The Study Of Extraction Method For Aerospace Sensitive Bearing Manufacturing Defect Feature And Software Development

Gyroscope is a key part of spacecraft guidance system. The momentum wheelbearing and frame bearing are very important parts of gyroscope. But the size,uniformity and stability of the sensitive bearing friction torque have great influence onensuring the operation precision and performance of the whole space system. However,In actual fact, bearings often appear the manufacturing defects such as clearancetolerance, ultra-poor gully shape roughness ultra-poor roughness, which make thefriction torque of bearings appear larger friction torque or stuck phenomenon. Bearingfriction torque uniformity and stability cannot be guaranteed, Due to the dynamicfriction torque of the bearings can directly and sensitively reflect the characteristic ofthe defect bearing, so through the analysis of bearing friction torque signal by usingreasonable method, extract the signal characteristics which reflect bearingmanufacturing defects, judge the type of bearing manufacturing defects, provide abasis for improving the manufacturing process for bearings, guarantee the normaloperation of the spacecraft and reduce unnecessary economic loss.The test instrument is the LSD-I low-speed dynamic friction torque tester,sensitive bearing defect types in this paper are the inner in bad gully shape, outer ringin bad gully shape, the inner ring in bad roundness, bad roughness in inner ring andouter ring, the outer ring in bad roundness, grinding silk flow in inner ring and outer ring,The comprehensive application using the power spectrum, the probability densityfunction and the empirical mode decomposition (EMD) to analyze friction torquesignal in frequency domain analysis and extract the manufacturing defect of inner andouter rings. Combined with the statistics analysis, judge the manufacturing defect typeof the bearings. At last, on the basis of the above analysis methods to develop a set offault diagnosis software which works for this kind of bearing, initially formed a set of intelligent diagnosis method for this is good for sensitive precision bearing.At the same time, in view of the complexity of bearing manufacturing defectsinfluence on dynamic friction torque, based on the previous analysis, a variety ofextracting random signal analysis methods are theoretically discussed, and furtherrefinement of the way to judge bearing defect types, laid the foundation for the futuredevelopment of a deeper research work.