CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Timestep (

Joe August 11, 2006 07:47


When running a simulation and fails at a specific timestep can this be interpeted at the length of the domain which has got stuck?

I mean if i have a domain of 10m and velocity 1m/s, physical time 10s. Inputting a physical timescale of 0.5s and the simulation fails at timestep 5, does it mean that it was at 2.5m when it was running?

I am running a steady state simulation and usually fails at 8,9,10 timestep and when i reduce the timescale it can go maximum to 17.

Thanks in Advance.

Mike August 11, 2006 08:04

Re: Timestep
Hi Joe, failing at a particular timestep cannot be interpreted in this way. Your best bet is to look at the results before it fails and see what the flow is doing. The solver diagnostics in the out file should also provide some clues - e.g. which equations were struggling, was the solver stable before it failed. Mike

alex August 11, 2006 08:26

Re: Timestep
HI joe, i had the same problem and, as said mike look at which equations were struggling. My issue was solved by changing turbulence model. ALex

Ravi August 12, 2006 05:28

Re: Timestep
I want information for Time step

Houman August 23, 2006 06:51

Re: Timestep
Hi, as Mike has written, you can not judge the results for the error runs like that... as far as it's about Time step settings, you better take a look at the CFX manual :

ANSYS CFX-Solver, Release 10.0: Modelling | Advice on Flow Modelling | Timestep Selection |

there you'll find some good general information about Timesteps, but your problem can also be because of something else, so don't ristrict your view to "timesteps" but it can be because of one of the boundary conditions or ...

best regards, Houman

All times are GMT -4. The time now is 18:49.