ENGINEERING PROGRESS at the UNIVERSITY OF FLORIDA Vol. XIX, No. 3 March, 1965 Technical Paper No. 300 Generalizing and Measuring the Hvorslev Effective Components of Shear Resistance by Jobs H. Schmertmann Reprinted from Laboratory Shear Testing of Soils. ASTM Spe- celal Technical Publication No. 361, pp. 147-162. © 1964 by the American Society for Testing and Materials. Published monthly by the FLORIDA ENGINEERING AND INDUSTRIAL EXPERIMENT STATION Entered os second-class matter ot the Post Olice ot Gainesville, FloridaGENERALIZING AND MEASURING THE HVORSLEV EFFECTIVE COMPONENTS OF SHEAR RESISTANCE, By Jouw HT. Scimrertann' Synopsis ‘The Hvorslev effective components are reviewed, and it i argued that the usual assumption of equal cohesion at equal void ratio can be in error because void rato is not stficient description of sll structure. To avotd this error, and to permit investigating the components as functions of strain, it Is neces” sary to redefine and generalize them, This is done. An experimental procedure, talied the IDS test, is described which permits separation of the generalized ‘components in a direst manner from a single specimen, The test is versatile as wel as practical. The few available comparisons between the Hvorsley and. seneralized componeats are presented and are favorable. Hvorstev Coupoxers ‘Hvorslev (1.2? presented and later re- viewed his work in determining shear strength* components, which he called effective cohesion, <., and effective friction, g.. He formulated the results of his experiments with two remolded clays, and the resulting ecuation for each haad the simple Coulomb form, thus de- fining cy and gs. Because these com- ponents were based on strength behavior at different effective stress but equal void ratio on the failure plane at failure, and Ge at a given void ratio appeared inde- pendent of whether in the virgin, re- ‘bound, or recompression phase of stress ‘ Amociate profeor of sul! engineering Univeratyof Florida, Gainer, Pia The boldface nuinbere im parerthones refer to the ist of reterences appended to thls paper. ue term abaur "renarea™ refer tothe sheay ste ablaed by the ail at any en ‘ondiion, wile shane “strength refers ony 20 {he shear rsintanoe tt a deed failure cond history, they have also become com ‘monly known as the “fundamental” 0 “true” components of soil strength, ‘The concept of these components ha endured with few corrections or ad since the work of Hvorslev. Bjerrun G, p. 80) Bjerrum and Wu @), anc Jakobson (5) have indicated that fo Some undisturbed clays, or clays re molded at water contents below th liquid limit, there exists an “or cobesion” to be added to the cohesion which is a function of only void ratio Gibson (6) showed that the component determined from drained tests must b corrected for volume-change effects, i any. Because the determination of c, ane ¢¢ isa laborious procedure often requir ing unobtainable duplicate specimens, i thas been done in research for only very few undisturbed soils, and perhaps neve for purely practical work, Yet the effec tive components have had great practice2 Lanoratony Suzan Tesrinc oF Sorts significance. ‘Their mere existence has ‘been the rallying-point for the continued ‘use of the componest philosophy of soil behavior. Their use permitted the mathe- matical studies which are the theoretical (@ womstaric ONS DATION pokes Toa eS ‘Bay Bung (8) oveORENT SEPARATION fective Components and bases for some common empirical meth- ods of applying soil strength analysis in practice. Some examples are Skempton’s () analysis of undvained strength and they = Oanalysis, Hansen and Gibson's (@ extension to the case of anisotropic consolidation and consideration of the comparisons between different types of tests for undrained strength, and Jakob- son’s() further extension toin-lude origin cohesion and octahedral normal stress. 99) ( weasimen Powe PressuRES Hoo. Jc Hydestately Comlate, Undine Tia Tests toa) Deterine E- ste Their Physi Meaning Bjerrum (3, p. 79) and Skempton and Bishop (9) used the effective components to show the relationship between drained and undrained strength envelopes and between water content and undrained strength, Scuaeatuanw ox Couponents oF SHEAR Resistance 3 ‘Vor Ratio ano Sreucrume Intended Physical Meoning of Bective ‘Sirengh Components ‘The well-known experimental method ‘suggested by Terzaghi (0) to determine ‘end gs, which also is generally con- Sidered” to ustrate excellently the physical meaning of the effective com- ponents, is reviewed in Fig. 1(a) and (6). "This method and its relationship to four other methods is also discussed by Bjer- rum @). Following Horsley 0, the a8 sumption is made that because tests and Bhave the same void ratio at failure ‘they have the same c,, and all strength differences are due to (84) tan . Most investigators would agree that efective cohesion, beause it is believed to be “fundamental” is ome function of the noneffetive siresses or bonds between particles and of the details of the ex- tremely complex geometry of particle spacing and contacts. We can say that Gea function of the catchall term oil structure” or just. “structure,” ‘Methods-of-testing will be considered a constant here, and. possible dlatancy corrections, at suggested by Gibson, are also not considered for simplicity ‘When assuming equal’ cobesion at equal void ratio it is therefore also as- sumed that equal void ratio means equal structure. Of course, Horsley realized this point. For example, he wrote: “Fur- thermore, it must also be assumed that there is no significant diference in the geometric structure of the test specimens inva given seis at the time of failure” @,p. 206. ‘The meaning intended for these com- ponents is then: ‘These are the twocomponentsof a soils sear strength in which one, effetive friction, is com- pletely and linearly response for the difference in shear strength between two specimens with the same structure but ‘with different effective stress ‘on the failure plane. The other, effective cohesion, is the strengh remaining after ext ‘lating to zero effective stress. But this intended meaning may not be clearly achieved by an experimental determination using the Terzaghi method ora variation of it. Some ofthe following discussion refers to Fig. 1, in which is presented imaginary data from a simple Dest series to determine the effective components by the Terzaghi method. For illustration, the author bas imagined test data which involve and. perhaps exaggerate several dificulties in evaluat- ing the signicance or accuracy of the components thus separated. Treo Factors Influencing Structure: Void rato is of course an important ross measure of soll structure. How- ever it tells us nothing about the impor- tant details of structure. For example, an undisturbed and a remolded sensitive clay can have the same void ratio, but structures certainly very diferent. Even assuming void ratio is necesary to de- seribe structure the question is whether it is sufficient. ‘The author argues that in general itis not, but in the special circumstances originally encountered by “Hyorsle it could be suficient. Onerconsoldation—Fig. 1(e) and (2) ile lustrate stresestrain and pore pressure- strain curves that might be expected from two consolidated-undrained triaxial tests such as A and B. The initial modu- Jus is a measure ofthe “sifness” ofthe soil structure. Tt could well be stifer and therefore obviously different for test Bin spite of lower efiectve stress. Pore pressure is a measure of the tendency of 4 saturated structure to reduce volume with strain, The stricture of wants ta txpand while that ofA wants to contract. Daring consolidation it is necessary to4 Lanorstony Suzan Teste oF Sons Smagine the occurrence of miero-shear failure, or slippages, at many contact 3s thrcughout the mass. These slp- pages permit. the platey particles to assume a more stacked (dispersed) at- Zangement which isa necessary corequ site to the void ratio reduction. Horsley already noted this efect. Ie impossible that this many slippages are all reversible with rebound swell when the consolida- tion stresses are reduced in test B. The consolidation rebound modified structure at B musi be different, perhaps very uch $0, from the simple consolidation vodifed structure at ‘Consider now the practical case of an ‘undisturbed soil with always posible signiGcant brittle cementing, oF other bonding, between its parties. Tt then would have a true tensile strength in terms of efective streses. What remains of this strength at failure mast be part of “origin obesion” and be included in fe. Te also seems likely that the strains of overconsoidation will irreversibiy de Stroy some of these bonds that would otherwise not be destroyed, and then the total cat B wil be less than at .t (Fig 1). Thos the assumption of equal ¢ at equal void ratio would be in error and Tead to vahes ofc that ate too low and of es that are too high. Mencl qu) has rade the same observation. Iti known iat determinations of ge for some un- disturbed days yield greater values than after remolding which may, in part, bbe due to the above effect i283). An’ other possibility is discussed below. Also implicit in the component de- termination illustrated by Tig. 10) is the assumption that both specimens fail fon the plane of common tangency, which has the same inclination (a = 48 + e/2) in both specimens. But, as ius. trated in Fig. 1(0, specimen may fail along a plane with a different inclina- tion than B. If (a #45 + gy/2 = a, then the use of the common tangent ives tooshigh values of p,. On the other hhand, if (4 = 45 + 4/2 x a), then one obtains too-low values of gs. It seems likely that the less-disturbed specimen, A, will retain more of any natural brittle and/or anisotropic struc- ture and thus [ax — (65 + es/2) > \ ay — (35 + ¢x/2)|, and ge will be too high ‘Shear Sirain'—One has oly to observe the development ofa failure plane in a compression test to realize that shear Strain can produce a structural change in the soil, The slickensides observed in some natural clays are evidence of the permanence of such structural changes. Structural distortion accompanying shear strain causes many local and ireversble slippages between particles as the structure struggles to mobilize its in- ternal resistance to the forces producing the distortion. Recent investigations in- dicate that shear strain causes the particles to assume a more parallel, or Gispersed, structure with increasing strain (141816. Refereing to Fig. 1(0, suppose one tried to determine the effective com- ponents at diferent values of strain. At strain below «* this would lead to the ‘unplausible result of negative ys. At sreater strain it might lead to negative e- Obviously, this is not a correct pro- cedure, Yet to consider together in Fig. 4() the two failures, each at diferent strain, is seemingly conrect—at least in Hyorslev’s investigation. Why? It can only be s0 if the inital diferences in structure due to the overconsolidation were erased by the structural changes of strain, so that at failure atthe same void ratio the structures were essentially identical. Under what conditions might this occur? Horse's Experiments — Hivorsev “Ta writer doce not diferente betnoen recente and nonrcoverle deformations “incr slat and ne sing.‘ em arin” ed ere sane teal dormation divided by a appropriate andeforned dines ScantenTMANN ON CouronENTs oF SHEAR RESISTANCE 5 started with soils remolded at the liquid limit, so that brittle bonding or a highly floceulated initial structure were not possible, His use of direct and torsion shear machines tended to disperse the particles during consolidation in the Te AXIAL STRAIN AT FAILURE, PER CENT, Nomaly ores st ‘Sek Sas, ‘Yanna clay o 4 ite Bet > 8 (e-03)- [2&\Lores)] sin (ton'd,) coe NOTE: (oa) AMO FROW SANE, uy eITWERY cuacLe, ag. 2 Separation of Generalized Components, Iustrated for Plane of Envelope Tangeney ia a “Triana Tet. same direction as the subsequent shear Strains would tend to do and thus make a high degree of dispersion at failure all the more likely. Translating Hvorsy's failure stain from direct and. torsion machines into the now more familar Compressive strain and ignoring the pos Silty of greater strain in failre planes, ‘the author obtained the following (17, p26): Tt seems reasonable to conclude that failure strain was large in these tests, ‘making it still more likely that presshear differences in structure were erased by the shear strain to failure. Can we rely on this structural sequence to occur with other soils, such as undis- turbed soils with a highly developed ini- tial locculated and/or anisotropic struc ture, or brittle soils which do not fail at6 Lavoratory SHEAR TESTING oF Sous high strain, or in other tests such as the triaxial where consolidation and shear dispersion directions are less likely to coincide? The author thinks not. In a recent paper Olson (1) indicates he may hhave similar thoughts. Even if we could rely on this for the failure condition, we definitely should not for the important strains less than failure, Any considera- tion of other strains must be preceded by more generalized definitions for the effective components. ey Fi, 3—Curve Hopping with One Specimen. GeneRazizarion Definitions: Because structure is a function of stress history and strain, the components, or atleast c,, must also be. Furthermore, there is no reason to believe that ce ot ¢. (or both) are maximum values or that they both reach a maximum at the same strain, or that they are not indepenent during strain. Strain must be considered an independent variable. Then the con- cept of fale is no longer of special significance, nor is the need to rst component seperti to some fare failure plane. Complete generalization requires that any plane can be consid- ered. Using the intended meaning of the effective strength components as a basis, more generalized definitions for the effective components can now be written: D, = Dependent Component, compo- nent of shear resistance mobilized on any plane and at any strain e, which is de- pendent on effective stress on that plane according to the equation: =F OL-be-« when, at ¢, the shear resstance on that. Moncton even enna Boat | secs e plane changes by Ar due to Ag. Because ‘there must be no change i soil structure, ‘4g must approach zero.+ 1, = Independent Combonent, remain- ing component defined by: Lenn De ® Tn the incremental form necessary in experiments, the definitions are illus trated by Fig. 2 for the plane of envelope tangency (note similarity to Fig. 100) ‘The equations for this case are also given. ‘The author has taken the liberty of The cara, but perhaps minor, additional lndusace on ar of nora sre changes on the lanes orthogonl othe plane cnatdered fe 20m [gored for simply. Toerefr, this gern ton must be considered incomplete. ScHMERTHANN ON ComPONENTS OF SHEAR RESISTANCE 7 ‘changing the symbols for these general ized effective components. Tn previous ‘publications he merely changed 10 c, and tov, to emphasize they were func. tions of stain, «. However, the physical meaning ofthese components may prove to difer considerably from the usual con- cepts of sol cohesion and fiction. There- fore, continuing the philosophy of gene. aliing, they are merely. called. the components that are “dependent” and “independent” of effective stress, oF D, and I, respectively. Experimental Approach The ebjective when measuring D, and 11s to obtain two effective stress condi tions with the soi tthe identically same structure and to do this at any strain, “Thisseems impossible, anditis. Structure changes with effective stress and cannot ‘be exaclly the same, but pechaps it can be suliently so for a good approxi ration af the ideal. The measuring prob- lem becomes one of achieving a relative maximum of effective stress change for a minimum of the inevitable structural changes, Tmagine that we can obtain two shear test sperimens that are genuinely ident- cal at the start ofa shear test (aero strain in Fig. 3). We shear specimen 1 by some test method and obtain the curve shown. ‘Then just before shearing specimen 2 we impose a slight change in the starting effective stresses. We then shear withthe same lst method and obtain the other stress-strain curve at all points of which Affective stress is slightly diferent from 1. At any strain chosen we have two mobilized. shear resistance values. and two effective conditions and can there fore separate D, and J , as shown in Fig 2 for the plane of envelope tangency. How great must the slight change be? ‘The answer is the minimum necessary to separate the two circles in Fig. 2,50 that a tangent can be determined with suficiert accuracy. Curve-Hopping and Use of One Specimen: ‘The great practical dificulty in a component measurement method such as that just suggested (Fig. 3) is that it is often impossible to obtain two or more specimeas that are sulficently identical. Consider now a different technique with specimen 3. The small change in the ini- tial effective stress condition of specimen 2 could also have been imposed at any point during strain (assuming it could be {imposed without change of strain) such srt éy. If the subsequent shear resist- ance of specimen 3 at é is the same as ‘that for specimen 2 (point y), then only specimen 3 is needed for component separation at ey. The only possible difference in struc ture between points + and y is due to the imposed decrease ineffective stress which in turn results in a decrease in the shear resistance at é,y. The author's experi- ence is that an adequate “hop” bk ‘ween curves results in a void ratio crease Ae of less than 1 part per 100, It has been a pleasant surprise to discover that it is possible to make several such hops during a strain-controied test, and to make them in both the xy and y- di- rections, without greatly changing the subsequent position of the stress-strain curves (09). This curve-hopping tech- nique is then at least approximately re- versihle with continuing strain. Either very litte structural change occurs with cach hop or the shear resistance effects of any change are almost recoverable with continued strain. Tt is the author's opinion that the 1. specimen, curvehopping separation of the generalized effective components results in the minimum structural dif. ference between two conditions of ef- fective stress that is presently possible with common soil mechanics techniques. Tt is at least an approximation of the ‘minimum necessary.8 LABORATORY SHEAR TESTING OP SomLs The IDS Test: ‘The objective of this testis to sepa- rate the J and D effective components at any strain—hence the abbreviation IDS (CFS was used in previous publica tions because of previous cohesion and friction terminology). Curve-hopping can be used to the extent required. Two examples are given in Fig 4 Ifthe com- pponents are only desired for the after- failure shear resistance condition, then ‘hopping sequence such as aed-ef in 4(@), with interpolation between be, is sufficient for a separation at ey. Ifa (2) 105 Ts ATER eon aR tests, or two levels of constant volume in ‘undrained tests with pore pressure meas: ured, Wu et al used the test, in mult specimen form, to study relaxation and creep. Bea 20), and Schmertmann and Hall (21) also studied time phenomena with Lspecimen tests using a short ex- trapolation to obtain pre-test conditions. Equipment Repuirements and Tech. ‘nigue—To separate elective stress com- ponents on a particular plane at a par- ticular strain it is only necessary to know mobilized shear resistance for two known effective stress conditions on that glane at that strain, The design of the equip- (co) ort 1 TST Tic, 4Exanpls of After Failure and Complete OnesSpectnen 1DS Tests. continuous curve of the variation of D and J with strain is cesired, then many Jhops may be required, as shown by the solid line in Fig. 4(8) After interpolation, again indicated by dached lines the com ponent separation canbe made at strai ssuch as «to « and a curve fitted to Iustrate their continuous variation with strain. Variations in IDS Test—The concept ofthis test has proven to be applicable in ‘many variations. Is the imposed change in effective stress that triggers the curve- hhop. This change can be performed in any convenient manner as for example, ‘two levels of a1 wherein pore pressure is suitably controlled, or two levels of pore pressure or confining stress in drained ment does not matter provided it can. give this information. Presently, the tri axial is most suitable because of the rela: tive ease of effective stress measurement. Any plane can be chosen, The direct shear apparatus is also suitable if the analysis i restricted to drained tests and the horizontal plane, as done by Landva. (2). Theoretically even the undreined field vane test would be suitable to deter- mine the average d component for the failure cylinder if somehow one could im- pose a controlled effective stress incre- ment. The measurement precision re- quired with any equipment can be estimated by review of the example given by the author in Table 5 of Ref. a9) Scuueataann on Components oF SHEAR Resistance 5 ‘The most difficult part of the Lspeci- ‘men technique isto ensure that the speci- men has come to stress equilibrium after the imposition of an effective stress hop. ‘The difficulty increases with decreasing permeability and increase in swelling tendency. ‘The experimental techniques for the IDS test are intended to overcome this problem. =a Remolded toot ‘TABLE |-AVAILABLE COMPARISONS OF tan gs AND du fixed to a specified strain nor are they necessarily masimum values. As shown in Fig. 1(0, itis likely that the NC and OC specimens fail at a different strain. ‘Then to what strain should c, and tan ys be assigned for comparison with J and a? At conventional failure the sum of both components is a maximum, but it is very likely that both are not at their sade ‘ite 0.1 rained teat, constant lume O'Neil 5) Geico ans Sc et otto ‘ales lundroinedtriasal —Sehertmann O82 an os alae constant volume TDS arr fal. conttant 3, 1DS Undinarted... 0.19 (a) undmined triaxial Wu etal 8) eas G)ultapetnes, is Remolded 0.2 (v= awh 10 undesined tla er eont axel oan mult speien onsard olume ibs Undistrbed Leda hy 0.28 (oo) and cantare TDS_—_Beeemann es a) ost Tadiectly infeed before else Crawford (6) Comparison BETWEEN HvoRste Axo GENERALIZED COMPONENTS In this paper the author attempts to show that the effective shear resistance components, D, and I. can logically be considered a generalization of the effec- tive strength components originally proposed by Hvorslev. If true, then at failure one should expect to be able to demonstraje that I = cand d= tan gy But the comparison is more difficult be- cause the ¢, and y, components are not maximum values so one cannot compare with both Zs and dans» These compari- son. dificultis are illustrated by the author 7. A further complication is that Hvorslev used drained tess, and his work does not speci'y whether to use (6: — e:) or (@/a,) failure in undained tests. Tt has become common to use (61 ~ ¢:)aae Dut the structural condition of the soil is probably more unique at (6i/é:)n0< (3.1824) Table I lists the presently available10 ‘comparisons of tan gs and daeg. This Comparison was chosen es most ely to show agreement IDS tests indicat i is very likely that Z is past its maximam nd decreasing atthe strain of maximum Shear restance, which i also the strain ‘of (I + D)mx. Therefore if & is not hanging too rapily with strain, this is also approximately the stain of dans Maximum shea: resistance is also the point for the ten yy determination and therefore should usualy be in approx: rate agreement ith Guae. ‘The data bresented generally support this argo. rent. Unfortunately thelist is bref and nore comparisons are certainly needed ‘Wa Gn offers some additional, though indirect, evden. Using data from four remolded clays found in the literature, he plotted the variation of dae mith pas: ticity index and found agreement with a published tang versus PY curve aver: ged ffom many other remolded clays. Lavoratory Suzan Tesrinc oF Sous Perhaps significantly, a similar compar son using undisturbee clays showed poor agreement, with dayx much less than tan g, As discussed Lere, the determin tion of ye for undisturbed clays can it 8 errors with resulting too- she values. Conc.vsi0N ‘The author recommends that the D, and Tz shear resistance components be considered as possible generalizations of the Hvorslev effective strength compo- nents. If true, their determination is ‘greatly simplified by the various forms Of the IDS test. However, regardless of the merit of this argument, itis neces- sary for investigators to keep in mind the possibilty that equal void ratio may not mean equal soil structure (cohesion) when determining the effective strength components for the conditions in their tests, [REFERENCES () MJ. Honey, “Ueber die Festgke! algenschatten gestérter bindiger Boden’ ‘Thess, 199 pages Publahed by Danmavks Naturvidensabelige Samfund, "Ingei- Widenabele Suter, Series A, Nr AS Copeniagen, 1997. @) Me, Hyorsey, “Physical Components of the Shear Soength of Saturated Clays” American Sicety of Civil Engineers eseatch Conference on Shear Strength of Cohesive Soi 1960, 9.168. @) L Bjerrum, "Theoretical and Experimental vestigations on the ‘Shear Strength of Soll” Norman Geotechnical Inte, Publiation No.5 1984 ( L. Bjerum and ‘TW, “Fundamental Shear Strength Properties of Lila. Edet Chay,” Geoectniae, Vl. 10, No.3, 1960. (8B. Jakobeon “Influence of Sample Type and Testing Method on Shear Stength of Clay Samples" Preceding, No. 8, Royal ‘Swediah Geotchnial Ist, tocol, (© RCE, Geo, “Experimental Deternina. tion ofthe True Cohesion and True Fec- tion in Clayy” Procedingr, Sed Ta ‘ational Conference on Soil Mechanics and Foundation Engineering, Vol 1, 1953, 126 (0 A.W. Skempton, "The Bective Steses in Saturated Cys Strained at Constant Volume,” Procedigs, Intemational Con: rea of Applied Mechanics, Vol. IE, 198, pas © J.B. Hansen and RE, Gibson, “Une rane Shear Strengths of Anisotrpically Consolidated Clay,” Geteciniqu, Vo. 1, No.3, 1948, (©) A.W. Skempton ant A.W. Bishop, “Sil” in Buslding Matias, by M." Reiner, chapter X 1984, p-470, U0) K. Terzaghi, "Die Colombsche Glichung fir den Scherwidentand bindiger Boden,” Bavtechit, Vol 16, 1938, pp. M3-M6. ‘Translated in Fron Theor) to Practice jn ‘So Medians, by 1. Bjerrum (a1) Vi Mench, Discusion in’ American Society of Chil Engineers Journal Sol Mechanics (nd Foundations Diion, Feb, 1963, p. 289 (42) E"Bjersum and N. =. Simons, "Comparison. fof Shear Stength Characterstin of Nor tally Conmifdsted Clays” American Socety of Civil Engineers Research Con- ScHMERTMANN oN Components oF SuEAR RESISTANCE n ference on Shear Strength of Cobesive Sails, 1960, p 720. 13) K, ¥.La, "Shear Stength Properties of a Sursple of Voleanic Material of the Valley of Mexico," Gtecnigue, Vol. XIT, No. 4, Dee. 1062». 316 146) HB. Seed tnd C.K, Cha, “Structure and Strength Characteritics of Compacted Clays” Anercan Society of Civil Engi. tera; Jowral Sod! Mechanics and Poundo- lions Division, Vol. 8S, No. SMS, 1989, po, sr-128 (05) T.-H. Wa, 4 G.Dougls and R.D. Gough- nour, “Ficion and Cobesion of Saturated Cty” American Society of Civil Engi- nett, Jonrsal Sol Mechonice ond Founda ions Dison, June, 1982 (16) K Pusch, "Clay "Partces—Thelr Size, Shape and Arrangement in Relation Some Important Physieal Properties of Clays," Handlingae N. 40, Tronsecions, ‘The Nat. Swedish Council for Bids. Res, Stock, 1962. (07) J-H. Sehmertnans, Discusion, in Ameri. ‘an Society of Civil Enginges, Journal ‘Sol! Meckovier and. Foundstione Division, Feb, 1963p. 28, (08) RE. Olse, "Shear Suength Properties of Sodium Mle,” American Society of Civil Engineers, Journal Soll Mechanics ond Foundation: Dison, Feb 196, p. 206. (49) J. Schnaremann, *"Comparions of One and Two Specimen CFS Test,” American Society of Civil Engineers, Journal Sol Mchonics ond Foundations Division, Dec 1962, (20) Re G. Bes, “An Experimental Study 0 Cohesion and Friction During Creep in Saturated ‘Cly,”" Masters Thesis Uni versity of Forde, June 1960. Partly Dublished in discuss, American Society “Teal Sol Mechones (21) FH. Schmertmann, and J.'R. Hall J "Cohesion. After NeaHydrosiatic Con soliaton,” American. Society of Civil Engineers, Jounal Soi) Meclanice ond Foundations Dison, Aug. 196. (22) Ar"Tandvs, "En eapeimentell_wnder- selse av’ skjaerfatheten 1 normal- "Norwegian Geotech- ‘lel Tnaltute, Intemal Report F- 175, 1902. 23) L-Bjerum, “The Efctive Sheae Strength Parameters of Sensitve Clays” Proced- ingr, Filth Tnternatonal Conference on Soll’ Mechanics ané Foundation Eng eerng, Vol p. 27; and Norwegian Geo- technical Institute Publ No. 45,1961. (24) NE. Simons, “the Efect of Overeon, salidation on ‘the, Shear Strength Char ‘eters of an Unisturbed Oslo Clay,” ‘American “Society of Cull Engnees, Proceedings, Research Conference on Shear Strength of Cohesive Sls, 1960, p. 761. 25) HM, O'Nel, "Diet Shear’ Test for American Society of Civil Engineers, Journal Sot ‘Machaics ond Foundations Dison, paper 3282, 1962, (26) C. W. Crawford, “The Infuence of Stain on Shearing Restase of Senive Clay,” ‘Proceedings, Ar. Soe. Testing Mats, Vol. 61,196. nT! H. Wa, Discusion, Am. Soe. Civil Engrs Journl Sol Mechanics and Foun dations Division, Jt, 1963, p. 250DISCUSSION Roserr L. Konpwes'—The review of the various aspects of the strength and resistance of cohesive soils under applied stresses given by Schmertmann is a very interesting study in contrasts. Although the writer agrees with many of the points reviewed by Schmertmann, there are many points of disagreement. Although he uses several parameters such as D, and 7, or the previously used ppararneters, ce $e yn considering shear- ing resistance, Schmertmann is actually dealing with the stress-strain aspect of the strest-strain-time response of cohesive soils Tn studying material behavior itis important, if possible, to express the re- sponse in terms of the fundamental vari ables under consideration, which in this case are stress, strain, and time, or in terms of parameters that are directly defined or calculated from them rather than artificial or indirect parameters. Schmertmann is attempting to study stress.strain-time behavior using the techniques or methods of representation used for studies of ultimate strength or failure stresses. Such techniques, as used by Schmertmann, are not compatible ‘with material behavior at strains below failure The most widely used formulations of ultimate or failure strength of soils have their basis in the modified Mobr.Cou- lomb criterion expressed as a failure en= velope in a two-dimensional stress space. ‘This failure criterion concept does not Amocinte profesor of civil engineering. ‘Technologie! Instat, NorUuwestern Univer sity, Branston, IL 2 refer to or correlate with the strain space; that is, it doss not provide a rela- Uionship which species the coaxiality or orientation of the tress tensor relative tothe strain tensor; nor does it provide a basis for an equational relation between them. The so called “cohesion” and “angle of internal fiction” are fttious soil properties which in reality are sin ply expedient parameters that have been Used £0 approximate the representation of the failure envelope. Te is indeed un fortunate thatthe names “cohesion” and internal friction” were ever given to these parameters which are an “inter: cept” and a “slope,” respectively, of an empirical curve fit. This has been recog nized by some people for a long time and has been ably expressed by Lambe? Tn terms of failure stresses or shear strength in the plane of failure at the time of failure fr wse in a form of limit or ultimate analysis, these two empirial parameters are useul engineering ine dices. However, there is no theoretical basis for extending the failure envelope concept. to variols stress strain states Ielow failure because strains at such stres states are not unique but are func tions of the history cr loading path of the soil up to the stres state as well as the nonlinear viscoelastic nature of the soil. Schmertmann seems to indicate a vague auwareness of this at various points in his paper but then proceeds to ignore it by YT, W, Lambe, “The Enginering Behavior of Compacted Cay," Precedings, Ams Soe {ini Boers, Paper No, 1055, Vol Bi, No. S13, 1088, p 1085 20, Discussion on Couronswas oF Siar RESISTANCE B ters in introducing the quantities, D, and 1, The fact that there isa soil resistance to the applied loads or deformations is beyond question, but to divide sistance into various parts with associ- fated mechanisms of behavior is purely speculative and subject to considerable question. In addition, it would be highly unlikely for a number of investigators to agree on basic definitions and importance of any parts of such a division, The use of io, $—Void Ratio-Preture Rehtion. artificial parameters instead ef funda- mental variables can be very misleading in attempting to explain the stres-strain- time response of soils and, hence, is not to be recommended. This does not mean that all such parameters have 20 use as ‘engineering indices for particular situa- tions. In the future it might well be ed visable for Schmertmann to consider stress-strain-time response of soils in terms of the fundamental quantities under consideration, namely, to express his results in terms of stress strain, and time directly, in addition to the path and history variables ‘The writer agrees with Schmertmann’s contention that, in general, void ratio alone is not sufficient to describe the “state” of a soil. The term “state” may contain a variety of effects including such items as path dependence, history, and structure, The CFS or IDS vest de- ‘veloped by Schmertmann is certainly an ingenious test and may prove to be a very valuable experimental tool for the soil mechanician. This may be particu- larly true in light of the difficulty of try- ing to obtain a number of genuinely iden- tical soil specimens. However, the manner in which the test is conducted using one specimen with the curve-hop- ping technique is such that the instantan- ‘cous effective mean hydrostatic stress is continually changing; hence, one is con- tinually operating on a portion of the recompression branch of the void ratio. pressure relation. This is shown qualit tively in Fig, 5, in which the test speci men is hydrostatically consolidated to point A. During the CFS curve-hopping test, the effective hydrostatic pressure Aecreases, and one is operating approxi- mately along the recomprestion path AB. This is a consideration of significant im- portance. Both in the present paper and in other papers dealing with the CFS test, Schmertmann has indicated that it was possible to obtain approximately the same stress-strain curves for the single specimen with curve hopping as he ob- tained using a number of individual test specimens. All of these specimens are initially hydrostatically consolidated to the same void ratio as illustrated by point 4 of Fig. 5. The implication of the validity of the curve-hopping technique is an implication of the validity of the principle of superposition of effects, that is, linearity of effects. It is important to realize in making the substitution of one specimen for a number of specimens that all are related or associated with the vicinity ofa single region of the e-p space4 such as the region AB of Fig. 5. For such a situation, the overconsolidated nature of the soll for the recompression branch would indicat that in general the testis ‘being completely conducted on a portion of an unloading curve with strain-bard- ened effects built into the soil structure and, hence, one should expect a quasi linear behavior. I s well known that the general principle of superposition does fot hold for soil? Indications are that thecurve-hopping technique is valid only for small variations inthe loading history of & soil. Thus, one must ascertain the effects of “slight” changes in the stress and deformation paths both before and during the tet. Although Schmertmann states that void ratio alone is not sufi cient to describe structural effects in sil response, he uss it as an indication of structural change and notes that the void ratio variation i less than one percent ‘Thus, the extent of the general appica- bility of the CFS test, and the curve- hopping technique remains to be shown Of equal importance is the form in which the test results are presented and the possible interpretations that are applied to these results Jou H. Scmsemncann (outhor's dlosure)—Mr. Kondner’s discussion does not concern itself with the main thesis of this paper, namely, that the J and D components can be considered generaliza- tions af the Hvorsey effective com- ponents. Presumably he bas no opinion bout, of agrees with, this thesis. He does question in some detail the signs cance ofthe I and D components and the test technique to determine them. By implication he is also questioning the Horsley effective components. For the purpose of eficent discussion take the liberty of ofering this concise restatement of Rondnee’s criticisms: (1) VIE Terughl, Theoraical Soi Mechanien, Joh Winy and Sony, New York, 1983, pp 304° saa Lanozatory SHEAR TESTING oF Sous ‘My omission of consideration of time, whichis an important variable. (2) My perpetuation of cohesion and. friction philosophy, which he considers to be non- fundamental. 3) In his opinion the use of ‘Mohr czces for separation of compo- nents at strains less than failure i the- oreticlly unjustified. (4) He sees this paper as implying the validity of super- position in sols, and he disagrees. shall discuss these in this order. 1. One cannot but agree that time is an important independent variable, However, to simplify the main thesis of the paper I noted that methods of testing were to be considered constant. This was intended to be interpreted in the broad sense of including sample storage time, strain rate or rate of steess application, and any other time-dependent effect. T believe that a useful experimental ap proach toward a better understanding of the mechanism of shear resistance in soils isto consider stress-strain, strain- time and stress-time individually. with the time, stress, or strain (respectively) held constant. "This paper deals with stres-strain ‘The generalized components can be, and have been, evaluated as functions of various time effets. Already published examples are Bea's (20) study of creep (suain-time), the studies of exeep and relaxation (siress-time) by Wu e al. U5), and Schmertmann and Hall's (a) study of creep and rate-ofstzain (in closure). 2. Kondner seems to have missed the essential difference between the “inter- cept” and “slope” parameters obtained from the empirical straight line ft to ap proximate the Mobr failure envelop over therange of sre interest, and the gener- alized effective stress parameters , and D,. With the ZDS test the sil structure remains comparatively much more con “The boldface numbers in parenthene ree tothe lta rafroncs appended tote paper Discussion on CoMpowents oF SHEAR RESISTANCE, 6 stant per unit change in effective stress. Parameters J and D are measures of @ soils shear resistance sensitivity at a given structure to 2 probing, seemingly nondestructive change of effective stress. ‘This must be contrasted to the ordinary “ohesion”” and “angle of internal ric tion” terms which represent the stength parameters from tests usually encom- passing a large range ineffective stress, with the different tests involving differ ent failure modes, dilatancy behavior, failure strains, and void ratio at falure— in short, greatly diferent structure. The T.and D,aymbols are used to help avoid rising this point. Tt is clear from the history of soil mechanics that a great surge in our un- derstanding of the engineering behavior of sols occurred with Terzaghis labora- tory and field demonstrations ef the im- portance of effective stress. tis also clear that because soil deformation results pri- marily from sliding between paticles the shear resistance of sols fundamental to all shear and consolidation problems. Consequently, itis my opinion that the shear resistance sensitivity of a. given soll structure to © change in effective stress, and the change ofthis sensitivity with strain, are of fundamental import: ance. T developed the IDS test as a actical means of determining this sen- Siivty 3. The Mohr stress circle is simply a sraphical representation of the dstibu- tion of shear and normal stres on any plane ina two-dimensional stress feld, oF there two ofthe principle streses have the same magnitude. Static equilibrium is assured-which is approximately true av all times with conventional rates of strain, The les thestrain the mcre closely ‘eis true because of reduced creep effects ‘ACany strain, such as a strain less than failure, the two Mohr circles obtained fom an IDS test permit a simple deter. rmination of how the normal tress and shear resistance on any plane changed in response to an imposed change inthe ef fective stress conditions. The T and D components express this change, quanti- tatively, for that (any) plane, ‘The theoretical justification for the above use of the Mor circles sas sound, and perhaps more so beeause of reduced Geap a5 hat fo hel ie a dened 2” condition. Confusion due to concern with strew history and time ef- fects is unnecessary, because the gener- alized components are determined for the condition of the sol as found atthe in stant of the imposed effective stress change. As mentioned, stress history and time effects can be and have been studied independently "The common procedure of indicat- ing the progress of a triaxial test by means of stress paths assumes the valid- ity of the use of the Mobr stress circle representation throughout the test. To iy knowledge, this assumption has not been questioned previously in. written discussion. This is « good assumption, with a degree of validity much greater than most assumptions made in sols en- sineering “4 The experimental success of the curve-hopping procedure isa fact which the reader can verify (7). However, from this the author makes no implication of the validity of superposition or linear elasticity in soils. This implication is made by Kondner and be then argues againat it, ‘Without here supporting the posible applicability of superposition, Kendaer's * a. Cuangrande, and 8. D. Wilton, “Prestros, Tadiead in Conoldated-Quick Trina Tost, Drooesings ofthe Third International Con ference gu Soll Machaniesand. Foundation TGnginerng, Switseriand, 1089, Vol,» 100, ‘Methods of Eaineting Settlement,” preprint subtad tothe Atria Seciety of Civil Engineers Conference om The [Besgn of Foundation for the Contel on Sette rents Sesion, Evanston I, Juve, 100816 Lanonatony Suxar Tesvine oF Sons argument against itis weakened by choice of reference. On the pages cited ‘Terzabgi said essentially (in 1942) that superpoition isnot valid for the case of layered systems in which each layer has a diferent modulus. Curtis and ichar’? Showed in 1955 that it was valid for this ‘Referring to the succes of curve hop- ping, I stated “Either very litle struc- tural change occurs with each hop othe shear resistance effects of any change are almost recoverable with continued strain.” Ifthe latter (partial collapse of sol structure, recovery with strain, par- 7 A.J. Curtis and. Richart, Jt, "Photo- siaatie Ansogy for Nonbomogenents’ Founda: tong, Tranwacions, American Socety of Cv nioer, Vol 120,088 35 tial collapse, etc.) explanation is correct, then the sol’s seemingly elastic behavior is only a superficial observation. Mention of void ratio change during a single curve hop serves two. purposes. ‘The small magnitude of this change ie « necessary, though of course not & sufi cient, condition for indicating a small change in structure, However, it should be remembered that the main thesis of the paper isthe relationship between the Hvorslev and [DS effective components land that the Hvorslev components are determined with no change in void ratio. To compare with the JDS components, wherein a small change in void ratio is permitted, it was considered useful to indicate just how small this change usu ally is.PUBLICATIONS OF THE FLORIDA ENGINEERING AND INDUSTRIAL EXPERIMENT STATION “The Engineeting and Industrial Experiment Staion, College of Engineering, issues five setiee of publications under the general title, ENGINEERING PROGRESS at the University of Florida: (1) Bulletin Series ~ original publica tions of research, or conference proceedings, usvally on problems of specific interest to the industries of the Stace of Florida; (2) Technical Paper Series ~ eprint of technical articles by staff members appearing ia national publications 16) Leaflet Series ~ nontechaicsl articles of a general nature writen by staff smerbere; (4) Technical Progress Reports ~ reports of reseasch which is still in progress, and (5) Florida Engineering Series ~ books oa tecisical engineering. problems. A complete list of these publications ix avaliable upon request ‘The Bator Florida Engineering and Industrial Experiment Station University of Florida Gainesville, Florid