Three-dimensional finite element analysis of braced excavations in clays
Date of Issue2016
School of Civil and Environmental Engineering
Braced excavations are commonly used for constructing underground infrastructures in built-up metropolitan areas such as Singapore. Current design of braced excavations in clay soils relies mainly on two-dimensional plane strain finite element analysis. However, due to the inability to take account of the effects of the length of the excavation and the associated restraints from the secondary walls at the two ends of the excavation, the plane strain analysis may not be reliable to predict the excavation behaviour accurately. This thesis presents some practical guidelines and approaches for the design of braced excavations in clays by carrying out three-dimensional finite element analysis with non-linear constitutive soil models. In addition, some comparisons are carried out to assess the accuracy of conventional design methods and charts with the results from two-dimensional and three-dimensional finite element analyses. The total stress Mohr-Coulomb constitutive soil model was used to investigate the basal heave stability for braced excavations in clay. The factor of safety against basal heave was determined using the shear strength reduction technique. The results generally show that the factor of safety from three-dimensional analysis is larger than that from plane strain analysis, due to the restraining effects from the secondary walls and bracing system. In addition, the wall system stiffness is found to influence the factor of safety. The increase of the excavation length to width ratio (L/B ratio) also leads to a reduction of the factor of safety. The factor of safety determined using the conventional modified Terzaghi’s method which accounts for the wall embedment depth D agrees well with the results from the plane strain analyses. Based on the results of this parametric study, a simplified empirical equation is developed for determining the FS that takes these 3D effects into consideration. Plane strain and three-dimensional finite element analyses using the non-linear hardening soil model were also carried out to investigate the wall deflection behaviour, the wall bending moments and the strut forces. The results indicate that the increase of L/B ratio leads to the increase of the maximum horizontal wall deflection when L/B ratio is smaller than 3.4, and the increase of D/B ratio helps to restrain the maximum horizontal wall deflection. By incorporating the L/B and D/B ratios into a relative stiffness ratio, a semi-empirical closed-form equation is proposed to estimate the maximum horizontal wall deflection in clays. For the bracing system, the strut forces and apparent earth pressures are examined. The results show the rate of increase of the strut force diminishes after the installation of the next level of struts, so that the lowest strut carries the maximum load before the next level of struts is installed. As Peck’s apparent pressure diagram (APD) is found to under-predict the strut forces, a modified APD has been proposed in this study. In the design of braced excavation systems in Singapore, one of the requirements by the building authorities is to perform one-strut failure analyses, in order to ensure that there is no progressive collapse when one strut is damaged due to a construction accident. This evaluation is normally carried out using plane strain analysis. The three-dimensional one-strut failure analysis carried out in this study shows that the load of the failed strut is transferred horizontally, vertically and diagonally across to the adjacent struts. Consequently, the plane strain analyses would result in fairly conservative (i.e., larger) estimates of the loads transferred to the adjacent struts from the failed strut.