基于分数阶Fourier变换的往复式压缩机故障诊断方法研究

发布时间：2018-09-10 11:21

【摘要】：往复式压缩机是石油化工行业中常用的设备,设备一旦发生故障,除了影响生产效率、缩减经济效益而且还存在严重的安全隐患。因此,对往复式压缩机进行故障诊断,是十分必要的。然而,往复式压缩机结构复杂、激励源较多,给往复式压缩机的故障特征的提取带来了巨大的困难。阀片是往复式压缩机最容易发生故障的关键部件,其振动信号中隐藏着丰富的调频信号。由于传统时频分析方法难以有效提取阀片的调频故障信号,本文提出基于分数阶Fourier变换的往复式压缩机故障诊断方法。首先对传统的时频分析方法进行分析,通过传统时频方法的研究,发现Wigner分布虽然具有一定的能量聚焦性,但是多分量信号分析中存在着难以解决交叉项问题。接着通过对标准正弦信号、单分量LFM信号、多分量LFM信号以及含噪声干扰的LFM信号进行仿真分析,验证离散分数阶Fourier变换算法的准确性和有效性。分析结果表明,分数阶Fourier变换不仅可以有效提取LFM信号的频率特征,具有高度的时频聚焦性,而且还可以有效的抑制噪声干扰。最后将分数阶Fourier变换应用到阀片故障诊断中,实验结果表明,不同的阀片故障对应不同的频率特征,从而实现往复式压缩机阀片故障的定量诊断。
[Abstract]:Reciprocating compressor is a commonly used equipment in petrochemical industry. Once the equipment breaks down, it not only affects the production efficiency, but also reduces the economic benefit. Therefore, the reciprocating compressor fault diagnosis, is very necessary. However, the structure of reciprocating compressor is complex and there are many exciting sources, which brings great difficulties to the fault feature extraction of reciprocating compressor. Valve blade is the most prone to failure of reciprocating compressor key components, its vibration signal hidden rich frequency modulation signal. Because the traditional time-frequency analysis method is difficult to extract the frequency modulation fault signal of valve chip effectively, a fault diagnosis method of reciprocating compressor based on fractional Fourier transform is proposed in this paper. Firstly, the traditional time-frequency analysis method is analyzed. Through the research of the traditional time-frequency method, it is found that although the Wigner distribution has some energy focusing, it is difficult to solve the cross term problem in the multi-component signal analysis. Then the accuracy and validity of discrete fractional Fourier transform algorithm are verified by simulation and analysis of standard sinusoidal signal, single component LFM signal, multi-component LFM signal and LFM signal with noise interference. The results show that the fractional Fourier transform can not only extract the frequency characteristics of LFM signal effectively, but also suppress the noise interference effectively. Finally, the fractional Fourier transform is applied to the valve plate fault diagnosis. The experimental results show that different valve plate faults correspond to different frequency characteristics, thus realizing the quantitative diagnosis of valve plate faults of reciprocating compressors.
【学位授予单位】：东北石油大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TH45;TH165.3

【参考文献】