| The absorbing material is a kind of material that can be used to absorb electromagnetic radiation energy and has great application value in technology and life.Absorbing materials could be divided into two types:coating and structural materials.The basic principle of coating materials is to apply absorbing paint on the surface of the structure.Structural absorbing materials mean that in addition to absorbing ability,they can also bear the load.Based on genetic algorithm,the electromagnetic parameters of coated absorbing materials and the geometric parameters of honeycomb structure absorbing materials are optimized in this paper.The geometric parameters of honeycomb structures absorbing wave are based on deep convolutional neural network?CNN?.The inverse prediction is performed,and the optimization results of genetic algorithm and deep convolutional neural network are compared.The main research contents of this article are as follows:?1?The joint simulation of MATLAB and FEKO is established,and the advantages of genetic algorithm and fast multipole simulation algorithm are fully utilized,which effectively improves efficiency of the optimization process of coated absorbing materials.The relative permittivity and relative permeability of the coated absorbing material are optimized by genetic algorithm.The reduction of the radar cross section?RCS?is used to measure the optimization effect.The results show that the genetic algorithm has a significant effect on the optimization of the coating material'sloapytiermemleocdter iclo mofazoaatt iionnggthnee imntisca e t thheo drmpaer as imzeeasnigg,et ne trhsiofif.e c t Caboomvahnet liyn pea-rmeedcrieddeunct n essatwio tmhpeitnhe d t hGeliRngC SaAR-bCaSsnvgall e u evd ea-o4f5leule t hee c torf o 4 m5otmhae,?dg eeslnuen tcicopaefct atpeiearl ldyadr a mmeettearloiun btlhee-?range:-1010,the reduction is about 25dBm2.?2?The MATLAB-HFSS joint simulation interface is established,and the characteristics of the genetic algorithm and HFSS software are used to improve the efficiency of the optimization process of the honeycomb structure absorbing material.By writing macro commands to control HFSS software to realize MATLAB-HFSS co-simulation,using MATLAB to control electromagnetic simulation software to automatically create honeycomb structure-type absorbing materials,the structural parameters of honeycomb structure absorbing materials are optimized based on GA.The optimization effect is reflected by the reduction degree of the reflection coefficient of the absorbing honeycomb,so that the reflection coefficient is reduced as much as possible.At the frequency of 8.412.2GHz,the simulation result of the reflection coefficient of the absorbing honeycomb model with initial structural parameters is roughly distributed at about-2dB.After the genetic algorithm is used to optimize the structural parameters,the reflection coefficient of the absorbing honeycomb decreases at all simulation frequencies.The maximum absorption peak of-20dB was reached at10.4GHz,which was reduced by about 18dB,which verified the effectiveness of the genetic algorithm in optimizing the parameters of the absorbing honeycomb structure.?3?An optimization method based on deep convolutional neural network was proposed to quickly and simultaneously optimize the structural parameters of honeycomb absorbing material height and impregnating thickness.Based on the simulation results of the reflection coefficient of the honeycomb model as input samples,a convolutional neural network consisting of 16 layers of neurons in the input layer was built.Batch normalization was used to avoid overfitting.The CNN model with the best performance on the verification set is selected to optimize the prediction of the honeycomb structure type absorbing material.The results show that in the 8.412.2GHz frequency band,all the reflection coefficients at 20 sampling frequencies are reduced,and the lowest frequency of the-5dB effective absorption bandwidth is9.6GHz,the highest frequency is 12GHz,and the effective absorption bandwidth is2.4GHz.The-10dB effective absorption band ranges from 9.6GHz to 11GHz,and the maximum absorption peak is reached at 10.4GHz,and the reflection coefficient is close to-25dB.In order to further explain the correctness of the geometric parameters of the CNN model for optimizing the honeycomb structure absorbing material,it is compared with the chapter 3 genetic algorithm for optimizing the reflection coefficient of the absorbing honeycomb.The reflection coefficients of the absorbing honeycomb optimized by the two methods are roughly consistent in the trend,and the performance of CNN in the frequency range of 9GHz11GHz is obviously better than that of genetic algorithm,and has higher optimization efficiency. |
PDF Full Text Request