Assumption of steel connection being rigid or pin 1

[ynnuseel0106]

When assuming the fixity of a steel connection, how should I assume whether it is fixed or pinned? For example, a rectangular hollow section welded all round to another hollow section. Is there a way to classify/justify whether a connection can be considered as rigid or fixed??

Thanks for reply.

 

[canwesteng]

Welded all around is pretty much fixed. The stiffness of the connection justifies whether it´s pinned or fixed. For example, typical clip angles have very little stiffness, which allows end rotation of the beams.

 

[Guastavino]

I agree with canwesteng. Although, I think if it's a truss you can argue pinned connection. In reality of course it's neither, it's somewhere in between. You'll also want to make sure whatever you assume that the connection is capable of transferring the loads. If you assume it's fixed, make sure the connection can take that moment.

With all that said, I don't think there is a cut and dry one size fits all answer. I'd consider looking at it both ways and get a comfort level with what's going on. For example, let's say you model it fixed and your analysis shows it transfers 10 k-ft of moment but the connection has a capacity of 100 k-ft. I'd think that could be considered fixed because you have so much reserve capacity that clearly it's going to be stiff.

 

[canwesteng]

To add to that, you should probably check the weld for moment from a fixed connection whether you design as pinned or fixed, since the weld will have a brittle failure.

 

[Leftwow]

When you say welded whole, I'm assuming you mean that the whole beam is welded to the column, which you could design as rigid. Here's the concept, if you don't have bracing, you fix your main connections in the directions where bracing doesn't exist. The reason you do this is to make sure that your connection can handle the moment. There was a time when engineers would predict the "fixability" of a connection, saying that it would have a certain percentage of rigidness. Then the majority of engineers decided to change that to either pinned or fixed, this way its much safer. If everything is fixed on a complex structure, your results will not give you the proper beam and connection loads, then the results will show that there is little or no moment transferred to the base of your structure, which in turn is not what one should assume. Many engineers that I have worked with like to pin their base on pipe racks and etc, however I think this should only be used when determining instabilities. It's ultimately all about where you want your moment to be transferred.

 

[jeffandmike]

This may be a bit more than what you were looking for, but it is an interesting read nonetheless:

http://www.ejse.org/Archives/Fulltext/2010/2010v1/20103.pdf

 

[ynnuseel0106]

Thanks all for your reply. So there is no clear cut between a pin joint or fixed joint, just an assumption. Then I have a question, if I assume an all round welded joint, say a hollow section to another hollow section, to be pinned, will there be a problem? Since in that case I won' t check the weld as I assume the moment at the joint is zero.

 

[jayrod12]

A full weld will act more fixed than pinned. So although it may not be a problem, it is incorrect to assume a pin and then fully weld it.

 

[Tomfh]

The connection is rigid and brittle and cannot deform into a pin without failure, thus you must design for the actual moment.

 

[MacGruber22]

Then I have a question, if I assume an all round welded joint, say a hollow section to another hollow section, to be pinned, will there be a problem?

ynnuseel0106:
1. If your beam is designed to be simply supported and the end connections are welded all-around, then those connections will need to appreciably deform/fail in order to re-distribute moments to the midspan.
2. Transversely loaded welds are extremely stiff and brittle, so in order for number 1 to occur it would take a great amount of end rotation. Until that fuse forms, moment will be transferred into the column.

So, yes it will be a problem if your column (and base plate, etc.) isn't designed to resist the moments that develop from the rigid beam end rotation.

You should review the Fully and Partially Restrained Moment Connection configurations in the AISC Steel Construction Manual. There you can get a better feel for those connections configurations that are rigid, partially restrained(traditionally called "wind moment connections"), and by process of elimination non-rigid or simple.



"It is imperative Cunth doesn't get his hands on those codes."

 

[SAIL3]

I ran into this question a few years ago....to solve it, I ran the structure in RISA first as pinned connection and then as fixed....this way I eliminated the uncertainty and designed for the most conservative result....

 

[Ron]

SAIL3....why would you want to throw common sense into this discussion?

 

[MacGruber22]

I agree that the method Sail3 recommends will likely work. But it will get you a cranky detailer if repeated many times over. Particularly if it means going from single pass fillet welds to heavier fillets, PJP, or CJP welds.

"It is imperative Cunth doesn't get his hands on those codes."

 

[SAIL3]

my bad...probably, should of used the Direct Analysis Method Not(DAMN) and introduced an intial deflection with reduced stiffeness and checked it all for Lateral Torsional Buckling(LTB) which seems to be the current flavor of the day.....

 

[canwesteng]

No matter what you do, that weld needs to be designed for the moment, since it's going to have a brittle failure mode. It's also all kinds of stiff so I see no harm in just calling it fixed. But occasionally doing what SAIL suggest is a good idea, although poor practice to be doing all over the place as you´ll have beefy beams and piles of bolts and welds at connections.