Rapid Forecasting Method for Flood Process by Using on Physically Based Numerical and AI Model

With the acceleration of urbanization, complex underlying surfaces, pipe networks, river channels, and hydraulic facilities (gates, sluices, pumps) have significantly increased the number of computational grids and physical processes, making the computational efficiency of physical rainfall-runoff models insufficient to meet the timeliness requirements of emergency management for flood disasters. This necessitates further research on new technologies to enhance the computational efficiency of flood simulation and forecasting models. The development of AI technology provides new approaches for rapid flood disaster simulation and forecasting. This study proposes three innovative methods to address these challenges. First, GPU Accelerated Model for Surface Water Flow and Associated Transport. Second, AI Based Rapid Predicting Method for Flood Process. Third, Model Application for Dam Break Flood Simulation.