An Ensemble Model Based on Fusion of Multiple Machine Learning Algorithms for Remaining Useful Life Prediction of Lithium Battery in Electric Vehicles

As the demand for renewable energy solutions increases, electric vehicles (EVs) have become a critical component of sustainable transportation. Lithium-ion batteries, the core of EVs, determine vehicle performance and efficiency. Accurate Remaining Useful Life (RUL) prediction of these batteries is essential for effective battery management, reducing unexpected failures, and supporting sustainability goals. Traditional RUL prediction methods often fail to capture the complex degradation processes of batteries. To address these limitations, we propose a novel machine learning framework combining Artificial Neural Networks (ANN) for feature extraction and ensemble modeling with Random Forest (RF), K-Nearest Neighbors (KNN), and Gradient Boosting Decision Tree (GBDT). The final ensemble fusion further refines predictions by leveraging the complementary strengths of the models. Experimental results demonstrate that the proposed framework significantly improves prediction accuracy, with the final ensemble model achieving an R2 of 0.92 and reducing MAE and RMSE to 74.52 and 110.07, respectively. While promising, the framework faces challenges in real-world generalization and computational efficiency, highlighting the need for further research.