Fault diagnosis of rolling bearing based on PSO and continuous Gaussian mixture HMM

DOI

10.2991/mmeceb-15.2016.183How to use a DOI?

Keywords

Particle Swarm Optimization; Hidden Markov Model; fault diagnosis; LPC

Abstract

As the hidden Markov model (HMM) has a strong ability of time sequence modeling, the continuous Gaussian mixture HMM is used to establish a model base of the rolling bearing fault. An adaptive particle swarm optimization (APSO) with extremum disturbed operator and dynamic change of inertia weights is introduced to the traditional training algorithm for solving the local extremum problem. The vibration signal is collected for extracting 12 order LPC coefficients as a feature vector through the dispose of adding window. In the given feature vector, the HMM is built for bearing fault condition monitoring and fault diagnosis. Then, different fault conditions experiment are carried out on the motor bearing test-bed. The experiment result shows that the method can use a small amount of samples for training HMM, and it is more effective and has higher classification accuracy in fault diagnosis compared with the traditional training algorithm.

Copyright

© 2016, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Guangchun Liao
AU  - Haiping Zhu
AU  - Kangjun Liu
AU  - JiaWei Liao
PY  - 2015/12
DA  - 2015/12
TI  - Fault diagnosis of rolling bearing based on PSO and continuous Gaussian mixture HMM
BT  - Proceedings of the 2015 2nd International Conference on Machinery, Materials Engineering, Chemical Engineering and Biotechnology
PB  - Atlantis Press
SP  - 911
EP  - 917
SN  - 2352-5401
UR  - https://doi.org/10.2991/mmeceb-15.2016.183
DO  - 10.2991/mmeceb-15.2016.183
ID  - Liao2015/12
ER  -