Vibration Signal-based Diagnosis of Defect Embedded in Outer Race of Ball Bearing using 1-D CNN

Abstract

This work is for developing a deep learning-based model for design and diagnosis of fault embedded in ball bearing. To overcome disadvantages of traditional methods for fault identification and diagnosis of ball bearing, method of 1-D Convolutional Neural Network (1-D CNN) is used in this work. 1-D CNN is developed for identification and classification of faults embedded in outer race of ball bearing. Adaptive design of 1-D CNN model presents an ability to fuse extraction of features and classification of fault in single learning body. Open source data "Society for Machinery Failure Prevention Technology (MFPT bearing fault dataset)" is used in this work for training and testing purpose. Main focus for using 1-D CNN approach is to get higher accuracy in fault diagnosis and less computational complexity for results.