Evaluation of pore-pressure monitoring setup

Abstract

In this study soil parameters are back analyzed by using the finite element method and two optimization algorithms, the Non Linear Least Squares (NLLS) and Coupled Local Minimizers (CLM) method. Soil permeability parameters are back analyzed to match the observed values of the variation of pore-water pressures at the Fosso S. Martino Landslide in Italy. Piezometer readings are used to back analyze the permeability coefficients of three soil layers and the left-hand side water level boundary. Furthermore, the performance of each piezometer is evaluated through neglecting the information of other points and performing the optimization analysis. It is found that both optimization algorithms determine soil parameters that represent the pore-water fluctuation of the study area, with their advantages and drawbacks. The NLLS computes values that represent the performance of the pore-water fluctuation in the piezometer readings however, the selection of the starting values is important in order to reduce the computational time. The CLM method was also able to compute adequate soil parameters however, the computational effort is much higher due to the fact that this method uses several search points in the process. For the minimum monitoring setup needed for the adequate determination of soil parameters, the piezometers in the middle of the slope provided the best information. These piezometers are essential to adequately back analyze the soil properties in this case study. Great attention has to be considered in maintain these piezometers in order to continue monitoring the slope.