142. IDENTIFICATION AND MODELING OF TURBINE ENGINE COMPONENTS FOR FAULT DETECTION AND HEALTH MONITORING

Department: Mechanical & Aerospace Engineering
Faculty Advisor(s): Raymond de Callafon | Robert Bitmead

Primary Student
Name: Chad M. Holcomb
Email: cholcomb@ucsd.edu
Phone: 858-405-2349
Grad Year: 2012

Abstract
This research discusses the use of system identification techniques to estimate and monitor the dynamic and non-linear static behavior of main components in turbo-machinery. The research is sponsored by Solar Turbines and in the first phase of this project the focus is on fault detection and correction of time-based degradation of industrial flow control valves. In this application, a closed loop estimator is used to identify deviations in the non-linear flow characteristic of the device by comparing discrete time adiabatic-flow calculations based bulk pressure and temperature of the gas and nominal flow characteristic of the device to the expected temperature drop based on the Joule-Thompson effect. Real-time estimation techniques are proposed to be used as a degradation metric for use in online health monitoring. The closed loop prediction can then extended to adaptive correction of the flow characteristic.

« Back to Posters or Search Results