Department of Aeronautics & Astronautics
Stanford, CA 94305
Ph.D. Candidate, Aeronautics & Astronautics, Stanford University, 2009-present.
M.S., Aeronautics and Astronautics, Stanford University, 2004.
M.Sci., Physics, University of Bristol, 2001.
You can find a more detailed version of my CV here.
- Computational Fluid Dynamics (CFD)
- Adjoint methodology (Continuous/Discrete/Hybrid)
- Uncertainty Quantification (UQ) of fluid flow simulations
- Goal-oriented Error Estimation and Grid Adaptation
- Simulation of turbulent, compressible, reacting hypersonic flows
A Hybrid Adjoint Approach for Complex Flow Simulations
I am currently implementing a discrete adjoint solver into the ADL's open-source code, SU2. The software already includes a continuous adjoint solver, and adding this discrete capability will help compare these two methods and fully understand how to build a hybrid adjoint, and what features it would want to incorporate. This coding project will also create, alongside the existing continuous code, much of the framework that will be needed in writing the hybrid adjoint solver.
At the same time I am working through the required mathematical derivations for the hybrid adjoint in two- or three-dimensional flows. Many of the difficulties with this have already been explored in a quasi-one-dimensional case, but there are some unique differences in higher dimensions, especially with regard to the treatment of boundary conditions.
Though I have already implemented the hybrid adjoint in the case of supersonic quasi-one-dimensional flow with a simple combustion model, I am also adapting this code to handle shocked flow cases. This requires more careful treatment at the boundaries, including the development of true hybrid boundary conditions.
||Taylor, T.W.R., Palacios, F., Duraisamy, K., Alonso, J.J., "Towards a Hybrid Adjoint Approach for Arbitrarily Complex Partial Differential Equations", AIAA Paper 2012-3342, 42nd AIAA Fluid Dynamics Conference, New Orleans, Louisiana, June 2012.