A Hybrid Approach for Day-Ahead Forecast of PV Power Generation

In the past years, the applications of solar energy have grown significantly in electricity generation. However, the fluctuations of PV power output create different negative impacts on reliability, stability, and dispatch in the connecting grid. The exact PV power generation forecast is thus crucial to stabilize the operation of a power grid. This paper presents a radial basis function neural network with decoupling method for day-ahead PV power generation forecast. Results are compared with autoregressive integrated moving average (ARIMA), back propagation neural network (BPNN), and radial basis function neural network (RBFNN), and the actual measured PV power outputs. It shows that the proposed model leads to more accurate and the computational efficient forecast on PV output.