基于多分形特征的枪械自动机裂纹故障诊断  

Crack Fault Diagnosis of Gun Automatic Mechanism Based on Multifractal Features

作　　者：任海锋[1] 潘宏侠[1] REN Hai-feng;PAN Hong-xia(School of Mechanical Engineering, North University of China, Taiyuan 030051, Shanxi, China)

机构地区：[1]中北大学机械工程学院,山西太原030051

出　　处：《兵工学报》2018年第3期457-462,共6页Acta Armamentarii

基　　金：国家自然科学基金项目(51675491)

摘　　要：为更好地利用振动信号对枪械自动机的裂纹故障进行诊断,提出基于振动信号多分形特征的故障诊断方法。该方法利用Wavelet Leader来估计振动信号的多分形谱,通过6个特征量描述多分形谱的形态特征以实现多分形谱的降维,并使用基于Mahalanobis距离的分类器对不同的裂纹故障进行分类。应用该方法对某12.7 mm高射机枪自动机闭锁机构的裂纹故障进行了诊断,诊断正确率达到了82.5%,验证了将振动信号的多分形特征用于自动机裂纹故障诊断的可行性。In order to make better use of vibration signals to diagnose the crack faults of gun automatic mechanism,a fault diagnosis method based on multifractal features of vibration signals is proposed. The proposed method uses Wavelet Leader to estimate the multifractal spectrum of vibration signals. 6 feature quantities are used to describe the morphological features of multifractal spectrum,and the dimensionality reduction of multifractal spectrum is realized. A classifier based on Mahalanobis distance is used to classify different crack faults. This method is applied to diagnose the crack faults of locking mechanism in a 12. 7 mm antiaircraft machine gun,and the diagnostic accuracy is up to 82. 5%,which verifies the feasibility of applying the multifractal features of vibration signals to the crack fault diagnosis of gun automatic mechanism.

关 键 词：枪械自动机 多分形特征 WAVELET Leader方法 MAHALANOBIS距离 裂纹故障诊断 

分 类 号：TJ203[兵器科学与技术—武器系统与运用工程]