Composite Delamination in Nastran

[venkith]

Hello All,
Have anyone done a Composite delamination problem in Nastram before. Any small help would help a lot.

Thanks
VENKIth

 

[dmacx]

I am using Nastran now with 3d layered solid elements which give interlaminar tensile and shear stress. I use these stresses and compare them to the bond material allowable.

 

[GBor]

Using laminated plates, you have to look at out-of-plane directional stresses. The local "z" direction is your tensile/compressive stress. The local "xz" and "yz" shear stresses are interlaminar. You have to have a well-developed allowable and I would recommend strains instead of stresses since strains are continuous through the thickness.

Garland E. Borowski, PE
Borowski Engineering & Analytical Services, Inc.
Lower Alabama SolidWorks Users Group

 

[venkith]

Thanks dmacx and Borowski.

DMACX: Modeling as 3D layered in my case would be tedious in my case[too many variations in number of ply and properties]. But, I will use it relatively in smaller test cases. Thanks again!

Borowski: I was thinking of about this method[well, I am doing it actually], but the problem is I donot have a good number to compare with. Well, that's my problem . Thanks again!

 

[GBor]

What material system are you using? Carbon/Epoxy? E-glass/Polyester? Is it a material system that might be contained in Mil-HNDBK-17?

If it is a unique material system with little or no data, you need to seriously consider a test program. Even if the data is a well-known material, but you are processing it differently (you are using VARTM where others have Autoclaved) or if you are processing using a facility with an unknown "track record", you may still need the test program.

I would comment that DMACX's procedure can probably still be used in your situation. There are many programs that will "smear" the properties of laminates so that you can use 3-D orthotropic elements. In your case, you may use 5 layers with the two outer and the center-most layers being thin. This would allow you to check the tensile stresses and shear stresses at what are generally the critical locations...just a thought.

 

[SWComposites]

venkith - what specifically are you trying to do? what are you modelling (structure, material, loads, boundary conditions)? what are you trying to predict with the model? delam buckling? delam growth? or something else?

 

[dmacx]

What I was referring to was using a single hex element through the thickness and a laminate definition (PCOMP) to define the layer properties. This capability is possible in NEiNastran. I was not sure if you thought I was referring to using a single hex layer for each ply.

 

[GBor]

I was referring to a mix of the concepts. First layer of solid elements representing a few plys near the surface, thicker layer representing plies toward the middle, thin layer for the center ply or two (possibly an orthotropic plate), basically, smearing the properties of "less-interesting" areas and refining the critical locations.