Constant curvature and constant twist patch tests

[illini8181]

I am trying to understand the details of how a constant curvature and a constant twist patch test are applied, and what the desired results are.

Doing some internet browsing, I found the linked book, which has an example of a constant curvature test and a constant twist test.

http://upetd.up.ac.za/thesis/submitted/etd-03062006-114313/unrestricted/05chapter5.pdf

In the constant curvature patch test, the boundary conditions are zero displacement in x,y,z directions at the upper left node, zero displacement in x and z directions at the lower left node, and zero rotation about the y-axis at both the upper and lower left hand nodes. A bending moment, M is then applied to the upper and lower right hand side nodes.

FEM is then used to calculated the curvature across the patch. If the curvature is constant, the patch test is passed. Is that correct?

In the constant twist patch test, all nodes are restrained in the x and y directions, and three of the four corners are restrained in the z-direction. A force is then applied to the fourth corner in the z-direction.
With this test, I am not sure what the outputs of the test are, and how the test is passed? Can anyone give me a hand?

 

[rb1957]

for me a patch test is a single element test, and the idea is to compare the FE output with your hand calcs. for me it's not "If the curvature is constant" but i'm more interested in the stress results (the displacements should match well as these are typically the assumption that the FE makes).

Quando Omni Flunkus Moritati