【作者】 李宁；

【导师】 秦树人；

【作者基本信息】 重庆大学 ， 机械电子工程， 2010， 博士

【摘要】 旋转机械运行中产生的振动信号包含了丰富的故障信息,通过对其进行处理与分析,可以得到机械设备零部件的状态变化信息,从而判断出机械的运行状态或故障类型。旋转机械测试信号的处理一直是故障诊断和特征识别的重要研究领域,在此背景下,本文对以下内容进行了研究和阐述:首先介绍了课题的来源、背景、研究现状及其研究意义,对旋转机械故障诊断的基本过程,研究内容做了详细阐述;介绍了论文的主要工作和创新点。研究了旋转机械的典型故障及其振动特征;由于故障而引起的振动其表现形式是多种多样的,为了准确判明引起故障的原因,一般来说,应将振动故障的类型区分清楚,对不同类型的故障以及各自对应的振动特征有明确的了解,只有这样,才能取得故障诊断的成功。因此,在第2章中对旋转机械中的常见典型故障产生的基本理论和相应的振动特征进行了研究。研究了旋转机械测试信号的处理方法,并用于旋转机械典型故障的特征提取与诊断;对旋转机械故障诊断中的振动信号处理方法进行了研究。首先研究了谱分析的最早形式——傅里叶变换,并将其用于汽轮发电机组的轴瓦以及发电机组给水泵的轴瓦振动信号的分析与故障诊断;在比较了传统的离散频谱校正与细化技术后,提出了离散频谱的频率抽取校正法,仿真算例验证了该方法的有效性;研究了包络分析技术、全息谱技术、高阶统计分析技术;研究了主要的时频分析技术:短时傅里叶变换、小波变换、Winger-Ville分布、Hilbert-Huang变换、Chirplet变换;研究了旋转机械的阶比跟踪滤波技术,并以某汽车的加速振动信号为例,说明了改进后的Gabor阶比跟踪滤波方法的正确性和可行性。研究这些先进的振动信号处理方法,对正确地提取旋转机械的故障特征,保证大型旋转机械设备的安全可靠运行,避免巨额的经济损失和灾难性事故发生,提高经济效益和社会效益有重要的意义。介绍了隐马尔科夫模型的基本理论,研究了隐马尔科夫模型的算法及其在故障诊断中的应用;介绍了Markov链基本理论,并通过一个简单的实例把它扩展到了隐马尔科夫模型(HMM);然后重点介绍离散HMM的基本概念、理论以及设计时遇到的三个基本问题;分析了HMM在实际应用中所面临的问题并提出了合理的解决方案,最后,分析了HMM在故障诊断中的作用,并介绍了HMM故障诊断的方法。介绍了一体化旋转机械特征分析仪的研制与应用;介绍了虚拟仪器的产生和发展;对“秦氏模型”进行了简单地描述;根据“秦氏模型”虚拟仪器的思想开发了虚拟式旋转机械特征分析仪,并与一体化仪器技术相结合,形成了QLVC-RM3型一体化旋转机械特征分析仪,将QLVC-RM3型一体化旋转机械特征分析仪应用到了现场,对直流电机驱动的涡轮减速箱振动信号以及转子实验台转子升速振动信号进行了测试与分析,并与B&K公司的PULSE 3560C系统的计算结果进行了对比。最后,总结了该论文的内容,并提出了进一步的研究方向。更多还原

【Abstract】 Vibration signals generated by rotary machinery contain lots of fault information. By analysis into them, state changing of the parts in the mechanical equipment can be recognized, and the fault type can be found out. The test signal process of rotary machinery is always important research field in fault diagnose and character feature recognition. With this background, the following content are researched and described:Firstly, the origin, background, current situation for the paper are introduced, especially the significance and research content of fault diagnose for rotary machinery. The primary work and innovative work for the paper are also contained in this chapter.The typical fault and vibration feature for rotary machinery are researched in detail.The exterior form of vibration caused by fault is various. In order to exactly judge the root, which can cause fault, the vibration fault type should be distinguished clearly. The different type of fault and vibration feature respective should be comprehend too. On in this way, can we achieve the success of fault diagnosis. Then the basic theory and vibration feature of familiar typical faults for rotary machinery are researched deeply in chapter 2.Analysis and process methods for test signal of rotary machinery are researched deeply.Process methods for test signal of rotary machinery fault diagnosis are researched. Firstly, Fourier Transform (FT), the pioneer form of spectrum analysis, is introduced. And FT was used to signal analysis and fault diagnosis for bushes of turbogenerator and water pump. The spectrum correction method of frequency extracting is proposed after compares the traditional spectrum correction method and spectrum zoom, simulation result proves the method validity. Envelopment analysis technology, holospectrum technology, high-order statistic analysis technology are also researched. Time-frequency analysis technologies including short time Fourier transform, Wavelet transform, Winger-Ville distribution, Hilbert-Huang transform, Chirplet transform are also researched. Finally, order tracking filtering method for rotary machinery is investigated deeply. Take one car’s vibration signal during run-up as an example, the correction and feasibility of improved Gabor order tracking filtering method are approved. Research those above process methods for vibration signal is of great importance to extract fault feature of rotary machinery correctly, assure safe and reliable operation of large rotary machinery and equipment, avoid the huge economic losses and catastrophic accidents and improve the economic and social benefits.Introduce the basic theory of hidden Markov model, research algorithm and application in fault diagnosis of HMMIntroduce the basic theory of Markov chains, and through a simple example to extend it to the Hidden Markov Model (HMM). The basic concept, theory and three problems encountered in the design of discrete HMM are the focus in the following content. Analysis the problems faced in practical applications of HMM, and made a reasonable solution. Finally, analysis the role of HMM in fault diagnosis and presents fault diagnosis method based on HMM.Introduce the development and application of integrated rotary machinery feature analyzer.Describe the emergence and development of virtual instrument. Introduce the“Qin’s Model”simply. Develop virtual rotary machinery feature analyzer according to“Qin’s Model”. Combine virtual rotary machinery feature analyzer and integrated instrument technology, QLVC-RM3 integrated rotary machinery feature analyzer is developed. Application of QLVC-RM3 integrated rotary machinery feature analyzer includes testing and analyses of vibration signal generated by turbine reduction gearbox and rotor during run-up, and compare the compute result with PULSE 3560C system of B&K company.Finally, summarize the contents of the paper, and made further research directions.更多还原