Battery Dual Extended Kalman Filter State of Charge and Health Estimation Strategy for Traction Applications

The growing market of electrified vehicles requires efforts from car manufacturers to build robust systems to deal with all types of situations their products will face in customers' hands. A major system of electrified vehicles is the energy storage unit. The complexity of batteries lies in their nonlinear behaviour that is highly dependent on external factors such as temperature and load dynamics. To handle these conditions, the battery management system relies on algorithms that estimate the state of the storage unit. State of charge (SOC) estimation, which is widely studied in industry and academia, is commonly considered one of the most significant functions of a battery management system (BMS). State of health (SOH) estimation is likewise important as it is necessary to support more consistent SOC and state of power (SOP) estimation. In this paper, a dual Extended Kalman Filter (DEKF) model is proposed to estimate the battery state of charge and capacity state of health across the battery lifespan. The DEKF model is demonstrated to accurately estimate SOC as the battery ages, with an average RMS error of 1.0% for SOH varying from 100% to 80%. The model is also shown to be robust against initial SOC and sensor error, demonstrating its applicability to real world conditions.