Development of a DEM-Based Method for Modeling the Water-Induced Failure Process of Rock from Laboratory- to Engineering-Scale

The time-dependent hydromechanical behavior of rocks is crucial to assessing the long-term stability of rock engineering, such as reservoir rock slopes and underground waste storage facilities. This research aims at developing a discrete-element method (DEM)-based numerical method to simulate the whole water-induced carbonate rock failure process, which is highly sensitive to water with a maximum 70% loss of mechanical strength when saturated by water. First, strength degradation laws for carbonate rock induced by water were derived based on the different behaviors of two types of cement inside the rock. The degradation laws were coupled with universal distinct element code (UDEC) to build the progressive damage model by developing a customized code using Fish language. Numerical simulations were then carried out on laboratory-scale tests and an example of an engineering-scale rock slope. In laboratory-scale simulations, the predicted lifetime (time-to-failure) of the rock sample generally aligns with that of experimental results, and well-known physical phenomena, such as the pattern of macroscopic fracture and stress–strain responses, can also be reproduced. In the engineering-scale simulation, the long-term (nearly 5 years) progressive failure of the slope toe under the influence of the fluctuation reservoir level, and the consequent rockslide can be well modeled. The results indicate that the proposed model is capable of predicting the strength properties for carbonate rocks submerged in water for different time periods, as well as simulating the whole failure process of rock from the engineering scale due to water erosion.