Multi-Fault Diagnosis of Lithium-Ion Batteries Using DTW-Transformed EIS Data and CNN

Le résumé fourni par la source

Ensuring the safety and reliability of battery systems requires effective fault diagnosis. However, conventional fault diagnosis methods based on equivalent circuit models (ECM) have inherent limitations in quantitatively assessing pattern variations between normal and abnormal states. To overcome these limitations, this study utilizes electrochemical impedance spectroscopy (EIS) data, which are processed using a dynamic time warping (DTW)-based recurrence plot (RP) transformation and subsequently analyzed through a convolutional neural network (CNN) model. Building on this approach, a multi-fault detection framework is proposed to enable real-time identification and classification of deviations between normal and abnormal conditions. Experimental results indicate that the proposed method significantly improves diagnostic accuracy, achieving a mean squared error (MSE) of 0.003496% and a root mean squared error (RMSE) of 0.059131%.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.