bei Google ScholarTitelangaben
Siva Dev, Rabin ; Sivasubramanyan, Akshay ; Wycisk, Dominik ; Oldenburger, Marc ; Moos, Ralf ; Schmidt, Jan Philipp:
Modelling of OCV Hysteresis.
2026
Veranstaltung: MODVAL 2026: 22nd Symposium on Modeling and Experimental Validation of Electrochemical Energy Technologies
, 10.-11.3.2026
, Lausanne Switzerland
Lausanne Switzerland
Lausanne Switzerland.
(Veranstaltungsbeitrag: Kongress/Konferenz/Symposium/Tagung
,
Poster
)
Abstract
Open circuit voltage (OCV) hysteresis refers to the path-dependent difference in equilibrium open circuit voltage between charge and discharge at identical states of charge (SOC). Anode materials such as graphite and silicon show a hysteresis of 30 mV and 200 mV respectively while cathode materials such as LFP shows up to 30 mV. Modelling this is crucial not only for accurate voltage prediction, but also for state estimation in EVs thereby affecting its range prediction and safety. OCV hysteresis can be modelled using operator- based approaches like with the Preisach model, or differential equation based empirical approaches like with the Plett model. For EVs, where the use of pseudo-2D (P2D) models are very prevalent, the Plett model is more compatible owing to the similar differential equation-based framework of P2D models. While the Plett model is fully empirical and requires extensive parameterisation, some physics-motivated inclusion was achieved in the modified Plett model where the shapes of hysteresis transitions are captured from the differential capacity of the system, thereby significantly reducing the extent of parameterization. However, the Plett model and all its derivatives are accurate only for hysteresis transitions of the first order and fail to account for complex behaviours observed during higher order reversals and nested current loops. Such scenarios are frequently encountered by EVs especially during multiple recuperations. The limitations of the model originate from the innate asymmetry and simplicity of the differential equations used. In this regard, our work proposes physically motivated hysteresis representations achieved by correcting the Plett model and by incorporating experimentally observed closed loop and nested loop voltage responses. While the proposed models capture first order transitions with comparable accuracy to the Plett model, they show significantly higher accuracies with higher order transitions and successive recuperation pulses.