Research context
I am an MSCA Doctoral Candidate in the Parallel CFD & Optimization Unit at NTUA, supervised by Prof. Kyriakos Giannakoglou. My project is FairCFD DC8, Adjoint-based Optimisation for Flows exhibiting Chaotic Dynamics, with Toyota Gazoo Racing as the industrial partner.
The problem
Aerodynamic optimization needs gradients that tell us how a change in shape affects a quantity such as drag. Conventional adjoint calculations become difficult when a flow exhibits chaotic dynamics, particularly when the objective depends on a long-time average.
Least Squares Shadowing offers a way to calculate sensitivities in this setting. Its storage and computational requirements are a major obstacle for industrial CFD.
My research
I work on shadowing-based sensitivity analysis and adjoint methods in OpenFOAM. The research programme includes reformulating the shadowing problem and investigating iterative solution methods and data compression to reduce its cost. The intended application is aerodynamic shape optimization around bluff bodies and automotive geometries.
This is ongoing doctoral research. The image below illustrates the automotive application.
Automotive flow visualization illustrating the application. This is not a completed result from my PhD.
Project information
The FairCFD DC8 research programme describes the scientific objectives and industrial collaboration.
Marie Skłodowska-Curie Actions Doctoral Networks.