Flexible Moment Connection

[CJJS]

The connection detail I posted is being used as a FMC. How would you go about calculating the capcity of this connection to transmit gravity load moments from the beam? I am trying to justify that this connection is flexible enough that it will not transmit more moment than the moment due to seismic loading.

 

[ash060]

That doesn't look like a FMC. It looks like a WT shear connection with a erection seat. You could assume that the connection does not transfer any moment.

 

[Ron]

I disagree. This connection will transmit moment to the column as a force couple exists between the 4-bolt connection and the seat connection.

It is a relatively weak moment connection with the ultimate transfer capability defined by the lateral shear of the bolts or deformation in the web at the 4-bolt connection.

 

[CJJS]

Ron,
That is essentially how I determine its capacity to resist the seismic moment. This connection is being used as a FMC. So, it is assumed that this will yield under gravity loading. I am basically trying to justify that it will in fact yield under gravity loads. If it's too rigid, the columns will be underdesigned as they are only designed for the moment due to seismic loading only.

My thinking is that the top of the flat portion of the wing tee (that portion welded to the column) will yield first as it wants to pull away from the column (notice there is no weld along the top). I am looking for a simplified way to analyze that. Thoughts?

 

[Teguci]

I'm getting an allowable moment of 97 inK for the W14 and 73 inK for the W12 (bending in the 3/8" plate). The failure mechanism will be shear in the bolts with a maximum moment transfer after a great deflection (considering the overstrength factor) SWAG of around W14 - 228 inK, W12 - 171 inK.

 

[CJJS]

Teguci,
Once the 3/8" plate yields, how can it transmit any more moment? You say that the maximum moment is 228k-in (W14) and 171k-in (W12). How did you calculate these values?

 

[Teguci]

Allowable is based on Fy/Omega x Z, Ultimate for the plate is Fu x Z, where Fu can be 80 ksi.

 

[CJJS]

Are you assuming the plate is simply supported for bending? And are you assuming the entire plate is in bending or just a portion? Using 36ksi yield, I am calculating an allowable moment of 39 k-in (w12).

 

[Teguci]

fixed at the weld and no bending (infinite stiffness) over the 3/4" distance between the 3/8" stem plate and welds. or for half the plate, fixed - no rotation over 2.125" distance

 

[CJJS]

Thanks. That explains our differences. I assumed a pinned connection at the weld.

 

[connectegr]

This is not a moment connection. AWS/AISC does not permit a one-sided fillet weld in prying. Which is the condition if a tension force is applied to the WT web connection. The beam web connection can develop only shear. As shown the seat and beam web connection cannot share the shear load. In order to load the seat, there must be yielding of the beam web connection. And if the seat is in bearing, the web bolts may not be carrying any load.

Typically a flexible moment connection will provide angles at the top and bottom flange. Or flange plates (with controlled weld lengths) at the top and bottom. These connections provide flexibility due to the ductility of the connection material. Rotation of the heal of the angle or elongation of the flange plates.

Best professor I ever had, had a simple guide. "If it looks wrong, it probably is"

http://www.FerrellEngineering.com

 

[Ron]

The one premise that we are all overlooking is that the physical condition will prevail without regard to how we model the connection. No, it is not a "moment connection". Will it still transmit moment to the column? Of course it will...until the components fail...not just yield.

Backcalculate the moment that will be transmitted until component failure occurs and make sure the column can handle it.

We can assume the idealized conditions of pinned or fixed all we want....the actual restraint will dictate how the members respond.

 

[connectegr]

So does a clip angle connection or a shear tab. Nearly any shear connection lacks the theoretical flexibility we rely on for a pinned analysis. This does not allow us to reverse engineer a shear connection to find a little more fixity. This is OK for a doctorate research project. But, there is also valid research for the intended applications of AISC recommended shear connections.

http://www.FerrellEngineering.com

 

[CJJS]

Ron,
Yielding of the connection will cause rotation of the beam end. Once it rotates sufficiently it acts as a simply supported beam and therefore will not transmit any more moment.

 

[connectegr]

In the case of the detail provided, yielding of the WT flange will result in bending of the flange. Bending of the flange will result in stress concentrations at the root of the fillet welds. These stresses will result in a crack, propagation of which significantly reduces the tensile and shear strength of the weld.

http://www.FerrellEngineering.com

 

[southard2]

I'm not an expert in this subject but I'm think I agree with connectegr. Assuming the connection is only located in a few areas my advice would be to remove the seat and then add plates or angles as needed to get the design you need. At any rate by removing the seat you would know three things that would help make your problem simpler.

1) You would know that the connection is not transfering any moment, and all shear is going through the tee part of the connection.

2) You would know that you only have to design your moment plates for live loads in the structure whether they be floor live loads, roof live loads, or seismic live loads. All dead loads are all ready being transfered in simple shear. This helps you in not overstressing the columns. Hopefully your live moments are less than your seismic load moments.

3) You don't need to analyze the tee connection anymore as the moment will be handled by the added on plates.

Assuming you don't have a lot of P-delta issues in your frame analysis you could superimpose the dead load case onto the other cases and see if your columns work as if the connection were traditional moment connection. If that didn't work than you could size the plates or angles just right so that they would qualify as a flexible moment connection.

If by removing dead load moments your column overstressing problem is solved than you could probably get away with just removing the bolts from the seated connection and then replacing them. It would be like releasing the dead load moments from each end one at a time. Friction might prevent this though.

Living in Florida I don't have to deal with the seismic requirements thank God. I honestly don't know how you can make ends meet dealing with all the seismic provisions but that is another story. For me wind is enough. And I'm still not yet comfortable with the whole flexible moment connection idea to be honest. You would really have to do it just right and have really good inspectors in the field. In my practice I either assume its a moment connection or simple shear. Of course this doesn't help you situation out much.

I agree with other comments above also that we never really have a pure simple shear connection unless of course we one bolted everything. I always try to be conservative with my columns. This was a fun question to think about and I'm glad its not my problem. I've gotten to a point though where I'm tired of trying to do physics gymnastics trying to justify things I'm not really sure about. Hence my advice above. Do something that makes this problem simpler and lets you sleep at night.

John Southard, M.S., P.E.

http://www.pdhlibrary.com/

 

[BAretired]

Is the column a W Shape? One detail indicates the angles connected to the web. The other detail indicates the angles connected to the flanges.

Or is the column a HSS?



BA

 

[Teguci]

ConnectEgr - I tried finding the "no moment stress through fillet weld" statement in both AISC and AWS but couldn't find it (J2.2b.(2) seems to suggest that end returns aren't necessary?). Makes sense though. Is the attached diagram appropriate for what you are referring too?

In this case, the prying action of the plate should be avoided with the use of a returned fillet weld along the top (length = 2 W) similar to single angle connections (J2.2b). This will add to the bending stiffness of the connection but should remove the fillet weld as the failure mechanism.

 

[CJJS]

This is an HSS column.

This detail has been used in storage platform design often in seismic design categories C or less (R=3). The lateral loads are relatively small. I have reviewed calculations on these mezzanines from several different engineers. It seems that all of the engineers have designed the mezzanines similarly. The beams are designed as simply supported. The columns are designed for axial loads due to gravity loads and moments due to seismic loads. The connection is designed to take the moment due to seismic only; end moments from the beam are never considered. Usually, engineers check the shear capcity of the bolts and/or bolt bearing acting over the moment arm (top of seat connection to center of beam) as the moment capacity. If the connection is capable of transmitting moment due to seismic, logic tells me that it must also be able to transmit moment due to gravity loads. I understand the theory behind the flexible moment connection design, and the key to that working correctly is that the connections yield under gravity load, thus my whole concern of determining when the connection will yield.

Generally these platforms are designed for 125 psf LL, however, they rarely see anything close to this loading. I have never heard of a platform failing, although I'm am sure there are failures out there that we never hear about.

connectegr,
You stated that "AWS/AISC does not permit a one-sided fillet weld in prying." Can you point me to the refrence?

 

[connectegr]

See One-Sided Fillet Welds in AISC Chapter 8. The detail shown is a "T" joint and shows where cracks initiate at the root of the weld. A similar condition occurs due to prying. In some cases the prying condition can be worst, since initial the stress in concentrated at the top of the weld and not evenly distributed. If the flange of the WT is extremely stiff, then it can be argued that tension only is in the welds and no rotating is allowed at the root of the weld.

http://www.FerrellEngineering.com