BO‐XGBoost‐based voltage/var optimization for distribution network considering the LCOE of PV system

Abstract

AbstractWith large‐scale renewable energy connected to the distribution network, the traditional method to solve the distribution network voltage/var power optimization model can hardly meet the needs of online optimization of the distribution network. Therefore, a Bayesian optimization eXtreme Gradient Boosting (BO‐XGBoost) based optimization strategy for voltage/var of the distribution network considering the cost of the photovoltaic (PV) power is proposed to improve the solving speed of the optimization model. First, the voltage/var optimization model of the distribution network considering the LCOE of the PV system is established. The optimal reactive power output command data set of the PV generation in the distribution network is obtained by solving the established optimization model. The XGBoost model is used to mine the nonlinear mapping relationship between the state information of the distribution network and the optimal reactive power output of the PV system; then the Bayesian algorithm is used to complete the adaptive optimization of the hyperparameters of the XGBoost model. Finally, the PG&E 69 node power distribution system is used to verify the advantages of the proposed strategy in improving the performance of the XGBoost algorithm, increasing the solving speed of the voltage/var optimization model, and reducing the LCOE of PV generation.