Deformation and early warning of reservoir landslide subjected to water level fluctuation and rainfall

During the reservoir filling and operational periods of the hydropower station, periodic changes in reservoir water levels and sudden heavy rainfall could adversely affect the hydrological and geological conditions, resulting in slope deterioration and increased susceptibility to reservoir landslides, which may cause project shutdowns and even catastrophic failures. Therefore, this study investigated the mechanisms of reservoir landslide evolution and proposed an innovative early warning model to provide warnings for reservoir landslides. First, this study designed a physical model experiment for reservoir landslides, employing Terrestrial Laser Scanning (TLS), pore water pressure sensors, and inclinometers to monitor the slope model. The research investigated the complete evolutionary process and failure mechanisms of reservoir landslides under two conditions: fluctuations alone and reservoir water level cycles combined with rainfall. Results indicate that hydrodynamic actions, fine particle loss, and bank reconstruction are the main factors leading to the instability and destruction of reservoir landslides. Then, a novel early warning model was proposed based on statistical methods for the early warning of reservoir landslides. This method can effectively address the issue of conventional methods failing to accurately capture the true destabilization of reservoir landslides exhibiting step-like evolutionary characteristics. The findings from this work are crucial for enhancing disaster prediction, prevention, and early warning strategies, thereby safeguarding the lives and properties of communities adjacent to reservoirs.