Structure small failure probability calculation method based on double-layer nested optimization and subset simulation

Abstract

The invention provides a small failure probability calculation method based on a double-layer nested optimization support vector machine and subset simulation. The method comprises the following steps: 1, determining a research object; 2, determining a random variable influencing the key part, and establishing a finite element model; 3, solving a response corresponding to the finite element modelaccording to the finite element model; 4, constructing an initial model according to the current test design; 5, performing inner-layer optimization by adopting the effective set to construct an initial model; 6, carrying out outer layer optimization on two parameters of a penalty factor and a kernel function in the support vector machine through multi-path particle swarm optimization in order toobtain optimized support vector machine parameters; 7, constructing a final support vector machine regression model according to the optimized parameters, and obtaining a final limit state equation; and 8, performing failure probability solving on the final limit state equation by utilizing a subset simulation method to obtain a final failure probability. The method is scientific and good in manufacturability and has wide application and promotion value.

Description

technical field [0001] The title of the present invention is "Calculation Method of Structural Small Failure Probability Based on Double-layer Nesting Optimization and Subset Simulation", which provides a calculation method of structural small failure probability based on double-layer nesting optimization support vector machine and subset simulation A method for calculating the small failure probability of a structure relates to modeling of a failure model and solving the small failure probability, and belongs to the field of calculation of the small failure probability of a mechanical structure. Background technique [0002] With the rapid improvement of global advanced manufacturing technology, high-end equipment innovation fields represented by electromechanical products such as aerospace and medical equipment have put forward high requirements on the failure probability of products. At the same time, relying solely on field tests has been unable to verify the failure pro...

AI Technical Summary

Problems solved by technology

But in the case of unknown checking point, it can only be solved by Monte Carlo or analytical approximation method, which greatly limits the scope of their use

In terms of modeling, the current response surface method has certain limitations in the case of high-dimensional nonlinearity

Images