Imbalance Charge Reduction in the Intra-Day Market using Short-Term Forecasting of Photovoltaic Generation

The growing integration of photovoltaic power into electricity markets poses new challenges, particularly in terms of system balancing and forecasting accuracy. How-ever, due to the strong dependence of PV output on weather conditions, forecasting the actual energy injected into the grid has become increasingly complex. Therefore, within intraday trading across Europe, short-term forecasting accuracy is crucial to reduce imbalances and improve market performance. While forecast performance is tradi-tionally assessed using metrics such as RMSE or MAE, under the assumption that lower errors yield higher economic benefits, this relationship is not always linear. A thorough understanding of market conditions and the link between forecast deviations and imbalance penalties is essential. In this context, this study proposes a methodology to evaluate, analyze and, if necessary, reduce the charge imbalance in the Intra-Day market, using a short-term photovoltaic power forecasting model based on a hybrid physical–artificial neural network approach. The analysis focuses on the Italian elec-tricity market, particularly the MI-XBID continuous trading session, where participants adjust their energy positions through continuous intraday trading. The results demonstrate that improved forecasting help in reducing charge imbalance-related costs and enhancing market efficiency. Furthermore, the evaluation of charge imbal-ances can serve as a key input for assessing the improvement that can be achieved by integrating a storage system with suitable sizing and charge/discharge control logic. By enhancing the efficiency and reliability of renewable energy integration, the proposed approach contributes to the broader goals of sustainable development and energy transition.