A Combined Model-Based and Data-Driven Multi-Fault Diagnosis Method for Series–Parallel Battery Packs

Multifault diagnosis is of great significance for the safe operation of lithium-ion battery (LIB) systems. A combined model-based and data-driven multifault diagnosis method based on an improved nonredundant cross-over voltage measurement topology is proposed for series–parallel connected LIB systems in this article. First, a reference module is selected and dual H-infinity filter is employed to estimate its state of charge (SOC). Rint model is adopted to evaluate the resistance inconsistency of modules. Second, Pearson correlation coefficients (PCCs) are calculated to detect faults considering inconsistencies in SOC and resistance between modules. The amplitude of auxiliary square wave is automatically determined based on the current variation. Standard deviation of voltage difference is employed to determine the threshold of PCCs. Finally, cell open-circuit (COC) faults, connection (CXN) faults, short-circuit (SC) faults and voltage sensor (VS) faults are comprehensively diagnosed. CXN faults and VS faults are identified according to the improved measurement topology and PCCs. Then, SC faults and COC faults are identified based on PCCs and fault modeling. The proposed method can also diagnose multiple faults that occur simultaneously or successively. The performance of the proposed diagnostic method is validated through experiments on a 2P-6S LIB pack.