Enhanced Diagnosis of Lithium-Ion Battery Health in Electric Vehicles Through Advanced Hybrid Deep Learning Model Incorporating Attention Mechanism

Abstract

With the rising adoption of lithium-ion batteries(LIBs) in electric vehicles, significant research efforts have been dedicated to ensuring the safe and reliable performance of Battery Management Systems (BMS). Accurately estimating the State of Health (SOH), a crucial aspect of BMS functionality, is imperative for the secure operation of electric vehicles. To tackle this challenge, we propose a time series model designed specifically for forecasting the State of Health (SOH) of Li-ion batteries. Our approach integrates a CNN with a Long Short-Term Memory (LSTM) architecture, augmented by an attention mechanism. Experimental findings demonstrate that this LSTM-based approach significantly enhances prediction accuracy. The outcomes of this investigation present a promising avenue for improving the dependability and efficiency of electric vehicle battery systems, ultimately advancing the widespread adoption and sustainability of electric transportation.