cell testing in the TWAICE battery research center
Research Papers

Modeling capacity fade of li-ion batteries

Aging models are fundamental tools to optimize the application of lithium-ion batteries. In this work, we show that the CAP-method models capacity fade more accurately when applied to dynamic cyclic aging tests with periodically changing mean state-of-charge, depth-of-discharge, ambient temperature and discharge rates for a commercial NCA cell with a silicon-doped graphite anode.

TWAICE / Jul 13, 2022

twaicetech

TWAICE helped me to learn more about: Modeling capacity fade of li-ion batteries read article here:

www.twaice.com/research/modeling-capacity-fade-of-li-ion-batteries

#thinktwaice

Modeling capacity fade of lithium-ion batteries during dynamic cycling considering path dependence

Alexander Karger, Leo Wildfeuer, Deniz Aygül, Arpit Maheshwari, Jan P. Singer, Andreas Jossen.

Highlights

- Two methods for modeling capacity fade during dynamic operation are compared.

- Current capacity instead of charge-throughput as reference point is more accurate.

- Experimental validation with commercial NCA cell with silicon-doped graphite anode.

- Non-commutative capacity fade revealed as insufficient sign of path dependence.

Abstract

Aging models are important tools to optimize the application of lithium-ion batteries. Usually, aging models are parameterized at constant storage or cycling conditions, whereas during application, storage and cycling conditions can change. In the literature, two different methods for modeling capacity fade during such dynamic operation are proposed. These methods use either the cumulated charge-throughput (CCT-method) or the current capacity (CAP-method) as reference points, when aging conditions are changing.

In this work, we show that the CAP-method models capacity fade more accurately when applied to dynamic cyclic aging tests with periodically changing mean state-of-charge, depth-of-discharge, ambient temperature and discharge rates for a commercial NCA cell with a silicon-doped graphite anode. However, in cases where the difference between actual and reference charge-throughput of the CAP-method becomes large, the capacity gradient is modeled more accurately with the CCT-method. Because the relative capacity fade error of the CAP-method is small at it with 6%, we assume that capacity fade behaves path-independently for the dynamic cyclic aging tests since the CAP-method assumes path independence through history independence.

Moreover, because the measured capacity fade is non-commutative, which is sometimes labeled path-dependent, we recommend to not consider non-commutative capacity fade as a definitive sign of path-dependent degradation.

Access the paper here.

Related Resources

February 14, 2023

Evaluation of transmission losses of various battery electric vehicles

Transmission losses in battery electric vehicles have compared to internal combustion engine powertrains a larger share in the total energy consumption and play therefore a major role. In this paper, three simulation models of the Institute of Automotive Engineering are presented.
December 2, 2021

Combining EIS and time-domain data

This work proposes a method to combine time-domain and frequency-domain measurement data for parameterization of RC elements by exploiting the full potential of the distribution of relaxation times (DRT).
September 2, 2021

Quantifiability of Cell-to-Cell Variations

Motivated by the question of why impedance variation is consistently reported to be higher than the variation of other parameters, we show that measuring inherent parameter variations caused by production tolerances is superimposed by effects of an imperfect measurement setup.