CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   Main CFD Forum (http://www.cfd-online.com/Forums/main/)
-   -   Question about convergence (http://www.cfd-online.com/Forums/main/109449-question-about-convergence.html)

Anna Tian November 18, 2012 05:07

Question about convergence
 
Hi,

Sometimes when I use large time step to calculate, the simulation doesn't converge. So I change to use a smaller time step to run it for several iterations then change back to large time step. In this way, my simulation converge easily.

May I ask theoretically why those several iterations with small time step bring the simulation to convergence?

Thank you very much.

vicarious November 18, 2012 06:09

Hi,
Large time steps cause in an appropriate Courant condition. For exmaple in explicit time stepping, large time step will increase the CFL number and the solution diverges specially at the beginning of the procedure due to the solver guess initialization.

Ahmed November 18, 2012 09:22

Quote:

Originally Posted by Anna Tian (Post 392751)
Hi,

Sometimes when I use large time step to calculate, the simulation doesn't converge. So I change to use a smaller time step to run it for several iterations then change back to large time step. In this way, my simulation converge easily.

May I ask theoretically why those several iterations with small time step bring the simulation to convergence?

Thank you very much.

When using small time step, the solution is not likely to diverge and hence it advances and the initial values are modified towards the path of convergence.


All times are GMT -4. The time now is 04:39.