Prediction of Power Grid Rainfall-Induced Flood Inundation Risk Based on Machine Learning and HEC-RAS

With the continuous expansion of power grids and the frequent occurrence of extreme weather events, rainfall-induced floods pose a significant threat to the safe operation of power grids. Predicting the risk of rainfall-induced flood inundation for power grids can effectively enhance their resilience to such disasters. Taking Fangshan District, Beijing as the study area, this research designs rainfall processes for different return periods based on the rainfall intensity formula, constructs rainfall-induced flood inundation processes using HEC-RAS, employs the Random Forest algorithm for learning and training, and validates the results using rainfall data from the “July 23” event in Beijing. The results indicate that the rainfall-induced flood inundation processes can be effectively predicted. Furthermore, by integrating power grid site data, early warnings for rainfall-induced flood disasters are implemented for the power grid. Through this approach, the prediction of power grid rainfall-induced flood inundation risk is achieved, providing a scientific basis and decision-making support for power grid planning, operation, and maintenance.