A Novel Three-Stage Battery Cell Anomaly Detection Approach for a Frequency Regulation-Energy Storage System in Edge-Cloud Computing

Energy storage systems (ESSs) have increasingly become important, and an electrical grid upgraded as a smart grid with the widespread use of renewables and electric vehicles needs to be stabilized considering the grid's safety, stability and reliability requirements. In this paper, a new screening approach using three-stage battery cell anomaly detection is proposed. This approach more precisely quantifies the relative deterioration of battery cells, allowing battery cell outliers to be traceable during operation inside battery modules constituting battery racks in a (frequency regulation-)ESS. These outliers, which are represented by cell voltage imbalance to deteriorate over time, pose potential safety hazards and need to be targeted and prevented. The proposed approach is based on an edge-cloud computing framework. The progress of a conventional ESS suggests an advanced toward a next-generation ESS, where geographically distributed systems can be monitored and managed over edge-cloud computing in a distributed, decentralized fashion. The approach proposed in this research is a preliminary implementation that has been experimentally validated by an on-site, in-service FR-ESS.