Solar Photovoltaic Energy Forecasting Using Improved Ensemble Method For Micro-grid Energy Management

Solar photovoltaic energy is one of the renewable energy sources which can be used in microgrids. The inherent drawback of solar photovoltaic energy is the intermittent nature caused by the seasonality and cloud cover. Therefore, forecasting methods are used to account the effect of intermittency of solar energy. This paper describes about the forecasting method that used improved ensemble method in order to forecast the variation of power generation of solar panels due to the cloud cover. The ensemble forecasting model is a combination of a sky imagery model and a statistical model. The combination of two models is done taking the weighted average of forecasted solar photovoltaic (PV) generation output given by each model. The ensemble forecasting model was tested and validated with the data of days of different weather conditions such as sunny day and rainy day. The accuracy of model is determined by root mean square error of predictions. In sunny day and rainy day and the RMSE value of predictions are 1.066W and 0.351W respectively.