Adaptive State-of-Charge and State-of-Health Estimation Using Dual Kalman Filtering Techniques

Abstract

Proper estimation of State-of-Charge (SOC) and State-of-Health (SOH) is of the essence in the efficient functionality of battery systems of energy storage, its safety, and long durability, especially in use in electric vehicles, renewable energy integration, and smart grid infrastructures. The available charge in the battery is shown by SOC, whereas SOH is required to show the degree of ageing and degradation of the battery, which is a vital indicator when it comes to dealing with the battery. Nonlinear battery dynamics, parameter variations and measurement noise however tend to diminish the accuracy of conventional estimation methods. Moreover, most of the currently available methods estimate SOC and SOH separately, without considering the interplay of the two, which puts a constraint on general estimation accuracy. This paper is aimed at solving these problems by proposing a Kalman filtering structure that is adaptive by combining the estimation of SOC with SOH. The suggested algorithm uses two parallel Kalman philtres where one is used to determine SOC and the other one is used to determine battery parameters that are related to SOH. Moreover, a dynamic adjustment of the covariance tuning technique is proposed to dynamically adapt the noise statistics depending on the estimation residuals to increase the robustness of this technique in different operating conditions. Simulation works prove that the suggested approach is characterised by a better estimation accuracy, higher convergence speed, and stability as compared to traditional methods. The designed architecture can be used in real-time to manage high-tech batteries.