| Nowadays,with the rapid developing of the mechanical equipment,the unexpected failure of mechanical equipment would bring in serious consequences.Hence,it's of great significance to develop fault diagnosis technology to guarantee the long-term safe and smooth running of the mechanical equipment.Considering the great influence of the high-speed railway transportation to the national economy,it's very important to ensure the railway transportation's high-speed,safe,smooth and constant running.As train bearing failure is the most common train failure,it's urgent to develop train bearing fault diagnosis techniques.At the same time,as the low cost wayside acoustic diagnostic system takes advantage of non-contact measurement and it could detect initial failure in time,it has drawn a lot attention in recently years and developing wayside acoustic diagnostic system is in line with our country.To solve the existing problem for developing wayside acoustic diagnostic system,this dissertation proposed short-time based microphone array techniques to separate and correct the Doppler distortion aliasing acoustic signal acquired by the microphone array.Afterwards,the monitored train bearing condition is accurately evaluated based on the separated and corrected acoustic signal.Finally,a fast,accurate and low-cost train bearing wayside acoustic diagnostic system was developed.First,this dissertation thoroughly introduced the process to acquire the wayside acoustic array signal of moving train bearing,which involved the type and the detail parameter of bearing in this paper.The bearing failure related characteristic information was also described.Then the self-design experiment set-ups to acquire the wayside acoustic array signal was presented in details.According to some time-frequency analysis method,it was clear that there existed serious Doppler distortion and aliasing in the obtained wayside signal,which makes it more difficult to detect the train bearing failure according to the acoustic signal acquired by the trackside.Besides,the necessity to develop circular microphone array based techniques to tackle the wayside acoustic signal was further emphasized by analyzing the limitations of the diagnosis method that based on signal microphone and linear microphone array acoustic signal.Subsequently,to separate the aliasing signal acquired by the trackside,this dissertation established the basic model of the moving target acoustic source for the circular microphone array set by the trackside.Then a short time oblique projection based spatial directive filter was proposed in this dissertation.The proposed method firstly calculated the array time center for different moving acoustic source based on the constructed time center oblique projection based directive filter.Afterwards,combining the moving speed of the acoustic source,the aliasing received signal was separated and corrected according the short-time oblique projection based spatial directive filter successfully.The simulation and experiments proved that the method could be employed to separate and correct the Doppler distortion aliasing acoustic signal.Besides,the application prospect of the proposed method is more extensive because of taking full advantage of the source spatial information.Afterwards,to accurate correct the Doppler distortion in the received circular microphone array signal,based on the previous established basic model,a short-time sparse singular value decomposition method is proposed tin this dissertation.This method first transformed the sparsity of the direction matrix into the sparsity of the original moving source by introducing the overcmoplete direction matrix.Then the short-time sparse singular value decomposition method was employed to estimate the time-varying spatial position of the moving acoustic source.With the relationship between the acoustic source emission time and the microphone acquisition time,the Doppler distortion signal is finally corrected by interpolation resampling in the time domain.Simulation and experimental verification has displayed that the proposed method could effectively correct the Doppler disortion in the received circular microphone array signal.Moreover,this method show potential application to distinguish the coaxial acoustic source as the circular microphone array employed in this paper could realize the two-dimensional localization.Besides,this method doesn't need the accurate number of the moving acoustic source and has better environment adaptability,which show greater potential application in practical when compared with other method.Finally,considering the conditions of our country's railway transportation,the cost of the wayside acoustic diagnostic system should be further decreased and the diagnostic speed should be further improved.In this dissertation,a simplified online Doppler distortion removal method is realized in a self-designed embedded system.The related parameter for Doppler distortion elimination was calculated with the aforementioned microphone array based method or acquired by the corresponding transducer and then transferred to the embedded system in advance.Afterwards,a simplified online unequal time interval sampling strategy was proposed to resample the Doppler distorted signal acquired by the embedded system.Besides,online enveloping method in the embedded system was introduced into the embedded system to further analysis of the corrected signal.Both simulation and experimental verification indicate the effectiveness of the proposed method and the self-designed embedded system,which shows a new potential realization of fast and low-cost high speed train bearing wayside acoustic diagnostic system.In this research,we tried to develop short-time array based signal processing techniques to separate and correct the Doppler distorted aliasing signal and employed this methods into embedded system for decrease the cost of wayside acoustic diagnostic system,which shows some research foundation and reference value for developing reliable,fast and low-cost wayside acoustic diagnostic system. |
PDF Full Text Request