Enhanced SOC Estimation Performance Through Noise Covariance Analysis in Nonlinear Kalman Filters
Hari Prasad Bhupathi, Srikiran Chinta, Swarna Kumari Yeditha, Vijayalaxmi Biradar, L. Bhagyalakshmi, Sanjay Kumar Suman
Abstract
The effect of noisy matrix covariance on determination of SOC in batteries with lithium ion utilising not linear Kalman filter techniques is investigated in this technical paper. An essential part of electric vehicle (EV) battery management systems (BMS) is state-of-charge (SOC) estimate, which guarantees security as well as optimum performance. Batteries devices are inherently unpredictable, which calls for state-of-the-art estimate methods like Particles Filter, Extended Kalman Filter (EKF), and Unscented Kalman Filter (UKF). Kalman filter performance is shown to be dependent on tweaking the measurements as well as noisy covariance matrix values (Q and R), according to the study. Examining the effects of Q and R on converging periods and estimate errors via simulations highlights the compromise among accuracy and computing efficiency. To optimise Q and R values, the research suggests alternate tuning procedures. This will ensure that there is low estimate errors and that converging is enhanced. The effectiveness of the ECM-based EKF concept was shown in practical uses by validating it utilising discharges curve using batteries the datasheets at different present rates. When comparison to the traditional EKF, the suggested adaptive EKF achieves an overall accuracy that is around 89% higher. Its performance is superior to that of the UKF and PF, showing that the SOC estimate precision is much improved. The efficacy of adaptive tuning in enhancing the suggested method's effectiveness is demonstrated by this enhancement.