Fault Diagnosis and Size Estimation of Rolling Element Bearing under Time-varying Speed Conditions

Abstract

Rolling element bearings are one of the most significant elements and frequently-used components in mechanical systems. Rolling bearing fault diagnosis and failure prognostics are helpful for preventing equipment failure and predicting the remaining useful life (RUL) to avoid catastrophic failure. Therefore, reliable fault detection is necessary to ensure productive and safe operations. Spall size is an important fault feature for the RUL prediction, and most of research work has focused on estimating the fault size under constant speed conditions. However, estimation of the defect width size under time-varying speed conditions is still a challenge work. In this paper, a novel method is proposed to solve this problem. The influence of speed variation on fault size estimation was investigated in the follow-up study. A resampling method was used to eliminate the effect of speed variation. The defect size can be calculated with the angle duration, which is measured from the identified entry and exit points. To obtain better understanding of defect size estimation, a dynamic vibration model of a defective rolling bearing is established. The changes of contact deformation and force of the defective bearing pattern are obtained. The entry and exit events can be identified by these illustrations. Then, two defect size estimation models are introduced from the small size model to the large model. Based on the edited cepstrum and LMD (EC-LMD) algorithm, entry and exit events were enhanced to achieve a better diagnosis result than the classical methods. An improved LMD method is proposed to eliminate the end effect by the DTW technique. In order to verify the effectiveness of EC-LMD method, the experiment was performed with a fault bearing. The diagnosis results from the experimental data illustrated that the EC-LMD method could improve the diagnosis performance. With the previous research work, a novel signal processing method combining ECLMD, resampling and continuous wavelet transform was proposed for estimating the fault size of rolling element bearing under time-varying speed conditions. The combination method could not only diagnose the bearing fault but also estimate the fault size under time-varying speed condition. In order to prove the effectiveness and stability of this combination method, the real experiments were carried out. The estimation results show that the proposed method can effectively estimate the defect size on the outer race under time-varying speed conditions. An intelligent rolling bearing fault diagnosis method was proposed. EC-LMD was used to pre-process the signal for extracting good features. The feature extraction was done by the MFE. Laplacian score was used to select the fault feature by reorder the scale factors. SVM is used to evaluate the classification performance. The experimental results showed that the different categories of rolling bearings are effective identified by the proposed method.