BEARING CAPACITY OF FOOTINGS OVER TWO-LAYER FOUNDATION SOILS By Radoslaw L. Michalowski,' Member, ASCE, and Lei Shi,? Student Member, ASCE Aasrnact: The bearing: apacity of strip footings over a two-layer foundation sil is considered. The kinematic approach of limit analysis is used to calculate the average limit pressure under footings. The method is applicable tony combination of parameters of the two Ia ers Dut the re 2 footing is placed on & layer of granular sol esting on la Its are presented anly for specific ease when The depth of the collapse mechanism is found to he very much dependent on the streneth of the clay. Very weak elay can “attrac the mechanism even at teat depths, The results are presented as limit pressures rather than traditional bearing-capacity coefficients, ‘The later are strongly depen ent not only on the internal fiction angle ofthe nd, but also on the thickness ‘of the sand layer, cohesion of the clay. and surcharge pressure. Results are presented jn the form of dimen Sionless charts for diferent internal fiction angles of sund, Te way found that linear interpolation sithin 5 increments i acceptable inthe rat INTRODUCTION The bearing capacity of footings comprises quite an exten: sive literature today. Most design methods are based on sem- ‘empirical formulas. The original Prandtl (1920) and Reissner (1924) solution to limit pressure on a strip punch over a per- fectly plastic cohesive-ftictional weightless half-space is ust= ally altered to accommodate departure from symmetcical loads, to account for different footing shapes, and to include the resistance due to the soil weight (Hansen 1970). Either limit jonsooran empirical approach isusially used to account for the conditions not included in the Prandtl Reissner solution, The methods for calculating the bearing capacity of mul- tilayer soils range from averaging the strength parameter [ct Bowles (1988)]. using limit equilibrium considerations (Reddy and Srinivasan 1967: Meyerhot 1974), toa more rigorous limit analysis approach (Chen and Davidson 1973; Florkiewice 1989) ‘The finite-element method can capture the complexity of the boundary conditions and soil nonhomogeneity quite accu- rately, but itis more elaborate and has not found a wide acceptance in foundation design practice. The effort pre- ‘ented herein is restricted to finding the bearing capacity (oF ‘average limit pressure) on symmetrically loaded strip footings ‘over a two-layer foundation soil. The method presented can be applied to any combination of two different soils, but the specifi problem presented here is that where the footing rests an granular fyerunderun by a cohesive, possibly weak “The approach to solving for the besring capacity over a two-layer foundation system is presented in the next section, This is followed by a description of the collapse mechanisms considered in the analysis, Some comments about the solution are given next. Results of calculations are then presented in the form of design charts. and generic examples are given ‘A comparison with experimental results and other methods 's alo shown, Final remarks complete the paper. APPROACH The kinematical approach of limit analysis is used here This approach yields the upper bound to the true limit loads, ‘cane: Prof Dept of C: Engrs Jobs Hopkins Univ, Balin Mp IRIs ‘Grad. Student, Dept. of Civ. Engrg. Jobs Hopkins Univ. Bal: ‘Nate, Discussion open until October 1, 1995. To extend the closing sate one month waten request must he Hed wih the ASCE Manager ‘Tour. The mamascrit Yor th paper ws suited for review and posse publican on August 3, 18-Ths papet pet othe Journal of Genecical Engineering, Vol 121, No May. HS. ASCE. IGN ImsRstinsinmesas1 ONS a) 65.55 por page. Paper No 67 ‘of ¢ from AP 10 45" but. once reasonable collapse mechanisms are considered, it yields failure loss that are very close to the true collapse Toads on elasto- perfectly plastic bodies. This hus been con: firmed by comparing the upper-bound results to exact Solu- tions (such as the bearing capacity of weightless soil, and problems in metal-forming mechanics). For the bearing ¢¥ pacity of weightless sol the least upper bound is identical to the exact solution (Shield 1954), while for & ponderable soil (half-space) an exact solution is not known, but the discrep: ancy between the upper-bound and the slip-line solutions is almost negligible. The slip-line solution in this case can be proved to be the iower hound, provided the soil is restricted 10 a finite volume (not a half-space) limited by rough bound- aries. and deformation is governed by the associative flow rule. Close upper-bound and lower-bound solutions also can bbe found for ultimate loads over nonhomogeneous clay layers (Michalowski and Shi 1993). The kinematic approach of limit analysis leads to solutions identical to those from the “limit equilibrium” approach so widely accepted in design: the two are equivalent (Mroz and Drescher 1969; Michalowski 198; Salengon 1990; Drescher and Detournay 1993). The authors chose the kinematic approach as it has more appeal (0 e gineering intuition The advantage of the kinematic approach over the static, lower-bound approach, is that itis based on the construction of collapse mechanisms verifiable by experiments or practical experience. In the static approach, however. the stress fields ‘constructed without any clear relation to the tue sttess id other than the stress boundary conditions. (Note the difference between the lower-bound approach based on con structing statically admissible stress fields and the limit equi librium approach where only the global force equilibrium is required.) ‘Yielding ofthe soil is described here by the Mohr-Coulomb condition (compression taken as positive) Flo 8.04.) = (a, + a,sin g ~ VO FTE + recog = 0 o where g = the internal friction angle; and e = cohesion, The upper layer of the granular soil is deseribed by (1) with ¢ (0. and the strength of the cohesive soil (lower layer) is d scribed by the undrained shear strength equal to ¢y (¢ = ") The deformation is governed by the associative flow rule = 5 fs) ‘ hemnj= has a where €, and Bi he strain rate tensor: &, ‘a nonnegative multiplier (2)implies ditataney for pressure-sensitive granular soil the stress tensor: JOURNAL OF GECTECHNICAL ENGINEERING / MAY 1995 421and incompressibility of clays daring deformation. Conse ‘quently. in Fgid-block mechanisms where all deformation takes place along interfaces between blocks (within “rupture lay- ers"). the velocity “jump” between two blocks in granular Soil must be inlined at the angle of internal fiction @ to the Liscontinuity, and in clays it must be tangent ‘The upper-hound theorem states that the rate of energy Alisipation is larger than or equal to the rate of work done by fexternal forces in any kinematically admissible mechanism. Thus. if the material properties and geometry of a collapse smecchanism a Known, one can find an upper bound to the the rate of work of external forces tothe rate of internal energy dissipation. Optimization ‘of the geometrical parameters ip the failure mechanism pro> Vides the best approximation of the collapse load (least upper pound ‘COLLAPSE MECHANISMS ue 10 limited space, considerations are restricted here to ‘8 ease of practical importance when a layer of granular soil ited in such tthe velocity discontinuities originating at point A are bent at the interface between the layers, ‘The angle at Which they bend is equal o the difference in the internal friction angle of the soils in the two layers. The velocity-jump vector along segment AG, for instance, is then parallel 0 i along G:D.. Their magnitudes are also the Same. and block AG.D:D,G,A moves as one rigid body with veloci The hodograph for the entire mechanism is shown in Fig. 10h). A similar mechanism was anticipated by Karal (1979) No analysis or numerical resulls were offered by Karal, how: For the mechanism in Fig. 1. the rate of work by the un= known traction under half of the footing is equal to p¥)B!2 (Vc being the magnitude of the vertical velocity of the foot ing). The upper-bound theorem is used to calculate the av. crag limit pressure 7 along boundary OA, with traction q given along AF. The upper bound to the average traction p fean be weiten as pe [Jeme fava fave] where €, and V, = the strain rate tensor and velocity vector inthe kinematically admissible mechanism of collapse: ey the stress tensor: q, = the known traction vector on boundary Si(here AP): y, = the unit weight vector; and 2 = volume fof the collapsing mass. The first integral in (3) represents the rate of energy dissipation in the mechanism (discontinuities included in volume 1), the second one shows the rate of work fof 4 given traction on boundary S (AF), and the last one fepresents the work rate of the soil weight, "The mechanism consists of rigid blocks, and the energy is © TL Nee ° i, ® Fig. 1. ‘ism of Two-Layer Foundat ‘collapse Mechanism: (a) Rigid-Block Collapse Macha- jon Sol; and (b) Hodograph 4222 JOURNAL OF GEOTECHNICAL ENGINEERING / MAY 1995 clissipated in the shear layers between the blocks, The 1 rate of energy dissipation in the mechanism [first intext G)] was caleulated ay the sum of dissipation along all dis continuities. Since the associative flow rule is used. the Uis- Spution rate ong the velocity discontinuities inthe granular ‘material is zero, and in clay (g = 0) its equal 10 the product fof the shear strength and the magnitude of the veloxity-jump vector (dissipation per unit area ofthe discontinuity surface} More specifically. (3) can be written, for the mechanism Fig. 1-1 form po gfe S is — anew. Sav] w wtiere f= length ofthe th veosty discontinuity in ty (or Instance. G,D,. Dy Ds ete) [V], = magnitude ofthe ele iy-jump vestor slog that dictinuiys umber ot ftaghtline Gacontinuy tegmente; Vi vena compar ent the velocity of The Ath lock (negate tf upwae) ‘A, = aren ofthe Kth ack within the snd layer only nd f= numer of blocks (ey sandrained shear strength of ay ‘bunt weight of sand) All elotes ean he bund fom comets tlattons inthe hodograph in Fig. 1) “The second collapse mechanisms shown in Fig. 23) This imechansm iy ako symmetrical. with blocks AGDON snd AAGEJA moving a isd bodies, and with region GDE de> fovming in'a continous fashion, The fale pattern in lay resembles that nthe casa Pani mechanism the nergy dsipation rate nt integral in (3) ow inches ie Shation bh long uscontiutic and sith the continaly ‘tiorming region DE tn isms shown in Figs. and 2, nism confined only [simi in Michalowski (1993)} were considered. The objective of the calculations was 10 ‘obtain the least upper bound to the hearing capacity. and ich of the mechanisms considered is kinematically salmise sible failure partern. It was not clear prior to calculations, however, which mechanism would predict the minimum bes ing capacity (best upper bound), and all four different mech- tanisins were considered in order not to predefine the collapse ‘mode. For all cases analyzed here. either the mechanism in Fig. Lor the mechanism in Fig. 2 ensured the minimum hii load; therefore. the other two mechanisms mentioned ie not presented here COMMENTS ON SOLUTION ‘The results of calculations are presented inthe next section, Here, however. some consideration is given to the feasibility ‘of presenting the results in terms of traditional bearing « pacity coefficients. Also, it is indicated wien a weak clay “underlying a granular soil hasan adverse effect on the bearing cexpacily FIG. 2, Collapse Mechanism: (a) Fllure Mechanism with Contin- tal Deformation Field in Clay; and (6) HodographBearing Capacity Coefficients It follows from the incompressibility of clay ‘ometry of the problem considered that the net work done by the weight of the clay must be ze10. The solution to the ‘beating capacity must then be independent of the specific ‘weight of the clay. Dimensional analysis allows one to con clude that the bearing capacity for the two-layer foundation soil can be represented as, 6) p = average limit pressure under the footing: B footing width; 1 = thickness of the sand layer; y and the unit weight and the internal friction angle of the sand, repectvay = ‘undined shea strng of he cya surcharge load at the boundary adjacent to the footing fig. 1G): Sotutions based on (3) can be writen #8 p Snr Spe Wes aeh t Bren NEN +N 6) ‘The inequality sign in (6) indicates that the solution yields the upper bound to j'yB [as do all solutions based on the consistent limit force equilibrium method... Terzaghi 1943)] By comparing (5) and (6) one concludes that f, are not nec- ‘essary functions of the internal friction angle alone, as Sug igested in bearing capacity formulas for uniform soils. For instance. f; and f, (N.. Ny) for the case where q = 0 and ‘g = 35° are shown in Fig. 3 for different /B, as functions of c,/yB. These functions were calculated from the specific terms i @) applied to the optimized collapse mechanisms, where a minimum of bearing pressure was sought, and independent angles describing the geometry of the mechanism were var- iable, Coefficient N, drops down with an increase in strength of the underlying clay, and, for sufficiently large ¢,/¥B, it becomes zero. This occurs when the entire mechanism of failure is contained within the upper layer of granular ma terial. Coefficient N,, on the other hand. increases with an increase in ¢,/yB. and, when the mechanism becomes re- siricted to the upper layer. N, reaches a value independent of c,/yB (itis dependent, however. on @ and qiyB), Numerical calculations with optimization of failure mech- anisms show that all functions f,in (6) (which ean be inter- preted as bearing capacity coefficients N, Nj, and Ny) are dependent on q/yB. /B. ¢,/yB, and . Presentation of these functions would be more elaborate than presentation of the average beating capacity alone. Therefore, the results are Presented in terms of pyB rather than coefficients NN, and N, Critical Depth of Weak Layer Fig. 4(a) shows the average limit pressure (dotted line) and the depth of the collapse mechanism ft, both dependent ‘on the thickness of the sand layer (1/B), fora specific internal friction angle of sand and parameter ¢,/yB. Cohesive soil ‘weakens the foundation system, but the limit pressure (in- creases when the depth of the clay increases. When this rel ative depth reaches 2.7 (for this specific case) limit pressure ‘becomes constant, since for greater depths ofthe clay layer the critical collapse’ mechanism becomes contained entirely in the upper layer of sand. Atr/B = 2.7. however, the same least upper bound to p is associated with two failure mech- tanisms: a deep mechanism extending. into the clay’ (hiB 3.75), and a much more shallow mechanism, restricted en- tirely to the sand layer (h/B = 1.4). Fig. 4(a) indicates that, for g = 40, a clay layer with ¢,/yB = 3.0/is a "weak" soi It follows from Fig. 4(a) that the depth of the failure mech- ‘anism increases with an inerease in the depth of the weak clay layer, all other parameters being constant, A weak clay at a certain depth below the attracts” the failure mechanism, However, when parameter ¢,/B is large, the close proximity of the strong clay to the’ surface increases limit pressure p relative to that for the sand ‘The critical depth of a clay layer is defined here as the largest depth of the clay layer that still has an effect on the limit pressure. For the specific case in Fig. 4(a), the dimen sonless parameter (1/B) representing the critical depth is 2.7 Fig. 4(b) shows calculated critical depths for a varity of pa rameters (but all for a surcharge load equal to zero, q = 0), Iv indicates thatthe weaker the clay ayer, the larger the depth ‘up to which the clay has an adverse effect on the bearing (a) 4 70 Iss é 10 8 a 5 “ ap g nots wad £ “Tp al} Jed z PAA nes or mot wo eee $ $ FIG. 3. Bearing Capacity Costicients as Functions of Cohesion (of Undertying Clay for Diferent Thickness of Sand Layer, 1, and 49 = 0:(a} Coetielent N.; and (o) Cooticlent N FIG. 4. Critical Depth: (a) Depth of Collapse Mechanism and Limit Pressure as Functions of Thickness of Sand Layer; and (0) Critical Depth of Clay as Function of Cay Shear Strength JOURNAL OF GEOTECHNICAL ENGINEERING / MAY 1995 429,‘capacity ofa footing (critical depth). Also, the stronger the sand layer (the larger is). the larger the critical depth This observation indicates that the offen-suggested caleu- lations for estinnation of the eritical depth of the weak layer ay. function of the width of the footing and parameters of the soil immediately under the footing should be abandoned. Calculations such as those suggested in the present paper should be used [for case where q = 0 use Fig, 4(b)). Design charts for bearing pressure yB given in the next section however. do not requite that the critical depth be calculated first. This critical depth is accounted for in the charts, DESIGN CHARTS Results ofthe calculations are presented in a dimensionless {form in Figs. 5=7. All diagrams represent least upper bounds obtained from analyses hased on mechanisms in Figs. | and 2. whichever yields the minimum, ‘The number of blocks in the mechanism in Fig. Lused in the calculations was 20, The increase of the number of blocks beyond 20 affected the least er bound by less than 1%, ‘The independent angles de- the geometry of the mechanism were vatied in an ‘optimization scheme (with the smallest angle inerement being 105°), and (3) was used to calculate the bearing pressure p for each combination of these angles. The minimum of f was mensionless coefficient piyB representing the average limit pressure under a footing is function of four parameters, (5). nd i cannot be represented conveniently (without mak: ing erude approximations) as a function of bearing-capacity coefficiems dependent on the internal friction angle alone. Therefore, diagrams for piyB are presented without sepa rating components dependent on cohesion, surcharge load. sid the unit weight of the soil (this would require three times more diagrams) 4s. 5-7 present results for surcharge load q/yB equal to 0.115, and 0, spectively, Results are independent of the unit sight of the clay: 9 i the unit weight of the granular Soil gis its internal friction angle (g = 0 for lay), and cy is the undrained shea strength of the clay (no cohesion in the upper kiyer) In each Figure separate diagrams are shown for the internal friction angle of the upper layer: 3". 35", 40. and 4S?. The separate curves on the diagrams represent the ‘dependence of the average limit pressure on the shear strength ‘of the clay for a single depth of the clay soil. These curves are shown for ferent 1/8 the range trom O10 5.550 an adverse effect on the bearing capacity. Ax expected, the limit pressure increases with an increase in the strength of the ely and with a inctease inthe clay depth, For most cases the fimit pressure reaches a constant value and further crease in the clay strength does not improve the bearing ¢3 pacity. This Limit is equal to the bearing capacity ofthe gran- lular soil alone. Only when the clay is strong and the layer of sind overlying the clay is thin relative to the footing with ‘an the bearing capacity increase beyond that expected for homogencous granular soil [sce Fig. S(a) and (b), and Fig. ‘o(a)]. In such cases the bearing pressure reaches higher constant level at large ¢, yf (beyond ¢,/yB shown in diagrams here). Inthe case when wcity nereases proportionally to the clay st The critical depth is implicitly included in the diagrams in 5-1. Ifthe value of pyB read from these diagrams for ven 1B is les tha the limit set by the granular soil alone {horizontal lines), it implies that this specific #/B is less than the critical depth. the value read from the diagrams is equal tw the constant limit independent of ¢,/%. the specific 1B 424, JOURNAL OF GEOTECHNICAL ENGINEERING / MAY 1095 is larger or equal to the critical depth. Using the charts in Figs. 5-7 does not require estimation of the critical depth “The following two examples show the application of the charts in Figs, 5~7 in the design of footings. Example 1 Find the bearing capacity per unit length of a 2-m wide (B 2 m) strip footing placed on the surface of a -m layer (1 4m) of sand ( — 35° "The sand layer rests of clay whose undr mm From Figs 3 for = 35. cy'yB = 1 sand 1B obtain xB = 22.5. hence the’ average limit pressure 765 kim (calculations using the computer program: 7) = 764.16 kN). The bearing capacity is then equal to PB 765 2 = 1,530 kN, the bearing capacity of a4-m wide strip footing placed ata depth of 2m (D = 2m) in a 12-m deep layer of sand (t 12 = D= 1m) with g = 42° andy = 17.5 KNim The clay beneath the sand has an undrained shear strength of G, ~ 10S KNim®. The surcharge load is q = -yD = 35 kNim*? ey'yB = 13 and 1/8 = 2.5, Prom Fig. 6: Bene bape wl ca CO) 120 Te ower | 300 | ~-2~-bean creamed ye (1880) “Shoe ra (1980 ~sienae ae Meyerot (1980) B txn/m’) Ft. ‘Tests for Footings with 8 = 0.05 mand q = 0 in view of the low rule in (2), is kinematically inadmissible. This leads to conservative estimates with respect to the fig ‘rous upper-bound calculations, for some combinations of parameters (see curves 1/8 ~ 1.0 in Fig, 9). Overestimation of the bearing eapacity by Hanna and Meyerhof (1980) be: yond upper bounds calculated here (for other combinations Of parameters: curves for 1/8 ~ 2.0, q/yB = 1.0), is de to 425 JOURNAL OF GEOTECHNICAL ENGINEERING / MAY 1995 FIG. 9. Comperison of Computational Results not allowing for adjustments of the collapse mechanism (op- "The design charts presented in Figs 5~7 were derived under assumption that the deformation of the sand layer is gos- cerned by the associative flow law (normality rule). It can be12, a yo] _° Satis (1982) FIG. 10. Comparison of Bearing Pressure Calculations to Finite. Element Computations (Nonassociative Flow Rule) argued that the normality rule does not accurately describe the deformation of the granular soil, and a nonassociative law be described by (2) O replaced by a plastic 0. Further, function g(a,,) = 0 can Coulomb function. (1). where al fiction angle gis replaced by the dilatancy angle (Davis 1968). Theoretical considerations (Drescher and De- tournay 1995) allow one to conclude that for such a nonas- sociative model of sand (¢ = 0) a solution can be obtained using the same technique as presented herein, but with the internal fiction angle ¢ replaced by angle y* calculated from should be used for sand. Such law c with the tlre funetion #(0,) potential. say. (04) fi.) show he comparison of bearing apis calelted by Gite (NS wing the hinted suming tenant mompesibe (=), hte one calcu thing te dnp cans presented hte lange alate ton Oi The two cine rma wll FINAL REMARKS ‘A method was presented for calculations of the bearing capacity of strip footings over # two-layer foundation soil system, Design charts are shown for the case where 2 layer Of granular soil overlies the cohesive soil. either weak or strong. The method can be applied easily to a general case Where the strengths of both layers ate characterized by both internal friction and cohesion “The formulation of the bearing-capacity problem and the solution using the upper-bound theorem of limit analysis is conceptually straightforward. The solution does not require introducing eoncepts or assumptions [such as the “coefficient ‘of punching shear" (Hanna and Meyethot 1980)] used in some limit equilibrium approaches to avoid staical indeterminacy. ‘The advantage of the upper-bound approach is in the clarity ‘of the concept and the small effort needed to generalize the ‘olution fo include both friction and cohesion in both layers, inclined loads. and so on, “The simplicity ofthe solution to the relatively complicated problem of the bearing capacity of footings over two different Soils was achieved by introducing a failure mechanism where the velocity discontinuities were bent at specific angles atthe interface between layers. This allowed one to consteuct simple hodograph, ss in the case of uniform soil. Opti zation of the geometry ofthe mechanism led tothe least upper ‘bounds. [t was found, however, that, for a strong first layer (large ¢) and weak underlying clay. mechanism with 2 con: tinual deformation field in the weak layer was more effective “The collapse mechanism that assures the least upper bound to the bearing capacity can attain a very large depth, far exceeding that for homogencous soils. This depth becomes particularly large for a strong granular layer (high internal friction angle) and a weak clay underneath, ACKNOWLEDGMENT The results presented in tis paper are bused on work supported by the National Setence Foundation under grant No, MSSHA01498 Thssop- ports gratefully acknowledged APPENDIX |. REFERENCES: Bowles JE (198) Foundation analeicand design th ., MeGiraw il, New York, N'Y (Chen. WF and Davidhon, HL L973). “Bearing expacty de terminakin by limit analyi” J. So! Mec Found. Die 946) au Danis. EH. (1968). “Theos of patty and the fare of a mass Soni mchanestleed opis. 1K Le, Ea Butterworth, London England. 341 Drescher. A. and Detourmay, (1998). “Limit tin Tallre mechanisms for anocitve and non-isc Geotchnigue, Landon, Eglin. 33), 0-330 Floksewies A. (089)"""Upper hound to heating capacity of tayered oil Can Goer Jn 8), T= T36 Grits, 'D. V-"1842). "Computation of bearing capacity’ on layered teil" Pro, hf Conf Nom. Meth Geomeeh, 2 Enemtcin, 6 Balkema, Rottentam, The Netherlands. 18-110 Hann, A.M sand Meserhol GG. (190), “Desgn chars for aimate Tearing capacity of foundations on sand overjing st ly.” Can Geoteh 1702). 98-8 Huser, JB. (1930). "A revived and extended for opacity” Grvtekaik frst Bl. 28 8-11 Karls. (1979). "oegy mid for soit." PRD thes, ‘Noregin Insitute of Fechnolgy. Trondhcin. Nora MesertolG.-G. (1974). “"uiimate heating pac footings on sin layer overying clay." Cun. Grote Jos 11(). 238-229 Meyerhot, G. G:-and Hanna. A.M. (197), “Uimate beating capacity of foundations om layered sas under inclined al" Gan, Grech, J 1544). 868-872 Michawaic RL (1889). ~“Theceme sional aay foal sloncs = Génie Lotion. England. 39.27" 3% Michlowshie RL (1993). "Limit amas of ak Tayers under em Trinkets” Ss wd Pond. 33(1). 153-108 Michalak RL. and Shi, L- (1999) “Bearing capacity of mobo ogencox cay ers ander embankments" Gaaerh Eng ASCE, TT) 1687" on Mess Z. and Drescher, (1969). “Limit plasticity appre to some ise of low of lk sola" Eg I S12). 387" So Prandd. L. (1020), "Uber die Harte plaster Karp” Nachr. Gx Wisensch, Goningen. math-ph. Kae. 7-45 (m German). Reddy. A-5.-andStinitanan 1 1967) "Bearing espacio footings ‘on kyeredstyn "J. Sot Mech. Found. Dis. ASCE. 942}. 43-, Resists (1923). “Zum Exddtuckprohioa.” Proc, Frat In Cone Mor Appl Mech »C. 1B. Biczeno in}, M, Burges, Edk., Technische Bockhandel en Drgkkeri J. Waltman Ir. The Netherlands, Del mea gon J 1980). “AR spc taal mecha asi Shick, RT. (1983), “Plastic potential theory and Prandi beatog city solution" Apple Mech 2112) 193-13 ‘Teraghi, K. (1949), Pheoretcl so mechani 3, Wiley & Sons. New York: NY. rduction tthe yet design theory and is fo Ey. J Mech. Ser A'S. 13). APPENDIX I NOTATION The following symbols are use in this paper: B = width of footing: 6, ~ undrained shear strength of clay JOURNAL OF GEOTECHNICAL ENGINEERING / MAY 1995 427D = depth of footing fla,) = yield condition velocity vector: "f, = beating capacity functions (( = 1, 2.3) unit weight of sands, 1h = depth of failure mechanism: Strain rate tensor: NNN, = bearing capacity coefficients; B= aserage imi presi (bearing re 4g = overburden pressure (yD): silataney angle of sand 428 JOURNAL OF GEOTECHNICAL ENGINEERING / MAY 1905 thickness of sand layer or depth of clays = internal fiction angle of sand: and