Short-term load forecasting method based on structural neural network

Neural network can increase forecasting accuracy of power system load , but canpsilat provide explanation for forecast reason, so this paper proposes a short-period load forecasting method based on structural neural network. The paper respectively set up such models as historical load data forecasting model, weather forecasting model and date type model. First three models are respectively learned and then are combined and learned again. The examples indicate that the method can not only improve forecasting accuracy but also analyze load factors. Therefore the method provides a feasible basis for quantitative study of how various load factors affect load.