Aircraft wing structure global sensitivity analysis method based on support vector machine

Abstract

The invention discloses an aircraft wing structure global sensitivity analysis method based on a support vector machine, which comprises the specific links of: 1, obtaining sampling points by adoptingan experimental design method based on a CVT (Centroidal Voronoi Tessellation) structure, and obtaining an output value of an aircraft wing finite element model at each corresponding point; 2, carrying out normalization on input variables and output responses; 3, according to distribution of random input variables of a wing structure and numerical characteristics thereof, sampling two groups of input samples to denote as matrices A and B; 4, constructing a matrix Ci, wherein the matrix is obtained after an ith column of the matrix B is replaced with an ith column of the matrix A; 6, by utilizing an aircraft wing proxy model constructed by the support vector machine, calculating wing structure output responses corresponding to the matrices A, B and Ci, and after normalization inversion, respectively denoting as yA, yB and (with reference to the specification); 6, according to a simplification MC method, calculating importance degree indexes si and (with reference to the specification)of each input variable; and 7, carrying out sorting on importance degrees of the input variables so as to provide guidance for reliability analysis, prediction and optimization of the wing structure.

Description

Technical field: [0001] The invention relates to the technical field of safety and reliability, in particular to a reliability sensitivity analysis method, in particular to a support vector machine-based aircraft wing structure global sensitivity analysis method. Background technique: [0002] With the development of science and technology, the demand for critical and complex designs of modern structures is gradually increasing, and machines and equipment in various industries are becoming more and more complex. Among them, the complexity of aircraft as the flower of modern industry is self-evident. At the same time, more and more attention has been paid to the reliability of aircraft, because its reliability not only involves the economic interests and reputation of the country and enterprises, but is also closely related to the safety of life. [0003] In structural safety analysis, two aspects are mainly concerned: reliability analysis and sensitivity analysis. Reliabili...

AI Technical Summary

Problems solved by technology

[0007] The Kriging model has complex matrix calculations and difficult model construction

The radial basis model requires a lot of calculations and is expensive

On the other hand, the neural network has the problem of overfitting

Images