Title :
Combining partial least squares and feed forward neural network technologies in a fault detection system with large number of correlated sensors
Author :
Fischer, Daniel ; Szabados, Barna ; Poehlman, Skip
Author_Institution :
Kinectrics, Toronto, Ont., Canada
fDate :
6/24/1905 12:00:00 AM
Abstract :
The paper describes the issues that have to be dealt with by Failure Detection Systems that process a large number of highly correlated signals. As an example of such system, a Failure Detection System responsible for detecting flow restrictions in liquid cooled stator windings of electric power generators is studied. Field data is presented.
Keywords :
electric generators; fault location; feedforward neural nets; least squares approximations; machine testing; stators; correlated sensors; electric power generators; failure detection systems; feedforward neural network technologies; field data; flow restrictions; highly correlated signals; liquid cooled stator windings; partial least squares; Fault detection; Feedforward neural networks; Feeds; Fluid flow; Least squares methods; Neural networks; Power generation; Sensor systems; Signal processing; Stator windings;
Conference_Titel :
Instrumentation and Measurement Technology Conference, 2002. IMTC/2002. Proceedings of the 19th IEEE
Print_ISBN :
0-7803-7218-2
DOI :
10.1109/IMTC.2002.1006949
