SimScale CAE Forum

'cavity check' simulation project by pakanatiakash


I created a new simulation project called 'cavity check':

Checking my own solver against SimScale solver

More of my public projects can be found here.


@pakanatiakash - seems like you’re looking to run a 2D flow simulation on SimScale. While the simulation setup of a 2D simulation is possible with SimScale, the meshing of a 2D structure is not yet (working on it as I write). If you want to run a 2D CFD sim, you’d need to upload the mesh yourself. I’d recommend checking out this project:


Hi @dheiny. Thanks for the prompt response. Actually with reference from previous posts (Modelling of Solidification Process), I have constructed a preliminary solidification solver for binary alloys. My results are fairly good but I am slightly suspicious of convection during temperature gradient. So I am looking for a simple test case with 2D cavity. No solidification and imposed temperature on the left wall is slightly lower than the initial temperature of the liquid. All other walls are adiabatic. This should drive some convection. This simple case can verify my velocity field, especially magnitudes. But it might sound silly but I have no access to CAD tool. Is is possible to use the mesh I generated in Openfoam using blockMesh? If so I think my life would be little easier :smile:



sure, you can simply generate the OpenFOAM mesh locally and then upload it. For uploading, just pack the contant folder, find more detail here:




@dheiny, thanks a lot. I guess I am one step closer. I have a small problem though. I am not able to give density in the material properties. Rather I can give only amount of substance and molar weight. Since I am using a Al-Mg alloy, I only have reference density value and nothing related to molar weight. I looked in the forums but could not really find something on giving density. How do i proceed?

Also, how is the bousinessq approximation treated? I am using the equation rho = rho_o*(1 - betaT*(T-Tref)) to account for thermal convection in my solver. rho_o is my reference density, betaT is my coefficient of thermal expansion and Tref is my reference temperature. I think to compare simscale with my solver, it is crucial for me to have the same values.



@dheiny figured out the density issue! I am still in dark regarding reference temperature.


@pakanatiakash - so is your question where to set the reference temperature or how the density is being computed within the solver? My colleagues @gholami and @jprobst would be the right folks to ask here. Babak, Hannes do you have a hint for Akash?


Yes, my question is how the density variation due to temperature difference( or more precisely how the bousinessq approximation) is treated in the solver.


@pakanatiakash, nice work with the setup.

To use the Boussinesq approximation in your simulation, you should use buoyantBoussinesqPimpleFoam or buoyantBoussinesqSimpleFoam solvers. These two solvers, for now, are only available through case upload (see case upload documentation). Therefore, you could

  1. Download your simulation from SimScale.
  2. Make modifications according one of OpenFOAM tutorials, e.g. the hotRoom tutorial.
  3. Specifically, you could define Thermal expansion coefficient and Reference temperature in constant/transportProperties.
  4. Use case upload to run it.

Give it a try and let us know how it went. No worries if it didn’t work on the first try, we could look further into it together.

We are looking into adding these solvers to our default set of supported ones.


@gholami Thank you for the information. I have come across Boussinesq solver is OpenFOAM tutorials. I will look into it. Thoug I have used a melting solver to compare my thermal convection case and the results look reasonable.

@dheiny Just to update you on solidification solver, I have a solver based on volume averaging method -

I am also referring to numerical benchmark case ( ) to compare my results. As of now I am using the same mathematical framework mentioned in this paper. I have to resolve issues with groovyBC as i am using 3.0.0 Openfoam version and i am not able to get the libraries for it. This can help me in having variable heat flux BC to compare with the case in the benchmark case. Also, I would be happy to test my code against any solver which might be developed at SimScale in near future!