SimScale CAE Forum

Compressible Flow Simulation - Retro Propulsion

Hey!

I am trying to do a Retro propulsion CFD analysis.

  1. For case one, which is free stream without any opposing flow. That has been computed successfully. But facing issues with counterflow conditions. I tried to use “Outlet Inlet” Boundary condition and getting errors for it. Can anyone please help me out with this.

Free Stream case can be found in Simulations>Free_stream>Run 1.
And Opposing flow in Retro_Mach_0.5

  1. As density based solvers are to be used for the problem. While in workbench density based solver is available for the laminar model. Will this be able to capture the turbulent flow properly?
    Or should I use pressure based solver with other Turbulence Models?

My Project Link
https://www.simscale.com/projects/ojagtap/supersonic_retro-propulsion_1/

For more details, please follow the research paper at the provided link
https://commons.erau.edu/cgi/viewcontent.cgi?article=1262&context=edt

Thanks in advance!

Om

Hi @ojagtap!

That’s a good project you have chosen @Get_Barried and I can work on (still have to deliver some more compressible validation cases anyway :smiley: ).

For the density based solver it is only available with a laminar model and afaik nothing has been changed so far thus the laminar assumption has to be used and tested how far that works for this case. A nice discussion can be found here: Trans Super Sonic Compressible Simulation. We will see what we can achieve here and let you know asap!

Cheers!

Jousef

Hi @jousefm!

Thanks for replying and sharing the link!

  1. As mentioned by Mr. Arafat in " Trans Super Sonic Compressible Simulation".

Trans Super Sonic Compressible Simulation by Ali Arafat
(now the case was tested locally with ‘rhoCentralFoam’ Transient solver ( density based ) assuming Laminar flow ( for the sake of simplicity )).

a) But when the flow would come out of the nozzle the behavior will be turbulent forming shock diamonds. Will the laminar model be able to properly capture the whole phenomenon.

b) As the properties of the fluids are to be changed due to the restriction of Courant number. How can I calculate those?

  1. Still facing issues with counterflow boundary conditions.

Regards

Om

Hi @ojagtap!

Give me some time to do see what we can do here. I remember supporting a user doing compressible simulation also having issues with the shock diamond - keeping you up-to-date and feel free to inbox me if you have other questions in the meantime.

Best,

Jousef

@ojagtap I’m also working on Retropropulsion CFD analysis and having some issues lately, I think we should collaborate on this project and work it out together!
If you’re in, feel free to inbox me.

2 Likes

Hi @ronnie007 and thanks for the help here!

@Get_Barried and I would love to help you guys out as we also discussed compressible simulations (shock diamonds & Hyperloop simulation) in the past and it would be more than great to deliver more projects involving compressible flows. Let us know how you want to proceed!

Best,

Jousef

2 Likes

Hi @ronnie007 Ya sure! we can do that.

1 Like

Hi everyone,

Yes compressible simulations have not been very easy to get working. More work in terms of projects and other people tacking all sorts of different problems would do good to benefit our collective understanding in this area!

More than happy to help out where I can.

Cheers.

Regards,
Barry

2 Likes

Hey @ojagtap

You need a Riemann solver for what you intend to do. A pressure-based solver on collocated grids needs modified Rhie-Chow interpolation to be shock capturing, because the traditional RC interpolation for smooth flow is simply too dissipative for shocks. The pressure-based solver in OpenFOAM does have a transonic formulation, but it is only able to capture weak shock ( not even gonna get the weak shock correct ), let alone strong shock in your case.

The density-based rhoCentralFoam uses an explicit euler time stepping, so you will have to limit CFL to <0.2. I think you can use the laminar model, just set wall BC to slip instead.

2 Likes

Hi @dylan,

Thats great insight. No wonder I’ve been having issues with getting strong shocks to show from a compressible diamond aerofoil project.

Cheers.

Regards,
Barry

If you don’t model the flow boundary layer, you can use rhoCentralFoam and set wall to slip. But if you want to include shock and boundary layer interaction, you should consider using a Runge-Kutta 4 stage time stepping, or the computation effort is 4~5 times.

2 Likes