부산대학교 도서관

The power prediction of photovoltaic (PV) power generation is important to reduce the impact of PV grid connection and maintain the grid's secure and stable operation. However, the power prediction for PV power station still has the problem of poor prediction accuracy. To address the problem, this paper proposes a back propagation (BP) neural network-assisted power prediction method for PV power station, which can achieve realtime and high-accuracy power prediction by effectively training a BP neural network model. First, the power prediction architecture for PV power station is constructed. Then, the BP neural network-assisted power prediction algorithm is proposed to provide accurate PV power prediction. Simulation results demonstrate the superior performance of the proposed algorithm in prediction deviation.