'roundabout flow' simulation project by dgeoghegan


#1

I created a new simulation project called 'roundabout flow':

What happens when you flow liquids through this geometry


More of my public projects can be found here.


#2

@dgeoghegan - interesting CAD model, looking forward to see the flow simulation of it. The automatic internal mesh operation might work, but the meshing and simulation workflow would be easier if you had a solid body geometry and not just free faces. You can see this in the Geometry Event Log after uploading:

In the case of IGES that happens quite often, which is why I’d recommend using STEP for uploading. What CAD software do you use?


#3

@dgeoghegan - mesh now looks good. I quickly parsed your sim setup. If you remove the 4th and 5th boundary condition and make the outlet a “pressure outlet” it’ll run!

The CFD BCs in SimScale cover all physical quantities at a given face. This means if you assign one BC to a face, all quantities (U, p, k, Omega) will be defined. Therefore you “overconstrained” the simulation with your current setup.


#4

@dgeoghegan - very odd. The solver complains about the decomposition approach - I rarely use the Simple method and always stick to Scotch. If you have no specific reason for using simple, give scotch a try


#5

Just want to say thanks for keeping an eye on this one, really appreciate the support


#6

@dgeoghegan - glad I can help! So now the sim seems to start, looks good. However under Simulation Control, you chose these settings:

Now the thing is that under Analysis Type you chose this simulation to be steady-state. This means the actual end-time is irrelevant as the solver is only looking for the “equilibrium” state via pseudo-timesteps. This means only the number of iterations is relevant, not the actual end time. With 10 seconds end time and a time step of 0.1 this will be 100 iterations. A good place to start is typically 1000, so I’d suggest:

  • Start time = 0
  • End time = 1000
  • Time step length = 1
  • Write interval = 1000

Since it’s steady state the intermediate time steps will only show you an intermediate solution without any good meaning (except for debugging purposes).

Looking forward to see that flow streamlines :smile:


#7

Getting there! This is your last simulation run a bit post-processed…