Artificial Intelligence-Based Hardware Fault Detection for Battery Balancing Circuits

A battery balancing circuit is a key component of a battery management system (BMS) that ensures safe and reliable operations of the multicell battery where imbalanced cell states are present, specifically as more battery cells are aged or eXtreme fast charging (XFC) is adopted. This paper explores how to apply artificial intelligence (AI) methods on measured battery cell data from a BMS to detect a defective battery balancing circuit. Several AI algorithms are evaluated by simulation and experiments using a real BMS hardware. Among them, Convolutional Neural Network (CNN) model-based detection meets the highest accuracy and about 95.31% of F-1 score.