基于振动图像特征的机械状态异常检测算法  

Method of Anomaly Detection of Mechanical Operating State Based on Vibration Image Features

作　　者：卢国梁[1,2,3] 陈光远 曾世琛 李苏[4] 闫鹏 LU Guoliang;CHEN Guangyuan;ZENG Shichen;LI Su;YAN Peng(School of Mechanical Engineering,Shandong University Jinan,250061,China;Key Laboratory of High-Efficiency and Clean Mechanical Manufacture of MOE,Shandong University Jinan,250061,China;National Demonstration Center for Experimental Mechanical Engineering Education,Shandong University Jinan,250061,China;AtomHorizon Electric(Jinan)Company Limited Jinan,250061,China)

机构地区：[1]山东大学机械工程学院,济南250061 [2]山东大学高效洁净机械制造教育部重点实验室,济南250061 [3]山东大学机械工程国家级实验教学示范中心,济南250061 [4]美核电气(济南)股份有限公司,济南250061

出　　处：《振动．测试与诊断》2022年第1期148-153,201,共7页Journal of Vibration,Measurement & Diagnosis

基　　金：山东省自然科学基金资助项目(2018JC010);山东大学青年学者未来计划资助项目(2015WLJH30)。

摘　　要：针对实时监测得到的振动信号,从图像角度出发,提出了一种新的特征提取方法。首先,将实时监测得到的连续信号划分为独立的循环片段,将各个片段内的一维振动信号重构成二维灰度图像,在此基础上对灰度图像进行特征提取;其次,通过计算相邻周期图像间的相似性建立能描述机械运行动态特性的量化指标;然后,采用拉依达准则对机械设备运行状态进行实时监测与异常决策;最后,基于转速变化检测、(外部)负载变化检测以及早期轴承故障检测这3种典型的工程应用,对提出方法进行了验证。实验结果表明,对以上3种典型的工程应用场景均可实现100%的准确检测,证明了本研究方法的有效性。The real-time monitoring technology of operation state is an important means to ensure the safe and stable operation of modern mechanical equipment.In view of the vibration signals obtained from real-time moni⁃toring,firstly,the one-dimensional vibration signal is transformed into a two-dimensional gray image.On this basis,a new feature extraction method is proposed,and then by calculating the similarity between adjacent peri⁃odic images,a quantitative index that can describe the dynamic characteristics of the machine is established.Fi⁃nally,the Laida criterion is used to determine whether the operating state has changed.Based on three typical en⁃gineering applications:speed change detection,(external)load change detection,early bearing fault detection,the proposed method is verified.The experimental results show that the proposed method achieves 100%detec⁃tion accuracy in all three tested application scenarios,which reveals the effectiveness of the method.

关 键 词：机械状态监测 信号处理 特征提取 振动信号 

分 类 号：TH17[机械工程—机械制造及自动化]