Prediction of the State of Health (SOH) of Lithium-Ion Batteries in Electric Vehicles Based on Temporal Convolutional Networks (TCN) with Integrated Attention Mechanisms

The State of Health (SOH) is a crucial indicator for ensuring the stable and reliable operation of lithium-ion batteries in electric vehicles. Real-time and accurate prediction of SOH is an effective monitoring method to ensure the normal functioning of electric vehicles. This paper focuses on predicting the SOH of lithium-ion batteries in electric vehicles, based on Temporal Convolutional Networks (TCN) and attention mechanisms. Initially, health factor data from electric vehicle lithium-ion batteries are extracted and processed. Subsequently, a prediction model for the SOH of these batteries is developed, utilizing Temporal Convolutional Networks (TCN), and its predictive performance is further enhanced through the integration of attention mechanisms. Finally, the prediction model is further optimized using Bayesian optimization. Through discussion and analysis of the prediction results, it is found that the Mean Absolute Error Percentage (MAEP) of the SOH prediction for electric vehicle lithium-ion batteries is less than 5.00%, thereby demonstrating the feasibility of the method.