No abstract

Two-phase boiling flows are used in a wide range of engineering applications. One of the most common applications is cooling systems. A refrigerant is injected into a closed-loop under saturation conditions extracting the unwanted heat from the corresponding system and boiling as a consequence. However, flow boiling is a complex process governed by several flow characteristics making it challenging to predict. It is strongly related to the turbulence and shear stress occurring between the two phases. Correlations are usually used to model this process. These correlations are based on global averaged parameters in a certain range of operating conditions, while boiling is also dependent on local effects. Thus, the implementation of a more sophisticated numerical approach is mandatory. Computational Fluid Dynamics (CFD) can provide the needed local flow characteristics to predict all aspects of boiling. In this study, a two-phase model implemented in an open-source CFD solver has been tested on two different fluids: Freon R12 and Carbon Dioxide CO2. R12 was chosen since it is one of the most common fluids employed in cooling technologies and its behaviour has been widely assessed experimentally. In addition, CO2 fluid has been selected as it is the new refrigerant with promising characteristics to replace CFCs such as R12. In particular, CERN/CMS is heading in that direction as they are updating their detector cooling system to CO2 evaporative cooling. The CFD model has been tested for the two fluids in different configurations and under various levels of simplification of the model. This work provides a detailed analysis of the key physical aspects dominating the boiling process such as turbulence and buoyancy effects. Finally, some recommendations on best practices in modelling boiling flows are provided.

<p>Ocean submesoscales are characterized by horizontal scales smaller than approximately 10 km that evolve with timescales of O(1) day. Due to their small size and rapid temporal evolution, they are notoriously difficult to measure. In particular, the associated velocity field is not resolved in current satellite altimetry products. At these scales, surface ocean flows are populated by small eddies, and filaments linked with strong gradients of physical properties, such as temperature. Several recent studies indicate that submesoscale fronts are associated with important vertical velocities, thus playing a significant role in vertical transport. On that account, these fine-scale flows are key to the dynamical coupling between the interior and the surface of the ocean, as well as to plankton dynamics and marine ecology. In spite of their importance, the understanding of submesoscale ocean dynamics is still incomplete. In particular, a relevant open question concerns the role played by the ageostrophic components of the surface velocity field that manifest at these scales.</p> <p>By means of numerical simulations, we investigate ocean submesoscale turbulence in the SQG<sup>+1</sup> model, which accounts for ageostrophic motions generated at fronts, and which is obtained as a small-Rossby-number approximation of the primitive equations. In the limit of vanishing Rossby number, this system gives surface quasi-geostrophic (SQG) dynamics. In this study, we explore the effect of the ageostrophic flow components on the spreading process of Lagrangian tracer particles on the horizontal. We particularly focus on the characterization of pair-dispersion regimes and particle clustering, as a function of the Rossby number, using different indicators. The observed Lagrangian behaviours are further related to the structure of the underlying turbulent flow. We find that relative dispersion is essentially unaffected by the ageostrophic flow components. However, these components are found to be responsible for (temporary) particle aggregation in cyclonic frontal regions. These results appear interesting for the modelling of submesoscale dynamics and for comparison purposes with the new high-resolution surface current data that will be soon provided by the satellite SWOT.</p>