convegence accelerated ?
HI i want to know if we can accelerate a convergence ( reduce the iteration number ) without affecting the physical meaning a the results,in this case how?
|
Re: convegence accelerated ?
Hi,
This is discussed in the manual. For a steady state simulation you use the timestep size as the acceleration parameter. Glenn Horrocks |
Re: convegence accelerated ?
hi thanks alots Glenn,i have to do it because you've answered to many of my questions. actual i know that parameter,i think that a larger timestep would accelerate the convergence,but would it affect the result ( i mean the physical one),CFX compute the the auto,to which extend can i deviate from.
thanks again. |
Re: convegence accelerated ?
A converged solution does NOT depend on the timestep. Choose a value that gives you optimal convergence.
|
Re: convegence accelerated ?
i haven't understood?,thanks
|
Re: convegence accelerated ?
Hi,
A converged steady state solution will not change no matter how large or small the timestep. The psuedo-time terms go to zero. As Hekler says, you adjust the timestep to get optimal convergence for your simulation. If you are finding the timestep size is changing the solution then you have a problem in your simulation. Glenn Horrocks |
All times are GMT -4. The time now is 10:41. |