World Model for Battery Degradation Prediction Under Non-Stationary Aging

Fuente: Zenodo
Guardado en:
Detalles Bibliográficos
Autores principales: Lim, Kai Chin, SEE, KHAY WAI
Formato: Recurso digital
Lenguaje:inglés
Publicado: Zenodo 2026
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866902141088563200
author Lim, Kai Chin
SEE, KHAY WAI
author_facet Lim, Kai Chin
SEE, KHAY WAI
contents <div> <div>Degradation prognosis for lithium-ion cells requires forecasting the state-of-health (SOH) trajectory over future cycles. Existing data-driven approaches can produce trajectory outputs through direct regression, but lack a mechanism to propagate degradation dynamics forward in time. This paper formulates battery degradation prognosis as a world model problem, encoding raw voltage, current, and temperature time-series from each cycle into a latent state and propagating it forward via a learned dynamics transition to produce a future trajectory spanning 80 cycles. To investigate whether physics constraints improve the learned dynamics, a monotonicity penalty derived from irreversible degradation is incorporated into the training loss. Three configurations are evaluated on the Severson LiFePO4 (LFP) dataset of 138 cells. Iterative rollout halves the trajectory forecast error compared to direct regression from the same encoder. The physics constraint improves prediction at the degradation knee without changing aggregate accuracy.</div> </div>
format Recurso digital
id zenodo_https___doi_org_10_5281_zenodo_18952631
institution Zenodo
language eng
publishDate 2026
publisher Zenodo
record_format zenodo
spellingShingle World Model for Battery Degradation Prediction Under Non-Stationary Aging
Lim, Kai Chin
SEE, KHAY WAI
<div> <div>Degradation prognosis for lithium-ion cells requires forecasting the state-of-health (SOH) trajectory over future cycles. Existing data-driven approaches can produce trajectory outputs through direct regression, but lack a mechanism to propagate degradation dynamics forward in time. This paper formulates battery degradation prognosis as a world model problem, encoding raw voltage, current, and temperature time-series from each cycle into a latent state and propagating it forward via a learned dynamics transition to produce a future trajectory spanning 80 cycles. To investigate whether physics constraints improve the learned dynamics, a monotonicity penalty derived from irreversible degradation is incorporated into the training loss. Three configurations are evaluated on the Severson LiFePO4 (LFP) dataset of 138 cells. Iterative rollout halves the trajectory forecast error compared to direct regression from the same encoder. The physics constraint improves prediction at the degradation knee without changing aggregate accuracy.</div> </div>
title World Model for Battery Degradation Prediction Under Non-Stationary Aging
url https://doi.org/10.5281/zenodo.18952631