Diagnostic-free onboard battery health assessment

2025-03-10

AI Summary

Lithium-ion batteries exhibit complex aging behavior under dynamic operating conditions; conventional state-of-health (SOH) diagnostics rely on offline diagnostic cycles, which perturb degradation trajectories, incur significant time overhead, and hinder onboard deployment. This paper proposes a real-time, onboard SOH estimation and remaining useful life (RUL) prediction method that requires no offline testing or historical data. The core innovation is a mechanism-constrained, interpretable encoder–decoder architecture that directly maps electrode-level aging states onto a physically interpretable latent space—enabling diagnosis-free operation, online execution, and cross-condition generalization. By synergistically integrating electrochemical mechanism priors with operational data-driven modeling, the method is validated on a large-scale dataset comprising 422 battery cells across three representative operating conditions. Results show SOH estimation error <1.2%, degradation trajectory reconstruction RMSE <0.8%, and computational efficiency suitable for embedded real-time deployment.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Life-Long and Continual Learning
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Diverse usage patterns induce complex and variable aging behaviors in lithium-ion batteries, complicating accurate health diagnosis and prognosis. Separate diagnostic cycles are often used to untangle the battery's current state of health from prior complex aging patterns. However, these same diagnostic cycles alter the battery's degradation trajectory, are time-intensive, and cannot be practically performed in onboard applications. In this work, we leverage portions of operational measurements in combination with an interpretable machine learning model to enable rapid, onboard battery health diagnostics and prognostics without offline diagnostic testing and the requirement of historical data. We integrate mechanistic constraints within an encoder-decoder architecture to extract electrode states in a physically interpretable latent space and enable improved reconstruction of the degradation path. The health diagnosis model framework can be flexibly applied across diverse application interests with slight fine-tuning. We demonstrate the versatility of this model framework by applying it to three battery-cycling datasets consisting of 422 cells under different operating conditions, highlighting the utility of an interpretable diagnostic-free, onboard battery diagnosis and prognosis model.

Problem

• Enables onboard battery health assessment without diagnostic cycles
• Uses operational data and machine learning for accurate health diagnostics
• Applies interpretable model to diverse battery usage patterns

Innovation

• Uses operational measurements for diagnostics
• Integrates mechanistic constraints in encoder-decoder
• Applies interpretable machine learning model