Time-Frequency Squeezing Order Tracking For Time-varying Speed Bearing Fault Diagnosis

Rolling bearings are important parts of rotating machinery and are widely used in mechanical equipment.In the event of failure,it may cause significant property damage or even casualties.Therefore,how to perform fault diagnosis and condition monitoring on the key components of rotating machinery under variable working conditions is of great significance.This research is financially supported by the Natural Science Foundation of China(No.51405320,No.51875376 and No.51705349),Jiangsu Province Natural Science Foundation(No.BK20160318)and Science and Technology Project of Suzhou(No.S YG201511).With the aim of local fault features extraction for bearing under variable speed condition,a research on the application order tracking technology based on a time-frequency extrusion analysis for variable speed bearing fault diagnosis is proposed in this thesis,which are detailed in the following.(1)The failure types and characteristics of vibration signals of the bearing with a local fault are analyzed,which provide a theoretical basis for the thesis.Aiming at the problems of time-frequency ambiguity in traditional time-frequency methods,based on the theoretical analysis of traditional time-frequency methods,a new time-frequency squeezing method is introduced to improve the resolution of time-frequency analysis results of variable-speed bearing signals.The time-frequency analysis and comparison results of a simulated signal verify the validity and applicability of the novel time-frequency squeezing method for improving the time-frequency aggregation,also give the basic research ideas for the whole paper.(2)Considering the problem of the cost increase and application limitation caused by the traditional order tracking method,the time-frequency squeezing method is introduced to extract the shaft rotating frequency in the vibration signal,based on which the traditional resampling analysis method for the variable speed bearing fault diagnosis without tachometer is realized.The method can reduce the cost of diagnosis hardware and ensure the efficiency and accuracy of the method.(3)To solve the problems of calculation error in resampling process,this paper obtains the instantaneous fault characteristic frequency in the vibration signal through the time-frequency extrusion method,and introduces the generalized demodulation order tracking method,free from tachometer and resampling.The error caused by resampling is effectively avoided.At the same time,the method also ensures the accuracy of the diagnosis,and provides an idea for the engineering application of bearing fault diagnosis.(4)Time-frequency analysis of the original vibration signal usually fails when the background noise of the mechanical system is strong.With this question,this paper introduces a sparse representation method to filter and denoise the faulty bearing signal,based on which the time-frequency squeezing and generalized demodulation order tracking method of the reconstructed signal can be applied.Therefore,the fault diagnosis of variable speed bearing under strong noise background is a further extension of the application of the order tracking technology.In order to test the validity and practicability of the proposed method,both the traditional method and the proposed method are applied to experimental signals.The comparison results show that the methods are effective and accurate in variable speed bearing fault diagnosis.which is of certain theoretical and practical value for fault diagnosis of rotating machinery under variable rotational speed.