This article is concerned with the multilevel secure warning problem for lithium-ion battery systems (LBSs) subject to multi-fault scenarios. First, the long short-term memory (LSTM) network is employed to design a series of cell voltage estimators. In comparison with the estimation of the cell voltage and its measurement, the estimation residual is constructed in terms of the Euclidean distance (ED), which plays an essential index for the multi-fault detection issue. Furthermore, in combination with the voltage envelope relationship, a specific type of fault is subsequently identified. Then, for the subsequent fault warning purpose, a novel fault quantitative analysis method is proposed to further determine not only the fault amplitude but also its duration. In addition, on the basis of specific characteristics of the fault such as the type, count, duration, and amplitude, a fuzzy-logic-based multilevel secure warning strategy is put forward. Finally, the effectiveness of the developed multilevel fault warning approach is substantiated through comprehensive experimental validation.
Multi-Fault Diagnosis, Quantitative Analysis, and Warning Strategy for Lithium-Ion Battery System: When LSTM Meets the Fuzzy Logic Theory
IEEE Transactions on Transportation Electrification ; 11 , 3 ; 8224-8235
2025-06-01
2722739 byte
Article (Journal)
Electronic Resource
English
Vehicle power battery system fault diagnosis method based on LSTM-GAN
European Patent Office | 2023
|Model-Based Multi-Fault Diagnosis for Lithium-Ion Battery Systems
British Library Conference Proceedings | 2022
|Model-Based Multi-Fault Diagnosis for Lithium-Ion Battery Systems
SAE Technical Papers | 2022
|