Research on Aero-Engine Vibration Fault Based on Neural Network and Information Fusion Technology

Wu Yahui, Zhang Dazhi

Abstract

It is an important means for determining the conditions and making fault analysis of the aero-engine by measuring its vibration. Because the different features give different analysis results for vibration fault, in order to integration these information, the results of the different Back Propagation (BP) neural networks were fused by applying the Dempster-Shafer (D-S) evidential theory of the information fusion and the basic belief assignment function was established according to the statistical parameters of the networks. The analysis results from the aero-engine vibration signals show that the information fusion method can improve the reliability of the diagnosis and decrease the uncertainty.

Keywords

D-S evidence theory, vibration, BP neural network, Aero-engine

Download

IMEKO-TC22-2017-017.pdf

DOI

-

IMEKO TC

TC22 - Vibration Measurement

Back to the proceedings