Optimal storage sizing of energy storage for peak shaving in presence of uncertainties in distributed energy management systems

The rapid development of eco-friendly technologies such as energy storage system (ESS) and peak-shaving technology in smart grid plays a significant role and shapes the future electricity consumption patterns. Distributed energy management system (DEMS) can be utilised to shave the peak load and reduce the users' electricity tariff. In this paper, a robust analytical method is presented to determine the size of ESS and its scheduling strategy. Firstly, extreme learning machine (ELM) and k-means algorithms are employed to classify customers into groups according to their characteristics. For each group, a support vector regression (SVR) model is developed for improving accuracy of load forecast. The whole storage system is then divided into schedule-based capacity and emergency capacity for different optimal objectives. A mixed integer linear programming (MILP) model considering the reliability constraints, peak-shaving requirement, and linearisation method is constructed to optimise the management of the DEMS. Verification and comparison studies demonstrate the effectiveness of the proposed scheme.