Granite residual soils (GRS) are often encountered in geotechnical projects in the Guangdong-Hong Kong-Macao Greater Bay Area (briefly written as the Greater Bay Area, or abbreviated as GBA). The rea experiences frequent rainfall, leading to wetting-drying cycles that progressively diminish the shear strength of GRS. This weakening effect is not only significant but also accumulates, exhibiting a direct positive correlation with the number of cycles. Current studies on the soil strength attenuation due to wetting-drying cycles are typically limited to no more than 10 cycles, which is rather insufficient to uncover the long-term water-weakening behaviors and their accumulative impacts on GRS. To address this gap, typical GRS samples were first taken from the GBA and then prepared by making them go through a certain number of wetting-drying cycles (maximum of up to 100). Next, a total of 552 small- and large-scale direct shear tests were conducted to investigate the mechanisms of water-weakening effects on soil internal friction angle, cohesion, and shear strength. The degree of saturation and number of cycles were also examined to see their effects on the cumulation of water weakening. Based on results from the small-scale direct shear tests, a model was developed for assessing the weakening impact of water on soil strength. The accuracy of the model prediction was statistically evaluated. Last, the effectiveness and efficiency of the proposed model were demonstrated by validating against the results from the large-scale direct shear tests.

Despite the prevalence and validity of the universal distinct element code (UDEC) in simulations in geotechnics domain, water-weakening process of rock models remains elusive. Prior research has made positive contributions to a presupposed link between modelling parameters and saturation degree, Sr. Nevertheless, this effort presents inaccurate results and limited implications owing to the misleading interpretation, that is, devoid of the basic logic in UDEC that modelling parameters should be calibrated by tested macroscopic properties in contrast to a presupposed relation with Sr. To fill this gap, a new methodology is proposed by coupling a computationally efficient parametric study with the simulation of water-weakening mechanisms. More specifically, tested macroscopic properties with different Sr values are input into parametric relations to acquire initial modelling parameters that are sequentially calibrated and modulated until simulations are in line with geomechanical tests. Illustrative example reveals that numerical water-weakening effects on macroscopic properties, mechanical behaviours, and failure configurations are highly consistent with tested ones with noticeable computational expediency, implying the feasibility and simplicity of this methodology. Furthermore, with compatibility across various numerical models, the proposed methodology substantially extends the applicability of UDEC in simulating water-weakening geotechnical problems. (c) 2025 Institute of Rock and Soil Mechanics, Chinese Academy of Sciences. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/ 4.0/).