Ranking and transformation error assessment of shear strength parameters correlations

Abstract

Shear strength is responsible for soil ability to withstand applied loads and considered the most important engineering properties of soil. Shear strength can be measured directly in the field or in the laboratory, however many empirical correlations were developed to predict shear strength from simple basic and index properties of soil. The main reasons for using these correlations that direct measurements are costly, time consuming, and not always applicable. The correlated shear strength parameter deviates from its actual value, by a transformation error, causing larger uncertainty that affects the overall design. The current commonly used transformation error values have shortcomings that warrant the need for further research. To offer avenues for improving the current values, a database contains geotechnical investigation results for 5087 boreholes collected from Egypt, UAE, Iraq, and Indonesia were created. The database served to assess inherent variability, measurement error, and transformation error more consistently. The most commonly used empirical correlations were ranked based on amended Theil inequality coefficient method, then the transformation error was determined using uncertainty propagation combined with the second moment probabilistic method (RUP-SMP). The adapted methodology can be used as standard procedures to evaluate different empirical correlations for other geotechnical properties. Framework for combining the predicted transformation error with other uncertainties was introduced to determine the overall uncertainty in shear strength parameters.