Multi-class Fault Diagnosis of Rolling Element Bearing Based on Empirical Mode Decomposition and Entropy Features

Swarup Kumar Laha

International Journal of Manufacturing and Materials Processing

Multi-class Fault Diagnosis of Rolling Element Bearing Based on Empirical Mode Decomposition and Entropy Features

Swarup Kumar Laha

Abstract

Rolling element bearings are one of the most commonly used machine elements in engineering industry. Fault detection and diagnosis of rolling element bearings is essential for prevention of malfunction and failure during operation. The present work deals with rolling element bearing fault diagnosis by using Support Vector Machine (SVM) and Artificial Neural Network (ANN). Four types of bearing conditions are considered in the present analysis: inner race fault (IR), outer race fault (OR), ball fault (BF) and healthy bearing (HB). The vibration signals from bearing housing are acquired through accelerometers. Since the vibration signal from faulty bearing is non-stationary and non-linear Empirical Mode Decomposition (EMD) is a suitable method for analyzing such signals. The vibration signal is decomposed into intrinsic mode functions (IMF) by using EMD method. Statistical features like Shannon Entropy and Approximate Entropy from the IMFs are extracted for training and testing of the SVM and ANN. The trained models are able to classify different kinds of faults with good accuracy.

Full text

The full text of this article is available to subscribers.

On a subscribing campus you are let in by network address, with no login at all.

One-off purchase. Permanent access to this article, no account needed.

A year of online access to every article in International Journal of Manufacturing and Materials Processing. Or choose several journals →

How to cite

Swarup Kumar Laha. Multi-class Fault Diagnosis of Rolling Element Bearing Based on Empirical Mode Decomposition and Entropy Features. International Journal of Manufacturing and Materials Processing. 2020;2(1):01–10.

Download the citation:

Article details

Multi-class Fault Diagnosis of Rolling Element Bearing Based on Empirical Mode Decomposition and Entropy Features

Abstract

Read more of it

Publish your article

Abstract

Related articles

Experimental Investigation of FSW on Al 6061-T6 MG Dissimilar Joints

Tri-Route Orchestration of Sr-Modification, MoldVibration, and T6 Treatment on the Microstructure andPerformance of A356 Alloy

An Overview on Metal Processing using Machine Learning

Adaptive Production Scheduling in Multi-Machine Manufacturing Environments

Characterization of AL-2014 TIG Welded Joints

Second-Life Battery Applications and Recycling Technologies of EV in India: A Comprehensive Review

Read more of it

Publish your article