Optimal Flexibility Provision of Electric Vehicle and Photovoltaic Systems through Probabilistic Forecasting and Unsuccessful Day-Ahead Market Integration

Integrating electric vehicles (EVs) and photovoltaic (PV) systems into power grids presents challenges due to uncertainties in EV charging demand and PV generation, especially with limited real-time data. Managing these uncertainties, particularly in providing optimal flexibility for EV late arrivals, early departures, and PV forecast errors, remains a key concern. This study proposes a proactive flexibility provision framework that leverages probabilistic forecasting and cumulative adjustment factors (CAFs) to dynamically refine flexibility bounds. The method optimizes local resource utilization and ensures flexibility availability, effectively managing real-time power imbalances. The program also integrates unsuccessful day-ahead market participants into real-time operations, enhancing local resource use and reducing dependence on the main grid. Numerical simulations on a 380-V low-voltage distribution system in Thailand demonstrate that the model successfully resolved all real-time power imbalances. Compared to the no-CAF baseline (61.59%), the proposed method increased utility reduction to 69.10%. Importantly, log-normal distributions with small variance (e.g., σ = 0.1 for PV and σ = 0.5 for EV) achieved consistently strong performance across mean absolute error, root mean square error, and coverage probability metrics—delivering accurate forecasts while supporting long-term contracts and grid planning. Incentives were fairly allocated and fully funded within the local energy market.