CFD Online Discussion Forums

CFD Online Discussion Forums (https://www.cfd-online.com/Forums/)
-   CFX (https://www.cfd-online.com/Forums/cfx/)
-   -   How to export raw residual (https://www.cfd-online.com/Forums/cfx/199886-how-export-raw-residual.html)

lostking18 March 18, 2018 15:47

How to export raw residual
 
Hi, may I ask if we can use CFX to output the raw residual? I say raw because CFX can only output the normalized residual. Since the normalization method is not public, I cannot obtain the real residual.

I have already set in the .pre with: 'output control' -> 'output equation rediduals' and obtained the residual results. However the residual I obtained is normalized by a ANSYS self designed parameters(a_p and delta_phi) as shown in the help documents. Could you tell me how can I get the non-normalized residual?

I also have tried the 'expert parameter' --> 'convergence control' --> 'imbalance normalization type' --> 'true'. But the residual I get is still non-dimensional.


If you are interested in the CFX help documents about normalization, here is the link: https://www.sharcnet.ca/Software/Ans.../i1312559.html

Thank you very much for your help!

Lance March 19, 2018 04:14

I think you answered your own question.
Quote:

Originally Posted by lostking18 (Post 685620)
Since the normalization method is not public, I cannot obtain the real residual.

Why do you want the raw values?

lostking18 March 19, 2018 06:40

Quote:

Originally Posted by Lance (Post 685679)
I think you answered your own question.

Why do you want the raw values?

Hi, Lance, thanks for answering. My current research really need raw residual to analyze the imbalance of the flow. I am wondering if there is any method to obtain the residual/ imbalance of each cell.

ghorrocks March 19, 2018 16:36

The variables are probably in the MMS, alternately you can probably calculate it from MMS variables. So you will need fortran to access it, and you will need a detailed guide to the MMS to find it. There is no detailed guide to the MMS in the public domain that I am aware of, so your only option is to talk to ANSYS support about this. Be prepared to answer the "why?" question with them as well :)

lostking18 March 19, 2018 18:19

Quote:

Originally Posted by ghorrocks (Post 685783)
The variables are probably in the MMS, alternately you can probably calculate it from MMS variables. So you will need fortran to access it, and you will need a detailed guide to the MMS to find it. There is no detailed guide to the MMS in the public domain that I am aware of, so your only option is to talk to ANSYS support about this. Be prepared to answer the "why?" question with them as well :)

Hi, ghorrocks, Thank you very much for your suggestions! That's a good point. I do need to learn MMS in the future. However now with the supporter's help I have successfully solve this problem:

EXPERT PARAMETERS:
resid normalisation type = 0
END

will do the work!!! I'm really grateful to the supporter named Holger who gave me a fast and efficient reply. Hope this will help other people visit this page.


All times are GMT -4. The time now is 15:21.