Paper Titles

Compensation of Input Disturbance in EPS System Using Sliding Mode Observer
p.4992

Design of Neural Network for Satellite’s Attitude Control Systems
p.5001

Application of Neural Networks and PID in on Line of Real-Time Temperature Control System
p.5009

A New Hybrid Design of Experiments Approach for Optimization in Structural Acoustics
p.5015

Machine Fault Diagnosis Using MLPs and RBF Neural Networks
p.5021

Design of Gas Injection Unit in an Iranian Onshore Oil Field Development Project
p.5029

Design and Simulation of MEMS Based Tuning Fork Micro-Gyroscope
p.5036

Two Output OTA Biquad Filter
p.5044

Solving Problems in Chaos Control though an Differential Evolution Algorithm with Region Zooming
p.5048

HomeApplied Mechanics and MaterialsApplied Mechanics and Materials Vols. 110-116Machine Fault Diagnosis Using MLPs and RBF Neural...

Machine Fault Diagnosis Using MLPs and RBF Neural Networks

Article Preview

Abstract:

- Fault detection and elimination in industrial machineries can help prevent loss of life and financial assets. In this study four common faults in rotating machineries namely: 1) Mass Unbalance 2) Angular Misalignment 3) Bearing Faults and 4) Mechanical Looseness have been considered. Each of these defects has been created separately on a test rig comprising of an electrical motor coupled to a rotor assembly. A Vibrotest 60 vibration spectrum analyzer has been used to collect velocity spectrum of the vibration on the bearings. Eleven characteristic features have been chosen to distinguish different faults. Based on the acquired data an Artificial Neural Network Multi Layer Perceptrons (MLPs) and Radial Basis Functions (RBF) Neural Network has been designed to recognize each one of the aforementioned defects. After training the Neural Network, it was checked by new data gathered by new experiments and the results showed that the designed network can predict the faults with more than 75% reliability, and it can be a good assistance to an ordinary machine operator to guess the problem and hence make a good decision.

You might also be interested in these eBooks

Mechanical and Aerospace Engineering, ICMAE2011

Info:

Periodical:

Applied Mechanics and Materials (Volumes 110-116)

Pages:

5021-5028

DOI:

https://doi.org/10.4028/www.scientific.net/AMM.110-116.5021

Citation:

Cite this paper

Online since:

October 2011

Authors:

Gholam Hassan Payganeh, Mehrdad Nouri Khajavi, Reza Ebrahimpour, Ebrahim Babaei

Keywords:

Fault Diagnosis, Neural Network (NN), Rotating Machineries, Vibration

Export:

RIS, BibTeX

Price:

Permissions:

Request Permissions

[1] R. Beale , T. Jackson, Neural Computing: An Introduction, Department of Computer Science, University of York (1998).

[2] M. Kia, Neural networks in the content, Printing, Publishing Kian Green Computing, (2008).

[3] A. Fotovati, and T. Goswami, Prediction of elevated temperature fatigue crack growth rates in TL-6AL-4V alloy– neural network approach, Materials & Design 25(2004) 547-554, (2004).

DOI: 10.1016/j.matdes.2004.03.003

[4] M.T. Hagan, H.B. Demuth, and", M. Beale, Neural Network Design, PWS Publishing Company, (1995).

[5] A. Muszynska , Rotordynamics, Taylor & Francis, (2005).

[6] P. Girdhar, Practical Machinery Vibration Analysis and Predictive Maintenance, IDC Technologies, Working together to grow Libraries in developing countries , (2004).

[7] C.C. Wang, Y. Kang, P.C. Shen, Y.P. Chang, and Y.L. Chung , Applications of fault diagnosis in rotating machinery by using time series analysis with neural , , Expert Systems with Applications Journal, (2009).

DOI: 10.1016/j.eswa.2009.06.089

[8] V. Wowk, Machinery vibration Measurement And Analysis, McGraw-Hill, Inc., 1991. TABLE 2. NETWORK TRAINING RESULTS Number of neurons 1 2 3 4 5 6 7 8 9 10 Training percentage 25.

[21] 42.

[28] 57.

[21] 42.

[21] 42.

[28] 57.

[28] 57.

[28] 57.

[28] 57.

[35] 71 Number of neurons 11 12 13 14 15 16 17 18 19 20 Training percentage.

[35] 71.

[53] 57.

[60] 71 50.

[21] 42.

[57] 14 75.

[28] 57 50.

[28] 57 TABLE 3. NETWORK TRAINING RESULTS epoch 200 400 600 800 1000 1200 1400 1600 1800 2000 Training percentage.

[42] 85.

[46] 42.

[57] 14.

[46] 42.

[60] 71.

[64] 28.

[57] 14.

[57] 14.

[60] 71.

[64] 28 epoch 2200 2400 2600 2800 3000 3200 3400 3600 3800 4000 Training percentage 75.

[67] 85.

[60] 71.

[64] 28.

[64] 28.

[67] 85.

[67] 85.

[71] 42.

[67] 85.

[67] 85 Fig 13. Network training percentage Epoch*200 Fig 14. Network training percentage.