Cunzhi Zhao developed this program. Xingpeng Li supervised this work. This work is licensed under the terms of the Creative Commons Attribution 4.0 (CC BY 4.0) license.
What is sample data in a battery aging test?
Sample Data (Data_1067_Battery_Aging_Test.xlsx) inlcude 1067 (rows) groups of battery aging tests with different SOC, Temp, DOD and DC per test. The 1067 is split to 889 groups of training dataset and 178 groups of validation dataset. All the data are normalized. In this case, 1 represests the fully capacity.
Note that the state of charge (SOC) can be tuned inside the setting of "Battery". You can simiulate the battery degradation by the BatteryTesting100.slx itself for certain setups of (Temp DOD and DC). The Matlab file (BatteryMain.m) is able to simulate several groups of battery aging tests and record the data to the excel.
How to simulate a battery aging test in MATLAB?
The Matlab file (BatteryMain.m) is able to simulate several groups of battery aging tests and record the data to the excel. Sample Data (Data_1067_Battery_Aging_Test.xlsx) inlcude 1067 (rows) groups of battery aging tests with different SOC, Temp, DOD and DC per test.
What are the ageing tests for Li-ion batteries?
This table covers ageing tests for Li-ion batteries. It is made in the European projects eCaiman, Spicy and Naiades. 7.6.1 Storage tests - Charge retention test. 7.5 SOC loss at storage / 7.4 No-load SOC loss. 7.6 SOC loss at storage / 7.5 No load SOC loss.
Some of the capacity of the battery aging test starts from 0.98/0.97, these are due to the low ambient temperature effects. Each cell represets a charging/discharging cycle. Degradation for each cycle is calacuted by the difference between thecapacity of the current cell and previsous cell.
Are battery aging datasets a problem in data science?
Battery aging datasets are not immune to the issues faced by the data science community, such as a lack of data or poor data quality. In fact, data gathering and data cleaning have grown to take a significant role in data science, as it is important to have high-quality data before building a data-driven model.