State of Health estimation algorithm of LiFePO4 battery pack based on differential voltage curves for BMS application

This paper discusses a novel differential voltage curve capacity estimation to determine the state of health of LiFePO4 cells. Differential voltage curves are used because of their ability to detect and quantify degradation mechanisms. The estimation is carried out through partial charging or discharging tests, and is specifically designed for battery management systems, due to the trade off between accuracy and low computational effort. This means the method can be effectively executed online, in a real application. The technique is also able to accurately detect the end of life of the cells.
Aging datasets of 18 cells with identical chemistry were used for both parametrization and validation. The cells were subjected to a wide range of cycling and storage conditions, including temperature, state of charge, charging and discharging rate, depth of discharge and state of health. The performance and robustness of the estimation are validated by means of the degradation datasets from more than 25 different scenarios at the cell and battery pack level. The related results indicate that the proposed health management strategy has an average relative error of 1.5% at the battery pack level.