0

MINC problems in post processing

Hi, everyone.

I`m studying dual permeability problem and I`ve got a dual permeability modeling result.

Anyway I`ve ran into difficulties doing post processing after running.

Output file that result from the modeling have two types in each grid position, one is matrix and another is fracture. Firstly, I carried out both of them at once using Tecplot, but the graphcal results came out strangely.

I think the reason of this strangely result came from sharing a mesh between matrix and fracture.

How do I resolve the issue?  Do I have to do post processing after dividing fractures and matrix in the resulting file?

Thank you in advance,

HA

9replies Oldest first
  • Oldest first
  • Newest first
  • Active threads
  • Popular
  • Hi Jaechul,

    Since you are using EXT (I presume) in combination with Tecplot, here is what I do:

    Take the MINC file, and remove all the matrix elements (i.e., only keep elements that have a blank as the first character in the element name). Under Unix (or Cygwin), all you need to do is:

    grep "^ " MINC > MESH

    (You need to edit file MESH and add back in keywords 'ELEME' and 'CONNE'). Then use this MESH file to run EXT to get the fracture continuum response.

    Next, you do the opposite, i.e.,

    grep -v "^ " MINC > MESH

    (You need to edit file MESH and add the blank lines before 'CONNE" and at the end of the file). Then use this MESH file to run EXT to get the matrix continuum response.

    Hope this is clear enough.

    Stefan

    Like
  • Thank you very much Stefan, but unfortunately my laptop is under Windows.

    Is there any other manners?

    Like
  • Jaechul,

    I (purposely) mentioned Cygwin, which you could download to your PC to get a Unix-style environment under Windows, and then run the commands I proposed. But that is just one way to throw out certain elements. The solution to the problem as you stated it is to edit the MINC file and create two MESH files, one for the fracture elements, one for the matrix elements. I have to leave it up to you to find a way to edit your files.

    Hope the rest of the community can help to find a convenient way to solve the simple problem of deleting every other line (or lines with a specific pattern) using a PC text editor.

    Stefan

    Like
  • Thank you very much Stefan. I will try.

    Like
  • Mr. Ha,

    I wrote a short program (minc2mesh.exe) for you that splits MINC meshes into separate MESH files (one for each continuum) for subsequent EXT processing on a PC. The program is attached and will be available at http://esd.lbl.gov/research/projects/tough/licensing/free.html.

    Let me know whether this works for you.

    Regards,

    Stefan

    Like
  • I`m sincerely thanks to Stefan for your kindly response.

    I`ve just got two mesh files using you attached.

    But it is a case of mesh, I`m still confused because how to divide the output file after running (including two types of results all at once one file).

     

    Actually, I have got the divided result files depending on the time change for using Tecplot,

    but each result files (divided under the time) are still including two types (matrix and fractures).

    I want to divide each result file into two files: the matrix and the fracture.

    I really do not want to disturb to you.

    Best regards,

    HA

    Like
  • Dear Mr. Ha,

    Dividing your MINC file into separate files (one for each continuum) is exactly what minc2mesh.exe does. After running minc2mesh.exe, you should (for a double-porosity or dual-permeability model) have two separate files, named MESHfracture and MESHmatrix2. You can now visualize these separately (by first renaming MESHfracture to MESH, run EXT, then visualize with Tecplot; repeat the procedure for MESHmatrix2). 

    If you don't get the two files, please send me your MINC file and I will process it for you.

    Stefan

    Like
  • Dear Stefan,

    Yes, I could get the two files (MESHfracture and MESHmatrix2) as you mentioned.

    But it does not running EXT with the MESH (changed from the MESHfracture).

    I think MESHfracture and MESHmatrix2 files do not have CONNE. So it does not work, does it?

    Like
  • Jaechul,

    You should have a CONNE block, at least for MESHfracture. Note that in case you use a dual-porosity model, the CONNE block for MESHmatrix2 will be empty (as there are indeed no matrix-to-matrix connections!). If you set variable DUAL in block MINC to MMALL, you will get a CONNE block for all continua; if you use MMVER, you will get a CONNE block for the fracture and first matrix continuum, but the other matrix continua J > 2 don't have a CONNE block.

    Here is my next set of questions:

    (1) Do you get a CONNE block in MESHfracture?

    (2) If so, can you run EXT and visualize the fracture-continuum variables? You definitely should be able to do that (assuming you have set KDATA=2 or higher in the TOUGH2 input file). If not, what are the error messages?

    (3) Do you have a dual-porosity or double-permeability model?

    (4) If you have a dual-porosity model (i.e., DUAL is blank), you indeed may have a problem visualizing it (as no elements are created in EXT). There are ways around it: you  have to take the element orderings (the lines containing 4 integers at the end of the EXT output file created for the fracture continuum file) and copy these lines to the end of the EXT output file created for the matrix continuum. You also need to replace the ZONE lines from the matrix EXT output file with those from the fracture EXT output file. There are other options (e.g., creating a regular zone in Tecplot and interpolating it, etc.; too complicated top explain here).

    In short, there are definitely ways to visualize both the fracture and matrix continua using minc2mesh.exe and Tecplot. I don't know how your model looks like, but I presume you'll have to go through Step (4). If my instructions are not clear enough, send me (SAFinsterle@lbl.gov) the answers to Questions (1)-(3) along with the MINC and TOUGH2 output files, and I will do Step (4) for you once, so you can see how the final output should look like. 

    Regards,

    Stefan

    Like
Like Follow
  • 6 yrs agoLast active
  • 9Replies
  • 782Views
  • 2 Following