摘要
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.
The aperture coupling problem plays an important in electromagnetic compatibility （EMC）. Both analytical and numerical methods have been developed to solve these problems. However, each of those methods faces more and more challenges since electronic systems are becoming increasingly complicated. In order to find some other approaches to solve the aperture coupling problem better, a prediction model based on artificial neural network is proposed in this paper. Using a lot of numerical results as the training sample, a prediction model for circular aperture electromagnetic coupling with any radius is determined. The predicted results of diffraction cross section show agreements with the simulation results which are never used in the training. The proposed model is expected to extend to solve more complicated aperture coupling problems of cavity.
出处
《四川大学学报（自然科学版）》
CAS
CSCD
北大核心
2011年第5期1071-1074,共4页
Journal of Sichuan University(Natural Science Edition)
基金
国家自然科学基金项目-青年科学基金项目(60801035)
关键词
圆孔电磁耦合
人工神经网络
等效散射面积
circular aperture electromagnetic coupling, artificial neural network （ANN）, diffraction cross section
