基于特征频率的天线替代模型优化方法  

Optimization of antenna surrogate model based on feature frequency

作　　者：李坤来 李廷鹏 陈志峰 杨娜娜 林文婷 王健 闫淑霞[3] LI Kun-lai;LI Ting-peng;CHEN Zhi-feng;YANG Na-na;LIN Wen-ting;WANG Jian;YAN Shu-xia(State Key Laboratory of Complex Electromagnetic Environmental Effects on Electronics and Information System,Luoyang 471003,China;School of Electronic Information Engineering,Inner Mongolia University,Hohhot 010021,China;School of Electronics and Information Engineering,Tiangong University,Tianjin 300387,China;School of Microelectronics,Tianjin University,Tianjin 300072,China;Qingdao Institute for Ocean Technology,Tianjin University,Qingdao 266200,China)

机构地区：[1]电子信息系统复杂电磁环境效应国家重点实验室,河南洛阳471003 [2]内蒙古大学电子信息工程学院,内蒙古呼和浩特010021 [3]天津工业大学电子与信息工程学院,天津300387 [4]天津大学微电子学院,天津300072 [5]天津大学青岛海洋技术研究院,山东青岛266200

出　　处：《陕西科技大学学报》2023年第3期180-184,214,共6页Journal of Shaanxi University of Science & Technology

基　　金：天津市自然科学基金项目(19JCQNJC03300);电子信息系统复杂电磁环境效应国家重点实验室开放基金项目(CEMEE2022G0201)。

摘　　要：目前天线优化设计主要采用数值仿真方法来实现,设计过程主要依靠全波仿真,但其不仅仿真时间长而且耗费大量计算机资源.为提高天线设计效率,提出一种基于特征频率的天线替代模型优化方法,利用神经网络建立的替代模型代替耗时长的全波仿真模型进行优化,引入了特征频率技术作为替代模型优化辅助条件,可将特征频率移位至目标频率范围内,从而为优化的过程提供了更为直接的方向,加快了基于替代模型中的优化速度.此外,本文引入了信任区间更新算法,提高建模范围的有效性,减少优化次数.双频天线设计实例证明:所提出的替代模型优化方法仅用较少的迭代次数就可获得最优设计参数,缩短了天线设计周期.Antenna optimization design is mainly realized by numerical simulation algorithm.The design process of antenna mainly relies on full-wave simulation,which costs long simulation time and huge computer resources.In order to improve the design efficiency of antenna,this paper proposed an optimization method of antenna surrogate model based on feature frequency,which utilizes the surrogate model established by neural network to replace the time-consuming full-wave simulation model for optimization.As an auxiliary condition for the optimization of the surrogate model,the feature frequency technology is introduced,which can shift the characteristic feature frequency to the target range,so as to provide a direct direction for the optimization process and accelerate the optimization speed.In addition,a trust region updating algorithm is introduced to improve the validity of model modeling range and reduce the iteration times of surrogate model.The experiment of dual-frequency antenna proves that the proposed optimization method can obtain the optimal design parameters with fewer iterations and shorten design cycle of the antenna.

关 键 词：替代模型 神经网络 天线优化 

分 类 号：TN820[电子电信—信息与通信工程]