An aero-engine high-pressure rotor assembly error prediction method

Abstract

The invention discloses an aero-engine high-pressure rotor assembly eccentricity prediction method. The method comprises the following steps: measuring radial and end jump error contour tracks of a matching surface; carrying out feature extraction on the contour track by using an image recognition method; establishing a complex domain autoregression model to obtain an error track feature vector; taking the seam allowance error feature vector as input, the eccentric position coordinate as output, the Gaussian function as a primary function of the intermediate layer, determining the center position of the hidden layer through K-means dynamic clustering, and estimating the weight of the output layer by using a minimum mean square algorithm to construct an RBF radial primary function neural network model; and substituting the measured data into the model for learning and training, performing error evaluation and model correction by using finite element simulation data, and determining neural network parameters so as to complete aero-engine high-pressure rotor assembly eccentricity prediction. According to the method, the seam allowance morphology error and the assembly deformation areconsidered, and the assembly eccentricity of the high-pressure rotor is predicted more quickly and more accurately.

Description

technical field [0001] The invention belongs to the technical field of mechanical assembly, and in particular relates to an assembly error prediction method of an aero-engine high-pressure rotor. Background technique [0002] At present, the high-pressure rotors of aero-engines are designed as multi-stage combined rotors. Due to the existence of parts manufacturing errors, there must be assembly errors such as eccentricity in the rotor. These assembly errors will seriously affect the running performance of the engine, so there are extremely strict requirements on the assembly eccentricity of the rotor in engineering. [0003] Establishing an error transfer model and predicting the assembly eccentricity according to the part accuracy test results can effectively reduce the work of trial assembly and debugging on the assembly site and improve work efficiency. Many researchers use the homogeneous coordinate transformation method, Jacobian matrix method, vector projection meth...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problems of the low prediction accuracy of the rigid body error transfer analysis method and the long and complicated calculation time of the finite element method in the current aero-engine high-pressure rotor assembly error prediction. A New Method for Predicting Assembly Error of Engine High Pressure Rotor

Images