飞机复合材料结构损伤的预测方法  

Damage Prediction and Verification of Aircraft Composite Materials

作　　者：崔建国[1] 张文生[1] 蒋丽英[1] 朴春雨[2] 周志强[2] 

机构地区：[1]沈阳航空航天大学自动化学院,辽宁沈阳110136 [2]沈阳飞机设计研究所,辽宁沈阳110035

出　　处：《材料科学与工程学报》2016年第5期755-760,共6页Journal of Materials Science and Engineering

基　　金：辽宁省自然科学基金资助项目(2014024003);航空科学基金资助项目(20153354005);国防预研资助项目(A0520110023);国防基础科研资助项目(Z052012B002)

摘　　要：针对复合材料结构损伤机理的复杂性,很难准确预测结构损伤状态,本文提出一种基于动态主元分析(DPCA)和最小二乘支持向量机(LS-SVM)相结合的复合材料结构损伤演化预测新方法,并针对复合材料结构损伤特性,采用疲劳振动试验进行结构损伤预测研究。首先,采用经验模态分解(EMD)方法对多传感器采集的复合材料结构健康信息进行自适应分解,得到不同传感器下的多个本征模态分量(IMF),并通过计算各阶IMF分量的奇异熵作为各传感器的特征信息;然后采用DPCA对多传感器的奇异熵进行降维融合,得到融合后的奇异熵特征,再对其采用距离形态相似度方法定义结构健康指数;最后将结构健康指数作为建模数据,创建LS-SVM预测模型,并通过预测模型对飞机复合材料结构健康指数进行预测,其预测结果直接反映了飞机复合材料结构的健康状态。试验验证表明,该方法可有效地实现飞机复合材料结构损伤预测效能,具有很好的工程应用价值。Due to complex mechanism behind a composite structural damage, it is difficult to find a clear function to characterize the structure damage state of the composite. In this work, a new method to predict the composite structural damage veolutionby combining Dynamic Principal Component Analysis （DPCA） with Least Squares Support Vector Machines （LS-SVM） was proposed. At the same time , vibration fatigue test was performed for the investigation of the composite structural damage preidiction. In this whole approach, the health information of the composite structure acquired via multi-sensor was decomposed into Intrinsic Mode Function （IMF） by Empirical Mode Decomposition （EMD）; then, multi-feature dimension reduction fusion was made to get the characteristic singular entropy using DPCA, and the singular entropy was defined as the structure of the health index using the method of DMS; finally, regarding the structure of the health index as modeling data, LS-SVM model was established to predict the health status of the aircraft composite structures. The experimental verification shows that this method can be used effectively to predict aircraft composite structural damage, with good engineering value.

关 键 词：复合材料 结构损伤预测 动态主元分析法 最小二乘支持向量机 

分 类 号：V257[一般工业技术—材料科学与工程]