Application of Neural Networks to the F/A-18 Engine Condition Monitoring System
Abstract
Neural networks were applied to the Engine Condition and Monitoring System of the F/A-18 aircraft. Due to recent fleet experience with compressor blade failures in flight, neural networks were applied to three engine conditions, flameout due to compressor failures, normal operating conditions, and low oil pressure conditions. An attempt was made to predict compressor failure using the neural networks. A back propagation and back propagation/ Kohonen network were successfully tested in recognizing the various conditions with data previously unseen by the networks. Both networks demonstrated promise in predicting failures although not enough data was available for conclusive results. Keywords: Theses; Data acquisition; Fortran; Naval aircraft.
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1989
- Accession Number
- ADA219820
Entities
People
- Joseph T. Gengo
Organizations
- Naval Postgraduate School