Affiliation:
1. Tianjin University, 300354 Tianjin, People’s Republic of China
Abstract
The complete reduced-order model of an aeroengine and diagnosis recognition of the static eccentricity fault of the elastic ring squeeze film damper (ERSFD) is the research topic in this study. The finite element method was to set up the complete dynamic model of the “ERSFD–rolling bearing–double rotor–casing” by considering nonlinear bearing and oil film forces. Based on the Karhunen–Loève orthogonal decomposition reduced theories, the reduced-order model of the complete double-rotor systems is obtained. The improved short-time Fourier transform (STFT) + pseudocolor technology and lightweight network mode EfficientNetV2 realize the static eccentricity fault diagnosis and automatic recognition of the ERSFD. The general expression of oil film force and effective reduced-order model are given. The numerical comparison shows that the degree of freedom was reduced to [Formula: see text], and the computational running time was shortened to [Formula: see text] of the full-order model by analyzing the amplitudes and bifurcations at the front and back supports of the casing. It has the highest recognition accuracy and the shorter automatic recognition time of EfficientNetV2 than other network models. The theories and methods used are of guiding significance for effective vibration modeling and order reduction solution of the complete aeroengine system, as well as dynamic simulation and fault mechanism research.
Funder
National Science and Technology Major Project
National Natural Science Foundation of China
Publisher
American Institute of Aeronautics and Astronautics (AIAA)