State of Charge Estimation of Ultracapacitor Modules Based on Improved Sage-Husa Adaptive Unscented Kalman Filter Algorithm

In the field of new energy electric vehicles, ultracapacitor modules are often used as energy storage batteries. Precise estimation of state of charge (SOC) of ultracapacitor modules is critical to the secure operation of vehicle power supply. In this paper, equivalent circuit models and SOC estimation algorithms are compared and analyzed. The forgetting factor recursive least squares (FFRLS) is employed to supply precise equal circuit model parameters for SOC estimation algorithm. On this basis, an improved Sage-Husa adaptive unscented Kalman filter (IAUKF) online SOC estimation algorithm is proposed. The improved unscented Kalman filter algorithm solves the problems of poor robustness and large computational effort of the conventional Sage-Husa adaptive unscented Kalman filter algorithm (AUKF). The experiment verification is carried out in UDDS test and FUDS test respectively. The experiment verified that the SOC estimation error of IAUKF algorithm is less than 1.076%, and the average relative error is reduced by more than 50.574% compared with the conventional algorithms. The FFRLS-IAUKF joint SOC estimation algorithm has high estimation accuracy and good robustness.