Bearing fault detection based on interval type-2 fuzzy logic systems for support vector machines

Author

Zarei, Jafar ; Arefi, Mohammad Mehdi ; Hassani, Hossein

Author_Institution

Sch. of Electr. & Electron. Eng., Shiraz Univ. of Technol., Shiraz, Iran

fYear

2015

fDate

27-29 May 2015

Firstpage

1

Lastpage

6

Abstract

A method based on Interval Type-2 Fuzzy Logic Systems (IT2FLSs) for combination of different Support Vector Machines (SVMs) in order to bearing fault detection is the main argument of this paper. For this purpose, an experimental setup has been provided to collect data samples of stator current phase a of the induction motor using healthy and defective bearing. The defective bearing has an inner race hole with the diameter 1-mm that is created by the spark. An Interval Type-2 Fuzzy Fusion Model (IT2FFM) has been presented that is consists of two phases. Using this IT2FFM, testing data samples have been classified. A comparison between T1FFM, IT2FFM, SVMs and also Adaptive Neuro Fuzzy Inference Systems (ANFIS) in classification of testing data samples has been done and the results show the effectiveness of the proposed ITFFM.

Keywords

electrical engineering computing; fault diagnosis; fuzzy logic; fuzzy neural nets; fuzzy reasoning; fuzzy set theory; induction motors; machine bearings; mechanical engineering computing; pattern classification; stators; support vector machines; ANFIS; IT2FFM; Interval Type-2 Fuzzy Fusion Model; SVM; T1FFM; adaptive neuro fuzzy inference systems; bearing fault detection; defective bearing; healthy bearing; induction motor; inner race hole; interval type-2 fuzzy logic systems; size 1 mm; stator current phase; support vector machines; testing data sample classification; Accuracy; Fault detection; Fuzzy logic; Fuzzy sets; Kernel; Support vector machines; Testing; Bearing; Fault Detection; Support Vector Machines; Type-2 fuzzy logic system;

fLanguage

English

Publisher

ieee

Conference_Titel

Modeling, Simulation, and Applied Optimization (ICMSAO), 2015 6th International Conference on

Conference_Location

Istanbul

Type

conf

DOI

10.1109/ICMSAO.2015.7152214

Filename

7152214