ABSTRACT
Probabilistic modeling of geotechnical strength parameters for large scale in-situ characterization of geological formations is presented in this paper. The actual measurements of strength parameters from a limited sample test (hard data) are combined with other sources of information (subjective soft data) to characterize the localized probabilistic model of in-situ strength parameters. The sequential Gaussian cosimulation technique was introduced to achieve the localized probabilistic characterization of in-situ strength parameters. The bivariate probability distribution was applied for the probabilistic joint model of strength parameters (c-ϕ). In other words, the full statistical distribution of strength parameters are obtainable for every local point within the geological domain. These localized probabilistic models provide the necessary input for probabilistic structural analysis, such structures as slopes, dams, foundations or underground openings, where c and ϕ are applied for the structural design.
