IMSA-SIR Developing Reduced-order Models for Predicting Aerodynamic Performance of Novel Wing Shapes for Aircraft

Argonne National Laboratory

  • Lemont, Illinois
  • 30 days ago

    Highlights

    This dataset is available for developing surrogate models using machine learning and dynamical systems tools to provide predictions for off-design conditions, as well as inform potential geometries for optimal performance. Computational fluid dynamics (CFD) is a widely utilized tool in Aerospace Engineering that can provide detailed information regarding aerodynamic performance; however, it is too costly to implement in real-time performance assessment.

    Numbers & Facts

    LocationLemont, Illinois

    Description

    Internship Description

    Background:

    Computational fluid dynamics (CFD) is a widely utilized tool in Aerospace Engineering that can provide detailed information regarding aerodynamic performance; however, it is too costly to implement in real-time performance assessment. The use of data-driven surrogate models, which can be trained from CFD and experimental datasets, can be utilized for generation predictions for changes in the system. Additionally, these tools can be utilized for extracting complex dynamics into simpler representations which enable greater understanding and utilization. CFD simulations for various airfoil profiles have been performed and data characterizing their aerodynamic performance has been extracted. This dataset is available for developing surrogate models using machine learning and dynamical systems tools to provide predictions for off-design conditions, as well as inform potential geometries for optimal performance.

    Description of Student Internship:

    The flow over 2D airfoil profiles has been setup in the open-source software OpenFOAM. We will provide the structure for running simulations and post-processing the results using tools such as ParaView, as well as in-house python scripts. There is a small database of aerodynamic coefficients that has been collated for these airfoils for various operating conditions which is available for training data-driven surrogate models. The structure of the input and desired outcome of these models will be provided to the student, with the opportunity to explore various ML and related models for predictions of aerodynamic performance.

    Education and Experience Requirements

    Required skills:

    Desired skills are a fundamental understanding of classical mechanics

    Experience with aerodynamics is a plus

    Interest in computing in Linux/Unix environments

    Programming or interest to learn Python, Matlab

    Interest in utilizing machine learning for engineering projects

    Internship Family

    Visiting Student High School Research

    Internship Category

    IMSA Student

    Similar Jobs