An online fault diagnosis method for lithium-ion batteries based on signal decomposition and dimensionless indicators selection

Accurate and timely battery fault diagnosis can effectively ensure the safe operation of lithium-ion battery systems. In response to the problems of high false alarm rates and low accuracy of fault diagnosis methods using only a single feature, a battery safety warning and fault diagnosis method based on multi-dimensional Dimensionless Indicators (DIs) is proposed. Firstly, the Sail Fish Optimizer (SFO) is used to determine the optimal parameter combination of the Variational Mode Decomposition (VMD) to realize adaptive VMD for battery voltage signals, and then the Optimal Mode combination (OMC) is calculated. Sliding windows are introduced to enable the online application of the method. Secondly, DIs are extracted from OMC. High-dimensional DIs are reduced using Self Organizing Maps (SOM) which is an unsupervised feature reduction method so that effective features that can be used for identifying faulty cells are screened out. Finally, an adaptive Local Outlier Factor (LOF)-based faulty cell identification method is proposed. The rules for selecting LOF adaptive thresholds are defined, and the adaptive setting of battery fault threshold parameters for different vehicle data is achieved. The method proposed in this paper is applied to three thermal runaway vehicles, and the faulty cells are identified before thermal runaway, which verifies the effectiveness and accuracy of the method.