Abstract
Lithium-ion battery energy storage systems (ESSs) occupy the majority share of cumulative installed capacity of new energy storage. Consistency of an ESS significantly affects its performance and efficiency. Thus, accurate consistency evaluation for ESSs is vital to the operation maintenance management. This article proposes an integrated framework of evaluating the consistency of battery groups and identifying the inconsistent battery packs. First, low-dimensional feature representations are learned from charge–discharge voltage curves by the approximate low-rank representation (ALRR), which can realize the dimension reduction and also preserve the spatial structure relationships in the low-dimensional subspaces. Second, a hypersphere concentration-based consistency evaluation index for battery groups is constructed with the von Mises–Fisher distribution and optimal representations. Third, the inconsistent battery packs with voltage anomalies are rapidly identified by the hypothesis testing of chi-square statistics. Finally, the proposed method is validated based on real datasets of a battery ESS.
Funding Information
  • National Natural Science Foundation of China (72001138, 72071127, 72471143)

This publication has 32 references indexed in Scilit: