March 3, 2026Open Access

Prédiction de durée de vie restante de batterieslithium-ion : une approche frugale à base de données

Q: What is the clinical evidence from this study?

Study design: Other. Population: lithium-ion battery degradation prediction (n=6). Intervention: Polynomial approximation model for battery capacity decay to estimate Remaining Useful Life (RUL) vs. No prediction or ground truth cycle count. Primary outcome: Accuracy of Remaining Useful Life (RUL) prediction defined by the prediction horizon N0.7 - Ntube10 in cycles.

Key Result

The polynomial approximation method predicted battery remaining useful life with an average prediction horizon of 227 cycles, representing over 30% of battery life remaining at prediction entry (within ±10% of true RUL).

Key Points

Remaining lifespan of lithium-ion batteries can be predicted using data-driven approaches, improving efficiency.
Key evidence shows a reliable prediction method based on limited data and resources can enhance battery management.
Analysis using data approaches enhances the understanding of battery lifespan predictions across various conditions.
These findings highlight the potential for more sustainable battery usage methods, driving future technological advancements.

PICO

Population

lithium-ion battery degradation prediction (n=6)

Intervention / Comparator

Polynomial approximation model for battery capacity decay to estimate Remaining Useful Life (RUL) vs No prediction or ground truth cycle count

Primary Outcome

Accuracy of Remaining Useful Life (RUL) prediction defined by the prediction horizon N0.7 - Ntube10 in cycles

Limitations

Small sample size of 6 batteries from a single dataset
Results apply only to lithium-ion batteries type CS2 with cobalt oxide cathode
The approach underestimates RUL in early cycles and overestimates in intermediate cycles before stabilization
No clinical trial or real-world clinical validation context

Abstract

International audience

Bookmark

View Full Paper

Bookmark

View Full Paper

Cite This Study

Gaignard et al. (Thu,) conducted a other in lithium-ion battery degradation prediction (n=6). Polynomial approximation model for battery capacity decay to estimate Remaining Useful Life (RUL) vs. No prediction or ground truth cycle count was evaluated on Accuracy of Remaining Useful Life (RUL) prediction defined by the prediction horizon N0.7 - Ntube10 in cycles. The polynomial approximation method predicted battery remaining useful life with an average prediction horizon of 227 cycles, representing over 30% of battery life remaining at prediction entry (within ±10% of true RUL).

synapsesocial.com/papers/69a76230c6e9836116a306fe