Battery connection fault diagnosis method based on enhanced voltage entropy and real vehicle data

The reliable operation of an electric vehicle (EV) is largely dependent on the integrity of the battery pack connections, which, due to the large number of connectors involved, are prone to connectivity failures, posing a serious threat to the safety of the entire vehicle. To this end, this study proposes a battery connection fault diagnosis method based on enhanced voltage entropy and real vehicle data. Firstly, based on the connection fault mechanism, a fault feature enhancement filtering strategy is introduced to achieve double filtering of redundant signals in the voltage signal in both frequency and time domains that are not related to connection faults; secondly, the Shannon entropy is used as a high mapping characterization measure to comprehensively describe the disordered nature of voltage fluctuations caused by connection faults; Finally, a new voltage entropy peak residence algorithm is proposed to dynamically accumulate and ablate the abnormal behaviors of historical information and real-time signals to accurately capture transient fault pulses while suppressing pseudo-feature interference triggered by operating condition fluctuations, forming a chain of evidence of faults with temporal continuity, realizing the accurate detection of connection faults and the protection against misdiagnosis. It is validated by ablation experiments and on real vehicle fault data of five different fault types. The results show that the proposed method significantly improves the detection accuracy and anti-interference ability, and provides a fast detection and high-precision identification solution for the diagnosis of real battery connection faults in EV.