Research On Measuring Signal Analysis And Hidden Markov Model Application Of Rotary Machinery

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.