Local Optimal Reactive Power Control of PV with Day-Ahead Load Forecasting

In view of the difficulties brought by the strong uncertainty of photovoltaic (PV) output to the optimal control of reactive power in distribution networks, a local optimal reactive power control method of PV with day-ahead load forecasting is proposed in this paper. First, some basic assumptions about operation conditions and optimal reactive power state in distribution networks are put forward, and a two-sided reactive power control framework including master station and PV controllers is established. Secondly, long short-term memory (LSTM) neural network is used for day-ahead load forecasting, and the expressions of optimal reactive power reference of PV plants are derived. Finally, a case study is conducted to prove the effectiveness of the proposed method. The results show that the proposed method can rapidly and optimally control the PV reactive power output, reduce the energy loss of the distribution network, and improve the voltage quality.