Prediction of the Electromagnetic Shielding Effectiveness of Metal Grid Using Neural Network Algorithm

Huiyu Chang,Jiale Gao,Senfeng Lai,Yanghui Wu,Chen Fu,Wenhua Gu

DOI: https://doi.org/10.1109/jphot.2021.3107298

IF: 2.4

2021-08-01

IEEE Photonics Journal

Abstract:With the advantages of simple structure, high electromagnetic (EM) shielding effectiveness (SE), wide shielding bandwidth and adjustable transmittance, the metal grid is a popular tool in the field of electromagnetic shielding. In this work, the deep learning algorithm of artificial neural network was used to predict the EM SE of the metal grid. After certain number of trainings, both the SE spectrum and the minimum SE value in a given frequency band can be predicted accurately and effectively. On the other hand, for a target SE, the design parameters of the metal grid structure can be immediately predicted. The results show that the relative mean square deviation of the predicted SE curve is about 5%, the relative error of the minimum SE value of the curve is less than 5%, and the average relative error of the minimum SE value in the predicted frequency band is 11.5%.

engineering, electrical & electronic,optics,physics, applied