Distributed collaborative agent model method for approximation analysis of complex engineering structure system

Abstract

The invention belongs to the technical field of mathematical modeling and approximation analysis of engineering systems, and particularly relates to a distributed collaborative agent model method forapproximation analysis of a complex engineering structure system. According to the method, a composite engineering structure system with large-scale parameters and high nonlinearity is classified intoa composite function, an engineering structure system approximation problem is converted into a composite function approximation problem, and the composite function approximation problem is decomposed into an approximation problem of a plurality of levels of sub-functions with few parameters and low nonlinearity; a corresponding sub-agent model for each sub-function of each level is established,and sub-function approximation is performed; a collaborative agent modelis built according to the relationship between each sub-function and each hierarchical function; composite function multi-levelmulti-sub-model collaborative approximation analysis is carried out based on the collaborative agent model; through the decomposition coordination strategy, composite function approximation analysis is realized, the approximation precision can be improved, and the analysis efficiency is improved.

Description

technical field [0001] The invention belongs to the technical field of mathematical modeling and approximation analysis of engineering systems, and in particular relates to a distributed cooperative agent model method for approximation analysis of complex engineering structure systems. Background technique [0002] In many cases, the mechanical structure is a structural system assembled from several components, such as the gas turbine engine rotor structural system, which is assembled from the main shaft, disc, blade and other components. The object of mechanical structure analysis often involves multiple components. Therefore, structural system optimization and reliability design is a complex analysis problem involving multiple objects and disciplines, and the amount of calculation is huge. If the structural system is directly optimized and designed as a whole, the general computing platform cannot bear such a large computing burden, and the design efficiency is unacceptabl...

AI Technical Summary

Problems solved by technology

If a "large" surrogate model is established for analysis and design of a real complex structure, although the calculation load is reduced compared with the finite element (or real structural model), the established "large" surrogate model is difficult to reasonably reflect the output response The nonlinear relationship between the model and the input parameters cannot effectively deal with large-scale parameters and other problems, the accuracy of the model cannot be guaranteed, and it cannot meet the engineering requirements

For structural optimization and reliability design problems with tens of thousands of iterative cycles, "big" surrogate models will also face severe computational efficiency problems

Images