Soil–pile separation effect on the performance of a pile group under static and dynamic lateral loads

Abstract

The effect of soil–pile separation is studied with respect to the performance of a laterally loaded pile group. Full-scale tests, which consist of a combination of a single and a 3 × 5 group pile under static and dynamic lateral loads, present a unique opportunity and allow a rigorous study without arbitrary parameter back-fitting. The coupled soil–pile system is idealized through two-dimensional finite elements with soil models idealized by a hyperbolic-type multiple shear mechanism. Nonlinear spring elements are used to idealize the soil–pile interaction through a hysteretic nonlinear load–displacement relationship. Joint elements with a separation–contact mechanism are used to idealize the separation effect at the soil–pile interface. Ignoring soil–pile separation in static tests overestimates the ultimate lateral load–carrying capacity by 43% for a single pile and 73% for the trailing pile in a closely spaced pile group. Moreover, neglecting soil–pile separation in dynamic tests overestimates the total group load–deflection curve in both the loading and unloading phases.