In this paper the simulation of vibroacoustical signals radiated by the engine turbine at the stationary vibration excitation is carried out for situations when all turbine blades have no defects and one blade has a small fatigue crack. Bispectral analysis is used for diagnostic information processing. It demonstrates that appearance and evolution of the fatigue crack in a blade change intensity of global and local extremums of bispectral modules. Simulation and higher-order spectral analysis of vibroacoustical signals radiated at stationary mode by an engine rotor with defectless and cracked blades indicate changing of global and local intensity extremums of a bispectrum module at initiation and initial development of a fatigue crack. Application of a stochastic neural network provides test object condition classification using results of bispectrum analysis (the crack parameter was theta=0.005,...,0.1). Received results are new and justify efficiency of test object condition recognition using results of higher-order spectral analysis. These results can be used to create a vibroacoustical monitoring system for aircraft engine rotor components.
Decision making of aircraft engine blades condition based on bispectral analysis of the vibroacoustical signal
Entscheidungen über den Zustand von Flugturbinenschaufeln durch zweispektrale Analyse vibroakustischer Signale
2005
7 Seiten, 4 Bilder, 1 Tabelle, 4 Quellen
Aufsatz (Konferenz)
Datenträger
Englisch
Objective Evaluation of Vibroacoustical Quality
SAE Technical Papers | 2004
|Objective evaluation of vibroacoustical quality
Tema Archiv | 2004
|Objective Evaluation of Vibroacoustical Quality
British Library Conference Proceedings | 2004
|