Interpolation Based Reduced Order Model of Unsteady Aerodynamics
Researcher: Rahul Halder
Sponsor: Airbus
Unsteady Aerodynamics involves nonlinearity due to large shock motion over the wing surface and viscous interaction and therefore high fidelity computationally expensive aerodynamic solver like computational fluid dynamics-based model is essential for the accurate prediction of aerodynamic load and moment at transonic regime. Whereas a Machine Learning based Interpolation Reduced Order Model can predict the aerodynamic loads and moments at several orders of lower computational cost at desirable accuracy.
Development of Neural Network based linearized Aeroelastic Solver for the prediction of the Aeroelastic Instabilities.
Neural network based Parametric Reduced Order Model development.