Dear Sir/Madam,
I am working on simulating a custom designed propeller to fly in mars atmosphere conditions to see thrust plot and CFD data. It is for my master thesis. I have taken the reference of simulating a simple propeller tutorial but the difference here is I am doing it with compressible flow with k epsilon model and transient time dependancy simulation for mars conditions. The tutorial was helpful but it was for incompressible flow.
The issue I am facing is that the simulation starts as usual, it shows the job was prepared successfully but the when simulation run starts it never progresses above 0%
I have spend last couple of days waiting and now I think it’s time to ask for expert help.
I am having really hard time with this issue. I tried modifying the boundary conditions etc but issue is still the same.
Can you look at the project and give me a suggestion on tackling this?
I would be really grateful!
@jousefm @cfd_squad @dschroeder
Best regards,
Akash
I believe you gave up too early at only 65 core hours, I would expect over 10 times that for a 21 million cell mesh using 32 cores.
That 0% is not a linear thing, it takes a lot of percentage of the run time before it starts to move towards 100%…
What is your reasoning to not generate boundary layers in your mesh?
Hello Sir,
Thank you for replying. I have 21 M cells in the mesh and as per your suggestion it is taking longer may be 10 times the time. I only have a propeller in the CAD geometry and do you know why I have big number like 21 M cells ? It is not really complex model still it took a long time to mesh.
I am new to CFD, I do not know how to generate boundary layers in mesh. Can you help?
Can you tell me how can I reduce the number cells in the final mesh? I know it calculates automatically but if there is some parameter I could change to have a simple mesh?
Is there any other way I could simplify the simulation?
Edit: the new warning is Courant number exceed 1, decrease timestep. I already tried, Can you help do something else?
Thanks again!
Sorry, I am not familiar with mesh requirements for compressible flow or transient time simulations, lets see what some other CFD power users can add… calling @cfd_squad
1 Like
Hi @patelakash9429,
Dale is spot on with the mesh issue. Your mesh is much much too fine for transient based simulation. You have alot of optimization to do for the mesh. I suggest simplifying your mesh to several refinements with the main refinement refining the blade profile mesh, the layers and the immediate areas surrounding the propeller.
Start from a very coarse mesh and get the results first before looking at where to refine. If need be, you can just add a surface refinement to the propeller itself and have one level of refinement for the MRF and start with that first.
For a ballpark rough mesh count, maybe you can try to hit about 500k cells first.
1 Like
Hey @Get_Barried
I will try to generate the mesh just for propeller surface refinement and MRF surface refinement at very coarse level meshing. Let’s see how the mesh comes out to be.
In case I get less number of cells, how can I generate simulation for my purpose after that?
For now I have deleted Ext RR, Wake RR etc. But for the final simulation I will need them to be there to generate simulation. How could I achieve that?
Thanks!
Hi @patelakash9429,
You want to refine incrementally. There is no rule stating you need your additional meshing zones as long as you can reach some suitable results.
However, considering you don’t even have a first run, having a well refined mesh is the last of your problems.
Start with a basic mesh, get some results, deduce how far off you are from the experimental or reference results, then proceed to to refinements and adjustments as needed.
1 Like
