Close

A note on cookies

We use cookies to improve your experience of our website. If you want to find out more see our Privacy Policy

Menu

Multidisciplinary ADjoint Design Optimisation of Gasturbines

People menu

Research Fellows / Open Positions

Early Stage Researcher 4 at ENGYS

Automated CAD-free adjoint shape optimisation based on harmonic coordinates

Note: this position has been filled.

The complexity associated with coupling multiple tools and processes can in many instances constitute a barrier to entry for optimisation methods. In this work task we aim to produce a self-contained fully automated optimisation method based on adjoint guided surface morphing. Central to this approach is the procedural definition of optimal morphing control points (parameters) via analysis of the adjoint surface sensitivity data. A generalised form of the harmonic coordinate morphing method will be employed for this purpose. In particular the method will be extended to allow for arbitrary control point placements and nested morphing and constraint elements.  The final tool will further incorporate a robust re-meshing/mapping workflow so that extreme mesh deformation events will not place a bound on the attainable optimal shape. Strong collaboration with other participants of WP2 and WP5 is expected, especially in the areas of mesh morphing (ESRs 3, 7) and harmonic coordinates (ESRs 3, 6, 7, 15).

Objectives:

  • M6-M12: Verification, adjoint coupling and testing for harmonic coordinate based shape optimisation engine (WP2). Primary objective is to evaluate in more detail the characteristics of and potential issues with the methodology.
  • M12-M18: Development of effective engineering constraints within the harmonic coordinate framework. Potential avenues include: sequential differential constraints and least squares residual minimisation. Generalised formulation of control source elements in the finite volume framework. Development of control element influence scaling laws for complex (1D, 2D and 3D) entities, Secondment to VW (WP5).
  • M18-M24: Optimisation of control point location based on adjoint sensitivity field maps. Methods based on adjoint and stochastic optimisation will be considered. Further structure will be delivered via constraints and inter-control-point forcing functions. Decoupling of control points from cage mesh topology to allow for effective positional optimisation. Automated control mesh adaptive refinement to adequately capture target geometry, Secondment to NTUA  (WP2).
  • M24-M30: Advanced control elements. Support for nested hierarchical control structures for feature subsets, i.e. control element of control elements with built-in constraints. Non-uniform diffusivity of harmonic interpolation to allow intelligent shaping and/or blocking of control element influence. (WP5)
  • M30-M36: Automated and solution integrated re-meshing and solution interpolation. Re-meshing for large deformations, improved ease of use, robustness and speed. Re-meshing triggers, re-meshing of confined zones. Conservative solution interpolation, zonal interpolation and smooth restarting. (WP2)
  • M36-M42: Industrial application and workflow integration. (WP1, 6)

This position is hosted at Engys' Italian office in Trieste. For further details, contact Dr. Paolo Geremia, p.geremia@engys.com

^ Back to Top