Analyzing and Predicting the Relationship Between the Coupling Capacitance and the Plates Misalignment Based on SSA‐CNN Method in Dynamic Charging of EVs

ABSTRACT Electric vehicles (EVs) have become increasingly popular among new energy vehicles. Therefore, the cost of charging devices and the efficiency of power conversion are major challenges to the large‐scale development of EVs. Capacitive power transfer (CPT) technology uses an electric field for transmission, which is low‐cost and insensitive to metal objects. Usually, the vehicle chassis is made of metal. It is relatively easy to deal with the chassis when we use CPT. To analyze and predict the relationship between the four‐plate capacitive coupler (FPCC) lateral and longitudinal misalignment, vertical distance variation, and the coupling capacitance, a prediction model of the coupling capacitance is proposed, based on the Sparrow Search Algorithm optimized convolutional neural network (SSA‐CNN). Firstly, replace the four‐plate coupling structure with the six‐capacitor model, derive the π‐shape model based on Kirchhoff's current equation, and transform the π‐shape model into the T‐shape model through the star‐delta transformation method. Next, the coupling capacitances for the plates misalignment are simulated utilizing Maxwell. Finally, through theoretical analysis and experimental validation, it is demonstrated that the proposed prediction model in this paper can predict the relationship between C N and the plates dynamic misalignment within a 15% error in the given range.