A high-temperature aerodynamic non-equilibrium factor correction method based on microcosmic data

By using a non-equilibrium factor correction method based on microdata and employing conditional variational autoencoders and symbolic regression partitioning to fit model parameters, the problems of insufficient prediction accuracy and physical interpretability in high-temperature regions are solved, and high-precision simulation of hypersonic flow fields is achieved.

CN121996896BActive Publication Date: 2026-07-14COMP NETWORK INFORMATION CENT CHINESE ACADEMY OF SCI
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
COMP NETWORK INFORMATION CENT CHINESE ACADEMY OF SCI
Filing Date
2025-12-31
Publication Date
2026-07-14

AI Technical Summary

Technical Problem

Existing computational fluid dynamics models for engineering have poor prediction accuracy in the high-temperature region, and purely data-driven models lack physical interpretability and generalization ability, making it difficult to accurately describe high-temperature aerodynamic non-equilibrium effects.

Method used

By constructing a high-temperature aerodynamic non-equilibrium factor correction method based on microscopic data, using a conditional variational autoencoder to learn the data distribution, and combining symbolic regression and hybrid expert ideas, the model parameters are fitted in partitions to construct a non-equilibrium factor correction model, which is then embedded into the CFD engineering code.

Benefits of technology

It achieves high-precision, physically interpretable, and computationally efficient non-equilibrium factor correction, improving the accuracy and efficiency of hypersonic flow field simulation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121996896B_ABST
    Figure CN121996896B_ABST
Patent Text Reader

Abstract

The application provides a high-temperature aerodynamic non-equilibrium factor correction method based on microcosmic data. The method firstly acquires and pre-processes microstate-state simulation data, and constructs a feature space; a conditional variational autoencoder is used to learn data distribution and perform directional sampling, so that a reconstructed data set conforming to physical distribution rules is generated in the current temperature interval; then, a constraint equation is constructed based on the physical relationship of the non-equilibrium factor, the pseudo-temperature is solved by inversion, and a function skeleton containing undetermined parameters is obtained by symbolic regression; subsequently, the function skeleton is globally fitted to obtain a benchmark model, the error field is analyzed, the high-error region is automatically identified according to the set error threshold, the error field is divided into multiple physical sub-regions, and the analytical boundary of each sub-region is extracted; finally, the parameters in the function skeleton are optimized independently for the physical sub-regions, and are integrated into a hybrid expert model composed of a unified function skeleton, multiple partition boundary equations and multiple partition parameters.
Need to check novelty before this filing date? Find Prior Art