You are on page 1of 9
The overall stability of free and propped embedded cantilever retaining walls by J.B, BURLAND", MSc (Eng). PhD, CEng, MICE, MiStructE, MSAICE, FGS. DM. POTTS?, BSc, PhD, & N.M,. WALSHS, MA, MSc, DIC 1, Introduction THE DEVELOPMENT OF slury trench and Jecant bored pile construction methods hae led to the widespread use of free Gnd propped "(or tied back) embedded Cantilever retaining walls in 2 variety of Ground conditions, The free earth support Srethod of analysis (Which was developed {or 2 sheet pile wall design) is widely Used for assessing. the overall stability St'such walls. It is well known that some Sesign methods which have proved suit- bie for sheet piles in granular soils 2p- Fear 10 lead to excessive wall penetro- Fins and expensive designs for clay soils with angles of fiction less than about 3s"" Moreover, there appears to be some Uncertainty about the definition and mean ing. af “actor of safety’ for the overall ‘Stability of embedded cantilever retaining walls. Tn this Paper some of the commonly used design methods for overall stability Based on factoring the passive resistance fof the toe are examined. These methods feed to widely difering levels of safety with respect to shear strength (some Gangerously low) for # variety of condi- fore and’ sometimes give. illogical re- Sits (ep. o decreasing factor of safety Srith increasing. wall penetration for Yorm. ‘ground conditions). The present position is clearly thoroughly unsatistec- ry % revised definition of factor of salery with "respect to. passive failure is pro- Posed which appears to overcome the Sitfeulties encountered with present de- fnitions. The revised definition is based fon an” analogy with bearing capacity Gheory and is shown to be consistent juin the factor of safety with respect 0 shear strength. 2. Comparison of conventional factors of safety Fig. 1 shows an embedded propped cantilever wall retaining © material with 7 eee Sal ees {roaches Engineer, Ove Arup Parte = PROP — both cohesion and friction. At timiting equilibrium the active pressure behind the ‘wall and the passive pressures in front of the wall ere axsumed to be fully mobilised Tike prop is compressible. For horizontal equilibrium PROP + P,P, = 0 o ond Prete - Par tg =O @ Eqn. 2 is used to determine the depth of penetration dja of the wall to give Timiting equilibrium: The values of P, and Pr are usually obtained using eerth pres Sure coefficients derived from Coulomb's theory, or some other method such 9s Coquot & Kerisel (1948), which takes ac- Count of wall (ition. "The problem is to determine the depth of penetration of the wall to ensure an Sdeguate margin of safety against overall instability. Various design methods are in current use and four of the more com- mon ones are described briefly as follows: Method 1. Perhaps the most widely used method is to specify that the moment of the active forces P,. Ly should not ex- Coed @ Iraction 1/F, of the total passive fesisting moment Fp. Lp. The quantity Fr ie the factor of safety? with respect 10 the total passive resistance of the toe land could ‘be reparded ss 2 load factor Values of between 15 and 2.0 are usually employed (NAVFAC (1871), CP2 (1951) Method 2. An approach advocated in the British Stee! Corporation Piling Handbook is to derive the net horizontal pressure distribution acting on the wall (see Fig. 3). The design condition is such that the moment of the net active forces should fot exceed @ fraction 1/Fy, of the mon nt of the net passive forces. Fay is 2 foad fector and value of 2 is normally feet. be made ‘bout o Active zone: Fig. 1. Propped embedded cantilever retaining well a limiting equilibrium Motos 3. A paricuary simple methos reo determine dq tom ean. 2 890 Mu Moly. t byt iedor Fy. Tachebowriot (873) recommends 2 vdive of Fy = 17 [oefranctar suis ond 2 fr undained cohe ecole wheress the USS. Stes) Shee: Fling’ Marval recommends Fy = 12 10 se moy be thovght of 86» scaling factor Mathod 4. A more fundemental por than the previous thee methods fe 0 Buse tae the depth of penetration os Sich thot the average strength recuiee To give limtng equilbsim does net es See “acto W/F, of the avaiable Stengin. Te quanti’ Fy isthe factor ot Seer with respect 19 shear strength. For eecrely ticltonal soi" Fy = tan @/t Soot Fe) where tym 1 the angle o fizion eevired to. We limiting ea: tom, For undrines cohesion Fy = Col cree Fe When the. sol posses! Satie ond & soparate values of Fane F fray be used (eg Taylor, 1948; Hanser 1580). Values Of Fp between 1.25 and 15 TAPE, between 18 and 20 are common aed Te isnot the purpose of this Paper t argue the monte of toad factors. vets Eigen factors as bot have saventane She" aascventages, Whichever method Stes thould nowover be consisten! 2 logical Por example, the adequicy ol + {eating with respect to, Bearing cops" foiure may be expressed eer #5 + food tector Fy = Qy/0 or a6 9 factor ‘Sten’ Fon the avergge avaiable tes FI tne Tomer beng. more. frequen Soba. On the other hand the steity et slope is invariably expressed as #1 fof of salty on senor ‘aimouah’ methods 1 to. 3 are, wit used ee nn ny 10 i Tie the’ physical sigmifeance of a giver tht of F tn relation ‘to the more common pro? atat'ot slope stabity and bearing OP Ir the tree methods wil thereore SBinpared with method & (actor 3 5 Se” erengtn) for two simple bot |! General cases. ruined concions with © “re tne labelled (L) in Fa. 3(¢) 2072 tne imiting reisonahip between H/? Ser’ proppes embedded wall 0 2 eae Fig, 2. Net horizontal pressure cist! stan '/ 109 Bim (b) = Relationship between o' and (a) H/h and (b) factor of safety on shear strength vious design methods form tictions! soil with the water tele at excavation (or deedge) level. Line (L) is of course comman to all methods and Conesponds to 8 lactor ol salety of one Line (2) corresponds to 8 factor of solely on strength f= 15(=F4) ond is easily Gexved tom line (L) ne (3) is obtained using method 1 for a load tector F, = 2. For values of above about 30° line (1) ives. similar Selves of H/h to line (4). This probably ecounts for the success of the method when used for embeddes wall in granuler ols However, 08 decreases below 30° S vaive of F, = 2 gives increasingly con- fervative values of Hh in comparison win F, = 15. Indeed it canbe shown thet at g = 10" the volue of H/h 9p. roaches” infrityS (or Fy = 2 compared aha valve of 65 for F, = 15, Line (2)"is given by ‘method’ 2 with 2. te'can be sean to lie close to (U) implying @ very low factor of tafery on shear strength, In contest line (G). which is. given by method 3 Fz "17, early coincides with line (2) Tlying an simost constant fector of sale- ty on strength of 15 In Fig. 3(b) tho above lines have been replotted giving» more direct compar Son between F, and the three other me ods. It is clear that both methods. + snd 2 are unsoistoctory. A value of F =z ‘gives very variable levels of sefety with feapect 10 shear suongth snd. very Conservative values of H/h for values of $ less than sbout 25°. A value of Fay 2 gives 2 constant level of safety but the {actor of satety on strength is always le than 13, Surprisingly method 3. with 1.7 eppeers to compere, most fevou bly with method & The main objection qo the use of method 3 would be for var- ble ground conditions beceuse the me hod dose not take account of ground eon. ions Below he lmting depth of pee: Undrained” ays. in Fig. 4 Ine (L)_ shows the limiting relationship between “H/h and 2ey/yh for" propped. embedded cantiover watt in ely. with cy constant with eptn "(ignoring tensile stresses). Stove ‘be noted that there iS a teal depth of excavation Ayn (e=4ex/) Below which instability. occurs Wespec: tive of the “depth of penetration of the wal. Line (4) represents. the relationship Between H/m and 2ey/9h for Fy m= 18 ale, ua). Lines” (1) are given by metned 1 id correspond to Fy = 1.5 and 20. it can be seen that they difer érama- tially from line (@). A panicularly stan. ling festure i that when 2ey/yh fe held constant the value Of F, decrasses with increasing wall penetration — 9 result which i clerly not logical The lines cor- Tesponding 10 Fy = and F, = 2, hough not identical, are go similar thet they ore Shown 08 8 single line n Fig. 4 labeled 2 and’ respectively. iis evident that mete hode 2 and 3 bath give very low lactors OF salery with respect to strength General remarks “The use of Both F, and Fy, can lead to very unsatistactory results, There oppe to be no logical or consistent relationship between F and the factor of ssfety on strength and ite use can lead to very con Servative values. of wall penetaion for daines maser Joly, 1881 Passive Game ae HAMA 2 Kp, Vg. 82g Ka) Ye Rea 5 owt resting pressure yp Vest (pM) ds Ka (ed drained conditions with 9° less than 25° nd tor undrained conditions. With regard fo F> its use with currently recommended values of about 2 leads to factors of safe- ty. on shear’ strength generally less than Yh Tor both undrained and drained condi tions, It should. therefore only be used with great caution and with much higher Velues which are compatible with accept- ed values of F,.. For drained conditions {h uniform ground the use of the simple Seale. factor Fy appears to be entirely Entisactory. However, it, should not be Sed for undrained conditions or where the strength properties of the ground Vary Significantly with depth. Tne concept of factoring passive resis- tance of the tos of an embedded retain- ing. wall is. attractive in principle as the Overall stability can be expressed as Single number. This is not the case for Shear strength where, for all but the sim- Shest ground conditions, the engineer may Be faced with the possibility of 2 number Of factors for strength and wal friction. Hence there are considerable benefits in Beveloping # definition of factor of safety fon passive resistance which can be shown {0 be logical and consistent. 3. Simple bearing capacity analogy ‘There is # reasonably close analogy be- tween the overall stability height of ground and the be Of 2 stip loed. kt was decided to pursue fhis analogy to see it 2 more logical de fition of factor of safety on pessive re- Gistance for an embedded retaining wall ‘Could be developed. The activating pressure Aetve zone 1 ta 5 Re Yee he ecg reine lte Barbas * Fae Ra, Fig.6 (above). Revised factor of safety for on embedded wall ine frietional soil Consider a uniform surcharge @_ acting fon the surfece of @ granular material hav- ing an angle of shearing resistance ¢° end rho cohesion, At limiting equilibrium cles- Bical bearing. capacity theory (see Fig 5s) ives an active fallure zone | immed: iately, beneath the loaded ares coupled by @ zone. of radial shear, zone Il, 10 8 passive fale Zone Ill outside the loaded brea, For present purposes this can be tive and passive zone seperated by 2 rigid we Fimiting equilibrium the passive resistance Ppa dey Po Ke and the ective foree on the wall Py may dK at bey Ky For equilibrium P, must equel P, and hence solving for @, givest Hoye (KK) a= e aK, Now the factor of safety with respect to bearing capacity is given by: pe wt NI ok) q asd where F; isa load factor, The quantity (Ky, “6 Ey ot eensting peeeve Fig.5 (elt), Bearing capacity of 2 strip load (a). classical failure mechanism (b) simplified failure mechanism (2). defintion of factor of satecy FB cinta oresare o =K,)_may be thought of as the ne: Tesiltant) passive pressure coefficient. t an be seen from Fig. Se that the desir Eondition is such that the horizontal fore Setivated by the applied pressure @ (= GK, -d) must not exceed the fraction Tit! of tne. net availble passive cess tance (= do 7s8t (Kp = Ka)) of ground. 4, Embedded retaining wall ‘The problem of an embedded ret2 ‘wall may be weated in 8 we the bearing capacity problem. In eda” to # vertical pressure q acting on ground surface, there will be @ sheer 1 Fond # bending moment M acting 2 top of the embedded portion of the *™ due to the horizontal pressure of the ov! Iying retained soll, see Fig. Ga. A revs: Gefnition of factor of safety which * Consistent with the bearing capacity 2° fem would be that the moment ac? by the retoined material should not <= 2 cerin traction 1/F, of the moms" the net available passive resistance © ulated using the net passive pressure < eficient. (Kp—K,). For the problem = Picted in Fig. 6b the revised factor salety F, is Pre brs B cS Pay -ban + Pag ba ‘The broken tine in Fig, 7 she relationship between H/h and @: <-> ponding to F, =2 for the problems & Viously: analysed in Fig. 32. Lines (4) * Fs, 1(a) Comparison between revised factor of safety F, and factor of safety of shear strength F, fore frictional sol. (a) Reaconap between, snd for various vas of (4) are taken from Fig. 32, Also shown ‘i Fig. Ta are the relationships between F, (= tan 4/190 6 ua) and g” for F, 1'5 and 20. is evident that given valve ol F, corresponds to ® reasonably con- stant’ factor of safety with respect to ‘Soer strength, For exemple, when F, = 2) the values of F, vary trom 1.4 to 1.52 2 varies from 30” to 20°, When F, = 15" tne. corresponding variation in Fy is ftom 1.22 10 127. Kt can be seen from Fg. 7b that there is an approximately linear Frionstie beneeen Fang W/m tons yen value of ". This’ is a feature that an be usetul in design Tk appears that the revised factor of salety on passive resistance F, provides 2 satisactory method. of obtaining the ath of penetration d in 8 purely tri 'cnal soil, The application of F, to soils, Bossessing cohesion will now “be con- sored, 5. Soils with cohesion and friction Wthe wall ig constructed in # soil which ossesses both cohesive and. frictional srength “componente (c'-_ ¢ material ‘he "active and passive horizontal pres sures at limiting equilibrium may be ex ‘essed in the following manner, where o,' is the vertical efective stress and Ky. Kp. Kye and Kye are earth pressure coeficients which will Usually include the effects of wall {ction and adhesion, Considering again the simplified bearing ‘capacity analogy discussed in section 3, ‘we obtain the following expression for Py and Pat limiting equilibrium, see Fig. a, Pam ay Ky St Hy Ky elieikine Pp hey Ky 2.0 Kee Equating the sbove expressions and re- arranging yields, Feed Kp Ky) 42-004 (Kae + Kre) a= ema 6) ‘and the factor of safety with respect 10 bearing capacity is given by: Foye dt pa Ky) + Bled (Kye Kre) ea 6) aK, ‘Again this may be defined as the net available possive resistance divided by the horizontal force activated by the applied pressure g, see Fig. 86. Inspection of the ‘above ean. (6) and Fig. 8b indicates that the ‘cohesive active pressure coeficient Kge contributes positively to the net ble passive resistance ‘Tne limiting horizontal pressure. distri- butions ‘acting on @ propped embedded wall are shown in Fig. Se. ll, ae is cus Tomary, it ig assumed that the’ soll cannot sustain tensile stresses the zone of nego et resisting presse Keke oyye8) = 2k On Ye 6.1 Korky) be 2k cag Ke2eiKae en 2cihec net ectiating pressure ry Fig, 8. Definition of factor of safety for a strip load on a c'—¢" soll tive active pressure near the top of the ‘wall ig ignored. By making use of the Dearing capacity analogy we obtain the following expression for the revised factor ff safety for the wall shown in Fig. 9. Prxs-boxs + Prva boxe fn a” Pasta + Pas Las The pressure distributions used to eval- vate the thrusts Py., Pas, Pry, 29d Pras are indicated by the ‘shaded ‘zones int Fig. 9b. 6. Undrained conditions the undrained stability of the wall is under investigation the clay will have an Undrained shear strength c,. The horiton- {al sctive and passive prescures then re vce (assuming no wall adhesion) to = Pr =2e, +202 where of is now the total vertical stress. “The distributions of limiting horizontal pressure acting on the wall are shown in Fg, 102, lanoring tensile stresses it can be seen from Fig. 10b that the revises fec- tor of safety F, is given by Pox Lox Pay ober + Pas Las S-Lpy 4 P= 2-6)? Lal Tee yeh bg” te, @ In the above example it has been impli July, 198131 Keyed 22Koc e) Tk L Be ret resisting pressure Few Fame bon* Protons Patay * Paacbas © citly assumed that the setive pressure at excovation lovel is positive, tht isto 53 Fen SM this int the case ond Beth Sit follows trom the. con proctice of ignoring. tensile iS no active thrust cling on the wall sbove excavation level. ‘The activating moment on the wall then Consists solely of the moment. derived fom the surcharge eect ofthe sil above txcevation level, The revised factor of DD stiay’'s" thon evened by substring Pe, 0 in the above expression for Fy ‘lich then reduces 0 eqn. 6 with @ re placed by 94h ing, Vt the revises factor of safery Fis compared with the factor of safety fon strength F, (= ¢4/eq jue) f0F 8 prope ped embedded wall"in' slay with ¢, Constant with depth and no wal adhesion eecan be seen that a2 H/b increases for 2 given valve of 2-cq/yh Both factors of Safety tend rapidly 18 the. same. limiting alue (4. ty/9h), Both methods show That tere is normaly Ile sdvontege in Seepening the wall beyond sbout H/n $B to 20 ac thoreser tore is litle gain in factor of stety. Fora given value of 2.¢,/9h the re vised method tends 10 the liming valve Dt factor of tery more rapidly than the Strength tector method. The reason for this ciflerence can be explained by refer fing 10 Fg. 100. The vole of Fis obtained ang the setual strength e, of the sol However, Fe is derived by" factoring cy To obits e! yy. In doing” this not only ore he forede'B., Pac and Pye chenged But also the lever or Ly. is itered, ie the geometry of the loading a2 well 96 the magnitude is chenged “tne Case when 2-cq/yh = 11s of par ticular interest The retained material ex: 8 Yd. lhenka) be 2e(hpe-Kac) = Kehoe ae fe 2elkie net cctivating pressure Eee Yyh-2cy iF ® Fe = Poko as obag? alas () Fig. 9 (left). Revised factor of safety for an embedded wall in « ena sail Fig, 10 (above). Revised factor of safety for undrained conditions 18 no pressure (P,, = 0 in Fig. 10b) fnd the revised method eats the prob- em as purely one of bearing capacity with no Benefit derived from wall pene- tration, However, if ¢, is factored, wall pressures develop and hence the strength factor approach requires some wall pene- tration and i intrinsically more conserva- ‘The above example is of course artificial ‘as in practice the wal is always design- fd to resist earth pressures, These can be derived either by reducing the value ofc, above dredge level to allow for Softening or by assuming partial or com- Dlete drainage and carrying out an effect. Ive stress earth pressure caleuation, In Factor of Safety os 1 her case the revised method provides logical approach by separating the 2c ‘ating moment of the earth pressure ex feted by the retained material from the fctivating | moment generated by ils bearing pressure. For comparison the dotted line in Fi 11. shows the variation of F, (method 1 in Section 2) for 2 . cy/yh = 0.7, As H/h increases F, increases’ to s maximum of 1153. and then decreases tending to ont as H/h becomes very large. Such verie tions in factor of safety are clearly not logical "The results given in Section 4, § and 6 demonstrate that the revised factor ol safety F, gives logical and consistent re 2eu/Yh =1 H/h Fig. 11. Comparison ofthe revised lector of safety F, with F, and F, for undrained conditions Pe ry YeEHR piKeekg CaP Ka, emt cessteg pressure et cetnating pressure ) (e-6hke rs) ‘ig. 12. Definition of factor of safety for a strip loed with sur- charge acting s v Las ies EE net resisting pressure = 5. Kesh Pour = 6 Veed?( = Yee ¥s.h? Ky Poy = P-d (Kp-Ka) = (Ss syph-p).d. Ky Bs (Ke Ka} 1% Ys-d- Leys P» Lewe K.[5.h. Lars 4 -Ys-h?. Lap + (5+ Ye, h-p).d Las] * 5. 18, Revised factor of safety for embedded wall with surcherge net activating pressure ) rs sults for both frictional and cohesive soils, Ivis peninent to examine how F, may be used to take sccount of surchrge. and seepage pressures and this will be done ia the following. Sections, 7. Surcharges Wa surcharge is placed on the soil sur- face behind the wall it causes an increase Jin the active thrust and reduces the sta- bility: of the wall, When ealculating the revised factor of salety F, the moments of the lateral earth pressures due to # charge are added to the ecti ‘Surcharges acting at the Bottom of the excavation tend to increase the. st2 ff the wall by increasing the pessive istance of the soil. The evaluation for such a sivation isnot immeditely nt. However, the situ sn be lated by onaicgy once again with the bearing capacity problem. Considering the simplified ‘ilure condition discussed in section 3, (see also Fig. 5b) with sur. ‘charge loading p we arrive at the situation ‘depicted in Fig, 122, where, et limiting ‘equilibrium, Pe Ded Kp F474 8 Ky and Pam ay OK t by Ky ‘As these forces must be in equilibrium, we obtain Ded-Kp tH. y0 088 (Kp —Ky) % OK, (9) Now the net ultimate pressure that may bbe applied, a, is given by ey = 4-P ‘ond noting that it is customary to define the llewable bearing pressure @ we obtain the following expression for the Toad factor Fy: Pd. RoR) + 088 Kp K,) 2 (10) (e-p).4-K, The forces contributing to the numers- tion, see Fig. 13, we ob- tin the following expression for F, Prny Lous + Pres ons an z Lat Pa 8. Seepage ‘To illustrate how seepage pressures may be retaining wall embedded in 2 cohesionless ‘soll with @ water table at a distance below the ground surfacs 't excavation level in front of the ‘As there is @ difference in hydraulic ‘across the wall seepage will occur into the excavation and hydrostatic cond ns will not prevail, The exact nature of the pore pressure distribution in the vic ‘ay of the wall may be determined using fone of several seepage calculation proce: dures, For illustrative purposes it is 2% July, 198135 wr wT tv) Ryn 2 ew e(H=i)8 (@H-h=1) a) = (Kp Kaled-Yy-2eYunds [Mer Kade( HD) + Ka D-I)] (2H be heeYyelhe2eYethei ad ‘Yse(2H-h—j ). ce jerks Fo Len Fe Bars Las® Bigs bag ¥ Bg: Lag + Pes wi Poa twa Fig, 14, Embedded wall with seepage pressures @ (b) effective earch pressures (le) net water pressures ‘sumed that the hydraulic head varies Tinearly down the back and up the front fof the wall which in turn gives the pore pressure distribution shown in Fig. 148". Faking the datum for the hydraulic head Coincident with the water table behind the ‘wall, the hydraulic head at any point round the wall is given by: Heed =—x. (b= i) Yel (2H-h-i) where x is the distance around the wa rom the datum level, Knowing the magni- tudes of the hydraulic and elevation heads the pore pressures may be evaluated, For ‘ample at the tov of the wall x = H=/ land the pore pressure is given by: bur =(H=D ye (Hi) (=i) pol =h= 1) 22. 99: (HA) -6/(QH-h= I) ‘Dore water pressures around the wall Adopting this pore pressure distribution the effective horizontal soil pressure and het water pressure diagrams can be con- Stucted 83 shown in Fes. 140 and 4c. ‘The expression for, is then Pry bes where the forces Pps, Pas etc. are again tvalusted from the shaded part ofthe pres. ‘sure diagram shown in Fig. 146. 9. Embedded cantilever walls inthis seetion the application of the re vised load factor to unpropped cantilever Wells will be described brief. It is come ‘monly assumed thet at limiting equilibrium passive and active pressure istributions Exist down the front and back of the wall fand that a concentrated force acts at the foe as shown in Fig. 15 for 3 wall em- bedded in 8 frictional soi, This concen trated force is assumed to represent the passive resistance acting on the back of fhe wall fora shor distance above the toe. The limiting depth of penetration Ge, can be ‘obtained from the condition wnat Peter, Ly A more rigorous approach is outlines by Tertaghi, 1943 (page 357, 8th Edition te take account of the length over whici Q sets. ‘The revised design method requires that the "moment activated by the retained materisl should not exceed the rection T/A, of the moment of the net available postive resistence (given by the net pas Sive_ pressure coefficient K,—K,). Reler fing to Big. 15bit can be seen that Pry try A= ae Pact +P, Water pressutes, surcharges and co- hesion may be included by following the procedures outlined. previously, Tris of interest to note that Krey (1836) appears to have proposed essentially the Seme approach for the design of embedded Eamulever walls with the sdded refinement ‘of including the length over which Q acts (ihe passive resistance at the heel of th2 wall), Similarly the revised approach: ‘sentially the same as that proposed by Ferzaghi (1943) for the design of embed: ed ghd walls subject to lateral loads 3t for above ground level 10. Concluding remarks Te is important to emphasise that the efnition of fector of safety is a matte: fof convention which varies from one class ‘Of ‘problem to. another. It should, how ever, be both logical and consistent 'ot the full range of conditions likely to bt fencountered. In this Paper the customary Sefnitons of factor of salary with ces pect to passive failure of the toe of 9° ‘embedded retaining wall have been shows fo be inconsistent, illogical and at times Unsafe. A revieed defination has been pre- posed which is based on an analogy with {he bearing capacity of 2 strip load. In tssence the method identifies. the lim ing horizontal forces on the back of * wall that are activated by the presence of the retained material, The maximum Sultant force that the underlying grouné tan offer to resist these activating forces {3 erived from the net passive pressur Coeficients (XK, — Kj) for friction and Kr. “F Kae) for cohesion. The method requires that the moment of the activating forces Should not exceeds factor 1/F, of Moment of the net available passive * Sistence of the underlying ground “The revised. definition leads to res! which are both logical and are consis with the factor of safety on shear sirens!” Tor soils renging from purely fricuona! (© Cohesive. Procedures for including Cfiects of seepage pressure and surche fre presented ‘The main benefit of the revised def" tion of fector of safety is that it leads © more. ational values which are consi Tent with bearing capacity problems 2° tre. free from the inconsistencies" tin the conventional definitons fiso worth noting that in many 97° particularly those for which the cots fe small or negligibie, the revised tion leads to smaller depths of pene! (2s oA (0) ret resisting pressure oo) Fp. 18, Free cantilever wall tion for 2 given value of factor of safety than method 1” (described in Section 2) which is widely used. This may lead to Significant savings particulary in situations where the angle of friction is low and the groundwater pressures are high—2 situation that is ot uncommon for cut- tings and excavations in clay sols. Regarding the magnitude of F, to use we design it would appear trom the work described in this Paper that values of F, between about 1.5 oné 2.0 would normally be sppropriate. These values are less than the load factors normally used in the de- 97 of footings where uncertainty of load ing ie usualy greater and the consequen- ces of excessive settlement are often sev. ‘re, When selecting # value or F, not only should the uncertainty of the sol strength and its variation with depth be considered Bi algo the uncertainty attached to. the ‘imation of seepage pressures and their ‘aluence “both on. the water pressures ‘gainst the wall and the effective stresses which contol the earth pressures, Two. exemples which incorporate the majority of the work discussed in this Paper ate-presented in the Appendix and Serve to illustrate the calculation proced- re in detail In the frst example a propped wall retaining a cohesionless sil with cit- ‘event water levele on either side has been ‘sidered and inthe second 2 short ‘sem analysis for 2 propped wall retaining # clay overain by 8 granular fll and a sure Shrge has ‘bean invesieted_ For Both ‘amples the values of Fy. Fa, and F, have siso been calculated for comparison fupouas, Whereas Fy and Fae fenson- iy consistent it can be seen thet F, Generally conservative end Fy is ery As stated previously, it has not been net activating pressure the purpose of this Paper to debate the felative ments of load factors and strength factors. twill, however, have been noted ‘that the consistency of F, (which is es- Sentialy s load factor) ha’ boon assessed largely ‘by comparing it with F,. It can bbe argued that for this type of problem the use of load factors should be discard. fein favour of factors of safety with re- Spect to strength, Alternatively the use of partial factors on all the major variables ‘may be thought to be more rational, Such 12 debate is outside the scope of this Paper. At present the concept of factor ‘of safety with respect to passive follure Of the toe of an embedded retaining ws is widely used in design and hae the traction that the overall stability can ‘expressed by 8 single factor. The revised Gefnition is logical end consistent for & ‘wide renge of conditions and overcomes the major deficiencies that are inherent in current definitions, E GeomeTRY Fig, 16. Example 1 NETWATER PRESSURE References Toncen). Girt Engraerng Cote ot Procice Ne? (18. tanga, 48 (8 fey 0): Eset, niger Ere, aan Pring Handbook (187), 8 Tones, Baty: Meare" Teaten YMG) Thorn! Sot Mechanic, ing ond arn suchas Hore Mapua, (168) APPENDIX Example 1: Propped wall retaining cohesionless soil (Fig. 16) Soil propenies: 9° = Ws 5° = 92, y, = 196KN/m?, Je = 98iN/m= “Ky ='03, Kp = 48 Net water pressures: Position Net water pressure (kN/m*) Co 267x88-267x267x98/ 1733 = 24 D 533x98-533x267x88/ 1733-26798 = 180 Net loceral effective earth pressures: ACTIVE SIDE Position Earth Pressure (kN/m*) B 03x267%196 D O3%8x 196~ (5.33x98- 533x267 x98/1723)) = 338 PASSIVE SIDE Position Earth pressure (KN/m*) (Passive E 49(267%98 + 6x 196 pressure = (1133x98- 11.33 ae) ¥287%98/1733)) (active 03 (146x196 (1133 Pressure 98~1133x267% se) 88/1733) (active 403 (8x 198~ (533% pressure 98-533%267x98/ sD) 1733)) = 2238 = LATERAL EFFECTIVE EARTH PRESSURE Joly, 198137 Surerorge 8210 KN/me sestaert ELS GEOMETRY var Fig. 17. Example 2 ‘Moments about tie position Moment due to net available passive re- Zone Moment (kNm/m run) Eanh pressure 5 05%2239x bx (446) =87170 Moments setivated by retained material Zone Moment (kNm/m run) enh pressure 1 05% 157x267 x (087x287-2) 157x533 (087 +267) 279 05 x (338-157) x533x (067 + 067x533) = 205 8x8x i (+3) = 182 05 22.1267 x (057 + 067 287) = 2 418x267 x (087 + 15x267) = 724 05 (221-18) 267 (087 S267 +033 x2) 05 x 18x8x (+2) = Earth pressure 2 Eenh pressure 3 Earth pressure 4 Water pressure 6 Water pressure 7 Water pressure 8 Water pressure 9 a2 Total fetained materi For comparison the factors of s esponding to Methods 1, 2 and 4 respec- tively are: Fy = 15 B= AT Paers Example 2: Short-term analysis of a propped wall retaining Clay overlain by a gran- ular fill and a surcharge (Fig. 17) Soil proper () FW: 9) =35",5'=0, re) (027, 97, = 2OKN/m* 38 Ground Enoineerina = ERAL EARTH PRESSURES S8KN/m 9, = IBRN/me (i) Chay se, Net laeral earth pressures: ACTIVE SIDE Position Earth Pressure (KN/m) a tox0m) y=t 27 B 027 (10+ 20 x2) = 135 D 10+ 2x04 3x18-2x35 = 4D 1042x204 3x18 = 5 PASSIVE SIDE Dd 4x38 = 40 (Moments about tie position: 28 Length y= — 6 1.67% Moment due to net avaiable passive, re- Zone Moment (kNem/m run) Jo com 2M, cm cay en aoe en es eaeso | Seta Sd"eimatil Rate Sit Se Se Eee oooh Moments activated by retained material | wane! Ossian = Oe Sh Sin Sac te ts Zone Moment (sXim/m ron) SP bein Mar ee aE ool put axaix@y) = 54 Fil -2 08x(195-27)x oe oen = 2 ome cay 3 osxasxis x a Vosreiey = sus chy 4 Sexsxss | =26tas | naira z, im, “mete pegnin 3 Tout “Post | arn gem M18 oma oe Best uecktnce Ou edits e'™ Factor of ety Fy: pia me fim come moment of at aveloble tg ig il rng Sacsaeresionce 3650 | semen We For comparison the {actors of safety cor Tesponding to Methods 1, 2 and 4 res. & —_— 5 | Soil mechanics and tunnelling | {continued from page 26) sed, and the method and rate of exc Vation. It is recognised that it is often UGitfeult to assess the influence of each of these factors in isolation, but it has been the aim of this Paper 10 show how it's possible to separate some of the me's Tnportent parameters and interpret ie Deheviour of 2 tunnel in soft ground in terms of well-established principles of soi mechanics. With this approach to soft ‘ground tunneling, @ more rational under Standing of tunnel deformation behaviour Ean be Weveloped and a framework provi ed for tunnel design and interpretation of Field data Acknowledgements ‘Much of the work referred to in ts Paper was carried out at Cambridge Uni versity as pan of 8 programme of inves Tigations into the behaviour of tunnels in soft ground, supported by the Transport ‘and Road Research Laboratory, (TARL) fe authors are grateful to Mr. M. P. OrReily, Head of the Tunnels Division at the TRAL, for his continual interest and constructive erticisms. The authors ere itdebted to Professors A. N. Schofield and CP. Weoth for their overall direction > the research work References aera a a henge tsb, a ge de ee eee i a a (108), “The sheng sai? Sa

You might also like