|Gas Turbine Condition Monitoring and Fault Diagnosis|
|A range of techniques available for assessing the condition of gas turbine components and detecting faults will be covered.|
|Date:||January 13, 2003 - January 17, 2003|
|Organizer:||von Karman Institute|
|Type of Event:||Conference, International|
The well-known benefits from implementation of condition monitoring and fault diagnosis methods have made them essential for efficient management of modern gas turbines in aero-propulsion or ground applications. Of the different techniques, those employing aerothermodynamic and performance measurement data are the most important, since they provide information not only on the condition of individual components but also on engine performance and its ability to achieve the design target. Although the first such techniques made their public appearance in the early seventies, this is still an active research area today, and methods are still being proposed for improving diagnostic capabilities.
In the present Lecture Series, a wide range of techniques based on aerothermodynamic and performance data will be covered. The first lectures will provide the basic principles for fault diagnosis and the theoretical background of the methods known under the name of Gas Path Analysis. They will be followed by a description of the way these methods are applied to accommodate the characteristics of real life situations and, in particular, the problem of limited available measurements. The background and application of the most commonly used methods, as employed by the major engine manufacturers today, will be presented. Applications will also be presented to illustrate the abilities of the methods to detect anomalous situations. Representative test cases of gas turbine fault conditions and their identification will be discussed.
Subsequent lectures will give an overview of recently developed techniques that are finding their way to application today. These include non-linear methods and methods based on artificial intelligence techniques. The use of different types of neural networks and the application of genetic algorithms to gas turbine diagnostics will be discussed. Methods providing probabilistic information will also be presented, including probabilistic neural networks and Bayesian belief networks.
In lectures covering application aspects, the impact on engine reliability and availability will be addressed. Possibilities for improved engine usage and management through the incorporation of advanced diagnostics in next generation engine controls will also be discussed.
This Lecture Series is intended to accommodate attendees of both novice and advanced levels of technical expertise. The Director of this Lecture Series is Professor K. Mathioudakis of the National Technical University of Athens and the local coordinator is Professor C.H. Sieverding of the von Karman Institute.
|Event record first posted on November 13, 2002, last modified on November 16, 2002|