A fast topology optimization method based on two-stage deep probabilistic model

By combining a two-stage deep probabilistic model with probabilistic intrinsic orthogonal decomposition and an encoder-decoder convolutional neural network, the problem of high computational cost in existing topology optimization is solved, and fast topology optimization on small sample datasets is achieved, which is applicable to two-dimensional and three-dimensional designs.

CN117540517BActive Publication Date: 2026-07-24浣江实验室 +1
0 Cites 0 Cited by

Patent Information

Application Number
CN202311563975.0
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-11-22
Publication Date
2026-07-24
Estimated Expiration
2043-11-22

AI Technical Summary

Technical Problem

Existing deep probabilistic models are computationally expensive in topology optimization and are overly dependent on training datasets, making it difficult to achieve rapid design on small sample datasets.

Method used

A two-stage deep probabilistic model is adopted, which combines probabilistic intrinsic orthogonal decomposition and an encoder-decoder convolutional neural network. The training dataset is reduced by the linear assumption, and prediction is performed using prior information and Gaussian process. The encoder-decoder network enhances fine features.

Benefits of technology

Achieve fast topology optimization on small sample datasets, reduce computational costs, are applicable to 2D and 3D designs, shorten optimization time, and improve design accuracy.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The application discloses a kind of fast topology optimization methods based on two-stage deep probability model, the steps of this method are as follows: determine the case to be generated design;According to the actual working condition of case, determine the load and boundary condition required in generation design;The combination form of load and boundary condition is sampled using Latin hypercube sampling;According to prior constraint, construct data set;Two-stage deep probability model is constructed, and the first stage design generation is carried out using probability eigenvalue orthogonal decomposition, and the second stage enhancement is carried out based on coding-decoding convolutional neural network;Two-stage deep probability model is trained based on the constructed data set;The trained model is used to carry out generation design under unknown load and boundary condition for case;The structure predicted using model is used as the basis for subsequent further topology optimization, so the optimal topology structure can be obtained.The method of the application is suitable for two-dimensional design and three-dimensional design, and can greatly shorten the optimization time while ensuring the design accuracy.
Need to check novelty before this filing date? Find Prior Art