张宝印,董恩生.基于PCA-GA-RSPSVM的复合材料损伤检测技术研究[J].电子测量与仪器学报,2017,31(9):1402-1407
基于PCA-GA-RSPSVM的复合材料损伤检测技术研究
Research on damage detection technique of composite material based on PCA GA RSPSVM
  
DOI:10.13382/j.jemi.2017.09.008
中文关键词:  支持向量机  主元分析  同面电容传感器  损伤检测  复合材料
英文关键词:support vector machine(SVM)  principal component analysis (PCA)  uniplanar capacitance sensor  anomaly detection  composite material
基金项目:装备维修科学与改革项目(2011325)资助
作者单位
张宝印 空军航空大学长春130022 
董恩生 空军航空大学长春130022 
AuthorInstitution
Zhang Baoyin Air Force Aviation University, Changchun 130022, China 
Dong Ensheng Air Force Aviation University, Changchun 130022, China 
摘要点击次数: 816
全文下载次数: 846
中文摘要:
      针对复合材料损伤检测数据少、效率低等问题,提出一种基于主元分析(PCA)和改进的轮换对称分块支持向量机(RSPSVM)的损伤识别算法,并用其进行飞机复合材料构件损伤检测。首先,算法对平面多电极电容传感器检测模型等面积剖分,获取足够多复合材料检测样本;然后引入遗传算法(GA)改进RSPSVM获得更好的分类性能,并且结合PCA提取主特征向量用于降低特征向量维度和缩短训练时间,将新的特征集送入改进的RSPSVM算法,实现PCA GA RSPSVM识别算法;最后,用3种复合材料样板的实测值对算法进一步验证。经过仿真数据与实测数据的验证,有效的验证了PCA GA RSPSVM算法应用于飞机复合材料构件损伤检测的有效性。
英文摘要:
      Aiming at solving lacking of failure data and low efficiency of composite material damage detection, a fault diagnosis method based on principal component analysis(PCA)and support vector machine combined with the rotation symmetric partition(RSPSVM)was proposed. Firstly, the model of uniplanar multi electrode is partitioned into equal area units with rotation symmetry partition, and fault data is acquired adequately. Secondly, genetic algorithm(GA)was introduced into RSPSVM in order to promote the classification performance, and PCA was used to reduce the dimension of feature vector and shorten training time, the final features were put into improved RSPSVM so that PCA GA RSPSVM was achieved. Finally, the measured data of three composite material samples were sent to PCA GA RSPSVM for verification. After verification of the simulation data and the measured data, the effective certificate of general PCA GA RSPSVM algorithm is applied to the diagnosis of damage of aircraft composite material.
查看全文  查看/发表评论  下载PDF阅读器