Please use this identifier to cite or link to this item:
Title: Development of fault detection & prediction system for bearing fault analysis in an industrial motor using sensor data
Authors: Sairam, Sekaran
Keywords: DRNTU::Engineering::Electrical and electronic engineering
Issue Date: 2016
Abstract: Fault analysis of Industrial Motors has been an area where there has been tremendous focus in the past few decades. The prime motive behind the research activities in this area is to reduce the maintenance cost and decrease the motor downtime so as to enhance the production and profit. There are various factors that cause the fault in these industrial motors. Some of them are:- fault due to mechanical imbalance; fault due to loose belt; fault due to cracked shaft, etc., but the major contributor to the fault of an industrial motor is the fault due to defects in bearing. Hence, more focus on analyzing the effect of defective bearing on the industrial motor has been increasing among the industries. The age old technique used for bearing fault analysis to utilize the vibration sensor to collect the data and analyze them. But the major drawback in utilizing the vibration sensor is that, this technique is invasive and it requires the vibration sensor to be in physical contact with the environment to collect the vibration data and analyze how the defects in the bearing affect the vibration. As an alternative, a non-intrusive means of obtaining the fault signature is focused in this project. This project focuses on developing and testing code while using data processing tools available in MATLAB to read, process and analyze the changing Stator Power Signature of the fault simulator under the effects of various defective bearings using FFT. A GUI has been developed that provides a clear picture on the effect of each of the different bearing faults at a specific selected frequency. Vibration data is also taken for comparison sake. Further, from the above analysis a machine learning algorithm called the Extreme Learning Machine has been used to automate the FFT analysis, i.e, given the data, the ELM can predict if the faults have occurred or not. The use of a machine learning algorithm introduces the concept of predictive maintenance that further decreases the maintenance cost and increases the efficiency. The Extreme Learning Machine is a novel tool in machine learning that was introduced recently and has gained a lot of research interest. In this project ELM is used for profiling and classifying the bearing faults.
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:EEE Theses

Files in This Item:
File Description SizeFormat 
  Restricted Access
Main article15.34 MBAdobe PDFView/Open

Page view(s)

Updated on Jun 21, 2021

Download(s) 50

Updated on Jun 21, 2021

Google ScholarTM


Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.