The Monitoring And Fault Diagnosis Of The BLT5 Artificial Grass Tufting Machine

This paper is based on the transverse project ?The monitoring and faults diagnosis of BLT5 CNC equipment(artificial grass tufting machine)? of our university and the large enterprise in Shandong.Based on the analysis of the common faults of tufting machine,the analysis of monitoring process of tufting machine was completed.The transmission bearings of tufting machine spindle were selected as the research subjects and the general design of monitoring and fault diagnosis system was completed.The researches of structure,vibration mechanism,fault character frequencies and threshold value of bearing were completed.Then the research of analytical methods of vibration signals of bearings-time domain,frequency domain and time-frequency analysis were completed.Then the I-EEMD method was proposed to decompose the signals based on EMD,thus obtained information of different frequency bands.The nonlinear dynamics parameter – fuzzy entropy was proposed as the characteristic value of the signals,SVM based on GA was proposed to accomplish the mode recognition of faults and acquired preferable result.Eventually the design of system database and general program was completed in Lab VIEW.The specific research contents and methods as follow:(1)Completed fully understanding of the past and present development of machine monitoring and fault diagnosis domestic and overseas after looking up a large number of related document and literatures.The research of mechanism,common faults and excitation sources of spindle system was completed.Based on the analysis of monitoring process of tufting machine,the monitoring objections – drive bearings of spindle were confirmed.The general structure design of the system was completed,the mode of data transmission and remote access method was confirmed,and the hardware of system was designed.(2)The time-frequency methods such as STFT,WT and EMD were studied and the ends affections and mode mixing which affects the performances of EMD were studied.Aiming at these problems and former improvements,the adaptive Improved-EEMD(I-EEMD)was proposed.The simulated signal was adopted to testify the effectiveness of I-EEMD.And the I-EEMD was used to decompose the bearing fault signals and the Hilbert envelope spectrum of IMFs validly gained bearing fault information.(3)Based on the study of support vector machine(SVM),the adaptive selection of parameters of SVM based on genetic algorithm was presented.The nonlinear dynamics parameters-Entropy functions were introduced and the research of parameters of fuzzy entropy was completed.Furthermore,EEMD decomposition was proposed to accomplish multilevel analysis of fault signals,use the sensitivity coefficient to select effective IMFs and calculated the fuzzy entropy of them as the fault signal characteristics.Finally they were used as the input to support vector machine to accomplish fault diagnosis.The example verifications were completed and the trained model cans accurately classifying the testing samples and identifying the bearing fault information.(4)The general monitoring and fault diagnosis system was completed based on Lab VIEW by the queue message handler as the program structure.The design of database was accomplished by Access.Matlab was used to accomplish the fault diagnosis methods.The system is able to accomplish acquisition,transmission,storage,analysis and fault diagnosis of signals.