Fault Detection and Diagnosis

Published: 19 November 2012

Condition monitoring techniques are receiving more attention in the area of Fault Detection and Diagnosis (FDD) in the electrical machines.

Due to their robustness, low cost and easy maintenance the induction machines are widely used in modern industry applications, therefore, we are considering the FDD of induction motor one of our group approaches. The objective of our current research activities is to propose and experimentally evaluate novel methodologies for the cases of broken rotor bars, short circuits in induction motor, or general fault trends that could help predictive maintenance cycles in induction motors and reduction in maintenance cost, while preventing unscheduled downtime which means increased system availability.

Until now, our proposed methodologies were based on the theory of Set Membership Identification (SMI) and Uncertainty Bounds Violation for fault detection and diagnosis approaches. The proposed schemes have been experimentally evaluated for many cases of broken rotor bars such as gradually or partially broken rotor bar, caused by drilling 1/4 bar, then drilling ½ bar in the same rotor bar, and additional to cases of one broken bar, two broken bar, three broken bar.

On the other hand, Principal Component Analysis of the start-up transient and hidden Markov modeling for broken rotor bar fault diagnosis have been successfully utilized for identifying the presence of a broken bar fault.

Moreover, the Convex Discrimination theory for the model free fault detection approach has been also investigated.

The group has already managed to create an electrical motors’ laboratory setup for further real life experimentation.