Estimation of Pore Water Pressure of Soil Using Genetic Programming

Abstract

Soil–water characteristic curve (SWCC) is one of the input components required for conducting the transient seepage analysis in unsaturated soil for estimating pore water pressure (PWP). SWCC is usually defined by saturated volumetric water content (θs), residual water content (RWC) and air entry value (AEV). Mathematical model of PWP could be useful to unearth the important SWCC components and the physics behind it. Based on authors’ knowledge, rarely any mathematical models describing the relationship between PWP and SWCC components are found. In the present work, an evolutionary approach, namely, multi-gene genetic programming (MGGP) has been applied to formulate the relationship between the PWP profile along soil depth and input variables for SWCC (θs, RWC and AEV) for a given duration of ponding. The PWP predicted using the MGGP model has been compared with those generated using finite element simulations. The results indicate that the MGGP model is able to extrapolate the PWP satisfactory along the soil depth for a given set of boundary conditions. Based on the given AEV and saturated water content, the PWP along the depth can be determined from the newly developed MGGP model, which will be useful for design and analysis of slopes and landfill covers.