No description available.

The primary focus of the present work is to model the entire casting process from filling stage to complete solidification. The model takes into consideration any phase change taking place during the filling process. An implicit volume of fluid (VOF) based algorithm has been employed for simulating free surface flows during the filling process and the model for solidification is based on a fixed-grid enthalpy-based control volume approach. Solidification modelling is coupled with VOF through User Defined Functions (UDF) developed in commercial fluid dynamics (CFD) code FLUENT 6.3.26. The developed model is applied for the simultaneous filling and solidification of pure metals and binary alloy systems to study the effects of filling process on the solidification characteristics, evolution of mushy zone and the final macrosegregation pattern in the casting. The numerical results of the present analysis are compared with the conventional analysis assuming the initial conditions to be a completely filled mould cavity with uniform temperature, solute concentration and quiescent melt inside the cavity. The effects of process parameters, namely the degree of superheat, cooling temperature and filling velocity etc. are also investigated. Results show significant differences on the evolution of mushy zone and macrosegregation between the present analysis and the conventional analysis. The application of present model to simulate three dimensional sand casting is also demonstrated. The three dimensional competetive effect of filling generated residual flow and the buoyancy-induced convective flow pattern cause significant difference in macrosegregation pattern in casting.

L’objet de ce travail est de proposer un modèle adapté pour la simulation du remplissage de moules qui réponde au meilleur compromis entre temps de calcul et précision des résultats. La difficulté est double. Il faut prendre en compte le phénomène de remplissage qui est un problème complexe à frontières libres et les spécificités liées au Verre : viscosité fortement thermodépendante et température de fusion élevée qui nécessite de prendre en compte le rayonnement. Le Chapitre I est consacrée à la partie écoulement du Verre liquide. La bibliothèque numérique Aquilon/Thétis, adaptée pour traiter ce type de problèmes et les couplages thermique air/verre/parois, a été utilisée (Méthode V.O.F pour le suivi de l’interface, méthodes de type Lagrangien augmenté/Projection vectorielle pour le couplage Vitesse-Pression). Pour l’aspect radiatif, différentes approches sont proposées : conductivité radiative équivalente (Chapitre II), méthode explicite directe pour la validation (Chapitre III) et méthode d’harmoniques sphériques ou méthode PN (Chapitre IV). Dans le Chapitre V, la méthode PN retenue est validée dans des cas simples et est appliquée ensuite à des cas avec couplage convectif en géométries complexes et obstacles semi-transparents (1D, 2D et 3D, 2D axi-symétrique et milieu non gris). Une version P1 modifiée est présentée. Les résultats sont assez proches de ceux donnés par la méthode P3 avec des temps de calcul modestes. L’intérêt de ce modèle est qu’il est facilement intégrable dans des codes numériques existants : une seule équation différentielle du second ordre stationnaire à résoudre en 3D / The aim of this study is to propose an adapted model for the simulation of mould filling that must be a compromise solution between computational time and results accuracy. The double difficulty is to take into account the filling phenomenon that is a complex problem due to the presence of free boundaries and to the Glass specificities: viscosity that is highly thermal dependant and high melting temperature that requires taking into account radiation effects. Chapter I is devoted to the melting Glass flow. The numerical libraries Aquilon/Thétis, adapted for solving such type of problems and the thermal coupling between Air/Glass/Walls, has been used. (V.O.F method for front tracking, Augmented Lagrangian/Vector Projection methods for solving Pressure/Velocity coupling). For radiative aspect, different approaches are proposed: equivalent radiative conductivity (Chapter II), direct explicit method for validation (Chapter III) and spherical harmonics method or PN method (Chapter IV). In the Chapter V, the selected PN method is validated through simple cases and is then applied in other cases with convective coupling in complex geometries including semi-transparent inclusions (1D, 2D and 3D, 2D axi-symmetric and non grey medium). A P1 modified version is presented. The results are close to those given by P3 method but with reduced computational time. The main interest of this model is that it can be easily implemented in existing numerical codes: a single stationary second order partial differential equation to solve in 3D