基于信息融合的旋转机械全信息时频分析方法研究

【作者】 王丽雅；

【导师】 韩捷；

【作者基本信息】 郑州大学 ， 机械电子工程， 2005， 硕士

【摘要】 大型旋转机械是石化、能源、冶金等许多行业中的关键设备,随着科学技术的进步与发展,设备的性能越来越先进,结构也越来越复杂,这就使得设备状态监测和故障诊断技术变得越来越重要。转子一个截面上的振动信息一般是由相互垂直的两个径向传感器和一个轴向传感器拾取的,由于转子的涡动特性,任一探头检测到的信息无论在能量的量值还是频谱结构上均不能全面反映机组运行的实际状况。在很多情况下,各通道信息差异很大,从而导致对机组故障的误判误诊。因此采用全信息技术融合转子两通道甚至多通道的信息可以更全面地了解转子的运转情况。另外,工业生产现场还存在大量的非平稳信号,用传统的信号分析方法已远不能满足故障诊断的需要,研究相应的时频分析方法并应用于生产实践,具有重大工程意义。 本文阐述了基于信息融合的全信息技术理论及其在故障诊断中的应用,在此基础上全方位地介绍和探讨了全谱、全息谱、全矢谱技术的数值方法、图谱表达以及工程应用,并对其实用性作以对比,最后得出结论:全矢谱技术具有直观、计算简单、易用、易拓展的特点,能更好地反映转子运转的客观现实,对故障更敏感、更具特色,因此,具有更重要的工程意义。结合工业现场存在的大量非平稳信号,本文还对短时傅里叶变换和魏格纳-威尔分布等时频分析方法进行了理论上的阐述,并对其工程实用性作了探讨,在此基础上,结合全信息技术之一的全矢谱技术和短时傅里叶变换,创新性地提出了短时矢功率谱的概念,对其进行了算法推导,并用工程实际信号验证了该方法的实用性及准确性。最后得出结论:短时矢功率谱可以得到非平稳信号的能量、频率等随时间不断变化的情况,更全面、准确地反映了信号的能量大小及频率构成,完全可以应用于工业生产现场,为故障诊断实践服务。全矢谱技术和短时傅里叶变换是实现该分析方法的关键技术。 本文在理论探讨的基础上,应用MATLAB软件完成了相关功能的实现,包括幅值谱、相位谱、功率谱、全矢谱、全谱、全息谱、短时傅里叶变换、魏格纳-威尔分布、以及短时矢谱、短时矢功率谱等。更多还原

【Abstract】 Large scale rotating machinery is the key equipment in many industries, such as petrochemical, power, metallurgy etc. With the development of science and technology, the capability of the equipment is more and more advanced, and the structure more and more complex. Correspondingly the equipments condition monitoring and the technology of fault diagnosis become more and more’ important. The vibration information of one section usually is obtained from the two sensors that are vertical to each other, and the third one that is horizontal. The information from the single sensor cannot represent the actual spectrum and its power because of the vortex of the rotor. Most the information which come from different sensor is differently so it is easily to get the wrong result .So it is necessary to adopt the full information technology to fuse information from the two or more sensors in order to obtain the factual state of the rotor. In the other hand, there is a mass of non-stationary signal in real production, the traditional analysis method cannot satisfy the need, so it has significance to study the relevant Time-Frequency Analysis method, and put it into the industry.This article expounds the full information technology and its application in fault diagnosis which based on Information-Fusion, farther, it introduces and discusses adequately, numerical method, illustrative plates and application of Full Spectrum, Holo-Spectrum, and Vector Spectrum, follows the comparison of the practicality, and then, it gets the results: the Vector Spectrum method is evident, simple, it can reflect the actual status better, it has better sensitivity, so it has more meaning .Take the mass of non-stationary signal into account ,this paper expatiate the Time-Frequency Analysis such as Short-Time Fourier Transform (STFT) and Wigner-Ville Distribution(WVD) in theory, and discuss its application in industry. This paper bring forward a new conception: Short Vector Power Spectrum (SVPS), and derive out it in arithmetic, and then approve the practicality and veracity by locale signal. Then it gets the conclusion: SVPS can reflect the variety of the power, frequency, etc., along with the time .It reflects the structure of the power and frequency more comprehensively, more precisely, sowe have sufficient reason to put it into industry application. The Vector Spectrum and STFT is the key technology to realize this analysis method.This paper realize the correlative function using MATLAB based on the studying of the theory, including Amplitude Spectrum, Phase Spectrum, Power Spectrum, Vector Spectrum, Full Spectrum, Holo-Spectrum, STFT, WVD, and SVPS, etc.更多还原