Fault Diagnosis for Gas Turbine Engine Systems Using Model-Based Techniques

The ability to find fault in engine systems and proactively monitor their progression to remedy the root-cause before it fails is of paramount importance in today’s industry safety. An effecting change in current engine monitoring methods will require insight and understanding of the level of robustness and engineering rigour required to maintain safety and airworthiness standards. A model-based fault detection method compares the engine’s output data to that of a model running simultaneously. A suitable threshold should be selected to avoid false alarms due to measurement non-repeatability and current model uncertainty. Any difference beyond a threshold value signifies that a fault has occurred. This project aims to develop novel methods for fault diagnosis of gas turbines that require fewer data/parameters to work while maintaining or improving the detection and isolation schemes’ accuracy.

Faculty Supervisor:

Afshin Rahimi

Student:

Farshid Bazmi

Partner:

Reticom Solutions

Discipline:

Engineering - mechanical

Sector:

Professional, scientific and technical services

University:

University of Windsor