BANDED/UNIFORM TENDON LAYOUT 1

[PSR_1]

After reading multiple responses about how the banded/uniform tendon layout doesn't perform well under service conditions since the elastic moment distribution tends to prefer column/middle strip arrangement, I decided to start this tread with an ugly sketch attached.

If, for whatever span and column arrangement stiff beams are provided on column center line in one direction and the other is free from beams, the slab would tend to act as one way slab supported on beams(correct me if am wrong) and if it is reinforced following this one way action, then service issue won't there.

If we agree on the above statement, then I would say the banded tendons would act as a strong band in conventional RC system creating a hidden beam and the uniform tendons acting as rebar taking care of cracking issues. And the argued cracking won't be a problem...this is how I see it,please correct me if am wrong.

The other thing we must not forgot is that the code is asking the top rebar around columns to be placed within 1.5h, which is a very small region and it would tend to take care the cracking associated with high moments that are being transferred to the column

 

[rapt]

This has already been answered in those other threads you mentioned. The answer is still the same.

The strong bands you talk about only influence things after everything has cracked sufficiently for the moments to redistribute to those strong bands.

 

[PSR_1]

I understand nothing is going to change and I have got the logic behind why the performance of such layout is bad under service load given elastic behavior of two way slabs. what I want you, rapt,to help me understand is that If one way beam slab arrangement(the one shown in the sketch) works okay under service condition even if the slab layout and the ratio between longer to shorter dimension suggest otherwise, why wouldn't the banded distributed perform well under service conditions? after all you choose your load path and design accordingly?

I want to be convinced enough to change my tendon layout to uniform-unfirom( the basket weave), that's why am asking.

 

[rapt]

Because the banded distributed tendon pattern does not stiffen the slab like a beam slab system does. The banded/distributed tendon slab still works like a flat flat slab for the gravity loads.

The extra stiffness of the beams completely changes the way the slab bending works, and depending on the stiffness of the beams relative to the slab, can make it a one way slab system supported by stiff supporting beams or close to it.

The moment pattern without the beams is purely a flat slab moment pattern. Flat slabs have concentrated moment regions that one way slabs do not have.

In USA, they use unbonded PT and it is almost impossible to lay a weaved tendon pattern in a flab slab with unbonded PT (economically at least) so they adopted the banded distributed system which allows them to lay the banded tendons first and then the distributed tendons over them and there is no weaving. And then to make theri design life easy, they opted a design method that ignores the variation of the moments across the slab between column and middle strip regions. As long as the tendon layout is a load balanced layout, that system will work at ultimate. To use the same logic at service is not logical, but they do it. So they do not know if the slab is really cracked ot not, cannot do partial prestressing and have no real idea of the effect of cracking on deflections. In the banded direction only the tendons withing the band over the columns can be included in the calculations, but unfortunately some designers are stretching this a long way! I once saw the tendon band actually the whole width of the slab panel for a design in Thailand. So in effect it was distributed/distributed and that does not give you a load balance pattern
To do banded/distributed for the Australian code, you still have to check crack control based on real moment distributions around the slab using as a minimum the column/middle strip logic and you still have to provide the 25% capacity over the column mentioned below in both directions, so you column strip width over the column ends up with probably about 70% of the capacity in that direction anyway.

Australia and many other countries have generally adopted a column/middle strip pattern in both directions (essentially banded/distributed in 2 directions, so that the tendon pattern reflects fairly closely the elastic moment distribution in the slab. By doing this, you can do proper partial prestressing calculations for crack control and deflection calculations. The tendons have to be weaved in laying them, but that is possible as the duct lengths are only about 6m so they can be put into the weaved pattern where necessary before being completed.

The banded/distributed system has traditionally only been used in Australia when the columns only line up in one direction, so the band would be in that direction.

The Australian code does not require all top reinforcement be within 1.5D from the column. It requires that 25% of the total negative moment capacity be within D either side of the column. This can include some of the tendons and must be supplied in both directions.

One other thing when you are using RAM Concept, make sure you have Mxy moments turned on. They are always calculated, but only included in the design effects if you tell them to. Unfortunately some software people, especially from USA, think you can ignore them. You cannot, they must be included in design.

 

[PSR_1]

Thanks for the explanation! In RAM I always make sure the wood and armer method is there while defining Strips to make sure torsion is included in the design( your explanation on how static would be lost if Mxy isn't included in the design has awoken me, thanks).

Three things I want to ask you:

1. ACI 318-11 section 18.8.2 requires the rebar provided "shall be adequate to develop a factored load at least 1.2 times the cracking moment" to insure enough reinforcement is there to prevent cracking. Wouldn't this requirement have positive effect on the service behavior of flat plates? since it means more rebar.

2. Since the "hypothetical" tensile stress limit on two way flat plates is lower than that allowed for one ways slabs(ACI 318-11).to attain this limit you would be required to use more tendons( more cowbell as Dirk bondy would put it). since am working with bonded, I usually use 0.5m or so distance between tendons in the banded direction. If I have 5 sets( each let's say with 5 tenons in them) then I would place one set over the column, and 2 in each direction from the face of column. so I have a tendon 1.5m away from the column. for shorter span we can agree that 1.5m away from the column is almost middle strip, how do u see this, your comments?

3. The Stress around column are very high and then vanish very rapidly as we go further away from the column. my thinking is the most likely place to crack in two way flat plates is that area around column. which would usually be provided with rebar to control cracking and to provide strength. so 2 and 3 together would serve as crack proof strategies?

your thoughts please!

 

[rapt]

efr,

I thought you were Australia based and AS3600. Are you designing to ACI in USA and using unbonded or somewhere else and using bonded PT?

1 That is simply a minimum reinforcement rule. It has nothing to do with crack control and will not prevent cracking as it is not related to the applied loads.

2 That is the sort f high level technical logic I would expect from a Bondy! The flat plate limit is lower but not sufficiently lower. Why fudge it when you can calculate it. You are doing an FEM analysis which is giving you an elastic stress distribution from which you can see the real stresses that are acting on the concrete.
You are then taking the average of those and comparing that to a fabricated illogical stress limit. The British, when they were playing with unbonded design in banded/distributed systems and using average moments set that limit to 0, not .5. That was a lot more logical.
When we design it with 2D software, we still look at a column strip stress rather than an average over the full panel. You have the real stress distribution before cracking and which is the pattern that will cause cracking. Use it.
The Band is supposed to be concentrated over the column. If it spreads as wide as the full column strip, it is not a band!

3 If you are using the average stress relative to the .5root fc stress limit and that stress is less than the limit, that does not mean that you do not have cracking in that area around the column. Yes, having your reinforecment there will help control any cracks you get, but it is still cracked. And you are doing your deflection calculations on the basis of the averaged moment and assuming it is still uncracked.

 

[PSR_1]

ACI and am in Africa. I actually don't agree with the averaged moment over the entire strip width. Even for RC we try to capture the moment distribution in column strip by dividing the column strip into smaller strips inorder to avoid averaging over a large stip and losing the actual moment distribution across the design section. I am more into the banded distributed logic. From your explanations it seems some of the tendons in the banded direction shall be placed in the middle strip and more of the tendons cconcentrated over the column.

About the concentration over the column to qualify as banded, anchorage sspacing shall come into the picture, you can't just bundle them.

 

[rapt]

From your explanations it seems some of the tendons in the banded direction shall be placed in the middle strip and more of the tendons concentrated over the column.

Not sure where I said anything like that.

About the concentration over the column to qualify as banded, anchorage spacing shall come into the picture, you can't just bundle them.

The whole idea of a band is that it is transferring the load directly to the column. If it is spread out over several metres width in a relatively thin slab, then that is not goign to happen, so it is not a band. Unfortunately the clever people who developed the logic did not define things like this, and you get bands that are far too wide and therefore not really a band, as I mentioned above.

The logic only works if the banded direction tendons are in a relatively tight band over the columns. They need to be significantly tighter than a normal column strip and cannot be out anywhere near the middle strip. Probably not more than half of a normal column strip width.