Transmission Efficiency Prediction And Coil Parameter Estimation Of Magnetic Coupling Resonance System

In recent years,as a relatively safe and convenient non-contact power transmission technology,wireless power transmission technology has satisfied people's power requirements in various aspects,and has received more and more attention.Among them,the magnetic coupling resonance system has become a research hotspot in recent years due to its long transmission distance,high transmission efficiency,and small impact on the environment and organisms.The magnetic coupling resonance system has become a research hotspot in recent years due to its long transmission distance,high transmission efficiency,and small impact on the environment and organisms.In magnetic coupling resonance system,the influence of coil parameter design on system performance is particularly important.To achieve the desired effect,the coil parameters need to be adjusted continuously.In order to quickly find the coil parameters that meet the conditions,the following research is performed in this study.First,through theoretical analysis of the circuit principle of the magnetic coupling resonance system,the factors and system parameters that affect the transmission efficiency of the magnetic coupling resonance power transmission system are determined.According to the analysis,the simulation model is established based on Maxwell and HFSS(High Frequency Structure Simulator),and an experimental device is made to analyze the influence of each parameter on the transmission efficiency.And the raw data of the neural network model used to establish the rapid prediction of transmission efficiency was obtained.Secondly,a transmission efficiency prediction model based on the BP neural network and RBF neural network are established respectively.Since the accuracy of the model is related to the initial value,20 BP and RBF prediction models were established.The mean absolute percentage error(MAPE)of the best-precision BP and RBF prediction model are 1.63% and 1.72%.The average MAPE of these 20 BP and RBF models are 1.795% and 1.91%.Finally,a method of inversely adjusting the coil parameters based on the transmission efficiency prediction model is proposed,and the coil parameter estimation models based on the BP neural network and the RBF neural network are established respectively.When inputting the desired transmission efficiency and transmission distance,the model can output reasonable coil parameters.Corresponding coil models were established in HFSS,and the network output results were verified experimentally.Compared with the expected transmission efficiency,with a transmission distance of 90 mm,the error of the BP model is 0.35% and the error of the RBF model is 0.58%.Change the expected output and transmission distance,other settings remain unchanged,after many experiments,the average error of the BP parameter estimation model is 0.339%,and the average error of the RBF parameter estimation model is 0.571%,which are both smaller than the expected target.The experimental results show that the transmission efficiency prediction model of the magnetic coupling resonance system based on BP and RBF neural networks has a faster estimation speed than the HFSS simulation model under certain accuracy requirements.Under the given transmission efficiency,the coil parameter estimation model of the magnetic coupling resonance system based on BP and RBF neural networks can quickly obtain the required coil parameters,providing a basis for the coil design of magnetic coupling resonance system;whether it is the transmission efficiency prediction or the coil parameter estimation,BP model is better than RBF model.