Simple Computational Battery Aging Models for Heavy-Duty Electric Vehicle Applications

Abstract

In contrast to liquid fuels and conventional internal combustion engines, in electric vehicle applications, the power source, i.e. battery, degrades with time. The mechanisms of battery degradation are typically divided into cycle aging (resulting from usage) and calendar aging (resulting from storage) [6]. The parameters primarily affecting battery degradation have also been widely documented in the academic literature (e.g. [1, 4, 5, 2]). They are: temperature, average state of charge (SoC), variation of SoC, C-rate current, time and number of full equivalent cycles.For many battery-powered heavy-duty vehicle (BHDV) applications, such as harvesters or mining dumpers, the above battery usage and storage parameters can be reasonably well estimated during system design. Consequently, a BHDV designer can attempt at model-based optimization of the battery system with respect to performance and cost, while taking into account performance degradation and the expected lifetime. Unfortunately the effect of the above parameters on battery degradation is different for each cell chemistry. Moreover, there appears to be no one widely adopted approach for numerically modeling and simulating battery aging (see e.g. [1, 7, 3] and the references therein). Practical engineering design necessitates, in particular, simulation methods that are simple and quick to use: robust low-parameter solutions requiring a minimal amount of experimental data for identification.The purpose of this article is to identify simple computational battery aging models that can be used for optimization of battery systems in BHDV applications. We describe minimal models, based on statistical techniques such as stepwise regression, for the most prevalent cell chemistries used in BHDV applications. The statistical techniques employed herein enable systematic exclusion of spurious variables from the models, and provide data-based justification for the chosen functional form.As an application, we carry out model-based cell chemistry optimization to maximize the lifetime of a battery on a real heavy-duty vehicle using real-world cycle data. We demonstrate the aging models through an easy-to-use and open access web interface.[1] Carnovale, A. and Li, X. (2020). A modeling and experimental study of capacity fade for lithium-ion batteries. Energy and AI, 2:100032.[2] de Hoog, J., Timmermans, J.-M., Ioan-Stroe, D., Swierczynski, M., Jaguemont, J., Goutam, S., Omar, N., Van Mierlo, J., and Van Den Bossche, P. (2017). Combinedcycling and calendar capacity fade modeling of a nickel-manganese-cobalt oxide cell with real-life profile validation. Applied Energy, 200:47–61.[3] Dubarry, M., Qin, N., and Brooker, P. (2018). Calendar aging of commercial li-ion cells of different chemistries–a review. Current Opinion in Electrochemistry, 9:106–113.[4] Naumann, M., Schimpe, M., Keil, P., Hesse, H. C., and Jossen, A. (2018). Analysis and modeling of calendar aging of a commercial LiFePO4/graphite cell. Journal of EnergyStorage, 17:153–169.[5] Naumann, M., Spingler, F. B., and Jossen, A. (2020). Analysis and modeling of cycle aging of a commercial lifepo4/graphite cell. Journal of Power Sources, 451:227666.[6] Redondo-Iglesias, E., Venet, P., and Pelissier, S. (2018). Calendar and cycling ageing combination of batteries in electric vehicles. Microelectronics Reliability, 88:1212–1215.[7] Schmalstieg, J., Käbitz, S., Ecker, M., and Sauer, D. U. (2014). A holistic aging model for Li (NiMnCo) O2 based 18650 lithium-ion batteries. Journal of Power Sources, 257:325–334.