Ultra-short term hybrid power forecasting model for photovoltaic power station with meteorological monitoring data

In order to realize the real time tracking and prediction of ultra-short term power of PV power station, according to the actual monitoring data of the photovoltaic power station, an ultra-short term power forecasting model based on autoregressive moving average (ARIMA) and support vector regression (SVR) is established in this paper. The ARIMA model has good adaptability to the linear trend. The results show that the method has good tracking and prediction affection in the prediction of irradiance and temperature. It can be a reliable input to regression models. The SVR model realizes the highly nonlinear fitting of data mapping. The effectiveness of the ARIMA-SVR model is verified by the prediction of the ultra-short term output power of PV power plants under four different weather conditions.