Short-term load forecasting model for metro power supply system based on echo state neural network

Abstract

The paper presents a short-term load forecasting model for metro power supply system based on echo state neural network. Echo state neural network composed of input layer, reserve pool, the output layer. Reserve pool as a dynamic network is connected by a large number of random sparse of neurons. Reserve pool is used to overcome the slow convergence speed and avoid neural network into the local minimum. Using the actual historical data of the metro power supply system to simulate, the simulation results show that the short-term load forecasting model for metro power supply system based on echo state neural network has good prediction accuracy.