Recognition For Particles In Lubricating Oil Based On Microfluidics And Image Processing Technology

Oil wear particles monitoring is one of the important means for condition monitoring of mechanical equipment,which can provide information for condition-based maintenance and fault diagnosis of mechanical equipment.Based on the study of the current technologies for oil wear particles monitoring,for the shortage of the current oil wear particles monitoring technologies and the demand of oil wear particles monitoring,a new oil wear particles monitoring method based on microfluidics and image processing technology for particles in lubricating oil is innovatively proposed.(1)an monitoring method of oil wear particles that combines micro-fluidic technology with ferrography technology is proposed.Applying micro-fluidic chip to replace the traditional ferrogram and microscopic imaging realize the analysis for particles in lubricating oil.This method has not only the advantages of high integration,fast result analysis and less sample consumption of micro-fluidic technology,but also the advantages of direct observation of particles morphology,size and other characteristics,mastering the wear condition of friction pair surface,judging the wear types and positions of ferrography technology.(2)an on-line monitoring system of lubricating oil pollution and wear particle recognition that is based on microscopic and image technology is designed.This paper has analyzed some contents including image processing,feature distilling and particles recognition.Firstly the restoration method of the blurred wear particles image is studied.The wear particles image segmentation method based on oil back ground image and Otsu is proposed.Secondly,on the feature extraction aspects,six morphological characteristic parameters are extracted including:area,perimeter,long axis,the ratio of the short-long axis,round and compact degrees.Besides,this paper discusses fractal theory and apply it to measure surface texture of particles,also innovatively extracts HSI' s hue mean and value mean and.It is simpler to apply them to describe particle color than RGB format.So as to make wear particles analysis more diverse than ever before.Finally,automatic recognition of wear particles applying the theory of gray relational analysis is realized.The detective results of oil pollution is quantitative by the meaning of application of this experimental system.Wear particles type can be distinguished.There are two functions in the systerm.And its accuracy reaches 92%.