基于循環(huán)自相關(guān)函數(shù)的滾動(dòng)軸承故障診斷方法研究.pdf

1、華中科技大學(xué)碩士學(xué)位論文基于循環(huán)自相關(guān)函數(shù)的滾動(dòng)軸承故障診斷方法研究姓名：陳智伶申請(qǐng)學(xué)位級(jí)別：碩士專業(yè)：機(jī)械電子工程指導(dǎo)教師：史鐵林;軒建平20090521華 中 科 技 大 學(xué) 碩 士 學(xué) 位 論 文 華 中 科 技 大 學(xué) 碩 士 學(xué) 位 論 文 IIAbstract Rolling element bearing is one of the most important and common parts in the machin

2、ery and equipments， its working condition is directly related to quality and security of the whole machinery as well as the whole production line. How to extract weak fault characteristics of the rolling element bearing，

3、and how to reveal the occurrence of early， weak， potential faults and its development and transfer， are the enormous challenges in the field of equipment condition monitoring and fault diagnosis. For the fault of the rol

4、ling bearing components, the vibration signals show cyclostationary characteristics. According to these characteristics, the fault feature extraction methods of the rolling bearing based on the cyclic auto-correlation fu

5、nction are studied theoretically and verified by experiment in this thesis. The demodulation performance of the cyclic autocorrelation function is analyzed firstly. The mathematical calculation and demodulator process o

6、f the cyclic auto-correlation function for the AM signal model are presented. The simulation analysis for the common AM, FM and AM-FM signal are carried out by using envelope spectrum analysis and cyclic auto-correlation

7、 function respectively. Compared with the envelope spectrum analysis，the cyclic auto-correlation function has a more excellent demodulation performance and is suitable for the fault feature extraction of the rolling bear

8、ing. Based on the sensitivity of the cyclic autocorrelation function to colored noise and noise that has cyclostationary characteristics, the method of cyclic auto-correlation function analysis based on the wavelet packe

9、t feature energy is proposed. After the adoption of wavelet packet decomposition, the original signal is reconstructed using the frequency band in which the feature energy changes larger, so that the interference signal

10、components can be filtered out，and the high frequency vibration related to modulation information is highlighted. The excellent combination of wavelet packet analysis and second-order cyclic statistics enriches the appli

11、cations of wavelet packet in fault diagnosis. Combined with the fault characteristics of rolling bearing, simulation experiments of several typical faults are designed, and the data acquisition and analysis software for

12、rolling bearing is developed with Labview. The comparative analysis of the experiment data is carried out between the envelope spectrum analysis method and the cyclic auto-correlation function method, which verify the ef