Online state of health estimation for seriesconnected LiFePO₄ battery pack based on differential voltage and inconsistency analysis

Low-complexity and accurate state of health estimation of series-connected batteries has always been a difficult problem to solve in a well-designed battery management system. Lithium iron phosphate battery has been widely used as energy storage carrier due to its better safety and longer cycle life. In this paper, we proposed an online state of health estimation method for lithium iron phosphate battery pack based on differential voltage and inconsistency analysis. According to the aging mechanism of lithium iron phosphate battery, the region capacity in differential voltage curve is extracted as health feature to establish an aging model. Furthermore, all four inconsistency cases that affect pack capacity are analyzed, and the calculation formulas are concluded relying only on three representative battery modules, thereby reducing the computational burden. The pack capacity is estimated to verify the feasibility of the proposed method at different aging cycles. The estimation errors of the state of health of eight in-pack battery modules are less than 2%, and further the maximum error for the battery pack is only 2.27% in all aging cycles. In addition, the maximum error of the state of health for the other battery pack is only slightly over 1% at different inconsistencies.