Beam to Column Connection Identification

[USCeng09]

All,

I have a steel structure (built approx. 1974 to 1978) with some column deterioration at a girder connection. The girder to column connection is a bolted extended end-plate connection which is welded to the web of the girder. At first glance it appears this is a moment connection, however the flanges of the girder are not welded to the end-plate. In fact, the top flange is not even in contact with the end plate. I've attached a photo of the connection (in good condition). Note, we do not have any documentation of the original construction, however it is known that the beam seat and stiffener plate were not part of the original construction and added by another engineer about 10 years ago.

I would like to know if anyone has seen a similar connection and if they can give some quidence into its original design. I have done some research, but cannot confirm if it is intended to be a partial-strength connection to reduce the simple span moment on the beam or if it was originally designed to be a shear connection. Any one have some insight into this?

 

[ishvaaag]

As you describe the thing quite likely the initial intent was to get a simply supported beam ... even if not well conceived for that. To reduce positive moments looks to me quite strange.

 

[marinaman]

I do not believe this is originally designed as a simple shear connection.

A couple things lead me to believe that:

1.) The bolts are grouped high and low. Typically, shear connections only are not grouped this way.
2.) There is a stiffener across the column web at the elevation of the girder bottom flange.
3.) Original or not, the bottom flange seat appears to be welded to the bottom flange, thus
adding to the rotational restraint of the girder end.

How much rigidity was originally there?....hard to say. With the flanges not originally welded, it would be more rigid than many typical shear connections...but less rigid than welding up the flanges to the face of column. Also, like I say, coming back 10 years ago and welding up the stiffened plate to the bottom flange made the end connection even more stiff.

In analysis, you may have to look at it twice. Once, assuming a more flexible end condition, resulting in more moment out in the girder span.....and again assuming a more rigid end connection, resulting in more moment in the column, girder end, and girder across the column web.

 

[ishvaaag]

You can model one extreme behavior of your connection by assuming entire fixity at ends of the beam modeled with 3 segments, one central with the ordinary whole section properties' and very short segments of only the web properties at both ends (as for a coped beam).

The alternative to this is to get proper rotational fixity stiffness data for the connection. I must say that the connection is not altogether unknown to me but I don't remember where I have seen it (apart from when for simply supported cases) and an inspection looking for the case in what I have may take some time.

Maybe the rotational stiffness of the connection, since bearing on a quite rigid front end plate and restricted by the flanges can be initially extrapolated from that the double tee semirigid connection and then factor by the quotient of inertia of the web divided by inertia of the whole double tee. Data for the semirigid connection double tee in the Chen and Goto text (even if I doubt applicable to front plates restricted to the height of the double tee, but it is a start). Probably the approximate estimate as per above should render even less rotational stiffness than actual, because the front plate and flanges must be influencing the stiffness more than proportionally as assumed.

A third approach and also reasonable in effort for such single question is to make a 3D solid FEM model. This will get quite well the behavior.

 

[BAretired]

It would appear that the original designer intended to develop the web in bending. The more recent addition of the seat and stiffener appears to have been intended to develop a bit more negative moment, i.e. the tee section consisting of the web plus bottom flange.

If you need to rely on negative moment, it might be prudent to develop the top flange as well. If the beam works as a simple span, leave it alone.



BA

 

[paddingtongreen]

The AISC used the general rule, back then, that if the plate was less than 5/8" thick, it was flexible. That one looks thick enough to consider it stiff.

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[ishvaaag]

You can also allocate any bolt capacity not used for shear to stand moment action, (some attention required if only part of some bolts available for that at the checks ...) so in the end you derive what moment both the connection and web stand to the purpose of reducing positive moment. Put simply, some bolts would be standing moment, and some others shear. Adding the seat as someone did might relieve all the capacity for moment and so enhance the moment strength to the 3+3 bolts in the web moment capacity.

 

[ToadJones]

Original connection was likely a flexible shear end plate connection. I believe there are still design examples of this on the CD's that come with the steel manual...2005 or 2010