Integrated Wind-Solar-Hydro Power Prediction Method Based on Deep Neural Network

With the increase of the proportion of renewable energy in the power system, the randomness and volatility of wind power and photovoltaic output and the seasonality of hydropower output have increasingly prominent impacts on the security and reliability of the power system. In order to guide the day-ahead generation plan of wind-solar-hydro complementary system, it is necessary to accurately predict the future output of wind-solar-hydro, analyze the generation benefit and output reliability of distributed power supply side, and finally realize the economy and reliability of power grid operation under high-density distributed new energy integration. In this paper, an ultra-short-term power prediction method based on deep neural network for regional wind-solar-hydro integration is proposed. Considering the spatio-temporal correlation of the stations, the prediction accuracy is improved. Historical wind-solar-hydro power series are decomposed by empirical mode and reconstructed into high frequency and low frequency components. On this basis, based on LSTM deep neural network, the integrated prediction models of the high frequency and low frequency power series of the wind-solar-hydro are constructed, respectively. By fusing the prediction sequences obtained from the two models, the integrated prediction of the ultra-short-term power of the regional wind-solar-hydro is realized.