Study On Variable-scale Evolutionary Adaptive Denoising Method And Diagnosis System For Gearbox Vibration Signal

Gear transmission system is widely used because of its fixed transmission ratio,high transmission efficiency and compact structure.Gearbox is one of the key parts of mechanical equipment,its operation is often affected by the fault of gears and bearings,which can even cause significant loss of life and property.If the fault can be realized or predicted accurately in the early stage,it will be helpful for the maintenance of the equipment and can recover a lot of economic losses.Therefore,it is of great theoretical and practical significance to improve the early fault diagnosis accuracy for equipment and monitor their operation status through intelligent condition monitoring system.The fault diagnosis technology based on vibration signal has been widely used in the engineering field.However,the early weak fault information is usually submerged by the multi-scale heavy noise,and the distribution characteristics of multi-scale heavy noise on different scales are distinct,which bring some difficulties to the early weak fault feature extraction.As a complex system,the compound fault of mechanical equipment often occurs,which results in the shutdown of the equipment.Therefore,it is urgent to carry out the research of multi-scale heavy noise elimination and compound fault diagnosis.This thesis aims at the problem of multi-scale heavy noise elimination and compound fault diagnosis of gearbox.The improvement of signal-to-noise ratio,compound fault separation and intelligent state monitoring system are studied in depth.The main research work of this thesis includes:(1)To solve the problem of improving signal-to-noise ratio based on adaptive denoising algorithm,the denoising effect of the adaptive denoising algorithm based on the evolutionary digital filter on the simulation signal with white noise,band limited noise and multi-scale noise is analyzed.And to solve the problem of multi-scale heavy noise elimination,the characteristics of target information distribution and the difference of noise in different scales are considered.And a variable-scale evolutionary adaptive denoising and reconstructing algorithm is proposed based on the target information identification ability of traditional evolutionary digital filter and the theory of variational mode decomposition.The problem of multi-scale heavy noise elimination is solved by the proposed algorithm,and the effectiveness of the proposed algorithm in noise elimination is verified by simulation and experimental analysis.(2)For the separation of compound signal with gear and bearing fault,a new characteristic separation method based on variable-scale evolutionary adaptive decomposition algorithm is proposed.By calculating the bearing fault characteristic frequency indexes of the decomposed signals,the gear and bearing fault signals are separated from the compound fault signals successfully.In addition,fault features of gear and bearing are also extracted effectively.The simulation and experimental results show that the proposed method is advanced in fault signal separation.(3)Based on the theory of variable-scale evolutionary adaptive denoising algorithm,a smart on-line monitoring system for high-speed railway bearing is developed,in which three kinds of modules are established: data acquisition,data analysis and early warning of bearing fault.The functions of acquisition and storage of signals,time domain signal analysis,spectrum analysis of vibration signals and early warning of bearing fault are realized in the system.By monitoring a high-speed railway bearing experimental platform in an enterprise,the application ability in practical engineering of the variable-scale evolutionary adaptive theory is verified.