论文部分内容阅读
孔缝电磁耦合问题历来是电磁兼容领域中的一个重要部分,多年来人们已经采用多种方法对孔缝电磁耦合问题进行了深入研究.随着电子系统的日益复杂,无论是全波分析还是解析方法在孔缝电磁耦合问题研究中都存在很大的局限性.为了更好的解决这个问题,本文探讨了人工神经网络用于孔缝耦合分析的可行性.以大量数值计算结果为训练样本,建立了适用于分析任意大小圆孔电磁耦合的人工神经网络模型.该模型以圆孔半径的电尺寸为输入参数,等效散射面积为输出参数.利用该模型预测其他尺寸圆孔的电磁耦合,结果证明采用基于人工神经网络的圆孔电磁耦合预测是可行的、高效的.该方法可以推广到更为复杂的腔体孔缝电磁耦合的分析中.
Hole-hole electromagnetic coupling problem has always been an important part in the field of electromagnetic compatibility. For many years, many methods have been used to study the hole-hole electromagnetic coupling problem. With the increasingly complex electronic systems, whether full-wave analysis or analytical In order to solve this problem better, this paper discusses the feasibility of using artificial neural network in hole-hole coupling analysis.With a large number of numerical results as training samples, An artificial neural network model which is suitable for analyzing the electromagnetic coupling of circular holes of arbitrary size is established.The model takes the electrical dimension of the radius of circular hole as input parameter and the equivalent scattering area as output parameter.The model is used to predict the electromagnetic coupling of circular holes of other sizes, The results show that it is feasible and efficient to predict the electromagnetic coupling of circular holes based on artificial neural network, which can be extended to the analysis of more complex cavity-hole electromagnetic coupling.