Hybrid Deep Learning with Attention Mechanism based Health State Intelligent Diagnosis of Lithium-Ion Batteries

Abstract

With the increasing utilization of lithium-ion batteries in electric vehicles, substantial research has been undertaken to ensure the secure and dependable performance of the Battery Management System (BMS). The precise estimation of State of Health (SOH), a pivotal function of the BMS, is critical for the safe operation of electric vehicles. Addressing this challenge, we present a time series model tailored for predicting the State of Health (SOH) of Li-ion batteries. Our approach combines a CNN with a Bidirectional Long Short-Term Memory (BI-LSTM) architecture, enhanced by an attention mechanism. Experimental results demonstrate that this enhanced LSTM approach notably enhances prediction accuracy. The results from this study offer a promising solution for enhancing the reliability and performance of electric vehicle battery systems, ultimately contributing to the widespread adoption and sustainability of electric transportation.