Έμβλημα Πολυτεχνείου Κρήτης
Το Πολυτεχνείο Κρήτης στο Facebook  Το Πολυτεχνείο Κρήτης στο Instagram  Το Πολυτεχνείο Κρήτης στο Twitter  Το Πολυτεχνείο Κρήτης στο YouTube   Το Πολυτεχνείο Κρήτης στο Linkedin

Νέα / Ανακοινώσεις / Συζητήσεις

Παρουσίαση Διδακτορικής Διατριβής Γεώργιου Λυγιδάκη

  • Συντάχθηκε 04-03-2015 09:50 από Georgios Lygidakis Πληροφορίες σύνταξης

    Email συντάκτη: glygidakis<στο>tuc.gr

    Ενημερώθηκε: -

    Κύρια: απόφοιτος ΜΔΕ/Διδ. ΜΠΔ. Άλλες ιδιότητες: απόφοιτος προπτυχιακός ΜΠΔ
    ΘΕΜΑ ΔΙΔΑΚΤΟΡΙΚΗΣ ΔΙΑΤΡΙΒΗΣ:
    «On the Numerical Solution of Compressible Fluid Flow and Radiative Heat Transfer Problems»

    Τριτη 10 Μαρτίου 2015, ώρα 15:00 στην
    Αίθουσα Συνεδριάσεων Σχολής ΜΠΔ

    Επταμελής Εξεταστική Επιτροπή:
    Δρ. Ιωάννης Νικολός
    Αν. Καθηγητής Σχολής Μηχανικών Παραγωγής & Διοίκησης, ΠΚ (Επιβλέπων Καθηγητής)
    Δρ. Αργύρης Δελής
    Αν. Καθηγητής Σχολής Μηχανικών Παραγωγής & Διοίκησης, ΠΚ
    Δρ. Δημήτριος Ρόβας
    Επ. Καθηγητής Σχολής Μηχανικών Παραγωγής & Διοίκησης, ΠΚ (Fraunhofer Institute, Nuremberg, Germany)
    Δρ. Σωκράτης Τσαγγάρης
    Καθηγητής Σχολής Μηχανολόγων Μηχανικών, ΕΜΠ
    Δρ. Γεώργιος Παπαδάκης
    Reader in Aerodynamics, Imperial College, London, UK
    Δρ. Ιωάννης Αναγνωστόπουλος
    Αν. Καθηγητής Σχολής Μηχανολόγων Μηχανικών, ΕΜΠ

    Περίληψη
    In this study the development of a methodology for the numerical solution of steady-state compressible fluid flow and radiative heat transfer problems is reported. For the representation of computational domains, three-dimensional unstructured hybrid grids with tetrahedral, prismatic and pyramidical elements are employed, discretized by a node-centered finite-volume scheme. Flow modelling is achieved via the Reynolds-Averaged Navier-Stokes (RANS) equations, along with appropriate two-equation turbulence models, namely, k-ε (in three versions), k-ω and SST. For the computation of the inviscid fluxes an upwind method, applying Roe's approximate Riemann solver, is implemented, jointed with a higher-order accurate spatial scheme, while for the viscous ones the required gradients are evaluated with an element-based (edge-dual volume) or a nodal-averaging approach. The iterative approximation of the aforementioned equations is achieved with either an explicit scheme, applying a second-order temporal accurate four-stage Runge-Kutta (RK(4)) method, or an implicit one, implementing the Jacobi or the Gauss-Seidel algorithm. For the prediction of radiative heat transfer in general enclosures through absorbing, emitting, and either isotropically or anisotropically scattering gray media, the time- or non-time-dependent Radiative Transfer Equation (RTE) is employed. Similarly to fluid flow a second-order accurate spatial scheme along with appropriate slope limiters is applied to increase accuracy of the solution, especially at the boundary surfaces' regions, while time integration is obtained with simple iterative approximations or the same to flow model explicit scheme. In order to increase the efficiency of the proposed methodology additional enhancing techniques are used, namely, an edge-based data structure, a parallelization strategy based on the domain decomposition approach and MPI library functions, and an agglomeration multigrid method employed in isotropic or directional formulation for the flow solver and in spatial, angular or nested spatial/angular one for radiative heat transfer algorithm. Finally, the h-refinement technique is incorporated to increase accuracy at pre-selected regions of the examined grid, by enriching them with more nodes during the solution procedure; as a result, the generation of a new denser mesh from the very beginning is avoided. Based on the pre-mentioned methods, an academic CFD code, named Galatea, was developed; it has been validated against three- and quasi-three-dimensional benchmark test cases presented in the open literature, while its results have been compared with wind tunnel experimental data as well as results obtained by acclaimed corresponding solvers, confirming its capability to effectively perform such simulations in terms of accuracy, geometric flexibility and computational efficiency.

© Πολυτεχνείο Κρήτης 2012