Research On Diagnosis Method For Sliding Bearing Based On Cyclostationarity Analysis

Due to the advantages of high speed, stability, impact resistance and strong load capacity, it is widely used in in chemical, metallurgy, iron and steel, aerospace industry. Carrying out the fault diagnosis of sliding bearing is very important for ensuring the stability operation, reducing and avoiding accidents. The main work done in this paper is as follows:First the method that decomposes a ellipse vector into a pair of circle vector is introduced. Then the traditional first order statistics(cyclic mean) and second order statistic(cyclic correlation)are introduced. The traditional cyclostationarity method deal with the rotational machine through only one signal channel which mean that the interrelation of signals in different directions is neglected. To overcome the weakness this paper designed a new method called directional cyclic mean and directional cyclic correlation based on cyclostationarity and the procedure of drawing axis orbits from directional cyclic mean is studied. As the directional cyclic mean and directional cyclic correlation is neither unable to deal with the rotational machine during start up nor closed down process. To solve this problem this paper combined the computed order tracking technology and directional cyclic method. The analysis to the simulated signal in the method presented in this paper gives ideal result.Considering that oil foul whirl and oil whip is the most common fault type the oil foul whirl and oil whip experiment is carried out based on the INV1613 experimental bench so that the detail procedure of oil foul whirl and oil whip is clearly revealed. The analysis of the displacement vibration signal the acceleration vibration signal showed that directional cyclic mean and directional cyclic correlation is able to discover the fault characteristics besides the cyclic correlation has the time-frequency functionality, which proves that the methods in this paper have practical values.Based on the Lab VIEW software platform and hardware produced by NI, a journal bearing measurement system is developed which transform the theory outcomes into project application. The system can be used fo r data acquisition, data logging and analysis.