Statistical Modeling of Geotechnical Parameters of a Shallow Foundation Located in Daral Peulh (Thiès, Senegal)

Abstract

In general, the analysis and design of civil engineering structures are based on deterministic approaches where uncertainties in various parameters (soil characteristics, loading, etc.) are not taken into account. In recent decades, the description and analysis of uncertainty in geotechnical engineering, with a focus on the safety of structures, have become increasingly important issues. To fix the level of risk for which geotechnical parameter values must be determined, the Eurocodes have introduced an approach based on the statistical processing of soil data. This paper presents the statistical modeling of geotechnical parameters (cohesion, friction angle, specific weight) of a shallow foundation soil from Daral Peulh area located in Thies in Senegal. Laboratory tests for soil identification were conducted on 26 samples. After outliers, values determined and treated with boxplot method, graphical methods and Kolmogorov and Shapiro-Wilk tests were used to determine the probability distributions associated with geotechnical parameters and the bearing capacity of a shallow foundation. The results conclude after outliers treatment with 5% risk, the probability distributions of the cohesion, the friction angle and the specific weight are normal distributions and can be used for probability design. However, the variability around their means is high for the cohesion, the friction angle and the bearing capacity. For the bearing, the most suitable probability distribution in this case is the lognormal distribution.