基于PCA-GA-RSPSVM的复合材料损伤检测技术研究

发布时间：2018-04-22 21:10

本文选题：支持向量机 + 主元分析　；参考：《电子测量与仪器学报》2017年09期

【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。
[Abstract]:Aiming at the problems of less data and low efficiency of composite material damage detection, a damage identification algorithm based on principal component analysis (PCA) and improved rotational symmetric block support vector machine (RSPSVM) is proposed, and it is used to detect the damage of aircraft composite component. Firstly, the algorithm divides the detection model of planar multi-electrode capacitive sensor into equal area and obtains enough samples of composite material detection. Then genetic algorithm (GA) is introduced to improve the performance of RSPSVM to obtain better classification performance. And the main feature vector extracted by PCA is used to reduce the dimension of the feature vector and shorten the training time, and the new feature set is sent into the improved RSPSVM algorithm to realize the PCA-GA-RSPSVM recognition algorithm. Finally, The algorithm is further verified by the measured values of three kinds of composite material samples. The validity of the PCA-GA-RSPSVM algorithm applied to the damage detection of aircraft composite components is verified by the simulation data and the measured data.
【作者单位】：空军航空大学;
【基金】：装备维修科学与改革项目(2011325)资助
【分类号】：TB33

【相似文献】