Structure non-gradient topological optimization method based on sequence Kriging agent model

Abstract

The invention provides a structure non-gradient topological optimization method based on a sequence Kriging proxy model, and belongs to the field of structure and multidisciplinary optimization design. The method mainly comprises three parts of reduction series expansion of a design domain material field, establishment of the non-gradient topological optimization model and solution of a sequence Kriging proxy model algorithm. According to the method, the topological optimization problem design variables are effectively reduced through reduced series expansion of the material field function, and the topological optimization problem described by the material field within 50 design variables can be effectively solved by using a sequence Kriging agent model algorithm with an adaptive design space adjustment strategy. The method does not need a performance function to design sensitivity information, and is suitable for solving complex multi-field coupling, multidisciplinary and high-nonlinearity topological optimization problems. The method not only inherits the simple form of density method topological optimization, but also enables the boundary of the topological structure to be clearand smooth, facilitates the integration of various commercial and self-developed finite element software, and is a topological optimization method convenient for popularization and engineering application.

Description

technical field [0001] The invention belongs to the field of structure and multidisciplinary optimization design, and relates to a structure non-gradient topology optimization method based on a sequence Kriging agent model. This method is suitable for topology optimization design of complex equipment such as machinery, instrument equipment, aerospace and so on. Background technique [0002] Topology optimization is an important tool to solve the problem of optimal layout design of structural and multidisciplinary optimization materials. The current mainstream methods include variable density method, level set method and progressive optimization method. Most of these methods derive the sensitivity information of the adjoint method based on specific optimization problems, and use gradient optimization algorithms to solve them. They have been successfully applied in the innovative design of machinery, instruments, aerospace and other equipment. However, many problems in actual...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the shortcomings of traditional topology optimization methods that require complex mathematical operations such as deriv...

Images