Learning Forum

[veera]

influent, I am running unsteady case, initially, I am started with 1e-6 in the time step size

and after a few iterations, I started to change the 2e-6.

with the help of continuity, momentum (x-velocity,y-velocity,z velocity) and energy data in the console, how I can confirm my self to the problem is converged for 1e-6 and like that, I will do for other incremental times step size

[Kalyan Goparaju]

Hello, 

Convergence in unsteady simulations is only  relevant within a time-step. Since the problem is unsteady, you can run for as long as you want, however, ensure that residuals converge within a time-step. One possible approach to judging an equilibrium state for an unsteady simulation is to have some monitors for appropriate variables in the flow and ensure that these attain a statistically stationary state (unchanging mean value). 

Thanks,

Kalyan

[veera]

I am not getting sir, 

actually same problem I am started with 1e-5 ..after few iteration its giving nan problem, so I am started to initiate the problem with the 1e-6 time step size ..after some iteration, I can change to 2e-6 or 3e-6 like that, but my question is before changing the time step size 2e-6 or 3e-6  from 1e-6 ..at what condition i need to look through consoile

[Rob]

You need to converge each time step. If the step is too big the solver may fail, or may not converge so your answer will be wrong. If the time step is too small it'll be stable but take ages to reach the desired solution time.  Aim to converge a time step in 10-15 iterations AND make sure the time step is small enough to pick up any/all features of interest.