小波包变换和支持向量机相结合在发动机气门间隙故障诊断中的应用  

Application of combining wavelet transform and support vector machines on fault diagnosis for engine valve clearance

作　　者：张梅军[1] 石文磊[1] 赵亮[1] 李曙光[1] 

机构地区：[1]解放军理工大学工程兵工程学院,江苏南京210007

出　　处：《机械研究与应用》2009年第6期115-117,120,共4页Mechanical Research & Application

摘　　要：为了对发动机气门间隙进行故障诊断,在对振动信号进行采集和预处理的基础上,运用小波包频带能量分解技术提取发动机故障的特征向量,以此作为支持向量机分类器(SVM)的训练样本,用经训练的SVM多分类器对发动机不同故障进行自动识别和诊断,实现了信号特征向量提取与故障模式识别的有机结合。实验结果表明,该方法能在机械故障样本少的情况下准确的识别和诊断出发动机气门间隙的故障类型,具有实际的工程应用价值。For diagnosing fault of the engine valve clearance,the fault eigenvector was distilled by using the wavelet packet transforms,based on gathering and pretreating the libration signal.The eigenvectors were input to SVM multi-classifier to recognize and diagnose different faults;The purpose was presented by combining distilling signal eigenvectors with identifying fault modes.The result shows that this method can identify and diagnose the fault types of the engine valve clearancet exactly,and is of a applied value to engineering application.

关 键 词：小波包变换 支持向量基(SVM) 故障诊断 

分 类 号：E92[军事—军事装备学] TK428[兵器科学与技术—武器系统与运用工程]