CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Varying time-step and Results freequency (

jstaylor89 April 16, 2012 11:29

Varying time-step and Results freequency
I have a CFX simulation that is basically a pipe with inlet and outlet orifice's. the simulation runs over 4.5ms, with a time step of 0.1ms. I now need to run the same simulation but with inlet events before this to try and replicate initial conditions. This brings the whole simulation to 1050ms (1.05s), but I only need the results from the last 4.5ms as before.

I have it all set up but running it now would give me 10500 time steps and take 600+ hours and 240GB results.

The accuracy of the first 700ms is not so important so ideally I want a time-step of 10ms for the first 700ms, then change to a 0.1ms time-step after this. Is this possible and if so, how?

I would also like to only record the results near the end. I have tried setting trn results to, Time Interval: t>0.99[s], but this gives me a resolves to dimensionless error. Any ideas on this?


ghorrocks April 17, 2012 02:00

Why do you use 0.1ms time steps? Have you confirmed these are adequate?

I would run the whole thing using adaptive timesteps, adapting to 3-5 coeff loops per iteration. This means you will be running a sufficiently small time step at all times, and the time step will automatically get bigger when it can and smaller when it needs to.

All times are GMT -4. The time now is 05:33.