I've created several surface contacts between shell elements that are part of a fairly complex 3 dimensional sheet metal surface object that continuously but gradually twists in all three axes over a large span (several feet). The part is not orthogonal to any plane in the global coordinate system, and there are no true flat planar surfaces anywhere between the elements.
Surface to surface, non-symmetric 'offset bonded' contacts are placed manually between the surfaces (there are no automatic contacts in the model, everything is manual). I selected 'A' surface as the primary, 'B' Surface as the secondary. (colors correspond to the Primary and secondary entries int the Nastran workflow dialog). The activation distance is set at
[ 1.1 * sqrt(element_size^2 + thickness^2)],
where thickness in the sheet metal thickness and the element size is on the order of half an inch. The meshing and contact nodes seem to resolve just fine.
After loading and constraining, I am able to output the Nastran File and have been able to use the FNO reader to extract a subset of the generated surface contact nodes X Y Z coordinates into excel, along with each node's force vector [202] T1, [203] T2 and [204] T3 (similar to the method outlined in another post about method to sum contact forces across a surface).
Here's where I get fuzzy...I assume the force vectors [ T1 T2 T3 ] are output in the global axes with T1 being force along the positive x axis, T2 being force along the positive Y axis and T3 being the force along the positive Z axis. But I'm not clear on whether the nodes are essentially "attached" to the 'B' surface and imparting a force in the summed vector direction onto the 'A' surface, or are the nodes "attached" to the 'A' surface and imparting a force to the "B' surface in that vector direction?
For a very simplified example, if the 'A' surface is the global XY plane, and the 'B' surface is a parallel plane 0.25 inches above the 'A' surface (figure two pieces of 1/4" plate sitting on top of each other, and the shell elements are the midplanes), and connected via offset bonded contacts across the surface.
Now assume I apply a pressure downward on the 'B' surface in the [0 -1 0] direction (downward along the global Y axis), and I constrain the 'A' surface to the ground XY plane.
Would the resulting force vectors from the contact nodes generated by the surface contact be reported in the Nastran card as
A) [202] T1 = 0; [202] T2 = *some POSITIVE number* ; [203] T3 = 0
or would it be reported as
B) [202] T1 = 0; [202] T2 = *some NEGATIVE number* ; [203] T3 = 0
It's important, because I have to run a 3D transform of all the global force vectors at my subset of nodes to correspond to a coordinate system roughly parallel with the subset that I'm interested in (one of our brilliant engineers has already set up a VB program to run in excel and generate these transform matrices for me as I need them). Otherwise the global force vectors don't allow me to understand the net shear across the plane and the net normal contact force between them. I at least need to know if the surfaces are being compressed together or pulled apart! (And it's NOT intuitive, there's a lot of forces an moments and eccentricities pulling stuff every which way on this guy).
Thanks for all your help!
