Comparative Analysis an Early Fault Diagnosis Approaches in Rotating Machinery by Convolution Neural Network

Abstract

In several industrial applications, rotating machinery is widely utilized in various forms. A growing amount of study, in the academic and industrial fields, as a potential sector for the confidentiality of modern industrial labor systems, has been drawing early fault diagnosis (EFD) techniques. However, EFD plays an essential role in providing sufficient information for performing maintenance activities, preventing and reducing financial loss and disastrous defaults. Many of the existing techniques for identifying rotations were ineffective. For the identification of spinning machine faults, many in-depth learning methods have recently been developed. This research report has included and analysed a number of research publications that have higher precision than standard algorithms for detecting early failures in rotating machinery. In addition to the artificial intelligence monitoring (AIM) model, detecting the defects in rotating machine was also realized through the simulation output. AIM framework model is also testing the rotating machinery in three different stages, which is based on the vibration signal obtained from the bearing system and further it has been trained with the neural network preceding. Compared to other traditional algorithms, the AIM model has achieved greater precision and also the other performance measures are tabulated in the result and discussion section.