ABSTRACT Due to factors such as aging and manufacturing variations, the parameter of compensation components in the IPT system inevitably drifts. This drift can affect the accuracy of parameter identification, such as mutual inductance and load resistance, which could impact the precision of the output control. To address this issue, an offline multiparameter identification method based on a multiforest regression model is proposed. This method aims to proofread the parameters of the compensation components before charging, thereby reducing the identification errors caused by parameter drift in the compensation components. The LCC–LCC‐type IPT system is utilized as an example. First, the impact of the secondary‐side compensation components' parameter drift is analyzed; then, a multiforest regression model is introduced to proofread these parameters. By only measuring the effective value of voltage and current at multiple frequencies on the primary side, all the parameters of the secondary‐side compensation components can be estimated. Simulation and experimental results show that the proposed method achieves an average identification error of approximately 1.4% for the LCC–LCC IPT system, and the maximum identification error does not exceed 3% under various coupling and load conditions.
Liu et al. (Fri,) studied this question.