基于支持向量机的数控机床进给系统滚动轴承故障诊断研究

发布时间：2018-01-06 23:35

本文关键词：基于支持向量机的数控机床进给系统滚动轴承故障诊断研究　出处：《青岛理工大学》2013年硕士论文　论文类型：学位论文

【摘要】：近年来,我国大力发展装备制造业,对数控设备关键部件的研发、生产、制造以及故障诊断方面加大了研究。为了满足目前生产中对加工质量和精度更高的要求,实现装备制造业的全面发展,国家已将数控设备研发列入十一五重大科技专项。数控机床加工中,旋转部件的工作状态是保证加工质量的的关键因素,数控机床在装备制造业中占据着重要的地位,滚动轴承是数控机床中十分关键的部件,当机床中轴承发生故障时,不可避免的对机床产生影响,有数据显示,目前数控设备中30%的故障是由于轴承故障引起的,轴承一旦发生故障会严重影响生产效率,且状况严重时造成重大安全事故,因此,开展对数控机床滚动轴承的故障在线诊断,监测机床轴承的运行状态,具有重要的实用意义和经济意义。本文分析了数控机床进给系统中轴承部件的常用种类,以及滚动轴承的主要结构,在此基础上确定研究对象为角接触球轴承,分析了轴承的主要故障形式和形成机理以及故障振动频率,提出了基于振动信号和支持向量机的轴承故障诊断方法,研究了振动信号的小波包分析方法,将检测振动信号进行小波包分解,得到小波包能量作为特征向量,最后通过实验分析对该方法进行了实验验证。首先,从滚动轴承的主要种类、结构特点入手,在此基础上着重分析了数控机床进给系统滚动轴承的主要失效形式以及各失效形式产生的原因和现象,对失效形式、原因、现象进行了归纳总结；当机床进给轴承发生故障时,会对数控机床产生影响,本文分析了滚动轴承常见故障对数控机床进给系统的影响；对滚动轴承振动信号的特性进行了深入详细的研究。其次,研究了滚动轴承故障诊断软硬件系统的设计,基于LabVIEW和MATLAB软件实现故障信号的数据采集和数据分析处理,数据分析主要实现了振动信号的小波包分解以及小波包能量的计算,然后利用MATLAB支持向量机工具箱建立支持向量机模型,实现滚动轴承故障的模式识别。再次,在西门子数控机床802Dsl上搭建实验台,对进给系统滚动轴承的故障进行了实验分析,验证了本文基于支持向量机的数控机床滚动轴承故障诊断的有效性。最后,研究了基于PXI-6281多功能数据采集卡的滚动轴承振动信号的数据采集技术及基于Microsoft Access数据库的数据管理技术与方法,以LabVIEW为平台,开发了由数据采集模块、数据处理模块和数据分析模块组成的数控机床进给系统滚动轴承故障诊断系统。
[Abstract]:In recent years, our country vigorously develops the equipment manufacturing industry, to the numerical control equipment key component research and development, the production. In order to meet the requirement of higher machining quality and precision in current production, the equipment manufacturing industry can be developed in an all-round way. The research and development of numerical control equipment has been listed as a major scientific and technological project in the 11th Five-Year Plan. The working state of rotating parts is the key factor to ensure the quality of machining in NC machine tool processing. The CNC machine tool occupies an important position in the equipment manufacturing industry. The rolling bearing is a very important part in the NC machine tool. When the bearing in the machine tool breaks down, it will inevitably have an impact on the machine tool. At present, the fault of 30% in the NC equipment is caused by the bearing fault, once the bearing failure will seriously affect the production efficiency, and when the situation is serious, there will be a serious safety accident. It is of great practical and economic significance to carry out on-line fault diagnosis of rolling bearings of NC machine tools and monitor the running state of machine tools bearings. This paper analyzes the common types of bearing parts in the feed system of CNC machine tools and the main structure of rolling bearings. On this basis, the object of study is angular contact ball bearings. The main fault form and forming mechanism of bearing and the frequency of fault vibration are analyzed. The method of bearing fault diagnosis based on vibration signal and support vector machine is put forward, and the wavelet packet analysis method of vibration signal is studied. The wavelet packet energy is obtained as the eigenvector by wavelet packet decomposition of the detected vibration signal. Finally, the method is verified by experimental analysis. First of all, starting with the main types and structural characteristics of rolling bearings, the main failure forms of rolling bearings in CNC machine tool feed system and the causes and phenomena of each failure forms are analyzed emphatically. The form, cause and phenomenon of failure are summarized. When the feed bearing of the machine tool breaks down, it will affect the NC machine tool. This paper analyzes the influence of the common fault of the rolling bearing on the feed system of the NC machine tool. The characteristics of rolling bearing vibration signal are studied in detail. Secondly, the software and hardware design of rolling bearing fault diagnosis system is studied. The data acquisition and data analysis of fault signals are realized based on LabVIEW and MATLAB software. The data analysis mainly realizes the wavelet packet decomposition of vibration signal and the calculation of wavelet packet energy. Then the support vector machine model is established by using MATLAB support vector machine toolbox. The pattern recognition of rolling bearing fault is realized. Thirdly, the experiment platform is built on 802DSL of Siemens NC machine tool, and the fault of rolling bearing of feed system is analyzed experimentally. The effectiveness of bearing fault diagnosis of NC machine tool based on support vector machine is verified. Finally. The data acquisition technology of rolling bearing vibration signal based on PXI-6281 multi-function data acquisition card and the data management technology and method based on Microsoft Access database are studied. Based on LabVIEW, a rolling bearing fault diagnosis system of NC machine tool feed system is developed, which is composed of data acquisition module, data processing module and data analysis module.
【学位授予单位】：青岛理工大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TG659;TH165.3

【参考文献】