.png){kind=icon}
.png){kind=icon}
Return
Tracking thermophysical drift improves latent thermal-runaway risk evaluation in lithium iron phosphate energy storage batteries under low-rate overcharge
.png){kind=icon}
PRE
AI
G
X
K
X
Q
P
Y
W
DOI：10.1016/j.ensm.2026.105422.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Detecting latent overcharge-induced instability prior to observable thermal runaway (TR) during low-rate overcharge in lithium iron phosphate energy storage batteries remains a critical challenge. Thermophysical-parameter drift and safety-boundary evolution can obscure early thermal evolution and render traditional static criteria insufficient. Here, we propose a risk-scoring assessment framework integrating a physics-informed neural network with adaptive filtering. By embedding a thermodynamic model as a physical constraint, this approach improves the robustness of parameter inversion under strongly nonlinear thermal conditions. It accurately reconstructs core temperatures to compensate for shell temperature observation lag. We further develop a dynamic risk evaluation scheme utilizing parameter uncertainty propagation. Validation at 0.5 C demonstrates that the core-temperature RMSE was 2.2 °C, reducing RMSE by 80.0% relative to static-adaptive extended Kalman filter. Comparing the 0.1 C to 0.5 C overcharge conditions, the normalized safe observation window was lower at 0.5 C (53.5%) than at 0.1 C (81.8%) of the overcharge duration preceding the 6 V experimental abuse-test cutoff. This work is expected to provide a physically interpretable basis for developing active safety management strategies in battery management systems for advanced energy storage applications.
Journal
IF:
20.2
Papers:
5.6K
Citations:
6.3W
Organization
Cited Papers