Bilinear-Law of cohesive element in LS-DYNA

Hello, 

I'm coding a user materils for cohesive elemnt in dyna, But I met some problems, hope anyone could give some guidance or help . 

I'm trying to code my own user material subroutine for cohesive element in ls-dyna.  The cohesive element formulation is already predefined and exists  in *SECTION card ( elform=19). 

By default, four integration points are used for one cohesive element , so my first question is that (1)how do we change number of integration points for one cohesive element ?  Because this is not what we can decide in the subroutine part. And I couldn't find any option in any card which would do.

So with 4 points integration, i tried some basic tests on one cohesive element . For example , I fixed the 4 nodes of the lower surface and then I applied traction and / ou shear displacement on the upper surface nodes. Here comes the problem. When the displacements applied to upper surface nodes are uniform, the output displacement vs. traction/shear force corresponds well to the expected bilinear law. HOWEVER, when the applied displacements are not uniform ( say 2mm/ms at two adjacent node 1,4 and 1mm/ms at node 2,3 ),  the displacement saved for the element is by default the one interpolated on the first integration point ( -0.577,-0.577), which means i'll get a 1.7887mm/ms. But apparently, a average displacement of 1mm/ms and 2mm/ms should be more reasonable.   Meanwhile, I did the same tests with Mat 138 Cohesive mixed mode by using all same parameters, I get the same result, which is, the constitutive law of a cohesive element is reflected on the first integration point (-0.577, -0.577) instead of the point in the middle of plane.  

So sum up a little bit I raise my second question (2) Have anyone came accros the same problem ? and how do we cope with it ?

Thanks in advance for anyone's help !  

if anyone is intersted in this subject , I would like to have more discussion on it . 

daniel.cheng.chen@gmail.com

CHENG