Welded Connection Fixity

[cazkoop]

We are currently load rating a truss where the connections are welded using gusset plates. When we input the truss into STAAD we assumed fully fixed connections at the ends. The resultant moments in the members were relatively small, but since the members bend in the weak axis, the load rating is very low. Is there any information out there that discusses partial fixity of such connections?

 

[J1D]

cazkoop,

Do you want to get information of study or experiment about the connection regidity, or how to model semi-rigid connections?

If it is a truss, the connection rigidity (either it is 100% rigid or simple a pin) shouldn't be a significent factor to affect the load rating.

 

[cazkoop]

Thank you J1D for your response. If there are studies out there on this matter that you know about, that would be great. Unfortunately, our load rating does change significantly when we add rigidity to the connection. Our STAAD model gives us moments in the members. Even though the moments are small, the members are bending about the weak axis. When applying the beam-column interaction effects in AASHTO, the capacity does decrease significantly. In the pinned analysis, there is no moment, so the interaction equations don't come into effect.

 

[J1D]

cazkoop,
I don’t have any document or article about rigidity of steel connection handy. I once wrote a program of concrete flat-plate structure analysis with non-linear “springs” to model the column to plate connections. For steel structures, because of the variety, it is harder to have analytical expression of connection rigidity. I know somebody did this kind of study, heavily based on experimental results. Maybe more study results are available now.

I don’t know whether your truss is 2D or 3D. But if the loads are applied to the joints the bending moment anywhere should be pretty small, and even with the beam-column criteria the strength should be dictated mainly by the compression or tension of the critical member. Or you just use pin connection approach. With gusset plates the connections are a "pin" in one direction already.

 

[Ron]

cazkoop....remember that your selection of an analysis mode does not dictate the action of the member. The way it is fabricated, erected, and loaded will dictate its reactions, without regard to the way you decide to analyze it.

If there is fixity in the joints, then analyze it so. I'm surprised that your load capacity goes down as compared to pinned joints. Are you sure your model is correct? I have seen the opposite in similar analyses. STAAD is pretty quirky so be careful with your model. Do an animation and see how that looks. That will usually tell you if something is amiss in your model.