基于自适应形态滤波的滚动轴承故障检测与识别

发布时间：2018-03-27 20:49

本文选题：滚动轴承　切入点：声发射　出处：《沈阳航空航天大学》2017年硕士论文

【摘要】：滚动轴承是旋转设备中最重要的机械组件之一,但具有易损特性,一旦发生故障会极大程度地影响整个设备的运行安全。本文主要研究如何检测滚动轴承运行过程中是否发生故障,并识别故障类型,在机械设备安全维护方面有着重要意义。实验所用数据来自江苏千鹏QPZZ-II型滚动轴承故障模拟平台,通过该平台模拟轴承的四种运行状态,分别是正常、滚动体故障、外圈故障以及内圈故障。滚动轴承声发射信号由SAEU2S声发射系统采集。首先,研究自适应形态滤波方法对声发射信号进行降噪。设计一种加权级联形态滤波器,利用开运算抑制声发射正脉冲峰值噪声,利用闭运算抑制声发射负脉冲谷值噪声;比较不同的去噪指标,引入去噪前后信号的相关性指标作为实测声发射信号的降噪评价指标,实现对形态闭开和开闭运算权值的自适应调整,完成形态滤波器的设计。对模拟声发射信号的仿真结果表明该方法可以有效滤除信号中的脉冲噪声。通过信噪比与相关性指标的比较表明二者随权值变化的趋势一致,从而验证相关性指标可以代替信噪比用于评价噪声强度未知的实测信号去噪效果。其次,针对自适应形态滤波方法在白噪声抑制方面表现欠佳,将自适应形态滤波与集合平均经验模态分解(Ensemble Empirical Mode Decomposition,EEMD)进行去噪组合。应用EEMD对形态滤波处理后的信号进行分解,将信号从高频到低频分解为有限个反映不同振动模态的本征模态函数(IMF),从而分离出高频干扰信号。对模拟声发射信号的仿真和比较结果表明了组合方法的可行性与有效性。最后采用组合方法对滚动轴承故障模拟平台的实测声发射信号进行处理,并分析比对不同轴承故障状态下的频域特征,识别故障。实验结果验证本文方法在实测轴承信号处理中的有效性。
[Abstract]:Rolling bearing is one of the most important mechanical components in rotating equipment, but it has the characteristics of vulnerability. Once the fault occurs, the safety of the whole equipment will be greatly affected. This paper mainly studies how to detect the rolling bearing failure and identify the fault type. The data used in the experiment come from Jiangsu Qianpeng QPZZ-II rolling bearing fault simulation platform, through which the four running states of the bearing are simulated, which are normal and rolling body faults, respectively. The acoustic emission signal of rolling bearing is collected by SAEU2S acoustic emission system. Firstly, the adaptive morphological filtering method is studied to reduce the noise of acoustic emission signal. A weighted cascade morphological filter is designed. The peak noise of acoustic emission positive pulse is suppressed by open operation, the valley noise of negative pulse is suppressed by closed operation, different denoising indexes are compared, and the correlation index of signal before and after de-noising is introduced as the denoising evaluation index of measured acoustic emission signal. To realize the adaptive adjustment of the morphological closed operation and the open and close operation weight, The simulation results of simulated acoustic emission signal show that this method can effectively filter the pulse noise in the signal. The comparison of SNR and correlation index shows that the trend of the two changes with weight is consistent. It is verified that the correlation index can be used to evaluate the denoising effect of the measured signals with unknown noise intensity instead of SNR. Secondly, the adaptive morphological filtering method is not good for white noise suppression. Combining adaptive morphological filtering with set average empirical mode decomposition (Ensemble Empirical Mode Decompositiontion EEMD), the signal processed by morphological filtering is decomposed by EEMD. The signal is decomposed from high frequency to low frequency into a finite number of intrinsic mode functions reflecting different vibration modes, and the high frequency interference signals are separated. The simulation and comparison of simulated acoustic emission signals show the feasibility of the combined method. Finally, the combined method is used to process the measured acoustic emission signals of the rolling bearing fault simulation platform. The characteristics of frequency domain in different bearing fault states are analyzed and the fault is identified. The experimental results show that the proposed method is effective in the field of bearing signal processing.
【学位授予单位】：沈阳航空航天大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH133.33

【相似文献】