CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   ANSYS (https://www.cfd-online.com/Forums/ansys/)
-   -   Improving the CFD Case Imbalances (https://www.cfd-online.com/Forums/ansys/242143-improving-cfd-case-imbalances.html)

arv2995 April 7, 2022 04:43

Improving the CFD Case Imbalances
 
I have created a CFD Steady State Case and have run the case for 3000 iterations.

After the case was fully run, I checked the .res file in CFX Solver Manager to check the imbalances of the variables. I want to achieve imbalances below 10, if possible below 5.

But for a few variables, the imbalances are as high as 80(mainly the mass imbalance).
My question was, what are the methods I can use to achieve better imbalances in the next simulation for the same case? ( I want to keep the entire mesh constant).

Looking forward to your reply.

Gert-Jan April 7, 2022 05:10

In the solver manager, you need to create a plot in which you monitor all imbalances. Then you can see that the imbalances are not correct, so you need more iterations.
You can also see the imbalances at the bottom of the output file. There you'll probably get confirmed that the imbalances are not within your requirements.
To be on the safe side, you should specify 10000 iterations in Pre and add a convergence criteria of 0.01 for the Conservation target = 1%

BTW, better ask these CFX quesitons in the CFD forum

Also I can recommend to read this: https://www.cfd-online.com/Wiki/Ansy...ated_questions

arv2995 April 7, 2022 05:50

Thank you for the reply.

I already checked the individual imbalances values. The problem is that the imbalances are reducing and increasing in a cyclic manner from -80 to + 80 for mass flow imbalance specifically speaking.

If I run the case again the same pattern is getting repeated unfortunately. Do you think any other change should be made , than running the case for more iterations. Like, will changing the physical timescale improve the imbalances?

Gert-Jan April 7, 2022 06:09

It can depend on anything. You need to give details on the case if you want more help.
Most likely, the solution is transient, so you need to do a transient calculation.

arv2995 April 7, 2022 06:29

The solution is a steady-state model.

Also, what other points does this depends on? Also what details would you like to know?

Gert-Jan April 13, 2022 17:27

You can tell the software to look for a steady state solution.
But if there is no steady state solution, it will never find it, and your numerical solution will fluctuate for ever.


All times are GMT -4. The time now is 07:29.