Learning Forum

[Sifiso Mngonyama]

Hi Everyone 

When you have initialized a transient flow by first running it as a steady state and then switching to transient is it possible for the simulation to continuously converge at only just 2 iterations per time step or does this mean something is wrong? 

 

[Essence]

Hello,

It depends on the problem you are solving. It would be better to reduce the time step size and increase the number of iterations adequately. Once, the solution is getting converged, you can then increase and reduce the time step size and number of iterations per time step accordingly.

[Sifiso Mngonyama]

Thank you for your reply let me explain the problem. I'm simulating the effect of ventilation (air) in a slightly complex mine tunnel that has toxic gas (thoron) coming from the walls. I wrote some macros to introduce the toxic gas in the domain from the cells next to the walls. I'm only interested in the concentration distribution of the gas when there are different ventilation speeds and nothing else. When I switched from steady state to transient I didn't change the default max iterations/time steps which is 20 cause I thought it would be enough. My concern is that the simulation is converging continuously at just 2 iterations/ time steps out of the default 20 max iterations/time steps which I think is too soon. Furthermore when I play around with the time step size by decreasing/increasing it still converges at 2 iterations/time step hence why I think maybe I'm doing something wrong. 

 

[Essence]

What is the convergence criteria you have set? Did you try changing them?

[Sifiso Mngonyama]

I somehow forgot to set the residuals accordingly?. Thank you very much?

[Sifiso Mngonyama]

Hi again, I'm trying to do a mesh-independence study of the simulation setup I mentioned above. Will it be possible to do it with steady-state results even though it's transient since this would be faster? If yes, would using the average concentration of the toxic gas in the domain and outlet be correct, or would I need to use different variables? 

[Essence]

I would not recommend steady state if the goal is to study how the toxic gas propagates in the domain.

Please post a new Forum question for the queries out of the scope of this Forum thread.

Thanks.