Compression Steel in a T-beam 1

[Lion06]

The post is a bit of a misnomer, it is a concrete T-section, but behaves rectangular. That being said, I need a decent amount of compression steel to control long term deflections. I can't find any guidance in ACI regarding placement of compression steel in a flanged section. It gives guidance on how to place negative moment reinforcement in a flanged section, but not for compression steel.
What I ended up doing is spacing the compression steel equally along the flange. Does this seem reasonable when it is really only to control deflections?
The reason for the compression steel (which was a reasonably significant amount) is that I was very limited in depth and had a significant load on the member approximately 1.5 k/ft service over a 28' span being limited to a 18" deep section.
Deflection was getting killed, especially long term, so I needed to add some compression steel to get Icr up (and, in turn, Ie up) and also reduce the long term multiplier.

 

[JAE]

Sounds logical to me to disperse the steel as you did.

 

[hokie66]

I would place compression steel either in the beam width or as close thereto as possible. If you place it too far away from the beam stem, you can lose part of the effectiveness due to differential shrinkage of the beam and slab.

 

[asixth]

StructuralEIT,

Just a comment about this long-term mutiplier that you mentioned.

In the Australian code (AS3600), we have a similar long-term multiplier that is used to estimate the effects of creep and shrinkage (we call it the kcs term).

This term is used to calculate the long-term deflection by multiplying the immediate deflection. The term itself decreases proportionally as the ratio of compression steel to tensile steel increases.

There has been alot of critism about the use of this approach in the Australian industry because this term doesn't fully take into account the main contributors to long-term deflection such as the actual creep and shrinkage properties of the concrete.

My understanding is that this approach came from a paper written by Branson titled, 'Instantaneous and time-dependent deflections of simple and continuous reinforced concrete beams', Bureau of Public Roads, Alabama.

This research published by Branson was based on tests done for heavily reinforced concrete girders, hence the use of the compression steel to tension steel ratio to estimate the long-term behaviour of the member.

I believe the re-write of AS3600 which is to be out soon (hopefully) will actually scrap this use of a long-term multiplier.

As for the distribution of steel across the flange, I agree with with JAE, I guess you should make sure the steel is within the effective flange area so it is effective.

 

[csd72]

You want to put the compression steel inside the shear links so that it is restrained against buckling. If you put it in the flanges then it is technically unrestrained against buckling.

 

[Lion06]

csd,

I considered that. I assumed since I wasn't using it for strength at fy and only for control of deflections that it wasn't critical to restrain it against buckling.
I didn't use it for phiMn calcs, but did use it for Icr and Ie calcs, but Icr and Ie are at service loads, not ultimate.

 

[csd72]

Yes it is fiddling the numbers a bit, but I see your point.

That steel will be in compression though and as it is stiffer than the concrete it will tend to take the load first.

I would not be concerened about the distribution of compression steel as much as tension steel.

If there is no compression steel in a certain region it will not result in cracking and I find it hard to believe that it will fail by concrete stress rather than take that stress up in the much stiffer reinforcement.

Tension steel is a different story. Regions in tension without reinforcement will crack.

I would put it inside the shear links and be done with it.

 

[hokie66]

When using compression steel to limit long term deflection, you are trying to make the concrete beam act like a steel beam, so it makes sense to me to place the compression steel as compactly over the web as possible.

 

[JAE]

Now that the point has been made that the compression steel requires confinement (I knew that but typed out too soon)...I'd agree that keeping it within the ties is best.

 

[Lion06]

The point is well taken, however, the question that I immediately have is this:
If you keep all of the "compression" reinforcement in the stem won't the flanges (which extend 28" to each side of the stem) undergoe excessive creep/shrinkage causing not only larger long term deflections than expected, but also some differential creep/shrinkage?

Additionally, I took this into consideration: The section is MORE than adequately designed for strength such that at the nominal moment capacity of the beam the compression bars will truly be in compression, BUT at the factored moment applied to the beam, the "compression" bars are below the neutral axis....just barely, but still below - and even if they were right at or just above, they wouldn't be taking much load. They are stressed nowhere near fy. They are not truly in the compression zone, but are in the flanges such that I would expect them to act as "compression" steel for purposes of helping to restrain creep/shrinkage, but are not truly taking compressive load even at the factored moment the beam will be subjected to.
I suppose it is possible that the creep/shrinkage of the concrete will induce some compressive stresses in the top steel, but I don't see that being close to stressing the bars to the point of causing buckling.

 

[hokie66]

The concrete in the flanges will shrink more than the concrete in the beam, thus it will crack, making it less effective in resisting compressive force. This is true regardless of the amount of reinforcing in the flange, as the reinforcing controls the width and number of cracks, not the gross amount of cracking.

 

[Lion06]

Fair enough.
That being said, would you still consider the T-beam MOI, when calc'ing deflections?

 

[hokie66]

Personally, I rarely try to calculate deflections in concrete structures. The answers obtained depend on so many variables that I get no comfort from them. Instead, I rely on deemed to comply span to depth ratios for the most part. When Rapt tells me I have a deflection problem, I look harder at my depth or add some post-tensioning.

To answer your question, the flange would contribute a bit at midspan, but not at continuous supports. Have you considered posttensioning this beam?

 

[rapt]

Asixth,

Unfortunately kcs is still in the next AS3600 in the approximate design methods. Fortunately, most people in Australia use more accurate methods.

StructuralEIT,
Can you explain your first comment that it is a Tbeam that is acting like a rectangular beam? It is one or the other! You are saying it acts like a rectangular beam then wanting to put the reinforcement in the flange! Make up your mind.

In your later post, what factored moment are you talking about? Factored for deflections as strength has already been considered?

I would
1 calculate it properly, not using multipliers as they are very misleading, especially when you start adding compression steel but even the starting figure of 2 a huge guess and is based on rectangular beam tests and specific concrete mixes. Concrete properties vary with concrete dimensions, mix design, aggregate and sand types, weather conditions, etc. And it all comes down to "2". It is a joke. As a world renowned professor specialising in long term concrete effects at UNSW likes to point out, you may as well toss a coin if you are using this method to calculate deflections.
2 put the compression reinforcement in the shear cage. There will be some in the flange anyway as slab reinforcemernt and the major part of the compression force will be concentrated in the web width plus a bit, maybe D to 1.5D width of the flange either side of the beam.

If the compression bars are below the neutral axis then they will do absolutely nothing to reduce creep deflection and will increase shrinkage deflections.

 

[Lion06]

rapt-

I'm not sure what you mean that it can't be a T-beam that acts like a rectangular beam. That is quite common. Any flanged section would be a T-beam, but will only "act" as a T-beam if the compression block depth, a, is greater than the flange thickness. If a<hf, then it will behave (and should be designed as) a rectangular section regardless of flanges.

The point I was making about the factored moment was that this beam was not designed such than phiMn is just slightly > Mu. phiMn >>> Mu, such that at Mu, while the section is cracked, both concrete and steel are still linear. The point of that statement was merely to say that the "compression" bars will not be taking load even at Mu.

I agree that this is not the best way to check deflections, but it is required in ACI if you don't meet the minimum span to depth ratios (which is not possible in this case). Additionally, I can't say that I know how to calculate the long term deflections "properly" (i.e. without the use of the multipliers), can you shed some light on this?

I'm not sure I believe that bars that lie just below the neutral axis will not help with creep and shrinkage deflections. The bars are helping to restrain the concrete within a reasonable distance around it - not just at the same depth as the bar

 

[csd72]

I think this paper covers all you need, also it disusses the shortcomings of kcs, enjoy!

http://www.ejse.org/Archives/Fulltext/200101/02/20010102.htm