[nepomnyi]

Hello dear CFDOnline community.


I've been trying to find the equation behind the function MRF.correctBoundaryVelocity(U) given in the file UEqn.H on line 3.


I concluded that the dot operator, in this case, means that we are accessing the method correctBoundaryVelocity of the class MRF.



So - not with ease - I found the class MRF. And, indeed, it has correctBoundaryVelocity as a member function. See line 174 in the file MRFZoneList.H. The definition of correctBoundaryVelocity is given in the file MRFZoneList.C on lines 376-382.


There - again - correctBoundaryVelocity is given as a method of the class operator[]. So, I found that operator[] class in the file UPtrListl.H on lines 95-108.



operator[] happens to be templated from the class T. That's where I have a problem: I haven't been able to find class T. Moreover, in the header of the List.H file, I read that it is, actually, a class of type T which I have no idea what that means.


I read here that Foam::T has to do with transposing matrices, but that's the only thing I was able to find. Searching Foam namespace list didn't reveal T to me.


Thus, I'm wondering 1) what MRF.correctBoundaryVelocity does, 2) what formula it uses and 3) where I can find the underlying code for it.


Thank you in advance.
Ivan


P.S. Regarding suggestions to learn C++ first which I anticipate. They are right indeed. But I don't have time for that. I need to complete the task of figuring out the formulas behind the code that I intend to use. That's an urgent task. I'm learning C++ simultaneously. But I'm at the very early stages of learning C++. People in academia should understand me. When you're in Ph.D. you do a lot of tasks simultaneously and the idea is to get the task done by any means on time with minimum knowledge. That's why I'm seeking help from the community while I - myself - don't have knowledge, yet, to figure out the solution.

[nepomnyi]

I suspect I have found a partial solution.


MRF seems to be a folder containing several files with code (see here the MFR tree on the left hand side of the page).



Evidently, correctBoundaryVelocity method may occur in several files within the folder.


Like I mentioned in my original post, I had found correctBoundaryVelocity method in the file MRFZoneList.H. Following that file, led me to a dead end. I'm not sure what those lists are.


I searched the rest of the files within the MRF folder and I found correctBoundaryVelocity declaration in the file MRFZone.H on line 255 and definition in the file MRFZone.C on lines 541-568.


Even though I still have to decipher what's written on lines 541-568 in the file MRFZone.C, preliminary, it looks like what I need: the formula for correctBoundaryVelocity.


If it is so and once I've deciphered it, I'll post the formula in my blog.


Ivan


P.S. For those interested, <class T> is a special thing in C++ that allows you to give any type to the variable. I.e. T is a C++ keyword meaning any type (type can be int, double and so on). See here (scroll down to "Class templates" section).

[crpvn]

Hello Ivan,

I think that correctBoundaryVelocity method changes velocity for the patches that are boundary of the MRF zone.

It works correctly when fixedValue 0 is applied as BC for U, but it doesn't when rotatingWallVelocity is applied, as explained in this issue:
https://develop.openfoam.com/Develop.../-/issues/1810

I hope it helps.

[crpvn]

Also this document could be useful:

https://www.google.com/url?q=http://...fvp1OfoXWH8H-K

[nepomnyi]

Originally Posted by crpvn

Also this document could be useful:

https://www.google.com/url?q=http://...fvp1OfoXWH8H-K

Hello @crpvn, thank you very much.


The link wouldn't open. The Internet search didn't give me the document as well. I was wondering if you could attach the document please.



Also I was wondering if you could clarify your previous answer please.


First is MRF. People seem to use it as something obvious. However, I have no idea, whatsoever, what that means. Digging through the Internet, I found out that MRF stands for multi reference frame (which, again, doesn't tell me anything). Apart from that, I've got a feeling that people refer to non-stationary flow boundaries when using the term MRF. E.g., flapping airfoil is MRF, or rotating cylinders are MRF, but just a regular stationary pipe is not MRF. Am I right?


Second, you said "I think that correctBoundaryVelocity method changes velocity for the patches that are boundary of the MRF zone." What did you imply saying "changes"? What change did you mean? Did you mean that the user assigns boundary condition (velocity in this case) in the absolute frame of reference, and the function converts it to the relative frame of reference? Or something like that. I don't need the exact answer, I just need to get a very general feeling of what's going because I don't have any non-stationary boundaries in my problem.


Thank you in advance.



Ivan

[crpvn]

Hello,

The document is "Methods for wheel rotation modelling" by Chalmers University, where MRF approach is explained in details.
http://www.tfd.chalmers.se/~hani/kur..._Josefsson.pdf



At this link you can find the mathematical details for MRF development.

https://openfoamwiki.net/index.php/S...RF_development


If you apply fixedValue 0 as BC for a wall inside an MRF zone, wall velocity will be changed as omega X R, instead 0 value.