Research Fellows / Open Positions

Early Stage Researcher 3 at QMUL

Robust CAD-free and CAD-based optimisation

Note: This position has been filled.

While all the elements needed in the industrial design chain have reached a level of capability that allows application to challenging industrial cases, routine application of adjoint-based optimisation is today still limited to a few examples. Robustness of the workflow needs to be improved in many areas. The About Flow project aims at improving the robustness of the adjoint solvers themselves, such that converged and reliable adjoint sensitivities can be obtained for complex geometries, flows with limited stability and simulations on poor meshes. Here we focus a) on issues of maintaining or improving mesh quality under large deformation, or even moderate deformation on high-aspect ratio grids for high-Reynolds flows, and b) on obtaining optima that are insensitive to parameters such as changing build-space constraints, manufacturing tolerances or other boundary parameters that may vary.

Objectives:

• M6-M18: Robust large mesh deformation via elastic formulation, harmonic coordinates and optimisation-based smoothing. Automatic remeshing steps with local repair or global remeshing, Secondment to ESI. (WP2).
• M18-M26: Integration into workflows, Secondment to OCC. (WP6).
• M26-M34: Multi-objective adjoint solvers using vector AD, robust optimisation using second moment and inexpensive Monte Carlo techniques Secondment to KTW, RRD(WP6).
• M34-M40: Integration with in-house solver for demonstration/evaluation, linking as a plugin into ParaView using the IODA API (WP6).

For further details, contact Dr. Jens-Dominik Müller, j.mueller@qmul.ac.uk. To apply, follow this link.