HHT-SVM在滚珠丝杠副疲劳点蚀失效诊断中的应用  

Fault Diagnosis Method for Ball Screw Pitting Corrosion Based on HHT and SVM

作　　者：聂从辉 周长光 刘迪一 冯虎田[1] 王志荣[1] 欧屹[1] NIE Cong-hui;ZHOU Chang-guang;LIU Di-yi;FENG Hu-tian;WANG Zhi-rong;OU Yi(School of Mechanical Engineering,Nanjing University of Science and Technology,Nanjing 210094,China)

机构地区：[1]南京理工大学机械工程学院,南京210094

出　　处：《组合机床与自动化加工技术》2020年第12期80-84,89,共6页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：国家重大科技专项(2017ZX04011001);国家自然科学基金(51905274)。

摘　　要：针对滚珠丝杠副振动信号的非平稳特性的特点,提出了一种基于希尔伯特黄变换(Hilbert-Huang Transform,简称HHT)和支持向量机(Support Vector Machine,简称SVM)的滚珠丝杠副疲劳点蚀诊断方法。该方法首先对滚珠丝杠副进行了运动学分析,推导出了正常状态及疲劳点蚀失效状态下的特征频率;通过经验模态分解(Empirical Mode Decomposition,简称EMD)方法将原始振动信号分解为若干平稳本征模函数;利用有效固有模态函数(Intrinsic Mode Function,简称IMF)分量重构信号并进行希尔伯特变换获得振动信号的频谱图,提取特征频率作为输入建立支持向量机,从而实现对滚珠丝杠副点蚀失效的诊断。试验结果表明:支持向量机的诊断正确率达到95%以上,表明了了该方法可有效对滚珠丝杠副疲劳点蚀失效进行诊断,在工程上具有一定的实用值。Aiming at the non-stationary characteristics of the ball screw vibration signal,a ball screw pitting corrosion diagnosis method based on HHT and SVM was proposed.Firstly,the kinematics analysis of the ball screw pair was carried out,and the characteristic frequencies of the ball screw under normal condition and pitting corrosion state were derived.The original vibration signal was decomposed into several stationary eigenmode functions by EMD.The effective IMFs were used to reconstruct the signal and the Hilbert-Huang transform was used to obtain the spectrum of vibration signal.Then the characteristic frequencies were extracted as the input to establish the support vector machine network,thereby realizing the diagnosis of ball screw pitting corrosion.The results showed that the diagnostic accuracy of support vector machine is over 95%,which indicates that the method can effectively diagnose the fatigue pitting failure of ball screw,and hence is of applicative value.

关 键 词：点蚀 滚珠丝杠副 故障诊断 希尔伯特黄变换 支持向量机 

分 类 号：TH165.3[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]