LAUSR

A state-of-charge (SOC) balancing method which accounts for state-of-health (SOH) status of battery cells is presented in this article. The data collected from aging experiments conducted in the laboratory indicates that there is correlation between the minimum impedance of the battery and capacity fading which can be used to predict capacity capability of the battery. However, this relationship is complex and nonlinear. In this article, artificial neural network (ANN) is utilized to learn this relationship and an ANN-based capacity estimator is developed to predict available capacity of battery cells. An online impedance measurement method with improved measurement resolution is presented to obtain a more accurate minimum impedance for the ANN-based capacity estimator. The estimated capacity from the ANN capacity estimator is fed to an SOC balancing controller to calculate SOC values for the battery cells. A battery cell with worse health has lower available capacity than a battery cell with better health, and, therefore, its SOC value is adjusted to a smaller value when SOH or capacity estimation functionality is activated. Because of this mechanism and the control principle of the presented SOC balancing controller, the system draws energy at a slower rate from the battery cell(s) with lower SOH and draws energy at a faster rate from the battery cell(s) with higher SOH such that all battery cells reach their end of discharge at the same time. This results in extending operation time of the system, makes best use of energy from every battery cell, and avoids over-discharging battery cells with lower SOH. The presented method is evaluated using results obtained from a laboratory experimental setup.