多模块构架式空间可展开天线结构参数优化

内容简介：为降低卫星天线的发射成本,提高天线的展开刚度,以多模块构架式空间可展开天线结构的质量和1阶固有频率为目标函数,基于误差反向传播（BP）神经网络和遗传算法对天线的结构参数进行了优化.运用ANSYS软件对支撑桁架的结构参数进行了数值模拟,得到了与设计变量对应的目标函数值;通过正交试验设计,构建了用于神经网络训练和检验的样本集;按照BP算法的基本思想,调整网络模型的参数,建立了用于优化的预测模型;采用分目标乘除法,将多目标优化问题转变成单目标优化问题;采用遗传算法进行了优化分析,得到了支撑桁架各杆件的设计参数.结果表明：该优化方法在降低天线质量的同时,使结构的刚度得到了提高,为天线的结构设计提供了参考.

Abstract：In order to reduce the launching cost of deployable antenna and improve its deployment stiffness,by taking the mass and the first order natural frequency of deployable truss antenna with multi-module as the objective functions,the structural parameters of truss structure were optimized based on BP（back propagation） neural network and genetic algorithm.The numerical simulation of structural parameters was studied by software ANSYS,and the objective function values corresponding to the design variables were obtained.The training samples and test samples were obtained by orthogonal design.According to the basic idea of BP neural network,prediction model was derived by adjusting the parameters of network model.Sub-goals multiplication and division were adopted to simplify the multi-objective optimization as a single-objective function.The optimization analysis was performed by genetic algorithm,and the design parameters of truss structure were obtained.The results show that the mass is reduced and the deployment stiffness is improved simultaneously.This optimization method provides a foundation for the structural design of deployable truss antenna.

作者：田大可, 郭宏伟, 邓宗全, 刘荣强,

关键词：空间可展开天线, 支撑桁架, 多目标优化, 反向传播(BP)神经网络, 遗传算法,