摘要
为解决飞行器关键结构部件裂纹损伤的有效监测,及时发现潜在的安全隐患,避免灾难性事故的发生,采用先进的声发射技术对某军用飞行器真实关键结构部件的健康状态进行监测。使用小波包分析方法对所募集的飞行器结构部件声发射信息进行分解,提取能反映结构裂纹损伤信息的范数特征向量,作为支持向量机健康状态监测器的输入,对其进行训练和健康诊断研究。提出了一种由声发射信息范数特征向量与支持向量机相结合对飞行器结构裂纹损伤进行有效识别的新方法。在某军用飞行器真实结构部件的裂纹损伤试验中,运用该方法对其健康状态进行监测研究表明,该方法可准确诊断其裂纹损伤,为飞行器结构部件健康状态的有效监测提供了新途径。
To monitor effectively the aerocraft key structure components crack damages, discover in good time hidden trouble and avoid fearful accident occurring, a new kind of crack damage recognition method, based on norm feature vector from acoustic emission （AE） information and SVM, is proposed in this paper. The advanced AE technology is used to get health state information of the aerocraft key structure components. And the new kind of norm feature vector is extracted from the AE information by wavelet packet transform. The health state monitor for the aerocraft structure components is designed by using the support vector machine （SVM） in ＂one versus one＂ classification strategy to diagnose its health state. The monitor is trained and tested by the new kind of norm feature vectors. Experiments show that the method has good performance in recognizing and diagnosing the crack damages of the aerocraft structure components. It presents a new approach to monitor health state of aircraft structure com ponents.
出处
《压电与声光》
CSCD
北大核心
2009年第2期266-269,共4页
Piezoelectrics & Acoustooptics
基金
国家航空科学基金资助项目(2007ZD54006)
中国博士后科学基金资助项目(20070421062)
辽宁省教育厅科研基金资助项目(2008544)
沈阳航空工业学院博士启动基金资助项目(06YB19)
关键词
飞行器
健康诊断
范数
小波包
支持向量机
aerocraft
health diagnosis
norm
wavelet packet
support vector machine
