1.1 Motivation Over the last decade, a lot of research effort has been put into the understanding of the physics of liquid sprays and atomization processes. This research field has been mainly driven and sponsored by the automotive industry since, especially in western Europe, the environmental regulations have imposed on the car industry over the years more and more restrictive requirements in terms of pollutants (NOx, CO) and particle (soot) emission. One of the key points to achieve these requirements is to produce during injection a spray which is as homogeneous as possible and produces as small drop lets as possible, in order to obtain the greatest possible reacting surface for the later combustion process taking place in the piston bowl of the car engine. For typical applications in the chemical industry, fluid atomization has to produce droplets of uniform size with high efficiency and low maintenance costs. The fulfillment of such criteria is most of the time achieved by designing or choosing an appropriate atomization nozzle. But this process still requires substantial experimental work to find out the best nozzle adapted to a particular purpose. Some of the effort spent on such process could be spared if one could understand the nozzle internal flow pattern and its subsequent influence on the jet breakup behavior. Unfortunately, even though the physics of jet breakup are in principle weIl understood, the nozzle internal flow is not and it is very often simply assumed that the flow leaving the nozzle outlet is fully turbulent. This assumption may be justified in some cases, but it demonstrates how little is known about the flow history before the fluid exits from nozzles to form a jet. This is often related to the tiny dimensions of the nozzle - typically in the range of 200 to 500 micrometers -used in practice for which a non-intrusive and detailed measurement of the flow features is very difficult if not impossible. On the other hand, numerical methods capable of dealing with the flow simulation involving gas-liquid interfaces subjected to surface tension effects have also made their appearance over the last decade. With the gradual increase ofthe computational power, these methods are slowly proving that they are becoming a worthy tool for the analysis and understanding of free-surface flows, as Finite-Element methods for structural analysis have. With the help of the numerical methods, it is nowadays possible to accurately simulate in a reasonable amount oftime, depending on the size of the computational grid, all the flow details that are currently not accessible to measurement procedures. This does not mean that numerical simulations can replace experiments, since they have to provide data to validate or invalidate the computational models. But, the purpose of this work is to show that numerical methods based on the Navier-Stokes equations for fluid motion are able to complement experiments and provide better insight into the physics of jet breakup and into the related nozzle internal flow.weiterlesen