Fault Diagnosis For Key Components Of Rotating Machinery Under Varying Speeds

Rotors,bearings and gears are important components of rotating machinery,they play an important role in support,power conversion and power transfer during operation,and their performance status has an important impact on the operation and efficiency of mechanical equipment.In the phase of startup or shutdown,the vibration signal of the rotating machinery no longer have periodicity due to the varying speed,At the same time,non-stationary characteristics such as frequency modulation,amplitude modulation and phase modulation appear in the vibration signal,which results in the failure of the fault diagnosis method based on the stationary operation.In recent years,the research of fault diagnosis of rotating machinery under variable rotating speed has become one of the research focus.The vibration signals collected by the mechanical system have different non-stationary characteristics when the faults occur in different parts of rotor,bearing and gear,lots of methods have been put forward.However,these methods have a low adaptability due to the vibration signals have different characteristics in various rotating speed.Therefore,takes rotors,bearings and gears as the research target,it is of great significance to study the fault diagnosis method of rotating parts suitable for different ratating speed.Supported by the National Natural Science Foundation of China "Muti-timescale Hybrid Signals based Health Management of Rotating Machinery under Varying Operation Conditions"(No.51475455),this dissertation takes rotors,bearings and gears as the research target,aiming at the problems of low signal-to-noise,dependence on the speed measuring device and weak fault information contations,the author researched the noise elimination,speed extraction,fault sensitive frequency band location and fault feature extraction method based on vibration signal is carried out.Main research work is as follows:(1)The common fault types of rotor,bearing and gear are studied.The fault features corresponding to different faults of components under variable speed conditions are pointed out,and the fault vibration models of different components are analyzed and constructed.The time domain,frequency domain and order domain characteristics of the monitoring signal under variable speed conditions are also analyzed by means of simulation,which establishes the basis for the follow-up fault information extraction based on order component identification.Meanwhile,some simulation experiments are also designed,and the corresponding data are obtained under different operating conditions.(2)The vibration signal collected in the engineering practice has strong noise interference,which will overwhelm the useful information about the fault.Address this problem,the SVD denoising algorithm is introduced and the defects of SVD in the vibration signal denoising analysis under variable speed conditions are pointed out.Considering the characteristics of the vibration signal under variable speed conditions,an iterated SVD(ISVD)denoising method is proposed.The effectiveness of the proposed method is verified by a number of simulations and experimental bearing fault signals,which is quantitatively expressed by comparing the amplitude magnitudes of the fault order components.(3)In order to obtain the rotational speed information of the rotor,bearing and gear during the startup and shutdown operation without the aid of tachometer,a solution to for extracting the rotational speed information based on the time-frequency ridge feature of the vibration signal is proposed.In the Short-time Fourier Transform analysis,the minimum information entropy criterion is introduced to adaptively determine the window length,which effectively improves the time-frequency resolution of the signal.At the same time,the advantages and disadvantages of the existing rotational speed extraction method based on ridge detection are analyzed.Based on this,a rotational speed extraction method based on optimal dynamic path planning ridge detection is proposed,which is used under different operating conditions.The efficacy of the proposed method is demonstrated by some experimental signals respectively.(4)In order to overcome the defects of common sensitive frequency band positioning methods in locating the fault information frequency bands of rotor,bearing and gear under the condition of variable speed,combined with the signal order characteristics an order spectrum correlated kurtosis index in proposed based on the analysis of the influence of variable speed on the kurtosis and other indicators.Considering the angular resampling is easy to cause the distortion of the signal resonance band,it is pointed out that the determination of the signal resonance must be completed before the angular resampling.Replacing the spectral kurtosis index in the fast kurtogram with OSCK,an improved kurtogram called Fast Order Spectrum Correlated Kurtogram is proposed.The location of sensitive frequency bands corresponding to different faults of the rotor,bearing and gear under different operating conditions is realized.(5)In actual production applications,there are single or multiple faults in mechanical equipment,hence the rapid fault pattern recognition for single and mixed faults of rotors,bearings and gears is studied.The spectrum characteristics of fault vibration signals of different components have significant differences under the variable speed operating conditions.The frequency spectrum random statistical average feature and the order spectrum random statistical average feature are constructed by means of the random statistical average algorithm for the fast and slow variable speed conditions,respectively.With the Nearest Neighbour Classifier(NNC)algorithm,the pattern classification of spectral features is carried out.A repid and efficient classification of single and mixed faults of one or more components of rotors,bearings and gears under variable speed conditions is achieved.There are 137 figures,23 tables and 191 references in this dissertation.