Recursive ARMAX-Based Global Battery SOC Estimation Model Design using Kalman Filter with Optimized Parameters by Radial Movement Optimization Method

Abstract

Precise estimating battery state-of-charge (SOC) is an important factor in determining vehicle range in an electric or hybrid vehicle. However, the parameters of battery are highly dependent on environmental conditions such as temperature. One of the main drawbacks in battery SOC estimation methods by using existing approaches is inability to adapt to the variable environmental operating conditions. In this study, a hybrid technique which combines the Kalman filter and recursive autoregressive exogeneous moving average model is proposed for the online battery parameters estimation. To design robust SOC estimation to varying battery parameters, the parameters of Kalman filter are optimized with the radial movement optimization metaheuristic algorithm. The proposed approach is implemented considering both different temperatures (0 °C, 25 °C, and 40 °C) and driving test cycles, UDDS, LA-92, and US06. Second-order RC battery equivalent model-based approach are compared with the proposed method, and the result shows that the proposed method is better than the conventional method in the aspects of average statistical parameters for all the driving test cycles.