Learning Forum

[sailaja]

Hi,

I am running transient simulation and implicit formulation of vof model. I have selected adaptive time advancement method. I set the global courant number as 0.5,initail time step size 0.00001 s , number of time step 10,000 , max iteration 60. when i run , it says , 230 solution has been converged in fourth time step . I have been running my simulation since yesterday and its progressing very very slowly and its only 3 % in 8 hrs. My questions are:

1. once its says after some iteration like it says 230 solutions has been converged, what is the meaning of physical time step in range of 10^-4 in between the iteration and also what is the meaning of global courant number 0.1 while running simulation in console ?
2. Based on what and after which time we will know that we can increase the initial time step size which in my case is set to 0.00001. I want to increase the time step size because it will take too long time for complete simulation and I have very limited time.
3. Once the solution has reached convergence ,can we also decrease number of iteration ? (Suggestion with explanation on this please)
4. Is it based on the courant number and the physical time step , we can increase the initial time step?
5. Also , what is the reason behind not getting very low value of residual for continuity ?

Thank You in advance.

[Karthik Remella]

Hello Just a couple of thoughts on your issue:
You could increase the Global Courant Number to 1. This may speed up your simulation as the solver may take larger time steps.
Regarding your initial time step, this depends on the problem you are attempting to solve as it is intimately dependent on the physics you intend to capture. Sometimes, numerics also play a role in the stability of your solution.
Max iterations / time step - the value depends on your convergence criteria (your residuals you have set for your problem). If you have modified the residual criteria to obtain a deeper convergence, you will need a larger number of iterations per time step. You can relax this number to 1e-3 (the default values). Your goal should be to choose appropriate values for time step size as well as max iterations per time step so that you get faster convergence. You need to ensure that you are reaching the set residual criteria every single time step - that's all.
Karthik