| Title |
Method of Selection Health Indicator according to the Charging Profile of Lithium-ion Batteries and Estimating State-of-health based on LSTM |
| Authors |
Dongcheol Lee ; Miyoung Lee ; Pyeongyeon Lee ; Jonghoon Kim |
| DOI |
https://doi.org/10.6113/TKPE.2024.29.3.233 |
| Keywords |
Health indicator; Pulse-constant voltage charging; Pearson correlation coefficient; Entropy weight method; Long short-term memory |
| Abstract |
Constant current(CC)-constant voltage(CV) charging, which is commonly used in various applications, has the advantage of simple implementation and low cost but is unsuitable for rapid charging technology. Accordingly, various rapid charging technologies are being developed, and many studies have emerged for their applications. However, rapid charging technology can cause rapid battery performance degradation and safety problems. To overcome this problem, estimating the health of the battery is important. To estimate the health condition within battery, the health indicators of the charging section are mainly used. However, given that the patterns of current, voltage, and temperature input data into the charging section are different for each charging profile, the health indicator selection method associated with state-of-health(SOH) must also be taken differently. In this paper, we propose a method of selection health indicator according to the charging profile and evaluate the health condition estimation performance according to each health indicator selection method using the long short-term memory(LSTM) model. |