Electrode ageing estimation and open circuit voltage reconstruction for lithium ion batteries

Open circuit voltage (OCV) test is an effective way of ageing diagnosis for lithium ion batteries and it constitutes a basis for state of charge (SOC) estimation. However, onboard OCV tests are rarely feasible due to the time-consuming nature. In this paper, we propose a method to estimate the results of offline OCV based ageing diagnosis, including electrode capacities and initial SOCs, termed electrode ageing parameters (EAPs). In this method, parts of daily charging profiles are sampled and directly fed into a convolutional neural network to estimate EAPs without feature extraction. Validation results on eight cells show that the estimated EAPs are very close to those obtained by using offline OCV tests. Therefore, this method enables a fast ageing diagnosis at an electrode level. Furthermore, we can use the estimated EAPs to reconstruct OCV-Q (charge amount) curves of batteries at different ageing levels over the entire battery life. The error for the OCV-Q reconstruction is within 15 mV compared with actual OCV-Q curves. Based on the OCV-Q curves, we show that battery capacity can be accurately obtained with an error of less than 1% although it is not explicitly considered as a target. The influence of voltage ranges on estimation results is also discussed.

Automated summary

The proposed method in this paper utilizes offline OCV test results to estimate aging diagnosis of lithium ion batteries at an electrode level, achieving fast diagnosis. The estimated aging parameters are close to the results obtained by offline tests, enabling reconstruction of OCV-Q curves for battery capacity estimation with high accuracy. The influence of voltage ranges on estimation results is also discussed in the study.