Interlaminar stresses with Solid46

[plvachon]

Hi to all,

I'm trying to calculate interlaminar stresses in a composite plate (Sxz, Syz), and I compare my results with those from other authors. I'm using Solid46 elements with 2 elements through thickness for each layer. the case is very simple: a rectangular plate under uniaxial tension. Here is my problem: when I extract interlaminar stresses at nodes between 2 layers, the values are not the same as for the other authors. Here are the code lines for extracting stresses:

 NSEL,S,LOC,X,0
NSEL,R,LOC,Z,0.375     ! Select nodes
*GET,sxz,NODE,node_number,S,XZ   ! Sxz

I guess that the problem comes from the way I extract the stresses. I mention that all nodes on the interface between two layers are merged. Consequently, the solution at a node is the mean between the solution from the element above that node and the solution from the element under that node. So what would be the right way to get interlaminar stresses between 2 layers?

Thanks in advance. If it's not clear enough, please let me know!

 

[plvachon]

I just want to give you some precisions about my model. Let's consider a rectangular plate (with Length >> width). The length is -Length/2 < X < 0 and the width is -width/2 < Y < 0. There are two symmetry planes (length and width), so only a quarter of the plate is modelled. The layup is [45/-45]s (4 layers of 0.125mm thick.). The plate is loaded with a uniaxial tension of ux = 0.01 (1%). I want to get the SXZ stresses on a line that begins on the free edge of the laminate, going towards the center of the laminate (at X=0, with -width/2<Y<0). Moreover, the line is situated at the 45/-45 interface. The values of the stresses should be high at free edge and they should vanish as we go towards the center of the laminate (according to other authors). My problem is that I obtain a logical curve near the free-edge (although the values are too small), but at the center of the laminate, there is a peak where I should see 0... I guess that the problem lies in the way I extract interlaminar stresses with the *GET command on nodes. Here is a piece of my code:

ET,1,SOLID46
KEYOPT,1,1,1		! Suppress extra displacement shapes
KEYOPT,1,2,0		! constant thick.
KEYOPT,1,3,0
KEYOPT,1,4,0
KEYOPT,1,5,2
KEYOPT,1,6,4
KEYOPT,1,7,0
KEYOPT,1,8,0
KEYOPT,1,9,1
KEYOPT,1,10,0

(...)

*GET,SXZ,NODE,ndnum,S,XZ ! extract SXZ for all nodes on the line
(...)

Could anyone help me?