RNN and CNN Based Ensemble Models for State-of-Health Prediction of Li-Ion Batteries

Accurate prediction of the State-of-Health (SoH) of lithium-ion batteries is essential for optimizing performance and preventing failures in critical applications such as electric vehicles and energy storage systems. This paper presents an ensemble model integrating Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN), including Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRU), for enhanced SoH prediction. The ensemble approach leverages CNN's ability to capture spatial features and RNN's strength in handling temporal dependencies, providing a more comprehensive model for battery health monitoring. Using a NASA battery dataset, the proposed model outperforms traditional methods, achieving significant improvements in prediction accuracy, with up to 1.60% MAPE. This hybrid framework offers a promising solution for improving the reliability and efficiency of Battery Management Systems (BMS), enabling better decision-making for battery life cycle management. The results suggest that deep learning-based ensemble models can be a robust approach for SoH estimation, with potential applications across diverse industries.