Abstract
To evaluate battery conditions and predict the remaining operational lifespan of a battery, traditional techniques use empirical rules or model-based approaches based on electrochemistry. However, the complex and non-linear degradation patterns observed under diverse real-world operating conditions are frequently not captured by such empirical rules or model-based approaches. This disclosure describes cloud-based battery-health diagnostics and prognostics techniques that leverage machine learning (ML) to provide accurate and reliable assessment of the state of health (SOH) of the battery and predict its remaining useful life (RUL) across diverse battery chemistries, applications, and operating conditions. The cloud-based architecture aggregates data from large battery fleets, allowing scalable, adaptive, and cost-efficient diagnostics without burdening local battery management systems. Continuous learning improves accuracy as models adapt to new chemistries and usage patterns. Applicable to electric vehicles, energy storage systems, consumer devices, etc. the techniques support predictive maintenance, performance optimization, and end-of-life planning, enhancing reliability, efficiency, and sustainability across battery-dependent industries.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Yeh, Tim and Kang, Kang, "Cloud-Based Battery Health Diagnostics Using Machine Learning", Technical Disclosure Commons, (November 04, 2025)
https://www.tdcommons.org/dpubs_series/8833