Feature Extraction And Fault Diagnosis Of Acoustic Emission Signals From Rolling Bearings

Rolling bearing is one of the most widely used parts of the rotating machinery. If afailure occurs on the rolling bearing, it may cause a particularly pernicious effect on the entiremechanical equipment’s production efficiency and production safety, resulting in equipmentdamage and economic losses. As a consequence, the condition monitoring and fault diagnosisof rolling bearing are very important. This article is committed to use acoustic emission (AE)signals generated in the rolling bearing to detect the fault. In this study, the AE signalsgenerated by rotating machinery fault simulation testbed under normal condition, bearing balldefect, and inner raceway and outer raceway defects, were respectively measured by SAEU2Sacoustic emission system.Firstly, characteristic parameters of rolling bearing AE signals were computed.Concerning diffetent rolling bearing faults, the typical acoustic emission parameters includingamplitude, energy, rising time, duration time, ring down counts and RMS were compared. Thevarying trend and the sensitivity of these parameters to diffetent rolling bearing faults wereconcluded. Meanwhile, the correlation diagram of amplitude-energy was used to analyze therolling bearing AE signals. And the time domain information of acoustic emission signals wasalso described in detail.Then wavelet time-frequency analysis of rolling bearing AE signal was applied to extractthe rolling bearing fault characteristic frequency more effectively. It can be regarded as thebasis for rolling bearing fault diagnosis.On the basis of the above, the ensemble empirical mode decomposition (EEMD) methodwas applied to adaptively decompose the AE signals in time-frequency domain, and theintrinsic mode functions (IMFs) in different frequency bands were analyzed. Then the energyof IMFs were computed, and a failure feature vector was formed by the principal componentsselected according to the energy contribution analysis. Moreover, using the probabilisticneural network (PNN), the intelligent diagnosis method was applied to acoustic emissionsignals of rolling bearings. Compared with other methods, the proposed method has better recognition rate of faults.