A FIRST COURSE IN CONTINUUM MECHANICSThird Edition A FIRST COURSE IN CONTINUUM MECHANICS for Physical and Biological Engineers and Scientists —— { Y. C. Fung Profesor of Applied Mechaxcs end Bening University of California, San Diego PRENTICE HALL, Englewood Clits, New Jersey 07632vary of Congres Catia Polain Dita Fang, ¥.C.(Yumeben) "Af coue in conn means: for py nd ‘loop eens en engineers Y.C. Fang, 3 Bom ies ibogrpi! efeenes a ind. ISBN Orsa6Is02 1. Coninoom metas, 1. Tie (anata 1551 sia nsei0 Acquistion tr Dou Haspey Prodcion er Bayi Mentor Lzon Cove deg: oe DiDomenio (Copy ed: Bran Baker Pras Byer Linda Berens Marafcarig bayer: Dave Dickey ‘Silene er les Dworkin Bala asst Sin Hai SE 019,197,180 by Prentice t,t, A Sion &Sehuser Company Englewood Cit, New Je 165. Allright revere No part of his oak my be repro in any foro bya eas, ‘tot perio in wig the pois Print inthe Unie States of Anes ISBN O-13-Ob1524-2 woeTésaszi 15264 ISBN O-13-Ob1524-2 straots Pre Hal Intron (UI) Limited, Landon Pree Hal of Asal Py Line, Syne Prentice Hal Cina ae, Toone Pron Hal Hipananericn, SA. Meso Prentice al fea Pvt int, Nor Deli rence Halo Jpe, Ine, Tojo Sinan & Shar Ass ei, Singapore "Bora Prete Hal do Bis a, Ro de onto a Dedicated to students who would share my enthusiasm {for the application of mechanics, and t0 ‘Luna, Conrad, and Brenda. oe ————Contents a Preface to the Third Edition Preface to the First Elton 41 Introduetion 1 The objective ofthis course 1 1.2 Applications to seience and technology 2 13. What is mechanics? 2 {14 Aprototye ofa continuum: The casal deiiton 2 15 Our definition ofa continuum 3 116 The concep of tres in our defiton ofa coninuum 4 4.7 Abstract copy of a1ealcontinoum 5 18 What ie continnum mechanics atout? 6 19 Axioms of continoum mechanict 7 1.10 A biological example of hier of continua depending on dhe size ofthe object involed in scent inquiry 7 1.11 lemeatary topic through whic basi ideas evohed 22cl Contents Vectors and Tensors. 39 21 esas 22 Vetorosis 4 23 Thesumtonconenion 24 Tindaon ation wordnass 25 Comte tannin ine 5 2 Anda esto ec et, Cates 27 Repiianeot oars aos ts an ess Balle orn? 29 Quotient rule 59. * * 20 Paral dvs @ ‘Stress oa MI Titastanes 4 32 Tectia 8 23 Guiysomin 6 34. Egutinsel tron 2 35 Chgeestes competi barman a 36 Ss copes ion cue ort ‘3.7 Stress boundary conditions. © 78 * Principal Stresses and Principal Axes 88 41 Introduction 88 42 Plane state of stress 89 43° Mol’s circle fr plane stress 92 44 Mohr’ citls for thnee-dimensional stress states 94 45, Principal stresses 94 46° Shearing stresses 97 47 Suress-deviation tensor 99 48° Lamé’s stress elipsoid 102 Analysie of Deformation 12 541 Deformation 112 52 Thestrain 115 53 Strain components in terms of displacements 117 54 Geometric interpretation of infiuitesimal strain components 119 55. Infinitesimal rotation 124 5.6 Finite strain components 122 5.7 Principal strains: Mobr’s cielo 124 Contents ‘58 Infntesimal strain components in polar coordinates 125, ‘39 Direct derivation of the strain-dsplacement relations in polar coordinates 128 5.10 Other strain measures 131 Velocity Fields and Compatibility Conditions 145 6.1 Velocity fields 145 62. The compatibility contition 146 63. Compaiblity of strain components in three dimensions 148 Constitutive Equations 154 ‘TA. Specifiction of the properties of materials 154 72 ‘The noniseous uid 155 73 Newtonian uid 156 74 Hlookean elastic solid 157 75 Bfect of temperature 161 71 Materials with more complex mechanical behavior 161 Isotropy 165 8.1 The concept of material isotropy 165 82 Isotropic tensor 165, 83 Isotropic tensors of rank 3 169 84 Isotropic tensors of rank 4170 85. sotropic materials 172 86 Coincidence of principal axes of stress and of strain 172 87 Other methods of characterizing isotropy 173 88 Can we recognize a materia’ isotropy from the nicrostructure? 173 Mechanical Properties of Real Fiuids and Solids 181 94° Fluids 181 92. Viscosity 183 93 Plasticity of metals — 186 ‘94 Materials with nonlinear elasticity 188 9.5 Nonlinear stress strain relationships of rubber and biological tissues 19 9.6 Linear viscoelastic bodies 193 9:7 Quasitinear viscoelasticity of biological tissues 197 98 Not-Newonian fuids 201 9.9 Viscoplastic materials 202 9.10 Sol-gel transformation and thixotropy 2041 2 B Derivation of Field Equations 10.1 Gauss’s theorem 209 10.2. Material description ofthe motion of acontinoum 212 10.3. Spatial description ofthe motion of a continuum 214 10.4 The material derivative of a velume integral 215 10.5 The equation of continuity 217 10.6 The equations of motion 218 10.7 Moment of momentum 219 108 The balance of energy 220 10.9 The equations of motion and continuity in polar coordinates 223, Field Equations and Boundary Conditions in Fluid Mechanics ILL The Navier-Stokes equations 231 11.2. Boundary condition ata solid‘luid interface 233, 113 Surface tension and the boundary conditions at an interface between two fids 235 14 Dynamic similarity and Reynods number 238 115 Laminar Pow in a horizontal cannel or tube 240 11.6 Boundary layer 244 11.7 Laminar boundary layer over «flat plate 247 118 Nonviscous fuid 249 11.9. Vortcity and circulation 251 11,10 Irotational ow 253, LIL Compressible noaviscous fis 254 11.12 Subsonic and supersonic low 257 11.13 Applications to biology 265 Some Simple Problems in Etesticity 12.1. Basie equations of elasticity for homogencous, isotropic bodies 270 122 Plane elastic waves 272 123. Simplifications 274 124 Torsion ofa circular ylindtial shaft 274 25 Beas 218 12.6 Biomechanics 281 Stress, Strain, and Active Remodeling of Structures 13. Introduction 285, 182 How to discover the zero-strest state of material in a solid body 285 Contents 231 270 Contents 23 B4 BS B6 BI BR ‘Remodeling the zerostres state of structure: A biological example of active remodeling due to change ia stress 288 ‘Change of zero-stress state with temperature: Material that “remem ber” their shapes 290 Morphological and strctural modi of blood vessel due to change in blood pressure 292 Remodeling of mechanical propertis 294. Sires analysis wth the zoro-stes sate taken into acount 296 Stese-rowth relationship 259 302-Theobentve of tisedion the same ath ofthe ai AS To emphasize cae maton of problems in mechanics, to reduce vague HT precise wre ematical statements, and to cla a habit of queso analyzing, design ing and inventing in engineering and sec. 1 ave stressed applications even ins aongy inthis eto. Ths, ate very Penn continua are defined With regard 1o real materia. Thoughcst the boa, no discrimination was made Again biological materials, Biology included in Soe» bioengineering is ariel in engnoeing, Mechanics ot Kite to physic. Tr hs book, [often ask the rae to formulate equations regardless of ats Peon ec solve them ort. Tv knowe many Sader ‘wo ave i erable exercises wouter formulating problem theie own. 1 op that they wir things the ter way f Enews mS problems oftheir ne ihe aie to dzver the netbods of soon 28 subtleties of the oon and they are encouraged to serve nar and 10 Ck of problems in > ‘Acourse incakydom mechanics then provides 2 foundation for suis ‘uid and solid in ry materl scenes, and ote branches of seen a SENET risa opiion ta, for a egies the approach sould Pe physical rather than mathematical, To engineers and physics who tse coninnurt mechanics ‘oat ly, the primary attraction fhe subj is in its iy cof conception cot apeteness in applications, Therefore the stuns shoud introduced to the applications as soon as possible. Poni ox egier, he important questions he must Gin ante to ae How shal [formulate the problem? How shell state = ‘governing field equations and boundary conditions? How hall choose alternate hypotheses? What aati iments woud july or deny improve my hypotiese How exaus: Xin the ivesigalion be? Whee might eos appear? How mich time is feud to obtain a easonale soon? At what co? hes qoestions which re ae investigator, and are questions of sates, which sr analyses ome “Gpmlte anne 1 these question ate peyon th pe fe “Gest wo ate ean male a god began, In this ook, ten a the reader ‘erate problems, regardless of heer he can solve is aor and under- {oor material xls, have known many students who Be read smaay books and worked innumerable exerises without ever formulating problem sree own, 1 hope they wil Iam the other way, to generale mony problems of sean ard the sive to discover the methods andsbdeties of sation ‘Theywi Profaco tothe Fis Elton should be encouraged to observe nature an to think of problems in engineering and then to take the first step to write down a posible set of governing equations 1nd boundary conditions. This “first step”—to derive the basic governing equa- tions—is the object of this book. Pethaps it sjustiiabe for a “fist course” to be concerned only with this first step. But the preparation required for taking tis step is extensive, For such a step to be firm, one would have to understand the basic concops of mechanis and their mathematical expressions. To be able to use these basic equations with confidence one must know their origin and their der ivations. Therefore, the discussions of basic ideas must he thorough. Its for this. reason thatthe first ten chapters of this book ae rather comprehensive and detailed. {As forthe organization ofthe book: At the outst, the concept of continua isexpleined. Then a thorough treatment of the canceps of stress and strain follows. ‘The practical technigues of determining the principal stress and strain, and the concept of compatibility, ere given emphasis in two separate chapters. The descrip- tion of motion is considered. In Chapter 7, cn idealized specification of fuids and Solis is presented, The important concept of isotropy is described in detail in (Chapter 8, Data onthe mechanical propertts of common Mud and solids appear in Chapter 9. In Chapter 10, a thorough testment ofthe basic conservation laws of physics is given, Boginning with Chapter 11, some features of perfect fuids, viscous flow, boundary layer theory, lineaized theory of elasticity, theories of bending and torsion, and elastic waves are described briefly. The last two chapters provide a glimpse into the rch fields of fd and solid mechanics; to teat them comprehensively would require many volures at @ more advanced mathematical level. Te introduction given here should prepa the student to enter thes fields with greater eas. If the reader obtains clear ideas about the stress, stn, and constitutive ‘equations from this book, I would consider tis introductory text a success. Beyond this, only sketch of some classical problems is provided. Many discussions are sven inthe exercises, which should be regarded as an integra part of the txt. T have quoted frequently and borrowed heavily from my previous book, Foundations of Solid Mechanics, which can be used for a eourse following the present one. The material fr this “first couse” was organized for my clas t the University of California, San Diego, where the curriculum offers emphasis on ‘general sciences before specialization. The book shouldbe useful for undergrad: ‘ates and younger graduate students who hove a reasonable background in math- ‘ematies andl physics ‘Tho writing ofthis book wasa pleasant experience. My wie, Luna, cooperated throughout the task. A mathematician, she gave up her teaching career when T ‘came to La Jolla. Wiling to learn some meckanis, she worked through the man- script very thoroughly. Many passages are clearer because of her decaration that she did not understand. My friend, Chi-Shun Yih, Timoshenko Professor at the ‘University of Michigan, reed through the manuscript and gave me many valuable ‘comments. Iam alo grateful to Drs. Pin Tong ofthe Massachusetts Institute of ‘Technology and Gilbert Hegemir ofthe University of California, San Diego, for tc tet ton Pra to Nicholas Romanelli of ter my thanks 1 wish to eter ating the index, Sie comments Fi ee ine Li or prep Soa fox otra = Pri Hl eo peg et ENN Sine woke pease 6 fig La Jolla, CaliforniaA FIRST COURSE IN CONTINUUM MECHANICS1 INTRODUCTION INIkop Ses Te defntion of conn or edn mater i peed eee ey examples trough which basi idea of mechanics evolve. 1.4 THE OBJECTIVE OF THIS COURSE ‘our objective is wo lean how 1 formate problems in mechanis and how to oa questions and ideas into pee material tatemen' a Woh res ae aio questing, anayng, desing, and iveting in ear neering and science. Te ora few questions. Sppose a panes ying above 8. The wing mst be under stain in nr to support he Pang 00 freight. How wigs Te are the wings subjected of you were Shing lier, m= anvil sy appeared, the thermal coment woud cry the eat Higher, Dove fly la arsed Have the wings sufcent sng? Acad yu se the Gated it ge cables suppor treendovs lod. How Goes one es such ote cloud contain wate andthe ouside needs that water, Ne cloud ee oul hat produce ran? And would the in fl where ned? Me pe aount of anal be adequte and not produce fod Inte ine, Pe gla eactr power station. How isthe est anspor ip the reactor? ‘ins ind of thermal reses are ther inthe reactor? How does ope SS the Wty ofthe power aon agin eariuakes What happens 2 arth in an sa) ake? Thinking aout he lobe. you may wonder pow ths pene float, carte apart. Ant how about oss: How dowe breathe? Wha hanes {ave place in ur hngs if we do yogt exerci and stand or 99° heads? Pe Manly, al these questions ae concerned with fre, moon, 0%, deformation, energy, propertis of mater, external inteacion Datwesy bodies, oF Seton between on part ody and anther part, nd chase aren tamporerly or permanent, reversibly ot ever. Ts changes, toga withthe axioms of coatinuum mechanics ca Be reduced to certain dif- rasa equations and boundary conditions. By solving such equatiTs obtain presi quantitative information. In hs book, we del wth he fundamental prin 12 Introduction Chop. ciples that underlie these differential equations and boundary conditions. Although it would be nice to solve these equations once they are formulated, we shall not ‘become involved in discussing their solutionsin detail, Our objective is formulation: the formal reduction of general ideas to a mathematical form. ‘12 APPLICATIONS TO SCIENCE AND TECHNOLOGY “The mathematical approach taken inthis book willbe aimed at serving science and technology. I want the applications to be apparent tothe student; hence, the examples and the problems to be solved are often stated in terms of scientific research or engineering design, A person’ frame of mind with regard to designing ‘and inventing things, devices, methods theories, and experiments can be strength ceed by constant practce—by forming a bit. 13 WHAT IS MECHANICS? “Mechanics isthe study of the motor. (or equilibrium) of matter and the fores that cause such motion (or equilibrium) Mechanics is based on the concepts of time, space, force, energy, and matter. A knowledge of mechanics is needed for the study of all branches of physics, chemistry, biology, and engineering. 14 A PROTOTYPE OF A CONTINUUM: THE CLASSICAL DEFINITION. ‘The classical concept ofa continuum is derived from mathematics. We say that the real number system isa continuum. Between any two distinct real numbers there is another distinct real number, and therefore, there are ininitely many real rnumiers between any two distinct eal numbers. Intuitively, we fee that time can be represented by a real number sstem ¢ and that athree-dimensional space can be represented by three rel number systems x,y, 2. Thus, we identity de and space together as a four-dimensional continuum. Extending the concept of a continuum to matter, we speak ofa continuous distibution of matter in epace. This may be best listrated hy considering the concept of density. Let the amount of matter be measured by its mas, and let us sssume that a certain mater permeates a certain space Ve, asin Fig. 1.1. Let us consider a point P in Ys and a sequence of subspaces Y, Vo, .. converging on P: CVn FET, (012-0), aH) Let the volume of Y. be V, and the mass of the matter contained in Y, be M. We form the ratio M/W. Then ifthe limit of MJ/V, exists as n> & and V, > 0, the Se. 15 Our Definition of @ Continuum LAA sequence of spatial omtias converging on ing valu is detind asthe deny of he mas dsbucon ate pin P and is denoted by p(P): ay) ti a crc “cr in ig Stiga a oer at nm fe en ce a ep no cell aggregate can either. 115 OUR DEFINITION OF A CONTINUUM ch dea mae acpinnin sro cal rs eg nee ioe pee te tered ot oo SOG see stl i Tp ube tad ve mihi lnm mt lr ey fe mnt oer iP sec fen Fo uscpaneet bcs on. Yaw in To chat none nal eng‘ Introduction Chap. t P. As n— «the lini of V, tends ta finite postive number w, Let the mass of the material enclose in be My. The sequence of the ratios My; i said to have a limit p with an acceptable variability «if as n> «, The quantity pis then said to be the density ofthe material at P with an ‘acceptable variability ¢ ina defining lit volume ‘We define the momentum of th: material particles per unit volume and the nergy per unit volume similarly, cach asorated with an acceptable variability and a defining volume. Later (See Sec. 1.6), we shall deal with the free acting on surface ofa material body, and it would be necessary to consider whether a limit ‘of force per unit area exists at any point onthe surface with an acceptable variability ina defining limit ara, Ifit does exis, then the limits called the traction or ste, and the collective entity of traction in every orientation ofthe susface i called the ses tensor. Further, in Chap 5 weshall consider the change of spacing between particles and define the strain tensor. The existence of strain components willbe ‘associated with an acceptable variabilty and a defining Limit length If, with a clear understanding cf acceptable variables and defining limit lengths, areas, and volumes, the density, momentum, energy, stress, and strain ‘ean be defined at every point nthe space, andi they are all continuovs functions ‘of spatial coordinates in Vo, then we say thatthe material in Vs isa continuum, fa material is a continuum in the classical sense, then it aso a continuum {in our sense. Fora classical continuum the acceptable variability andthe defining ‘init length, area, and volume are zo. In other books on continuum mechanics, the authors say or imply that to decide whether continuum mechanics is applicable to science and technology is a ‘matter forthe experimenters in each dscipne to decide. I say instead tht every experimenter knows that the classical taeory does not apply hence its the respon- sibility of the theorist o refine the theay to ft the real world. Our approach does fitmany field of scence and technology; the need to specify ecceptablevarabilites and defining dimensions isthe price we pay. ‘U6 THE CONCEPT OF STRESS IN OUR DEFINITION OF A CONTINUUM Consider a materat B occupying a spatial region V (Fg. 1.2) Imagine a closed surface $ within B, and consider the interaction between the material outside $ and that inthe interior. Let AS be a sal surface element on S. Let us draw, fom 4 point on AS, a unit vector v normal to AS, wih its direction pointing outward ‘rom the interior of S. Then we can cistinguish the two sides of AS according to the direction of v. Let the side to which this normal vetor points be called the Positive side, Consider the part of material lying on the positive side. This part exerts a foree AF on the other part, which is situated on the negative side ofthe Sec. 1.7 Abstract Copy ofa Real Contintum 5 igure 1.2 Stress principle sm The es AP depen on aon ni he eh Mt of te somal Welnds than ho 85 en oma ‘athe ea eat AS tdsos dfn RUS thon ple vty on he monet of eo go ene Sabet any Pt nts he ofa bt nel ean pl aly. The ning eri ten ire he supervising ode the dest of he normal» i Ting er sl ie rain ot oer et res eee et ei ra ting cot is "=P erin tat here dened pon ay imged ced sua Sin the indir ofseomnum ase vets Bk whose etn on tena o=cuPyig thse nr Segoe oe acon teeter mately tune pnp of Ele ond avy. We ep his pine 8 0 fovevert eno morta bssnpifeaton,Forezamp, hes pra ttestn ny th meacon ofthe net o he vo sie fhe sae AAS eu be manent Inet ome people we dost isthe ese then hat th monet oh floes gon these AS aba a pit inthe ee he i ave propo een of the Ss np Er anf Cay, ume cosy sl sue ema na vr te aon of he ete mae upon the nor equipolet 1 stand couple Te eu they seth ot of ope srs a 1 mre cml tn te cmentna they. 80 ft 0 elation fare fud fore couplets tea, hes, we hallo eos ut ints boot 117 ABSTRACT COPY OF A REAL CONTINUUM Cnc is desded that mati bdy canbe ead conus, oe cn tat hat copy fhe el mati according oh si tion ‘atopy opi wi hel umber ns es ab alton6 Introduction Chap. 1 ofthe rel material, The rues ofiealiztion areas follows: The mass densiy of thedeaized systems the same a the real one in the range ofits definition. When tet of forse applied to both the real material and the abstract copy, the stress fndstrain ofthe two stems are e same, excep tha the calel ofthe deaized Som ean be cred out goo}, wheeas hat of he eel material would have 2 limitation onthe lower bound of sizes anda statistical variability that mus be ualuted. The constitutive eqution ofthe real materi used to dserie the fechaieal properties ofthe idedized copy. The real tem sais the equations hates (ecoqalbrom),coaticty and blanc of energy ofthe idealized system with eros whose Bounds can be calculated. Often, library fll of results con- Teming the abstract copy exits and can be borowed. The known, acceptable “rnb and defining dmensixs wl allow sto evaluate the ference between the real material and the abstraction and let ws know something bout the real system, “The constutve equation of an abstract copy ofa ral materi in a certain range of ses may difr from the consiutive equation of another copy of the ‘am material ina diffrent range of sizes. Ihe constitutive equations of the real tateril were the same for conscuie ranges of sis, then the abstract copy has tne constitutive equation inthe total range. If the consttative equations of the eal mattal ae diferent n diferent ranges of size, then the abstract copies ia Trotsive ranges of sizes have diferent constitutive equations. This is indeed a tsefu feature of our system, because it allows us to recognize diferent structures ofan object at diferent dimensions of observation, to ask diffrent questions at itferent levels of sizes, and to ain a better understanding ofthe whole “The hierarchy ofthe consittve equations at diferent dimensions of obser- vation ofa material is related tthe sinrity or dsimiary ofthe structure of the material at diferent dimensions, The structure of « material may be fac, ie selesimilas in successive rages of sizes; ort may nt be facial, For example, the geomet pattem of the aivays of the lung, (rom the largest bronei tothe Smallest bronchiole, is fractal in structure, so the bronchi in this rane of sizes fan be expested to bey the sane constitutive equation, The structural patter of the alveolar ducts, from the respiratory bronco othe alveolar sacs, is another fractal; hence, different consitutive equation is expected to hold inthis range. “Te pulmonary alveolthe areolar wall—are not fatal at al nether are the callagen and elastin fiber in th abvenar walls. Hene, their mechanical properties ood ently diferent desciptos 1.8 WHAT IS CONTINUUM MECHANICS ABOUT? ‘We shall calla continuum endosed ina closed surface @ body. The surface my be real, like the skin of a man or the shell of an airplane, But it may also be imaginary, visualized to enclose a bt of space Roel woxld material objets are subjected to forees acting on their bodies (euchs gravitational and electromagnet ores nd forces acing on ther surfaces SSec,1.10 A Blologcal Example of aHierarchy of Continua 7 (euch as atmospheric prestue, wind and ai, burdens tobe carted, and oad 0 be transmitted to a remote place), I the body isa continuum in the sense described in the preceding section, then we woald want to know how the materi inthe brody racist the exteral force, The detemination ofthe internal condition of 2 body in response to external fore 's whet continuum mechanics about 119 AXIOMS OF CONTINUUM MECHANICS “The anious of physics are taken as th axioms of continuum mechanics. In pa- ticalar, we use Newton's lavs of moion and the frst and second laws of ther- rmodyniamics inthis book. "There are thee additional axioms of continuum mechanics. First, « material continu remains a continuum under the action of forces. Hence, two particles that are neighbors at one time remain neighbors a all imes. We do allow bodies to be breakable (ce. they can be fractured); but the surfaces of fracture must be identified as nowly created enteral surfaces. In ving bodies, we allow new growth (ee. cellar or extracellular mes increased, new cells migrated into an are of the body, or proliferated from existing cels by division) and resorption (cellar fr extracellular mass reduced, eels migrated out of an area of the body, or cell ‘ied and subsequently washed away by blood, or absorbed in issue), Each newly aulded or resorbed call ereates a new surface inthe body. “The second axiom of contingum mechanics is that tress, as described in Sec. 1.6, and stain, as described in Chap. 5, can be defined everywhere in the body. ‘The third axiom of continuum mechanics i tat the stress at & point is related to the stain andthe rate of change of train with respect to time at the same point. ‘This axiom i a great simplijing assumption. I asserts thatthe stress at any point in the body depends only on the deformation in the immediate neighborhood of that poin."This stessstrain relationship may be influenced by other parameters, toch a fomperature, electric charges, nerve impulses, muscle contraction, fon transport, e., but these influences san be studied separately. 1:10 A BIOLOGICAL EXAMPLE OF A HIERARCHY OF CONTINUA. DEPENDING ON THE SIZE OF THE OBJECT INVOLVED IN A SCIENTIFIC INQUIRY We ave familiar with telescopic views of the sky and microscopic views of eels tissues, metals, and ceramics. As scales of observation change, the object appears diferent. An example is the humax lung. Figure 1.3 shows that the lung may be considered to be composed ofthec trees: an airway tee, an arterial tres, and a enous tee, The airway te fr ventilation. The trachea is divided into bronchi, “Which enter the lang and subdivide repeatedly (in humans, 23 times, statistically Speaking) into miller and smaller branches and, ally, nto the smallest nits, felled pulmonary alveoli, Figure 14 shows a photograph of human alveob, 3 seem8 Introduction Chap. 1 he igure 1.3 A conceptual sketch showing the ung as composed of thre tres: the airway te» (raches-bronchi-bronchioleralveoer ducs-lvel), shown on te lf; the palmonary arterial ree (atres— artrioles-capilaies), shown at the eeater, andthe pulmonary venous tue (venles-sens-eft atu), shown onthe ight. Taal hight on the order of 40cm. in a microscope. The photo covers a small area within the fitle cre on the left border of the leftmost drawing in Fig. 1.3. The wall of the alveoli are capillary blood vessels. Every wall of each alveolus is exposed to ges on both sides. The pulmonary artery aso bifurcates again and again uni it becomes capillary blood vessels that lie in the alveolar wall. The venous tree begins with the capillaries in the alveolar wall. The veins converge repeatedly until they become pulmonary ‘eins, which enter the left atrium ofthe heart. The main function of the lang takes place in the alveoli. The venous lood takes up oxygea from the ga in the alveoli and releases carbon dioxide tothe alveolar gas. The gas exchange takes place across ‘the alveolar wall. Figure 1.5 shows a micrograph of the alveolar walls enclosed in 4 small circle on the lft border of Fig. 1.4. Figure 1.6 shows the collagen fibers inthe alveolar walls. The collagen fibers were stained wit silver and appear as black bundles. The collagen fibers are formed by ibis. Fibils are formed by collagen molecules. One could goon to smaller hierarchies consisting of molecules, ‘atoms, nucle, and quarks. Depending on what prope-ty of the lung we wish to investigate, we may consider the lung a a continuum at sucessve hiorarcies. For example, if one is interested in comparing the diference in strain in the upper part of the human hung from that in the lower par, then the individual alveoli can be considered infinitesimal and one can speak of deformation averaged over volumes that are large compared with the volume af a single alveolus, but small compared with the ‘whole lung Such an approximation would be appropriate in studying the interaction jag. Silver-stsined, thick ing sketched in Fig. 1.3 re on the left border fon of ling tissue, Black lines, Aiveclar walls: From Matsuda, M., Fong, ¥- ‘Bhstin Fibers in Fluman Puleionsty Alveolar Mouths snd Due Phynistogy 63): 1185-1194, 1987, Reproduced by permission. Figure 1.4 A magnified view of a tle tip of the rE within the smal G36 pt) se0 Introduction Chap. 1 Figure 5 A magnified view of the alveolar walls enclose in a small cle on the left border of ig. 1.4 showing the capillary blood vessels in he walls as called he iateralveolar septa) A = aleoir gas space. between the lung and the ches wal, the distribution of pleural pressure, or the distribution of ventilation inthe whole lng (because ventilation is elated toalveo- lar siz, which s proportional ty the strin) For these problems, a hierarchy with minimum dimension on the order of 1m wll suffice. On the other hand, if one is interested inthe blood flow in the large pulmonary artery, then the blood can be regarded asa homogeneous fd and the biood vessel wall canbe regarded as ‘continuum, witha minimum dimension forthe definition of tress and strain on the order of 10 um. If oe is iteested insted inthe stres in a single alveolar wall (with 9 plane area onthe order of 100 x 100 pom anda thickness of about 10 um), then even the icv eollagen and eatin fhers inthe wall exmnot he ignored, and the wall must be considered a compost structure made of several diffrent materials and constnte in a special way. What kinds of averages are _seful depend on what the purpose of one's investigation is. Engineer, biologists, and physicists are concerned about these questions. We bend the casicalcontn- ‘uum mechanics in this directon to make it useful for dealing with practical problems, ‘See, 1.10 A Biological Example ofa Hierarchy of Continua Wee eae te Ve TS i be ‘igure 1.6. Collagen fibers ina pulmonary alveolar wall of human Fung inflated to a transplmonay pressure (alveoli ges pressure mis pleurl pressure) of 1 cm 10. Seal is marked on Border, 60 piels re equal to 200 um in sue, O50, fied. Sve tained, Black fe collagen ber. Larger ones te about 1 un wide. From Sobin, SS, Fung, Y.C, and Trener, 4.M.,“Collageo and Eastin bers ism Poimonaty Alvelr Wale,” App. Pysilogy 648): 1659- 1675, 1988, Reproduced by permission2 Introduction Chap. 1 4.11 ELEMENTARY TOPICS THROUGH WHICH BASIC IDEAS EVOLVED ‘Asan introduction fo the ret ofthe book, let us consider some elementary topics that re simple and useful end that have been fundamental inthe history of mechan- ics. These include Newton's laws of motion, the equations of equilibrium, the use of reebody diagrams, the analyses of a truss, a beam, a block, a plate, and a shell, nd the classical beam theory, Ifyou ae familiar with these topics, you can 0 over them quickly. If some poins are new to you, Tean assure you tha learning hem would not be a waste of you ime, Newton's Laws of Motion Newton’ laws are stated with respect to material particles in athree-dimensional space that obeys Euclidean geometry. A material particle has a unique, positive measure, the mass of the particle. The location of the particle can be described with respect to e rectangular Cartesian frame of reference. Its assumed that an inertial frame of reference exists, vit respect to which the Newtonian equations of motion are valid. It ean be shown that any frame of reference moving with a uniform velocity with respect to an inertial frame is again inertial. Consider a particle of mass m. Let the position, velocity, and acceleration of this particle be denoted by the vector x, v, and a, respectively, all defined in an inertial frame of reference, By definition, 8 au Let Fb the tot fore sting onthe patie IF = 0, then New’ fa aw slates that = constant ar) IEF #0 then Novo’ second lav stats tat favs, or B= ma (ita) When Fag (1.11-3) is writer as F + (-ma) = 0, (1K) it appears as an equation of equilibrium of two forces. The term —ma i called the inertia fore. Equation (11-4 states that the sum ofthe extemal force acting ‘on a particle and the inertial free vanishes; i. the inertial force balanes the ‘extemal foree. The Newtonian equation of motion stated inthis way is called D’Alemibert' principle. ‘Now, consider a sytem of parties that interact with each other. Byery particle i influenced by ll the othr particles inthe system. Let an index I denote seo.1:11 Elementary Ties through Which Basic Ideas Evolved 8 tion exerted by pare numer hea pie Let denote tfc ofitencon xe : te ean ttt tie on parle. The Neos third inp sites tt Fy or Pot y= ass) ef in ageemet wih Ea (L319) Er = J then ve wt Fy ~ 0 ([ 20 se) in geomet » ae ter farts i en. Te fore hat ac 0 and an inert force ofan eteral loc uch svi, and am ints 1 patie on ron een pre. TS, H+ De utes) fy i ‘The equation of motion of the th patil is, therefore, ras Ang =H + To iz y quae te Each partie is deseied by such an equation, The totality of K equations sis the motion ofthe system. ert ye mast gst te foes fines Fi can be ecmpated. Sock spedeatn isa statement of the materi property of aca pies und lone os 2 conta equton ofthe ater system. Equilibrium ‘Aspecial motions eur, i, onein which there sno aoeeraton fr any particles of the system. ‘At equilibrium, Eq. (1.11-7) becomes adhe Cen OO ‘Summing over I from 1 to K, we obtain fers bdn de a ne de 2H “Thats, for a body in equiru, the zammaton of el eternal forces acing on (ars) (111-10)“ Introduction Chap. 1 Next, let us consider the tendency ofa body to rotate. Ifa body is pivoted ata point O andi acted on by afore F,, then the moment of the force about O that tends to cause the body to rotate about Oi given by the vetor product x 5, where it a rads vector from Oo any point on the ln faction of the force F,, Forming a veetor product oft with Eg (.11-8),i., with every term of that equation, seting J = 1,2... K,ading the results together, ., summing over rom 1 to K, snd wsing Bq, (.11-5) to simplify the grand total, we obtain (ut) ‘The choice ofthe point O is arbitrary. Hence, we obtain the second conition of cequilbcium ofthe body: The summaion of the moments of all the exteral forces ‘acting nthe body about any point is ero. Use of Free-Body Diagram in the Analysis of Problems ‘The word body othe phrase ass: of particles used inthe previous section can ‘be interpreted in the most general wey. Ifa machin sin equim, every part of tis in equim, By a proper ston ofthe pat tobe examined a varey of information canbe obtained. This method is ike a surgeon's exploration of a bn The product of wo 3 x 3 matrices A= (a) smatix defined as = @)isa3 x 3 square Jax 6a ah [on bw ba AD=|ay ax ta] [ba ba bo nt So) Yb bm by 3.) fobs + aba + ab = [cabs + abn + dab nbs + dab + aba] whose element inthe th row and jh column can be written, with the summation convention, as (ABs = (Gaba) ex1) ‘A ectoru may be represented by arow mtr (u), and Eq, (21-2) ean be written [ul = (a) = ak + ak + ad = ma 03413) By this rue th clr product of wo vectors u-y, Eq, (21-3), canbe itn 8 w= WOT = uta tun = mr OHM) SSec.2.3. The Summation Convention a “The determinant ofa square mate: is a number that is the sum of all the prodvcts ofthe elements ofthe matrix, taken one from each row and one from tach column, and no two or more from aay row or columo, and with sign specified by a mule given shortly. For example, the determinant of a 3.x 3 matrix A is written as det A and is defined as fou ar a det A= det (@) = Jon x a lon, @x as] (345) Ayla + Gynt + Attn ~ dude ~ datatn ~ ayer “The special rule of signs is as follows: Arrange the first index inthe order 1, 2,3. “Then check the order ofthe second index. I they permute a 1,23, 1,2,3,- +++ then the sign is positive; otherwise the sign is negative. Let us introduce a specal symbol, ¢, called the permulation symbol and defined by the equations 6) = Gan © Ga = Gin = fat = Gay = fat = 319) ea = Gy = 60 ey = 6a = 6 In other words, vanishes whenever the values of any two indices coincide; cya = 1 when the subscripts permute as 1,2, 3; and ej. = ~1 otherwise. Then the determinant of the matrix (a) can te writen as et (aj) = eritadata 347) Using the symbole, We can write Bq. (21-8) defining the vector product w x vas XY = ellie (23.18) “The e demtity ‘The Kronecker delta and the permutaton symbol are very important quantities that wil appear again and again inthis 200k. They are connected by the identity un = Bybu ~ Be A (2319) ‘This ¢8 identity is used frequently enough to warrant special attention here. Itcan be verified by actual trial8 Vectors and Tensors Chap. 2 Differentiation Finally, we shel extend the summation convention to differentiation formulas. Let flO 3a, 5%) be afaneton of variables x2... Then its diferential sll be waiten as ca ix, af Vans tins tans Lays af xa 23-00) PROBLEMS 2.48 Write Ba. 2-1) ot (22-3) in the index form. Lethe components of F be writen as FP b= 12,30, F Answer. 3 FP = 0 216 Show that @)&=3 (0) Bi =3 8 eae 0 ends, (2) 5:5. = Bs @ tun =0 217 Write Eas. (21-1) and (21-5) in te index form, e.,0-¥ = wi. ‘Note. For Bq. (21-1), ne may dothe flowing: Define three unit vectors” = ey, v= 6 0 = eg then a = a 2218 Use the index form of vector equations to solve Prob. 25 trough 29. 2219 The vector prodoct of wo veo = (1, yt) and ¥ = (yy ithe eOr ew X ¥ whos components ae = t= hy ‘Show that this can be shortened by writing = eat 1m = Way =a — 2. Bess gs (21-7) in he index fem, 2221 Derive the vesor entity connect thre arbitrary vectors A, B, Cby the method of vector anys AX (EXC) = (ACB ~ (A-B)C. Solution. See A x (Bx C)isperpendiular to B x Cy it must iin the plone ofBandC. Hence, wemay ite A x (BX C) » aB + bC, where, bareseaar quantities But A x (B x C) isa near function af A, B, and C; hence, a must be linear sal See,24 Translation and Rotation of Coordinates @ combination of A and C, and mutbea lina salar combination ofA andB. Accordingly, 1 are proportional @ AC and AvB, respectively, and we may wite Ax (BX C) = NAB + wAB)C where , pare pure mumbets, independent ofA, B, and C. We can therefore, eat by spoil cases, eg. if J, ae the unk vestors inte cretion ofthe x, > and ‘anes (a right-handed rectangular Caresian coordinate sjtem), respectively, we may put Be iC=j,A = ito show that y = ~Lyand B= 1, C= J, A j toshow that 1 1.22 Wate the equation in Prob. 2.21 inthe index frm, and prove is vat by means of thee identity 23-19. Note. since the equation in Prob, 2.21 i valid fr aritrary vectors A, B,C, this proof may be regarded aa proof f the €8 ety, Solution. [AX (BX C)], = euta(B X Cle = GeatatanBes = anteater By the eBientity, Eq, (23-19), thisbecomes(6, Bau ~ Buby) aye Hence, iis y6¢00, = Butabats * 86 ~ Mauss = (AC) CH), ~ (ABYC, 2.4 TRANSLATION AND ROTATION OF COORDINATES ‘Two-Dimensional Space ‘Consider two sets of rectangular Cartesin frames of reference O-ry and O'-x'y! ‘ona plane. Ifthe frame of reference O'x’y' is obtained from O-ry by a sift of gin without a change in orientation, then th transformation isa translation. If a point Phas coordinates (x, )) and (x,y) wit respect tothe old and new frames of reference, respectively, and ifthe coordinates ofthe new origin O° are (hk) relative to O-y, then A fravth o e+) poytk tayo k 1f the origin remains fixed, andthe new axes are obtained by rotting Ox and Oy through an angle @ in the countrlockwise direction, then the trnsfor- ‘mation of axes isa raion, Let P have coordinates (x, y, (x, y') relative wo the old and new frames of reference, respectively. Then (see Fig. 2.2), xn x ost y' sind 42) ! sin 8+ yc 8. cos 6+ y sin 8 043) xsind + y cos 80 Vectors and Tensors Chap. 2 1 Plgue 2.2. Rotation of o ae coordinates Using the index notion, we let xs replace x, y and xix replace 2, Y ‘Then obviously, a rotation specified by Eq, (24-3) ean be represented by the ‘equation = Bat, = 1,2) 44) where By are elements ofthe square matrix fox Ba {cose sin6) wf (Sp es ‘The inverse transform of Eg. 24-4) is ee) +9) ‘wher, according to Eg. (24-2), yi he element in the jth row and th columa ot emai) 1 er te matrix (fy) i8 the lanspase of the matt Brien Ga = a". ) (On the other hand, fromthe pont of view ofthe solution of the se of simultaneous linear equations (2.44), the matrix (fin Eq, (24-6) must be identified as the inverse ofthe matix (8), = Ga" 248) “Thus, we obiain a fundamental propery ofthe transformation mati (By) that deine a rotation of rectal Carian coordinates: a" = GI" e45) ‘A matric (Bi) fj = 1,2, thatsatses Eq, (24-9) clled an orthogonal ‘matrix, A transformation ssid tobe o-thogonal ifthe associated matrix isorthog- ‘onal, The matrix of Eq. (24-5) defining a rotation of coordinates is orthogonal. For an orthogonal matrix, we have (Pa)(Bi)" = Gi) (Bd = Gs ‘Sec.2.4 Translation and Rotation of Coorsinates 5 where & is the Kronecker delta. Hence, Bae = By (410) ‘To arity the geometric meaning ofthis important equation, we rederive itiretly for the rotation transformation as follows. A unit vector issued from the origin along the x-axis has direction cosines Bx, Pa With respect to the ar, sates, respectively. The fact that its length sunt s expressed bythe equation (Ga)? + GaP = 1 @= 1,2). Qa) ‘The fact that a unit vector along the x-axis is perpendicular to @ unit vector along the x-anis if j# iis expressed by the equation BoB + Bon =0, AD): (@A12) Combining Eqs. (24-11) and (24-12), we obtain Eq, (24-10), ‘Note: Alternatively, sine we know what the 8 are from Eq. (24-5), we can verify Bq, (2.4-10) by direct computation. ‘Three Dimensional Space Obviously, the preceding discussion can be extended to tree dimension without ‘much ado, The range of indices i, jcan be extended to 1,2, 3. Thus, consider two Tight handed rectangular Cartesian coordinate systems ty andy 2 5, With the same origin O. Let x denote the postion vector of point P with components ing 5 OFF, 1X. Let, se unit vector inthe directions ofthe positive we reeves. They ae called base vectors ofthe 4, a5 coordinate system. Let iek e be the base vectors ofthe x, 1,35 coordinate system. Note that since the coordinates are orthogonal, we have = yee) = By e413) In terms ofthe base vector, the veetorx:nay be expressed sfllows: x= 56-36. eu ‘Asealar product ofboth sides of Eq (24-14) with ies ere) = #(6-e). e445) But 68) = 3 = 85 therfore, x= (68s, 416) Nov, define (0) = Bs eamPy estos and Tensors Chap. 2 then, x= Bah, G2 123) (24-48) Next, dot both sides of Eq. (24-14) wit ef This gives a(ere) = ¥(¢e). Bot (je) hand (@ref) = By therefore, we obtain = Bay, C= 1,2,3). 4419) Equations (24-18) and (24-18) are generalizations of Eqs. (24-4) and (24-6) tothe thee-dimensional ease, Equation (24-17) shows the genetic meaning ofthe coetcent By. That gs (2-47) and (24-8) bold for ,j = 1, 2,38 lear becase Fs. (2.4-18) and (24-10) are invere transformations of eachother. Then, Eqs. (24-8) and (24-10) follow. Now, the numbers, %, 5 that represent the coordinates ofthe point Pin Fig. 23 are als the components of the radius vector A. A recognition ofthis fact igure 23. Radius vetor and ‘i cnordiotes. ves us immediately the law of transformation ofthe components of a vector in ‘ectangular Cartesian coordinates: Al= Bids A= Bud (24-20) in which fy represents the cosine of the angle between the sxes Or and Ox, Finally, let us point out thatthe three unit vectors along x, 2, x5 form the edges ofa cube with volume 1. The volume ofa parallelepiped having any three vectors u,v, w as edges is given either by the triple product u-(v X w) of by its negative; the sign is determined by whether the three vectors u,v, w, in this order, form a right-handed sere system or ct If they are right handed, then the volume is equal fo the determinant of their components: See.25 Coordinate Transformatior 58 lu mu Volume = (ux v)-w=|n vm] (24-21) mm ‘Let us assume that x, %, #5 and xi, xx} are right handed, Then iti lear tht the determinant of By represents the vdume of e unit cube and hence has the value 1: Bu Bu Bal Bu Bx Bal = 1 24-22) [Bn Bu Ba (ead PROBLEMS ‘225 Waite out Eq, (2410) in extnso, and iterpet the goometric meaning ofthe six resing equations; = 1,2, 3. Solution, Lette index stand for 1, 2,3 Iie 1,j= 1: then Bap + Bae + BoB = 1 0 Wi= 1,j = 2:then BBs + BoB * PoBa = 0 a ation (1) meas tha the length f the veto (BBs Bs). Equation (2) means hat the vectors (Bs Bs) (Bs Bu» Bx) ate othogonl to cach other. ‘Other combinations of jae similar. 1224 Derive Bq (24-10) bythe following alsmative procore Difreniat both ides (LEG, (24-4) with respect to, Thea we Eq. (24-6) andthe fat that d= 8 to simplify the resus, Soluion. Differentiating Eq, (24-4) wth respect to x, we obain 8 = Buds! a, Bot = By}. On ebngng te inex K and iereting, we have axl} = {Be Combining these results yes 8y = BB 28 COORDINATE TRANSFORMATION IN GENERAL 'A-sotof independent variables x;, 24 specifies the coordinates of a point in a frame of reference. A set of equations H=fGian), §=123) Qs.) describes a transformation from 2, % toa set of new variables %,%, 3 The inverse transformation = HG HH), C= 123) 52)5 Vectors and Tensors Chap. 2 proceed in the reverse direction. In orcer to ensure that such a reversible trans- formation exist and isin one-to-one carespondence ina certain region R ofthe variables (x, 2,5), in order that each set of numbers (&,%, ) defines @ unique se of numbers (x, 2), fr (x, 3) inthe region Rand vice versa— itis suticent that (1) The functions are single value, are continuous, and possess continuous first partial derivatives inthe region R. (2) Te Jacobian determinant J = dt (ex) does oot vanish a any point ofthe reson That, am a a am as 40, @5-3) Coordinate transformation with the properties 1 and 2 are called admissible ‘transformations. Ifthe Jacobian is positive everywhere, then a right-hand set of coordinates is transformed into another right-hand set, and the transformation is said to be proper. If the Jacobian is uegative cverywhere, a righthand set of coordinates is transformed into a left-hand one, and the transformation is said to be improper. In this book, we shall tacitly assume that our transformations are ‘adssble and proper. Significance of the Jacobian Determinant ‘To appreciate the significance of the Jacrbian determinant, let ws assume that we have found that (of, 24.4) corresponds to (F2,¥%,3%), ie, they satisfy Eq, (2.5-1), and ak whether we can find an inverse transformation ina smal eighbor- hood of this point, We diferentiate Eq (25-1) to obtain eta, G=129 (5-4) and evaluate the partial derivatives aff, atthe point (f, 2, x). The Eg. (@.5-4) defines a near transformation ol the vector dito a vector di, If we solve the st of near equations (25-4) for dx, we know thatthe solution exists onl if the determinant ofthe coeficents does tot vanish: ‘Sec.2.6 Analytical Definitions of Scalars, Vectors, and Cartesian Tensors 85 ‘Thus, an inverse exists inthe neighborhood of (x, 28) only if Eq, (2.5-3) is valid. Further, when J # 0, Eq (25-4) canbe solved to obtain = 0) 05-6) where gare constants. Hence, a smal neighborhood ofthe known pont, an inverse transformation (an approximation of Eq. (25-2)} can be found in a small borhood ofthe known point. Thus, conditions 1 and 2 stated eater are sufficient conditions forthe existence ofan inverse ina smal region around the known point, [By epested application of this argument to new known points away from the initial known point, one can extend and find the region in which a one-to-one inverse twansformation given by Bq. (2-5-2) exists PROBLEM 2.25 (a) Review the methods of solving linear simultaneous equations. Oe ofthe methods ‘ues determinants, Use tat mod to sale Eq, (25-1 for diy di, dk. Use the ‘ermtaton symbol», defined in Eq, (23-16), to express the final result. (6) Ris asegion in and on ac of unt rads on plane, Te equation of the cic isr = 1 polar coordinate ands” + y* = 1in rectangular Cartesian coordinates Show tht th Jobin J sequal to rand tat the aes ofthe cil is [foals [[frsan ff ean Here, an integration ofthe Jacobian iuliplied by the product of the diferent dr db ves the area 2.6 ANALYTICAL DEFINITIONS OF SCALARS, VECTORS, AND CARTESIAN TENSORS et sas) ad Of Jn) be to feds of etangolar Cartesian frames of reference tlt by te transformation lav it 6) ‘where isthe direction cosine ofthe angle between unit vectors along the coor- dinate axes % and x, Thus, fy = cose, 2), 25-2) ands forth, The inverse transform is = Ba 26-3)55 Vectors and Tensors Chap.2 A.gystem of quantities iscalled 2 scalar, a vector, or a tensor, depending upon hhow the components of the system are defined in the variables x, x15 and how they ae transformed when the variables x, 2,25 are changed to, 3, ‘A system is called a scalar iit tas only a single component inthe varibles 2 and a single component & in the variables % and if © and @ are numerically ‘equal atthe corresponding points, Gi,28) = OG, BB). 26-4) ‘Acgystem is called a vector fild or a tensor field of rank 1 i it has three components nthe variables and three component inthe variables and if the components are related by the characteristic law Ein = Ein 2 8) Ba 265) sn) = EG Hs BB Generalizing these dftitions a system that has nine components wien ‘and j range over 1, 2,3, we define a tensor field of rank 2 if tis system that has nine component inthe variables. 2, and nine components inthe variables 5, Jn Bs and ifthe components are related by the characteristic law Tuy FB) = foal 31) 06-6) 1 8y85) = Faas Sa BBB Frer generalization to tensor lds of higher ranks is immediate. These 0, and noting that, 6 €€ vanish with h and dS, one obtains ee en 634 which isthe fist eomponent of Eq (3-2). Other components follow similarly ‘Cauchy's forma asures 1s that he nine components of stresses are nec- essary and sufcient to define the traction across any surface element in a body. ‘Hence theses atin ody i charecrzed completely by he et of uate sy Since Tis vector and Eq. (33-2) valid fran arbitrary vector vy it flows that zis a tensor. Henceforth, x wl Be called a ste tensor. Checking Acceptable Errors InSeo, 1.5, we defined continua on the basis of acceptable variability anda limiting approach that has lower bound of dimensions. Tn Sec. 1.6 the concept was applied te the definition of stress. In Section 1.7, we adopted an abstract copy of the real tnateial asa way of idealization. Inthe proof of Cauchy's formula, Eq. 33-4), tre have used the abstract copy and followed the usual method of calculus to throw ‘away a number of terms in Eq. (33-3) and reach Eq. (.3-4), We claimed that the sum of the terms em ten ton tet Ime — 0¥) e325) js smal, compared with the terms tha are retained i.c., Try Tad Yates Tas G34) sven we take Eq; (33-3) t the limit sk» 0 and AS > 0. Now, ifwe ae nat ‘owed to take the limit as h> 0 and AS» 0, but instead we are restricted to ‘hing smaller than a constant hi and AS no smaller than a constant multiplied by Ge} then the guantty sted ini (33-5) must be evaluated fr h = ht and ‘AS = consti) and must be compared with the quantities listed inline (3.36). ‘A standard of how small is neglgble must be defined, and the comparison be made under that definition, If we find the quantity inline (33-5) negligible com- pared with those listed inline (33-6), then we ean say that Bq. (33-3) oF Ea {8.32} i valid. This tedious step should be don, in principe, to apply the con- timum theory to objects ofthe real world,td Stress Chep. 3 34 EQUATIONS OF EQUILIBRIUM ‘We sll now transform the equations of motion, Eqs. (22-7) and (32-8), into differential equations. Tis can be dove elegantly by means of Gavss's theorem and Cauchy's formula, 28 is shown in Chapter 10, but here we shall pursue an elementary course to assure physical city. ‘Consider the static equilibrium sate of an infinitesimal parallelepiped with surfaces parallel tothe connate planes, The sreses ating on the various surfaces are shown in Fig, 3.6. The fore yd diy acs on the lefthand side, te free in, (eof dx] ats on therght-hand side, te. As it lle expaned below, these expressions are based on the assumption of continuity of the stresses. ‘The body free is Xi dr de ds. Figure 3.6. Equlibratng ses components on ‘nfaiesinalparalleped. ‘The stresses indicated in the figure may be explained as follows. We are concerned witha nonuniform sess Sel, Bvery ste components afneton of poston, Ths, the sess component i a fnction of, 1 ult X42). Tae a point slighty tothe right ofthe point (22), namely, at (83 + dB An “ps the value ofthe stress ry i-a(t, + dx, Xn). But if isa continuously dliferenable fonction of sy 2 then, according to Taylor's theorem with a remainder, we have sults # iyi) = tl 38) > iY 3038) 1a + ai SE a + eden) see.34 Equations of Equilibrium a where O:< a < 1. If ost is init, then the lst term canbe made abitanly weet eampare with the oer terms by choosing ds suey smal, With ssh 1 choiee, we have sy + by 30,38) = Tut X09) + me Gms mdde In Fig, 36 we wit, or shor, ands + (uf on he surat whee ge ot The lft, boom, and rear surfaces are lested a yn THE tegen ofthe element have lengths dr, di di elas and thes Gervtvs ae erated at (2,1). Dgulibiom of he body demands tht thereat fre ais. Cnsr the foes he xo Pee own in Fig, 7, We have sx component of surface force and one eamponent of by fore. Te sum is ay de, iy . a 4 rn an * ey dts — Tada dy + tn + ands + ft Ba) od = ad 0, + Hi dd ds eu) bon igre 37. Canponet twos in dieion Diving by ddd, we obtain fe 4) a epi emu te nics kao sini qos fein of A. gle peu ions, Te woke st, wten come ms x A G43)” Stress Chap. 3 ‘This isan important result. A shorter derivation willbe given later, in Sec. 10.6. ‘The equilibrium ofan element requires also thatthe resultant moment vanish, If there do not exist external moments proportional toa volume, the consideration ‘ofmoments wil ead to the important conclusion that he sressensor is symmetric, ie, a G44) “This is demonstrated a follows. Referring to Fig 3.6 and considering the moment of all the forces about the z-axis, we se that those components of forces parallel to Oxy or lying in planes containing Oxy do not contribute any moment. The 2 Fy, = t 1 Hl 1 |“ i pe 38. Componeats of ta tos ht ores nomet fat Gays Te about the Ox,-axis. compones tht do conibut a mona abou te raise sown in Fi 38 ‘Rete ig ewe he monet ra pop, ve kee itu dy ey wou Ber nt + de ta sno Be sft Be Se=.35 Change of Stress Components in Transformation of Cooninates 75 On dividing through by ds, drs dy and pasing tothe iit as +0, drs 0, and drs 0, we oblain 643) Sinilar considerations of resultant moments bout Ox and Ox, ead to the general result given by Eq. (3.43). A shorter derivation willbe given later, in Se. 107 So far, we hve considered the condion of equilibrium. I itis desired to 0, > 0, thn 6, {isthe largest stress and (oy ~ o;)/2isthe largest shear for al posible surfaces. The ‘lane on whic the largest shear ats incined at 45 fram te principal planes on which, and 0 ac ‘The meaning ofthe three bounding its in Fig. 4.4 can be explained easy Let sy, zaxes be chosen in the directions ofthe principal axes. On a plane perpendicular to the a-axis, there acs a normal stress, say, and no shear. On ‘plane normal to, there acts 2 norma stress, sa, and no shear, Now consider all lanes paatel tothe z-axis. For these planes, the normal and shear stresses acting 0 them are given exactly by Eqs. (42-3) trough (12-5) or Eqs. (42-6) through (42-8). Hence, the Mohr contruction described in Sec. 43 applies, and the ccle pesing through o, represents the totality of all stress tates on ‘hese planes. Similarly, the other two circles (one passing through 0, andthe other through o;, 0) represent the tality of al stes states ating on ll planes pall to either the ans or the yi. It remains only to show that the steses on all other plans le in the shaded are. The proo is given in eater edtons of this book, It is quite lengthy and is cmited here Figure 44 Moh’ cre. 45 PRINCIPAL STRESSES In a general state of stress, the stress vector acting on a surface with normal » depends onthe direction of v, At a given point in a body, the angle between the Sires vector and the normal v varies wth te orientation of the surface. We shall show that we can alway find a surfacs so oriented thatthe stress vector is exactly ‘normal to it Infact, we shallshow that there are atleast three mutually orthogonal Sec. 45 Principal Stresses . ee ee ss dace papa meri pe ee ee cre % iis ‘the stress tensor. Hence, writing v; = 8jy, we have, on equating these two expressions and transposing to the same side, (uu - o8)y=0, tic for » and 5. Since va nt veer, we tus finda tof onl aun or which ae 2+ v9 = 1. Thus, Eqs. (45-1) pose/an eigenvalue problem. Since x as a matrix Frm tai We loony fel etn ey of il ee ae an ici i. ce ra a crt br ee fr See caer a ; [ry - of = 0. (45-2) ation (45-2) isa bic equation in ot 08 athe princi stress Foren s of e pinpal sy at tral etre Be mined (On expanding Eq, (45-2), we have 1,2,3). as) le-o=| ome mf (45-3) =-0 the’ -loth view tutta tte 454) ' | fe 35 bel et (568 Piincipl Stresses and Principal Axes Chap. 4 On the other hand if, o, os ar the roots of Eq (45-2), which can then be writen as (© = a)(o~ 0,)(0 - «) = 0, Gs) itcan be seen thatthe following rations between the roots and the coefficients must hold: Laat toy (45-8) b= OG + os + 0m, (45-9) = ows (45-10) Since the principal stresses characterize the physical state of stress at pont, they are independent of any coordirats of reference. Hence, Eq. (4.57) is inde- pendeat of the orientation ofthe coordinates of reference. But Bq. (45-7) isexaclly the same as Eq. (4.5-3). Therefore, Eq. (45-3) and the coefficients 1, ss are invatient with respect to the rotation of coordinates , Isao called the invariants ofthe stress tensor with respect to rotation of coordinates ‘We sll show now that, fora symmetric stress tensor, the three principal stresses are all real and the three principal planes are motually orthogonal. These important properties cen be establihed when the stres tensor is symmetric, ie, (45-11) ‘The proofs follows Let, 3, 3 eu vector in the erection ofthe principal axes, with eomponents 5 3 (= 12, 3) that are the soltions of Eg (45-1) corresponding othe 10s, oc espetvely, thea (ry ~ o18)¥) = 0, (ru oy, = 0, (45-12) (ru ~ ei, = 0. Maliphing te fis equation by andthe second by smi ove, nd sabacg the esiog equation, we ban (: - oi = 0 (45-13) ‘on acount ofthe symmetry condition (45-11), which implies hat ahh = hh = nahh, (4518) the lest equality beng obtained by nterchonging the dummy ines i andj ‘Now, if we assume tentatively that Eq. (45-3) has a complex roo, then, sine the coefficient in that equation are rel valued, a complex conjugate root ‘musta eit and the set of roots maybe writen a aati 2=a- o ‘where a, B, 6; are rea numbers and i stands forthe imaginary number VT. Ia ee, 46 Shearing Stresses a ths ca, Bs (45-12) show that nae conjygst teacher ad canbe witen jegtin h=6-% Therefore, s+ 8) ~ 8) Gta eg ths + BHO. ovr Eq (45-15) ato, 0 29 = 0.508 = 0.Duttiscontas {fons he oo cmpls Ths, the asumpon of Be 2 pet ot etna, dh os 0 a a acne Bq (45-1) sina equaios imply tat sha, fi 0 4545) seth prac vest ae mtu athogonl to echt. Iay = 01 #8, see ted butwe eon determine if mui of prs af veers and Fhogna to Ia, = 0: = 0 er any setof orthogonal ae maybe taken tsi np es, : Sr exes, 2 een 0 cine ith th pins, then the matrix of stress components becomes i 0 0) =|) o 0 45419) =0, 3h 20» 46 SHEARING STRESSES vith a yt outer normal» (th We have seen that onan element of surface with a uit components), there acs traction (with components = )-The component ot ine dt of te oma ang on thesia ones Ta vr tess be denoted by dy Since the component ofa veto inthe ths aaa ee pen soa rd fhe 8 ver, ME obtain = i= 100 66) i into two orthogonal the other hand, since the veto canbe decomposed into ng eats eng ey get oe ie (Gee Fg 4.3), ese tat the magni ofthe shearing tess on a surfs having the normal is given by the equation 462)8 Prinepal Strsses and Principal Aros Chap. 4 Figure 43, Notations Let the principal axes be chosen ss the coordinate axes, and let o, 0 95 be the principal stresses. Thea =om =m, t= om, ; (66-3) ID = (oun? + (orm)? + (oa), and, from Eq. (4.6-1), oq = out + ont + oss, (46-4) of = [owt + oh + on 65) (On susttuting into Ea. (4.6-2) and noting that (9) - (9) = OF 1- I= FOF +O 46-5 we se that P= OOM oF + OFONG- oF en + FOF ~ oF. Forexamplo, fy = m= UVEand = then r= ( ~ 0. PROBLEMS fi ‘Show that tax = $(Com ~ ou) and that the plane on which tog acts makes an angle of 4 with the dieton ofthe kgs and the satlest principal stresses, Solution. ‘The problem isto fad the maximum or minimum of. Now, is ten by Bq. (46-7), We must Gnd the etemum of 1 function of», my vs under the reson hat» + of +o} = 1. Using the method ofthe Lagrangian ml, we srk to misinize the oneton LE ARG, ~ oF + RG, — HI + HG ~ of GEA HE = Dh By the usual method, we compute the para derivatives fn, af, equate them to 0, and ol for yy yh Tis eds he following equations: x Bylo, ~ 0) + DoaKloy ~ off + Do, = 0, 0 $00.47 tres Devatin Tensor = Bente ale ~ oy + tives! += ® Bete tater 0) 206-9 +B ® Hog gamegod 0 a (One of he solution of Ea, (1) i ebvioly », = 0. On sting» = 0, Eqs. (2) and (3) tesome a 4, Bg, (4) Becomes 0 + ler oN H.-F +8 “These equations ate consent ony = x. On sting Bed ooros = IVE Hen he st et of lions [ae eo ( apne de + Subsitling this back into for Eq (4.6), we find the extremum of Co 7 a 7 ‘other sets of solutions of Eqs. (2, @), (3), and (4) can be obtained by sting, in turm, vy = O and y= 0, We hav then te eatin maxima or minis ‘he gs oft thesis heim of. Reutimann te pan or ihe soe nun set isgoety te appopitr. Wee elon ne be my UeD, hich pis 245° ininaton the x, me. le ‘47 STRESS-DEVIATION TENSOR The tensor | y= 14 00 m4) | is called the sess deviation tensor, i whch 8, isthe Kronecker dla and oss the mean sess: =H tor toy = Hn tte tt) 3h, 472) where J; is the fist invariant of Seo. 4.5. The separation of x into hydrostatic100 Principal Stresses and Principal Axes. Chap. 4 Pa ony andthe devon i very inp in ering the pai bebe of met - ‘The first invariant ofthe stressdeviation tensor always vanishes: 0 (4-3) ‘To determine the principal stress deviations, the procedure of Sec. 4.5 ma followed. The determinantal equation = * oa iy oy = (4-4) Nath thts may be expanded inthe form oP = La! = hs 7-5) Its easy to verily the following equations relating J, Js to the invariant Ie cay tr ing ea 18 Jy Jato the invariants Jy Js I= 308 ~ by ar9 I= by ht + 20% = b +10) ~ 0, 42-7) tis ao easy to verify the following altemative expresios on acount of 473) ™ ms —- i — thts — nth + (ta)? + (ta) + (ra)? (0 — ea) + (om — 20)? + (om — tu) + Coa? + Gal? + Od Uri)? + (ra)? + (6) + Ca)? + (a)? + (On)? (478) Hence, d= bin 47) ‘To verify al four equations, we note first that sine JJ and, fs areal invariants, itis sufficient to verity Eqs. (4.7-6) nd (4.77) with ¢ particular choice of frame of reference, We observe that the principal axes ofthe stres tensor and the stress- 0, v=( cose 0720 cose <0, v=0 sary Mage 54 Deoratn pans ery il ‘5e0.55 Infinite Rotation m hem aly + fr epresens the hanein the angle x0, which was oily a right ange. Thus, sfau , 2 lay Inengnerngusege, he strain ponents yout, Pay called aa aeering sans or derusions. The nane is parselal suggestive Case 3 of Fig 5, whichis alle the case of simp shea. Fan ngeofange x0), 642) 55 INFINITESIMAL ROTATION consider an initesialipecement el (22). From form the Cae sian tensor 6s) which is antisymmetric .€., 6s) ene, the tensor ay bas only tre independent component, Woy nd - ata build dal vector - w= teu 653) fee ys the permutation syle. 23), On the other tan fom Pa 15 a hr cei, Ea, 218), flow ta cm = Hoy — which, by Eq. (55-2), is uy Hence, oy = Gn 6.4) ‘Tus, may be cal the du! ansymmetiz) esr of 2 vestor oe shall aa edn respecte, te rotation vor nd rotton for ofthe dspace ment field u ‘cig modfstion ofthe poo vena the end af Sex, 5.2 wil envi vy that he vanishing of the symmenri sain tensor By or ey nace, and ‘pata condtion fora neighbors of «parc fo be moved Ie & rigid body. =z motion consis of tanto anda rtaton, Ts wens Arig) aon? We hl show ta on infieial plcemen fil fr Wi enor ves aa pit Phe roion of neighborhood of ‘poten bythe vector To show this considera pont Pi neighborhood of iy oun of P and Pb andx\ + dy respectively. The eave displacement of P with respect to P's ay, x, uy, 655)wm Analysis of Deformation Chap. § ‘This can be written as glee B)e ‘The frst quantity in parentheses isthe infinitesimal strain tensor, whch is zero by ‘assumption. The second quantity in parenthesis may be identified with Eq, (55-1). Hence, ty = ~ yt = oy = ~ cindy by. 654) 659 = (0 x dy), (by definition. Thus, the relative daplaement isthe veto produc of w and de, This exactly hat would have bets produced by an initesinal rotation fa about an ax Srough Pin the don of Tr shouldbe ated ht we tae rested ourselves to inintesinal angular Gispleesens. Angular messes fo ft dilaements ae related fo yn tore compiatd ay 536 FINITE STRAIN COMPONENTS ‘When the strain components are no smal itis also easy to give simple geometric interpretations for the components ofthe strain tensors, ‘Consider a set of rectangular Cartesian coordinates with respect to which the strain components are defined s in Se. 5.2. Let line element before deformation be da, with components da; = dy do clement be defined by he or eT 652) From Ea, (52-10), we have 9 ~ dij = 2B da day = 26 (da). (663) ‘Combining Eqs. (5.6-2) and (5.6-3), we obtain (+B) -1= 28 664 hich gives the meaning of Bin erms of F,. Conversely, B= VIF - 1. 665) oy Sec.56 Finite Susin Components ‘This reduees to B*Ey 665) ssl compe “— ‘get the physical significance of the component En, let us omar two. tec hd ata eight ne nial te da, = diy d= 0, ds flag = diy day = (66-7) to, itn componess sn 8 em, ase ne cnet ee dit cmos At ig slr rt he tome cen obtzin But according to the definition given in Eq. (5.2-8), we have, since Bx any ae Hence, 8089 = 2 i ce) But, fom Bes, (56-1) and (6.65), ve have d= VOFIEn dy, di = VIF TE dhe Heace, Ba (5.68 ys = 665) 008 = SIE, Vi = Oe ‘The angle Othe ange between thin lenis ds and ater deformation ‘The change of angle betwen the wo line cma, whic nth orig nthogona, san = m2 ~ 0. From Eq. (5.69), we therefore obtain ee (56-10) Sinn = 7S 3B, VI + Tea oy Te ine « eqatns exit he relation of Exo ges # and Ta ml nthe cs aso te mohenet of Ey and Ex in these equationscy ‘Analysis of Deformation Chap. 5 A comple eds intepetin ante ion an ema fr the alan compet Deng toner fred egos io = (eu) wwe find that =~ Via Bs, (56-2) Sater ie devin fom ht nl wee wo seme in he ‘which, after deformation, become orthogonal is denoted by Bre, We tae cobs Yee VeVi a) In case of ites stain f infinitesimal strains, Eqs. (516-10) and (5.6-13) reduce to the familar results tem BtEs ox* 2a, Batten (56M) 5.7 PRINCIPAL STRAINS: MOHR'S CIRCLE tout mc ad, we cnexand he nected the sf Se, 41 ugh 4810 So pts een ep te od ome: Al We hnfo os echuge he wor re wih rah, Th (@) There exist three in ae i ie el a ae He ~ 0 =o. i (6041) "Mesa th dm vl ni - Associated with each principal st : h each pica sta, sy, thee is pip ai th Secon cose oo at we els be eutons @-aaw=0, (=1,2,3), or ‘The three sets of solut fof, op) hea oa tutions (>, of, v2), (oP, of, oP), (of, vf), oP") are com- pes wee unit vectors. If the roots ee, ¢s of Eq. (5.7-1) are di (ent is i ne ebro ee ire set sn ies ten) yi ttn ns nti so hte ete a es aera ne Pom eel (oA pls pps spl se inp plo en 88 lnfintesimal Stain Components inPoar Coordinates 5 (athe coordinate aes, 3 cine wth he principal ae, hen the strain tensor assumes the canonical form i 0 0) o 4 0 0 0 (@ We en ete nin deiton tes =~ ab a an Dove the tlowing independent sie veins heel, P= a=, B= teat, = Hele (673) y= Heatontee = Sehehtar (9) Mon’s cle may be se forthe rahi analy of Sin. Lamé’s ipso is also applicable to stra. 658 INFINITESIMAL STRAIN COMPONENTS IN POLAR ‘COORDINATES ~ ‘Aswe nated in See 36, it often eile to introduce eoreinear condos awe mince The sn component cn be refered tos lel etangsay ae toe rented in the diection of he curiinarcordinnes, For example, St par covets, 6,2 tests componens may be egies So ari ena they 2 lated 10 any yy EE BY HB ESE ‘wansfor- mre sa ta, as in the cases of stresses. (See Sec. 3.6.) a if diplacement vectors are resolved into component inte fos ons ofthe eotinearcordnate, the tran-deplacementreationip involves vo se ofthe dplcement components and teers nfiveneed Y) the sory of the coordinate system. The san-isplacement relations my 2PPSr Shute ciflecat rom the corzesponding fool in elas t coordinates. ier ofl mcd or kaniogenetneatcoornatsis that of ene tensor vale Te ade i referred to more advance eats. An puree even eathor's Fotos of old Mecha, (Y-C Poe 1965, Prentice Ben aglewood Cl, X). Ling curses in the preset bono Cartesian Ginates in an ad hoe manner. tensors, we must treat each set of curvilinear coor re Mesa sate two adhoc aproches i the case of elindiza Paoe coordinates by tansormation of coordaates and by detailed enumeration. The mer wil be discused inthis section, te later in Se. 5.9.128 ‘Analysis of Deformation Chap. 6 Inthe frst approach, we stat fom the relations between the pola coordinates 1,8, z and the rectangular coordinates x,y, 2 [x = cos 6, tao, Fores (681) ly = rsin 8, wey, Xela sng 682) yr a0 _ x _ cos 2.2.00 68-3) ecean 683) It follows that any derivative with respect to x end y in the Cartesian equations ‘ay be transformed into derivatives with respect to rand 6 by a_i 68-4) yy Now, in polar coordinates, we denote the components ofthe displacement vector 1 y t,t 28 shown in Fig. 5.5. The components ofthe same vector resolved _Mgwre SS Displacement vector in polar coordinates. in the directions of rectangular cocrinates are us, yt. Brom the figure it is seen tat these displacements are related by the equations o i, = n-008 0 ~ usin 8, a, = wan 0+ u08 8 685) wen ‘he sain component in polar coordinates ae designed as esa fe tnt 686) 2 c ex 58 tfritesinal Stain Componensin Pear Corina referred to a local frame of rectangular tin comonen ion 9 ining wth he ines between the two ses of coordinates These. are really the str ae 3 vedas aye, ‘with 2° coinciding with the r-direction, i ‘with z. The direction cos Seiretion, 2082" ae: Torx’ Bory’ Ge) “The tenor transformation law bods, and we have eccca? 0+ 6 Sit # + ey 828, cg cnc 6s ce eqenvineeg(Ft- HPO 684) on =u 005 OF by Si Sin 0 + 6 080 42% ry Oe ws) oe 6s fa, E a= har ie se, ion fp (2-4) 904) to EC) Yk ow at — a= (os 05. St Fuca wsn (68-10)29 Raton in Polar Coordinates = Analysis of Deformation Chap. 5 ‘Dirac Davin of the Sin Diaseent Substituting these and similar results into Bq, (5.8-8) and reducing, we obtain (su) ‘Thus, we see thatthe method cf transformation of coordinates is tedious but straightforward, Note that the structures of Eqs. (5.8-11) and (5.89) are diferent. t Inthe language of tensor analysis, the difference is caused by the differences in the fundamental metic tensors of the 1wo coordinate systems, ‘The reader should be warned again that we have adopted the tensor notations forthe strain, so that the shear strain components ee, are one-half of those ‘ordinarily given 28s Yee in mest books. i ‘59 DIRECT DERIVATION OF THE STRAIN-DISPLACEMENT RELATIONS IN POLAR COORDINATES ‘The results ofthe preceding section can be derived directly from the geometric definition of the infinitesimal strain components. Recall that the normal strain components mean the ratio of change of length per unit length, whereasthe shearing strain components mean one-half of the change ofa right angle, For infinitesimal displacements, these changes can be seen directly from drawings such as those es ee net nomena SETA Consider first the displacementin the r-dtestion, ,. We se from Fig, 5.6(8) ee aan te . seid telecast rd tion the a ae egw tet flr = eo) sind © runfeenel rain de 0 vrition of Sen ct ; fant Ce Sai (ue (URE aplenty rane Et fae gd asec ca Se or ‘From the same figure, we see als that a radial displacement of a crcumfer- ential element causes an elongation of that element and, hence, a strain in the 0 rection. The element ab, which wa orignaly of length rd, is displaced to a’b”130 ‘Analysis of Deformation Chap. § and becomes of length (r + 1) d#. The tangential strain due to this radial dis- placement i, therefore, Comme 69) rao (On the other hand, as shown in Fig 5.6(b), the tangential displacement u gives rise to tangential stain equal i eee ° do rad 693) ‘The total tangential strain is Bat 69-4) &, 695 as inthe case of rectangular coowtinates, ‘The shearing strain eis equal fo one-half of the change of angle ZC’a'b* ~ £Cab, as illustrated in Fig. 5.6(0). A direct examination of the figue shows that yim a ates oo ‘The frst term comes from the change i the radial displacement inthe dtitection, the second term comes from the change in the tangential displacement inthe racial rection, and the last term appears since part ofthe change in slope of the line °C’ comes from te rotation of the element asa solid body about the axis through 0. ‘The remaining strain components, e,s and e,,can be derived with reference to Fig. 5.6(4) and (¢). We have If (au,ta0)d0 iat #69 and = Wleuleedde - (ould fou, 5 ora HOM, Cali te rc) ‘These equations ar, ofcourse, the same as Eq. (58-11). Indeed, the dzect geo ‘metric method of derivation provides a much clearer mental picture thn the alge. braic method of the preceding section, ec, 5:10 Other Strain Measures 5.10 OTHER STRAIN MEASURES must not tink thatthe strain esos we have defined are the ony one suitable pnt enone hem tal ssw be rata f deformation onthe ange ofthe quar ofthe dtc between ny to parle (Se, 5.2), The sure ofits convenient stag pit became we ve Pythagoras thorn, whic tte tate quo te hypo nus of ight triangle sequal tothe sum ofthe squares of the lps. Using this tinino, we sas tht the equare ofthe distance between two points x and + dd, with coordinates referred toa rectangular Cartesian frame of reference, is de = di + dil + df, nec, 5.2, we based our analysis on this equation; the result wes natural definition ots cor mene ee dent tov hae te eid hi vay Fore, en ge ngs ed aed o wot erie a eth ofa ids he opment Bel point, uu xy ® yw» 6.104) wy aw aw aw ey H- i “deformation gradients,” are quite convenient. Indeed, these derivatives, called “deformation gradients, sit ‘We may separate the matrix (@u/ax) into a sum of a symmetric part and antsymmetc part: uu 44 You , a) =z. ytal Oe a f Afov | au) x oe, 2 wee aie Yaw , au) aa ae) % fou Dar fav 5.10) Daz & : ~ Hay a au , 3m ee1 ‘Analysis of Deformation Chap. evident thatthe symmetric part of the deformation gradient matrix is the matrix of the infinitesimal strain, es defined in Sec. 5.3. ‘Other wellknown strain measures are Cauchy's strain tensors and Finger’ strain tensors, When the mapping is gven by Eqs. (5:15) and (51-6), Cauehy’s strsin tensors are one 6103) ‘whereas Finger’ strain tensors are ay a oy Bs aac ay ia a) For these tensors, the absence of strain is inicated, not by the vanishing of C, oF Bip bat by Cy = Bip By) = By ‘We shall not discuss these strain measures any further, except to note that they may be convenient fr some special purposes in advanced theories of continua. PROBLEMS [51 A blood ves is incompressible, i, its volume does wot change. Under norm ondvons, a blood wes ean be crsidered asa cular ender. Suppose a person fs his or ber blood presue increased for some reason ad the inner radius ofthe ‘Hood vss increases from a0 2 + de, while the ail length is unchanged. Compute the changes inthe crcunerentl and radial strains throughout the blood vessel due tothe increas in blood pressure. 52 () Astute of defomation in which he dslacement i sa inearfneion of he noinses i elled a honegeneots deformaon. Whats the equation of surface that wil become a sphere © + y? + 2° = + after a homogeneous deformation? [Use an oguation of the pe fi, 7,2)» 0, in which x,y,z me recangulr Catan coordinates] (0) Asa special ease of homogeneous deformation, conser the following linear trans- formations of cordaats from (r, 2) (',)', 2), both of which fe othe same Cartesian frame ofreferene. (Se Fg, P52) (0) Pare shearsa? = kx,y' = Ie'y,2" = 2 (Q) Simple shear: x" = 2+ 23,9" ys 2" = 2 i of | |} oy ¥ = eo ‘We may regard (,y, 2) the coordinates of « material patil before a defor. ‘mations imposed ont nd (2, y', 28 the coordinates afer deformation. Show that 2 pre shear maybe regarded ae simple bear refered to anes inclined at tan“ (E~) chop. 8 Probie 1 vith On opi == 4°) Banke ain cc ny epee er A lh aol he sin psi inne a = Htanr! s = tant (87) with Ox. E ee ee er Pret rtsee cs cn bon} CBE nn dw. Lonto: Mee & C2, 185-32) vo We define a homogeneous eforao 3 defonaon which he die Soon: ee of ec, ha po raed 0 ee stration x “href a, we ents. Une itaoson ape 4" 2 Caren oo lipid fue + + aie + oon) =F 0 son pr arn single srr ed ye ling nt ab repose pupa fora sgue, sown nF P52 ‘reser Sunt et ate 0 par web, yay Fe ® Sil sear: verti yan Poh 4 it in at, they ate smi. “the wo transformations apese quite diet in he igure. But ft hey ‘The Siri shot shown by considering the strain lipids a a am ofrantn fw inthe ye By vy, Ghat y= 1 sme a lis e+e © weeBen, whereas by Eq, (the same clei transformed into ano lie a day + (U4 y= o Let ur simply Ea (6b rt of couriotes. By Eq, (42) ify raed 1 & sythroagh an angle 6, we baw = Eoosd—-nsind, y= sino + near d. o (Oa sbstitting ino Eq. (6) and sping, we btn ton o + Arcx asin 01 (1+ Ara? + wit ay sin cos 0 + (1-4 End 6] + Gof -2ens Bin & 0 4 Aso @ — sina) + Deas sin CL + A] = “The enti of vanishes = —eat 28, or 8 = —F tan (I. Wi ei vale of the coefficient of in Eq, (6) becomes cs 9 — Deot 20 sin 8 + (1+ 4 co? sin 8. 20s 28+ cos len? @ = tan? 814 ‘Anaiysis of Deformation Chap. 6 Sinilly, the coeticient of +n Eq. (8) can be rece to cot 0, Therefore, Eq. (8) becomes tan! 6 + oot Oxf = 1 o ewe wit k = tan 0, then Eq, (9s reduced exact to Eg. (5). Therefore, these two ‘srainelipsods are equal; one is rotted fom the ote by an angle @. Ths verifies the ‘equivalence of pure shear and single shear To find the relaon between k ands, we note that Sn 6 ” Dein Deon ~ 200? ~ tn ‘Therefore, sine s~ ~cot 20nd = tan 0, we have aft + and k= + VEFD. 9 53 A ste ppe of length 60cm, canetr 6 cm, and wall thickness 0.12 cm stretched 0.010 om ail, expanded 0.00 em in damter, and twisted trough 1. Determine the stan eomponen' inthe pipe 4 For the tras shown in Fg. P54, determine (4) The lads in the rod (0) The stress inthe rods, (6) Assume a onedimeasionalsrestai latonsbipe »o/8 forthe rds, nd ssuime that Youn’s modulus fortes 207 GPa (3x 10"psi. Determine the longitudinal strain ein the rods. (@) Determine the displacement vector a th poiat of lading, B. Ansiver: (0) = 503 MPa (7200 If’), one = ~88.2 MPa (— 12,80 In) (0a = 24 X 10? ge = 425 X10 (6) 0.640 cm (0252in) Solution: Te onsite rds ae dtrind by sac equim, a in Chaper 1, We atin exsn of 628 kg (V2 x IT) in AB and eompreion of 58g (1000) in BC The sees re ined by ing teas yt sos selon seal the members After rion y Young's mots gs the ine = 2.05 X10? andy = 045% 10 Toceimie he dplacennt Bw ote tht thes os ae in end. Asa comune of sorting the ed BC, he on Bmore t thee ut the od Bean ‘ing aun 0 tat he cs oft pombe laeston of es on an aro lth Cas cetera Caras For ver ol BC ( ompaed vith BC), hs bas 2 ne segment perpen to BC. Say, th or AB extends 8 AB in eh an the los of Hon AB ies on an ar perpen AB, The intesction of th, Bis the fea ein fhe place jt B 135 Tension 7 Dspace Fig BSAA spl rus and ametod of devin the— placement tthe iat B. (2) A simple tus. 0) The oc ofthe ‘movement ofthe poi Bon ths 7m BC, ray, BBR" aed that on ‘the arm AB, namely, BB'B*. (c) The area ofthe shaded region is i ‘proportional to the ssn energy sored in bar when it etched i ‘Tocompute the displacement BB, we ee tom Fig 5.40) that aos = fap a = fap FY = fae + Ps Pay = fa + P+ D+ HY (a + GO + Bw + Dry Now,136 Analysis of Deformation Chap. 5 Ne: Akematve Method of Fini Displacement a . Toe work done by the ond ‘sequal the stan enesy store in theres. When aol issubet to gradoaly increasing tein fom zr to 7, is ng hanes by amount el. = TLIEA, whee L isthe length etter and Ais is crossetionl aes The since stored inthe ois elt LPUIAB) (Se Fig, PSA(0),) Now, whe alond Ws paul applied ont the racket, {he work done by its equal fo N8 where isthe piconet inthe exon fhe ad, ifthe vet component ofthe dspeenet. The factors necessary bea, the Str big leary ela, he foreedeflctonrionstp ine, hat the area Under the carve, which eres te wrk doe is of load % éefton, Hee, on uating te wor done wih he sain ery sored, we obtxin 4 Doel, DEA * 7 he (On subsitting numerical values into this equation, we obtain displacement of the joint Bs (8° + & BC)!" = 0.640 en, 53 A rocketauncing towers afetedby thermal deletion nused by nunc beating ofthe rocket unde the sn (Fig, P52). Assume thatthe body ofthe rocket ia circular ‘finder, and estimate the horiontal splacement of the tip ifthe following asump- tins hold: (@) The linear thermal cetcen af expansion is = 107F = 0555 x 10°C. (6) The maximum temperature on te body ofthe rocket onthe side facing the sum is 2AFF hoster than the simu temperature onthe shady sie. (©) The temperature distbuton i wifum soog the length (longitudinal axis) of the rocket, but varies inert along the rai. (@) AS a consequence af), pla sexton ofthe rocket remains plane in thermal expasio, (@) The rocket is unloaded and isi o defor. “int; Compute the thermal strain and then iterate to obtain the defection. 1 pwae 85 om. The total Answer: Thermal strain difference from wo sides = aT = 20 x 10°. Tip detetion = 263cm, Ho. co 0" Orr o Figure P55_ Thermal defection ofa launching tower. (e) Toe Aimensons ofthe rocket. (2) The temperature distribution. chap. 5 Problems 137 156 Derive an expression forthe change involune ofan element of unit vohume subjected Pata ine, Sbow tat the invariant ' = ey + 6y + & fepresets he ehange ‘otvlume per anit yolume when the sain ae small. ‘Solution: According to Se, 5.7, we an fina se of etanglar Cartesian coordinates with expe © which the strain esr assumes te frm eB, (knot summed), where Tee the nial tans. Let ws considera dy undeaoingstsn and chose uit ‘eye tose edge are eniented along he princi! axes of rin Each ede, oigiaally of length 1, becomes 1+ after defomation, The new vlune i, theefore, (1b ep(t 4 el #6) = Ate 9 a + higher onder terms Hence, on ignoring th ihe ore terms, es tha fe chang of volume pr wit vole se, tet 6 “We know from Bg, (6.73) hat = 68, invariant Its equal to + es + és with reference tothe propel aves. Hence, 1, = e+ + 6 wih reference to any Catan Tasoatelerenee Ths, J = 68 meas the change of volume per uit olame when he strain are sal. 7 Given tes field oy with eomaponents refered to a system of coordinates X35 (a) What is the deiniton of principal stress? (0) What isthe definition of principal xe {6} Dest bry how the principal dso (Le. the dion of tbe pil ‘ne ean be determined in pipe (@ Comer a stain fmsr¢,refened othe same coordinate axe. How do you eee the principal stains andthe comespondng pincipal directions? {gt theses and strain esos ae related bythe elation y= ead + ey ‘he principal axes of tess coincide with he were Mand pare constants, prove t Principal axes of strain. 38 ina soy of earthquakes, Lord Rayleigh investigated a slaon ofthe linearized equations of elastin the form w= Ae exit ~ e Be exlitte ~ a ‘ie plane x ereents he ground, wiley represents he depth nto the sath, and aay ware the dsplaenent ofthe pares othe earth then Rayhgh's sation Tepacts a wave propesing the rection wth 2 speed cand an amp that “Fercaseaponataly rom the ground surface. Te waves assumed te pnerated inc cont The ground surface i fe; the tes vector ating onthe ground vaste ero After checking the equaions of motion sod the Boundary conitons, Rayleigh found the contuns A,B, , ad and obtained the solution Ae ~ 0.57 cas ke ~ Ede AC 0garse + L45T9e 9%) sin ~ Of) wea108 ‘Analysis of Deformation Chap. 8 “The constant ci the socalled Reylegh wave sped, whichis equal to 09194 times the sheer wave speed if Poisson's rato ij. Tis solton satis the conditions of « wave propagating ina seminfinit elastic sold witha re srface y = 0. The partes ‘move inthe xy-plene, with amplitude decreasing a the distance from the fe surface increases (ee Fig. PS). The Reyeigh wave represents one ofthe most prominent waves that canbe see ona sesmograph whe theres an earthquake. (@) Sheth the waveform. (@) Sete the path of motion of partes on the tes surface y = Oa several values (of Do the same for several parses tdfeent values of y > 0. {@) Stow thatthe mtion of he paris sro {(@) Determine the places where th maximum principal strsin occurs at any given instant andthe value ofthis stain. wae ae — oe sas Ne gue FSA Rayigh sure wave. arta! Solution: {G) Since w = 0, only the sin components eg a0 eaten. The exponential function e-, wth b> 0, shows thatthe largest values of uv, w and their ‘erivatves will oceur at y = 0. Ox tis plan anda = 0, we have ae = Bm —AR(L ~ 0578s, ao = at? ~ 1469 x 03985 in, 6, = AE (07s + 0573» 039) + (08S + LAD cos kx = 0. eno, the maximum picgl stains ae a= ORDTAR, Gy = 20.1084 AR 159 Comider a square plate of wit se defor 3 showa in Fig, PS. Find the stain compunen Solution: ‘he deformation can 3 desi by the fing equations: 1 neat aee, het Oy Gene BE Oe Chop. 5 Problems 139 a pe “ Ay Rigo P59 Deformation of a "| sare plate fe) -sJa 2a (B) -(B)- b= Gafbe—aseziaws[-o-fe a+ (Sm Sad ~ ba ‘uhereas all other components of strain ae 10. 5:0 Consider the square plate apan, bu his tine shear othe sgh only 2 very small, mount, s0 that Fea +00, gsm 00x, ma, =a, ‘Thea a ~ dij = 0.01 da, do, + (0.017 dof 0.01 dd, ~ (0.01 Heooe, Fy = 0005, Ey=5X 10, ey = 00005, eg ~5 x 10% Ta this case the By and e, measures are approximately the same SL A square plate is deformed uniformly rom configuration (a to configuration (8), 26 shown inthe thre css in Fg, PSL Determine the strain components, Ez, Ea and ean Ene Answer: The transformation tat leas trom configuration () to oonigtatio (6) in Cue Lis x, = Lay, = 120,15 = a. That in Case 2s x, = 120, + 050% = 1.2453) = Ia Case 3, we have x, = 1.014, + O20 25 = Olay, X= o From these, the sain components ate obained from E, (53-5). Case 3 qualifies for “infinitesimal” strains, as given by Eq. (63).“0 Anais of Deformation Chap. 5 2 ve aia) 4 : ; 1 come ‘ a) 4 — ° i ee wna) 7 i coe? | ae a Via a 1 oos 12 \ un 2 a02, Lo q u (108, LOH) coe’ } a ® o ol 4 — 0 ee —Tor [igure PSL. Three patterns of deformation ofa square plat, S312 A unit square OABC i dorado 4'B'C' in thre ways, a8 shown in Fig, 5.12 Th eath ofthe cases, wite down the displacement field uu of every pint in the square a futon of the loafon (of the pin in the original postion, Then ‘termine the sain Ey ¢y, Awame that uy = Oand that, ae independent of Sand a, In Cases () and (¢), assume that the lengths of OA, 04", OC, and! OC de all 1 Aso, obtain the Simplified expesions of the strins e 6 # ane mais ae 7 ate 4 4 6 ety y, a Figure P52. Dsfomation of OABC 10 OA'BC. 13 A wit square OABCis fst sbjcted toa steching, as sbown in Fig. P5.12(), then to a distosion, as shown in Fi P.12(0), and finally toa coatio, as shown in Fig. '5.12(¢). After he tre steps 'nsuxssion, what ar the values ofthe strain ye Chap. Problems at Answer this problem fist for fie vaués of, and then for ifinesial values of, 8, 4. 5:4 Find the strain component E, and ¢, ven one ofthe wedges in Fig. P54 i tone fovmed into the othe. The fit wedge has an apex angle of 0 theaters 9. The radi ae the same. 4 Ben ‘S45 Let ABCD be a nit square inthe 2-plane (Fig. P.15). ABCD is a ptt ofa large ‘deformable body subjected toa small rain hati uniform nthe entire body and is siren by h 24 1 ox s0% 0 ‘Whats the change in length of the ines AC and AB Answer: AC changes by 0.00423; AE changes by 0.00290, Figure P5.15 Change of length of line egments ia plate of known, strain ‘516 A square membrane, ~1-=.x=1, “1s y =, is stretched in such a manne that the displecement is desided by waasey), v= bo, wee ‘What are the strain components at (x, yf What the principal san at he arin (0, 0), Assume the constants o be infinitesimal, ‘517 A pimoined tras is sbowa io Fig, P17, where Lis the length ofthe vertizal and ‘orianntal members, The cross-sectional area of al he member ithe se, namely ‘A. The material of ll the members isthe sme, with Young’ modulus E. Te truss is loaded atthe center by a load P. What would bs the vertical deetion ofthe point under the loa? | Answer: Solve the problem by the stai-enery method ilstrated in Prob, 54. The Galeton it SSORPLIAE.a [Analysis of Deformation Chap. 5 7, ‘igure BST Cakalation of he 7 vertical deerton of jin of ‘5.8 The folowing may happen in¢ number of stations, suc as the flow of water, the foning of metal, and in cell membranes. The materia is incompressible. The dix placement component in the z-ection vanishes. The dsplcemeats wv re infine Aeximel andar fnetions of yf, in a cetan domain, we Know tht ua yyet br to, here, 5, eae constants, compute the dplasement inthe section. Hit Use the facts shown in Prob. 5.6 59 In the problem ofthe torsion ef indica! br of steel with an eit cross sexton (Fig 5.13), i was found thatthe displacement can be deserved by the equations vere ais th ange fbi rains per went ofthe bar. Let «= 2am and T= Lem, Comput th striata act at the point A (2 = 0, y= 8) What the tani hear am at A? On wat lane do the maximum tesog and maxim shea ae? Partial Answer: ae 8 8 ay = Op max sea stain = ~S max noma stain = 5 igure P59. An lpi cross sexton, 520 By cifecaiating an arity ataltc function fy 3 5), on can obtain is placement field wu, 1) fined by the equation & ee chap. Problems “a What are te strain components and sain invaiants ofthis placement fl? Consider a special case b= 54 DR where Regedeg and C,D ae onsans. Apply itt a bolo sphere whose inne rads is.a and ont ‘adi ib What are the ales ofthe trans inthis phere? 5:21 Figure P5.21 is another clas éranng trom Borlis 300 year ol book (se Prob. TLE). Here is shown Borei'sobsenatins onthe arangementof msc bers on igure PSL cls raving by Borel the surface of the heart, Muscle fibers contractor Ingen, change the shape and ‘rin of the organ. Geometric rcagemeat of fibers bas a great influence 00 the Focion of the most fa muse fshaped ike apralletgram shown the “simple hear” ase of Fi. 5.2 (Problem 3.2, with Sbers paral tothe ined borers thon a contraction ofthe mle whch furos dh muse block ito arestangle would Jeane the volume aod with uochanged. On the oter hand, i the muse fibers were paral othe y-axis ax shown in Fy. P5.2s “pure shear” case, then on contraction fhe wid ofthe bundle wil bulge ot. When the heart works t pump Hood, the ‘chambers ofthe ft ventricle end ight vente should expand and contact, ut the ‘myocardium does not changes volune, nor bulges out lal (alike the biceps lexor tule ofthe frat of the upper am). Boreli thought he knew how it could be done.us Analysis of Deformation Chap. 5 “Te up ths suggestion and discos it Add therefor epesientl deta far ts you can, Moder data on bear mule fier strctue ceo be fund in Ses, D. Jn. "Graxs morphology and fiber geonety of tbe hear,” In Handbook of Pysolgy, Ses 2 Cardianaedar Sytem, Vol. 1 The Heart (Berne, R. M, and Sperlaisy N., tis), Amin Physiological Society. Bethesda, MD, pp. 6-112. Expand the obser vation to explain how acebo ostercan move ts powerul laws by mses encased ina sig sel, What kindof strains hese mules must have VELOCITY FIELDS AND COMPATIBILITY 6 CONDITIONS We shal consider the velocity field and define the strain-rate tensor. Then wwe shall study the question of compatibility of the strain components or the stran-rate components. 6.4 VELOCITY FIELDS For the study offi flow, we are generally concemed withthe velocity l,i. i vith the velocity of every paticle inthe tody of the fui. We refer the location ‘ofeach uid particle toa frame of reference O-1yz; then te fed of ows described by the velocity vector field v(x, y, 2}, which dens the velocity at every point (2, 4,2), In terms of component, the velocity field is expressed by the functions ve ¥2)s if index notations are used, by v(8i 2 %)- For a continuous flow, we consider the continuous and differentiable functions | (gina). Tere are occasions, however, in which we mus study the relationship between velocities at neighboring poins. Let the particles P and P* be located ‘nstantaneously atx and x + dx, respectively. The difference in velocities at these ule, Ys 2), m2) i to points is (644) where the pata derivatives [ay are evaluated atthe particle P. Now, a on m - im -2 (63-2) eae oa 456 \Velociy Fields and Compatibity Conditions Chap. 6 Let us define the rat-of deformation tensor Vand the spin tensor My a8 Man 34) I vel 63 tam _ 3 nai 2| f Then Vi ~ (645) ca Itis evident that Vis symmetric and 0, is antisymmetric; Ve= Vn My = —Oy 616) Hace, the © tensor has only three independent elements, and there exits & vector 9 dua to 4 that is, Mees ie, = curly, ry ere ey the permutation tersor defined in Sex, 23, Eg, (23-16) The vector ‘Dis ead te vorily veto. Tees (elt ad (1) ae sir to Bas (55-3) and (55-8). Their geomet retains re ao Sn. Therefor, th anal the vloy Bar rea mh He the ana ofa ines deformation fed need, rasan ty enna interval ftine dt reais an infitesina eee ae yd Hence, hater ve are bo be infiniti tain ebb medal eens eorespnding tthe ra of age of tain, With te word lot lacing the wor dplceent. 62 THE COMPATIBILITY CONDITION Suppose we were given ase of two partial differential equations for one unknown function u(r, y, such a8 au iu gue F ‘We know that these equations cannot be solved: We have too many equations that axe mutually inconsistent, The inconsistency can be clarified if we compute the second derivative auld dy from the two equations: The fis yields 3, the second 2x. They are unequal ‘Therefore, when patti # 21) Aitfrental equations are given, the question of integrability arises. The difereatal equations aw fem Baas (622) Sec.6.2 The Compatbilty Condition 17 ‘ano be integrated unless the condition £8 a 23) is satisfied. Ths condition, a condition of integrability, is aso called an equation of compatibility "Now consider plane state of stain, such as may exis inthe solid propellant main of a rocket. Suppose that an engneer made a Iboratory model and obtained, by various instruments each a stein gauges, photoelastic equipment, lser holo rephy combined with Moire pattern alysis et.,a set of strain data that may be presented as fey), y= 8D (624) ‘The question arses whether the data are self-consistent. Could the consistency be checked? And if they are consistent can we compute the displacements u(r, y) and v(, y) from these data? If the strain is small the last question can be formulated as a mathematicel question of integrating the diferental equations ta) » y my ata (62-5) (= 2m) Now, if we differentiate the first equation with respect to y twice, the second with respect 10x twice, and the third wit respect tox and y once each, we obtain ue ay a aay yy” 624) mu, ey ah ayy ay (62-7) Subsituing Eq. (62-6) into Bg, (62-7), we have fia ay 628) “The experimental data mus satisfy ths equation. IFno, the data are not consistent and there must have been errs.