SimScale CAE Forum

Simulation Results view changes at the last moment

Hi all,
Why is it, that an image in the solution fields develops one way, and then at the last moment, produces a different result? This is consistent with 6 simulations (each with 6 sub sims) on one project and the same on another.
I don’t know which imagery to trust. The developing images look like the results I am expecting.

Please see images below

During development (and expected output)…
.

Then the final image…

during development…

Final image…

Project A
https://www.simscale.com/workbench/?pid=3020144088369128978&rru=cd548561-f2f4-4db1-9722-a221f4473640&ci=be8fd4c6-1e8d-461d-8c04-a91f1904dcfe&ct=SOLUTION_FIELD&mt=SIMULATION_RESULT

Project B
https://www.simscale.com/workbench/?pid=6846801220607138490&rru=5e9b825e-2a1d-448b-a963-a2f3ee86aa19&ci=c6c44a93-8622-4f5f-a50e-adb2442fceb5&ct=SOLUTION_FIELD&mt=SIMULATION_RESULT

Hi @jcdesigns!

You are looking at the steady-state solution. So assuming that your result have properly converged, the last step is the one that is relevant to you, the one that actually is the “steady state” of your simulation. However note that if the simulation is transient in nature you are making assumptions that might not fit in this situation.

Regarding convergence, not that this is not the only criterion to see if a simulation is fine. Usually to convergence plot has to drop by several orders of magnitude and forces have to converge as well. If there are for example oscillation in the force plots, that’s a quite good indicator that the phenomenon you’re looking at is unsteady. Good approach in any case! Because the simulation type you chose is way cheaper and if it is really steady-state, you have done a good first guess and saved the maybe unnecessary transient simulation. If it is unsteady you can still run a transient one and have got a first “cheap” impression from the steady-state one, hope that makes sense.

All in all (and to finally answer your question) - the first (zeroth state) is not of any interest but the 1000th step/iteration is assuming that you really want to have a look at your steady-state result (see above).

In case I missed something or some points were too confusing, I am tagging the @CFD-SQUAD as well.

Best,

Jousef

1 Like

Thanks Jousef,
let me reiterate.
I am not talking about from the first state to the last state.
these sims are all finished. when you start the solutions fields, there is already a particle trace ready to go (click on links to the projects).
As the particle trace is loading, it presents one image (the expected image), and then at 99% loaded, it then changes to a different image.
I trust the first image more.

Are you talking about the time it takes the traces to build up/render?

Cheers,

Jousef

yes. (I guess so).
The way they look whilst the traces are loading, look correct (air flow around structure), and then at the last moment the new refreshed image has the air flows looking different.