Research on Reservoir Safety Risk Based on BP Neural Network

Abstract

The work of forecasting the torrential flood has always been highly valued by the government departments. With the gradual improvement of the water level monitoring conditions, rainfall monitoring conditions and engineering monitoring conditions of the reservoir projects in the area where the torrential flood occurs frequently, the accuracy of the forecast will be greatly improved if we combine the relevant information of the reservoir with the work of forecasting the torrential flood. Based on reservoir's security risk, this paper analyzes the main influence factors of affecting reservoir inflow. On this foundation, this paper have discussed the relational functions between the reservoir inflow and influence factor, the principle and algorithm of BP neural network and the process of forecasting reservoir inflow with the method of BP neural network as well. Finally, this paper established a data-driven model based on the BP neural network. The scientificalness and rationality of the BP neural network has been proved in this paper. The research results of this paper can provide new ideas for the method of forecasting the reservoir's downstream torrential flood.