Out of Plumb Column

[bouk715]

A 3-story steel framed building that I designed is currently in construction. Lateral resistance for the structure are braced frames.

We were called out to the site because at one location, a column is out-of-plumb 3" within the first story. The column does not appear deflected over this length, just leaning to one side. The building is already framed and floor slabs are in place. The movement was noticed when a stud wall was being framed in front of the column. Architectural pilasters have been built around the columns on the second and third floors, so I can't tell if the column is deflected or out of plumb over those lengths. Since the pilasters were built, I'm assuming that the column is reasonably (within AISC tolerances) plumb on the upper stories. I've requested the steel survey to see if this is the case. We have the foundation survey, and the footing is in the correct location.

I've analyzed the column for the P-delta effect caused by the movement, and it still works. I just have additional lateral forces I need to transmit into the rest of the structure. However, I believe I need to determine possible residual stresses in the column caused by bending. My assumption is that the movement occurred during erection and before the floor deck/slabs were in place. I'm hoping that the survey will confirm that. What is the best way to determine these stresses? I've modeled the column with imposed deflections - is this an acceptable approach? Or should I specify some form of testing? Just wondering what the typical approach is to this problem (if there is one).

 

[kslee1000]

You need to wait for the survey. Actually it is not too difficult to DIY to get a first hand feeling. Just use the plumb bulb, or laser leveling device, both can get from local hardware stores.

I probably wouldn't leave a 3" displacement stays in place, unless you can tell the owner that it isn't a defect, and it might eventually comes vertical again, once the building frame moves in another direction.

 

[jike]

I wouldn't be surprised if there was one beam fabricated too long and one too short. I had a similar problem on an 11 story hotel.

As long as you can assure yourself that stresses are within allowables, it should be OK. I know of no testing that can be done.

Once the concrete is poured, it is unlikely that the column will ever be moved.

 

[BAretired]

Is it possible to reinforce the column in a way which would remove or reduce the eccentricity? If so, that is worth considering.

BA

 

[kslee1000]

Is this column stops below the beam, or it goes to the second floor? For the latter case with the column "Leaning", how could the second floor not affected? And how the column splice could be made without some fitting took place, if it had occurred at time of construction?

 

[jike]

Are the columns one piece or is there a field splice? There is generally some play in the splices and this is how an erector can pull the columns back into plumb.

Make sure the erector shims the column splices per AISC Fabrication, Erection and Quality Control, Section M4, "Fit of Column Compression Joints".

 

[bouk715]

Thanks for the replies, it's very helpful.

The column is a single piece. As far as how the upper columns might not be affected - it could be that they are slightly out of plumb (relative to 3" that is); however they are covered by a masonry pilaster, so determining how much the are out of plumb seems difficult - without removing the masonry obviously. With the floor slab already placed, removing the column is not an option. My thought is to reinforce the existing column if it is overstressed, as BAretired suggested.

My assumption is that the second floor beams tying into the column were fabricated incorrectly and that the column was pulled/pushed (with guys cables perhaps?) to connect the beams. So the column may just continue up reasonably straight (offset 3" from the proper location) on the upper floors. Or perhaps the column was "pulled" back slightly on these floors and is just not enough difference for the mason building the pilasters to notice. Just guessing since I wasn't out there when this piece was erected.

Again, I'm hoping that the steel survey will shed some light on the subject. My main concern is that if the column was adjusted during erection, there are inherent bending stresses remaining in the section that I need to add to my P-Delta and axial loads. At this point, I'm thinking that enforced displacements are the best way to model it and come up with the bending stresses.

 

[bouk715]

To All - Below is an answer that I received from AISC on this topic. Wasn't sure if anyone had an opinion on this. My question is first, followed by the response...

I understand that as part of the Direct Analysis Method, notional loads may be applied to a structural model to account for out-of-plumbness effects. However, how can one account for the stresses in the column due to out-of-plumbness? For instance, if a column is out of plumb for a single story, but is drawn plumb on higher stories during steel erection? It seems that there would be bending stresses present in the column; is there a point where these loads are negligible?


I think you’ve made an excellent observation. To my knowledge, the DAM does not explicitly account for residual stresses due to pulling the frame into a plumb position. I have not seen this mentioned in any of the papers on the DAM. However, consider that EI and EA are both reduced with a pretty substantial (and quite rounded-off!) 0.8 reduction factor that would seem to be enough to capture this effect.



Brad Davis, Ph.D., S.E.

Consultant

American Institute of Steel Construction

866.ASK.AISC

 

[lkjh345]

3" out of plumb?? Wow!! Hard to believe this was not noticed by the steel erector during erection. It would be hard not to to notice!!

What is the floor-to-floor height at this level?

I think your idea of using a forced displacement analysis is correct. My guess is the force required to displace a column 3" (assuming a 'norma' floor-to-floor height of something around 15 feet) is pretty substaintial, and will induce a considerable moment in you column member that a 20% reduction in I and A won't adequately account for. Just my hunch without knowing hte specifics of your situation.

 

[bouk715]

lkjh345 - floor to floor height is 14-feet. It is somewhat remarkable that the contractor did not notice it until the stud wall and pilasters were being built.

The column is an HSS6x6x1/4, not a massive member, but it would take substantial effort to move it 3". Resulting moment is about 12 kip-feet at the second floor level. That's about 40% of the capacity of the column. I did not list the actual out-of-plumbness in my question to AISC, so I'm guessing they're assuming it is within standard tolerance.

 

[lkjh345]

I'm thinking that you are correct in that AISC's answer assumed 'normal' tolerances. 3" in a 14 foot height is definately not normal.

Is it possible that the steel is fabricated correctly, and the anchor bolts were set 3" over from their proper location? Just a thought.

I think I would have a tough time living with leaving a column 3" out of plumb.

Is it possible to shore up the floors and roof adjacent the column, cut the base plate free of the anchor bolts, and the pull base plate over till the bottom of the column is aligned with the upper parts of the column? thus removing the induced moment. Just another thought.

 

[JKStruct]

I once had a column 3" out of plumb in a 10 ft flr/flr. They didn't notice until several stories were erected above. It was a lower column on an 14 story building... we made them "box in" the column section with plates welded to the ends of the flanges (parallel to the web). I'm still dumbfounded at how it happened, and to this day, I'm not happy about it.

 

[271828]

"I did not list the actual out-of-plumbness in my question to AISC, so I'm guessing they're assuming it is within standard tolerance. "

Gee, ya think? LOL. Talk about a misleading way to ask the question! I'm confident that AISC would've come back with a totally different answer if the actual out-of-plumbness had been known.

 

[bouk715]

271828 - Wasn't trying to be misleading; I wrote the question to them before we confirmed what the actual out-of-plumbness number was. I've re-submitted the question, so I'm anxious to see the answer.

lkjh345 - We have the foundation survey and the footing and anchor bolts are in the correct place (within a 1/2"). My concern with trying to move the column (beyond the shoring) is that the load will be eccentric to the footing, and the footing will have to retrofitted.

 

[bouk715]

A more complete answer from AISC, in light of the 3" out-of-plumbness measurement. Figured I would post it here in case anyone was interested/useful:

There's another question that I should've asked: Is this column part of
the lateral force resisting system? If not, then you won't need to
reduce the properties per the DAM. See the top line of Page 16.1-198.

If it's not part of the LFRS, then I think you're on the right track
already. Compute the moment that must have been applied to bend the
column into its current shape and add that to the applied moment (if one
exists) to get Mu for comparison with phi*Mn and for use in the Ch. H
interaction equations.

If it is part of the LFRS, then it's a bit more complicated. I'm sure
that no research exists to give us a solid answer, so we'll have to use
some judgment and mechanics. The idea behind the 0.8 factor is to
account for inelasticity due to bending and to capture "other" effects
that effectively reduce EI. (Tau_b accounts for inelasticity due to
axial load.) I have looked in (I think) all of the references on the
DAM and have never seen how much inelasticity due to bending is assumed
in the derivation of the 0.8 factor in Eq. A-7-2. It seems obvious,
though, that the developers of the method would not have considered such
a large initial bending stress, so it seems likely that the 0.8 factor
is too small for your application. I think it is reasonable to compute
a new factor in place of 0.8Tau_b and then bump it down a little to
account for those mysterious "other" effects.

Here's how I'd do it, following the Salmon & Johnson 5th Ed. Example
6.6.2. This is definitely not practical for everyday design use, but I
think it's probably justifiable in your case. Not knowing your loads
and section, I don't know how this calculation will turn out.

1. Assume a residual stress pattern in the flanges. S&J uses a residual
compressive stress of Fy/3 at the flange tips as shown in Fig. 6.6.6a,
which is a little higher than the residual stress assumed in Section F2
and F3--0.3Fy. The S&J residual stress pattern is a little on the safe
side, so I'd either use it or use 0.3Fy at the flange tips. Ignore the
residual stresses in the web, just as is done in the example.

2. Compute the applied stress at the flange tips, Pu/A + Mu/S using the
gross section properties. Use Pu as computed normally and Mu which
includes the moment that was used to pull the column plumb. If this
member is in a rigid frame and you have a model developed, run the model
with gross properties to get Mu to add to the moment to pull the column
plumb. Mu will change a little after you adjust the properties, but I
think going back and iterating is not justified.

3. Superimpose the residual stress with the applied flange stress and
see if the flange tips have a stress exceeding Fy. If the flange tip
stress does not exceed Fy, then use EI*=0.8*Tau_b*EI--this means that
the 0.8*Tau_b factor more than covers your situation for this member.
If the flange tip stress exceeds Fy, then compute how much of the flange
has a stress exceeding Fy, using the residual stress pattern in S&J Fig.
6.6.6a. Fig. 6.6.7 and the rest of the example show a somewhat similar
calculation.


4. Now pretend that the yielded part does not exist (it's on the flat
part of the stress-strain curve, so has no stiffness) and compute the
net A and Ix.

5. Compare the resulting EI to EI* from Eq. A-7-2 and use the smaller
one in your analysis. (Same with EA.) Technically, one should iterate
because the A and S used to compute Pu/A+Mu/S are not correct because of
the yielded flange tips. In a design situation, I would just stick with
the one iteration because I don't think the extra effort reflects the
precision with which we can predict the behavior.

There's also the issue of the force that was dumped into the diaphragm
to hold this column in its plumb position. If this force is of any
significant magnitude, then I'd include it as a lateral point load in
the model.

Please let me know if you want to discuss this more.

Brad Davis, Ph.D., S.E.
Consultant
American Institute of Steel Construction
866.ASK.AISC
kg/eng.

 

[ishvaaag]

If an analysis with the column at the actual position -even without the imposed deflection quality, which may be or not the case- shows the structure be valid, then only the transfer at the node remains the problem (I am oversimplifying), and reinforcement if any could be limited to its environment. In one column of this size maybe some stiffeners/doubleplaters may be enough. The advantage of this is that you need not care specially about member buckling differently from the usual ways, and you can concentrate just in the stresses surmised to be acting at the node.

 

[BAretired]

If I were the owner, I would expect a fix no matter what the numbers show and if I were the contractor, I would expect to pay for it.

The column is HSS 6x6x1/4. Why not use L 4x4x1/4 full height at each corner placed vertically and welded to the wall of the HSS so that the 3" slope is hidden from view? The resulting column would have an outer dimension of 6.5" x 9.5". No need to do any further calculations.

BA

 

[ishvaaag]

Certainly, BAretired. Too cheap a column and too clear a case to waste much brains.

 

[kslee1000]

Agree with BA. Retrofit may cause money, someone has to pay it now or later. As a buyer, one wouldn't accept goods with build-in defect, why the owner here? Unless this is an architectural column, the defect could get worse with time for matters unforseenable. (It really boggles my mind knowing there are two stories above resting on a column with such magnitude of tilt)

 

[271828]

If it is still strong and stiff enough, then why is it defective?