Prediction of Power Consumption of Hydroelectric Power Station by Levenberg-Marquardt-BP Algorithm

Abstract

Improving the predicting and monitoring of station power consumption of hydropower stations is of great significance to realize the fine management of energy efficiency of hydropower stations and reduce the level of station power consumption. The reliability of electrical equipment operation is very important for the safe and stable operation of hydropower stations.The power consumption of hydropower stations is closely related to the operating status of electrical equipment of hydropower stations. this paper establishes a BP neural network prediction model based on the Levenberg-Marquardt algorithm (Levenberg-Marquardt-BP) to accurately predict the power consumption of electrical equipment in a hydropower station. Field tests show that the RMSE of Levenberg-Marquardt-BP prediction method is 2.1%, which is much lower than the conventional BP prediction algorithm. The Levenberg-Marquardt-BP algorithm also can quicken the algorithm convergence speed and its convergence steps are 35% of the conventional BP prediction algorithm.The analysis of prediction examples proves the reliability and effectiveness of the Levenberg-Marquardt-BP prediction method.