Elasticity and Geomechanics Pe NN ETC) A..P.S: SELVADURAI ~ beanpea| TE] : TCI prom lw tip] : be‘This book provides a thorough examination of the use of elasticity in solving geotechnical engineering problems in a style that will be accessible to upper level students in civil engineering, geological engineering, and of the earth sciences, The first two chapters present a basic framework of the theory of clasticity and describe test procedures for determination of elastic parameters for soils. Chapters 3 and 4 present the fundamental solutions of Boussinesq, Kelvin, Cerrutti, and Mindlin, and use these to formulate solutions to prob- Jems of practical interest in geotechnical engineering. The book concludes with a sequence of appendixes designed to provide the interested student with details of the theory of elasticity that would be of considerable assistance to a deeper understanding of the main text‘This book provides a thorough examination of the use of elasticity in solving geotechnical engineering problems in a style that will be accessible to upper level students in civil engineering, geological engineering, and of the earth sciences. The frst wo chapters present a basic framework of the theory of| elasticity and describe test procedures for determination of elastic parameters for sols. Chapters 3 and 4 present the fundamental solutions of Boussinesg, Kelvin, Cerutti, and Mindlin, and use these to formulate solutions to prob- lems of practical interest in geotechnical engineering. The book concludes with a sequence of appendixes designed to provide the interested student with details ofthe theory of elasticity that would be of considerable assistance 10 a deeper understanding of the main texELASTICITY AND GEOMECHANICS.ELASTICITY AND GEOMECHANICS R. 0. DAVIS Chivers of Camertary A. P, 8. SELVADURAL MeGit Universiy | Branca DA CROLET NSTTUTODENGENEAK ge 0. 8 1997 ¥ 0 NSE sti e Be FOO RIDGE OGkannss‘bid bythe Pes Spee of the University of Cambie “In Pit Buiing, Trmpingtn Sect. Cumbedge C2 IP "a Wes ih Sear, New Yer. NY IDOI, USA 10 Sunfo Row. Gach Melo 3165, Aus Contents | SOI (© Cambie University Press 1996 | in ps 1986 | td in th Unie Stes of Areca Library of Congress Catalonia Publication Dat Davi. R.0. atic and geomschanes RO, Davis A. .S, Sau pcm lacus biiogrpilreerencs ISBN 02185067 (hel = ISBN 0521-49027-9 (bk) 1: Engng eslog) ~ Mathemaes,2-Ehsty. stad 8. IL Tie Preface oe sate TAMEDS 1 Some ideas from the theory of elasticity 624.1'51 ~de20 os.r07 TH tect! 7 1.2. The notion of a continuum 4 ‘Acatnlog recor fr this book i valable rom the Bits Library Ga ‘Defomalinne of conde ; ISBN O-S22-4950b-7 Hardback 14 Deformation and strain 7 ISBN D-52]-49827-4 Paperback 15 Volumetric stain 7 | 1.6 Compatibility of strains 15 LT Stress 16 1.8 Principal stresses a 1.9 Stress invariants 2 41.10 Equilibrium equations m4 LIL Formulation of problems 26 Exercises 30 References 3 2 The elastic constants 34 241 Hooke's law 34 1 2.2 Young's modulus and Poisson's ratio 35 2.3 Shear modulus aT 24° Bulk modulus and Lamé constants 2 2.5 Stress and strain deviators “4 2.6 Relationships between elastic constants 48 2.7 Bounds on elastic constants 48 \ 2.8 Determination of elastic constants 32 2.9 Laboratory tests 3B | 2:10 Field tests 6aut Contents Incompressible elasticity Brercses References 3 Fundamental solutions 31 32 33. 34 35 36 37 ‘Roussinesq’s problem Flaman’s problem Kelvin’s problem Cerruti’: problem Minatin’s problem Other fundamental solutions Gravity stresses, stress functions Exercises References 4 Applications of fundamental solutions 4 42 43 44 45 46 47 48. 49 4.10 Intoduction Uniform circular load on homogeneous half-space Uniform loads of other shapes, homogeneous half-space "Non-uniform Toads, homogeneous half-space Rigid foundations lane-strain problems Settlements in layered sols ‘Consolidation and settlement “Applications to in situ testing Gravity suesses in earth structures Exercises References Appendixes oumcom> Index. ‘The compatibility conditions Cauchy's stress principle ‘Traction vector on an arbitrary plane Uniqueness of solutions in classical elastic Saint-Venant’s principle Principles of virtual work [Betis reciprocal theorem ” n i 6 6 4 88 31 98 100 107 10 12 m2 ns 123 129 BI 138 145 148 1st 155 162 165 167 167 im 13 176 180 188 191 199 Preface “The linear theory of elasticity has enjoyed a fily long and profitable history in the field of foundation engineering. Geotechnical engineers have tumed (0 elasticity for answers to variety of questions, and despite the sure knowledge that the answers are at best approximate, they will continue to do so for some time, The reasons fr this le in the essential simplicity of the relevant elastic solutions. The point load solutions of Boussinesq and others, as well as the dis tributed oad solutions that derive from them, can provide answers with only @ few lines of calculations, This can be a significant advantage when compared With the ime and effort involved in obtaining numerical solutions that employ ‘one of the multitude of existing plasticity models for soil, Furthermore, quite a lot of information conceming soil properties is required to run any of the vari- us plasticity models In many practical situation, che information may simply be unavailable, and the geotechnical engineer is left with few alternatives. In contrast, elasticity solutions will generally requir ony a value forthe soil mod lus and Poisson's rao, and if the values given are known to be rough ap- proximation, then at last the solution method isin keeping withthe input dat “This book is about the use of elasticity in solving geotechnical engineering problems, Its directed toward upper level students in civil engineering or en- sincering geology, It was motivated when, in 1992, Patrick Selvadurai visited the University of Canterbury. There he found a course taught by Rob Davis, similar to « course Patrick himself taught at Carlton University. Both courses had evolved independently but were suprisingly similar in content, covering some basic applications of elasticity theory in geotechnical enginecring. “The book grew from lecture notes and is intended solely as a teaching tol tis not exhaustive in content but is suficienly complete to provide a grou ing in the linear theory of elasticity, together with an understanding of appli cations in foundation engineering. The presentation is deliberately informal ‘nd conversational in tone. From the standpoint of a student, it should not bex Preface intimidating, but should hopefully put a variety of new ideas into an aecessi- bie format Pour chapters form the main body ofthe book. Chapter I presents a frame- ‘work of asic ideas from the linear theory of elasticity: deformation, strain and stress, equilibrium and compatibility, and the formulation of problems. (Chapter 2 delves into Hooke's law and the elastic constants. The first half ofthis chapter describes the elastic constants and explains thei relation to the material in Chapter 1, while the second half discusses how elastic con- stants may be determined in a geotechnical context. In Chapter 3 the point load problems of Boussinesg, Kelvin, Cerrut, and Mindlin, as well as the line load problem of Flamant, are described. The solutions to these prob- lems are presented but techniques for finding solutions are not discussed ‘The thrust of the book is not to provide another elasticity txt. The point load problems are fundamental to geotechnical applications and white their solutions are examined, the corresponding solution methods are of only mat- ‘ginal interest, Chapter 4 uses the solutions from Chapter 3 to consider some basie problems in foundation engineering. The development progresses from ‘consideration of a simple uniformly loaded region on the surface of a ho- mogeneous elastic half-space to more challenging. problems involving nonuniform loads, rigid foundations, and layered half-spaces. Some remarks concerning consolidation and in situ testing complete this chapter. Finally, 4 sequence of appendixes bring the book toa close. These are an important ‘component of the book, designed to provide the interested student with de~ tails of elasticity theory that are peripheral to the main text, The compati- bility conditions, development ofthe stress tensor, Saint-Venant’s principle, ‘uniqueness, the virtual work principe, and reciprocity relations are all con sidered in the appendixes. ‘The book draws together material from a range of sources. The deserip- tion ofthe linear theory of elasticity is phrased in a physical rater than purely rmathematial context, While mathematical formulations are not avoided, they ‘are construed as the end result of physical thought processes. Examples are used to illustrate important concepts, and a set of problems is placed atthe ‘end of each of the four chapters. The book does not directly compete with any other text, While nearly all undergraduate soil mechanics texts contain some elastic analyses, none are devoted to elastic theory. Two other books, Poulos and Davis’ Elastic Solutions in Soll and Rock Mechanies and Sel- vadurai’s Elastic Analysis of Soil-Foundation Interaction, bath involve ge0- technical uses of elasticity; bu the former isa catalog of solutions, while the latter is a treatise, Neither would easily serve a a text for undergraduate stu- denis. Preface x Finally, we would like to express our sincere appreciation to two people ‘whose assistance made preparation ofthe book a far more pleasant task than it might have been. First, all the text and equations were typed with remark able care and accuracy by Mrs, Pat Roberts. Second, the figures were skill fully drawn by Mrs. Val Grey. To both these people we offer our thanks for ‘job well done. R.O. Davis ALPS, Selvadurai Christchurch ‘Montreal July 1995 July 19951 Some ideas from the theory of elasticity 1.1 Introduetion In this book we will use the linearized theory of elasticity to obtain solutions to some specialized problems in soil mechanics and foundation engineering. For beginning students in geotechnical engineering this may seem ikea strange ob- jetive. One ofthe frst things we lean about sol response is that, in general, itis neither linear no elastic. The stress-strain behavior of soils usually Found to be hysteretic and highly nonlinear. It seems git reasonable then to rae the question of how linear elasticity ean be profitably used in relation to soil “The answer to this question is two-fold, First, elasticity will be a conve- rient tool, and, in the busy and pressurized world of geotechnical engineer- ing, convenient, reliable, and speedy results are avery postive advantage. The Tinear theory of elasticity possesses a long history, which has been distin~ ‘uished not only by frm mathematical foundations but also by the solution ff a large number of useful practical problems, Some of these solutions are particularly well-suited to the types of loadings and geometries we encounter in foundation engineering, To have a ready-made solution that either can be tied immediately, or with only minor modification, is clearly » significant advantage. However, ifthe only advantages the theory of elasticity hd to of fet were convenience and speed, then it would remain an unused tool in the ‘geotechnical engineer’ tol kt. The second reason we use elastic theory is because most sols will behave approximately like a finer elastic materia, provided the stresses they are subjected to are relatively small. By small, we Inca the level of shearing stress within the soil deposit is considerably lower than is ultimate strength. Ths is often the case in many problems in founda: tion engineering, Usually foundations are designed with factors of safety of three oF more. Tis suggests the stress level, ina general sense, is about one= third of theultimate sol strength If we consider atypical stress-strain curve fora fine-grained soil, such as ilustrated in Figure 1.1, we might well con-Some ideas from the theory of elasticity Pak stress = 0.8890Pa Strose Deviate woe 00h CO Figure 1.1 Stes deviaor vers aval stain when from resus of anata! drained, ‘tres conrlled ails on uniform beach sand. Noe inca of inl stess- Stain response. clude the behavior is at least approximately linear over the lower one-third of the peak stess range. Thus, for problems in which the soil is not subjected to stresses larger than about one-third the ultimate strength, elastic theory is not ‘only convenient but also offers a rational approximation to the load- deformation behavior we may expect. ‘There ae two other attractions in geotechnical engineering associated with clastic analysis, Elasticity solutions will often give insight ito the “mectan- jes” of problem which might not be available using other methods. An un= ‘derstanding of elasticity and its use in geotechnical problems frequently has led to new and innovative solutions to old problems. This is particularly true in certain fields soch es insite testing of sols. The second atraction elastic~ ity offers is that we may use elastic solutions as a check on more sophis cated computer-based solutions. Computations methods can offer an ex- tremely wide range of problem solutions but i is always comforting forthe computer wser to see thatthe program uilized gives corect answers to clas- tie problems for which the exact solution is known a prior. “The types of geotechnical problems where elasticity will be useful are rostly confined to foundations of structures. Some typical examples are il- Justrated in Figore 1.2. A great many structure are founded on reinforced con- crete footings oF pads buried at relatively shallow depts Beneath the ground surface. For these types of structures, elastic theory is well-suited to yield es- ‘imates for both the stresses inthe foundation soil and the displacements or Introduction ple Applied load tae] np ented ave ‘Agog fala satire Pressure _e Stats 2 round anchor Figure 12 Some foundation engneeing problems tht can be fat take wsing slase analysi, setlement of the structure itself, Elasticity is also a powerful tol in assess ing the way in which the stuctue interacts with te soil Sine both the struc ture and the sil are deformable, but may have diferent stifinesses, this prob Jem, referred to as the soil-structure interaction problem, leads to many interesting results. Practically al text books on foundation engineering contain some mates ‘based on elastic analysis. There are wo books, however, which ae of pari ‘lar interest. The first book written by Harry Poulos andthe late Ted Davis called Elastic Solutions in Soil and Rock Mechanics, published in 1974. This book fs a compendium of solutions gathered from various sources and cata~ Ioged for easy reference. The second book is by Patrick Selvadura tite Blas- tie Analysis of Soil-Foundation Interaction, published in 1979, This is an ade ‘vanced text giving a very complete description ofthe soil-struture interaction problem with formulations and solutions to many problems of practical in- terest. Both ofthese books ate excellent sources for anyone interested in Us ing linear elasticity in geotechnical engineering. However, both are different from the book you are now holding. This book is designed to introduce both ‘ndergraduate and graduate stadens tothe subject. Itis not exhaustive in con- tent, and hopefully not exhausting to read. It will touch on a few interesting problems, but omit others. And it will give only a cursory description ofthe development of elasticity theory. Where we think slightly more detailed ex 74 Some ideas from the theory of elasticity positon may be of interest to the more mathematically minded reader, i will be presented in Appendixes ln tis text we primarily want to concentrate | how elasticity can be used rather than on why it eame about. Heace, the te ‘of this chapter, “Some Ideas From the Theory of Elastic what it ays 1.2. The notion of a continuum ‘The linearized theory of elasticity isa branch ofa large discipline called com- tinuum mechanics. A very luid definition of continuum mechanics was given by A.J. M, Spencer in his 1966 Inaugural Lecture us Professor of Theoreti- ‘eal Mechasics at the University of Notingham, CContimum mechenics means mechanics of continsous mia, and & continous ‘modium isa medium which occupies every point ofa continous region of space. In Comtinuum mechanics we treat oar materials ~ asc, ligude and solids ~ ax though they completely fill up te regions they occupy, with no oles or gaps. Thi is ot what ‘matrals are really Hk. They are made of molecules, at aoms, which in tn ae ‘male of stalls pales, and even these particles ae nl pats inthe seat of be ing lite hard lumps of mate. By far the greater par of any pecs of sny material empty, and onthe atomic scale the chance of any’ parca’ pint hein occupied by ‘mater is small Cominuum mechanis pretends ignore al his, ad nelle scars the atoms, molecules and soon smoatly and ifr over the regon wich we sp pose our material to occupy. In doing his, of course, we me once agin fckicing ‘mathematical ideation; we ae replacing a her complicated dea of what materi als are by ales acurate but much simpler one ia sch the sume way we make 3 idealization when we iat srs and pins us point masses when making astronom lca eaeastios. Thao are two jusifctons fr making this deaiation. The ist and mos importat that t works, We can perform cena experimcats On & matron the bass of theory ad hese experiments we cen make predictions about how them teria ol behave in oter experiments Let us now examine the relevance ofthe notion ofa continuum in the con- text ofa sol. If we were to divide soil into locks of one meter on aside we ‘would intuitively expecta typical block to represent the constitution of the soil in some average sense. If we then continued to cut the block into smaller and smaller paris we would quickly find @point at which one part might con- sist entirely of a particular soil mineral such as quartz, while another part right consist entirely of water, or for that matter, another mineral such as feldspat. So technically a soil can never be a continuum, but this is not a se= rious obstacle. What we must do in order to treat soil a8 a continuum, of a6 Spencer remarked "to make it work” is to agree to apply our results only to volumes sulliciently lage enough to encompass significant numbers of sil Paticles. This will elearly be the case for any reasonable problem in founda Deformations ofa continuum 5 tion engineering where the characteristic dimension of a foundation will be of the order of meters, whereas the characteristic dimension of soil could range from 2 pm (ea clay) to 0 mm (grave) (One consequence of dealing with a continuum is that we can use the concept of limit, For example, ina continuum the mass density pis defined as a limit Au eB av Where AV is a volume and AM the mass of material contained in AV. Inthe limit process we shrink AV toward zero volume around a particular point. We will identify the point by its postion in a three-dimensional Cartesian coor- inate system, and we'll denote the position by x. The components corre- sponding to x may he (x,y, 2) in a rectangular coordinate sytem, or (*, 8, 2) in a cylindrical coordinate system, or other coordinate names in systems of other flavors. Since the volume AV shrinks up to te pointx, we say p= px) In mote general problems, time may aso be involved and p = p(x). The only reservation we need in order to apply all ofthese ideas to soil is that, rather than take the Himit as AV—> O, we will agree to conser finite volumes and restrain the limiting procedure to AV— AV,, where AV, is sufficiently large to contain a significant number of sol particles. ‘The concept of a “permissible volume” AV, becomes quite important es- ‘pecially when dealing with the mechanical testing of soils for the determina tion of their strength and deformability characteristics. For example in the ti anil testing of soils such as clas, and sand, the dimensions ofa cylindeeal sample ofthe soil can be 75 mm in diameter and 150 mm in length. Clay, ‘the same sample dimensions will not apply when testing granular materials ‘such as gravel with particle sizes of the order of $0 mm. Here, the sample ‘size must be substantially larger if we are to be assured thatthe results do- ‘ved from the experiments conform to our notion of & continaum deserption, 13 Deformations of « continuum ‘When dealing with continua, one important aspect is the description ofits de- formations. We will need to be able to precisely describe the deformations that a continuum may experience as a result of action from outside forces, ‘The tem deformation relers both to the mation of a particular particle inthe ‘continuum and to the overall motion ofthe continuum itself. To be a litle more specie, suppose our continuum isa body, 8, which we could ilustrate a8 a generic potat-ish shape in Figure 1.3, Later on we will be much more specific about the configuration of the body, but for the time being the gen-6 Some ideas from the theory of elasticity Figure 13 The reference configuration B, al de deformed configuration B, fora ener elastic body. cal shape is all we need, We will let the body have a reference configuration, ‘and this willbe its shape and positon when itis at rest, fre from load, We ‘ill denote the reference configuration by B,, We can also tak about its de Jormed configuration, alter the loads have been applied, and we shall denote ‘this by Bi, The subscript r here refers to time. The configuration B, ives ws the shape and position of the body at met. We may need to be specifi about + in eases where the applied loads vary with time. In Figure 1.3, the reference tnd deformed configurations are both shown and are linked by « generalized ‘dea of deformation, shown as an arrow Linking the two configurations. “How can we precisely deseribe the deformation? One way to accomplish this is to introduce a vector, called the displacement veetorw, which joins the positon of a particular particle inthe reference configuration to its position inthe deformed configuration. typical displacement vector is illustrated in Figure 14. If we define a displacement vector for every particle inthe body, then the complete set of vectors forms a vector field an we ean write w= ule, ca) to show that w depends both on the position x and the time 1 Itisalso convenient to use an indicial notation wo presen the dependent and independent variable encountered in the presentation of basi results, Forex ample, in indicial notation, the position voctor is denoted by x; where the sub- soript¥can take the values 1,2, 3, and for convenience we can denote 1 = 1ysas = 2, Consequently, the displacement vector w can be written as 1 = aay (1b) ‘In using the indicia notation we employ the summation convention adopted by Einstein in that if two indices are ropeatod in an equation, summation is Deformation and strain 7 Postion of patile °° ral Figure 1 The dsplcement vector u joins the position “A for a particle nthe re: ‘rence Coniguation tos postion "A" nthe Geformed configuration carried out over the repeated index unless otherwise explicitly stated t0 the contrary. Tis natural to wonder about the position x in eq. (1-1). Is te position of the particle in the reference configuration, oF is it the position taken by the price in the deformed configuration? In a general development of contin- ‘wun mechanics, in eg 11) would be the pasiton in the reference config- tration, but, in the ines theory of elasticity, we assume the deformations are small s0 sal thatthe distinction between the position inthe two configu ‘ations wil not be important, That does not mean w isl is unimportant, only that whether we use the reference configuration or the deformed configura- tion positions as an independent variable is not important a is known fora certain deformation, then we really have everything we need to know, The vector field w gives us the deformation of the body as 2 whole, and, if we are interested in a specific particle, dhen u(x.) evaluated for the appropriate x gives us the displacement of the particle as a function of time, Lateron, when we are dealing wit foundations « wil ell us how much the foundation moves. The vertical componont of u, evaluated at a point im ‘mediately beneath the foundation, will give the settlement of that point 414 Deformation and strain ‘We intuitively feel that deformations may lead to strains within a ody. Be- {ore we can consider strains, however, we need to realize tha some defor- ‘mations will not eanse stains. These are deformations called rigid body de formations and they consist of either rig iranslations oF rigid rotations. A rigid translation is any deformation that does not depend on x, Thus, i'w for every x isthe same, the body must be undergoing a rigid translation, Rigid rotations rotate the body about a fixed axis. We ean be more specific aboutSome ideas from the theory af elasticity Figue 15 Any deformation canbe decomposed into sequence of sg translation followed by righ! ration followed by staining rotations in a moment, For the time being the important thing to realize is that any deformation can be broken down into, at most, rigid translation fol- lowed by a rigid rotation followed by straining This idea is shown sehemat {cally in Figure 1. “The difference between straining and the two rigid motion i this: Strains result in changes in length or shape within the body: rigid motions do not Suppose we consider two particles within the body that are quite close 10- gether inthe relerence configuration, Let the line that joins these particles be dx. We can think of dx as a small filament of the material, and we can ex famine what happens to this filament during the deformation, If there is any ‘gid translation, the filament will move, but its orientation will remain un- ‘changed and its length will not be changed either. If there is also rigid rota- tion, the filament will change its orientation, but its length will still be un- changed. Ifthe filament is stretched or compressed in length, then the body fs undergoing straining, We call the change in length divided by the original length the extensional strain, onal strain = ge in filament length ‘extensional strain =" ginal filament length Changes in shape also result in stains. Consider two material filaments dy nd da which, inthe reference configuration, leat right angles to each other, asin Figure 1.6, Ifthe angle between the tvo filaments changes during the ‘deformation, then, even if both filaments stil have their original length, there hhas been straining, Ths is ealled shear straining. The shear sts is defined asthe decrease inthe angle between the two filaments “The question we now need to answer is this: How can we go about sepa- rating the strains from the rigid motions? Suppose we were given a specified ssplacement field u(x). We are aware w may consist ofa rigid translation, Deformation and strain ey eo &, Shoat sain a a, Figure 1.6 Shear strain is defined asthe decrease in an inilly sight angle between two materi laments. ‘a rigid rotation, and straining, Only the strats will result in stresses within the body. If we want to characterize the stresses, we need to find the strains, and that means eliminating the rigid motions. The first step is to look at how tu varies in the neighborhood ofa point; ie., we look atthe partial derivatives fof w rather than w itself. There wil be nine partial derivatives in all. In a ree~ ‘angular coordinate system each of the components iy ty and ul, will have three derivatives, oe in each coordinate direction x,y,z. Ifthe deformation ‘were only a rigid translation, then 1 would be independent of the variables, ‘and all the partial derivatives would be zero. So by considering the deriva- tives we have concentrated on rigid rotations and strains and eliminated rigid translations. ‘Next we need to separate rotations and strains. We ean arrange the nine pat- ial derivatives of w into a matrix called the displacement gradient matrix, Vu may a ay ae x ay a “The mattix for Vu would look a litle different if we were using a eylin- vical or spherical coordinate system, but that need not worry us here, The ‘next step i to decompose Vu into two matrices, one symmetric and one skew- symmetric, The symmetric matrix i called the strain matrix, ¢, and is defined by yu (weit Liv + wh10 Some ideas from the theory of elasticity Here the superposed T denotes the transpose of the matrix. The skew. symmetric matrix is called the rotation matrix, ©, defined by = Livu ~ (Fu). aa Note that if we add ¢ and ©. we get Vu. This decomposition into symmetic and skew-symmetric parts i unique and can be accomplished with any square matrix. Just as their names imply, € will comectly account for the stains in the body and £2 will account forthe rigid rotations. ‘We will write the components of the strain matsx like his and bear in mind that symmetry of € implies Gy = 6x. 61. ~ és amd 6. y- The diagonal components of €, ax €, and € are called the extensional ‘scans. The off-diagonal components, ey, ee. are called the shear stains. In terms of the partial derivatives of u, we have an bem ey = A Me te ee OV ae * ay We will see how these names come about by considering two simple exam- pies. ‘The first example is shown in Figure 1.7. We have a material filament of length de initially aligned with the x-axis. (We can consider dx as just the scalar length of the flament whose vector description is dx = def, where f is the unit base veetor in the x-direction.) This filament joins points A and B in the ceference configuration. In the deformed configuration A has moved {0 A’ and B to B'The filament length has now changed to dx + (dude, Deformation and strain Bae rath = de (age Ba) Length = de (age Qa) - my, ax, = deo Bteee Figue 1-7 An example of extensional strain, as shown in the figure. Then, according to our definition of extensional stexin: ge in ent of lament ‘extensional sain =" sial length of filment (ce+ May) — a ae Which shows we were conect 10 refer to gq a an extensional strain. Ii, of course, the strain in dhe -diroetion, In a similar way we could show ey and ‘cate the extensional strains in the y- and z-coordinat dtectons respectively. “The second example is shown in Figure 1.8, where two filaments of lengths dts and dy are initially aligned with the x and y-cooedinate axes. Inilly the angle between the (wo filaments i 90°. Inthe deformed configuration the an- ‘ale will have changed, as illustrated in Figure 1.8. Eater we defined the shear strain as the decrease in the iniilly right angle between the wo filaments: shear strain = decrease in angle between filaments =atp = tana + tanB. = 2a 5 St, ay * oe2 Some ideas from the theory of elasticity ¥ = tan = SES , ar ay wha ay Figure 18 An example of shear stain "Note that we have un implied assumption that crand fare small angles and hence a = tanee and B~ tap If we compare this result with eq. (1.7) we see that ey i just one-half the shea strain. The reason for it being one-hal is to make the strain matrix e behave properly under coordinate transformation. In similar way we could show that ¢, and are one-half the angle decreases for filaments aligned with the x- and z-uection andthe y- and z-direction Ttis important to beatin mind here that these definitions depend on the pro- vision that the deformations, and particularly the components of the defor- ‘mation gradient matrix Vu, ate small. Only when we have small deformations ‘will egs. (1.6) and (.7) be good representations for extensional and shear inally we need to consider the rotation matrix $2, In component form we have 0 My Oe 22/0, 0 Oe as) 0, My 0 a9 Volumetric strain B _ my ae tomas GE 2-3 Figure 19 An example of sg roton. [ecause ofits skew symmetry, © has only three independent nonzero con ponents, simple example to illusrate the meaning ofthese components is shown in Figure 1.9. We have two filaments dx and dy initially aligned with the x and y-directions. If the body undergoes a rigid rotation, the filaments will be rotated but the 90” angle between them will remain unchanged. The angle of rotation ers, or small ottions, equal to Ax and equal to~ aud. ‘Comparing with eq, (1:9) we see that n= (SE Hare a= ax ay) 2 2 Had the rotation been clockwise rather than counterclockwise, the signs would be reversed and we would have found Oy equal 10 «Thos, My and represent the rigid rotation about the z-axis thatthe body has sustained. Similarly, Qj. = —Qee gives the (clockwise) rotation about the y-axis, and ‘Me™ ~My gives the eoction about the x-axis. Since there are only three de- grees of freedom for rotation, O. contains a eomplete description ofthe righ rotations thatthe body has suffered. Our earlier provisions concerning small ‘deformations apply hete too. 13 Volumetsie strain Changes in length within the body may result in changes of volume. We wit! define the volumetric strain ‘change of volume olumetic strain =" orginal volume“ Some ideas from the theory of elasticity - Tr, 65 + yh — Figure 1.10 Volume strain eis the change ia volume ofa material element divided bys orginal volume, ‘and denote this by , To see how e is related to , consider the elemental cube ‘of material shown in Figure 1.10, Inthe reference configuration the lengths ofthe sides ofthe cube area, dy, and dz. The volume inthe reference co figuration isd dy dz. Inthe deformed configuration the lengths of the sides ‘will be changed as shown in the figure. The volume in the deformed confi uration willbe: deformed volume = (ae + eadsNdy + end)de + ete) dyed + el * Gy) “The volumetric strain is therefore _ deformed volume ~ original volume - ‘original volume 1+ eal + eX ead I eat tt Et baby + Eke t Enter + Cant Now if we recall that all the components of Vu are small, we see that all of the stains must also be stall, andthe products of two oF more stains, such 8 ye ust be very smal, This suggests we can neglect all the products of siruins and have oF eat by tee (19) ‘The volumetti strain is simply the sum of the extensional strains, so long ‘asthe strains ate small. Often the sum of the diagonal components of @ ma- vin is ealled the rrace of the matrix, abbreviated fr. In that notation we have Compariiiry of strains 1.6 Compatibiity of strains “There ae centain mathematical relationships that the componeas ofthe sin rratrix mst obey, They ar called compat conlos, nd physical hey Cres the idea tha afer deformation, te bay is stl contin i he Sense that thas nt suddenly developed any cracks or g9p5 oF any points at ich the mati overlaps elf, Paces that wer originally lose together ‘rll be clove nthe deformed confguraon and, moreovet the reaive po- Stion of sever! panics wil emain unchanged sills in Figure Mathematically we can lok on compatibility this way. I we were given 3 reais placement ed afr), t would be ear tose cs. (1.6) an (1.7) to find te srain components, Bt what f we were ist given the sti and then asked to find the dspacements? This more treating ask beease row we have six equations (6) ad (1.7) to solve fr only re ukrown Components of displacement. Whether or not we can etegrte os. (1.6) and (1.7) to give consistent displacements turns out to be exactly equivalent to the physial problem of ensuring no gaps or overaps have developed. “Te whole problem was solved by the great Freeh mathematician Bate de Sin Vena n 1860, aint Venat showed that the trin components must satisfy these six differentia! equations. ay ae oy , Pe _ yi Pep , Pee ee ae ay? Pde Peay Pen Pes ae ae ase aay Peg, Bee, Bee , Pen aye ae * andy * ade Poy Pee, Peo, Pee aeix ay? | ayie | aye 4 Bae, Pee axay ae” azix © aay “These six eompatibity equations play an important role inthe solution of elasticity problems, but we will not make much use of them, since we will be ‘move concemed with applying known solutions than finding new ones. A de- luiled derivation of the compatibility equations is given in Appendix A.16 Some ideas from the theory of elastciny : . —~ CFE FEED Figure 1-11 The concept of strain compatibility. Poins 1,2, and 3 are found inthe ‘formed conguran inthe same eaonw One anotber as thy originally were in the reference configuration 17 Stress 1 the body is deformed and strains occu, then stresses will be developed ‘within the material. We'll explore the relationship between stress and strain presently, but first we need tobe precise about the concept of stress itself. We think intuitively of stess asthe ratio of foree to area on some surface in the ‘body. For example, consider the rope iustrated in Figure 112. If we make horizontal cu i the rope we can define the traction vector, a8 the fores in the rope (which isa vector quantity) divided by the cross-section area of the rope: force vector tmuotion vector = T “This idea of free por unit area is intuitive and clementary and well-known, to anyone remotely interested in mechanics. But the situation is mot as sim- ple as this. What if we were to cut the ope with another surface? Would the traction vector be the same? What ifthe surface were parallel to the rope in- Sead of perpendicular? These are questions that vexed the mathematicians fand mechanicians of the 18th century, Its clear the traction vector will de- ‘pend on the particular surface we consider. The problem is how to completely ‘esribe the state of stress ata given point in the body. We ean pass an i Finitude of surfaces through the point. In general there may be a diferent rac- tion on every surface. Hf we want to characterize the stess state, must we deal With all these tractions? ‘One such early attempt, describing force exerted by fluid on fui, and at- tributed to Archimedes (A.D. 250), is documented by Truesdell (1961). “Let it be supposed that a fui is of such a character dat its parts lying evenly and being continuous, that part which is thrust the less is driven along by that which is theust the more: and that each ofits parts is thrust hy the fluid which js above it in a perpendicular direction ifthe Muid be sunk in anything or Perc] Figure 1.12 The intiive idea of waction and sess pressed together by anything else." As Truesdell remarks, “What tis means is not clear” ‘The answers to all these questions were provided by Augustin Cauchy, a [French mathematician, in 1823, Cauchy showed how to find the traction on ‘any surface from knowledge of the tactions on three specific surfaces. Sup- ‘pose we tetum to a generic-shaped body as in Figure 1.13. We can look at & surface inside the body and consider an elemental area of that surface, the trea AA in the figure, Let AF be the force acting on AA. (This force is due {o the action of the material on the righthand side of the surface acting on the material on the left-hand side. We could look at the surface from the ather side and soe the same force caused by the Jeftchand material ating on the right-hand material. That force would atin the opposite direction to AF.) The traction vetoris defined by the limit Ar =i aa" [Next let be « unit vector that i normal tothe surface. Caueiy showed that ‘we ean determine 7 from the product of a matrix (the stress mati with the vector ft ta) Here ois the stess matrix. It contains all the information needed to find the traction on every surface that passes through the point in question. This inmocent-looking equation i very important. Just by knowing the nine com ppanentsof er (and only six willbe independent) we can completely determine the infinitely many tactions that act on the infinitude of surfaces passing through any point, Cauchy’s discovery broke open a Tog-jam of ideas in18 Some ideas from the theory of elastitty AA Figure 1.13 If we ct the body in some arbitrary sure, frees fom one prt of he ‘bal sting onthe oer pat wl Be revealed. These forces are vector quantities. mechanics. Before, researchers had continually struggled withthe description ‘of stress Eg, (1.12) clarified the situation tually. Caveh’s idew and the de rivation of Eq, (1.12) are disused further in Appendixes B and C. (Of course, we still don’t know what the components of «rare, oF how to Find them. Let's write in component form as =| oy oO 13) “To se physically what these components represent, consider a simple exam ple. Suppose the surface AA is perpendicular tothe x-axis, "Then, the unit nor mal vector wil be and, from Cauchy's eg (1.12), the components of the traction veetor on tis surface wil be 1.) [om oy on|*[1] [ow 1 |=| em oy e%| [0l=]o0 Tt] |e oy e} LO} Loe. ‘So the stress component dais the component 7, of the traction thar acts on the surface perpendicular to she x-axis. Similarly dy and cz ate the Y= tnd z-components ofthe traction T. The top row of the stress matrix consists ofthe three components ofthe tration vector that act on the surface perpen- cular to the x-anis. With parallel argument we can show that dhe second Siress 9 Figure 1.14 The components ofthe stress mati nd thitd rows of o are just the components of the wactions that act on sur= faces perpendicular to the y= and z-axes. Thus, the only information we need to find o are the three trations that act on the three surfaces perpendicular tw the three coordinate directions. (See also Appendix C:) “The subscripts on the componeuts of now tell us the physical meaning of each component. Consider oy for example. The fist subscript, tll ws this isa component of the traction acting on the surface perpendicular wo the ‘cx, The second subscript, y, tell ws we have the y-component of that trac> tion, All the nine «components ean be summarized by considering an ele~ ‘mental cube of the material with sides perpendicular to the coordinate axes, ts shown in Figure 1,14, Jus by inspecting this figure we see thatthe diago- fil components dys dj. a fe tho normal stress components, or simply the ‘normal stresses, while the of-diagonal components et. act tangential 10 the surfaces and are the shear stress components or shear stresses for shor ‘Cavchy went on o show, using limit argument, Uhat ois also a symmet- se matrix. Thus, Gy = aj ele. and there are only six independent stress ‘components needed to completely characterize the state of sess al any point inthe body. The components may change from point vo point though, and we an represent this by letting «be a function of both x and ¢ = ote. In the theory of elasticity, workers adopt the fllowing sign convention, ‘Consider surface with unit normal vector pointing in the postive coordinate tection, Any sss component acting on this surface is considered positive its direction is also in the postive coordinate direction. This is itustated in Figure 1.15, We ean sce from this thatthe normal sess cis positive in tension, Unfortunately, the opposite sign convention is usually adopted in20 Some ideas from the theory of elastcy Y0 ce yo Figure 1.18 Sign conventions fr sresses. geotechnical subjects because compressive normal stresses are more common than tensile ones. This difference is sometimes a small source of confusion in clastic analysis of foundations. We will not use either convention exclusively, ‘but we willbe specific about which convention is in use at each stage of the text For the present chapter we will use the elasticity convention and make tensile stress postive Returning again o eq (1.12), we seein general that the tration vector wil actin some arbitrary direction to the surface element, Cauchy's expression fives the components of Tin three coordinate directions. In some situations, itmay be more interesting to look atthe components of T that aet normal and tangential othe surface, To find these components, consider Figure 1.16. The vector component perpendicular to the surface is (7 «fit. The component tangential to the surface i given by the triple vector product X (TX i). The ‘magnitude of the normal component is called the normal siess ating on the ‘surface. The magnitude of the tangential component i called the shear stress ‘acting onthe surface. These are sometimes abbreviated o and 7. Note that and rare both scalar quantities and both are meaningless unless the surface fn which they act sas fully specified. We can express and ras, Tei =VETH A ag) ‘Aamiliar graphical representation forthe stess state at a point in terms of -rand eis the Mohr diagram. lf we take a certain point in the body and through it pass surfaces of every possible orientation and then graph the points (2.7) Tor each of those surfaces, we find the points all willie in @ well-defined re~ sion such as that shown in Figure I.17. This geometric interpretation of sess ‘vas discovered by the German engineer Ouwo Mohri 1882. Ofte, only points ‘om the outermost circle ofthe diagram may be of interest, but every point of Principal stresses a Figue 1.16 The components of traction normal an tangential 1 the surface Figue 1.17 A fypical Mob diagram, the shaded region coresponus to a stress pair (0,7), which sets on some sut= face pasing through the point. Different stress states ill have different Moke Aiagrams with circles of different sizes, hu in every case the circle centers will ie on the a-axis, and there will be three points where r= 0, corre- sponding to three surfaces that support no shear stress. These are the princi- pal surfaces we will iseuss next. 1L8 Principal stresses ‘An interesting question to ask at this point is this: Are there any surfaces for ‘which the waction vector Tis parallel tothe unit normal vector 7? That is, ‘ean we find a surface for whieh Teak ais) where 2 is scala. The answer, of cours, i yes. In general, there will be three such surfaces, called principal surfaces. The stesses that act on these surfaces willbe the prineipal srresses. ‘To find the principal stresses we can use og, (1.15) in eq (1.12) to have Ai= oT of2 Some ideas from the theory of elasticity which can be rearranged to become (a AMa~ 0, (116) where denotes the identity matrix, We recognize eq. (1-16) as an eigenvalue problem. A theorem from linear algebra assures us of solutions fr (1-16), pro- vided ais symmetric. The eigenvalues are found by solving a) eto ~ AN) Expanding the determinant gives cubie equation for 2. The roots ofthis ‘equation are the eigenvalues of er. We will denote them by oy 0 ado, fand rele to them as the principal stresses, The corresponding eigenvectors, fy dy iy, define the three principal surfaces. On each soface, the te tion is given by the principal stress multiplying the eigenvector. Thus, there fs no shear stress on the principal surface. Any surface free of shear stress ‘must be a principal surface. Iti convenient 0 number the principal stresses so that Oe eo “There is another theorem in linear algebra tha ells us te eigenvectors will tbe mutually orthogonal hs, the principal surfaces wil inessect at right an les, This also means we can find a special coordinate frame, say 2, y* 2% ‘whic aligns (athe point in question) with the eigenvector. IF we express the stress matrix in this coordinate frame it will be mo 0 om of. 8) 00 9 ‘tay be convenient to find this special coordinate frame in certain types of ‘roblem, but we need to be aware the principal directions may change a8 we ‘move from one point to another in the body. 19 Stress invariants ‘Suppose we have a body that is subjected to Toads and has 8 nonzero tess tmatrix at sonve point of interest. The components ofthis stress matrix ae, as Sires invariants 23 ‘we know, jast the components of the thee tations that act on surfaces per pendicular o the three coordinate axes, x y, and z, Bocause of this, we see thar the components of are dependeat on oue choice of coordinate system. If we had chosen a different coordinate system, say x*,y*, 2%, rotated rel tive 10 9,2 then the stess components would all be different. But even though the components would be different, there are cern combinations of components that would be the same in either coordinate system. These are called sess invariants (invariant implying something that does not change even when the coordinate frame does) (One invariant i the trace of or the sum of its diagonal components Ist Suess Invariant = J, Fat Oy tO (19) This quantity will always be the same regardless of how we ovient the x,y, and z axes. This is a fact that can be quite useful Basic laws of mechanics should not depend on what choice of coordinates we make atthe beginning of the problem. Since we called J; the first invariants sae wo assume there will be some others, ln fact dere are infinitely many others since products of 7; and pow cr of J ae aso invariants, bur there are only two others that are indepen dent of fy. They ae 2nd Stress Invariant (eran? — tro 420) and Sd tes Invariant = f= deter can Herero denotes he ace of empl by tft hat em hat makes Fpindependent off since the ler em Gro? is jest 2. The te nvaran’ yf ant fe called the principal invariants of be ste mat. They appr in nother cone with rep to the paca Seessex When we expnd the Jelena i (17, oe obtain he cat Scere cite aqation nA, and te cos oh oquation eth en cipal invariants, That, (1.17) 38 equivalent o He Figue 1.19 An example of a usetion bounary valve problem, Figure 120. An example ofa displacement boundary vale problem rust remain zero. Another, probably easier, specification would be to assume the outer boundaries extend so far from the cavity that they are Tocated ef fectively tifinity and then specify that the displacements must vanish asthe radial distance fom th cavity approaches infinity. Now we sre left with only the boundary of the cavity sel, and on this boundary we specify a constant radial displacement Ar. That displacement might be caused by (in some mag ‘cal way) increasing the pressure ina fluid that ills the cavity. One might then ‘object and sty that if the pressure were known, why not pose the problem as ‘traction boundary problem, and that might be done as well; but the point hte isto illustrate how a displacement problem might arise. Cavity expan- sion problems soeh as this play an important roe in certain areas of geoteeh- sical engincering, ‘Mixed boundary value problems are problems in which a part or pars of the boundary have tractons specified, while the remainder of the boundary has displacements specified, A funilar example isthe fixed-end beam illuscy Some ideas from the theory of elasticity foam Figue 121 An example ofa mixed boundary vale problem. trated in Figure 1.21. Traetions are specified on all lateral surfaces of the beam, ‘bt zero displacements are specified on both ends Mied-mixed boundary value problems involve boundary conditions that ae neither “ish nor fowl In these problems, at some pars of the boundary. ‘ve find a combination of both tactions and displacements specified. An ex ample isthe smouth rigid punch indenting an elastic plate shown in Figure 1122, The bottom surfece ofthe plate rests on a rigid base and displacements ‘tre apecified zero, The upper surface of the pate, outside the punch, has 220 teactions specified. Beneath the punch we find the mixed-mixed conditions. “The vertical component of displacement willbe specified, equal tothe puneh indentation. But ifthe face of the punch is smooth so tht Tater slip can o:- cut, then the horizontal components ofthe traction must be specified as zt Tn the other types of problems we have always hd either three components ofthe tration vecor specified or three components ofthe displacement Yec- for specificd. In mixed-mixed problems, either one traction and two dis- placement components are specified, or one displacement and two tractions ‘components are specified (as forthe smooth punch). ‘Note tha in waetion boundary value problems only some of the stress com- jponcats will be specified atthe boundary. To illustrate, consider atypical prob- Jem in foundation engineering «uniform circular load applied on the surface ‘of an elastic half-space, The half-space represents the ground and the circu far load might be an oil storage tank witha flexible foundation. The problem is shown in Figure 1.23, On the surface of de half-space we have tractions specified, either zero or, beneath the storage tank, vertical trations of mag- nitude p,. What do these specified tractions telus about the stress compor nents atthe boundary z = 02 To answer that question we turn to Cauchy's principle, eg (1.12). The unit normal veetor tothe boundary and the speci- Fied traction vecor beneath the tak are given by Formulation of problems _S- puen’ vaio ooundary Figure 1.22 Am example of « mixes Thus from (1.12) we have fo] ex ey [fo 0] =e oy me] ]0 pol Lox oy 2] [1 from which we can conclude Gn ™ Oy = 0 Oi2= Po ‘So one component of normal stress and two shear stresses are specified on the boundary. The remaining components of sess are undetermined. [nother problems, a more complicated condition forthe stress components may arise if the boundary is not perpendicular to a coordinate direetion or if the speci fied tection does not at perpendicular to the boundary; but in every case only thre equations result from eq (112), and these can ultimately contol ‘only three components of stress. ‘A problem that has beon propery formulated is solved by Finding sess, stain, and displacement fells for every point inthe body suet that (the equi ‘equations (123) are satisfied everywhere and (i) the specified! boundary condi- tions are satisfied at all points of the boundary. Finding the solution wil gener- ally involve solving one o move differential equations. We will not worry about solution techniques, indeed we do not yet have enough equations ta solve even the simplest problem Another ix equations that in the components ofthe stress and strait matrices aze sill required. These equations representa generalized form of Horie’ law, and ve will wait © discuss dhem until Chapter 2. The im- pprtant point is this: Solutions give the components of stress, stain, and dis placement everywhere inthe body such that equilibrium and all boundary con- ions ae fully satisfied. Itcan be shown, i solution ean be found, tha it will ‘be unique in terms of stresses and strains, andthe displacemenis wil be unique20 Some ideas from the theory of elasticity yo. cate Figure 1.23 An elastic half-space supporting a nifrm applied fad to within an acbitray rig motion. By that we mean two different solutions to the same problem cannot exist, except for solutions that differ only by gi trans- {ation and rigid rotation. This fet discussed futher in Appendix D. ‘One final point o end this chapter: The linear nature of dhe equations of classical elasticity allows vs to use superposition. Therefore, two solutions for the same body but with different boundary conditions can be added together to give the solution forthe problem where both sets of boundary conditions ae applied simultaneously. This isan extraordinary powerful tol, and we will exploit it continuously in later chapters Exercises ES 1.1. Consider the deformation defined by this displacement vector field my ba? ye ‘Write oat the displacement gradient matrix the strain matrix, and the rotation matrix. Verify that the Srint-Venant compatibility conditions {eqs. (1.11) ate satistied by this deformation. 1.2 The deformation illustated below is called simple sheas, The displace: meat veetor field is given by i= tanps, my =H = 0. Write out the displacement gradient matrix, the strain matrix, and dhe Exercises 31 rotation matrix. Verify that the nonzero components of the st lation matrices agree with our definitions for shear strain and ri tation, Skeich how the deformed configuration can be obtained by a rigid rotation followed by pure shearing Delos oeusaton 1.3 At some point nan elastic body the stress matrix is given by 0 0 0 10 0 oo Ss ‘where the stress components are measured in KPa, {) On one graph, plot the family of traction vectors that act on su faces wih the following unit normal vectors 1] [vae] [ua 12 0 of, | wv2}, | v3e}, |2 of Lo o o | fo ‘That is ind 7 = oF” for each @ and graph the resting vectors from common origin. (i) Demonstrate thatthe endpoints ofthe traction veetrs in part i) ie ‘on an clips. iy Generalize your result in part (i) to prove thatthe locus of endpoints ofall actions 7 that act on ll surfaces at any point in an elastic body ‘must le onan ellipsoid, This is called the Lam stress ellipsoid. 1.4 For te stress matrix given in question 1.3, find the principal stresses and the orientations of the principal plane 1.5 For the stress matrix given in question 1.3, find the three prinipel stress invariants, Use these together with the principal stesses found in ques tion 1.4 to verity tha the characteristic equation (1.2) is satisfied for ll ree principal stresses. 1.6 The octaedral plane is defined asthe surface whose unit normal vector is equally inclined tothe principal suess directions. This corresponds 102 ‘Some ideas from the theory of elasticity References 3 the plane ABC below, By considering equilibrium ofthe wedge OABC, 1.10 Ahollow sphere of diameter 2a and weight Wis pushed downward into erveexpresions (in terms of principal stesses) fr the following 4 sluny of thixotrpie clay used for drilling. The slurry can be treated {@) the normal stress onthe octahedral plane (as) fase ponviscous Nui. Wout using Archimedes’ Prinipl derive an ex- (i) the shear stress onthe octahedral plane (7x) pression forthe init weight of the slury if «force P is required vo im- mere the sphere just Below the slurry surface as noted in the figure pillow enero | oem 1.7 LeU T) and Ps be traction vectors at point x hat act on surfaces with anit normals fy and fz. Prove that aay ont reoter Tyo f= Tame References What propery ofthe ress mtx is necessary for your prot? “The two works mentioned in Seton 1. are: 1.8 The stress ali at some point na elastic ody is known to be Cinch Revs ur gi Moreen Ini des Compe “Dida res sigs ou nom Eger Bul Sec; Phomath V2, 116 500304 18D ae Mt 0, Zin, Ws od Sob, Bin 188 ° ‘Saint Venant, Barré de A.J. C., “Fablissement élementaire des formules et Squations: oo 4 unto 6d tar ore x Ape Rane serMGa‘as Pott Chauney su FApmcsion de Weenie pemie Consider the surface pasing through hs point whose nal vectors pure fan sect De to Reoanc er Cop Sides gar eM Sena nee Bd Pa (868). = Se a Poulos, H. G., and Davis, E. H., Blastie Solutions for Soil and Rock Mechanics, W' {0 Find th component fhe traction vector that at on this sface Se (G) nde magrites of noah sss tha et on his sae Stent PS: Hae Ans of Sion eration, Developments 19 A tue waa testing dovice hasbeen developed fr laboratory testing “it Geoeleal Engineering, Vc 7 evi, Asta (79 Gr cba samples of sis. in ths device, te principal stresses canbe “Tmesde's ears onthe conept of stess canbe fund in intpendoniy conrolled. A tests conde using the devise eval “Trcndell C. Stage inthe development fhe oneog of etn Problems of aera e anesoted ofa try smd, Asfulnon de we ater CeliSio Mecho Cotoutons Hone of he 70h Bh of Aca se ese hci of ly i Ae wt Sr ed os smart lowing ves or he ses invariants Pepe Pmt, 36-5 (61. ‘ ‘Tete ae many excellent references onthe thoy of elasticity. Tse mode | 1-6 10 kPa | em woiks tat develo the teary ae y= 11 108 Pa? ‘kin hand Fo, N An nado tthe Theory of Eas, Longman, . iano 980: [y= 6 X 10° (kPay ‘Burber, J., Flastcity, Kiuwer Academic Publishers, AA Dordrecht, The Netherlands ‘99. (i) Calculate the prinipal siesses at failure mg dan ofS Metric, ne Nea 095 Lite, R. Wi, Bley, Prentice Hall, New Jersey (197) (Gi) What is the greatest shear stress on any surface within the sample Mtvor Uc furducion o the Mechanisof 0 Contnaous Medi Pretce- at failure? ill, New Jersey (1960).