滚动轴承故障诊断与振动信号处理
发布时间：2019-04-01 20:51
【摘要】：滚动轴承作为一个基础零部件，它的运转可靠性直接影响到整个系统工作的稳定。能够及时较为准确的定位轴承故障部位同时诊断故障严重程度对于减少不必要的损失和确保人员安全起到至关重要的作用。本文主要通过实验和仿真模拟故障同时分析故障振动信号特点来达到这一目的。 本文首先理论研究了轴承故障振动信号的产生机理和信号特性，，同时结合实际205轴承利用运动学关系推导了信号特征频率。仿真部分建立了滚动轴承的Patran有限元模型和Adams动力学模型，通过有限元模型分析了轴承载荷分布、固有频率及模态在故障振动信号中的表现形式，通过动力学模型分析了不同故障发生部位的故障特征与故障信号影响因素：故障大小、载荷、转速、间隙。同时提出了测点优化方案，对于改变垂向测点为45度斜向的方式来提高故障脉冲信号的测量灵敏度。 对205轴承8个不同故障做了1200组实验信号采集，利用时频域分析与统计分析两大类方法做了定位故障与分析故障严重程度的诊断研究。统计量的应用以大量数据为背景，人为划分故障分析区间，用峭度方法诊断故障模式，用P/R方法诊断故障程度；时频域分析主要利用改进共振解调多次包络的方式验证信号的诊断准确性，诊断正确率在90%左右，同时利用谱峭度、小波包能量的方法来提取特征信号与滤除干扰信号，达到提高信号信噪比效果。
[Abstract]:Rolling bearing as a basic component, its running reliability directly affects the stability of the whole system. It is very important to locate the bearing fault location and diagnose the fault severity in time to reduce unnecessary loss and ensure the safety of personnel. The main purpose of this paper is to analyze the characteristics of the fault vibration signal through the simulation and experiment of the fault at the same time. In this paper, the generating mechanism and signal characteristics of bearing fault vibration signal are studied theoretically, and the characteristic frequency of bearing fault vibration signal is deduced according to the kinematic relation of 205 bearing. In the simulation part, the Patran finite element model and the Adams dynamic model of the rolling bearing are established. Through the finite element model, the bearing load distribution, natural frequency and modal representation in the fault vibration signal are analyzed. The dynamic model is used to analyze the fault characteristics of different fault locations and the influencing factors of fault signal: fault size, load, rotational speed and clearance. At the same time, an optimization scheme is proposed to improve the measurement sensitivity of the fault pulse signal by changing the vertical measurement point to 45 degrees oblique direction. 1200 sets of experimental signals were collected for 8 different faults of bearings. The diagnosis of locating faults and analyzing the severity of faults was carried out by using two kinds of methods: time-frequency domain analysis and statistical analysis. The application of statistics is based on a large number of data, divide the fault analysis interval artificially, diagnose the fault mode by kurtosis method, and diagnose the fault degree by Pur method. The time-frequency domain analysis mainly uses the improved resonance demodulation multiple envelope method to verify the diagnostic accuracy of the signal, and the diagnostic accuracy is about 90%. At the same time, the method of spectrum kurtosis and wavelet packet energy is used to extract the characteristic signal and filter the interference signal. The signal-to-noise ratio is improved.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH133.33
本文编号：2451897
[Abstract]:Rolling bearing as a basic component, its running reliability directly affects the stability of the whole system. It is very important to locate the bearing fault location and diagnose the fault severity in time to reduce unnecessary loss and ensure the safety of personnel. The main purpose of this paper is to analyze the characteristics of the fault vibration signal through the simulation and experiment of the fault at the same time. In this paper, the generating mechanism and signal characteristics of bearing fault vibration signal are studied theoretically, and the characteristic frequency of bearing fault vibration signal is deduced according to the kinematic relation of 205 bearing. In the simulation part, the Patran finite element model and the Adams dynamic model of the rolling bearing are established. Through the finite element model, the bearing load distribution, natural frequency and modal representation in the fault vibration signal are analyzed. The dynamic model is used to analyze the fault characteristics of different fault locations and the influencing factors of fault signal: fault size, load, rotational speed and clearance. At the same time, an optimization scheme is proposed to improve the measurement sensitivity of the fault pulse signal by changing the vertical measurement point to 45 degrees oblique direction. 1200 sets of experimental signals were collected for 8 different faults of bearings. The diagnosis of locating faults and analyzing the severity of faults was carried out by using two kinds of methods: time-frequency domain analysis and statistical analysis. The application of statistics is based on a large number of data, divide the fault analysis interval artificially, diagnose the fault mode by kurtosis method, and diagnose the fault degree by Pur method. The time-frequency domain analysis mainly uses the improved resonance demodulation multiple envelope method to verify the diagnostic accuracy of the signal, and the diagnostic accuracy is about 90%. At the same time, the method of spectrum kurtosis and wavelet packet energy is used to extract the characteristic signal and filter the interference signal. The signal-to-noise ratio is improved.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH133.33
【参考文献】
相关期刊论文 前10条
1 刘明利;傅行军;;一种对振动信号小波包能量算法的研究和改进[J];华东电力;2009年08期
2 李嶷,熊国良,张龙;经验模式分解法(EMD)在滚动轴承故障诊断中的应用[J];华东交通大学学报;2005年01期
3 唐云冰;高德平;罗贵火;;航空发动机高速滚珠轴承力学特性分析与研究[J];航空动力学报;2006年02期
4 梅宏斌,吴雅,师汉民,杨叔子;滚动轴承损伤类故障的一种精确诊断模型[J];华中理工大学学报;1992年S1期
5 丁芳;高立新;崔玲丽;张建宇;胥永刚;;共振解调技术在设备故障诊断中的应用[J];机械设计与制造;2007年11期
6 梅宏斌;杨叔子;师汉民;;滚动轴承故障诊断高频共振法的包络频谱分析[J];洛阳工学院学报;1992年04期
7 李国臣;李永强;;薄壁圆柱壳固有频率的计算[J];机械科学与技术;2010年09期
8 李自光;;机械故障的油样分析诊断[J];机械制造;1992年10期
9 徐国栋;周锡元;郭明珠;董娣;;典型振动信号时-频分析方法研究[J];世界地震工程;2007年02期
10 马新锋;顾海明;李泽豪;;基于EMD和Hilbert-Huang的滚动轴承故障诊断[J];通用机械;2011年08期
本文编号：2451897
本文链接：https://www.wllwen.com/kejilunwen/jixiegongcheng/2451897.html
