Airfoil profile aerodynamic force prediction method, device and equipment and storage medium

CN119537946APending Publication Date: 2025-02-28QILU INST OF TECH
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411596927.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-11
Publication Date
2025-02-28

AI Technical Summary

Technical Problem

In the prior art, airfoil aerodynamic prediction requires a large amount of labeled training data, and the cost of obtaining these data is very high.

Method used

Using a semi-supervised learning method, the data sets of airfoil parameterized control point values ​​and aerodynamic parameter values ​​are obtained, and they are divided into labeled data sets, unlabeled data sets and verification sets. Deep learning is used to build the first aerodynamic prediction model, predict the unlabeled data set and label the pseudo-labeled data. The unlabeled data set is then divided into simple and difficult data sets through the GMM model, the simple data set is iteratively updated and filtered, and a second aerodynamic prediction model is generated.

Benefits of technology

It reduces the need for large amounts of labeled data, reduces the cost of obtaining labeled data, improves the generalization ability of the model, and reduces the occurrence of overfitting.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119537946A_ABST
    Figure CN119537946A_ABST
Patent Text Reader

Abstract

The invention provides an airfoil profile aerodynamic force prediction method, device and equipment and a storage medium, and the method comprises the steps: a data obtaining step: obtaining a data set of an airfoil profile parameterization control point value and an aerodynamic force parameter value, and dividing the data set into a marked data set, an unmarked data set and a verification set; a first aerodynamic force prediction model construction step: constructing a first aerodynamic force prediction model based on deep learning on the labeled data set, training the first aerodynamic force prediction model, predicting the unlabeled data set by using the trained first aerodynamic force prediction model, and labeling pseudo-labeled data; a second aerodynamic force prediction model generation step: dividing the unmarked data set into a simple data set and a difficult data set through a GMM model, performing iterative updating on the simple data set and the difficult data set, and screening the simple data set through a set threshold value to obtain a second aerodynamic force prediction model; training the first aerodynamic force prediction model by using the screened simple data set and the marked data set to generate a second aerodynamic force prediction model; and a verification step: carrying out verification on the verification set by using the second aerodynamic force prediction model.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Cited By

  • Microstructure surface resistance prediction method

    CN121118747A