Development of Regional Soil Shear Strength Database and Its Application in Probabilistic Analysis of Slope Stability

Characterization of uncertainty in soil parameters constitutes an essential key aspect of probabilistic analysis in geotechnical engineering. While it is desirable to depict the soil variability for a specific site under a particular project, it is often difficult to obtain the site-specific probability distribution of soil properties mainly due to a lack of data. A pragmatic alternative to characterize the soil variability is to collect a large number of data for the same soil type from different sites across the region. This regional database would be useful for a preliminary reliability analysis of a site, or it could be served as prior knowledge to enrich the probability distribution of limited site-specific data. This study formed a database of over 3,000 data of shear strength parameters for various soils in Hong Kong. These data were derived from triaxial tests on soil samples retrieved from slopes in Hong Kong. The data have been compiled to form probability density functions of Mohr-Coulomb effective shear strength parameters (effective cohesion, c' and effective friction angle, ø') for different types of soils that are commonly encountered in the territory. Although the data were from various projects, the ground investigation field works and laboratory testing procedures are consistent as they follow the standard local practice. The application of these probability density functions is demonstrated through a case study of reliability analysis of a soil cut slope utilizing the Monte Carlo Simulation. The results of the reliability analysis incorporating the regional database are discussed against that undertaken based on limited site-specific data to articulate the differences in the calculations. Potential applications of the regional database are also discussed.

Keywords: Soil shear strength, Material variability, Triaxial test, Slope reliability, Monte Carlo simulation