Predictive Modeling of Photovoltaic Energy Yield Using an ARIMA Approach

This paper presents a method for predicting the energy yield of a photovoltaic (PV) system based on the ARIMA algorithm. We analyze two key time series: the specific yield and the total yield of the PV system. Two ARIMA models are developed for each time series: one selected by the authors and one determined by SPSS. Model performance is evaluated through fit statistics, providing a comprehensive assessment of model accuracy. The residuals’ ACF and PACF are examined to ensure model adequacy, and confidence intervals are calculated for residuals to validate the models. A monthly forecast is then generated for both time series, complete with confidence intervals, to demonstrate the models’ predictive capabilities. The results highlight the effectiveness of ARIMA in forecasting PV energy yields, offering valuable insights for optimizing PV system performance and planning. This study contributes to the field of renewable energy forecasting by demonstrating the applicability of ARIMA models in predicting the monthly performance of photovoltaic systems.