Learning to fuse: dynamic integration of multi-source data for accurate battery lifespan prediction

2025-04-25

AI Summary

To address the low prediction accuracy and poor interpretability in lithium-ion battery remaining useful life (RUL) estimation, this paper proposes a hybrid learning framework integrating dynamic multi-source data fusion and stacked ensemble modeling. Methodologically, we design an entropy-driven dynamic weighted data fusion mechanism; introduce a novel three-level stacked model combining Ridge regression, LSTM, and XGBoost; and employ SHAP for quantitative attribution analysis of aging-critical indicators. Experiments on public benchmark datasets achieve MAE = 0.0058, RMSE = 0.0092, and R² = 0.9839—representing a 46.2% improvement in R² and an 83.2% reduction in RMSE over baseline methods. Crucially, the framework identifies discharge capacity differential (Qdlin) and mean temperature (Temp_m) as the two most influential aging indicators, thereby significantly enhancing both the accuracy and interpretability of state-of-health assessment and aging trend forecasting.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: State Estimation
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Accurate prediction of lithium-ion battery lifespan is vital for ensuring operational reliability and reducing maintenance costs in applications like electric vehicles and smart grids. This study presents a hybrid learning framework for precise battery lifespan prediction, integrating dynamic multi-source data fusion with a stacked ensemble (SE) modeling approach. By leveraging heterogeneous datasets from the National Aeronautics and Space Administration (NASA), Center for Advanced Life Cycle Engineering (CALCE), MIT-Stanford-Toyota Research Institute (TRC), and nickel cobalt aluminum (NCA) chemistries, an entropy-based dynamic weighting mechanism mitigates variability across heterogeneous datasets. The SE model combines Ridge regression, long short-term memory (LSTM) networks, and eXtreme Gradient Boosting (XGBoost), effectively capturing temporal dependencies and nonlinear degradation patterns. It achieves a mean absolute error (MAE) of 0.0058, root mean square error (RMSE) of 0.0092, and coefficient of determination (R2) of 0.9839, outperforming established baseline models with a 46.2% improvement in R2 and an 83.2% reduction in RMSE. Shapley additive explanations (SHAP) analysis identifies differential discharge capacity (Qdlin) and temperature of measurement (Temp_m) as critical aging indicators. This scalable, interpretable framework enhances battery health management, supporting optimized maintenance and safety across diverse energy storage systems, thereby contributing to improved battery health management in energy storage systems.

Problem

• Accurate prediction of lithium-ion battery lifespan
• Dynamic integration of multi-source data for fusion
• Mitigating variability across heterogeneous battery datasets

Innovation

• Dynamic multi-source data fusion for prediction
• Stacked ensemble with Ridge, LSTM, XGBoost
• Entropy-based weighting mitigates dataset variability