Corrosion phenomena result in significant expenditures amounting to around 3 to 4% of the GDP of OECD countries. To assess nuclear installations’ safety, CEA develops, validates and uses thermal-hydraulic simulation tools to model boiling flows at different scales. To improve understanding of the involved mechanisms, the laboratory applies a multi-scale strategy relying on fine scale simulations (DNS, Direct Numerical Simulations of two-phase flows) as “numerical experiments” producing reference data. Then, these data are averaged to compare with macroscopic or averaged models. The dynamics of nucleation cycles and their impact on heat flux partitioning is a key factor in corrosion and critical heat flux triggering.
For many installations, their premature ageing is due to corrosion. In this PhD, the influence of nucleation cycles in pool boiling of binary mixtures (in particular, H2O+HNO3) on corrosion will be addressed. Experiments indicate that corrosion occurs in locations where heat and mass transfers are the strongest, i.e., below the foot of a bubble attached to the heater wall. DNS is crucial to capture and improve our understanding of the strong variations of temperature and concentration near the wall.
To produce these data, the lab has implemented a Front-Tracking algorithm in the open-source thermo-hydraulic (object oriented, C++) code TRUST/TrioCFD that is used to perform simulations of bubble growth in a pure liquid with constant properties. This approach has been coupled with a subgrid model describing the singular phenomena at the triple contact line (liquid/vapor/solid). Now, the candidate will extend the simulations to the binary mixtures [6] and validate them to consider
• Concentration evolution under convection, diffusion and interfacial evaporation;
• Evolution of interfacial temperature and concentration: the saturation temperature varies locally under the influence of local concentration and vapor partial pressure in the binary mixture;
• Variation of surface tension with concentration and temperature (Marangoni effect);
• The consequences of these mechanisms on the subgrid model implemented near the contact line.

After development and validation of these new models, the PhD student will produce DNS of growth and departure of a single bubble in a binary mixture (H2O/HNO3). The goal is to describe the evolution of thermal and chemical conditions over a boiling cycle, from which the consequences on corrosion rate will be inferred. It is expected to bridge these data with corrosion models. The candidate will then be in contact with the corrosion experts at CEA who have developed the experimental installation CONSTANSE UP (COrrosioN Study under heat TrANSfEr and Under reduced Pressure).