Helena Vreel-Koje'* and Stanislav Skrabl’ Limit analysis approach for passive earth pressure determination in three-dimensional conditions ABSTRACT: The paper presents non-dimensional coefficients of passive earth pressure influenced by the soil weight K,, and by the surcharge K,, in tabular form, regarding different geometric and soil parameters. Numerical values afford to reader pre cise applicability in geotechnical praxis, ‘The coefficients are calculated with a modified three-dimensional (31) translational kinematical admissible failare mechanism with metrical data. Therefore, these coeffic n the framework of the upper-bound theorem of the limit analysis, The 3D. coefficient is distinguished from the two-dimensional one, by its growing differenc nts are very useful at numerical analyses of different kinds of geotechnical problems, depending on soil properties and geo- where a 3D state gives more exact and realistic results. In comparison to other world-recognized failure mechanisms based on the limit analysis approach the results of this mechanism reached lower values, KEYWORDS: retaining structure, passive pressure, three-dimensional failure mechanism, limit analysis 1. INTRODUCTION Passive earth pressure acting on the rigid retaining wall hhas been widely studied in the past by the calculations based either on the limit-equilibriam method described by Coulomb (1776), Terzaghi (1943), Zakerzadeh et. al, (1999), and Skrabl (2006), on the slip line method described by Sokolovski (1965), Kérisel and Absi (1990), or on the limit analysis method described by Lysmer (1970), Chen and Liu (1990), or Soubra and Macuh (2002). ‘Three-dimensional (3D) problems of the passive earth pressure were presented by Blum (1932) toa restricted extent, by Soubra and Regenass (2000) with a multi-block translation failure mechanism using the limit analysis, and by Skrabl and Macuh (2005) with a rotational hyperbolical failure mechanism. his paper presents a new modified 3D translational kinematical admissible failure mechanism for determining the passive earth pressure coefficients within the framework of the upper-bound theorem of the limit analysis. The limit analysis theory determines the limit pressures that provide ‘Corresponding Author "PAD, University of Maiber, Faculty of Civil Engineering Smetanova lca 00 Maribor, SI-Skwenia Tel 043862 2296 327, fxs +438 2 24 79, E-mall: helen. vceleun-mb st ‘Ph.D. University of Maribor, Faculty of Ci Enginecting, Smetanova olca 17,2000 Maribor, Sh-Skovenia, Tel: +4886 222 94589, fx: 44386 2 252-4279, emai sanisaskrableaunimb si ring 200): (319-324) International Journal of Geotechnical Engin DOF 10.3328/8}GE,2010.04.05.519-324 strict lower or upper bounds to the true limit load by Chen and Liu (1980) or Michalowski (2001), This analysis consid- ers a general case of frictional and cohesive soils (@ and c) with the eventual surcharge loading q on the ground surface. ‘The numerical results of the 3D passive earth pressure are presented in the form of dimensionless coefficients K,, and Kyy representing the effects of the soil weight and surcharge loading, In conclusion the comparison with two world-rec- ognized failure mechanisms based on the approach of limit analysis is presented by Soubra and Regeniass (2000), and Skrabl and Macuh (2005). 2, THE DESCRIPTION OF NEW 3D FAILURE MECHANISM. A new modified translational three-dimensional failure mechanism within the framework of the upper-bound theo: rem of the limit analysis has been developed in order to opti- mize the 3D passive earth pressure coefficient as described by Vreel-Koje (2006). The 3D coefficient is distinguished from the two-dimen- sional one, by its growing difference, depending om soil properties and geometrical data. Thevefore, these coefficients are very useful when analyzing different kinds of geotechni: cal problems, where 4 3D state gives more exact and realistic results. Por example, it can be applied to retaining pile in the case of axially spaced piles, when the resi J. Ross Pablishing, Inc. © 201018 cross-section, plano view of the three dimensional failure mochariem lateral plane, and the ground plane of ‘he failure mechaniam along the x-axis piles along the embedment depth is analyzed by Vrecl-Kojc (2006). The following suppositions and limitations are applied: + soil characteristics present a homogeneous, isotzo- pic Coulomb material using the associative flow rule obeying Hill's maximal work principle by Hill (1950), + the translational failure mechanism is bounded by a polygonal sliding surface in the x-y plane, a rigid block of the dimensions (b = width, ht = height) in the y-2 plane, and the envelope of a family of half- cones at both Jateral sides, with a horizontal backfill, + the redistribution of the contact pressures over the entire height ft = | for the passive pressure due to the soil weight is triangular and is assumed to be inclined atthe constant friction angle 8 atthe soil- structure interface, + thevelocityat the soil-structare interface is assumed to be inclined at 8 to the wall in order to respect normality conditions used in Mroz and Drescher (1969), + the work equation is obtained by equating the rate of external work done by external forces to the rate of internal energy dissipation along different velocity discontinuities, 1 his + the resulting value of passive earth pressure is defined by Pek pba Kk gohb + Ksekb (1) ‘where is the unit weight of the soil, g is surcharge loading, and c is cohesion, The new translational three-dimensional kinematical admissible failure mechanism is shown in Big. 1, where the cross-section and plane view of the lateral part of the failure mechanism are schematically presented, The Cartesian co ardinate system is selected with the y axis along the wal A cross-section of the proposed failure mechanism with. the velocity field is presented in Fig. 2. ‘The optimal polygo- nal siding surface in the x-y plane consists of a final number of rigid segments, the mechanism is dimensionless with a height of '=1. The individual segment j has a starting point X go You) and a final point (0, Y,,) where the variables are calculated during the optimization procedure. ‘The kine matically-admissible velocity field is composed of j = 7 rigid segments bounded by the embedment point O (0, 1), and the final point (X,,. 0). In general, the number of segments can be varied, The kinematics of the segments velocities Vi are inclined at an angle of a, + ¢ to the horizontal axis, and the inter-segment velocities V, ,, are inclined at an angle of B,,..~@ to the horizontal axis The mechanism is defined by In-] angular parameters «(j= yu. n-1) and By, @= 1, +o}. The movement of each of the n rigid segments accom. modates the movement of the whole failure mechanism soil :mass, and its movement accommodates the movement of the retaining structure, ‘The seginent velocities V, and the inter-segment veloci- ties V,,,, are given by iB yay 24) @ Sin, =m, 20) sig ne) @) £8161 81> Figure 2, The cross-saction of the failure mechanism, and the velocity field of the failure mechanismfable 7. Non-dimensionat coefficients of K,, against g, 6 and bit H eee (aie bit 489 [vee = [02s | 25| raza] _essse| iiarn | 12.0088] Tessie 30] soee03 | 16,13 0657 | 22,0888 | 31.3718 35 | 19,0092 2075 | 42,2026 | 63,9804 40) sear (64,8862 | 86,3805 | 14,7490 [as [osser0 59,1052 | 202.2677 | 977,6050 Figure 8 Geometry ofthe frst Segment, and tne envelope of a farsi 050 | 25 40776 | 720655 | 6.1737 | 10,4917 ‘of hat-cones of other segments _| 30 [2.870 17,6088 [16,9926 |” 19,0086 Tas [iso ‘9.8400 280520 | 67.50 ‘The kinematical admissible velocity field is consistent een sae.orde [ait sat with the normality condition (at the angle @ to the sliding — | | 8.1954 191,9825 surface) not only in the x-y plane of the interface between | 1.00 | 25| 9.6731 pitta Figid segments but also on the interfaces perpendicular to [so] aero S665 | 12,9004 this plane. 36| "7a | sora1a| “13.5026 | 16.0201 | 26,2262 Fg. 3 presents the envelope of a family of cones of the 40| 14.2402 | 19,1600 |" 25,7527 | “sotto | 60.4705 first segment of the lateral body where the s-f is the local 45) 34Ba02 | 443705 83.0741 | 67,4000 | 122,0079 coordinate system and x-y the global system and the enve- | 200] 25| 2.9937] _3s7a6| ao7as | aeea7| S.aaag lope of a family of half-cones of other segments. 30|”s.a6ea[s.o76t| —o.0e64 | 7<0o4) e.re7e On the basis of these results, the volume of the separate 26|5.7a5¢| 9.0150] 9.9050] 12,0007 | 7.5401 block j and finally the volume of the whole failure mechanism 20 10.116¢[ 19,6290] “76,8606 | 21.6057] “Saa708 needed in the work equation of the upper-bound theorem of a6 | 26,1017] 29,5200 | a5arar] aa 7643) e081 the limit analysis can be calculated. The work equation and | "Eop[ a5] aver} sass) sas} aes apo the background of the limit analysis has been explained in 30] azee| aaa] ~4p0n1 | soon | 7enze detail by Soubra anc Regenass (2000), Michalowski (2001), 35 | 45800] 6.1690 | 7.7080| ~aaao8 | Taaoot or by Skrabl and Macuh (2005), therefore no attention will be EO ee paid to the background of the theory in this paper. a5 [ letra} zoams| zeasrel oraioe| cosa Considering the work equation for the condition of [aaa taet~sassst pest eee eee ee equality between the external rate of work and internal rate [10% : : : - of dissipation along the velocity interface for no cohesive SO} 811202 | 38820 | 4.6006] _ 5.4679] 7.2000 1igid-plastic material, the coefficients can be written by: 25} _42826| 5.9503 | 7.0781 | | s.esa6 | 12,1911 40| 68166 | 9.1607 | 17,3060 14,1600] 22,1161 kK -8 a @5| e.7e88 | 17,5408 | 20,0008 25,6529) 45,0012 a 2025] earra| 27e1e| 29802| sare] _4a779 ‘ 30] _3,0000| 3.5080] 4,1890| 5.0940 | _a7ors where f and g, denote the reduced values of the rate of work a5] a7610[ 5.1598) 6205) 7.4217 | 40.7000 due to the passive earth pressure and the rate of work due to 40] 47201 | 7.0403 | 0.79001 i210} 189306 the unit weight of the ground at K,, = 1, f, and g, denote the 45 | S109 | 1.8048] Te.1098 | 27.5000 | a7 2006 reduced values of the rate of work due 10 the passive earth pressure and the rate of work due to the surcharge loading con the backfill surface at K, The coefficient K,,, whith represents the effects of cohe- sion, can be determined using the coefficient of passive ‘earth pressure due to the surcharge K,y sce Kérisel and Absi (1990): 3. TABULAR FORM OF THE COEFFICIENTS The analysis gives the most critical non-dimensional three- dimensional passive earth pressure coefficient K,, and K, representing the effects of soil weight and surcharge loading, respectively, They ate calculated for different soil character- istics for 9 ranging from 15° to 45%, for three values of 8/8 (8/9 = 0, 0.50 and 1.00), and for three values of b/t (b/h = 0.25, 1 and 10).‘able 2. Noo-dimansionat coefficients of K.,, against g, & and bit L at aig (ans oe tee ae [pas [ io.eae6| i5.ea7a| i59488| 17,3758 rorea | za.0we| 25.070 2.008 25.1496] senee |-aasann] on0072 4.1068 02 a0 | 0.1000] 100 10 aa. 7 Seo | T6887 [teas eaeie| asrea]aaeee| ioasea co} a.saa1 | Taco0e “inaaee ce taarzo) eoance ar ao aaron [3.7008] 44.5708 [sizes | aese0e | caste] aB0r8 asl as) ates sees) acne af| “aai0 | a.s677 | Teaser | reoe70| 28.500 ao] 14,8564] 21,0604 | 26,9208| 344070| 52.6119 as] Base a7.as16 | a8.s008| 7.0000] 1219600 Tao] es] aasee | aasor] aera) saseo] —a2s80 GO| asise | ese) razea | —e7ea] to2s0r se] eer) a06s) Toros) Taawai| suze aa n0e0 [140007 ns00 | zaree7| “saree “| Teaies | 24 ie08 | si7ra7| aaeeee| 75.8788 Goo] 25 | 20000) saree | oaars| eaosi| awe 30.800 | ~aa0i0| s.n00 | ares] 70500 ce | aaoes | eseee | a.rea| acoso] aeses ao 70301 | toores | iesiee| isoora| e2.aze we Tizeae | veavea| ei orse | orstea| ariass roo] 28|2sraa| 33007] aces] acre] —4aaar ao asco | ater] sores] —seare| oro as| sees] cere] ra00a| aces] 10.9009 ao] 6037] a7oas| tossra | Ta005 | 1.8088 ae] aes] Taseoe) riseeo] Zasser! arava Bw] ae] ~B.es0 |” 20076) ~2s60 | aserel —aiea7 eal o| sooo | aasra] asis0| aaser | oaea7 ae] aroso | a ra4s| —eov00] 7.1604) 0.078 so[ sora] eposo| aod | yazeer| 16.7084 | [as 8.8051 | 9.6457 | 12,7800] 17,1684] 30.986 ‘Table | and 2 represent the values of the coetlicients K,, and K,, for different values of bl, different shear angles and friction quotients between the retaining strueture and the backfill sol ‘The analysis results show that for the values of K,, and Ky which decrease essentially by increasing the ratio of byl the coefficients resemble the value in the 2D state at b/h 10; likewise, the failure mechanisms of 21 and b/h = 10 have similar shapes, sce Kérisel and Absi (1990). Also the friction ‘ructure intertace 6 plays an important rate eat the soil as by increasing the ratio of Af the coefficients K,, and K increase essentially. ‘The resulls can be used in geotechnical practice Fig. 4 presents the critical failure mechanism in the «-y plane for different soil characteristics, The following conciu~ sion has been re-established from these results: any increase of the soil friction angle ¢ influences the failure mechanism; and while the volume of the failure mechanism increases, the shape of the sliding surface becomes more curved and the length of the last segment on the x-axis increases continu- ously. 4, COMPARISON WITH THE EXISTING SOLUTIONS ‘The three-dimensional passive earth pressure acting on a rigid retaining wall has been re-established using a simpli fied translational failure mechanism described by Soubra and Regenass (2000) and a rotational log spiral failure mecha- nism described by Skrabl and Macuh (2005). Soubra and Regenass (2000) published a truncated multicblock translational failure mechanism referred to as M,, which has been improved from his two previously pro- posed mechanisms, ie. the one-block mechanism M, and the multi-block mechanism M,. The improvement from M, has been obtained by a volume reduction of the final block, and from M, by increasing the number of blocks from one to 7. A comparison between this failure mechanism and the one presented in this paper can be made while both models are translational, using the same suppositions. The M,, mecha- nism of Soubra and Regenass (2000) has two major differ- ences from the presented modified failure mechanism, i. all locks have the same starting point, which is the origin of the x-y coordinate system, and just one portion of the right circular cones is used on each side in the lateral plane, Skrabl and Macuh (2005) developed this approach within the framework of the limit analysis theory. Similar to the mechanism described before, it is based on a three- dimensional rotational hyperbolical failure mechanism. The outer sides are laterally bounded by a curved and kinematical admissible hyperbolic surface, which is defined by envelop- ing the hyperbolical half cones and part of the ease surface of the leading half cone. Every point along the retaining wall height has an exactly defined hyperbolic frietion cone. A common velocity field connects all the bodies. Fig, 5 presents a comparison af the coelficients K,, for g 30° and 45°, and 6/6 =) aygainst the ratio bh.Figure 4. Critical falure mechanism bi and $= 90-40" 1000, 8451p Santa nas CBT | K, ’ | Shablan! Mos 2108) 1004 So 10 1 0 Figure 5. Comparison of the reslls for K, against ih for Sig lp = 0, ¢= 90" and 45" ‘The comparison of these results shows that the differ- ince in the coefficient is the greatest at g = 45°, at low ratio byh = 0.25, afterwards the difference is become smaller 5. CONCLUSIONS ‘The non-dimensional coefficients of passive earth pressure influenced by the soil weight K,, and by the surcharge K,, presented in tabular form that enable to reader precise applicability of these values in geotechnical praxis. ‘The coef- ficients are calculated on the base of modified three-dimen: sional translational failure mechanism within the framework of the upper-bound theorem of the limit analysis. In the past the three-dimensional passive earth pres sure was determined by a translational failure mechanism Oday described by Soubra and Reyenass (2000) and a rotational hyperbolical failure mechanism described by Skrabl and Macuh (2005). A comparison of the results for all three mechanisms shows that the values of the coefficients are improved 6. ACKNOWLEDGEMENT ‘The results of the presented paper perform part of the research work that has been supported by the Slovenian Agency for Researching, REFERENCES Blum, H, (1932). “Wirtschaftliche Dalbenformen und den Berechnung.” Bautechwik, 10(5), 122-135. Chea, W.E. and Liu, XL. (1990). Limit analysis in Soil ‘Mechanics, Elsevier, Amsterdam, Netherlands, 50-75, Coulomb, C.A. (1776), “Essai sur une application des régles de maximis et minimis 8 quelques problémes de statique relatifs & architecture.” ACAD.rSci. Mésn. Math, Phys. ar divers savant, 7, 343-382, Hill, R. (1950), The Mathematical Theory of Plasticity Clarendon Press, Oxford, U.K, 67-82 Rerisel, J. and Absi, E, (1990). Tables for the calculation of passive pressure, active pressure and bearing capacity of foundations, Gauthier-Villard, Paris, France, Lysmer, J. (1970). "Limit analysis of plane problems in soil mechanics.” Jourtal for Soil Mechanies and Foundation, 96{4), 1313-1334.whi, RL. C2001. “Ripper ber bead estimates on sane suid rectangular footings.” Geotecinique, 518), 87-798, Mrox, Z. and Drescher, A. (1969), “Limit plasticity approach to some cases of flow of bulk solids.” Journal of Engineering for Industry, ASME, 91, 357-364 Sokolovski, V.V. (1965). Static of granular media, Pergamon Press, New York, USA. Soubsa, A.H. and Regenass, P. (2000). “Ihree-dimensional passive earth pressures by kinematical approach.” J. of Geotech, and Geoeny. Eng,, ASCE, 2(2), 969-978. Soubra, A.H. and Macuh, B. (2002). “Active and passive carth pressure coefficients by a kinematical approach. Proc, Inst of Civil ng.- Geotech. Eng, 155(2), 119-131 Skrabl, S. and Macuk, B. (2005). “Upper-bound solutions, of three-dimensional passive earth pressures.” Can, Geotech. J., 4265), 1449-1460. Skrabl, S, (2006). “Interactional approach of cantilever pile walls analysis.” Acta Geotech. Slo., (1), 46-59. ‘Terzaghi, K. (1943). Theoretical soil mechanics, Wiley, New York, USA. Veecl-Koje, H. (2006). Upper-bound theorem used for opti- mizing a geomechanical model of a cantilever retaining, structure, PhD Thesis, University of Maribor, Slovenia, Zakerzadeh, N., Fredlund, D.G,, and Pufahl, D. E, (1999). “Interslice force functions for computing active and pas- sive earth force.” Can, Geotech. J., 36(6), 1015-1029.