Designing Columns for Concrete Buildings

Stazz,

1) Yes, the standard assumption is that you have to design the columns for all load effects on them. This includes vertical load effects and horizontal load effects, which should be significant in 1 15 storey concrete frame building, resulting in axial forces and moments on each column. You cannot design them as compression only!

2) I am not sure what fixed-fixed beams are. I hope you are only modelling the fixity that actually exists and can be generated by the frame!
You cannot deliberately over stiffen the connection to make the slab design work better. You should be analysing the frame as it will act in practice. Some designers reduce column stiffness for the frame design (in Asia they quite often ignore the columns completely in the slab/beam analysis) but that is dangerous as the column will attract moment and failures such as punching shear must be checked based on the real frame stiffnesses as redistribution is not possible. Columns in this case still need to be designed for the loads and moments that they attract.

Why do cantilevers make it difficult. I do not understand the NEED to design these quickly. You need to design them properly!

I hope you are doing this design under the supervision of an experienced concrete building designer!

What i think rapt is trying to say is "The consequences of overestimating or underestimating the actual stiffness’s of structural members depend on the type of structural system and the response parameter of interest” in your case, underestimating the column stiffness in a flat slab can cause problems with punching shear, however over estimating you column stiffness in a moment frame will lead to larger than expected deflections and cracking.

There are many ways to estimate this stiffness for columns, because of their compression bending interaction they are not the same as beams and slabs, thus you need to do a bit more home work. for a first run situation I normally use 0.7Ig as a starting point to size your columns then use have to use a empirical approach, there was a good discussion in a paper in the July 2009 ACI journal, by Kenneth. There is a good publication by Rice and hoffman that could help as well.

As for cantilevers these are something that you want to spend time on getting the stiffness correct and making sure they are more than strong enough as they have no redundant system.


When in doubt, just take the next small step.

Don't forget to design the columns for lateral loads, even if the lateral loads are taken out by shear walls, you will still need to check the moments that are put into the columns from lateral displacement.

1) Always consider the moment that is being transferred into the column from the slab. Also because the moment capacity of the column is dependent on how much compressive stress the column is subject to, look at a maximum and minimum compressive load envelope. The greatest moment in the column may occur for a patterned or checkered live load case which will reduce the axial stress and may have a reduced capacity depending on where the load points lie on the interaction diagram.

2) Are you referring to fixed-ended columns at the floor levels above and below the floor you are designing? Or are you referring to fixed-end moments used in frame design by moment distribution/redistribution?

3) To be on the conservative side, when designing columns design them with full stiffness. i.e. Icol=Igross. Some codes allow for the analysis to be conducted using 0.8*Ic. This allows for some stress to be redistributed because of the loss in stiffness that occurs when a section cracks.

Cantilevers can be designed quite easily, they are determinate structures, the moment cannot be redistributed. A spreadsheet can be set-up quite quickly to design overhangs.

Does risa analyse equivalent frames or meshes the slab and designs by FE?

When designing a concrete structure, in addition to gravity and lateral loads, you need to design for creep, shrinkage, temperature and, in your case, post-tensioning forces. The forces caused by restraining these strains is significant, particularly for the columns.

Let me guess, the old we just need foundation plans to get started catch 22.

There is no easy button for multi-story concrete columns but I've found Etabs does a nice job. Although I don't always agree with how it determines sway vs non-sway. Still a lengthy, input and multiple runs process to economize and back check.

Rowingengineer,

In a 15 storey building, under vertical loads, the columns at the lower floors will not crack as the axial compression effect will be far greater than the moment effect and the column will be in compression. Why are you using .7Ig if the columns cannot crack. You can only allow for reduced stiffness if there is going to be cracking.

The article you are talking about in ACI is discussing stiffness for sway cases as I understand it. You cannot use this logic for vertical effects.

So in other words, I should get another career?

CRAPT,

I want to design them quickly because this is just a preliminary design for estimating purposes for a design-build project. I need to get column sizes to the architect so they can lay out their guest rooms. In Risa the beams are fully fixed to the columns. I modeled the beams as the depth of the slab x the tributary width wide, then the program will calculate the correct Ig of the beams which will produce the correct stiffness in the equivalent frame analysis. The building has a high aspect ratio so I figured that the critical direction to analyze the columns was in the short direction where the column spacing’s are not the same, so I ran the floor slabs as one way so that the load went directly to the beams first and then to the columns. This is how the equivalent frames in the short direction are analyzed anyway so I figure that I can pull the end moments from the beams out of Risa and use these for the Post Tension slab ultimate strength design.

ROWENGINEER,

I’m using the full Ig for the column stiffness. I think this is conservative for the column design because this pulls the maximum moment out of slab and into the column. But for the slab design I would think that you do the opposite -> use the weakest column (0.7Ig) or maybe even no column stiffness so that all the moments stay in the slab so that there is adequate steel in the slab.

ASixth,

I’m referring to fix ended beams. There is no FE plate in the Risa slab, I’m hoping that these beams I’m modeling serve as column strips to collect the load and have a stiffness for the equivalent frame analysis. I have a spreadsheet now that I’m using for the PT slab design (since there are so many stress checks) and I’m pulling the end moments from these beam strips out of Risa to design for. For the lateral loads I might just model an entire 15 story frame and push the frame 3” to the right (h/500) and see what moments develop. What load case would I use for this? 1.2+1.6(moments developed from displacement)?

Spats,

To design for creep, you multiply the deflections by the factor in ACI right? To design for temperature, we limit the length of the building? TO design for shrinkage, we limit the length of the pours and use pour strips? But to account for the PT forces in the columns, aren’t these good for column? Don’t they relieve the dead load moments that the equivalent frame calculates? Because isn’t the PT force effectively a negative load that counteracts the dead load?

CTSENG,

If I prove that these gravity columns are non- sway, then do I have to analyze them for lateral loads?

If I prove that these gravity columns are non- sway, then do I have to analyze them for lateral loads?

Click to expand...

Yes, you still must include all applicable load combinations in the design of the columns. The sway vs. non-sway criteria affects how you magnify the column moments.

Now in a non-sway frame, moments from the lateral forces [red]might[/red] be low, but that depends upon the way the structure is braced, how combinations of alternating live loads and lateral loads interact, etc. But your lateral forces will possibly produce some moments in the columns and they would simply be included in the M[sub]2[/sub] values of the non-sway derivation of the [δ][sub]ns[/sub] values in ACI section 10.12.3.

Stazz,

You keep referring to Fixed Ended beams. That is mixing us up. A fixed ended beam has full fixity, which you do not have. I assume your beams have a full moment connection to the column and the frame analysis sorts out where the moments go. If they were fixed ended, the moment would be wl^2/12 at each end and wl^2/24 at mid span and would be very unconservative.

Agree with JAE that you need to allow for sway moments no matter what the framing system. In a braced building there is still sway and some sway moments and axial forces go into the columns.

You have to look at all of the possible combinations of wind, earthquake, and vertical load effects that the your loading/design code requires.

For the sway case, it is usual to use reduced column and beam stiffnesses as the members will usually be cracked and will loose stiffness, resulting in larger sway effects and more cracking.

PS why should you get another career?

Stazz,

Creep, shrinkage, temerature & post-tensioning forces all represent VOLUME CHANGES in the structure. Volume changes can significantly affect the stresses in your concrete frames... more so with taller buildings. Temperature is more important in exposed structures such as a CIP parking garage than it is in your instance. Temerature differences between the upper and lower surface of a beam/slab are very important when analyzing a garage, as are changes in member lengths.

You design for volume changes by including them as a load condition in your analysis program. I'm not familiar with RISA. I do this type of analysis using STAAD. With STAAD, you can apply temperature loads, pre/post stress loads, as well as strain loads (creep & shrinkage).

To obtain guideance on what kind of strains to figure for creep & shrinkage, I would reference ACI publications, and the ACI 209 Committee.

Hey Stazz,

I wanted to comment on your statement:

"But to account for the PT forces in the columns, aren't these good for column? Don't they relieve the dead load moments that the equivalent frame calculates? Because isn't the PT force effectively a negative load that counteracts the dead load?"

The PT "lift" really doesn't help the columns. The total load to be resisted by all of the columns in floor should remain the same. At best you might simply end up shifting some load from one column to another.

When you stress a PT floor slab, it tends to shrink axially. The specifics of this depend on tendon layout, shear wall distribution, etc. This shrinkage can result in large shear loads being applied to the tops of your columns as they go along for the ride. I believe that this is what Spats was referring to.

A colleague of mine designed a building similar to yours (mid rise, PT slab as lateral frame). We talked about it often. In retrospect, I'm not sure that I care for the system. Some of the reasons why:

1) An efficient PT slab is pretty thin. It doesn't do a whole lot for the frame stiffness. Especially when you consider the equivalent column business, which you should. Yeah, I realize that the prestressing stiffens the slab some.

2) We had no drop panels. As such the moments caused by lateral loads really exacerbated our punching shear problems. In general, I question the ability of these joints to accomplish the shear & moment transfer required of them when used in frames.

3) There isn't all that much mild reinforcement in a typical PT floor slab. In my mind, this seriously limits ductility and opportunities for load redistribution.

I try to steer PT floor slab buildings towards using shear walls for lateral resistance. In a high seismic area, that's the only way that I'd use PT floors.

KK

Rapt,
Thanks for the pickup.
I use 0.7Ig for my first run in Finite element modelling, just so I can have one model for the first estimate PRELIMINARY DESIGN (assumption was due to the need for quick and dirty this was a preliminary sizing design, not as it turns out a complete reo design). I wouldn’t normally use a finite element package to perform this task myself unless the building is complex and has a few transfer floors, with cokmplex column locations. I have noticed however that a lot of colleges, whom are heavy finite element users, prefer to do a complete model for preliminary design.

The reason for the 0.7Ig is for horizontal loads, I had assumed (probably wrong) this being a PT building, with thin slabs (180) that this was going to a quick column slab operation building with a small core. Here you want the columns to do a bit of work for you with regards to horizontal loads normally all I care about in preliminary design is getting serviceability criteria correct, because undersize members in relation to service requirements are very hard to fix during detailed design, however fixing strength problems is relatively easy.

However as the OP is doing a reo design no sizing design, I would suggest that he ignore my earlier advice, which he seems to be doing.


When in doubt, just take the next small step.

You dont need to analyze columns for lateral forces, UBC, ASCE and ACI allows the exemption of gravity only members from deformation compatablity due to lateral loads in low-moderate seismic categories. For high seismic categories you have to do certain checks in order to avoid deformation compatablity
In simple if you are in SDC A, B, C you can design your columns for gravity only, please do check the moments cranked in by slab. In most cases if you provide 1.5-2% reinf. you will be ok.

Kootenaykid,

I'm checking the punching shear of the slab now and there are incredably high moments getting sucked into the columns from the slab. This is killing the punching shear because of the unbalanced moments. I went back and analyzed the unbalanced moments again by considering that the PT force releives these unbalanced moments so in my model of the equivalent frame I simply superimposed the wb (effective uplift from PT) to the LL and DL in order to get unbalanced moment. Is this ok?

Rapt, Rowingengineer,

Sorry for the confusion, Fixed-Fixed beams refer to that the element is fully fixed at both ends in the model and the FE analysis sorts out where the moments go. A question reguarding Ig and Ie; aren't these close to the same? I always figured that if you use the min required flexure reinforcement 3 sqrt(f'c)bwd/fy then they should be close to the same since this requirement ensures the same ductility and strain compatability as a linear gross concrete cross section. If anything Ie should be higher. This is a separate study. I should get another career because this is rediculous.

StrucEng,

Thanks for the reference that will save me.


General,

In order to get the forces to design the slab and columns, will this procedure work (The procedure I have already implemented)...

1) I modeled a full 15 story section of the equivalent frame. 4 Column lines and 3 spans x 15 stories

2) Column sizes are as needed and beams between columns are slab thickness x L2 so that the propper Ig is used for stiffness.

3) If I run the DL+LL case can I use these moments to design the slab? + Reduced to face of support

4) To design the columns can I reduce the proper live load of each floor (Typically 40%LL for all floors) and use these axial forces and moments to design the columns.

5) Would it be wrong to reduce these moment further by allowing the wb (dead load balancing moments from PT)to relieve the moments being sucked into the column?

6) To get the unbalanced moments for the punching shear checks I'm simply subtracting the end moments from adjacent beams (which should give me the moment going into the column). Is this what the unbalanced moment is? And will thsi work? Can I let the PT balancing forces to releive these moments like I suggested in 5?

Also, requarding the shrinking of the PT floor and the effect this has on the columns...I read that the floor only shrinks 3/4" per 100' of floor. My one pull in the short direction is only 54' so the slab will shrink 3/8" total which will pull the exterior columns inward 3/16" each which is a deflection of L/640 which doesn't seem like anything to cry about?

It still seems like the PT forces will help the columns out.

Stazz,

You can try to rationalize away volume changes if you like. It only makes you terribly wrong.

Stazz,

You're right, the PT balancing forces will definitely help the columns out with respect to moment and therefore punching shear as well. The point that I was trying to make is that the PT won't significantly alter column axial loads. Sorry if I confused matters with my comment.

KK

Can I just confirm we are on the same page, Ig is the moment of inertia of the gross concrete cross section about the axis of consideration, while Ieff = effective moment of inertia for the cross-section.





Arguing with an engineer is like wrestling with a pig in mud. After a while you realize that them like it

Ok, that might be where I'm mixed up. ACI states that Ig = Moment of Inertia of gross cross-section WITHOUT considering reinforcement while Ieff = Effective Moment of Inertia the beam considering cracking. This is how it's defined in ACI.

To rephrase: I'm trying to say that Ig should approximately equal Ieff.