%0 Journal Article
%T 基于量子遗传支持向量机的流体惯容预测模型<br>Predicting Model of Fluid Inerter Based on the Quantum Genetic Support Vector Machine
%A 沈钰杰
%A 陈龙
%A 刘雁玲
%A 杨晓峰
%A 张孝良
%A 汪若尘
%J 振动.测试与诊断
%D 2018
%R 10.16450/j.cnki.issn.1004-6801.2018.05.003
%X 为进一步研究惯容器的实现形式与力学性能特征，研制了新型流体惯容器装置。首先，在单通道伺服激振台架上进行力学性能测试，分析其非线性因素及对力学输出的影响机理；然后，考虑采用基于统计理论的支持向量机方法建立流体惯容器的力学输出预测模型，分别研究了不同激振频率、不同位移输入条件下的流体惯容器力学性能输出；最后，利用全局搜索能力较优的量子遗传算法优化径向基函数的方差与惩罚因子，并将构建的预测模型与试验输出果相对比。结果表明：利用支持向量机构建的流体惯容器力学输出预测结果与试验吻合良好采用量子遗传算法优化的预测模型具有更高的预测精度，其方差降幅最大可达到61.36%说明构建的预测模型正确合理，可为准确掌握流体惯容器动力学特性提供新思路。<br>In the light of the dynamic characteristics of an inerter, a new fluid inerter is designed. First,the bench tests are carried out on the hydraulic serve exciting platform and the nonlinearities and their effect on the output force are also analyzed. The predicting model of the fluid inerter by the support vector machine is established because the accurate dynamic model of fluid inerter is hard to expressed by analytical methods. Then, the force output of the fluid inerter at different frequency and displacement is investigated, and the variance of the radial basis function and the punishment factor are optimized by means of quantum genetic algorithm. Finally, the comparisons between the results obtained from the predicting model and those of the experiments show that the predicting results are in good agreement with those obtained by the tests. The optimal predicting model has a higher precision, and the amplitude is decreased by 61.36%. The predicting model successfully provides a new method to study the dynamics behavior of the fluid inerter.
%K 流体惯容器
%K 量子遗传
%K 支持向量机
%K 预测模型&lt
%K br&gt
%K fluid inerter
%K quantum genetic algorithm
%K support vector machine
%K predicting model
%U http://zdcs.nuaa.edu.cn/ch/reader/view_abstract.aspx?file_no=201805003&flag=1