Stochastic Analysis of Piled-Raft Foundations Using the Random Finite-Element Method

Abstract

The inherent variability of soil parameters noticeably influences geotechnical structures such as piled-raft foundations. In this research the application of random fields is extended to a stochastic analysis of a piled-raft foundation embedded into an elastoplastic soil a problem that is known as complicated soil­structure interaction. For this purpose a stochastic framework is developed using a random finite-element method (RFEM) by coupling the random field theory and a finite-element-based program coded in MATLAB to model the stochastic plane strain piled-raft foundation behavior. The analysis is performed deterministically and then extended to the stochastic context to consider the spatial variability of Young’s modulus friction angle cohesion and unit weight of the soil media. The stochastic responses of the piled-raft foundation such as displacement axial force bending moment and shear force are assessed. The stochastic analysis shows that implementation of the uncertainties of soil parameters leads to more influence on the raft differential settlement compared to the other responses. Variation of the horizontal and vertical correlation length has a more considerable effect on the coefficient of variation (COV) of the bored piles’ responses compared to those of the raft. Based on the COV of the piles and raft load contribution it is demonstrated that the heterogeneity of the soil has more influence on the raft. Furthermore it is found in all realizations that the main percentage of the load is carried by the bored piles.