Column Eccentricity 6

[ronstruc]

I would like to get some feedback from fellow structural engineers on what eccentricity is typically used for square tube steel columns (HSS's) that are concentrically loaded and are not part of a frame (resist gravity loads only). The columns would support steel beam(s) on one or both sides with equal or unequal spans. The steel beams are connected to the column with a through knife plate, so they are not connected to just one face of the column. The columns sizes would be anywhere from 4" to 10" square.
I typically use 1" eccentricity. I have heard others use 10% of the column size, (i.e. 6" col. would have 0.6" of eccentricity).

Your thoughts?
Thank you,

 

[ronstruc]

I should clarify that the column ends would be pin connected, and the column may be one level, or multiple levels. If it was multiple levels it would be continuous through those levels.

 

[spats]

It doesn't matter if you are using a thru-plate or shear tabs welded to the face- the eccentricity is still the distance from the center of the connecting bolts to the center of the column.

 

[WillisV]

I typically do not use any eccentricity at all in those types of conditions.

 

[lkjh345]

From what you are describing, a knife plate through a steel tube column to connect beams, I would typically not use an eccentricity to my load. For me, e=0 if I am reading this correctly.

I don't add an eccentricity to a load when I connect a beam to the flange of a W-section, so in the same manner, I do not an eccentricity to the load when I have a simple shear connection to the face of a tube column.

JMHO.

 

[msquared48]

This is no different than using a knife plate to connect to a WF beam. Unless the load is high, I typically ignore the moment induced.

Mike McCann
MMC Engineering

 

[apsix]

By AS4100 (Aus) "beam reaction on a column shall be taken as acting at a minimum of 100mm from the face of column......except that for a column cap it shall be taken at the face..."

 

[asixth]

I would assume the beam reaction only occurs on one side of the column and would take the eccentricity as the distance from the bolt to the column centreline.

Apsix makes a valid point that the beam reaction should be taken as a minimum of 100mm (4") from the face of the column.

I assume the 1" you mention is from the column face and not the column centreline.

 

[ronster]

For the condition that you describe, I do not design the column for any moment as the beam on the opposite side will resist any induced moment that may occur. Any moment in the column would be negligable. I will use a 3" eccentricity from the face of the column when a beam is framed into only one side of the column. I will usually specify for a beam to sit on top of a tube or pipe column for a one sided condition.

 

[csd72]

I disagree with the notion that a continuous beam across the top of a column will reduce the eccentricity to neglible. If the beam is loaded on one side only then the beam will actually be sloped at the point of the column causing it to bear on one side of the cap pl.

Similar applies to single span on the top.

For this reason I would never take the eccentricity any less than half the width of the column.

Also what about eccentricity perpendicular to span, assuming zero ecentricity gives no room for erection tolerance.

 

[spats]

I don't understand why most of you guys think there is no eccentricity. With all due respect, you're wrong!... and how does the magnitude of the load dictate whether or not there is eccentricity! The only way you could claim no eccentricity in the column, is to design the connection for the eccentricity/moment.

 

[LowLax]

I think historically these types of connections were treated as having zero eccentricity. I know the more experienced engineers around here do it that way. If you don't have a computer it's much easier to assume no eccentricity and pick the size out of a table. I'm not saying that's the right way to do it.

I usually ignore any eccentricity on a column if beams frame into it in two opposite directions. If a load on one side was much higher than the other side I might consider eccentricity. If I have a one sided connection I use half the column depth plus 2 inches (approximate center of bolts) for the eccentricity.

 

[DaveAtkins]

spats,

Be careful about saying others are wrong--there are usually various ways to look at structural engineering issues!

If the bolts in a single plate connection are properly tightened, the row of bolts can take shear AND moment--and the point of inflection moves toward the column. In fact, considering the flexibility of the column wall compared to the stiffness of the row of bolts, the point of zero moment will be at the column wall. Therefore, the eccentricity is half the column depth, not the distance to the line of bolts.

DaveAtkins

 

[spats]

Sorry for the minor blow-up, DaveAtkins. Sometimes I get a little perturbed when I see responses to a question that I consider bad advice.

The load is applied to the plate at the location of the bolts. The plate is fixed to the column, so there's definitely a moment in the plate, and thus a moment in the column. This moment is rarely balanced by the beam on the opposite side... and it's appropriate to skip live load.

The point I made earlier was that you can design the beam connection to the plate to take moment as well, and this will reduce the moment in the plate & column. Think in terms of the plate being a beam, fixed at one end (column), and free to deflect but not rotate at the other end (bolts), with a concentrated load at the deflected end. Per standard beam diagram & formulas from the steel code book, the moment is PL/2 at each end.

 

[ronster]

Dave once again you get it and others just get perturbed. To look at a connection as a cantilever is just not correct. There is a beam attached to the connection that is not free to rotate, it is supported at its other end and there for the joint rotation is limited by the beam stiffness. The stiffness of the bolted connection, as Dave correctly pointed out, can easily tranfer the moment accross the joint to the face of the column. I would further argue that when you have a connection on the opposite side of the column that the same bolted connection will transfer the majority of any unbalanced moment to the beams rather than to the flexable tube column. I also do not agree with csd72 statement that the load from a beam over the top of the column will be transfered through the outside edge of the cap plate and by his reasoning would form a moment connection with the opposite bolts being in tension. The beam will have plenty of stiffness to remain in full bearing with the cap plate. Once again the beams are much stiffer than the tube column and the actual moment transfered to the column would be minimal. That being said the column moment will increase if you use large tube shapes for columns but the majority of structures that I design, the columns are typically 4 or 5 inch tubes and much less stiff than the beams. I will also use an eccentricity of 1" to account to account for these small induced moments and erection tolerances.

 

[DaveAtkins]

The way I look at it is that the single plate connection becomes part of the beam when bolted snugly to the beam. That is why a single plate connection can be used to connect a beam to a girder on the edge of a building without causing any torsion in the girder.

DaveAtkins

 

[spats]

Ronster, you seem to want to assume that everything is fixed, like it's a rigid frame, but do you design it that way? Do you design the connection for moment? Sounds to me like a case of rationalizing how it works, and then making completely different assumptions for the design. Assuming the beam is stiffer than the column doesn't mean the column is not being bent. You want your cake and eat it too!

 

[w16x26]

I have never designed a column for eccentricity if the connection is a "Simple Shear Connection" such as a single plate. In the 13th Edition of the AISC manual there is a statement on page 10-103 (bottom paragraph)"The design procedure for extended single-plate shear connections permits the column to be designed for an axial force without eccentricity."

 

[csd72]

ronster, it is the stiffness that is the problem.

No I am not saying that the continuous beam will form a moment connection, the plates cantilever out to the bolts and will therefore be more flexible than direct bearing.

What I am saying is that if the beam is at a slope (from uneven load flexure) and the cap plate is horizontal then there will be only one point of contact until the column flexes to even out the bearing. But the column has to flex to achieve this and that is the reason for minimum eccentricity.

Dave, That assumption is not strictly true. The single plate shear connection relies on the ability of the plate to rotate to match the beam end rotation (or the bolts to plough through the plate). Either action will induce a moment in the column. The reason why an edge girder works is because its torsional stiffness is very low and it easily rotates.

 

[ronster]

Yummm cake, I admit that I do eat my cake when I have it. Yes I I do make rationalizations but I do not change them to suite me. The connection is assumed pinned but with enough stiffnes to transfer the shear accross a 3" plate. Is this an absolute truth? Of course not but it is close enough to what is happening, to code, and I can seal my drawing and sleep at night. There is no way to completely simulate the actual amount of fixity in each connection nor would anyone want to. I feel your pain as I argued for the same point as you are to 2 partners in the consulting firm I worked at 28 years ago(ouch). Engineering is about making judgements to make models and formulas fit the real world. I am not wrong, just not as conservative as you. I am doing it the way most engineers in the US have been designing for centuries. I hope that someday you can eat your cake to. End of story.