On-line lithium-ion battery state of health estimation using aging-related impedance identification with optimization

This paper presents a method for on-line lithium-ion battery (LiB) internal impedance identification for accurate state of health (SOH) and state of charge (SOC) estimation. The proposed method extracts values of internal impedance caused by battery aging, using optimization linked to an extended Kalman filter (EKF). In a long update cycle, elements in an equivalent electrical circuit model are optimized during the storage of the logged measured data to identify fixed internal impedance that relates to the aging effect. An EKF process for optimization runs in each iteration to calculate a multi-objective function that effectively makes use of process parameters for the EKF. In simulations, the proposed method achieved internal impedance identification accuracy of 0.0450 and 0.0495 RMSE error under, respectively, square and triangle current patterns, as opposed to 2.54 and 5.81 with a conventional method. Experimental evaluations for actual applications, using a square wave of a LiB yielded a result of 0.0517 as opposed to 0.660 with the conventional method.