Skip to main navigation Skip to search Skip to main content

Internal/External Flow Coupling Design of a Flying Wing based on Hybrid parameterization and Adjoint algorithm

  • Northwestern Polytechnical University Xian
  • School of Aerospace Engineering

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

With advancements in science and technology, the aviation industry has significantly raised the expectations for aircraft design. Compared to traditional configuration, flying wing demonstrates exceptional aerodynamic and structural performance, making them a focal point of research in the aviation sector. However, to achieve optimal performance, flying wings usually incorporate the intake and exhaust systems inside the fuselage, substantially enhancing their aerodynamic impact. Consequently, a detailed, feature-based design of these systems is required. In gradient-free optimization design, feature-based parameterization methods for intake and exhaust systems are widely applied. However, these parameterization methods have seen limited application in gradient-based optimization due to challenges in solving the gradients of the objective function with respect to design variables. Therefore, constructing a parameterization method for intake and exhaust system that is compatible with existing adjoint-based gradient optimization methodologies remains a pressing issue. This paper addresses this challenge by establishing a gradient optimization design method based on a hybrid parameterization approach. This method employs the Free-Form Deformation (FFD) technique to model the aerodynamic shape of the flying wing and utilizes a two-dimensional shape blending function and B-spline method to design the cross-sectional shape, centerline, and cross-sectional area distribution of the intake and exhaust systems, enabling detailed design of these systems. On this basis, the gradients of the objective function with respect to the intake and exhaust system parameters are obtained through analytical derivation and are then coupled with the gradients of the objective function with respect to the aerodynamic shape parameters, obtained through the discrete adjoint method. This process results in the formation of a complete gradient matrix. Using the developed optimization design method, an aerodynamic optimization of a flying wing, accounting for the intake and exhaust systems, was conducted. Compared to the initial configuration, the optimization results showed a drag reduction of 6.67 counts. The total pressure recovery coefficient increased from 92.76% to 97.96%, the distortion coefficient decreased from 8.07% to 5.02%, and the optimized intake duct profile met engineering manufacturing requirements. These results, demonstrate the gradient optimization design method based on the hybrid parameterization approach effectively enhances the aerodynamic performance of flying wing layouts and significantly improves the airflow quality at the intake duct exit. This method provides theoretical guidance and methodological support for aircraft design considering intake and exhaust systems.

Original languageEnglish
Title of host publication15th Asia-Pacific International Symposium on Aerospace Technology, APISAT 2024
PublisherEngineers Australia
Pages2056-2065
Number of pages10
ISBN (Electronic)9798331323981
StatePublished - 2024
Event15th Asia-Pacific International Symposium on Aerospace Technology, APISAT 2024 - Adelaide, Australia
Duration: 28 Oct 202430 Oct 2024

Publication series

Name15th Asia-Pacific International Symposium on Aerospace Technology, APISAT 2024
Volume3

Conference

Conference15th Asia-Pacific International Symposium on Aerospace Technology, APISAT 2024
Country/TerritoryAustralia
CityAdelaide
Period28/10/2430/10/24

Keywords

  • B-spline parameterization
  • Gradient-based optimization
  • Hybrid parameterization
  • Shape blending function

Fingerprint

Dive into the research topics of 'Internal/External Flow Coupling Design of a Flying Wing based on Hybrid parameterization and Adjoint algorithm'. Together they form a unique fingerprint.

Cite this