code_saturne User's Forum

Hello,

I'm solving a problem that involves a phase change (water-vapor) in a heat exchanger. My question is if the Merkle (mass transfer) model implemented in the cavitation module can be adapted/modified to produce void fraction as a function of the saturation temperature.

The idea is to implement a wall boiling model like the one proposed in this article https://www.researchgate.net/publicatio ... ling_Model.

Regards,
Luciano

Hello Luciano,

I think it is technically possible (quite easily) by modifying the function cavitation_compute_source_term (src/base/cavitation.f90) and its call in navstv in order to introduce the temperature dependence and replace the "presat" constant with a temperature polynomial directly in the cavitation_compute_source_term function.

However, regarding your heat exchanger problem, it is not a really straightforward approach and it could be a bit hazardous…

The paper you mention uses a "semi-mechanical" vaporization term based on a difference at a saturation temperature, analogous to Merkle's term for cavitation, based on a pressure difference. So this approach appears, in my opinion, better suited for a heat exchanger problem but the boiling vaporization term is coupled with an equation on enthalpy. This equation is not taken into account in the isothermal cavitation model of Code_Saturne.

Best regards,
Benoît

Hello Benoît for your answer,

I modify the cavitation_compute_source_term as a function of the saturation temperature and this part is working.

As you mention in your answer, now I have to add to the energy equation the vaporization/condensation energy due to the phase change (latent heat).

Also, the heat exchange coefficient is different (greater) during the boiling process than during pure convection.

I will keep you updated.

Thanks for your time.

Regards,
Luciano

Hello Yvan and Benoît,

I have changed the approach and follow this article https://www.researchgate.net/publicatio ... ling_Model to avoid the "semi-mechanical" correlations.

I have added the vaporization/condensation terms in the energy eq. for the cells on the heated wall and also in the bulk for subcooled flows. Also, I have to modify (scale) the convective heat transfer coefficient to take into account the quenching heat flux.

According to the RPI model, the heat flux is partitioned into three terms, and these terms depend on the wall temperature, so I have to solve a simple non-linear problem for each heated boundary face when the heat flux is imposed.

The only problem is that I can not add the drag/lift and lubrication forces in the momentum eq. because depends on the relative velocity between the liquid and gas phases.

Also, the results are sensitive to the turbulence model, but using the Rij-* with 2 scale wall function the results are very good in comparison with the experimental one.

Regards,
Luciano