| Summary: | 碩士 === 國立中央大學 === 應用地質研究所 === 106 === Previous investigations have recognized the important role of groundwater variations on slope stability. Stochastic approaches are useful techniques to quantify the input uncertainty (i.e., the soil heterogeneity) on the output uncertainty (i.e., the displacement uncertainty). This study aims to develop a stochastic modeling workflow based on numerical Monte Carlo simulations. The Monte Carlo simulations involve a number of procedures including simulations of random hydraulic conductivity fields, random porosity fields, random unit weight fields. This study employs sequential Gaussian simulation method (SGSIM) model to generate random realizations of hydraulic conductivity fields, porosity fields, and unit weight fields. The commercial model FLAC3D (Fast Lagrangian Analysis of Continua in Three Dimensions) is then used for the simulations of slope stability. By collecting realizations of flow and displacement solutions in a slope with a fixed geometry and boundary conditions, the workflow can quantify the propagation of flow uncertainty on displacement uncertainty. The simulation results show that the variance of the logarithm of hydraulic conductivity significantly influences the water level variation in the slope system. The pressure head and values of water level variance values show one to two orders of magnitudes smaller than that of the logarithm of hydraulic conductivity, depending on the background flow gradients. The high-pressure head variance occurs near the downstream boundary. These high-pressure head variances also lead to high instability in the slope system. In addition, this study also discusses the influence of spatially variable porosity and unit weight parameters on the displacement uncertainty. Although the flow pattern is similar to homogeneous cases, the displacement uncertainty induced by porosity variation is relatively small as compared with the displacement uncertainty induced by spatial variation of unit weight.
|
|---|
