Document Type

Conference Proceeding

Language

eng

Format of Original

6 p.

Publication Date

5-15-2005

Publisher

Institute of Electrical and Electronics Engineers (IEEE)

Source Publication

2005 IEEE International Conference on Electric Machines and Drives

Source ISSN

0-7083-8988-3

Abstract

In this paper, a condition monitoring vector database (CMVDB) approach for broken bar fault diagnostics of squirrel-cage induction machines is presented. In this approach, a database of so-called "condition monitoring vectors" (CMVs) is generated for healthy and broken bar fault conditions using the time-stepping finite-element method. The CMV consists of the negative sequence components of winding voltages, currents, and impedances, the frequency spectrum sideband components of motor currents, and the space-vectors of motor terminal quantities (currents and voltages) from which the motor magnetic field pendulous oscillations are derived, as well as the motor speed and developed torque. This CMV will serve as the fault index (signature) for the faults under investigation in this work. This database is intended for use as a reference database in an on-line condition monitoring and fault diagnostic system. In this work, artificial intelligence (AI) techniques based on a statistical machine learning approach are used to detect and distinguish the type of fault and its severity based on the on-line measurements of the motor terminal voltages and currents, as well as the motor speed and developed torque, in comparison to the available CMVDB. To demonstrate the proof-of-principle of the database approach, simulation and experimental results for a 2-hp induction motor are given here to verify the viability of this approach

Comments

Accepted version. Published as part of the proceedings of the conference, 2005 IEEE International Conference on Electric Machines and Drives: 29-34. DOI. © 2005 The Institute of Electrical and Electronics Engineers. Used with permission.

demerdash_7784acc.docx (585 kB)
ADA Accessible Version

Share

COinS