Day-ahead Forecast of PV Systems and End-Users in the Contest of Renewable Energy Communities

Abstract

Efficient forecasting algorithms represent a key issue for monitoring and control of energy communities (ECs). After recalling the basic steps to perform a forecasting analysis, this paper reports the preliminary results obtained within the framework of the Italian ComER project, that aims to develop methods and tools for management and control of renewable ECs. Day-ahead forecast is applied to photovoltaic (PV) systems, residential end-users, and a public building. For each user, firstly an accurate preliminary analysis of public and private datasets of the target variables is performed. Then, forecasting methods based on persistence, multiple linear regression (MLR) and autoregressive integrated moving average (ARIMA) models have been implemented. Finally, results are compared to identify the more accurate forecasting method for each user belonging to the EC.