Jump to content


  • Content Count

  • Joined

  • Last visited

Posts posted by rzmehdipour

  1. 12 hours ago, Hyperman said:



    try this approach:

    • mesh the non-design component and assign 2.5mm thickness
    • duplicate the elements into the design component, equivalence nodes (F3) and assign 6mm thickness
    • set up topology optimization referencing the design component with 0 base thickness
    • to get distinct element densities increase the analysis>optimization>opti control>DISCRETE=2 or 3

    Optionally, after interpreting topology optimization results into patches you could perform size optimization to get discrete thickness (DDVAL).


    Thank you very much. By setting DISCRETE to 3 result was much better than the original one and was OK for me. But generally it would be useful if ALTAIR developers make it possible for user to define desired thickness steps for such optimization problems.


  2. Hi there,


    I wanted to know if it's possible to define thickness steps when performing topology optimization using Optistruct. For example assume there is a body of PSHELL elements, Initial thickness is 10 and base thickness is 2.5 and between these two values only 6 is permissible. Or in another case only 8.5 and 2.5 are possible for final part. How can I go with this kind of problems? The real problem is to decrease some deflections of sheet metal parts by adding welded patches so thickness in different regions can not be any value. For example if initial thickness is 2.5 mm then I want to improve strength by only welding some 6 mm thickness patches to it so optimization output must be only 2.5 and 8.5 mm thickness.


    Please let me know if my question is not clear.


    Thanks in advance.

  3. 36 minutes ago, GAJENDRA KUMAR NHAICHANIYA said:


    See the average Lankford coefficient R is calculated by R0 R45 & R90 (to consider plastic anisotropy)


    Here R0 R45 & R90 values are calculated by R = Exy/Ez along 0,45 and 90 degree (assume any one direction as 0 degree for R0 and start calculating R45 & R90 )

    Exy = in plane plastic strain & Ez = strain along thickness

    so no matter which direction u r choosing for calculation of R, because R value not calculated by axial strain. it calculated by share strain (Exy).


    For more u can go through link



    Thanks again.

    You're right. But physically material does not act according to average values, but real ones. I mean in real world material will behave differently in different directions. So how can we consider rolling direction in simulations? Sometimes considering rolling direction is quite necessary in first forming stage and forming the sheet in wrong direction will cause failure.

  4. On 2/25/2019 at 6:00 PM, GAJENDRA KUMAR NHAICHANIYA said:

    r0, r45 & r90 are lankford cofficient. Hope ur both questions answered by this.

    Thanks for your reply but my question is : As formability is not the same in different directions (relative to rolling one), how is it possible to predict the result without specifying rolling direction on raw material. Is there any default direction for rolling?

  5. Hi all,


    1- As sheet metal resistance to thinning is not the same in different directions (which is the subject of Lankford coefficients), how can we define FLD curve without considering rolling direction?


    2- In Radioss one-step forming simulation in which we define r0, r45 and r90 during material definition, is there a pre-defined sheet rolling direction? Is it X axis or what?


    Thanks in advance.


  6. Hi all,


    I wanted to know if actual springback after forming depends on sheet rolling direction. As anisotropy coefficients (r0, r45 and r90) affect thickness reduction, I think they must be important in springback analysis but during routine process (as defined in related tutorial) which is performing springback analysis following an incremental forming simulation, they are ignored at all. Am I wrong?


    Thanks in advance.

  7. And you made me another question. Can we compare thickness reduction percentage from simulation with max permissible elongation of material? I think the latter one is from one-directional tensile test while the first in resulted from complex loading state. I mean as it is apparent from FLC, permissible major strain can be higher when there is minor strain and it seems it can be even higher than max elongation of the material. Am I right?

  8. Dear kpavan,


    Thank you very much for your reply. Let me tell you that I have used tensile test results to define the material in simulation. I mean I have entered stress-strain data to define the material. As for max elongation, in fact it is very marginal in some regions at third forming stage (both material and simulation results are about 25%). In forming stages 1 and 2 I have used blank holder but stage three is a crash forming process. So my question is while thickness reduction from simulation is acceptable, why FLD shows failure in some regions. Are thicknesses from simulation unreal? Or do I have to define some properties of material other than stress-strain curve? 


    Thanks in advance

  9. Hi all,

    I was running a multi-stage forming simulation on 1mm thick 3003 aluminum sheet. In fact there are three stages on which two adjacent depressions are made. Depressions are ovals about 20mm x 16mm and are 4mm apart. During first stage minimum thickness became about 0.79mm and FLD shows forming process is safe. At second stage minimum thickness will become 0.77 and again FLD is OK, but during third stage at which minimum thickness turns to 0.75, FLD shows large areas of failure. Now I'm confused about the results. Which one is more important? Final sheet thickness or forming limit diagram? On desired part necking is not permissible but 0.75mm thickness is OK. Attached file is a sketch of the part. 

    I would be very grateful to hear your opinions.



  • Create New...