Print elements command

[castonguay.pierre]

Hello,

I have noticed something strange with OpenSees.

I have a model of a tubular steel brace with gusset plates modeled at both ends as zeroLength elements (rotational springs). Displacements are imposed at one end of the brace to evaluate a buckling load, and buckling takes place in the XY plane (brace longitudinal axis is along the global X, and geometric initial imperfections in the Y directions are used).

Here is how I defined my zeroLength element:
uniaxialMaterial Elastic 4 2445351.348
uniaxialMaterial Steel02 5 166388.866 2384217.565 0.01 20.0 0.925 0.15
element zeroLength 1 1 2 -mat 4 5 -dir 5 6 -orient 1 0 0 0 1 0;

So, material 4 is specified in direction 4 (torsion) and material 5 is specified in direction 5 (My).

I happened to print information about my elements, and here is what is specified about element 1:

Element: 1 type: ZeroLength iNode: 1 jNode: 2
Material1d, tag: 4, dir: 3
Elastic tag: 4
E: 2.44535e+006 eta: 0
Material1d, tag: 5, dir: 4
Steel02:(strain, stress, tangent) 0 0 2.38422e+006

Why is it telling me that materials 4 and 5 are specified in directions 3 and 4, respectively? Am I not reading this correctly?

My buckling loads for a pinned brace and a brace with rotational springs at both ends are the same when using direction 4 or 5. When specifying 5 and 6, these rotational springs increase the buckling load by about 6%, which makes more sense. Using direction 5 and 6, the print commands shows me that 4 and 5 were specified...

[castonguay.pierre]

Sorry, the definition of my zeroLength should read as follows to correspond with the print element command shown:

element zeroLength 1 1 2 -mat 4 5 -dir 4 5 -orient 1 0 0 0 1 0;

Sorry about that, I'd been playing around with it and forgot to put -dir back to 4 and 5 before pasting it on here.

[fmk]

for performance reasons, internally the dof are numbered 0 through ndof-1 .. we subtract the 1 on the i/p.