Prediction of long-term settlement and evaluation of pore water pressure using particle filter

Abstract The present paper showcases a numerical simulation of the long-term behavior of a foundation and an evaluation of the pore water pressure on Kobe Airport Island. To calculate the behavior of the ground and the pore water pressure, a numerical analysis is implemented using an elasto-plastic FEM for soil–water coupled problems with the Cam-clay model. However, numerical difficulties still remain in estimating the elasto-plastic parameters because of the nonlinearity of the materials and the uncertainty of the initial/boundary conditions. To account for these difficulties, a data assimilation technique, the particle filter, is adopted to identify the soil parameters using field measurements. The numerical simulation is carried out with the identified parameters to predict the long-term settlement and to evaluate the pore water pressure of the improved ground for two cases of boundary conditions. The outcomes show a good agreement with the direct measurements of the long-term settlement and the observed pore water pressure.