Columns @ Roof

[slickdeals]

What is the standard approach when detailing columns (which have unbalanced moments) and stop at the roof of a flat slab? For example, I have a 30' span next to a 18' span and the unbalanced moment is being resisted by the column. However, due to the slab being fairly thin, I cannot realistically develop any bars larger than #6/#7 in the slab.

Is it appropriate to assume that this column is "cracked" or to be a pinned joint for column design? The slab will be designed assuming unbalanced moments from the slab are being transferred into the column to preclude punching shear failure prior to redistribution. Also additional bottom reinforcement will be provided in the span to account for moment redistribution.

Would like to hear your thoughts.

 

[asixth]

For the slab design I would only assume a very notional stiffness of the column, probably not pinned but I wouldn't use the full stiffness.

For the column design I would have standard cogs to get at least 50% reo development and design the column based on this.

 

[KootK]

I have a bunch of thoughts on this, although no definitive answers:

1) Even if you could develop the vertical column bars that you'd like to, that wouldn't secure moment transfer. It would only guarantee that the joint failure mode would not be rebar pullout. To truly get the job done, you'd need a comprehensive joint design.

2) For better or worse, we have an almost identical problem when the column is constructed of concrete: incomplete rebar development, a joint insufficient for moment transfer, and rampant uncertainty about what moments need to be designed for.

3) For punching shear, I think that it's prudent to treat the situation like a column base plate attached to a footing. Use a punching shear perimeter located about 1/2 way between the edge of the cap plate and the steel section.

4) I know of no "standard" way of handling this. In new construction, I've only seen this detail for small canopy slabs supported on steel columns. There, I always see a cap plate with nelson studs. In retrofit, I've only seen this detail for slab reinforcement. Essentially permanent shoring. There, I always see a cap plate with post installed anchors. While I can't say for certain, the detailing I see in both situations suggests that designers are treating the column/slab joint as pinned for all aspects of design.

5) With insufficient bar/anchor development, I see things unfolding like this: a) slab tries to rotate b) tension rebar pulls out or fasteners breakout c) slab becomes truly pinned for flexure d) slab "hangs up" over compression side of column with slab lifting off on tension side.

6) Point number five above makes for an awkward punching shear situation. I would be tempted to modify the punching shear checks to reflect a bottom side punching area that is a line representing the compression face of the column rather than an area representing the entire cap plate. This may cause problems.

7) I would prefer nelson studs in this situation over welded vertical rebar. Nelson studs will stand a much better chance of being able to reach yield in service.

8) I think that the flexural design of the slab should proceed assuming that the column is 100% pinned. This doesn't take much and is one of the few parts of the design that you can make truly conservative without causing much fuss.

9) The only way to be truly conservative in your column design is to assume full moment transfer between the slab and the column. And this should include a reduction in slab stiffness of the "effective column" variety (fictional torsion members etc). This is an extreme position but I don't see how anything less could be reliably justified unless you're confident in your ability to yield your rebar (with over-strength). My gut tells me that a more realistic answer would be a column moment generated by the axial load at an eccentricity equal to 1/2 the column width.

10) I have some wacky ideas for detailing. They're likely to get you beat onsite.

10A) You could create a true hinge in one direction. Embed a plate with nelson studs into the slab. Weld on a couple of vertical plates with holes for a steel pin. Connect the column to that. While a two-way pin would be much more complex, it would be possible.

10B) Another approximation of a true pin. Weld a shear lug onto the cap plate and have that be the only connection mechanism. With some wrapping and a little luck, hopefully it would transfer shear but little moment. The two-way solution here is fairly simple: cruciform shear lugs.

10C) Place stud rails running over and through the columns in both directions and design them such that Vc is not relied upon at all. This way, come hell or high water, at least there isn't likely to be a punching shear failure.







The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[KootK]

10D) Roof buildup permitting, create a thru-bolt condition. Use anchor rods with double nuts on the bottom and a plate bolted over top of the slab. This is probably the best prospect for a true moment connection.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[Ingenuity]

KootK,

I think the OP is talking about RC not steel columns. "Slickdeals" - please confirm!

 

[KootK]

You're absolutely right Ingenuity. Upon reread, I'm not sure why I ever thought otherwise. Pin the slab. Fix the column and punching shear. Don't skimp on integrity steel.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[slickdeals]

Yes, concrete column.

 

[KootK]

So... column vert hooks turned in or out? In is technically correct; out is constructibility correct.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[slickdeals]

Out for ease of construction. I don't want to kid myself and provide reinforcement in the column that I cannot develop in the slab. I am sure others would have dealt with a similar situation. I am looking for industry practices based advise.

 

[KootK]

Well that makes it easy. Industry standard is hooks out. And that means that, analytically, you should probably disregard the hooks altogether. There's a very good chance that:

A) The hook will be too low and therefore almost worthless.
B) The hook will be too high and will need to be cut off or mangled.
c) The strut shooting out from the hook cannot be contained by the concrete.

For these reasons, I sometimes just omit the hook altogether.

Exterior columns exacerbate this issue.



The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[slickdeals]

In essence:

1. Design the column assuming it's pinned at the slab-column joint.
2. Design the slab assuming 50-70% fixity (0.5-0.7 Ig) and provide stud rails accordingly.
3. Provide bottom reinforcement in the longer span to allow for moment re-distribution.

 

[KootK]

Not me.

1) Design column and punching shear assuming full fixity. You'll have it until you've overcome the prestress effect from axial load. After that, you'll have a tough time knowing where in the load history fixity ends.

2) Design slab for flexure assuming no fixity. No need for extra redistribution steel if you do that.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.

 

[slickdeals]

I have a roof column, not sure if you misread that.

Designing the column assuming full fixity will generate moments for which the column rebar cannot be developed in the slab. A roof column should have very low axial loads and hence, very little beneficial effect.

 

[KootK]

I did catch the roof column part Slick.

In this context, I wasn't thinking of the axial loads having a beneficial effect. Rather, I was thinking of them as having a detrimental effect. They make it such that you can't avoid moments in your columns and at the punching shear interface.

I see things playing out like this:

1) Until axial load prestress effects are overcome, you have perfect moment connection. Column and punching shear designs must both consider this moment.
2) After prestress is overcome but before the rebar/anchors fail, you have have a moment connection with less than perfect efficiency. Column and punching shear designs must both consider this moment.
3) After the rebar/anchors fail, you have no moment connection at all. You have a slab that must be designed without the benefit of moment transfer to the column. You also have a column that must be designed for a moment equal to the axial load multiplied by half the width of the column.

For these reasons, I feel that a percent reduction in column stiffness makes little sense. It's a binary situation. You start off with a perfect moment connection and finish with no moment connection at all. The only exception would be if you can guarantee a ductile failure in the connection. In that case, the column and punching shear connection would both be appropriately designed for the yield moment + over-strength.

I realize that you're seeking practical advice here regarding what engineers actually do in this situation. In my experience, what they actually do is turn their hooks outward, assume the joint to be fixed in perpetuity, and pay no heed to all of this theory. Of course, if we leave it at that, then there really isn't much to discuss.

The greatest trick that bond stress ever pulled was convincing the world it didn't exist.