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

Abstract

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