A Fault Diagnosis Method Of Gearbox Based On The Identification Of Wear Debris Dynamic Characteristics

Gearbox is a significant part of transmission machinery for speed control,and wear is the main reason of failure.The lubricating oil in the gearbox contains a large number of wear debris carrying wear information.The online monitoring of the wear status can effectively avoid equipment failures caused by gear failures.However,interference factors such as light and shadows in the gearbox make it difficult to analyze the wear debris.Using on-line visual ferrograph(OLVF: On-Line Visual Ferrograph)for on-line wear monitoring can effectively improve the reliability of fault diagnosis.In order to improve the visual information extraction accuracy of the wear debris in the online monitoring of gear wear,an image segmentation algorithm of the wear debris with the interference suppression function of the online ferrographic reflectance spectrometer is proposed.Aiming at the problem that the reflected light wear debris image is too dark,a wear debris image enhancement algorithm based on brightness saturation enhancement is proposed.The wear debris image is converted into the HSI(Hue Saturate Intensity)color space wear debris image.The brightness and saturation space enhancement function are constructed to enhance the brightness and saturation of the wear debris image.The brightness of the wear debris image is increased by 0.2070-0.3294,and the saturation is increased by 0.106-0.2952.Aiming at the problem that it is difficult to segment the over-brightness and over-dark parts of the reflected light wear debris image,a reflection light wear debris segmentation algorithm based on watershed transform and H-minima transform is proposed,and H-minima change is used to suppress the extreme value in the wear debris image interference,using the watershed algorithm to segment the wear debris image to achieve OLVF reflected light wear debris image segmentation.The over-segmentation phenomenon of the traditional watershed algorithm is improved.The signal-to-noise ratio of the wear debris image is increased by 24.77%,and the root mean square error is reduced by 1.5%,which effectively improves the segmentation accuracy of the wear debris image.The feasibility of this algorithm is verified by the online wear debris concentration monitoring experiment.When the wear debris concentration changes,the wear debris can be accurately segmented and the wear debris concentration index can be obtained.Using the concentration change curve of the wear debris image to quantitatively describe the experimental process.Compared with other similar algorithms,the algorithm in this paper effectively suppresses the influence of reflected light and better retains the image information of the wear debris.Compared with other similar algorithms,the accuracy of the wear debris concentration curve is improved by 40.5%-56.8%,which verifies the feasibility of the algorithm in this paper.