A recurrent neural network for optimal energy management considering the battery cycle-life in smart grid

This paper proposes a combined battery energy management (CBEM) model against the background of smart grid, which focus on the absorptive ability of Battery Energy Storage System (BESS) distinguishing to the traditional energy management. To solve the CBEM model, an one-layer recurrent neural network (RNN) is recommended. Furthermore, by considering the peak and valley periods division prices, the generators absorption and the distribution network bought from power grid are optimized. Meanwhile, the emission cost is considered in total cost and there is a corresponding constraint to control the emission and the outputs of generators. Finally, the effectiveness of the CBEM model is verified on Matlab. With the fluctuation of electrical price, the charge model and discharge model of BESS are showed. Results indicate that this model performs well in cutting the load in peak time and storage energy in valley time.