Predicting shear strength of unsaturated soils based on soil–water retention curve

The complexity of unsaturated cohesive soil behavior presents challenges in directly measuring unsaturated shear strength, making it a complex and time-consuming task. Scholars have proposed indirect models to estimate unsaturated strength using soil–water retention curves and saturated shear strength indicators. However, scholars lack consistency in defining parameters to characterize apparent cohesion, resulting in a lack of standardized expressions. To establish a unified model for predicting the strength of different types of unsaturated cohesive soils, existing unsaturated shear strength models based on soil–water retention curves were systematically reviewed. Fifteen sets of experimental data were collected and utilized to analyze and compare the predictive performance of these models. It was observed that existing predictive models partially reflect the strength of unsaturated cohesive soils to some extent. However, they have applicability limitations and fail to predict the unsaturated shear strength of all soil types fully. An improved model for the shear strength of unsaturated cohesive soil was developed to overcome these limitations based on the Khalili and Khabbaz (1998) model. This improvement involved replacing a fixed empirical value in the Khalili and Khabbaz (1998) model with the water loss obtained from the soil–water retention curve. The average relative error (ARE) and normalized sum of square error (SSE) were used to quantitatively evaluate the predictive accuracy of the unsaturated soil strength model, comparing the improved model with existing ones. The analysis revealed that the improved model demonstrated higher prediction accuracy across fifteen types of unsaturated soils. Furthermore, soil–water retention curve tests and unsaturated triaxial tests were performed on two types of test soils, with sand-clay mass ratios of 3:2 and 1:4, respectively. By comparing the test data, the effectiveness of the improved model in predicting shear strength was evaluated, affirming its generalizability and accuracy in estimating the shear strength of unsaturated clay soils. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.