Battery States Co-estimation Methodology Using Dual Square Root Unscented Kalman Filter

Abstract

Real-time and accurate estimation of battery internal states is immensely critical for emerging applications such as Electric Vehicles (EV), smart grids, and space applications. Model-based state estimation methodology provides highly robust and accurate battery state estimation. However, separate estimation of states, such as State-of-Charge (SOC), State-of-Health (SOH), and State-of-Power (SOP), leads to erroneous estimation since the states are highly interdependent. A co-estimation methodology for SOC, SOH, and SOP using a highly accurate and stable formulation of the Kalman filter, i.e., the Dual Square Root Unscented Kalman filter (D-SRUKF) is proposed in this paper. The proposed battery states co-estimation methodology has been validated using experimental battery test data. The results show that SOC estimation error is 0.404 %, with an improvement of 77.60% compared to separate state estimation using the D-SRUKF estimator and 58.02% compared to state-of-the-art EKF-RLS co-estimation methodology. SOH and SOP are also co-estimated within the same filter, leading to accurate estimation without adding to the computational complexity of the system. The accuracy of SOH estimation is improved by 16.98% compared to the EKF-RLS co-estimation.