Multi-Fault Diagnosis of Lithium-Ion Battery Systems Based on Correlation Coefficient and Similarity Approaches

The continuous occurrence of lithium-ion battery system fires in recent years has made fault diagnosis a current research hotspot. For series connected pack, the each cell is same. Although there are differences parameters such as internal ohmic resistance, relative change between cells small. Therefore, correlation coefficient voltage signals different can detect faulty cell. Inspired by this, this paper proposes an improved Euclidean distance method and cosine similarity for online multi-fault packs, compares them with method. sensor positions arranged according to interleaved measurement design. involved study, including connection faults, short-circuit faults external will lead abnormal readings at positions, which turn cause changes coefficient, achieve detection. Fault experiments were conducted verify feasibility three methods pack.