. Detailed ‘letual work expressions are constructed for axial, torsional, and flexoral elements in Section 63. Although wo emphasize the theoretical foundations of virtual work, a proper appreciation of certain ofits features and subtleties ean be gained only through Epecific examples, Hence, Section 6:4 examines the principle of virtual displacements ‘a applied to a few classical analysis problems. “The principle of virtual forces is the alternative, or dual, ofthe principle of vistual displacements and is the basis ofthe direct formulation of element flexibility relation- ships. Although the stiffaess formulation is emphasized in computerized structorat ‘analysis, there are many cases in which the formulation of flexibility retationships and their transformation to stiffness (erms as deseribed in Section 44s simpler than the direct formalation of stiffness relationships, The theoretical basi’of the principle of Virtual forces is therefore established in Section 63, We (eat the principle in terms of its role in element flexibility formulation rather than as an exposition ofits historical place in classical indeterminate structural analysis at486 Chapter 6 Virtual Work Prtucipes GA PRINCIPLE OF VIRTUAL DISPLACEMENTS—RIGID BODIES ‘To develop the principle of virtual displacements wo must amplify the definition of ‘Work piven previously in Section 4:2. There, it was stated thatthe work of» force Fis we [eas on here A isthe displacement ofthe point of application of the fores in the direction of The force, with Ay the final amplitude of displacement. This development wi be limited to forces that are gradually applied so that the displacement grows in proportion 10 the force as shown in Figure 6.la, The figure refers ¢o conditions of linear elastic ‘behavior, This restriction is unnecessary in the development of virtual work principles etheir more general form, but itis adopted here because we deal only with Tinear clastic structures in this and the next chapter. Tor a given intensity ofthe load Fy the amount of work produced, Wis given by ‘the shaded area of Figure 6.1a, If seall increment of displacement, dA, is imposed through the action ofa small increment of force, dF, the change of work is dW = Fo(d) + U/2(aF (ad) (62) If we diseard the second term ca the right side, based on order-of magnitude consi trations, we have the “frst order" change of work: av = Feld) (628) ‘imagined “The same form of relationship is obtained if, as pictured in Figure 6.16, that the displacement increment occurs without the action of a corresponding force herement, Such en imaginary or hypothetical displacement is termed @ virtual dis. ‘Placement 88, andthe increment of the work of the existing force acting tbroush the Final displacement, 8%, = F(8A),is known as virtual work. The vetual displacement, ‘eing imaginary and unrelated to any motivating foree, ced not be restricted to small” ress-but wil, instead, be arbitrary in direction and undefined in magnitude. Ifthe “atual displacement is not mall, due account must be taken ofthis fat inthe resulting founmlation. Only small virtual displacements aze treated here. ‘in the following we Will use 84 and the like to designate virtual displacements." CO) o igure 61 Force-diplacement relationship. covebae was placed on al sta placement symbols to ini Svs are omited ins eton became we wll We the Saya an “steyentolopeator, Tesco, tee shout be no daoger of “in the et eon of ti text a ‘hem om el dplaceets. Ove {Ector o aul quanes and not tlsltepreaion61 Principle of Virus! Displacements Rigid Bode 139 Virtual work will be designated as 31, frequently with a subscript indicative ofits ‘We first consider a single particle acted upon by a system of forces Fy... Fs.» Fa which are in equilibrium. Denoting the direction cosines between an arbitrary ci- rection and the fo1e3s a5 Ay,-.-, Ay-++ Ay the condition of equilibrium of the forces is FA= 0. ‘Suppose the particle is subjected toa virtual displacement 84 in an arbitrary dines: tion. The virtual work is then aw= F(a) 3) where “denotes the dot product, Introducing the direction cosines that give the com- ‘ponents of the forces Fin the direction of A this can be wr BW = Fyay(BA) + +++ Fa(BA) + +++ FA(6A) + (sa) (632) and, since the bracketed terms that multiply 3A sre zero in accordance with the coa- dition of force equilibrium, i€ follows that SW = 0, This is the virtual displacemeat principle, which can be expressed as follows For a particle subjected to a system of forces in equilibrium, the work due to a virtual displacement zero. 11s of interest to ascertain whether the virtual displacement principle can be used to determine the converse, thats, ia system of forces acting on a particle isin equ librium. This would appear to be difficult, since the vitual displacement is arbitrary, implying. an infinity of directions in which tho virtual work can be assessed. Any dis placement, however, can be defined by components in the independent coordinate izections, such as three independent components in a general three-dimensionsl state. “Thus one can establish the conditions of equilibrium in the latter case by waiting vstual ‘Work equations for three independent virtual displacements. The converse of he above siatement ofthe principle of virtual displacements can therefore be written: ‘A particle i i equilibrium under the action ofa system of forces ifthe virtual work is tera for every independent virtual displacement. ‘Tho above statements apply aso toa system of partes under the ation of 22 equlibened sot of forces, sce the Vital work equations can obviously be writen aneithind every pare. Consideration of suchsstems of pte, origi bis, eats mors caret cxamination othe desxption of te viral placement ‘Comite tor purposes ofa eeamination, tbe seam pictured in Figure 624, The member shown nondstonnabie; its supported in stately deersinate mannze meri ubjcced to's conceteted applica Toad of yy. There would sppear to be no Sa orpoleto of he viral dlacement pace sine nodplacsmentsccms Per tee Sipport orcs can seplace the spport constrains, however, sin Figwre Poh i he condition of equiibum are watten and solved, hes apport forces can te Caltte Then the structure can be represented en Fgute 62, a8 fre body ve Jatahe ston ofan equilibrated force stom without anyconsidertoa of supper ‘Saniags Theseus represented in Pes 620 and 625 ae ell eqaleat an taint of wew of Hla alin, Hence, avira placement can be Ete fo Figure 620 os shown in Figures 62, Since the Dea iid hs viral140 Chapter 6 Vieual Work Principles igure 62 Rigid body analysis of simply supported beam by prnciole of virtual Sispncoments: (a) Actual structure. (@) Steveture with defined support reactions. (¢) Virval displacement splacement must represent rigid body motion, which isa straight line descxibed by ' the equation: w= (-3}m eta 6 : “The virtual dgplacemest principle can now be wsten at i BW ~ Byyb0, + Bata ~ PysBbs = 0 69 and, from Equation 64 with x = 35, i | BY = Fd + Baton ~ P(t = 2) — 3 ' (ox = = 8 a [te mt-4) + (seme ments ar arbitrary, the terms 2, and, can ke on any vals, inlding mero. Tis meas thatthe mips of, and in Eguaton Se must each equi zero, Tae we have 2) mont jr Pars se Application ofthe principle of virtual displacement has produced the two relevant ‘equations of equilibrium ofthis body. If we had introduced just one support foree (F,)GAL_ Principle of Vistsl Displacements Rigid Bodies 141 in place of the support constraint, then the admissible vitual displacement would lead, by means of the principle of virtual displacements, to only one of the above two equi [ibsium equations, On the other hand, ifthe virtual displacement state were generalized to include an axial component, then all three equilibrium equations of this rigid planar structure woutd be produced. ‘Example 6.1 demonstrates that the principle of virtual displacements can be ea ‘ployed to calculate directy the force in a member in the interior of a statically det fninate structure. The fore in the member is represented as acting on the joints atthe tends of the member, but the virtual displacement takes place as if the constraining tetion of the member itself were absent. CCleulate the force in member 3 6of the trssshawe, using the prineple of virtual work. ‘Replace monber 3.6by the equal ad opposite forces Fg acting onthe joints to which the members connected, Assim, for purposes of defining the vstual displaced sot, tht the eontisining action ofthe member is absent. Since the member has been ef {ectively semovedtheough this slep the indicated vstual displacement ean e enforced Fy the prinple of vstaleisplacements, Gu) + ea) ~ Ba Hg od] = Jn summary, the principle of virtual displacements can be employed to establish the equilibrium equations whose solution gives reactions or statically determinate internal member forces. In either case the zegetive or member force is represented, but the Support or constraining action is removed to permit participation ofthe corresponding ‘displacement component in the definition of the virtual displacement state. “The principle of virtual displacements, as an approseh to the construction of rigid body equilibrium equations is of a0 direc value in matrix stnicursl analysis, Blement stiffness formulations include all of the clément degrees of freedom and incorporateny ‘Chapter 6 Vital Work Principles ‘within themselves the element rigid-body equilibrium equations. Nevertheless, the ex- amples given above illstate the most fundamental application of the principle and ‘cmphasize its relationship to the conditions of equilibrium. “Our interest lies in deformable bodies rather than in rigid bodies becauso only by taking into account deformability are we able to calculate the response of statically Indeterminate structures, Therefore, we devote the next section to the extension of the principle of vistual displacements to deformable bodies. 62 PRINCIPLE OF VIRTUAL DISPLACEMENTS--DEFORMABLE BODIES “Application of the principle of vistual displacements to deformable bodies requires thet the total vttual work of a system be clearly defined, which means that distinctions must be made between extemal and internal work. The relationships that describe these terms aze developed inthe following. "To-stablish the desired relationships we examine the simplest arrangement of strac- tural elements, @ pair of axial members connected in series as showa in Figure 6.3 ‘The sight end Goint 3) is fixed, and loads P,, P, are applied to joints 1 and 2. The stiffnesses of the axial members are designated fy and ka. "Figure 63b shows the free body diagrams of the joints and elements. All forces are initially shown as ecting in the direction of the positive x anis. The internal forces ‘xerted by the joints on the members ae designated as Fp, Fay Ay while primes are ‘Smployed to distinguish the action of the internal forces on the joints.‘The force on a joint ond the counterpart member force must be equal aud opposite. Thus Fas -Fh Fy = Fa 6) Fa = ~Fo ‘Also, the condition for equilibrium of element 1 is Fat Bx=0 a ‘or, in view of Equations 6.6, FF (610) “The conditions of equilibrium of joints 1 and 2 are, respectively, P+ Fan0 (68a) Prt Pht Bis (680) ® igure 63 Two-slement axial structure. (a) Actua structure (2) Fre-body iapram of joins nd elements.62 Principle of Virtual Displacements—Deformable Bolles 143, Now consider the effect of virtual displacements &u, and Sy of the joints 1 and 2 ‘Work is a scalar quantity so that the total virtual work ofthe system of joints is simply the sum of the virtual work of the component parts. In this case, therefore, the total vistual work is BW = (B+ PA) + (Pp + Phy + Fe) (Pibia, + PyBie) + (Faby + Phe + Foydt) The first term on the right side of Equation 69 represents the virtual work ofthe applied loads and is designated a5 5Wag, that is Wag = (PrBe, + P,Bus) (610) “The second term on the right side isthe virtual work of the internal forces ‘acting on the joints, By substitution of Equations 64, this erm can be transformed into one expressed in ferms of the forces exerted by the joints on the member, that is (Pho8ty + Phd + Fxbia) = (Fadia + Bydty + Fads) (611) “The expression within the parentheses on the right represents the virtual work ofthe internal forces acting oa the members, or the internal virtual work, 6Wiq,” Heene> in the present ease 69) Wigs > (Fadi + Ryda + Baba) (612) Furthermore, in view of Equations 6.9 through 6.12, forthe conditions under stucy, BW = BW, — 8 By application of the above procedure to any other type of deformable structre, it can be shown that this is a general expression for the total vitual work. To establish the properties possessed by this expression itis only necessary to return to the earlier statement of the total virtual work (Eq. 69) and substitute into the second term on the sight (Le,, “8W,) the equilibrium relationships given by Equations 682 and b. We them have BW = (Py2u, + Pyan) + (-Pyem, ~ Pye) = 0 BW = BW ~ BW yg = 0 (613) ‘This isthe algebraic statement ofthe principle of virtual displacements for deformable structures, Again, although it has been established for a special cas, its validity can. be confirmed by application of the above procedure to any other type of structure It is completely consistent with the statement of the virtual displacements principle of the previous section. Its now necessary to recognize that it encompasses a system of parlices and the virtual work i the total vistual Work, both internal and external. The verbal statement of Equation 6:13 is'as follows: For a deformable structure in equilibrium under the action ofa system of applied {forces, the external virtual work die to an admissible virwal displaced suate i equal to the intemal virtal work due to the same virtual displacements Observe that the qualifying term, admissible, prefaces virtual displaced state. One vould expect that there ace some limitations oa the choices that can be made of the ris quotty may be viewed as energy sored within the member ax rl ofthe vital dslacenerts. ort razon eo eld the vir son eer, 5, designation tt ws used in eit eon seis ton,{ i I i 1 M4 Chapter 6 Victul Work Principles 63 VIRTUAL DISI victual displacements, and itis indeed the case. The term admissible is employed t0 umphasize this consideration. We define these limitations in the next section, where the method followed in using tho virtual displacements principle is outlined, and virtual ‘work expressions are given for the specific modes of structural action with which we are concerned. LACEMENTS ANALYSIS PROCEDURE AND DETAILED EXPRESSIONS 63.1 General Procedure ‘Although the stiffness properties of the elements in Section 62 were defined at the Guts, no se was made of them in the development of the principle of virtua dis placements for deformable bodies. This cicamstance can be used to advantage in Poytain situations, but the use to which the principle of virtual displacements wall be putin this text requires the introduction ofthe structural stifaess properties. They are Fitroduced to transforin the internal vistual work from an expression veitten in terms bf forces and virtual displacements into an expression in terms of displacements ond virtual displacement. "As au example, in the structure of Figure 6.3a, we have the following relationships between the member stiffness and the joint forces and displacements: Ra = a(t ~ 49) Fy = falls Hence, by substitution in Equation 6.12, Wg ~ Hee, ~ Hs )by ~ Kult, ~ te) Bia + Kat ‘This is an expression in terms of the displacements (a, 1:) and the virtuat displace ‘meni (84, 8). Tt is tompting to refer to displacements such as m, and w a5 actual displacements, ‘Allof the developments thus far have cither explicitly dealt with or implied @ proce- ‘ure leading to the solution that is exact within the limits of linear elastic behavior. For these eases the displacements are indeed the actual displacements of the linear syntom. However, tbe principe of virtual displacements finds its most powerful appli ‘ation in the development of approximate solutions. In these, the stato upon which the virtual displacement is imposed may itself be an approximation ofthe actual dis- placed state (sce Section 642). Thus, rather than calling them actual displacements, fve will use the adjective real—ia its sense of something rooted in natuxe—to distin- guish between tangible and virtual displacements ~ "The character of both the ral nal virtual displacements, and the Timitations imposed upon them, deserve careful attention. As emphasized previously, the solution (0 a problem in structural mechanics is exatt ift meets all relevant conditions of equlib- Finn and ciaplacement continuity, ‘The foregoing development of the principle of vir- tual displacements gives explicit atfention only to the condition of equilibrium. The ‘Condition of displacement continuity, which requires thatthe displacements are con- tinoous functions of the coordinates of tho structure (x,y, 2), must be met implicitly. ‘Thats the real ané virtual displacements must be of form such that the displacement continuity conditions are satisfied from the outset? Admissible Uisplaced states are those dhit meet these conditions. arabe be ated tht hs equlecent til permits e vita isplacoment of suppor and joints win Fone 2 sad Bente, povided that the sgpocs ae represented by supprtestion frees rather than by the atl consti agalat plement63 Virtual Displacements Analysis Procedure and Detiled Expressions 14S With the above in mind, en outline can be given ofa general procedure of structural analysis based on the principle of vistual displacements. One begins with a description (of the real displaced state by means of admissible functions that have undetermined ‘nultipiers. The virtue! displaced state ie simibarly described, except that the chosen ‘admissible functions have arbitrary multipliers (eg. Ou and ug in the development above). These functions are, in turn, used in the construction ofthe terms 6W7,q and ‘Vu Equating 6Waq and SW prodiices conditions that enable the evaluation ofthe ‘ondetetmined mollipiers-of the real displaced state. These values are such that the ‘conditions of equiibriom are satisfied. Te will be found that the arbitrary multipiers Inay be canceled from the resulting expressious or that they may lead to condit ons that can only be satisfied by certain relationships among the seal quantities of the problem, “The tasks that remein before this general procedure can be applied to specie rob: Jems are the formulation of explicit expressions for the internal virtual work BW. and the external virtual work 8W,q for the various modes of structural action, 632. Internal Virtual Work “The primary modes of structural ation with which we are concerned are axial, tor sional, and flexural behavior. Inthe following discussion, we develop relationships for '5Wj for members of length L sustaining these modes of behavior. ‘Consider fst an axial member, a differential segment of which is shown in Figure 6a. The stiess o, that acts upon the segment throughout the vistual displacement is the stress corzesponding to the real displacement. The virtual displacement of the left tend ofthe segment is designated as 6, The virtual displacement will in general, change from one point to the next along the segment so that at its right en it will be, tothe firs order in de, Bu + ((Bu)idx}e “To demonsiate the relationship between virtual displacements and virtual stains ‘we will ist treat the differential segment as 2 free body, The real forces at its e ° Figute 64 Vistual stain conditions. (2) Axial behavior (B) Torsional behavior. (c) Fexurl behavior145 Chapter 6 Vit Work Principles are Fs, with @ negative sign assigned to the force on the let face, since it acts in a direction opposite to the positive x direction. The vistual work ofthe foree on the left face is therefore —F,61and that ofthat force on the sight face is Pu + [d(@u)de]dx) ‘Thus for the segment as a free body, BWq = —8uF, + [Qu +b (a(Suyldxlde)F, = [aCGuplds) Feds and, since 5Wae = SWin from Equation 6.13, = 8 ow [Nov axial strain is defined asthe rato of change ofthe axial displacement with respect to the axial coordinate; that is, for linear behavior e, ~ duldx. In the present ease. ‘we have a virtual axial displacement 84. Defining the virtual strain de, in an analo- fous manner, we have 8, = d(Bu)ids. Also, F. = oA. Thus, Equation 6:14a can be swritton as Wg = Seon dx (1m) 1 segment dx, For a complete axial member I Virtual work due to virtual displace ‘This relationship pertains to a different ‘of length L, it follows thatthe change in inter ‘ment is ote ~ [ea (59 Finally, by use of Hooke’s law (a, Be,), this ean be written entirely in terms of strain si [faathea 9) Since the real and virtual strains are the desivatives of the real and virtual displace- iments [e, = duds, 8, = dds}, the intesaal virtual work can aso be written a8 fenetion only of the real displacements and virtua displacements ff [Aeon tt (9 Wn ‘This means that the internal work can be calculated if expressions are available for the real and virtual displacements, as will be the case in the work to follow: Tn the case of torsion, the “strain” is the rate of change of rotation of the crass section about the lougitudinal axis (see Section 4.5.2). If, in the differential segment of a torsion element, Figure 6.4, the eft face undergoes rotational displacement 8, the right face is displaced @, + (20,2), and the rate of change of rotational dis L 6, dO, p-E[(nethas) 9] wo 4(50,) os on ‘The change in internal vitual work due to a virtual twisting displacement in the presence ofa real tvisting moment M, is, in an element of lengtt L, ate ~ [epic = 82 aa «an64 Vist! Displacements Analysis Procedure and Detaled Expressions 147 cand, since M, = GI(d0dds), this can be written entirely in terms of displacement qualities fag = f° MO8D gy Me Hh = [C8 Gy Be de (6:88) In examining beam flexure we consider a member whose axis is coincident withthe seas, has a eross section symimetrie about the y axis, and is subjected to bending about the z axis. As explained in Section 4.5.3, the curvature dvidx’ = x can be regarded ‘as the strain for bending behavior. Figute 6.4c, which describes the displacements ot ‘differential sogment of a flexural member, is further illustration of this point. The eformation of the segment ie characterized by the rale of change of rotation 6, with respect to the x coordinate, The rotation ofthe cross section is equal to the slope of, the neutral axis, 8, = dofds.* ‘The rate of change of 8, with respect to x is, then llersea)-e]- 2 $e) 2] Bian om {It follows that the virtual bending strain 8x; is, for the virtual transverse displace- meat &, x, = LOD) (620) é 1 was alo explained in Section 4.53 thatthe “ses” im laure isthe bending moment Ma. Tecefore, in an element of length, the internal virtual work Gu (2 Cita! dipiceaent othe eal embers ax ie =f egiatgar » [HP mae (2) ‘Abo, since the moment is related to the curvature by M, = Ble{dulds?) {see Bq, 431b), this ean be written as at atu ‘A comment iin ordes regarding the conditions of continuity in lexus The tas vere daplacements, v and do, mst of eaure be continuo. The flexural deformation ats ao characterized by the rotations @ and 88, which, a8 we have noted, are agua tothe slopes ofthe eesti ani, dud and d( ide sce, in order forthe eal Sai vial dgplacement sates to be admisble, they must also possess continuous fiat Gerivatives Furtheamore, the displacement bousdary conditions in flestxe ia given problem ight invelve not ony » and By but sso dofds and ae). is uefa it the work that follows to have available @ genera fouls for applicable to any oral ofthe above cases and slo to problems of structural mechaniss eeia the sate of tess might be (wo: er threedimensional, A general designation Di the stat of stress at point in a structure is inthe form of column vector (a, Sich include ll compontats of tes that are presot. For a hree-imensional tate stan for example, we ave fo] Loy Oty Tye. Teal" Silay, the vice {hal stains can be designated bya row voctor| Se |which for three-dimensiona tate Data is| de, 8, es Bry Phe P| Now the internal virtue work peruuit Solum of &multiayial stvess sat ir pivem by Be," + 3,0 + oxy Which cen Be 1 ade (62a) “Aa ote ia Chapter 4 ti staterent ips neplet of deformation duet ansverse shear© capi Yemen Pane ren fo Fora sete who volumes yo b(t tea wa eee! ara = [setuid ox | ‘The symbolism of Equation 6.22 is consistent with each of the definitions given above Coa seed saafetons foregut recone [delete opnd a Aah ein sn [tea dente “The principle of virtual displacements in the formulation of stiffness equations re- {quires that dWiq, be expressed solely in terms of strain. For this purpose one can introduce the stress-strain relationships (= [B]fe) into Equation 6.22, resulting in Wt ~ [., LekEITeetvo (623) where [8] is the matrix of elastic constants. For framework elements [B] is merely the scalar value B, of in the case of torsion, £/2(1 + 0). 633. External Virtual Work ‘The calculation of the external virtual work 8Wax is 9 simple matter in the case of | concentrated Joads and can be represented symbolically as 2 oar where 64; refers to the virtual displacement ofthe ith degree of freedom and P; isthe Applied load in that degree of freedom, ina situation in which loads are applied to s egrees of freedom, For applied moments, of course, the virtual displacement isan. ‘angutor displacement 58, (0: 88,,or 88,) and the applied loads are moments, Png (OF Pray Of Pc) TNote especially those cates ia which the virtual displacements include rigid body motion to enable the constriction of equations for determination of support seaetious. “This wes illustated in Section 6.1 and wil again be illustrated in the next seetion. In such cases, the reaction at the support that is released to enable rigid-body virtual displacement must be designated and included in the expression for 3. Distributed load eases alzo deserve attention. Denoting the distributed loads as q and combining with the virtual work of concentrated loads, ath = J 88 gas +B (DP, (624) ‘where the limits ofthe iategsal are taken to define the portion ofthe member on which the load acts, Use of Equation 624 is illustrated below in Fxample 6.5 for a case of flexuce, Its of central importance in the exploitation of virtual work concepts in the formulation of fixed ené forces described in Section 75. ‘It was observed in Section 6:2 that work (and therefore energy) expressions produce sealer quantities. Thus, whea an element is subjected to the combined action of axial Toad, forsion, and flextze, the change in interaal virtual work due to virtual displsce. tent i simply the stam of the BWi,’s for the respective actions. Correspondingly, the ‘change in external virtual work is simply the sum of SW. forthe component external ‘actions. This point is brought out in Example 6:2. i | |