[visakhmg]

Hello everyone,
I am trying to solve 2D Rayleigh-Benard convection with OpenFOAM 7. I am using buoyantPimpleFoam with boussinesq model for equation of state. Does this make the solver incompressible? Does the solution differ in any way with that of the (deprecated) buoyantBoussinesqPimpleFoam? Also, does the value of molWeight play any role in the solution? Because I'm trying to get a Nu vs Ra, Pr correlation, my only input parameters are Ra and Pr, or indirectly, mu, Cp, rho0, beta, Pr and deltaT, L from the boundary conditions. How does sensibleInternalEnergy instead of sensibleEnthalpy make a difference in this context?
And finally, is there a better way to solve non-dimensional version of equations (than solving th usual dimensional equations by adjusting the fluid properties to get the desired Ra, Pr)?

Attaching my thermophysicalProperties dictionary:

thermoType
{
    type            heRhoThermo;
    mixture         pureMixture;
    transport       const;
    thermo          hConst;
    equationOfState Boussinesq;
    specie          specie;
    energy          sensibleEnthalpy;
}

mixture
{
    specie
    {
        molWeight   18;
    }
    equationOfState
    {
        rho0        996.5;
        T0          300;
        beta        3.03e-04;
    }
    thermodynamics
    {
        Cp          4195;
        Hf          0;
    }
    transport
    {
        mu          0.8e-2;
        Pr          6.0;
    }
}

[Krapf]

Some answers:

Originally Posted by visakhmg

I am using buoyantPimpleFoam with boussinesq model for equation of state. Does this make the solver incompressible?

Yes, since the density then no longer depends on the pressure, the flow is incompressible.

Originally Posted by visakhmg

Also, does the value of molWeight play any role in the solution?

It is used to calculate the specific gas constant.

[visakhmg]

Originally Posted by Krapf

It is used to calculate the specific gas constant.

Thanks. But I'm using Boussinesq appoximation for the equation of state. Why calculate the gas constant? What if the fluid is liquid?

[Bloerb]

nMoles and molWeight is not used for pureMixture. It is however used for the other options and therefore a parameter that needs to be present although it is not used in your case. Yeah c++ inheritance thingy. They got rid of some parts of it over the years though.

There was a debate about buoyantPimpleFoam and buoyantBoussinesqPimpleFoam and their differences on here a while ago. Since the Boussinesq approximation was incorporated into the equation of state it is not really incompressible if i remember correctly. However i am not certain if this is how it was implemented. You might find the thread though. What is true though is that the density is not a function of pressure.

[Krapf]

Originally Posted by Bloerb

nMoles and molWeight is not used for pureMixture. It is however used for the other options and therefore a parameter that needs to be present although it is not used in your case. Yeah c++ inheritance thingy. They got rid of some parts of it over the years though.

Ah, I didn't know that. Sorry for the misinformation!

Originally Posted by Bloerb

There was a debate about buoyantPimpleFoam and buoyantBoussinesqPimpleFoam and their differences on here a while ago. Since the Boussinesq approximation was incorporated into the equation of state it is not really incompressible if i remember correctly. However i am not certain if this is how it was implemented. You might find the thread though. What is true though is that the density is not a function of pressure.

There seem to be different definitions of "incompressibility". Many of the equations of states in OpenFOAM, which depend only on temperature, have an "ico" (i.e. incompressible) in their name.
The German Wikipedia article on incompressible fluid contains the following statement:

Ein Fluid, dessen Dichte nicht vom Druck abhängt, wird inkompressibel genannt – im Gegensatz zu kompressiblen Fluiden.

Dies bedeutet umgekehrt, dass Fluide, deren Dichte sich beispielsweise durch thermische Einflüsse ändert, inkompressibel sein können.

Translation:

A fluid whose density does not depend on pressure is called incompressible - in contrast to compressible fluids.

Conversely, this means that fluids whose density changes, for example, due to thermal influences can be incompressible.

[Bloerb]

Yes that is correct, sorry for the confusion. for rho(p,T) or rho(p) you get compressible for rho(T) incompressible behaviour. In other words, without pressure dependence you can't compress your volume by pushing on it. You can however change it's volume by heating it.



The topic i meant was the following.

Wrong implementation of the Boussinesq assumption in heat transfer solvers


the main argument here is probably that the density update shouldn't happen in all terms. I am however not familiar with the approximation and implementation to know which is the case.

[agustinvo]

This is a nice thread as I see, I am wondering the same and I built my own Boussinesq solver based on older OF versions.

Let's assume that buoyantSimple/PimpleFoam with Boussinesq is incompressible. What are the turbulence models employed on these simulations? Boussinesq solvers use incompressible turbulence models, but I assume that with this modification you're obliged use compressible models.

Most of the incompressible models for Reynolds stresses can be "translated into" compressible models, but if I remember correctly, only the turbulent Prandtl number is used for the turbulent heat fluxes. If you are using custom models for heat fluxes (THFM) , then some coding is needed to pass your models from the incompressible to the compressible framework.

This is why I'd build my own Boussinesq solvers until I find out a good way to add my custom THFM models