In this section, several ideas and methods will be proposed for the future studies in the hope of improving the study that have been done during this BEI and solving the issues that were discussed in the previous section 'NUCLEATE BOILING REGIME, Discussion' .

In order to avoid backflow into the studied domain, a different geometry as shown in figure could be considered to ensure all the gas phase can be ejected out of the domain without creating any backflow.

Figure 22: Geometry proposed for future studies

The flow was considered as laminar for single phase flow as well as liquid-gas phase flow. Nevertheless, it was remarked that the gas phase (generated due to boiling) can drag the liquid around and induce a higher liquid velocity that might cause the flow to become turbulent. In order to identify the dominant convection mechanism between the natural convection and convection induced by turbulence, Rayleigh number, as presented in the 'VALIDATION OF NATURAL CONVECTION - Correlation', could be calculated to compare with critical Rayleigh number ($Ra_c$~10⁹).

However, Rayleigh number could be inaccurate as the average $\Delta T$, by doing an integration around the tube, is not reliable due to nonphysical wall temperature at the stagnation points. Therefore, the idea is to increase the number of cells on the heating wall to have a more accurate average wall temperature.

Besides that, a more refined mesh around the tube can be considered for a better capture of boiling phenomena and gas volume fraction. Nonetheless, it is important for the wall adjacent cells to remain in log region of the wall boundary layer (30 < y⁺ $\leq$ 300) as the wall temperatures are estimated as a function of the liquid temperature of the wall adjacent cells (For more information, please refer to the article by Pérez Mañes J. et al [12]  pege 4 - equation (24)').

Additionally, if the flow is considered laminar, calculating y⁺ can be a very challenging task. It is found that if the wall adjacent cell is below the log region, NEPTUNE CFD tends to provide an inaccurate value for the heat exchange coefficient. This can be verified easily by access to the 'eb tube' value, (which corresponds to heat released by the tube to the surrounding [W]), in the 'listing' file. This 'eb value' has to be equal to the imposed heat flux [W] around the tube. In case where the 'eb value' is different from the imposed heat flux or keeps on decreasing, the mesh is believed to be below the log region.

As a consequence of all the propositions above, the closing laws for interface forces have to be questioned due to the magnitude of gas volume fraction produced by the boiling. Other than that, the hydrodynamics model for gas phase may need to be considered as turbulent according to the comparison of Rayleigh number and the critical Rayleigh number.