【作者】 王小玲；

【导师】 陈进；

【作者基本信息】 上海交通大学 ， 机械设计及理论， 2012， 硕士

【摘要】 随着科学技术的日益进步与现代工业的飞速发展,机械设备不断向大型、复杂、高速、高效及重载的方向发展;与此同时,其工作和运行环境也更加复杂和苛刻。这些设备一旦突然发生故障,不仅会增加企业的维护成本,降低企业的生产效率,还可能造成巨大的经济损失,甚至导致严重的人员伤亡,产生不良的社会影响。因此,如何对设备进行有效的状态监测和故障诊断,是当前亟需解决的问题。如何有效提取反映设备运行状态的特征,以及准确判断故障类别,一直是故障诊断领域的研究热点,新方法和新理论的研究也层出不穷,对丰富和完善机械故障诊断技术起到了重要作用。本文以滚动轴承为研究对象,提出频带熵方法,并对其在故障诊断中的应用进行了研究,旨在为滚动轴承状态监测提供一个新指标,为故障诊断信号预处理提供一种新方法,论文主要包括以下几个方面的内容:(1)从理论分析与工程应用的角度出发,阐述了论文的选题背景和研究意义。分析了机械设备故障诊断方法、滚动轴承故障诊断、时频分析与信息熵理论等方面的国内外研究现状,确立了本文的研究内容。(2)介绍了作为本文理论基础的几种时频分析方法及信息熵理论,借鉴谱峭度方法提出频带熵概念,定义频带熵为某一频率上(频带内)信号的复杂度,或者说不确定性,给出了频带熵的基本算法,最后从滤波的角度对频带熵概念进行了扩展。(3)介绍了滚动轴承的振动机理和故障特征。讨论了频带熵指标用于滚动轴承状态监测的可能性,对其鲁棒性进行了研究,证明其对奇异点的不敏感性。基于频带熵的上述特性,将其应用于滚动轴承全寿命周期数据分析,探讨了频带熵指标在性能退化各阶段的表现。介绍了为上述理论提供数据支撑的滚动轴承故障试验和加速疲劳寿命试验,通过对试验数据的分析,表明频带熵可作为状态监测指标的有效补充。(4)针对共振解调带通滤波中心频率难以确定的问题,提出了频带熵确定中心频率的方法。对基于STFT的频带熵,讨论了频率离散点数、时频分析窗长、窗函数类型等参数对频带熵的影响;对基于小波包变换的频带熵,讨论了小波包分解层数和小波包函数的选择对频带熵的影响。最后将两种方法应用于仿真和实际的滚动轴承故障诊断。分析结果证明,频带熵能够准确确定信号的共振频带,提升带通滤波和包络解调后的诊断效果。(5)提出频带熵与遗传算法相结合的方法,用于共振解调带通滤波器的优化设计。以频带熵最小为遗传算法的优化目标,通过选择、交叉、变异等操作,在取值范围内寻找最优的中心频率和带宽组合,设计优化滤波器。通过对仿真信号和不同信噪比实验数据的分析,证明此方法能够有效确定滤波中心频率和带宽,从而提高信号的信噪比,实现对轴承故障的诊断。更多还原

【Abstract】 With the rapid development of technology and industry, mechanical equipment has become more and more huge, complex, high-speed, effective, and heavy-load while they must face more and more harsh running conditions. Once they fail unexpectedly, the unexpected failure can increase maintenance cost, reduce production efficiency, and sometimes cause significant economic losses, or even catastrophic accidents. Therefore, it is necessary and important to realize effectively equipment condition monitoring and fault diagnosisHow to effectively extract features of the equipment running status and accurately determine the fault type has been a hot research field of fault diagnosis, and new methods and theory come out one after another to enrich and improve the mechanical fault diagnosis technology. Taking the rolling element bearing as the research object, this paper has proposed frequency band entropy(FBE) method and its application in fault diagnosis has been studied. The writer is aiming at providing a new indicator for the rolling bearing condition monitoring, and a new method for fault diagnosis signal preprocessing, the paper mainly including the following aspects:(1) From the viewpoint of theoretical analysis and engineering application, this paper’s research background and significance of present study are elucidated. A state of the art review is thoroughly completed, which consists of fault diagnosis of machinery and equipment, the rolling bearing fault diagnosis, time-frequency analysis and information entropy theory. The issues to be resolved are summarized and the research content of this paper are established .(2) Several time-frequency analysis methods and information entropy theory which the whole theory is based on are described, then the spectral kurtosis method is introduced to propose the frequency band entropy. Frequency band entropy is defined as the complexity (or uncertainty)of the signal on a particular frequency (or within a particular frequency band), the FBE algorithm is given. At the last the FBE concept is extended from the viewpoint of filtering.(3) The vibration principle and failure characteristics of rolling bearings are described. The possibility of rolling bearing condition monitoring for FBE is discussed as well as its robustness and insensitivity to singular points. Then it’s applied to the whole life data of rolling element bearing, and its performance in various stages of degradation is studied. Rolling element bearing fault test and accelerated fatigue life test are presented to support the above theory, and the analysis of experimental data indicates that FBE can be used as an effective complement of the existed condition monitoring indicators.(4) The determination of band-pass filter center frequency of resonance demodulation is a difficult problem. Time-frequency based FBE is used to solve this problem. For STFT-based FBE, parameters such as the discrete frequency points, analysis window length, window function are discussed; for wavelet packet transform-based FBE, the wavelet packet decomposition level and wavelet packet functions are discussed. Finally, the two methods are used in the simulation and test rolling element bearing fault diagnosis. The analysis results show that the FBE based on time-frequency analysis is able to accurately determine the resonance of the signal and enhance the effect of band-pass filter and envelope demodulation.(5) The method of combining FBE and genetic algorithm is proposed for the optimization of the resonance demodulation band-pass filter design. Minimum FBE as the genetic algorithm optimization goal, by selecting, crossover and mutation operations, genetic algorithm searches the optimal combination of center frequency and bandwidth within the range, then optimal filter is designed. It is proved by the analysis of the simulated signals and different signal to noise ratio experimental data that this method can effectively determine the filter center frequency and bandwidth, thereby improving the signal to noise ratio and the diagnosis of bearing faults.更多还原