Beam to Column Pinned Connection

[sunyaer]

There is a challenge task to design a steel transfer beam to support a few storey concrete hollow block masonry wall(190 thick)above. These masonry walls above are load bearing ones, giving a total force of 5,000KN on the beam. The beam spanning 8 meter will be supported on two steel columns, the reaction on one column would be around 2,500KN (factored). The initial size of the column is W250x115, while the beam is W760x484. This frame is in a lateral shear wall system, so the beam in question is laterally supported by the diaphragm. However, the columns in question will have to be braced by the beam or by other method.

The height of the columns is 5.2 meter (from base plate to top of the supported beam,the far end of the columns is considered pinned connection with base plate, although 4 anchor bolts will be used to comply with code for erection safety.)

There are some constraints that make us to not increase the size of the columns. This gives rise to finding a solution for a beam to column pure pin connection, which transfers no moment from the beam to column. If this could be achieved, a column of W250x115 with effective length of about 4.6 meter would support an axial load of 2,500KN.

Some have suggested a beam to column pinned connection like that in the attachment. Do you have experience using similar pinned connection to transfer a magnitude of an axial load of about 2,000 KN to 3,000 KN without considering moment? If a certain level of moment which would be transferred from beam to column, how to consider the magnitude? (Let's assume a 4 inches wide bearing plate as shown in the attachment is used.)

The design uses Canadian code.

Thank you in advance.

 

[dik]

Can you go to a thicker 'crown plate' and eliminate the strips at the side?

Have you considered any arching action in the masonry wall it's supporting.

Dik

 

[sunyaer]

Yes, thicker "crown plate" without the strips at the side is a good idea.

To be on the safe side, We don't want to consider any arching action in the masonry wall it's supporting.

 

[dik]

I've got several that have been standing for nearly 40 years... Relatively small steel beams, spanning 20'+/- and picking up 18 to 20 storeys of 8" CMU walls... laterally supported at each floor line.

Dik

 

[sunyaer]

Hi Dik, thank you very much for your inputs.

My spans are a bit longer than 20', about 25'. Technically, I think there will be a certain level of arching action. But this needs proper details for reinforcing and grouting the CMU walls. The 8" CMU walls being used are standard hollow concrete block.

Are your beams supported on steel columns? Would you mind elaborating a bit on the beam to column connection, as well as the reinforcing and grouting detail for the 8" CMU walls?

 

[dik]

Been a long time... but, if I recall, they are supported on masonry with the large portion of the wall weight being supported by other parts of the masonry wall bearing system. I don't recall using steel columns at all...

Dik

 

[sunyaer]

The challenge is in designing how the beam to be supported on the columns to achieve the beam transferring no moment to columns.

 

[Agent666]

There is no such thing as no moment being transferred to the column, you will presumably be governed by the minimum eccentricity requirements if they exist in your code for 'simple' construction. Typically for a cap plate like this the eccentricity is from the centreline of the column to the edge of the bearing plate, i.e. 2500kN x 'e', if you had 100mm eccentricity this is a moment of 250kNm for example. You would check your column for this moment + any about the minor axis with the axial load in accordance with your code equation for the axial/moment interaction

For stability there also needs to be some continuity between the column and the stiffeners to take the 2.5% of the axial load in the column about the minor axis of the column (i.e. the code restraint force). Alternatively another beam orthogonal to the beam in question might be sufficient to restrain the column if appropriately detailed. If its pinned in both directions it is unstable, top of column can kick out sideways.

 

[sunyaer]

If the bearing plate (the angle plank shown on the attached picture)on the bottom of beam is only 4" wide, which is at the center of the column, that is the only contact between the beam and column, how would the column get the minimum eccentricity and the moment?

Bracing the column is another thing to be worried about. I am thinking just welding the angle plank to the cap of the column. Please input your comments, thanks.

 

[Agent666]

This moment occurs by sitting on the edge of the bearing plate 2" off the centre of the column (i.e. half the width of your bearing plate), when the end of the beam rotates under loading. Alternatively fabrication tolerances could cause it to sit on the edge of the plate for example. Usually codes require a consideration of this (or some) minimum eccentricity for these reasons. The only way to design for zero moment would be a real pin assembly, but even then you'd probably still design for some moment (even nominally pinned connections develop some level of moment in real structures for example). Designing for only an axial load is in my view ignoring what is occurring in reality.

How you brace the column has a direct impact on the effective length, if you were to bolt it and prove there was sufficient clamping force from the 2500kN load its pretty much going to behave as if it was fixed in my view and be considered as continuous up to the top of the beam where presumably the beam is fixed to the diaphragm. I'd expect with the large axial load that this will be the case more or less.

For the column design I'd look at both conditions, minimum eccentricity and the fixed moment that might be developed. You can in these situations consider whether or not under the given axial load if the beam bearing plate lifts off the cap plate, or if these is effectively sufficient clamping force from the axial loads to effectively make the column continuous with the beam/stiffener arrangement (i.e. demonstrate that the column has no tensile stresses).

The only condition where I'd consider the effective length being 4.6m would be if the bottom flange of the beam was restrained in position via a fly brace system or similar depth beams in the orthogonal direction (or similar detailing). If its a standalone beam and column with no other structure then the full length is warranted in my view, and you have to address the stability of the system.

 

[sunyaer]

Hi Agent666, you are absolutely right. The top of the beam is braced by diaphragm. However, due to the fact that the beam is so deep (about 3 feet high), the bottom of the beam can not be considered securely bracing the column, which is the beam's supporting member. The bottom of the beam may side sway due to accidental forces. When this occur, the system suddenly become unstable.

By looking again into the beam supporting slab detail, there are stiffeners bracing the bottom of the beam from diaphragm, Although from what it shows the stiffener is 6mm (1/4"),which looks too thin for a 3 feet beam.

If the 6mm (1/4") beam stiffener, which is spaced 2'-8" apart, is increased to 3/8", and weld the 4" bearing plate to the cap plate of the column, would this look better?

Another option I am considering is to flip the column and double shear connect the beam to the column web. For this option, greater moment to the column needs to be considered while the moment resistance of the column about weak axis is smaller, making it an option I didn't got into it in the first place.

Your comments are welcome and appreciated.

 

[sunyaer]

Here is the sketch.

 

[Agent666]

Based on your sketch I don't really appreciate how you are resolving the beam stability into the slab. Based on your 2500kN axial load , lets say the force kicking out the column is based on 2.5% of this, therefore 62.5kN. Multiply this by 900mm (3') lever arm gives you a moment of ~56kNm which needs to be resisted by the connection of the beam flange to the slab, and by the slab itself in close proximity to the stiffener arrangement. Based on the sketch there isn't much to show how this moment is going to be resisted based on your proposed arrangement.

Whether the bearing plate is connected to either the beam or to the column via welding doesn't really matter, you still have the eccentricity moment to design for.

I like the idea of connecting into the web if that works, make the end plate flexible enough and it can rotate without developing a huge moment. This solves the stability issues and complicated loadpath to resolve the above point regarding moment cranking into the slab. You are dealing with a significant transfer structure and want it to be fairly robust even if it means a slightly heavier column to achieve better detailing.

 

[sunyaer]

Thank you very much, Agent666

I think the moment 56 KN-m we are talking about is actually the torsion acting on the beam about the centre of the top flange, is it correct?

You mentioned flexible end plate, are you saying the connection is like the attached picture? How thick is the end plate that could be considered flexible? And what magnitude of moment should be considered transferred to the column? (Note: the moment resistance of the column about weak axis is relatively small.)

Another thing is that If W250x115 is used, the web of the column is only 15.4mm, would it be too thin? For the column effective length, is it safe to take as 4.6 m (Floor height minus the beam depth) in this option?

There is another shear connection option, using double angle welded to column web and bolted to beam web. Compared to the above end plate connection, which is preferable?

 

[Agent666]

(Sorry for the late reply)

Yes, the torsion translates directly into moment in the slab.

When I say flexible I mean any end plate will still try take some moment, its just the moment developed is quite small in relation to the beam capacity. How thick it needs to be is up to your calculations. Most design procedures for flexible end plates look at transferring the shear and might have an upper limit on plate thickness, you usually don't specifically calculate the moment capacity, but you can work it out if you like to justify things to yourself. For example a 20mm end plate might transfer significantly more moment than an 8mm plate, and both might be suitable for transfer of the vertical shear.

Eccentricity would be very small in attaching to the web (half the web thickness).

Regarding 15.4mm web thickness, you would need to justify that it worked, I can't say off the top of my head. How much of the web is resisting the load and any applicable limit states might depend on the connection detailing and geometry.

I would imagine welded angles would be a lot stiffer, plus you cannot weld around both angles fully, it would behave more like a web plate/fin plate type connection I'd imagine with the bolt group in the beam web taking/developing any moment. Typically the angles would be bolted to the web of the column, not welded.

 

[sunyaer]

Thank you so much, Agent666.

For now, we have decided to use welding angles to column web and bolting them to beam web. The reason being is that, if welding the angles to beam web is used, the total length of the beam assemble with welded angles must be very accurate for easy erection and in order to not make the columns at both ends lean in or out for the beam assembly to fit in between the columns. The way of welding angles to columns and bolting to beam web might be a little easier since and holes on the beam web could be made a little larger horizontally.

In reality, the beam will deflect and the top and bottom flanges will be in a rotation at both ends, meaning that if the beam is bolted to the column web, the column web will get bent vertically. For this reason I think the column web is a bit thin for the bending and this is also another reason that makes it my second choice.

I attach the connection here for easy discussion. If any successful similar project experience can be referenced, I am still able to make some modification and revision to the connection detail.

 

[sunyaer]

I found that the space in the bolt group might be a bit tight for using a impact gun. Is there any socket extension that can get in to tight space to tighten bolts. The column is W250x131.

Hi Agent666, do you have any comments on my last and this post? Or any new finding for situations similar to this? Thanks.

 

[Agent666]

The issue I see with welded angles (which I noted previously wasn't a great solution) is how do you weld one of the angles between the angle legs. First one can be welded all round, but second one with just the gap of the web cannot be welded. If you look at the resulting weld group on the angle that can be fully welded around only one of the vertical runs is effective really for the moment, and on the second one its basically ineffective as a whole for moment, and a channel shape for the shear with the vertical leg located furthest form the point of load application. This is why these double angle type connections in most design models are usually bolted to the column web. You would be better having two separate fin plates, you can fillet weld both sides on one, and butt weld the other, it would be analysed as per standard means. If you were to do the welded angle detail then I'd make sure you have specifically looked at the loadpaths involved.

I have no issue with slotting the holes horizontally a little, I've done similar previously to create tolerance and in an effort to minimise any moment transfer. You still have the eccentricity from the bolts to the column center-line to deal with, i.e. design moment for the column will be eccentricity times shear, but at least you can in theory alleviate any additional moment developing due to the apparent fixity. You still need to design your column for this moment plus the axial load. If say the bolts are 75mm from the column center-line then 2500kN x 0.075 = 187.5kNm about the minor axis, no idea if your column can take that or not but it seems like a sizable load still depending on your actual bolt geometry. Just work out how much slotting is required based on the rotation under gravity or drift cases and provide plenty of clearance, be sure to provide sufficient gap between the end of the beam and the column to allow for the rotational movement.

Why do you need to use an impact wrench, for a simple 'pinned' connection you probably don't need tensioned bolts. Not sure what your code calls it, but in my region they call it snug tight, which is tightening all the bolts to bring the plys into contact, but not tensioning the bolts further to bring them into the bearing or friction state. Basically they note its the effort of a single man on a standard podge spanner.

 

[Agent666]

Even if you could weld it somehow, how do you get access to test it. An old engineer once told me, if you can't test it, weld might as well not be there!

 

[sunyaer]

Agent666, thanks for your input.

The moment magnitude about the minor axis you estimated is more or less what I considered. In Canadian code, when moment and axial load interact, there is also a 0.85 factor for the moment, so I was looking at about 120KN.m of moment about the minor axis, accounting for about 30% ratio in the axial load and moment interaction formula.

As for the weld, what I understand is that the weld on the angle needs to be only on the outside and a 2D return at the top and bottom. This may be the concept that is behind the shear connection, enabling the welded angles to bend a bit at ultimate load and reduce the moment transferred to the column.

I actually have no issue with the beam bolted to the web of column. The only concern is the web of the column is a bit thin, which may get bent. This is one of the reasons that I wanted to see experience from other engineers.

Snug tight is OK. for simple bolted connection, does the guy on site usually not use a tightening powered gun? I have never seen a podge spanner. Is there a picture that could be posted?