Non-Contact Fault Diagnosis of Bearings in Machine Learning Environment

Abstract

Timely detection of faults in bearings can save time, efforts and maintenance costs of rotating equipments. To avoid the physical connection of vibration pickup to the machine tool, a non-contact type vibration pickup has been designed and developed in this study to acquire the vibration data for bearing health monitoring under load and speed variation. Fault diagnosis has been accomplished using a Hilbert transform for denoising the signal. The dimensionality of the extracted features was reduced using Principal Component Analysis (PCA) and thereafter the selected features were ranked in order of relevance using the Sequential Floating Forward Selection (SFFS) method for reducing the number of input features and finding the most optimal feature set. Finally, these selected features have been passed to Support Vector Machines (SVM) and Artificial Neural Networks (ANN) for identifying and further classifying the various bearing defects. A comparative analysis of the effectiveness of SVM and ANN has been carried out. The results reveal that the vibration signatures obtained from developed non-contact sensor (NCS) compare well with the accelerometer data obtained under the same conditions. Classification accuracy achieved by the developed NCS with other sensors reported in the literature compares very well. The proposed strategy can be used for automatic recognition of machine faults which will help in providing early warnings to avoid unwanted and unplanned system shutdowns due to failure of the bearings.