Font Size: a A A

Research On Gearbox Compound Fault Extraction Based On Variational Mode Decomposition

Posted on:2019-05-28Degree:MasterType:Thesis
Country:ChinaCandidate:W X WuFull Text:PDF
GTID:2322330545991883Subject:Mechanical engineering
Abstract/Summary:PDF Full Text Request
The working signal of gearbox belongs to non-stationary signal,and it is easy to be disturbed by outside environment,especially in the background of industrial strong noise.At the same time,the faults in industrial production often appear in the form of multiple faults,and the faults with different positions and depths will influence each other,which makes it difficult to extract the fault information.So far,the gearbox fault diagnosis is still a major point and difficulty at the present stage.In view of the existing problems at this stage,and supported by Shan-xi Natural Science Foundation of China(2015011063),make the gear box as the research object,using the new fault diagnosis method,the composite fault signals of gearbox as the research target,take fault feature extraction of target signal in strong noise background,intensive study in the background of the current study on weak composite fault extraction research.The main contents of this paper are as follows:(1)The variational modal decomposition research found that the decomposition result directly affect by the intensity of noise in signals,the decomposition k and the penalty factor?,so the research has been carried out by how to determine the two parameters and how to reduce the noise.(2)Aiming at the problem of noise,using the minimum entropy deconvolution,combined with VMD method,the purpose is to eliminate the strong noise by the reduction capability of the MED,we found that after the introduction of MED method,VMD decomposition results have improved significantly,and the kurtosis improved.(3)To determine the parameters of the problem,the introduction of local mean decomposition as a tool,to determine the decomposition k,and use self correlation and cross(4)correlation coefficient as a criterion to determine the punishment factor ?,at the same time,the LMD method is a set of signal noise reduction ability.In view of the above problems and the current status of fault diagnosis methods,a fault diagnosis method based on MED and LMD is proposed to improve the VMD method.The simulation signal and actual fault are processed by this method,and the fault information is extracted successfully.Base the research above,a new idea for fault diagnosis is proposed.
Keywords/Search Tags:gear box, variational modal decomposition, minimum entropy deconvolution, local mean decomposition, strong noise background, compound fault
PDF Full Text Request
Related items