Characterization Of Subsurface Stratigraphic Heterogeneity And Reliability Updating Of Slopes Based On Fusion Of Site Exploration Data And Geophysical Data

Drilling or in-situ penetration test is usually used in slope engineering investigation,and the survey data obtained are very limited and sparse,so it is difficult to accurately describe the formation distribution between boreholes.In contrast,geophysical exploration techniques are inexpensive,have a wide range of measurements,and have large sampling Spaces.And it allows rapid acquisition of low-resolution exploration data that characterize the distribution of2D(or 3D)stratigraphic profiles.Characterization of slope heterogeneity based on the fusion of geophysical exploration data and drilling data is a key problem.In this paper,a series of related researches are carried out by organically combining random field,Bayesian,geostatistics and reliability theory.The main work and conclusions are as follows:(1)A characterization method of spatial variability of slope parameters based on the fusion of drilling data and geophysical exploration data is proposed,a Bayesian updating calculation formula considering various sources of uncertainties is deduced,and a modified Bayesian updating method with subset simulation is developed for the numerical solution of highdimensional Bayesian updating problem,which lays a foundation for the characterization of slope stratum heterogeneity and reliability updating evaluation.(2)A method for slope soil layer division and boundary recognition based on the fusion of multi-source test data is developed,and the performance of Bayesian method and co-Kriging method to generate two-dimensional stratigraphic distribution model is quantitatively evaluated,which effectively reduces the uncertainty of stratigraphic distribution description.(3)On the basis of inferring the posterior probability distribution of soil parameters based on multi-source data fusion,the updating of slope reliability considering the spatial variability of soil parameters is carried out,which effectively improves the accuracy of slope reliability evaluation.It is concluded that the posterior slope failure probability is closely related to the field test data and is not necessarily less than the prior slope failure probability.