A Voltage Control Strategy for Low-Voltage Distribution Network with High Proportion PVs Based on BP Neural Network

As large-scale distributed PVs are integrated into distribution network, the distribution network penetration rate continues to increase. The imbalance between photovoltaic output and load leads to reverse flow in the line, which easily causes the voltage to exceed the limit. The active distribution network can effectively prevent the voltage from exceeding the limit by coordinating and controlling the local controllable resources. This paper proposes a data-driven centralized voltage control method for distribution networks, and this method does not need accurately modeling the distribution system. Only the historical operation data are used as the training data of the BP neural network to fit the functional relationship between the injection power and the voltage of nodes. Then the voltage sensitivity can be obtained through trained BP neural network for voltage control in distribution networks. Effectiveness of the proposed voltage control method is verified by a simulation.