An intelligent scheduling control method for smart grid based on deep learning
Zhanying Tong, Yingying Zhou, Ke Xu
Abstract
Nowadays, data analysis is been the most important means to realize power scheduling in smart grids. However, the sharp increase in business data of grids has posed great challenges for this purpose. To deal with such issue, this paper utilizes deep learning to discover hidden rules from massive large-scale big data and particle swarm optimization (PSO) algorithm for generation of control decision. Therefore, an intelligent scheduling control method for smart grid based on deep learning is proposed in this paper. By modeling the historical data of the power company, the long short-term memory algorithm can effectively extract the effective features and realize the prediction of the coal consumption of the unit under certain conditions. At the same time, a kind of intelligent power scheduling algorithm is designed by using PSO, so as to save energy and reduce emissions as much as possible while fulfilling the real-time power generation task. Experiments on a real-world smart grid dataset show that the proposal can achieve a relatively good performance with respect to intelligent scheduling.