Bending simulation

JolanJolan
Hello guys !
Jolan here frome France :)

First of all, good job with Mecway, this software is marvelous. Keep up the good work :D

I'm not new in the FEM universe but i've only been working on linear problems on Ansys.

I've started a job in the "blanking and stamping of metal sheets" field, wich invole plastic behaviours, large displacements, long story short : nonlinearity.

So i'm trying to solve a few problems we have here with Mecway's help, and i've some issues with a bending problem. (More or less the same artifacts encountered in a previous thread concerning a sheet bending : Looks like it's exploding.)

I've tried to turn off auto time-stepping with a timestep of 0.01s but it crashes at 0.05s. :(

A few tips or advices could be usefull as i'm quite new in Calculix use. :)

Thanks !
Jolan

Comments

  • JolanJolan
    Oh I forgot to add some views to illustrate my concern.
    Here's the geometry, i've currently a simulation running and i don't have any screenshot of the "artifact" but it looks like the one attached (yet it's not this part).

    Thanks for any help :)
    Jolan
  • SergioSergio
    Hi Jolan, welcome on board! One of the mantra of my first FEA mentor (a venerable French guy also) was... "Sergio, remove all the unnecessary areas of the model" and he was very right! Removing the features and volumes that will not help you on the solution will lead to a small model that will run in much less time.


    I would reduce the tools shapes to the contact surfaces only (mesh the surfaces, only leave that and then extruded with only one element in thikness), and also remove the third block (that probably you put to guide the moving tool). Replace that block by boundary conditions on the moving tool.

    Another tip...when you have contacts, try to use hexa elements, they will converge better than tets.

    Regards
  • JolanJolan
    Hi Sergio

    Thanks for your support, you are right i'm currently removing every details i don't need. :)

    I want to try it in Hex elements, but i'm not sure to be able to do it with automatic meshing, as i am importing STEP from a solidworks part.
    Is it possible to do so (Gmsh command maybe ?) because i'm stuck with Tet elements as soon as i go for a generated mesh :(

    Best regards
    Jolan
  • SergioSergio
    Jola, do first a 2d quad dominant mesh on all the geometries, and then just extrude the faces that you need to create the hexas/wedges. Then you can delete all the rest 2d elements
  • JolanJolan
    Hi Sergio !
    Thank you for the tip ! I'll give it a try in Hex elements :)

    Regards
    Jolan
  • SergioSergio
    If you upload the steps incuded in the model I could give a look later. Take a look on the forum, there is another thread with something very similar.
  • JolanJolan
    Sergio

    Here are the step i used.
    (I moved the part called Poincon by 5.5mm in Z to have the starting situation)

    I saw the other thread about this artifacts, but as far as i understood i need to decrease the time step and desactivate the auto time stepping... Bit i gave a try one night at 0.0001 sec time step and it crashed :(

    In the ccx documentation they talk about middle nodes being source of problems :
    "If you include contact in your calculations and you are using quadratic elements, first avoid to include middle nodes in the slave surface."

    Again thank you for your help :)

    Regards
    Jolan
  • feazyxfeazyx
    Hello Jolan and Sergio,

    This is my very first post since I just discovered Mecway.
    I am a stamping simulation engineer. At this moment we are using a high-price Autoform software for our simulations. I am also using dataM Copra to make roll-forming simulations. I would like to replace both of this softwares with Mecway but I don't know if I can make it.

    Regarding your simulation, did you consider using a membrane element instead of solid for the blank? Also, don't you need to use a so-called "serre-flan" in french? An upper pressure to maintain the blank in it's position during forming?

    Regards
  • JolanJolan
    Hi Feazyx

    I too am a stamping engineer, and working in a very small company we can't afford Autoform. Mecway seems to answer a lot of questions we have.

    For the "Serre-Flanc" (The blank holder) you are right, i chose to add the geometry with contact constraints, because i did not found any "compression only" condition, but as i designed the real springs for the blank holder in my tool, i can clearly use it to guarantee the compression only condition :)
    I'll give it a try in the future : should reduce complexity :D

    I'm not sure to be able to use a membrane, because what we really have is a part stamped, you can see the radius in the middle, then cut along Z axis.
    So we have a particular geometry because we don't make a full "normal" cut but in the radius we have a Z-axis related cut line.
    (Not sure if i'm clear here ^^ )

    I can not upload the real blank model because of privacy ^^
    Best i can do is the picture attached with the real blank.
  • JolanJolan
    Hey there
    Double post for updating the feed :

    Finally reached convergence by drastically reducing the timestep and adding a blank holder.

    (Sim time : 0.5sec, timestep : 0.001sec )

    Bad new is i did not set a *NODE FILE,FREQUENCY card so it did stor every step and ended up being a 7GB file... And it crashes mecway if i intend to read the results. :( (16GB RAM btw)

    A new simulation with 0.002 sec timestep and 1 out of 25 steps stored is running on my computer right now :)

    I now need to try with a "virtual" blank older (only a pressure on the blank based on my real springs). It might speed up the calculation :)
  • feazyxfeazyx
    Hi Jolan,

    By doing as you are trying you will reduce all previous deformation from stamping to zero.
    Wouldn't be better to start from a flat membrane, stamp it and bend it after?

    Offtopic, did you made any deep drawing simulation using Mecway? I did't find much information on this topic.

    Regards
  • JolanJolan
    Hi Feazyx

    It is actually what we're doing : we stamp the radius, then cut the edge's geometry then bend it.

    Basically we want a square shape on the ZX Plane, conserving the radius along the Y axis.

    The fact is we have some issues with the geometry on existing tools and i need to know what are the elements involved, in order to change the geometry cutted just before the bending.

    So i don't have so much flexibility in my model : It already exists in our current tools. :)


    Concerning the deep drawing i did not, i'm new to mecway and started with simples (yet not as simple as i thought) bending sims.

    I'll later try axisymetrics deep-drawings to see how precise my timestep needs to be.

    But clearly it is a goal for my small company : being able to evaluate the force needed for complex dee-drawings :)

    Regards
Sign In or Register to comment.

Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In Register