Balancing the grid: mitigating the effects of renewable energy in Italy via skew modeling and forecasting

In this study, we introduce a conservative model designed to accurately predict renewable energy generation while addressing fluctuations and imbalances in power supply. We employ a time-inhomogeneous Vasicek model driven by both a skew Brownian motion and a compound Poisson process, providing proof of solution existence, uniqueness, and mean reversion. Our focus is on minimizing deficits caused by underestimating actual generation levels. By prioritizing a positively skewed distribution of deviations and aiming for them to cluster around zero, we achieve significant reductions in forecasting errors compared to established benchmarks. Utilizing hourly data from Terna in 2023, our model illustrates the effects of energy imbalances, which lead to financial losses, yet also highlights enhanced cost savings and significant reductions in CO2 emissions.