Fault detection based on modelling electromechanical drive chain

Author

Füvesi, V. ; Kovács, E.

Author_Institution

Res. Inst. of Appl. Earth Sci., Univ. of Miskolc, Miskolc, Hungary

fYear

2012

fDate

20-22 June 2012

Firstpage

1336

Lastpage

1341

Abstract

A nonlinear model of an electromechanical drive chain used to drive Gamma-log measurement equipment is presented. Locally linear neuro-fuzzy (LLNF) model was developed with LOLIMOT algorithm that is able to capture the dynamic properties of the system over a wide operating range. Some frequently occurring faults were artificially generated and the ability of the fault detection system to capture them was tested. Based on the detected faults, error signal generation was elaborated for end-users using differently structured neural network models. The structures of the used networks were briefly analysed and compared.

Keywords

electric drives; fault diagnosis; fuzzy set theory; neural nets; power engineering computing; Gamma-log measurement equipment; LLNF model; LOLIMOT algorithm; dynamic properties; electromechanical drive chain modelling; end-users; error signal generation; fault detection; locally linear neuro-fuzzy model; nonlinear model; structured neural network models; Circuit faults; Fault detection; Mathematical model; Sensors; Tracking; Training; Wheels; Electromechanical systems; Fault detection; Neural Networks;

fLanguage

English

Publisher

ieee

Conference_Titel

Power Electronics, Electrical Drives, Automation and Motion (SPEEDAM), 2012 International Symposium on

Conference_Location

Sorrento

Print_ISBN

978-1-4673-1299-8

Type

conf

DOI

10.1109/SPEEDAM.2012.6264441

Filename

6264441