Composite Freesize Optimization - Why did I get more Compliance when increasing possible orientations of superplies?

arh423
arh423 Altair Community Member
edited October 2020 in Community Q&A

Hey,

 

i am varying the possible orientations and so the number of superplies of my model.

For example I have two models:

Model 1: 0°,90°,+45°,-45°

Model 2: 0°,10°,20°,30°,40°,50°,60°,70°, 80°, 90°

With same boundarys and loads model 2 results in higher WCOMP.

In my understanding mroe possible orientations should result in a more efficient design and less WCOMP.

How is this explainable?

 

 

Thank you!

Tagged:

Answers

  • Adriano Koga_20259
    Adriano Koga_20259 New Altair Community Member
    edited October 2020

    I assume that you're keeping total laminate thickness the same in your comparisons, right?

    Esssentially you're giving way more freedom to your design by allowing it to have all these layers, so naturally the optimizer has much more freedom to find better designs.

    BUT, you need to take care, as adding these "non-traditional" orientations, you might also bring design issues such as non-balanced laminates, non-symmetric, and so on, that might mess up with your ABD matrices, thus resulting in some undesired couplings (extension-shear-bending-twisting).

    So use it with caution.

  • arh423
    arh423 Altair Community Member
    edited October 2020

    Yes right, total laminate thickness is the same.

    I added pattern grouping and balancing of the orientations to keep a symmetric laminate as achieved in model 1. But it don't helps to solve my issue.

     

    How does the inital thickess of my superplies does play a role?

    Of course, the inital laminate thickness should be higher than the optimized thickness and should let enough room for optimization. But I have the feeling that a "high" inital superply thickness or a "very high" initial superply thickness leads to a different optimized thickness which makes the choise for the right initial thickness difficult for me.


    Thank you!

  • arh423
    arh423 Altair Community Member
    edited October 2020

    Regarding the inital superply thickness issue:

    Starting with a very high inital superply thickness could be a good idea.

    Then I build a second model with inital superply thickness just a bit higher (maybe 50%) than the optimized results.

    The second model should perform better than the model with very high superply thickness I think.

    What do you think?

Categories