Uncertainty-Aware Scheduling of Multi-Use Battery Storage Systems

Large renewable power plants like solar farms are increasingly equipped with large-scale battery storage systems.Trading solar energy on day-ahead markets based on generation forecasts requires solar farms to take measures to balance deviations from forecast values, for example, to avoid penalties.For this reason, solar farms are increasingly trading on intraday markets, adversely affecting energy prices.Alternatively, an installed battery storage system can maximize day-ahead revenues while balancing forecast deviations, resulting in a multi-use storage system.The challenge is to trade off the scheduling of the battery storage system between maximizing day-ahead revenues and reserving sufficient capacity to balance forecast deviations.In addition, forecast deviations and the reserved capacity of the battery storage system are subject to uncertainty that must be considered in the scheduling problem.This work presents a solution to this multi-use scheduling problem using chance-constrained optimization.The chance constraints ensure balancing forecast deviations with high probability, i.e., considering uncertainty, while maximizing day-ahead revenues.An evaluation is performed based on open real solar farm data and market price data.Comparing single-use (revenue maximization only) and multi-use scheduling for an exemplary setup shows that average unbalanced forecast deviations decrease by 84 % if dayahead market revenues are reduced by only 4.6 % on average.These promising results highlight the ability of large-scale battery storage systems to be used for multiple purposes simultaneously.