Smart forecasting: Enhancing virtual power plant performance with analytical frameworks

• The study proposes an analytical framework for forecast electricity production in Virtual Power Plants (VPPs), focusing on a photovoltaic (PV) system installed at the Faculty of Electrical Engineering and Information Technologies in Skopje. • ARIMA models are utilized for short-term and seasonal electricity generation forecasts, highlighting their effectiveness compared to linear regression models. • The analysis evaluates key performance metrics, including Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), across different seasons, revealing ARIMA’s superior predictive capabilities. • Findings underscore the importance of precise forecasting for optimizing energy scheduling, enhancing grid stability and improving the integration of renewable energy sources into modern power systems. • The proposed framework serves as a foundation for further research and practical applications in the field of sustainable energy management. This paper presents a comprehensive framework for efficient electricity forecasting within a Virtual Power Plant (VPP) context. The focus is on forecasting both electricity consumption and production from photovoltaic (PV) systems, as essential components for successful operation of a sustainable energy infrastructure/system. In order to investigate the applicability of the framework, a case study approach was employed. The aim was to evaluate the effectiveness of various statistical modeling techniques, including time series analysis and regression analysis. Three datasets were analyzed: (i) electricity consumption at Ss. Cyril and Methodius University in Skopje (UKIM) facilities, (ii) solar radiation data for Skopje city region, and (iii) historical PV power generation data. The study delved into the application of Auto Regressive Integrated Moving Average (ARIMA) models to forecast PV power generation from a 22 kW PV system. A comparative analysis with linear regression highlighted the superior performance of ARIMA models, particularly in capturing seasonal patterns and handling the inherent variability of solar radiation. ARIMA (1,0,5) model was selected as the most efficient, with highest accuracy during summer, confirmed by the Mean Absolute Error (MAE) of 1.219 and R² = 0.899, likely due to stable and high levels of solar radiation. The highest error is observed in winter with MAE of 1.594 and R² = 0.631, which is attributed to the lower and more variable solar radiation typical of this season. The findings of this research contribute to the advancement of VPP management strategies by providing accurate and reliable electricity forecasts. These forecasts can be utilized to optimize energy scheduling, demand response, and improve the grid integration of renewable energy sources, ultimately leading to more efficient and sustainable energy systems.