Classification of in-Flight Fatigue Cracks in Aircraft Structures Using Acoustic Emission and Neural Networks

摘要

This paper describes the resutls of the analysis of acoustic emission (AE) signals from a fatigue crack growth specimen placed in the empennage of a T-303 Cessna Crusader aircraft. these signals were analyzed using neural networks in order to identify the waveform characteristics of fatigue cracks growing in flight. Aiding in this analysis was the incorporation of known fatigue crack growth waveform signals provided by a laboratory-tested specimen idnetical to the specimen used in the airplane. By analysis of the data, it will be possible to develop a system to monitor and predict the location of growing fatigue cracks in flight.