Very short-term solar PV generation forecast using SARIMA model: A case study

The increase of interests in low carbon energy resources has promoted the deployment of solar photovoltaic (PV) systems. However, the grid integration of solar PV generation has few technical challenges due to its intermittency and non-dispatchability. An accurate (very) short-term solar PV generation forecast is critical to ensure secure and economic operation as well as for grid energy management. In this paper, the Seasonal Autoregressive Integrated Moving Average (SARIMA) model has been adopted for the multi-step ahead forecast (20 minute resolution) of solar PV generation. The roof-top solar PV systems installed at academic blocks and student hostels at IIT Gandhinagar campus are used as a case study to implement and validate the forecast model. The developed model is compared with the persistence model.