interactional approach of cantilever pile walls analysis STANISLAV Skrabl About the author Stanislav Škrabl University of Maribor, Faculty of Civil Engineering Smetanova ulica 17, 2000 Maribor, Slovenia E-mail: stanislav.skrabl@uni-mb.si Abstract This paper proposes a new method for the geomechanical analysis and design of cantilever retaining structures. It is based on the limit equilibrium method, but it uses some additional conditions for interaction between the retaining structure and the ground, when referring to the distribution of the mobilized earth pressures on the structure. The greatest benefit of the proposed method is shown in the analysis of structures of layered ground (heterogeneous above the dredge level and homogeneous below it), embedded in frictional and cohesive materials, and in the possibility of considering the influence of surcharge loadings on the active or passive side of the retaining structure. When analyzing such cases in practice, the proposed method gives results which are in better agreement with the results of FEM based elasto-plastic interaction analyses than with the results of currently used methods. At the same time, its results are in accordance with those published for homogeneous cohesionless ground. Since in practice almost all retaining structures are erected in layered ground (heterogeneous above the dredge level and homogeneous below it), the proposed method is very convenient and applicable for the analyses and design of cantilever structures under arbitrary geomechanical conditions. Keywords retaining walls, embedment, cantilevers, soil-structure interaction, earth pressure, pressure distribution, wall friction, limit analysis, shear forces 1 INTRODUCTION In geotechnical practice, cantilever embedded retaining structures are specifically used for protecting permanent and temporary excavations, for highway construction, and sanitation of landslides. These structures are mostly sheet walls as temporary retaining structures, and pile walls and diaphragms as permanent retaining structures. Embedded retaining structures sustain overturning moments and horizontal forces, which are caused by backfill soil, ground water and surcharge loading. The contact pressures and resistances are distributed over the embedment depth due to the backfill loading, so that the entire retaining structure remains in equilibrium. The limit state of the retaining structure is achieved when the distribution of contact pressures and extensive regions of plastification in the ground are re-established at the embedded part of the structure, and the structure is no longer capable of taking additional backfill loading. The limit state can be defined by the limit shear loading of the ground or bending moment and shear loading of the retaining structure, respectively. Therefore, only retaining structures, which have a comparable level of safety and reliability for the ultimate limit state of both the soil and the structure, can be optimal. Bica and Clayton (1989) describe several different methods for the geomechanical analysis and design of embedded retaining structures. In these methods, different assumptions of soil pressure distribution, deformations and wall displacements are considered over the embedded part of the retaining structure. Most of the considered methods are limit equilibrium methods. They are based on a classical distribution of the limit soil pressure values. Some model studies of retaining structures are well known, e.g. Rowe (1951), Lyndon and Pearson (1985), Bica and Clayton (1993), as well as some empirical methods for the design of retaining structures. King (1995) proposed an original approach to geome-chanical analysis in homogenous cohesionless soil, which is based on the method of limit equilibrium, ACTA G6OT6CHNICA SLOVENICA, 2006/1 37. s. Škrabl: interactional approach of cantilever pile walls analysis polygonal distribution of soil pressures and one empirically determined parameter. This parameter is the depth, at which the net pressure is vanishing. King defined this depth by centrifuge test results. Day (1999) suggested improving King's method by determining the empirical parameter using the results of FEM analysis. Both methods, King's and Day's, are applicable only to analyses of retaining structures in homogeneous cohesionless ground without additional surcharge loading. This paper firstly presents a generalization of the existing standard methods for the analyses of rigid retaining structures embedded in frictional and cohesive layered ground (heterogeneous above the dredge level and homogeneous below it) with surcharge loading, and then compares the obtained results with the results of FEM-based analyses. Finally, it proposes a new method that considers significant interactive conditions between the retaining structure and the ground. 2 GENERALIZATION OF-CURRENT USED METHODS The aim of the analyses based on the method of limit equilibrium is to determine the critical excavation depth and the critical values of bending moment and shear force at which the limit equilibrium state can be reached. The first important step of this approach is to determine the soil pressure, which can be activated at the ultimate limit state of the ground. The soil pressures acting on the retaining structure at the ultimate limit state, depends on the complex properties of the interaction between the retaining structure and the ground. The most important interaction parameter represents equal displacements of both mediums along the contact surface between the retaining structure and the ground. As a rule, the retaining structures are designed by the limit values of soil pressures which cannot be activated over the whole region of interaction due to the condition of both mediums' equal displacements. The general shape of the mobilized earth pressures acting at the ultimate limit state of the homogeneous cohesionless ground is presented in Fig. 1. All standard methods are based on earth pressures at limit equilibrium. Methods differ only in the assumptions and simplifications used in the determination of these pressures. A generalization of the existing methods to the case of a layered strata is presented in this section. Layered strata means multilayer above the dredge level and a single layer below it. Figure 1. Activated Influences and Resistances: (a) Soil Pressure Distribution; (b) Net Soil Pressure Distribution 2.1 UK SIMPLIFIED METHODge This method, mostly used in Europe, is the simplest one. It is described in several publications (Padfield and Mair 1984, King 1995), and represents the basis of many computer codes for designing embedded retaining structures. The activation of all active pressures and passive resistances in this method is assumed, above the calculating rotation point 1 (Fig. 2). The resistances under the rotation point are not determined exactly; they are substituted with the fictitious concentrated force R. Figure 2. Influences and Resistances According to the UK Simplified, and UK Full Method By considering the equilibrium condition of the moments around the rotation point 1, we obtain: d*3(K-1) +1.5pX2 -3EV + d*) = 0 (1) where the normalized quantities are: d\ = dj h, K = KPy / Kai, E' = 2E/(Yh2Kai), c* = 2c/(7h) p0 = c* (Kc + Kpc )/Kai + P'Kpq / Kai - 2, q* = 2q/(jh), a* = ajh 58. ACTA GeOTeCHNICA SLOVeNICA, 2006/1 s. Skrabl: interactional approach of cantilever pile walls analysis K , K„ , and K , K„ , and K„ denote the coefficients ay > py > ac > pc > pq of active and passive pressures for the influences of soil weight, cohesion and surcharge (Kérisel and Absi 1990). The letters y, c and q in indices denote the unit weight, cohesion and surcharge loading of the soil layer in which the analyzed retaining structure is embedded. The required embedment depth d1 up to point 1 is determined by solving (1). The analytical solution (3) can be applied only for a retaining structure in homogeneous cohesionless soil without surface or other additional loading. d = 1/(3K -1) (3) The total required embedment depth d = d1 + d2 is approximately: d2 = ((K (y * + d* ) - d* + c* (Kpc + Kac )/2Kal )2/(K -1)2 + ((K -1)d*2 + pX - E')/(K -1))1/2 (9) - (K (y ' + d* ) - d* + c* (Kfc + Kac )/2Ka_, )/(K -1) E (3a + 3d* + 2d2) - p0d*(1.5d* + 2d2) + K y d22 -(K-1)(d;3 + 2d*2 -d;d2*2) + 0.5c*(Kpc + Kc)d?lKv = 0 (10) d2 = d2l h The required embedment depth d = d1 + d2 is calculated by solving the system of equations (9) and (10). The maximum values of bending moment and shear force acting on the retaining structure are determined using (5), (6) and (7). d = l.2d*, d = d/h (4) The maximum values of bending moment and shear force acting on the retaining structure are determined: M = M' YhhK /2, T = T'yH2K /2 (5) m m ' aY m m ' aY v/ where M* and T* , are the normalized values of the m m' maximum inner forces determined by using: Ml = e '(a' + xm ) - p0 x'm/2 - (K - l)x2 / 3 t: = E' - pd - (K -lK2 (6) where x*m denotes the distance between the point where the maximum bending moment acts and the dredge line: (P + 4(K-1)Et2 - p0 2(K -1) (7) 2.2 UK FULL MeTHOD66 UK full method additionally considers the influence of limit pressures under rotation point 1 (Fig. 2). The ground resistance at the toe of the embedded wall is determined: pb= 2pb /(jhKai) = 2Ky* + 2(K - l)d*+ c'(Kpc - Kac)/Kai (8) where y denotes the ratio of the average unit weight of the backfill ground above the dredge line for the considered problem to the unit weight of the ground under the dredge line. After considering the equilibrium conditions of the horizontal forces and moments we obtain the following expressions: 2-3 USA MeTHOD66 The USA method, which was first introduced by Bowles (1988) for homogeneous soil, supposes that the entire passive resistance is mobilized at the toe of the retaining structure. In the region between depths d1 and d (Fig. 3), the resulting resistances have the form of a straight line (polygonal net pressure distribution). q 111111,11 Figure 3. Influences and Resistances According to the USA, King, and Day Methods The individual earth pressures and available passive pressures above the calculated embedment depth can be estimated for arbitrarily layered ground. The unknown embedment depth d in the soil at the toe of the pile wall is determined by using the equilibrium of horizontal forces (11), while the depth d} is determined by the moment equilibrium condition (12). E* + (y"K + (K - \)d' + 0.5c*((Kpc + Kac)/Kai)(d' - d*) -0.5p0(d* + d*) -(K -l)d*< = 0 (11) d* = (E' (3a* + 2d' ) - 0.5p0dn)l((K-1)d*2 - E' + p0d' ) (12) ACTA GeOTeCHNICA SLOVeNICA, 2006/1 55. s. Škrabl: interactional approach of cantilever pile walls analysis 2.4 KINGS METHOD66 On the basis of centrifuge model tests, King (1995) established that not all available resistances at the toe of an embedded pile without anchors could be fully activated in the limit state. He, therefore, proposed modifying the USA method by considering a polygonal form of the active and passive pressures, and determining experimentally the value of parameter e = 0.35. This parameter defines the depth at which the activated contact-stresses in front of and behind the embedded pile are balanced. King only gave solutions for homogeneous cohesionless ground. The critical embedment depth d for layered ground can be determined using (13). E * (d * (1 - e) - d*) - 0.5 p (d *2 (1 - 2e) - d*2) -(K - 1)dd (d* (1 - 2e) - d[ (1 - e)) = 0 (13) The depth dj^ is determined by the moment equilibrium condition (12). 2.5 DAY S METHOD66 Day (1999) found that King's method gave too conservative results, particularly for lower values of parameter K. If K is lower than 7.90, the solution does not practically exist. He therefore proposed, on the basis of the interaction analysis results using FEM, introducing new value of parameter e. e = 0.047ln(K) + 0.1 (14) In Day's method the unknown embedment depth is also determined using (12) and (13). 3 FINITE ELEMENT ANALYSESS A set of two-dimensional plane strain analyses for the ultimate limit state of embedded retaining structures was performed to confirm the results of the described analytical methods. The critical excavation depth was determined by progressive removal of the excavated soils until the horizontal displacement of the wall was so high that further re-establishment of the retaining structure's global equilibrium could not occur anymore. A relatively stiff elastic concrete retaining structure of the total length h + d = 10.0 m was considered in the analyses. Its characteristics were E = 31106 kPa, I = 0.083 m4/m and A = 1 m2/m. The cross-section of the analyzed excavation is presented in Fig. 4. h + d = 10 m (p, c* = 0 y = 20 kN/m3 p* = 0 or 0.2 111111111111111111 (p, c* = 0 or 0.2 y = 20 kN/m3 Figure 4. Cross-Section of the Analyzed Excavation The elasto-plastic Hardening-Soil model with isotropic hardening (PLAXIS 1998) was used. The advantage of the Hardening-Soil model over the Mohr-Coulomb model is in the use of a hyperbolic stress-strain curve, and the stress dependency on soil stiffness. The Hardening-Soil model uses the theory of plasticity, includes dilatancy and introduces a cap yield surface. The following properties were considered: Ejef=E ,ref = 40 MPa, E ref= 3Ejef, 50 oed ' ur 50 ' vur= 0.2, m = 0.5, Rf = 0.9, pref = yh, where h denotes the final excavation depth in the limit state. The analyzed region was discretized in 2.000 six-node triangle finite elements with a refined mesh near the retaining structure. The finite element mesh with the deformed mesh of the analyzed excavation is presented in Fig. 5. Figure 5. The Cutting Out of the Finite Element Mesh with the Deformed Mesh (scaled 10 times) of the Analyzed Excavation The numerical analyses were performed for friction angles f =10° to q> =50° in steps of 5°, normalized cohesions c*=0 and 0.2, and normalized surcharges p*=0 and 0.2. All analyses considered the friction between the retaining structure and the ground S = f, dilatation f = f/2 and unit weight of soil y = 20 kN/m3. 58. ACTA GeOTeCHNICA SLOVeNICA, 2006/1 s. Skrabl: interactional approach of cantilever pile walls analysis The ultimate limit state, reflected in the critical excavation depth hc, was considered as a state when no equilibrium could be ensured anymore or when numerical results could not converge. The results of analyses have shown that the deformations obtained for the used elasto-plastic soil model for a rigid retaining structure are directly proportional to the square of the excavation, and inversely proportional to the deformation modulus E50 . The displacements of the retaining structure can be determined using (15). u = u y h2 / E50 (15) where u is the displacement of the actual structure and u* the displacement determined on a generalized system with normalized quantities h* = h / h = 1, d* = d / h, t* = t / h, y* = 1, c*, q* and p*, where t is the thickness of the retaining wall. The actual values of maximum bending moments and shear forces can be determined on the basis of their normalized values using (6). Variation in the normalized values of the horizontal displacements at the top of the retaining structure u*o with arbitrary excavation depths h* for p* = q* = c* = 0 (see Fig. 4) are shown in Fig. 6a. Variation in the normalized values of displacements u*o with arbitrary excavation depths h* for p*=q*=0 and c*=0.2 are presented in Fig. 6b, and for q*=c*=0 and p*=0.2 in Fig. 6c. The results of numerical simulations have shown that at higher friction angles the bearing capacity limit state occurs at smaller excavation depths than those determined by classic methods (UK simplified, UK full and USA). King (1995) and Day (1999) drew similar conclusions. Only in the case where p*= q*= c*= 0 are the results of FE analyses comparable with the results of King's and Day's solutions. The differences are considerable, especially for the required embedment depth (ultimate limit state of the ground), and for the maximum internal shear forces of the retaining structures (section strength limit state of the retaining structures). The deviations are observed for retaining structures in homogeneous and layered ground, as well as for surcharges. In geotechnical practice, retaining structures are most frequently embedded in soils of relatively high shear strength. Therefore, the choice of method for geome-chanical analysis is often decisive for safety, as well as for the economy of the designed retaining structure. For this reason, the next section presents a new method which yields more reliable results. It makes it possible to design retaining structures of comparable safety for the ultimate limit state of the ground, as well as for the section strength limit states of the retaining structures. a) 6 (h/d)7 10 11 b) (Wd) c) Proposed Day°E (1999) FE - Plaxis USAGE UK simplified"* UK fullGE KingSE (1995) 1 1 (18) After solving (17) and (18) we obtain: C = J(KP1 - Kai M (19) c 2 =-p0 - - Kai )dl (20) The boundary values of (18) that can only appear at the embedment in cohesive ground or when surcharge loading is in front of the retaining structure are: dp(z) dz z=0 n=1 C C2nzn-1 +——— d dn = - P0 (21) d dp(z) dz z=0 n<1 Cj + C2nz' d1 d1n (22) z=0 n<1 Figure 7. Influences and Activated Resistances at Limit State It has to be noted that the infinite slope boundary condition in (22) can only appear for higher cohesion or surcharge loading with a simultaneously high friction angle. The distribution of the mobilized earth pressures in the whole region of the embedment depth d1 follows the exponential function given below: p (z) = iM = + 2(K - Dd* (Z) - (p0 + 2(K - 1H)(4-y (23) lhKal di di The resulting horizontal force and moment around point 0 of the mobilized earth pressures are given in (24) and (25). = 58. ACTA GeOTeCHNICA SLOVeNICA, 2006/1 s. Skrabl: interactional approach of cantilever pile walls analysis R = M(-7) + (K-1M2(--) (24) n +1 n + 1 m:=px2( 2(n + 2) ) + 2(K —1)<3( n — l 3(n + 2) ) (25) In the region of embedment depth d2 the following distribution of earth pressures is considered: z — d p(z) = pb (^ )m (26) d2 where pb denotes the horizontal pressure on the wall at the toe of the retaining structure. The resulting horizontal force of the pressures in the region of depth d2 is given in (27), and the distance of the resulting horizontal force from point 1 in (28). d. R = Pb(^TT) (27) m + 1 , m + 1. ^2 = d2(—(28) 2 m + 2 Different values of parameters p*0, p*b , (K-1), m and n define a family of exponential functions that agrees with the experimental distribution of pressures along the embedment depths of the rigid retaining structures. The values of parameters m and p*b are determined from the results of numerical simulations of the limit states of retaining structures embedded in cohesionless and layered ground, as well as for the case of surcharge loading in front of the retaining structure. Parameter m is based on numerical interpolation using results of elasto plastic analyses using FEM and can be expressed as: m = 2 + ln(K )/4 (29) The comparative value of parameter p*b is determined experimentally for individual numerical solutions at the ultimate limit state: 2P0CH - c*=0.2, p*=q*=0 --c*=q*=0, p*=0.2 ---c*=p*=q*=0 Proposed ✓ □ c*=0.2, p*=q*=0 a c*=q*=0, p*=0.2 o c*=p*=q*=0 FE-Plaxis 100- T 20 200 K Figure 8. Numerical and Analytical Values of Parameter p*b In the analysis of retaining structures we can consider the following normalized values of pressures at the toe of the pile wall in a wider region of the value K (Fig. 8): pb =(1 + d' + c* )(ln K + 3.8 ln2 K) (31) Equation (31) is the approximation of the results using FEM determined by method of the least square. The mobilized pressures in front of and behind the retaining structure (defined by parameters n and m) are linked with (32). The relationship is determined by the slope of the tangent (line 2-3, see Fig. 7) to the net pressure function in front of the retaining structure through point 1. n = p*d* /m + 2(K — l)d*d*2 p'0d'2 + 2(K — l)d*d* (32) The embedment depths of the retaining structure d*1 and d*2 are determined considering the moment equilibrium around point 1 (33), and the horizontal forces' equilibrium condition (34). E*(a + d'*) - p^d'*2 ( n + 1 2(n + 2) iv n-1 2(n-1) , -(K — 1)d* (-- -—) - pb d22 - = 0 (33) n +1 3(n + 2) m + 2 . K\(m+1) Pb - (30) a2 E' - pd ) - (K -l)d'l2 (+ ptd"2 /(m + 1) = 0 (34) n +1 n +1 where T* denotes the maximum normalized shear m force acting on the retaining structure at limit state determined using FEM. The numerically (using FEM) and analytically (using Eq. 31) determined values of parameter p*b are presented in Fig. 8. The required embedment depth d* and the parameter n, representing the stage of pressure mobilization along the embedment depth d*1, are determined for each considered case separately by solving a system of three nonlinear equations (32), (33) and (34). The maximum bending moment and shear force acting on the embed- n n n ACTA GeOTeCHNICA SLOVeNICA, 2006/1 55. s. Škrabl: interactional approach of cantilever pile walls analysis ded retaining structure at limit state are determined using (35) and (36). ,», n s ^ ,»2/n — E Tm = E — p0d»(--) — (K — l)d12(--) (35) n +1 n +1 1 x" M= E'(a'+x'm) - p0x2(- — Xm m v ' m' 10 m ^ ~ 2 d\" (n2+3"+2) )- 1 2x ,*3 / 1 m (36) (K-M---¡"-rlf. 3 d"-1(" + 3" + 2) The distance of point xm, where the maximum bending moment acts, from the toe of the excavation is determined using (37). Numerical procedure i) We determine K = Kpy / Kay■ ; E* = 2E/(yh2Kay); a* = a/h,p*0 = 2pJ(yhKay), m = 2 + \n(K)/4 ii) Using Microsoft Solver real minimum solution of dj* + d2* of nonlinear equations (31) - (34) with constraints (dl* > 0 and + d2* > 0). iii) For obtained values dl* and d2* the position of maximum bending moment xm* (Eq 37) and values of maximum shear force Tm* (Eq 35) and maximum bending moment Mm* (Eq 36) are determined. For case c* = q* = p* = 0 of geomechanical analysis for shear angles f = 15°, 20°, 25°, 30°, 35°, 40°, 45°, and 50° are given in Table 1. x" 2x"-1 E' - P*x'J1 ) - (K - 1)xm(1 ) = 0 (37) d" (n+i) d"-l(" + 1 ) Table 1. The Results of Analyses for S/f = 1, for c* = q* = p* = 0 9 [°] Pb* m n d* d** M * m T * m (1) (2) (3) (4) (5) (6) (7) (8) 15 26.012 2.355 8.713 1.5630 0.2957 1.2892 2.293 20 36.708 2.488 8.252 1.0273 0.2060 0.8586 2.168 25 48.997 2.621 7.243 0.7302 0.1592 0.6594 2.155 30 63.926 2.762 5.996 0.5359 0.1297 0.5452 2.203 35 83.642 2.925 4.656 0.3942 0.1082 0.4703 2.306 40 108.645 3.102 3.472 0.2953 0.0927 0.4226 2.454 45 142.887 3.311 2.489 0.2224 0.0802 0.3905 2.657 50 191.143 3.559 1.774 0.1705 0.0697 0.3697 2.924 Table 2. The Results of Analyses for Different Combinations of h and p values Method UK Simpl. UK Full USA King Day FE Proposal (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (a) h=10.0 m, p=0 Total wall height_h+d (m)_11.77 11.52 11.53 12.01 12.32 12.44 12.43 Max. shear force T m (kN/m) 910.2 854.0 766.5 389.9 310.0 302.1 297.3 Max. bending moment M m (kNm/m) 373.3 373.3 373.8 373.3 373.3 379.3 377.9 (b) h=10.5 m, p=10 kPa Total wall height h+d (m) 11.85 11.67 11.68 11.92 12.03 12.56 12.52 Max. shear force T m (kN/m) 1011.7 945.5 860.5 559.2 491.3 366.0 317.3 Max. bending moment M m (kNm/m) 404.7 404.7 404.7 404.7 404.7 410.6 406.2 (c) h=11.0 m, p=20 kPa Total wall height_h+d (m)_12.14 12.01 12.02 12.17 12.24 12.92 12.89 Max. shear force T m (kN/m) 1205.4 1115.7 958.1 735.9 664.3 382.2 351.3 Max. bending moment M (kNm/m) 455.7 455.7 455.7 455.7 455.7 460.1 456.5 Note: All methods in columns (4) to (8) are generalized and extended (GE) 58. ACTA GeOTeCHNICA SLOVeNICA, 2006/1 s. Skrabl: interactional approach of cantilever pile walls analysis 5 VALIDATION OF THE METHOD A comprehensive set of results has been calculated for comparisons in order to validate the performance of the proposed method. Four cases are presented. In the first case, a cantilever reinforced concrete retaining structure of a certain width, embedded in homogeneous cohesion-less ground of certain properties at a certain soil-wall friction, was analyzed, where the height of the retaining structure and surcharge loading had been altered. The remaining three cases concerned a cantilever embedded retaining structure of certain height, where analyses were performed for different combinations of normalized c* and p*. The variation of (h/d), M*m and T*m with K is given in the form of diagrams. 5.1 CASE 1 A cantilever embedded reinforced concrete retaining structure (1.0 m in width) for the protection of a construction pit in homogeneous cohesionless ground was considered. The following ground properties were considered: f = 49.5°, c = 0, 8 = 20 kN/m3, and friction between soil and wall 8 = f. Three examples were analyzed: (a) h = 10.0m, p = 0, (b) h = 10.5m, p = 10 kPa (c) h = 11.0m, p = 20 kPa. The results of analyses are presented in Table 2. The results of analyses for example (a) show, that in ideal homogeneous cohesionless ground the classical analytical methods (UK simplified, UK full and USA) give similar results. They differ from the solutions of more recent methods (King, Day, FE and Proposal) in required embedment depths and maximum internal shear forces. The differences in embedment depths are up to 61%, for maximum internal shear forces up to 206%, and for bending moments up to 1.6% with regard to the smallest values. In examples (b) and (c) the weight of the ground at the toe of excavation was substituted with the pertinent surcharge loading (soil without shear strength). The differences with the smallest values for embedment depths are up to 76% in example (b) and grow up to 90% in example (c), for maximum internal shear forces up to 219% and 243%, and for bending moments up to 1.5% and 1%. In these cases King's and Day's method exhibit unacceptable solutions, because the required embedment depth is smaller than that of a ground with very high pressure resistance at the dredge level. For such cases, the results of analyses obtained by the two newer methods are comparable with the results obtained using classical methods. It can be concluded from the above examples, that all the presented methods give practically the same values of bending moments. This is due to the fact that bending moments depend mostly on wall height. When the surcharge loading in the excavation pit at the dredge line level is considered, only the results of the proposed method are comparable with the results of FEM elasto-plastic analyses. 5.2 cose 2 The limit states of a rigid cantilever reinforced concrete retaining structure embedded in homogeneous cohe-sionless ground were considered. The numerical analyses were done for ground with friction angles from f = 10° to f = 50° in steps of 2.5°, cohesion c = 0, and without surcharge loading. The analyses were performed for all considered methods. In all analyses the rigid retaining structure of total height h + d = 10.0 m, friction between soil and wall S = f, dilatation angle f = f/2, and unit weight y = 20 kN/m3 were considered. Fig. 9 shows the distribution of horizontal soil pressures acting on the retaining structure in homogeneous ground for f = 20, 35 and 50°. The solution according to FEM is presented in the same way as the results published by Day (1999). The excavation depths for each individual method are presented in the same diagram. It can be concluded from the net pressure distributions in Fig. 9, that the pressure distribution given by the proposed method agrees very well with the calculated net pressure points using FEM (Day 1999). Fig. 10a presents the variation of the required normalized values of embedment depth d* (d* = d/h), maximum bending moments M*m, and maximum shear forces T*m with parameter K. It is evident from Fig. 10a that the results of the proposed method are in agreement with the results of FE analyses and the solutions of Day (1999). The latest solutions show larger deviations at higher values of K ratios, which do not appear in practice. However, these deviations arise in further analyses for layered ground and when surcharge loading is presented. ACTA GeOTeCHNICA SLOVeNICA, 2006/1 55. s. Skrabl: interactional approach of cantilever pile walls analysis .6- gj > -600 4- Pressure [kPa] 1000 1400 D) a) X : b) : 9 = 350 ......./A 1 * (f> M' Jt\ A ^ 1 1 —1—1 *"* 1 1 8- 10 -600 4- 1000 1400 Pressure [kPa] „ 6E cp = 50 ° Proposed C) ----Day^ (1999) ♦ FE-Day (1999) ........................ USAGfc _ -UK simplified*^ f ------UK fullGE KingGE (1995) -600 1-1-1-H»— 1000 1400 Pressure [kPa] - Proposed - KingGE (1995) - Day** (1999) - UK simplified** - UK fullGE - USAGE - M* for all methods FE-Day (1999) □ Centrifuge-King (1995) 0 FE - Plaxis 2000 1000 1000 Figure 9. Net Pressure Distribution: (a) f = 20° (b) f = 35° (c) f = 50° Figure 10. Variation of (h/d), M*m and T*m with Kfor: a) c* = q* = p* = 0 b) c* = q* = 0 and p* = 0.2 c) P* = q* = 0 and c* = 0.2 56. ACTA G£OT£CHNICfl SLOVeNICñ, 2006/1 s. Skrabl: interactional approach of cantilever pile walls analysis It is evident from Fig. 10a that the results of the proposed method are in agreement with the results of FE analyses and the solutions of Day (1999). The latest solutions show larger deviations at higher values of K ratios, which do not appear in practice. However, these deviations arise in further analyses for layered ground and when surcharge loading is presented. 5-3 cose 3 The limit states of the rigid cantilever reinforced concrete retaining structure embedded in homogeneous cohesionless ground were considered again. The influence of surcharge loading in front of the retaining structure at the toe of the construction pit was considered, too. In geotechnical practice such loading represents the influences of soil layers with low pressure resistance, which are located above the foreseen embedment depth of the retaining structure. Fig. 10b presents the variation of the required normalized embedment values of the depth d* (d* = d/h), maximum bending moments M*m, and maximum shear forces T*m with parameter K. It is evident from Fig. 10b that only the results of the proposed method agree with the numerical results of FEM analyses. 5.4 CASE 4 The limit states of the rigid cantilever retaining structure in homogeneous ground were considered further, where soils under the dredge level have a part of the cohesion shear strength of the upper layer, c* = 0.2. Fig. 10c presents the variation of the required embedment depth normalized values d* (d* = d/h), maximum bending moments M* , and maximum shear forces T* Cm m with parameter K. It is again evident from Fig. 10c that only the results of the proposed method agree with the numerical results of FE analyses. ered. Furthermore, the largest differences are obtained in the region of the wall's fictitious rotation point at embedment depth, and at the toe of the embedded retaining structure. This paper presents the equations of the existing methods in generalized form, which enables analyses of retaining structures considering layered ground and surcharge loading. Based on the performed analyses it can be established that the results of the proposed method agree very well with the results of finite element elastoplastic analysis and with the results of centrifuge tests. All the existing methods for cantilever pile wall analysis, as well as the proposed one, give nearly the same values for maximum bending moments. Furthermore, only the proposed method yields a reliable embedment depth and maximum internal shear force and does not underestimate embedment depths and overestimate internal shear forces, as do the existing methods. The special benefit of this method is due to the realistic values for internal shear forces, which are, as a rule, strongly overestimated in the remaining methods. This benefit leads to a more economical design of pile walls, since larger pile spacing is possible. Consequently, the total bearing capacity of the reinforced concrete section can be exploited and a comparable safety of ground bearing capacity and strength of structure sections can be reached. In practice, almost all retaining structures are erected in layered ground; the existing methods only give exact solutions for retaining structures in cohesionless ground. Therefore, this proposed method is more convenient and more applicable in the analysis and design of cantilever structures in geotechnical practice. 6 CONCLUSIONS The results of analyses show that, due to the interaction between rigid cantilever retaining structures and the ground, the entire passive resistance is not actually mobilized, as is considered in classical methods for retaining structure analyses (UK simplified, UK full and USA). The greatest differences are exhibited for structures embedded in soils with high shear strength when higher friction between the structure and the ground is consid- APPENDIX I. R6F6R6NC6S Bica, A. V. D., and Clayton, C. I. R. (1989). Limit equilibrium design methods for free embedded cantilever walls in granular materials. Proc. Instn Civ. Engrs Part 1, 879-989. Bica, A. V. D. and Clayton, C. I. R. (1993). The preliminary design of free embedded cantilever walls in granular soil. Retaining structures, C. R. I. Clayton, ed., Thomas Telford, London, England, 731-740. acta geotechnica slovenica, 2006/1 57. s. Škrabl: interactional approach of cantilever pile walls analysis Bowles, J. E. (1988). Foundation analysis and design, 4th Ed., McGraw-Hill Book Co., New York, N.Y., 613-618. Day, R. A. (1999). Net pressure analysis of cantilever sheet pile walls. Geotechnique, London, England, 49(2), 231-245. Fang, Y. S., Chen, T. J., and Wu, B. F. (1994). Passive earth pressures with various wall movements. J. Geotech. EngngDiv., ASCE, 120(8), 1307-1323. Kerisel, J., and Absi, E. (1990). Tables for the calculation of passive pressure, active pressure and bearing capacity of foundations. Gauthier-Villard, Paris, France. King, G. J. W. (1995). Analysis of cantilever sheet-pile walls in cohesionless soil. J. Geotech. Engng Div., ASCE, 121(9), 629-635. Lyndon, A., and Pearson, R. A. (1985). Pressure distribution on a rigid retaining wall in cohesionless material. Proc., Int. Symp. on Application of Centrifuge Modeling to Geotech. Design, W. H. Craig, ed., A. A. Balkema, Rotterdam, The Netherlands, 271-280. Padfield, C. J., and Mair, R. J. (1984). Design of retaining walls embedded in stiff clays. CIRIA Report 104, London, Constr. Industry Res. and Information Assoc. (CIRIA), London, England. PLAXIS user's manual, version 7 (1998). R. B. J. Brinkgreve, and P. A. Vermeer, eds., A. A. Balkema, Rotterdam, The Netherlands Rowe, P. W. (1951). Cantilever sheet piling in cohesion-less soil. Engrg., London, England , (Sept.), 316-319. APPENDIX II. NOTATION The following symbols are used in this paper: a = distance of the resulting force of active pressures from the dredge line (see Figs. 2 and 3); a* = normalized quantity defined as ratio a/h; c* = normalized quantity defined as ratio 2c/(yh); C1, C2 = integration constants (see Eq. 16); d d1, d2 d*, d1*, d2* E E* E ref 50 EoJef E ref K Kpc, Kac Kpy Kay K ~pq m Mm M*m p0 p*0 pb p'b pref total embedment depth (see Figs. 2 and 3); partial embedment depths (see Figs. 2 and 3); normalized quantity defined as ratios d/h, djh and d2/h; resulting force of active pressures above dredge line (see Figs. 2 and 3); normalized quantity defined as ratio 2E/(yh2Kay); secant stiffness in standard drained triaxial test at the reference pressure; tangent stiffness for primary oedom-eter loading at the reference pressure; unloading / reloading stiffness in Hardening-Soil model (considered Ejef = 3 E50ref ); height of the retaining structure (see Figs. 2 and 3); critical (allowable) excavation depth; ratio between K and K ; py ay ' coefficient of passive and active earth pressure for influence of cohesion; coefficients of passive and active earth pressure for influence of soil weight; coefficient of passive earth pressure for influence of surcharge; parameter in elasto-plastic Hardening-Soil model (it is the input parameter in the relationship for stress dependent stiffness according to a power law) and parameter that defines family of exponential functions in proposed method; parameter that defines family of exponential functions in proposed method; maximum value of bending moment; normalized value of maximum bending moment; pressure on the pile wall at the dredge line (see Figs. 2 and 3); normalized quantity defined as ratio 2Po/(yhKa) net pressure at the toe of the pile wall (see Figs. 2 and 3); normalized quantity defined as ratio 2pb/(yhKa) reference pressure in elasto-plastic Hardening-Soil model; h h n 58. ACTA GeOTeCHNICA SLOVeNICA, 2006/1 S. SkRABL: INTERACTIONAL APPROACH OF CANTILEVER PILE WALLS ANALYSIS Rf = parameter in elasto-plastic Hardening-Soil model that defines failure ratio q/qa (ultimate deviatoric stress / asymptotic value of the shear strength), is derived from the MohrCoulomb failure criterion; R*i = normalized resulted horizontal force in region of depth di; R*2 = normalized resulted horizontal force in region of depth d2; M*\* = normalized resulted moment of acti- R i vated resistances in region of depth di; q* = normalized quantity defined as ratio 2q/(yh); Tm = maximum value of shear force; T*m = normalized value of maximum shear force; t = thickness of the retaining structure; t* = normalized quantity defined as ratio t/h; u = horizontal displacement; u* = normalized horizontal displacement; u*0 = normalized horizontal displacement at the top of the retaining structure; distance of the point where maximum bending moment acts measured from the bottom of excavation; independent variable that denotes depth; distance of the resulted horizontal force; Y = unit weight of the soil; Y* = ratio between average unit weight of the backfill ground above dredge line and the unit weight of the ground under dredge line; S = friction angle at the soil-structure interface; e = experimentally determined parameter that defines depth at which activated contact stresses in front of and behind embedded pile are balanced wall (see Fig. 3); vur = Poisson's ratio for unloading-reloading in Hardening-Soil model (considered vur = 0.2); f = angle of internal friction of the soil; and f = angle of dilatation of the soil. = z r*2 ACTA G6OT6CHNICA SLOVENICA, 2006/1 37.