Research on Optimal Scheduling Strategies of Energy Storage Systems for Peak Shaving in Power Systems with High Renewable Energy Penetration

Abstract

With the rapid growth of wind and solar, modern power systems face widening peak–valley gaps and variability that traditional dispatch cannot absorb. This paper presents a comprehensive review of energy storage for peak shaving under high renewable penetration. We synthesize modeling paradigms, including deterministic, stochastic, robust, and model predictive control, combined with degradation-aware, network-constrained, and carbon-cost formulations. This paper examines scheduling across multiple horizons: real time, intraday, day ahead, and seasonal; and across architectures such as utility-scale batteries, behind-the-meter fleets, and hybrid storage. The review catalogs evaluation metrics such as peak reduction, curtailment, reliability, cost, and emissions, and compares representative case studies. We assess implementation challenges, including forecast uncertainty, battery aging, coordination of heterogeneous assets, market rules, and cyber-physical limits at the distribution level. We identify research gaps in uncertainty-aware data-driven dispatch, valuation of stacked services, and scalable coordination of distributed storage. The contribution is a structured synthesis and a research agenda to guide more flexible, economical, and low-carbon grid operation, informing planners and regulators where storage delivers the highest marginal system value.