European HYFLEXPOWER project to demo first integrated power-to-X-to-power hydrogen gas turbine
Duke researchers boost electrolyzer productivity with microfibrous flow-through electrode; 12.5- to 50-times greater than conventional

Using cloud charging data to estimate EV battery life

A team from the University of Shanghai for Science and Technology, with colleagues from Tsinghua University and China Electric Power Research Institute, have developed a method to estimate EV battery life based on cloud data. A paper on their work is published in the Journal of Power Sources.

Automobile companies typically retain driving data from their EVs on the cloud for use in monitoring and management. The recording period of such cloud data is generally of 10–30 s duration at present; this makes it difficult to reveal the dynamic driving condition of EVs with cloud data, the researchers explain. However, the charging data are stable, making it possible to estimate the battery life based on the charging cloud data.

Li

Capacity estimation process. Source: Li et al.


In order to better detect and evaluate the performance of the power battery after it is applied in EVs, it is crucial to find an effective life estimation method for the battery pack. Battery life estimation includes capacity estimation and internal resistance estimation.

Capacity is an important indicator of battery state of health (SOH) estimation to evaluate the aging degree of the battery. When the capacity degrades to a certain extent, the battery reaches the end of life (EOL) and cannot continue to work normally. Battery life estimation methods can be divided into empirical model-based and data-driven. Instead of considering the complex physical and chemical reactions inside the battery, the empirical model-based method estimates and predicts the capacity based on experimental data. Because the general empirical model is an open-loop model, it is difficult to ensure the accuracy of the estimation results. The data-driven method typically involves establishing models to describe the battery degradation process, which can achieve higher estimation accuracy, but it is limited by the complexity of electrochemical modeling in real application. Besides, the method generally requires certain working environment to estimate the battery life.

… Aiming at the problems mentioned above, a life estimation method for EVs battery pack based on cloud data is proposed in this paper.

—Li et al.

The researchers estimated capacity by the ampere hour integral method using a large amount of historical charging data. The estimation results were then modified based on the temperature data to obtain the preliminary capacity estimation results.

Then the Kalman filter (KF) algorithm was used to optimize the estimation results; the observation noise of KF algorithm was controlled by a Fuzzy logic (FL) algorithm in real time to improve the accuracy of estimation results. Based on the capacity estimation results, the battery life is predicted by the Arrhenius empirical aging model.

The sudden changes of voltage and current in the charging data were used to estimate the internal resistance of the battery pack. The internal resistance prediction is derived using a similar process to that for the capacity prediction.

Data from 147 EVs provided by two data sources were used to verify the method.

The researchers say that the method provides data support for preventive maintenance for battery packs, residual value assessment in EVs, and as well as the rapid classification before the secondary use.

Resources

  • Kai Li, Ping Zhou, Yifan Lu, Xuebing Han, Xiangjun Li, Yuejiu Zheng (2020) “Battery life estimation based on cloud data for electric vehicles,” Journal of Power Sources, Volume 468, 228192 doi: 10.1016/j.jpowsour.2020.228192.

Comments

SJC_1

With all the cars, computers and wireless
we should get a good picture of mileage.

The comments to this entry are closed.