Structural reliability analysis method based on self-adaptive agent model

Abstract

The invention discloses a structural reliability analysis method based on a self-adaptive agent model. The method includes the following steps that firstly, according to distribution form of random variables, a small amount of initial input sample points are generated, and the response value of the system is calculated, and the obtained initial input and output sample points serve as initial training sample points for building a Kriging agent model; secondly, a newly added training sample point selection criterion is created, according to the distribution of random variables, a large amount ofrandom candidate training samples are generated and substitute into a learning function for calculation, candidate sample points with the smallest learning function value serve as newly added training sample points, and loop iteration is conducted until convergence occurs; finally, according to obtained final input and output training sample points, a final agent model is constructed, a Monte Carlo simulation method is adopted, the system failure probability and the reliability sensitivity are calculated, and the problems that an existing method only considers the limitations of the characteristics of sample points in an input space, and the accuracy is low and the calculation amount is large in the reliability analysis are solved.

Description

technical field [0001] The invention belongs to the technical field of reliability analysis and evaluation, and relates to a reliability analysis method of mechanical and structural products, in particular to a structural reliability analysis method based on an adaptive agent model. Background technique [0002] With the rapid development of science and technology, many products (such as: aero-engines, ships, automobiles, CNC machine tools, etc.) have more and more complex structures, many functions, and the working environment is harsh. Huge economic losses and casualties. Generally speaking, various uncertainties generated in the process of product analysis, design and operation are the key factors affecting product reliability (such as the uncertainty of material properties, the uncertainty of external loads, the design and manufacturing dimensions tolerances, etc.). Uncertainty is one of the most critical factors affecting high reliability and long life of products. ...

AI Technical Summary

Problems solved by technology

[0010] In order to overcome the problems of low calculation efficiency, poor robustness and only considering the characteristics of the samples in the input parameter space but not in the response space of the existing methods, the present invention proposes a structural reliability analysis based on an adaptive surrogate model method

Images