滚动轴承故障诊断的方法很多,从振动信号中提取反映轴承状态的信息进行故障诊断是最常用的方法,也是最有效的方法之一"这是由于滚动轴承在运转时会受到各种动载荷而引起振动,从而对机组的寿命!可靠性有重要影响;振动是最重要的信息来源,是机械设备运行状态信息的载体,它蕴涵了丰富的机械设备异常或故障信息,振动特征是机械设备运行状态特征的本质反映"利用振动信号进行测试也是最方便!最实用的"因此,研究滚动轴承振动信号的分析和诊断方法具有重大的科学意义和实际意义"。

    本文了解滚动轴承故障诊断的一般方法，重点了解利用数字特征分析进行故障诊断的方法，根据现有滚动轴承振动的实验数据进行初步的统计分析，计算数学期望、方差、概率分布密度函数（直方图估计）及相应的功率谱（PSD），提取有、无故障时的特征，进行初步的诊断识别分析，进而论述基于滚动轴承振动信号数字特征分析的故障诊断方法及其应用。利用加速度传感器开展测试实验。对实验数据进行统计分析，探讨滚动轴承振动信号的数字特征及其在故障诊断中的应用，进行数据的数字特征、概率分布密度函数、功率谱计算、分析，进行初步的特征识别及相关诊断方法的研究、总结。滚动轴承故障、失效的一般形式及其对信号数字特征的影响分析，探讨基于数字特征分析的故障诊断原理及其应用。

 

关键词：滚动轴承，故障诊断，MATLAB仿真

Many rolling bearings fault diagnosis, information extraction reflect the state of the bearing from vibration signals for fault diagnosis is the most commonly used method, one of the most effective way. "This is due to the rolling bearing in operation will be subject to various dynamic loads caused by vibration !, which have a major impact on the reliability of the life of the unit; vibration is the most important source of information is the carrier of machinery and equipment operating status information, which contains a wealth of information on mechanical equipment malfunction or failure, the vibration characteristics of the mechanical equipment running status feature reflects the essence of "the use of the vibration signal for testing is the most convenient! the most practical" Therefore, the analysis and diagnosis method of vibration signals of great scientific and practical significance. "

In this paper, for general rolling bearings fault diagnosis, focused on understanding the use of digital signature analysis for fault diagnosis, according to a preliminary statistical analysis of the experimental data in the existing rolling bearing vibration calculated mathematical expectation, variance, probability density function (histogram estimation) and the corresponding power spectrum (PSD), there is extracted, it features no fault, a preliminary diagnostic analysis to identify, and then discusses the vibration signals based on a digital signature analysis fault diagnosis method and its application. Experimental tests carried out using the acceleration sensor. The experimental data were analyzed to explore the digital signature vibration signals and its application in fault diagnosis, digital feature data, probability density function, power spectrum calculation, analysis, preliminary feature recognition and related diagnostic methods Research summary. Rolling failure, usually in the form of failure and its effect on the characteristics of the digital signal analysis to explore digital signature based on the analysis of fault diagnosis principle and its application.

 

Keywords: bearing, fault diagnosis, MATLAB simulation

1.1．滚动轴承运行状态智能化监测的研究意义1

1.2．滚动轴承故障形式分析1

1.3．滚动轴承主要的状态监测技术2

1.3.1 振动信号分析诊断法2

1.3.2轴承润滑状态监测诊断法2

1.3.3 油液分析诊断法3

1.3.4 温度监测诊断法3

1.4．滚动轴承状态监测常用传感检测方法3

1.5．滚动轴承运行状态特征信号提取处理方法3

1.5.1 时域分析4

1.5.2 频谱分析4

1.5.3 小波分析5

2.1．滚动轴承振动信号处理6

2.1.1MATLAB软件简介6

2.1.2 MATLAB软件应用实例7

2.1.3 滚动轴承振动信号数据预处理7

2.2．滚动轴承振动信号特征抽取9

2.2.1滚动轴承振动信号时域特征值抽取9

2.2.2 滚动轴承振动信号频域特征值抽取11

2.2.3滚动轴承振动信号的FFT谱特征值抽取12

2.3．滚动轴承振动信号特征值归一化14

3.1．BP神经网络工作原理16

3.2．滚动轴承故障的BP神经网络设计18

3.2.1输入层和输出层神经元个数的选择18

3.2.2隐含层神经元个数的选择19

3.2.3BP神经网络训练参数选择19

3.2.4滚动轴承状态模式识别19

3.2.5 BP神经网络模型测试20