结合数据驱动与等效电路模型的无线电力传输系统互感及负载识别

The variations of mutual inductance and load conditions affect the performance of wireless power transfer (WPT) systems. Identification of these parameters will be helpful for system control and condition monitoring. Circuit models are normally adopted in the existing WPT parameter identification methods, while the identification accuracy is easily impacted by the circuit parameter errors. In this article, a WPT mutual inductance and load identification method combining circuit and data-driven models is proposed. It has the advantage of a data-driven model that is not easily affected by parameters and a circuit model that is straightforward. Meanwhile, it can achieve accurate parameter identification only using the WPT dc input current and one voltage rms value without wireless communication. First, support vector regression (SVR) is adopted to establish the WPT data-driven model, and the mutual inductance identification algorithm is discussed. Then, parameter relationships are obtained from the WPT circuit model, considering the rectifier’s equivalent input impedance. Furthermore, the load identification method is presented based on the mutual inductance identification result. Finally, a WPT experimental prototype is built, and the experimental results show that the maximum identification errors of mutual inductance, load resistance, and load voltage are 2.61%, 4.10%, and 3.96%, respectively. They indicate that the proposed method can achieve high identification accuracy under the conditions of WPT mutual inductance and load variations.