Announcement

Collapse
No announcement yet.

Inverse calibration using Mcal and Abaqus

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Inverse calibration using Mcal and Abaqus

    I am modelling a 3D printed lattice structure so I need the material properties of the lattice build. therefore, I did some experimental compression tests on a single strand (material) to see what is the behaviour of material and seemed it is kind of crushable foam.

    The geometry in abaqus is very simple (single circular bar). I followed all steps in the" calibrate a material model using abaqus simulations" pdf file (exist in Mcalibration tutorial) and created the same geometry and BC and create .inp file. (attached is .inp file and data from experimental test)

    I saved .inp from abaqus along with .txt of experimental result at a location path :D:\abaqus_NSC\Mcalibration_abaqus

    Then I check at mcalibration preferences button to make sure the specified path to abaqus is correct which is: C:\SIMULIA\Commands

    The error faced when I click the run once button at Mcalibration is that:

    “Could not find the file: 'D:/abaqus_NSC/Mcalibration_abaqus/AbaExtSolver_LC_1.data' [4]

    Load case 1 ('abaqus_mcalibartion_f_d_2') finished: final stress=0

    Error: Could not calculate the fitness value.

    Simulation done”

    I check the mentioned path and found that there is no file called “AbaExtSolver_LC_1.data”. after click run, the only files that created are “AbaExtSolver_LC_1.inp”, “AbaExtSolver_LC_1.py” and “AbaExtSolver_mat_LC_1.inp” Files in the directory after running the calibration.


    I think the error probably related to the Mcalibration cannot run the abaqus file (or connect the abaqus) maybe due to installation path or something like that. What do you reckon? I attached .mcal file as well and you can run it.


    Can you please help me to find a solution as I don’t know what the main factor that cause the error and would be highly appreciated to help me?

    Best Regards
    Maedeh
    Attached Files

  • #2
    Interesting problem. Your zip-file contains a data file with 3 columns. I assume column 1 is time, column 2 is displacement, column 3 is force. Is that right?

    I then noticed that your Abaqus inp-file has the following step definition:
    Code:
    *Step, name=Step-1, nlgeom=NO
    *Static
    1., 1., 1e-05, 1.
    That is wrong. The STATIC step need to follow the same applied displacement history as in your experimental data file ("Specimen_RawData_1.csv"). The final time point in your experimental data file is "195.38802", so your step needs to have the same final time.

    Also, you should divide your STEP into multiple increments, and I think you should turn on NLGEOM=ON.

    There might be other issues, but try these changes first.

    -Jorgen
    Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

    Comment


    • #3
      Thanks Jorgen,
      I modified it (please find the attached modified .inp file) but there is still the same error. I think something is related to that Mcalibration cannot link to abaqus and find the .data file. is there any thing that I should do before running Mcalibratin e.g pre-installation requirement??

      I would say I only installed trial licence of Mcalibration (not Polyumod) and I sent many email to Veryst to purchase an academic licence but still waiting for their response.(after about 4 days).
      As my research got stock due to the inverse calibration using abaqus and Mcalibration, I highly appreciate if you can please assist me in this point.

      Regards
      Maedeh
      Attached Files

      Comment


      • #4
        There was something wrong with your mcal-file. I recreated it by: (1) starting a new MCalibration window; (2) reading in the experimental data; (3) copying over your plasticity model from the original mcal-file.

        Here's a screen shot:
        Click image for larger version

Name:	mcal_Jorgen_01.png
Views:	13
Size:	317.7 KB
ID:	25928

        I then clicked "Run Once" to see what your Abaqus FE simulation is doing. The simulation starts just fine, but the deformation field seems odd. This is what I see:

        Click image for larger version

Name:	AbaFig.png
Views:	7
Size:	123.5 KB
ID:	25929

        Is that what the deformation should look like?

        -Jorgen


        Attached Files
        Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

        Comment


        • #5
          Use the following web-page to purchase an academic version of MCalibration.
          Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

          Comment


          • #6
            Thanks Jorgen,

            I am just really confused and frustrated as I open the attached mcal file you sent and run it , the same error happened again !!!
            can you please let me know if the path of abaqus (in preferences) should be the same as the path of saved .inp file and .txt (experimental file).
            in other words, where should I save the .inp file, text file?

            about the deformation, it is wrong as I applied the 0.8 mm compression displacement on top of the strand line and fix the bottom line.
            Do you think it is the correct BC based on experimental test?

            Thanks a lot for all your hep on this point.
            Maedeh

            Comment


            • #7
              The path of Abaqus that is specified in the Options Dialog box should be the path to the script that starts Abaqus. Here is an image from my Linux computer.

              Click image for larger version

Name:	MCal_Options.png
Views:	15
Size:	338.9 KB
ID:	25933

              On a Windows computer the abaqus run script typically ends with .bat. For example: abq2019.bat

              Note that this path has nothing to do with the inp-file or the experimental data file. You can save the inp-file, the experimental data file anywhere you like on your computer. Note that you should save your mcal-file in that same directory.

              It sounds from your last message that the inp-file that you provided does not apply the deformation field that you had in mind. I recommend that you fix that.

              -Jorgen
              Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

              Comment


              • #8
                Many thanks Jorgen,

                I found that I have to save mcal and experimental file in to the folder that is Abaqus work_directory (for example my abaqus work directory set to D: Abaqus but I saved my mcal file and experimental file and .inp somewhere ales like H: test) that is why id was not able to find some files.
                Thanks a lot for all you help.

                I know that I should modify the abaqus (geometry, BC) to represent the experimental conditions (compression test) very well.

                One Mcal question is : why I cannot see prediction line (blue line) in the Mcal line as simulation done (below screen shot from my PC?? I used crushable foam model)

                Comment


                • #9
                  Hi Jorgen,

                  Thank you so much for your help. your comment was really helpful.

                  I asked one question in above post: One Mcal question is : why I cannot see prediction line (blue line) in the Mcal line as simulation done (below screen shot from my PC?? I used crushable foam model)

                  Two more questions?


                  1. can you please let me know if the inverse calibration using Mcal and abaqus is working for abaqus dynamic-explicit model?
                  2. I am validating the inverse calibration using a case study (a video I found from Veryst website) to understand how it is working. Can you please let me know where I can get files including ( .inp, Mcal and .txt for dogbone test in the video or any case study). I would say in the case_studies folder there is a .zip file but it did not provide the example (only template).


                  Thank you once again for your help on this point.

                  Best Regards
                  Maedeh
                  Last edited by maedeh8; 2019-05-15, 16:25.

                  Comment


                  • #10
                    Hello Maedeh,

                    Here are my answers:
                    1. If you don't see a prediction line in MCalibration then that typically means that the load case did not run to completion.
                    2. Yes, you can use Abaqus/Explicit also for inverse calibrations.
                    3. I am not sure I know what files you are looking for. Here are some files that I have used as an inverse calibration example.
                    -Jorgen
                    Attached Files
                    Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

                    Comment


                    • #11
                      Thanks Jorgen,
                      I just run the .zip files (including inverse calibration example ) you uploaded, here is the screenshot that shows no blue line and seems there is an error in the model you sent somewhere.I just wanted to validate the inveserse calibration.

                      I modified the abaqus model to has the same BC and loading conditions like experimental (compression loading on a circular fibre). but I don't know why I cannot get the inverse calibration results based on abaqus isotropic hardening.

                      Can you please have a look at the below url (I uploaded through a website as I only have 146KB remaining to upload) (including .inp and exp data) and let me know your thoughts?

                      https://ufile.io/7g445nox



                      One more thing is related to abaqus built-in (template model like crushable model, ): when I run I faced an error: Error: No sta-file exists. any tips?

                      Many thanks once again for your help.


                      Best Regards
                      Maedeh

                      Comment


                      • #12
                        I noticed that the example that I provided was based on an old file format. I have attached an updated file that works for me.

                        Click image for larger version

Name:	calibrate_inverse_example.png
Views:	11
Size:	194.6 KB
ID:	25970

                        -Jorgen
                        Attached Files
                        Jorgen Bergstrom, Ph.D. PolymerFEM Administrator

                        Comment


                        • #13
                          Thanks Jorgen,

                          I would say I purchased the academic licence and downloaded both Mcal and Polyumod softwares in my Virtual Machine account (Because at this time I am using VM not my personal computer).
                          I tested Mcalibration using Mcalibration_Test_cases_1 folder to make sure Mcal software is working. I wold say It is working for default solver.

                          But my problem is that when I want to use Mcal with Abaqus (as an either Abaqus external solver for inverse calibration of your files or build-in material model using Abaqus solver instead of default solver.

                          The problem is that when I click “run once” it takes ages (after even 2 hrs) nothing happen and in a simple words “seems that Mcal cannot connect to abaqus”.

                          I did the below changes at .mcal
                          1. In preferences : set “C:/SIMULIA/Commands/abaqus.bat” for abaqus command
                          2. In preferences use a temporary directory : “C:/Temp” as it is the same abaqus temp/working directory
                          3. Save .mcal and exp csv file in C:Temp (same as working directory).


                          I would say when I used the trial version of Mcal software at my PC (not VM account) and run your latset files, it worked fine and I did not have any problem.



                          Can you please help me to sort it out?

                          Best Regards
                          Maedeh


                          Comment

                          Working...
                          X