Geotechnical uncertainty analysis with sparse data using interval field method

Abstract

Spatial uncertainty in soil parameters has essential implications for dyke failure analysis. The currently popular random field methods used to characterize the spatial uncertainty of soils have been very influential. However, it requires a large amount of statistical data from Geotechnical engineering. The interval field method, as a complementary tool to the random field method, can be used to analyze geotechnical engineering and make decisions in the presence of sparse data. This paper adapts the method of generating interval fields with B-splines for Geotechnical engineering. The method of generating interval fields based on B-splines can consider the known values of the interval field at specific locations and impose zero intervals at these locations. Therefore, it is easy to model the interval field for geotechnical engineering considering the measured parameters of the borehole. Then, the stability of the deterministic dike is analyzed by the finite element method to estimate the safety factor. Subsequently, a Bayesian global optimization algorithm efficiently evaluates the upper and lower bounds of the safety factor. Finally, the validity of the interval field method is verified for a small polder field in south Holland. The results show that the interval field method can provide a reasonable basis for decision-making in geotechnical engineering.