An Online Detection Method of Short Circuit for Battery Packs

Abstract

Short circuit (SC) is a stumbling block to battery safety. The common battery management system (BMS) holding the fixed threshold focuses overly on the absolute magnitude of battery voltage, and therefore cannot detect the early SC. This paper proposes an online method for detecting SC based on principal component analysis (PCA), which possesses an adaptive threshold. First, we analyze the fault signatures of SC. Then, utilizing PCA, we generalize the adaptive threshold and design a detection flow. Finally, the effect of this method is experimentally verified. The results indicate that a early SC with the discharge current of 0.17C can be identified within 2s.