CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Number of Timesteps and Physical Timescale (http://www.cfd-online.com/Forums/cfx/81085-number-timesteps-physical-timescale.html)

Leuchte October 15, 2010 08:34

Number of Timesteps and Physical Timescale
 
Hey,

sorry but i have to ask some very fundamental questions. The first question concerns the physical timescale. What advisements have to be done before i adjust this value?
The second question is: Were in Ansys can i adjust the number of timesteps?
Under Solver Control in Ansys-Pre, there is only a possibilty to adjust the number of maximum and minimum Iterations.
Im using Ansys 12 in my Project.

regards
Nick

Josh October 15, 2010 08:54

For your first question:
http://www.cfd-online.com/Wiki/Ansys...gence_criteria

In Solver Control, you can choose to specify iterations, time, or timesteps.

michael_owen October 15, 2010 11:27

Quote:

Originally Posted by Leuchte (Post 279345)
Hey,

sorry but i have to ask some very fundamental questions. The first question concerns the physical timescale. What advisements have to be done before i adjust this value?
The second question is: Were in Ansys can i adjust the number of timesteps?
Under Solver Control in Ansys-Pre, there is only a possibilty to adjust the number of maximum and minimum Iterations.
Im using Ansys 12 in my Project.

regards
Nick

It sounds like you are doing a steady state simulation. There are no "timesteps", there are only interations. The name "timescale" is somewhate confusing; CFX uses a false transient algorithm in steady state to under-relax the solution as it iterates towards convergence. That's why it's a time "scale" and not a time "step".

For your first question, a good choice is usually a fraction (a third for example) of a representative timescale in the problem, 1/omega for rotating machines, the residence time L/V for through flow problems, etc.

michael_owen October 15, 2010 11:28

Also, Josh, I think you are confusing the Analysis Type settings for transient with the Solver Control settings.

Josh October 15, 2010 11:36

Indeed I am. 18-hour workdays fry your brain.

michael_owen October 15, 2010 11:47

Been there.


All times are GMT -4. The time now is 21:17.