Forecasting Solar Energy Production through Modeling of Photovoltaic System Data for Sustainable Energy Planning

The use of photovoltaic energy is critical for supporting the transition to sustainable energy systems and for reducing dependence on fossil fuels. This study provides an analysis and forecast of the monthly electricity production of four 30kW photovoltaic (PV) power plants located in the Southwestern region of Bulgaria. We used five years of data to consider seasonal variations in solar energy production typical of temperate climates, as well as peak summer production and significant declines in winter.

The prediction was carried out using ARIMA algorithms, which are based on time series models. Analysis of the residuals involves applying different statistical approaches such as autocorrelation (ACF) and partial autocorrelation (PACF) for the determination of a suitable model. The reliability of the models was confirmed by calculating confidence intervals and by applying standard precision metrics, which provides a basis for reliable forecasting of future electricity production.

The study demonstrates that ARIMA models can successfully capture seasonal dynamics and long-term trends in photovoltaic production. Building forecasting models provides valuable information for decision-makers, helping them manage capacity, optimize costs, and plan strategically. According to the results, this approach is capable of improving the efficiency and sustainability of small-scale solar installations for business and personal use.