The Boolean Stochastic Generation method - BoSG: A tool for the analysis of the error associated with the simplification of the stratigraphy in geotechnical models

Abstract

In geotechnical modelling, some minor stratigraphic features are usually discarded in order to simplify the problem, avoiding to deal with further uncertainties about their position, thickness and lateral extent. The study proposes a new method based on the stochastic generation of different soil layers configurations, following a boolean logic: the material is either matrix or layer (i.e., gravel lenses in a clay-rich matrix).The method has been called BoSG (Boolean Stochastic Generation). The methodology allows to randomize the presence of a specific material interdigitated in a uniform matrix thus enabling to gather a dataset which could be analysed automatically, in order quantify the error associated with the adopted simplification.The commercial codes FLAC and FLAC3D were used for the geotechnical modelling. A specifically-coded MatLab program allows to generate randomly the different soil configurations and then to automate the computation with the commercial software in order to maximize the sample number.In this paper the methodology is applied with reference to a simplified slope in 2D and in 3D. Results show that within a low resistance matrix, the presence of layers with higher friction angle can significantly affect significantly the stability and the displacement pattern of an unstable slope. Therefore, a method to investigate the influence of the spatial distribution of these layers can be particularly useful.