The Research On High-speed Electric Spindle Bearing Fault Diagnosis Technology Based On The Analysis Of The Time-frequency

The electric spindle technology is developed with the development and needs of themodern advanced manufacturing technology for high-speed cutting and high-speed CNCtechnology. Especially in recent years, with the rapid development of the aerospace,automotive and mold industry and technological progress, more and more ultra-high-speedelectric spindle applications. As the main support in the form of the electric spindle, rollingbearing close to the limit allowed by the design speed, and its life will be greatly influencedin conditions of ultra-high-speed operation of electric spindle. Once the bearing failure willinevitably lead to abnormal electric spindle and the whole machine running. Therefore,spindle bearing fault diagnosis technology is extremely important.In this paper, the more commonly used angular contact ceramic ball bearings for thestudy, make the dynamics simulation on the basis of the bearing vibration analysis and failuremechanism of the bearing. Use time-frequency analysis method to bearings of fault featureextraction and recognition for rolling bearing vibration signal non-linear characteristics.(1) The establishment of the finite element model of the angular contact ceramic ballbearings and bearing dynamics simulation analysis, Get bearing at high speed and easy to thedynamic response of the bearing of the failure of the danger point, for bearing and on-linemonitoring sensor arrangement to provide the necessary help and basis.(2) According to the characteristics of bearing vibration signal nonlinear,propose analgorithm based on EMD and Choi-Willams distribution on the bearing fault featureextraction, the results show that the method can fairly good extraction bearing the inner circle,outer ring and its roller fault frequency, and can be more effective eliminate cross terms.(3) Time-frequency analysis of spectrum to get when2d images.Accordingly, the firston the bearing fault signals S transformation, get when corresponding spectrum, will thegeometric moment invariants image processing applied to the bearing fault recognition,Thecalculation results show that, the same moment can effectively identify fault frequencycharacteristic vector in for the rolling bearing fault.