
Borealis
Duque, Josimar;
Phillip J. Kollmeyer;
Naguib, Mina
—
2023-09-25
This aging dataset was designed to be used for training/parameterization and testing of machine learning and conventional filter based state of charge and state of health estimation models. A number of characterization tests (HPPC, C/20 charge discharge, etc) are applied to each cell and are followed by repeating series of drive cycle discharges and a fifteen minute fast charge.
The characterization and drive cycle tests are repeated until the battery cells reach 70% SOH (around 1500 to 2000 cycles). The rate of aging for each cell is different because each cell is fast charged using a different method (standard CC/CV, boost charge - higher current at low SOC, and two pulsed charge methods). The four cells tested are brand new 3Ah Samsung INR21700-30T lithium ion battery cells. The testing was performed in a thermal chamber at 25 degrees Celsius using an Arbin battery cycler.