Comprehensive review and benchmarking of empirical open-circuit voltage hysteresis models for lithium iron phosphate cells including a novel Modified Preisach approach

Titelangaben

Elender, Valentin Ludwig ; Ludwig, Sebastian ; Horstmann, Yannis ; Erhard, Simon ; Schmidt, Jan Philipp:
Comprehensive review and benchmarking of empirical open-circuit voltage hysteresis models for lithium iron phosphate cells including a novel Modified Preisach approach.
In: Journal of Power Sources. Bd. 697 (2027) . - 241460.
ISSN 0378-7753
DOI: https://doi.org/10.1016/j.jpowsour.2026.241460

Volltext

Link zum Volltext (externe URL):

Angaben zu Projekten

Abstract

Open-circuit voltage (OCV) hysteresis affects state estimation in lithium iron phosphate (LFP) graphite cells. This work presents a review and benchmark of empirical open-circuit voltage (OCV) hysteresis models within a framework that treats OCV hysteresis holistically. The Plett Zero-State and One-State model, the Huria model, the Roscher model, and a Modified Preisach model adapted to the proposed framework are considered. Validation uses a real-world profile derived from fleet data (>75,000 vehicles, >20 million charging events) and augmented with OCV hysteresis validation points. All models are evaluated on a commercially available 2.6Ah lithium iron phosphate (LFP) graphite cell for comparability. The Roscher model achieves the best performance, with an overall mean absolute error (MAE) of 1.585mV and a mean absolute percentage error (MAPE) of 4.7 for the OCV hysteresis validation points, while also showing the highest robustness against drift. The Modified Preisach, Plett One-State, and Huria model achieve comparable accuracy but exhibit higher sensitivity to drift. Error decomposition indicates that polarization errors dominate under load, whereas hysteresis modeling becomes critical during rest. Sensitivity analysis demonstrates that reducing parameter interpolation decreases drift and measurement effort without compromising accuracy. The results provide guidance for robust hysteresis modeling in automotive battery management systems (BMSs).