MAT2/MAT8: question non answered. 2

[airmail]

I have the same question Vorwald requested a couple of years before and didn´t receive any answer. I quote it:

I have a mat 8 card derived from a composite laminate properties (inverse of the A matrix of the ABD), and I have PSHELL/MAT2 generated from the PCOMP card. The answers are quite a bit different. Can some explain how to compare a MAT8 card to a set of PSHELL/MAT2 cards.

Does anybody know something about it? How to create the MAT2 card which comprise the laminate properties? (With only one MAT2, the command ECHO=PUNCH in Nastran is not what I want because it creates three MAT2 cards)

THANKS A LOT.

 

[SWComposites]

Please post the MAT8 and PSHELL/MAT8 cards that you are using. Are you using the MAT8 cards with a PSHELL or PCOMP card? post the one that you are using. Also, post the laminate layup and ply properties. Why don't you want to use the full 3 card set of MAT2 cards?

Regards,

Steve

 

[airmail]

Hi Steve

Let´s suppose next material with next stacking sequence:

Material properties:
E11=50000 N/mm2
E22=10000 N/mm2
G12=2500 N/mm2
u=0.3
Stacking sequence (6 plies)
(+,-,0)S

I´d create the corresponding mat8 for the material and laminate properties with a pcomp card.

The thing is: Could I create one mat2 for the material and one pshell calling that mat2 for the laminate? In this way, I can manage laminate properties. If I leave the mat8/pcomp relationship, nastran calculates the properties for me, and I don´t want that.

I hope this is what you request. Thanks

 

[SWComposites]

The question is: do you want to include the effects of the laminate stacking sequence or not in your model? If you want to consider the stacking sequence on the bending stiffness, then you need to use separate MAT2 cards with the PSHELL property card for the inplane and bending stiffness (either calculated manually or automatically by NASTRAN from the PCOMP/MAT8 card input). If you do not want to consider stacking sequence effects, but want to model the laminate with smeared properties, then you can either use a PCOMP card with a single MAT8 card that has the laminate stiffness values (Ex, Ey, Gxy, nuxy), or use a PSHELL card with a single MAT2 card that has the laminate stiffness matrix (Aij*t). Either should give the same result.

Please explain what you are trying to model and why you don't want to use a PCOMP/MAT8 card set with individual ply input.

 

[airmail]

Thanks SW for your post.

Well, what I want to do is to consider only membrane properties.

The reason why I want to use PSHELL/MAT2 is to have the flexibility of entering the properties I want, not the NASTRAN ones (internal calculations).

My model is a general FEM, not a detailed one.

Thanks again.

 

[mdhayes]

Airmail,

Why not just use membrane elements with a MAT2 entry?

When NEiNastran calculates the MAT12 (solid) entries, it scales/normalizes them as follows:

MAT2 ID2 matrix =[A]/t where t is the laminate thickness
MAT2 ID3 matrix = [D]/t/(1/12*b*t^3)
MAT2 ID5 matrix = /t^2

It may be the same for the MAT2 (shell) entries...If you have any more questions, send me an email: mhayes@mmelab.com

 

[airmail]

Hi mdhayes

Yes, that´s what I want to do! Simmulate a laminate with only a MAT2/PSHELL. Is this possible? My idea is to calculate the laminate properties and create a MAT2 (with the E1, E2 and G12 obtained for the laminate, but how?). Then, create a PSHELL calling that MAT2 where only I define the thickness of the laminate.

Do you think make sense? I´d like to have the flexibility of changing the thickness of the laminate in a fast way just changing the thickness in the PSHELL (I guess if I keep orientation percentage I keep the old E1,E2 and G12)...am I right?

Thank you guy!