legacy-id: 311832 may refer to multiple records
Now showing 1 - 2 of 2
  • Some of the metrics are blocked by your 
    Publication

    DeepGeo: Deep Geometric Mapping for Automated and Effective Parameterization in Aerodynamic Shape Optimization

    (2024) ;
    Yang, Aobo
    ;
    Li, Jichao
    ;
    Bauerheim, Michaël
    ;
    Liem, Rhea Patricia
    ;

    Aerodynamic shape optimization (ASO) plays a crucial role in aerodynamic design by enhancing an object’s aerodynamic performance while satisfying design constraints. Traditional ASO is costly and time-intensive, requiring extensive manual tuning to configure parameterization methods that define the design space and manipulate the shape geometry. These conventional parameterization methods limit flexibility and require extensive domain prior knowledge on the design tasks and geometric processing, making efficient and accessible ASO a challenge. We present the Deep Geometric Mapping (DeepGeo) model, a neural network-based method to fully automate the parametrization that accelerates ASO and reduces the cost. By leveraging the universal approximation capabilities of neural networks, DeepGeo achieves large shape deformation freedom while maintaining global smoothness, facilitating efficient optimization in the high-dimensional design space. By integrating the Computational Fluid Dynamics (CFD) mesh deformation within its framework, DeepGeo simplifies the ASO pipeline. By eliminating the need for any training dataset and manual hyperparameter tuning, DeepGeo further significantly reduces implementation complexity and cost. We validate our approach through diverse case studies--including two-dimensional circle optimization, three-dimensional Common Research Model (CRM) wing optimization, and Blended-Wing-Body aircraft design--all using identical settings. Results demonstrate performance comparable to state-of-the-art free-form deformation method while requiring minimal human intervention. This work establishes DeepGeo as a transformative tool for automating ASO and democratizing advanced design optimization techniques.

  • Some of the metrics are blocked by your 
    Publication

    DeepGeo: Deep Geometric Mapping for Automated and Effective Parameterization in Aerodynamic Shape Optimization

    (2024) ;
    Yang, Aobo
    ;
    Li, Jichao
    ;
    Bauerheim, Michaël
    ;
    Liem, Rhea Patricia
    ;

    Aerodynamic shape optimization (ASO) is a key technique in aerodynamic designs, aimed at enhancing an object’s physical performance while adhering to specific constraints. Traditional parameterization methods for ASO often require substantial manual tuning and are only limited to surface deformations. This paper introduces the Deep Geometric Mapping (DeepGeo) model, a fully automatic neural-network-based parameterization method for complex geometries. DeepGeo utilizes its universal approximation capability to provide large shape deformation freedom with global shape smoothness, while achieving effective optimization in high-dimensional design spaces. Additionally, DeepGeo integrates volumetric mesh deformation, simplifying the ASO pipeline. By eliminating the need for extensive datasets and hyperparameter tuning, DeepGeo significantly reduces implementation complexity and cost. Multiple case studies using the same parameterization settings, including the two-dimensional circle-to-airfoil optimization, the three-dimensional CRM wing optimization, and the Blended-Wing-Body aircraft optimization, demonstrate DeepGeo’s effectiveness compared to state-of-the-art free-form deformation methods. This research highlights DeepGeo’s potential to automate ASO, making it more accessible and efficient.

      94  97