Cost Savings from Improved Forecast Accuracy in Distributed Energy Systems with PV-Demand Ratios and BESS Capacities

The rapid increase in photovoltaic (PV) deployment in distributed energy systems amplifies renewable integration but heightens uncertainty in supply-demand balance, especially as the PV to demand ratio changes. This study examines how enhancements in forecasting accuracy for PV output, power demand, and electricity price influence total system costs across varying PV-to demand ratios and battery energy storage system (BESS) capacities. We apply partial least squares for PV forecasting and the Prophet model for demand and price forecasts within a two stage optimization framework combining day ahead and real time scheduling. In this simulation, PV capacities vary from 1000 kW to 10000 kW (altering PV to demand ratios from 6.7% to 67%), BESS capacities vary from 0 to 2000 kWh. Additionally, electricity market spot price data representing stable (2020) and volatile (2022) conditions in Japanese electricity market. Results show that improved PV forecasts reduce normalized mean absolute percentage error by 35% and yield cost savings increasing from 0.5% at low PV to demand ratios to 8% at high ratios, while enhanced demand forecasts achieve from 6% to 3% savings although price forecast improvements contribute less than 1%. Under volatile pricing, greater forecast accuracy also justifies larger optimal BESS capacities. Unlike prior studies that assessed forecast improvements collectively, our analysis isolates each forecast’s individual contribution to cost reduction. These insights inform DES operators and policymakers on prioritizing forecasting investments and optimal storage sizing.