A significant level of research is currently being carried out in the development of driver support systems as they are expected to play a key role in minimizing road vehicle accidents and creating a safe driving environment under harsh weather conditions. However, the performance of some components used by existing driver support systems like LIDAR and visual cameras are affected by extreme weather conditions such as heavy rain fall and snow. Therefore, it is paramount to identify key locations in an automotive vehicle where such systems are least affect by external weather conditions, thereby, improving their overall performance. The field of research that deals with such questions from a simulation perspective is called contamination modeling. At the moment, one of the biggest knowledge gaps in this field is how to consider the effect of different materials on the movement of liquids such as water on different automotive surfaces like glass, plastic, rubber and painted metal. The work presented in this research study has been carried out to investigate and establish the most suitable simulation strategies to match numerical predictions with experimental data for flow of water over different automotive surfaces. Following a comprehensive parametric study of simulation parameters, it was found that the most suitable model that can be tweaked to achieve different flow properties with different surfaces is a dynamic contact angle model. The Blended Kistler model available in STAR-CCM+ required specific values for static, advancing and receding contact angles to optimize a surface for a given material. Therefore, droplet experiments of two droplet sizes were initially carried out for all tested materials at different inclinations and necessary flow parameters were recorded. All experiments were carried out using an approach known as light induced fluorescence imaging where the captured images provided a very convenient method for post processing in computational software. Results from droplet experiments showed that water moved quickest on plastic and slowest on glass. Static contact angle measurements were carried out first on horizontal surfaces. Afterwards, the surface was inclined at 15, 30, 45, 60, and 75 degrees to measure changes in contact angle and velocities. The surfaces for glass and painted metal were directly taken from the door of a Volvo S60 while a separate surface was used for plastic and rubber. These results were then used to create simulation setups for rivulets in STAR-CCM+ with the multiphase modeling approach known as volume of fluid. Rivulet simulations were carried out for all four materials at five different inclinations and the results were compared and validated with experimental data. The results show good correlation between numerical predictions for rivulet movement and experimental data emphasising on the possibility of fine-tuning the surfaces of a simulation setup to represent different material properties.

Depuis plusieurs décennies, une importante activité scientifique se concentre sur la description numérique, théorique ou expérimentale de l'hydrodynamique des écoulements multiphasiques. Ces écoulements sont caractérisés par l'existence d'interfaces, et d'une force à l'interface, la tension superficielle, séparant généralement deux fluides non miscibles. Un cas d'étude dans ce contexte est le problème du drainage d'une unique goutte dans une phase continue, l'ensemble étant soumis à la gravité. Ce système fait apparaître des écoulements récemment décrits pour une goutte d'eau dans l'huile de paraffine. Ce système constitue également un modèle simple pour l'étude des propriétés aux interfaces, Mais d'un point de vue numérique, se pose alors le problème de la stabilité des algorithmes pouvant être utilisés. Les effets aux interfaces impliquent en effet des domaines spatiaux très limités dans lesquels les grandeurs physiques entre les deux fluides sont discontinues. D'importants artéfacts numériques peuvent alors être générés dans les simulations et faire perdre la richesse de la physico-chimie du système considéré. Le problème de la simulation d'écoulements multiphasiques intéresse aussi bien le monde académique que le monde industriel. L'objectif de ce travail de thèse est donc d'implémenter les techniques numériques les plus récentes et de développer un code pour permettre la simulation de l'hydrodynamique de systèmes dispersés. Pour parvenir à ce but, il reste encore des problèmes algorithmiques importants à résoudre comme la prise en compte des effets thermocapillaires et thermosolutaux. Ces deux derniers points sont l'objet de cette thèse. / For several decades, an important scientific activity has focused on the numerical, theoretical and experimental hydrodynamics of drops. This work presents numerical results of a single droplet in the gravity field and in non-isothermal conditions. The simulation such a multiphase system is important in both academic and industrial world. This is particularly the case in the field of emulsions, wetting problems and evaporation. To achieve this goal, there are still important algorithmic problems due to the free moving interfaces and the description of capillary effects. Here, a Volume of Fluid technique has been implemented with high order temporal and spatial schemes to preserve the sharpness of the drop interface. The system under consideration is a simplified model consisting in a single water droplet in a continuous paraffin oil phase. These liquids are immiscible and non-compressible and the overall evolution is unsteady. Capillary contributions such as temperature and surfactant dependent surface tension are fully accounted for. This presentation is aimed to show the capabilities of VOF techniques for the simulations of unsteady multiphase systems in non-isothermal configurations. The role of the droplet initial position and temperature field is described with good numerical stability. There are still important problems remaining in the simulation of free interface systems with such a technique. Spurious currents induced by the description of capillarity can in particular come into play. But these latter can be controlled once the droplet average velocity due to drainage becomes large enough.