A minor-fault diagnosis approach based on modified variance for lithium-ion battery strings

In order to prevent battery accidents of electric vehicles (EVs), it is significant to quickly diagnose and recognize faults of lithium-ion battery strings. Nevertheless, the diagnosis of initial minor faults is an intractable problem because the minor fault voltages often do not trigger the cut-off voltage and are hidden in the normal voltage sequences which pose a great threat to the safe operation of electric vehicles. Therefore, this paper proposes a simple real-time minor fault diagnosis approach based on modified variance. By calculating the modified vari-ance of the battery voltage sequences in a sliding-window, the proposed diagnosis approach is able to effectively discern the type and the time of minor battery abnormities, which comprises open circuit and short circuit faults. The purpose of the sliding-window is to take a sliding reading of the voltage sequence in a window with appropriate length in order to improve the accuracy of battery fault diagnosis. Then, the experimental results and the contrast with the previously existing approaches verify the feasibility of the proposed method with low computational cost, easy realization, and without the battery model. Particularly, it can be used to effectively diagnose the minor faults of the battery under different ambient temperatures because of strong robustness and sensitivity. In conclusion, the proposed minor fault diagnosis method has practicable applications in electric vehicles.