Compound beams using rigid links

[PSSC]

I am analyzing a structure with vertical columns made from (2) w10x33 stacked on top of each other.
They are skip welded at the flanges and bolted through the flanges.
They are desinged to be one solid structrual member.
In my model I linked them with rigid links straight across ever foot or so.
When I run my model with a wind load the results I get are one of the pinned footings is in compression and the other is in tension.
I believe these are in reality acting as one member, so both should be in tension or compression.
Each of the columns are anchored in a pinned connection.

I started to put rigid links trussed diagonally from connection point to connection point but I am afraid this will make the column appear to be much much stronger than it is.

Any feedback would be greatly appreciated.

 

[JoshPlumSE]

When you link two beams together like that you are forming a built-up section. Where one beam is akin to a tension flange and the other beam is akin to the compression flange. Therefore, any moment that gets created / resisted by these combined members will result in tension in one side and compression on the other side. In order to get the total force in the built-up section you would have to sum together the forces in the two members. In order to get the total mometn in the built-up section you would have to sum together the moments and account for the force couple caused by their axial forces.

I would recommend one of two things:

1) Instead of using the rigid links, model this as a single member with the A, I and J of the built-up section. Then the force results for that member correspond to the design forces for the built up member. It's much easier to interpret your results in that case.

2) Continue what you have done. But, at the bottom of the column (where you have the boundary condition) replace the two boundary conditions with a single boundary at the centroid of the built-up cross section. This boundary condition will be connected via links to the two column members. That way, you only have one reaction result to worry about.

 

[PSSC]

Josh,

Thanks for the reply.
I think I will go with one connection at the centroid of the two members.
They are supported together at one pier, so it seems to me that one anchor point is more realistic.
Also I am done with the model, and that will be the easiest fix .