Research On Transient Component Extraction And Reconstruction Of Train Bearing's Wayside Acoustic Signal

Railway transportation plays an important role in the development of national economy.It is of great importance to ensure the safety of railway transportation.Railway accidents are mostly caused by the failure of train wheels.Therefore,it is of great significance to study the fault diagnosis of train wheels by train wheels.Wayside acoustic monitoring is an effective means of train bearing health monitoring,In this paper,the commonly used wheel bearing NJ(P)3226X1 in our country is taken as the research object,the failure of the inner ring and the outer ring of the wheel bearing is studied respectively,and the transient impact components in the signal of the wayside are extracted and reconstructed.In this paper,the wire-cut method is used to cut the outer ring,the inner ring and the roller of the wheelset bearing NJ(P)3226X1,by setting the point of failure to simulate the corresponding bearing a single point of failure.Wayside signal is obtained by static experiment and dynamic experiment.Due to the Doppler effect in the wayside signal,the time-frequency modulation of the wayside signal poses a challenge to the traditional transient component extraction method.To solve this problem,this paper presents a Doppler modulation algorithm.The algorithm based on Morse acoustic theory of time-domain modulation of the wavelet,the interpolation algorithm using the frequency domain modulation of the wavelet.Using this algorithm,the traditional single-transient wavelet is modulated in the time-domain and frequency-domain,respectively,and the Doppler modulated wavelet matched with the time-frequency structure of the wayside signal can be constructed.With the traditional extraction of rotating machinery fault signal transient impact components commonly used six wavelet:single Laplace wavelet model,single Morlet wavelet model,single harmonic wavelet model,double Laplace wavelet model,double Morlet wavelet model,double harmonic wavelet model,the Doppler modulation algorithm is used to modulate the six kinds of wavelets as a basis function for extracting transient impact components.By decomposing and reconstructing the signal,signal parameters identification and feature extraction can be realized.A common method of signal decomposition is to decompose a signal using a set of basis functions in accordance with certain rules.In this paper,Doppler modulated wavelet as a basis function,the use of matching tracking method to decompose the wayside signal.Taking into account the actual operating conditions of train bearings,the impact of wayside signal components will show the characteristics of microscopic non-periodic.In this paper,we propose a new method of using the single transient Doppler wavelet to peel off the impact components one by one from the wayside signal,the recognition accuracy is improved.The method of stripping the impact components one by one from the trackside signals using a single transient Doppler wavelet is a traversing search method.When the signal length increases,searching based on non-periodic traversal parameters,due to the need to search for more parameters,the search range,so a large amount of computation,practicality is not high.In the actual operating conditions,the impact components in the wayside signal show the microscopic aperiodicity,but the impact component intervals are not significantly different.Therefore,a method of rapid extraction of the impact components of train-borne wayside signals is proposed.First,the interval of shock components in the wayside signals is roughly estimated based on the maximum correlation coefficient method using the equal-period Doppler wavelet.Then the accurate search is carried out within the parameter range near the estimated cycle value to identify the oscillation damping ratio and the system resonance center frequency so that the impact components can be quickly extracted.