Welcome to the IODA Project Website
|Shape optimisation of a climate duct, duct shape and streamlines.
Top: original shape, bottom: optimised shape. The shape is parametrised through the CAD description of the surface, using 300 NURBS control points across 8 NURBS patches.
IODA is an Initial Training Network (ITN) funded by the European Commission running from January 2015 to December 2018. IODA follows on from the EC projects FlowHead and About Flow and aims to advance with the systematic integration of adjoint-based design optimisation with CFD into the regular development processes. The developments focus on:
- The missing link to CAD: part design starts from a CAD geometry and has to deliver the optimised geometry in CAD. However, there is currently no practicable way of either optimising directly on CAD geometries or generating a CAD model from the optimisation results. The currently practised manual capture is inefficient and loses important geometric details.
- The missing link between topology and shape optimisation: Topology optimisation delivers ragged surfaces as the topology is based on the volume mesh. The currently employed smoothing techniques mitigate this problem insufficiently and often deliver geometries that do not respect the packaging space or manufacturing constraints, making the optimal shape non-feasible.
- The missing option of considering constrains: Essentially all optimisation tasks in the automotive and turbomachinery industry are actually not primarily driven by pure performance but by constraints (packaging constraints for shape optimisation, trailing-edge radius for thermal life of turbine blades, manufacturability, etc), hence this shortcoming is especially critical. It restricts the usability specifically of shape optimisation methods to just a few selected applications, where constraints do not dominate or can be accounted for post-optimisation – and thereby diminishing the obtained gain in performance.
- 2-5 April 2017 Transferable and Complementary Skills
- 13-17 February 2017 Automatic Differentiation Training
- 15 to 17 November 2016 VKI Training - slides available