Eccentricity of Rotor Prediction of Aero-Engine Rotor Based on Image Identification and Machine Learning

Zihao Zhang,Junkang Guo,Yanhui Sun,Jun Hong
DOI: https://doi.org/10.1115/imece2019-10919
2019-11-11
Abstract:Abstract The eccentricity of rotor seriously affect the vibration and reliability of aero-engine. Due to the machining error of parts, it is very important to accurately predict the error propagation in assembly. A method based on image recognition and machine learning is proposed to predict the eccentricity of rotor. Firstly, by analyzing and calculating the axial and radial runout error data, the error is mainly concentrated in the first 30 orders of the Fourier series. Secondly, based on the mapping relationship between profile trajectory and eccentricity of rotor, the feature information of the profile trajectory is extracted by constructing the complex domain autoregressive (CAR) model for the radial and axial direction error profile trajectory. Then use the finite element method to calculate the rotor eccentricity. Using the feature information as the input of the neural network, the rotor eccentricity is assembled as the output of the neural network, and the radial basis function (RBF) neural network is built to predict the rotor eccentricity. Theoretical and experimental results show that the proposed method has good enforceability, high accuracy, short calculation time and high engineering application value. In addition, this method can not only be applied to predict the eccentricity of aero-engine rotor flange assembly, but also can be used in the general field of interference fit of assembly.
What problem does this paper attempt to address?