Machine Bearing Fault Diagnosis System using Tri-Axial Accelerometer

Bearing fault is the leading cause of the entire rotating machinery system failure. The fault diagnosis of machinery consists of features extraction and fault classification. The main purpose of this research is to develop a platform with less complexity and more accuracy for the early fault detection in machines and classification of faults. This study also focuses to present efficient, low-cost, and more reliable methods for fault detection and diagnosis. The vibrations data was collected from the rotating machinery (motor) with the AX-3DS wireless Tri-axial accelerometer. We obtained dataset vibrational signals of three states of the machines, namely, Normal, inner race bearing fault, and outer race bearing fault. For preprocessing and segmentation purposes we employed Empirical mode decomposition. Only two features namely Skewness (SK) and Root mean square (RMS) were extracted from three axes and fed to the Support vector machine (SVM) classifier. The proposed method yields an average accuracy of 99.8% on the dataset gathered. Such a compact, less costly, and the more accurate system will help industries for early fault diagnosis of machinery.