Adaptive State and Parameter Estimation of Lithium-Ion Batteries Based on a Dual Linear Kalman Filter

The estimation of the states and parameters of each cell, like the State-Of-Charge (SOC) or the State-Of-Health (SOH) in a battery pack significantly influences the efficiency of the system. Commonly used methods, such as current counting, show failures in their estimation, based on the accuracy of the used current and voltage sensors or an unknown aging behaviour of the battery cells. Other adaptive methods to solve these problems require such a high degree of processing power that there is no commercial implementation possible on a Battery Management System (BMS). Therefore in a previous work, an adaptive algorithm was created and implemented on a BMS. This algorithm enables an efficient adaptive method for estimating the states and parameters of each battery cell in an electrical vehicle, based on a simple battery model. A Dual Kalman Filter (DKF), was modified for the operation on the BMS. The deployment of a linear state estimator improves the filter performance. This modification allows a precise state estimation and with the dynamic of the battery, the DKF reveals a precise parameter estimation. In a previous work, the accuracy of the state estimator was validated, based on different cell temperatures and different SOCs. Thereby the algorithm enables an estimating error of the SOC less than 1.5%. This paper focuses on the behaviour of the modified DKF, based on the state- and parameter estimation and on the ideal tuning of the algorithm. Hence the derivation of the modified DKF will be shown. Improvements and differences in order to other KF systems will be illustrated. The modelling of the process and measurement noise for both KFs will be discussed. Based on different driving-charging cycles, the filter behaviour in the state and parameter estimation will be compared to commonly used methods. A stable filter response to dynamic and less dynamic loads is the goal of the paper.