National Technical University of Athens (NTUA)
Μέθοδοι βελτιστοποίησης στην αεροδυναμική και τις στροβιλομηχανές με χρήση συζυγών τεχνικών, υβριδικών πλεγμάτων και του ακριβούς εσσιανού μητρώου
Abstract
dc:descriptionThis PhD thesis contributes to the field of CFD with a complete methodology for solving turbulent flows on 2D and 3D hybrid grids. The method is validated through an adequate set of 2D and 3D case studies of internal and external aerodynamics. Proposed numerical schemes the interpolation of nodal values to the finite volume edges and their contribution to the numerical fluxes must be reported. The current thesis contributes to the field of optimization as well with the development of a novel method for computing first and second order sensitivities using the discrete adjoint method. The method covers functional used in inverse design and total pressure losses minimization problems and utilizes their so computed sensitivities in designing optimal geometries subject to geometric and flow constraints using deterministic optimization methods. A Newton method was implemented which uses the exactly computed Hessian matrix to rapidly locate the optimum in inverse design and total pressure losses optimization problems. Furthermore the hybridization of the Newton method with quasi Newton methods was also explored to further increase the efficiency of the optimization algorithm. In addition to the aforementioned Newton methods, the a posteriori error analysis method for hybrid grids was also developed which combines the discrete adjoint method and the refinement technique developed in this thesis to accurately predict quantities with practical interest such as the lift, the total pressure losses or the entropy generation in the flow field.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2011
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Zervogiannis, Thomas
- Ζερβογιάννης, Θωμάς
Subjects
dc:subject × 16Rights
- Language dc:language
- gre
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/24336