'Analysis Velox V2' simulation project by sliu


#1

I created a new simulation project called 'Analysis Velox V2':

Simulation für F1 in schools


More of my public projects can be found here.


#2

@sliu,

why did you cancel the mesh operation? I re-ran it and it worked smoothly - I would set it finer, but the mesh in general worked. Is there a specific reason why you canceled the operation?

All the best,

David


#3

Thank you so much, there was the message that the geometry is broke.


#4

How where you able to finish it?


#5

@sliu,

ah - you mean this message?

This is not a fatal error but a warning, so the the mesher is perfectly capable of finishing the meshing. Just re-start it and you’ll see a the mesh I had. Maybe make it one level finer, to not have those huge cells that I see :smile:

By the way, did you see the F1 workshop that starts tomorrow: https://www.simscale.com/simscale-community/f1-workshop/

Happy simulating!

David


#6

Now it says that my machine ran out of memory.


#7

Hey @sliu,

I see that in the meantime, your mesh finished:

Looks good enough to kick off a first simulation with it! Later you might want to refine those edges a bit and also refine the wake region of the model but generally looks reasonable.

I guess increasing the fineness level of the mesh led to the machine having insufficient memory - but you obviously already figured it out by bumping up the number of cores to be used. There is a little video that gives a little more insight into number of cores:

https://simscale-2.wistia.com/medias/rt0shguy1t

We’re working on making the core-choice automatic but as it’s hard to know ahead of time how many cells the mesher will produce and thereby knowing the RAM requirements, it’s not straightforward.

Hope that helps,

David


#8

@sliu - great! Congratulations on the simulation setup - looks good!

A few hints that might help to speed up your work: You could set the Initialize with potential flow setting under Simulation Control to true. That way your flow domain will be first initialized with a potential solution and then the actual RANS simulation will be started. Will give you a faster turn-around time on your simulations.

The residual convergence could be better, right - I typically aim for 3 orders of magnitude in pressure, 4 in all other quantities. Currently you’re at one order of magnitude. Possible angles of attack could be:

  • Numerics: Reduce the stop-criteria of the linear system solvers and switch to “GAMG”
  • Mesh: Refine the wake behind the car to make sure the turbulent region behind it is refined well

Hope that helps!

David


#9

Can you tell me where I can see the Cw-Wert?


#10

Hi @sliu,

so the easiest way to compute the drag coefficient (German C_w Wert) is to add Result Control items to your simulation before you start it. Those control items basically allow you to compute specific results you are interested already during runtime, where you most of the time don’t need to do any post-processing but just “read” the value after the simulation is completed. You’ll find the force coefficient result control item under Forces and moments of the Result control tree item:

That will give you the components drag, lift and pitch. You can also add only forces & moments which will give you the actual force values.

In case you already ran the simulation and have not assigned such a result control item, you can also compute the drag force retrospectively with Paraview. This blog post explains how to do so:

https://blog.simscale.com/blog/2014/07/aerodynamics-post-processing/

It’s a bit more elaborate but will do the job.

How is your experience so far? Do you get the results you need?

Looking forward to see your other projects completed as well!

All the best,

David