The Analysis Of Fault Predictable Diagnosis In Motorized Spindle Based On Green-remanufacturing

To deal with the environmental problem that becoming more and more severe,the manufacture industry mustn`t damage the environment without second thought for the economic interest.In order to achieve sustainable development of the manufacturing industry and society,the industry has proposed the development direction of remanufacturing,green manufacturing and intelligent manufacturing.This paper takes the core component in the manufacturing industry,electric spindle,as the research object,and analyzes its fault diagnosis method by machine learning algorithm.This paper then establishes remanufacturing evaluation indicators to determine the value of remanufacturing of every part.Research can be used to lay the foundation for remanufacturing of machine parts and to maximize remanufacturing value.The main research of this paper is as follows:(1)Select the subject.The main structure of the electric spindle and its main faults are analyzed through literature research and internship investigation.By analyzing every parts of electric spindle,we select bearings at two ends of spindle as our key research objects,which are the most vulnerable and most fatigued parts of electric spindle.(2)Establish a diagnostic algorithm.Summarize and analyze the working mechanism of traditional diagnosis methods,mathematical diagnosis methods and intelligent diagnosis methods,including waveform analysis method,spectrum analysis method,support vector machine,K-nearest neighbor and so on.It is determined to use the K-nearest neighbor algorithm as the diagnostic algorithm of this paper.Firstly,PCA processing is performed on the life cycle data of the electric spindle bearing,and then the K-nearest algorithm is applied to the dimension processed by the PCA.Finally,the PCA-KNN mode can accurately determine the working state of the bearing,and the accuracy is better than other machine learning algorithms such as SVM,decision tree,and random forest.(3)Establish remanufacturing evaluation indicators.Through research,it is found that after some parts are remanufactured,they are not bringing any benefits to the society.In order to achieve meaningful remanufacturing and achieve true sustainable development,establishing remanufacturing evaluation indicators is forward-looking.By considering four aspects,economic attributes,resource attributes,environmental attributes and remanufacturing potential,of the parts and referring to the indicators of replacing the new parts in these four aspects,the indicators this paper proposes can finally judge whether the parts have remanufacturing significance.In summary,the PCA-KNN algorithm proposed in this paper has a good performance in judging the working state of the electric spindle bearing.The established remanufacturing evaluation index of parts is also forward-looking and practical.It can play an active role in the development direction of remanufacturing.