A Predictive-Prescriptive Framework for Portable Energy Storage Operation in Real-time Market

Portable Energy Storage System (PESS) represents a promising business model of energy storage with flexible deployment options. It has the potential to shape a low-carbon and sustainable energy and transportation system. In the energy arbitrage applications, however, it has been proved that using the PESS schemes determined by the known day-ahead market prices to participate in the real-time market will lead to significant revenue deviation. To tackle the above problem, we develop a predictive-prescriptive framework for PESS operation in realtime market, which incorporates the real-time market price prediction and inventory routing planning of PESS. For real-time market price prediction, we propose an error-corrected hybrid forecasting model based on NuralProphet and eXtreme Gradient Boosting (XGBoost). Regarding the inventory routing planning of PESS, we develop a mathematical model for PESS considering battery degradation. In the case study, we apply the proposed framework to assess the profitability of PESS in the real-time market of California Independent System Operator (CAISO) in 2018. The findings indicate that the predictive-prescriptive framework can substantially reduce average revenue deviation to 4.4%, and improves evaluation accuracy by 14.6% compared to conventional methods. The predictive-prescriptive framework is expected to provide decision supports for managing battery assets in applications coupled energy and transportation sectors