2D and 3D Finite Element Analysis for Cantilever Slurry Walls in Cohesionless Soil

Support of excavation (SOE) design practice tends to model SOE elements using two-dimensional (2D) analyses to avoid the complexity associated with three-dimensional (3D) analyses. Such practice overlooks soil arching in the horizontal plane and prevents designers from accounting for elements such as capping beams as part of the statical system. This study simulates square excavations supported by cantilever slurry walls in dry cohesionless soil using 2D and 3D finite element (FE) models. First, side support wall behavior is compared from 2D and 3D FE models for variable excavation length to depth (L/H) ratios. Second, the effect of slurry wall lateral bending stiffness on wall behavior is studied by modeling the wall using plate element with isotropic and anisotropic wall stiffness. Finally, the sensitivity of cantilever wall behavior to the presence of capping beam is illustrated. The study concludes that unbraced excavations with L/H ratios greater than 5 almost resemble plane-strain conditions. Slurry walls are better simulated in 3D FE models by dividing the plate element into discrete panels or by assigning anisotropic bending stiffness to the plate element. Finally, the presence of capping beam affects the behavior of cantilever walls resulting in reduction of deflections and redistribution of bending moments. INTRODCUTION Support of excavation (SOE) design practice ranges in complexity from conducting simple hand calculations to performing three dimensional numerical modeling depending on several factors among which is the geometry of the design problem. This paper addresses walls that are designed using numerical modeling, and discusses the factors governing the choice of the most suitable modeling approach. It has been believed for a while that modeling excavations using 2D numerical models would lead to a more conservative solution, which turned out to be incorrect in specific cases. This paper attempts to describe cases where 2D finite element modeling could be on the conservative side and other cases where it could be less conservative than the more complex 3D modeling.