I am new to CFD and I wonder what I have done wrong.
I have made the car in the above link and I have run simulations, but it seems to me that the results are from the box around the car. I have followed the steps from the video below and I have used the site from below for assistance. But I wonder what I have done wrong.
Here’s a tutorial: https://www.simscale.com/docs/validation-cases/aerodynamics-flow-around-the-ahmed-body/ - @Ricardopg is currently answering so he probably already found the issue in your setup.
You have an issue with the boundary conditions. Please doublecheck the velocity directions for both velocity inlet and moving wall BCs:
Actually there was another thread very similar to this a while ago, if you want to take a look. Here’s the link Beginner CFD Analysis Mistake
Thank you for your answer , I have changed the Uy to -63.7 instead of the Uz, should this solve my problem?
Thanks for your answer, but it does not clarify it for me.
It would be +63.7 in the y direction. -63.7 points out of the domain
It will solve this one problem you were having. There could be other things wrong with the setup aswell, which I haven’t spotted while I was looking at your sim.
I can say, however, that it’s a good idea to use a coarser mesh at first to do test runs. Your mesh has 15M cells and if you make an error, it will be a costly one. First try to run the simulation on a coarser mesh and, as you get confident with your sim configuration, move to the bigger meshes. Keep us posted if you run into trouble
PS: you can also browse through the public projects section to see how other people have set up their cyber truck simulations
Thank you for your reply!
Is their a way to shorten my run-time without having to simulate only halve of my car?
Well, there definitely are ways to shorten the simulation time. I’d like to point out that, in theory, you should simulate until your convergence criteria has been met.
There’s a lot of testing and tweaking that you can do in Numerics to speed up the process (relative tolerances, under relaxation factors, schemes, solver selection etc.). Other things that also can help with convergence:
- having a good mesh for your problem;
- making good guesses of initial conditions;
- using “potential foam initialization” in simulation controls might also do that.
This link was really helpful, I used this link from the forum to alter my own settings https://www.simscale.com/projects/Ricardopg/cyber_truck/