Application And Study Of Fault Diagnosis Of Roller Bearings Base On The Wavelet Analysis

Roller bearings are one of widely used mechanical parts in rotatingmachines and vulnerable to damage. Many faults of rotating mechanism arerelated to roller bearings. The performance of roller bearings directly affect theperformance of axis,gear and whole equipment. The defectiveness of rollerbearings can result in abnormal vibration and yawp of equipment, even seriousdamage to the equipment and procreant disaster. Thus, developing faultdiagnosis of roller bearings has great practical significance. This paper summarizes procreant cause of faults of roller bearings in theory,constructs models of vary fault states. Vibration signal of roller bearings is verycomplex, including not only going information of bearings self but also manyinformation of other related parts and structures. It is difficult If we only usetime-domain or frequency-domain means analyzing vibration signal to find faultfore-and-aft transformation. If time-frequency speciality are provided at thesame time, diagnostic veracity and reliability will be greatly improved. So weput forward and study a new fault diagnosis technique— time frequencydiagnosis based on wavelet analysis. Because vibration signals of roller bearings in the gear reducer aresensitive to measure than time, we can draw characteristics--signals relative to中北大学学位论文the measure. The characteristic of vibration signal are to be translated intoenergy chart to measure through using series wavelet transform, so thateigenvector of measure-energy is set up. This provides a new method forspeediness diagnosis of roller bearings. In order to diagnose the fault typeprecisely, the fault character frequency of all parts of roller bearings and energycharacter of lowest crunode are drawn through wavelet package multiplayerdecompose and again compose. According to the properity of wavelet packageanalysis, low-frequency coefficient instead of signal growing tendency, we canapproximately estimate the going state of roller bearing. we can also drawcharacter frequency of every frequency segment, efficiently suppress noise andpioneer a new thought for extract the faint signals from the strong backgroundnoise. By way of qualitative estimate eigenvector whether or not goodexpression vibration signal, the fuzzy clustering methods are applied to detectand diagnose roller bearings. together with the character frequency exacted, wecan make certain the fault type. Through processing and analysis of many data,the diagnosis result is satisfactory. It shows that wavelet package analysis cansupply a convincing analysis means for roller bearings fault diagnosis.