PT design Ieff=Ig

[rowingengineer]

following from thread "Post Tensioning as a Deferred Submittal" a discussion about Ieff=Ig for PT has been sparked, trying not to take over this thread I have started a new thread.

Lion06 had mention that Ieff=Ig is allowed by the ACI code, i believe this is in reference to ultimate design, and not serviceable design as Asixth was describing.

In most flat slab PT designs in Australia the design is for PT bonded combined with reo, thus it is likely that the slab is going to crack under loading at some stage. so a Ieff=Ig assumption is not correct and needs to be reviewed for each loading case. However I limit my maximum Ief to 0.7Ig.

The reason we do this is for restrained shrinkage and other similar effects.

http://www.nceng.com.au/

"A safe structure will be the one whose weakest link is never overloaded by the greatest force to which the structure is subjected” Petroski 1992

 

[rowingengineer]

extract from previous thread:

Lion 06
"I don't think there's anything dodgy about using Ieff/Ig = 1.0 for slabs. ACI explicitly allows you to do this. I do this for PT design and every engineer I know that does PT (and precast, prestressed) does this as well.

Is there any literature out there that says that you should be using something other than Ig for slabs?"

Rapt:
"Lion06,

Just because ACI allows something does not mean it is correct. ACI and Bransons formula both grossly underestimate deflections in lightly reinforced RC members. This has been shown extensively in tests. But ACI still allows it.

Part of the reason is Bransons formula for tension stiffening which is completely incorrect at the point of cracking, saying that there is no increase in deflection when the first crack occurs (Ms = Mcr). The other problem is that a lot of stresses that are created in a memeber are ignored by most designers eg shrinkage restraint by internal reinforcement (in a normal member this will be between .5 and 2MPa, depending on how heavily reinforced it is). So some sensible codes have allowed for this in an attempt to get results from simplified code deflection calculation methods to match real world deflections. There was a paper several years ago in an ACI journal by Gilbert that gave an example of this. Actual deflection of the slab after 9 months was 29.5mm, while ACI predicted about 9.3mm (interestingly RAPT got 30.1!).

In PT design, shrinkage restraint by the reinforcement is much less as the percentage of steel is reduced. But it still occurs.
But the other thing that is ignored in most deflection calculations is tension stress induced by external restraints to shrinkage and temperature change shortening. This just about always exists and in many structures is very significant. It needs to be allowed for. One way, without calculating it, is to make a blanket assumption that Ieff is limited to a percentage of Igross. Another way is to add a tension stress into all sections in the member, or alternatively, reduce the tensile strength of the concrete by an equivalent amount.

The most stupid thing to do is to say ACI does not mention it so I do not have to worry about it. ACI does mention it where it says shringage and temperature effects must be considered in calculating deflections. Most designers simply ignore this as either "too hard to do" or as "I dont know how so lets forget it" and put their heads into the sand."

http://www.nceng.com.au/

"A safe structure will be the one whose weakest link is never overloaded by the greatest force to which the structure is subjected” Petroski 1992

 

[rowingengineer]

I use both the maximum 0.7 Ig and also the volume change design out of the PCI. I use the volume change to adjust the tension strength. I did like the idea of a tension force, due to it being non repeatable for all problems.

http://www.nceng.com.au/

"A safe structure will be the one whose weakest link is never overloaded by the greatest force to which the structure is subjected” Petroski 1992

 

[asixth]

0.9 Ig for stressed slabs. Push it to the limit.

 

[abusementpark]

I don't think there's anything dodgy about using Ieff/Ig = 1.0 for slabs. ACI explicitly allows you to do this. I do this for PT design and every engineer I know that does PT (and precast, prestressed) does this as well.

What section of ACI states this?

 

[slickdeals]

ACI requires that prestressed two-way slabs be designed as Class U with ft < 6 (fc')^0.5.

Table R18.3.3 states that gross section properties to be used for deflection calculations per 9.5.4.1

 

[rapt]

Slickdeals,

And clause 18.3.3 changes the class U limit to .5 for two-way slabs instead of .62 because an average moment is being used. This is still not adequate as the stress at the support in the "column strip" zone will still be much higher and this area will be cracked.

Ok, but in calculating ft, you must take into account restraint to shrinkage and temperature shortening.

Most designers ignore restraint stresses except in heavily restrained cases. And even in those case, they only consider it for crack control, not deflections! It needs to be considered for deflections also if you are going to assume full Ig.

 

[Lion06]

I'm skiing for the weekend, so I'll respond when I get back, but I'm almost 100% positive that ACI 318 allows use of full Ig for deflections for class U slabs, which a two way PT slab is, by virtue of limiting the tensile stress to 6(sqrt f'c).

 

[Lion06]

Just to ask the question. From a practical standpoint, where do you see a real benefit in PT over a mild slab? You're using less mild steel, but you're not really gaining anything with respect to deflections. This is not a big deal on interior spans, but what about the exterior slab edges? I try to keep the deflections at the slab edges to 3/8" to respect a 3/4" caulk joint with 50% compressibility.

The rules of thumb that I've learned (from my mentor and several PT suppliers, including at least two who sit on ACI committee 423) are that the slab thickness for a two way PT slab is the thickness is span/40 (to 45). This is significantly more lenient than for a mild slab.

Additionally, the same people have recommended using Ieff=Ig. One thing that I've learned is that it's not good to be viewed as an engineer who is overly conservative merely for conservatism's sake. I've been burned on that before and it didn't feel too good (and the reason for being conservative was cometely justified based on the schedule of packages).

I'm all about learning, and I admittedly don't know all there is about PT, but to reject out-of-hand something that ACI explicitly allows seems foolish, IMHO.

So my logical next question is this - if you assume Ieff<Ig then that implies at least moderate cracking (beyond shrinkage cracking, I get that concrete cracks). How does that get reconciled with the requirement to have the tensile stress in the concrete be less than 6(sqrt f'c)?

 

[asixth]

Will those same people call you too conservative when you have got to defend the engineering design for a floor slab which deflects too far to be serviceable? They didn't for me.

 

[rowingengineer]

Lion06,
your confusing me I must be getting old, with max Ieff=.7Ig your deflections will still be less than a mild steel slab of the same thickness. Yes for two way flat slabs the difference will be smaller . The main advantage for PT from my perspective is in the speed on construction. However I am a fan of the one way slab and band beam approach or drop panel slabs.

As for slab edges I am a fan of the edge beam, reduces deflections in this area and gives you some help with regards to punching shear.

as for tensile stress in the concrete be less than 6(sqrt f'c), one is a theoretical requirement and the other is practical knowledge.

I don't believe 0.7Ig is conservative, I very rarely have Pt designers come back to me and request thinner slabs. Mind you the guys in my local PT supplier have been burned a few times on promising deflections requirements and not delivering.

I like to have all my eggs in a basket, such that if I am ever taken to task by another engineer I can justify most things, funny thing is that rarely when you go to court do they care about about the calcs.

http://www.nceng.com.au/

"A safe structure will be the one whose weakest link is never overloaded by the greatest force to which the structure is subjected” Petroski 1992

 

[rapt]

Lion06,

You need to look in detail at the effects of different things on design.

PT will reduce cracking and curvature which reduces deflection.

If you are checking deflections on a flat slab/two-way slab on columns, calculating it based on the full panel width as per the ACI PT flat slab design method is useless. Column and middle strip zone deflections are completely different for PT flat slabs just as they are for RC flat slabs. The amount of load carried by each zone is completely different, the stresses and amount of cracking in the different zones is completely different, but ACI says to use an average, so that sklabs that are really cracked around the columns show as being uncracked according to ACI.

If the "experts" you are refering to are supporters of the ACI PT flat slab rules, then I would not be listening to them. The rules are completely wrong and very misleading. You cannot defy equilibrium and statics.

We are not rejecting it out of hand. I have been rejecting it for over 30 years because it is wrong. If you think a PT flat slab works the way ACI says, you need to go back and study plate behaviour.

You are not including all of the stresses in the slab in your calculation of the applied stress. So you cannot compare it to the tensile strength of the concrete. If you do include all stresses, then yes you can. This includes the distribution of stress around different areas of the slab, internal restraint to shrinkage shortening from the reinforcement, external restraint to shrinkage and temperature shortening due to columns, walls, beams etc. If all of these are calculated and included, then es you can use the tensile strength of the concrete. But all you are probably calculating is the vertical load stresses. And you are probably even using an averaged gestimate of the prestress losses rather than calculating them properly.
If you include all of the effects then there is no problem, but you do not. That is why so many PT buildings in USA have major cracking problems (though PTI will never admit it) and also deflection problems.

 

[Brad805]

The use of Ieff=Ig will depend upon the amount of prestress. I would agree with that assumption if you are fully stressed, but that is not always the case.

If you want some really good night time reading (ha!ha!), try this book.

http://books.google.ca/books/about/Concrete_structures.html?id=1M3-IlzKljMC&redir_esc=y

Brad

 

[Lion06]

The only thing not included in my PT analysis is restraint from mild steel and temperature. Any external restraint is captured by the FE analysis and the inclusion of the columns and shearwalls above and below.

Regarding the internal restraint to shrinkage from mild steel - I believe this is extremely small and essentially negligible. If you look at a 7.5" thick slab of 5000 psi concrete with #4 bars @ 18" o.c. bottom and assume the top bars over the column match the #4 bars @ 18", then, on a per foot basis, the EA of the concrete is 362750k and EA of the steel is 7714K. Therefore 98% of the precompression is being taken by the concrete.

Thermal volume changes are small if it's in a conditioned space and the temperature differential is small.

This really just leaves creep and shrinkage. I'm just not convinced that they play that big of a role.

I'm by no means an expert in PT, but just telling me that ACI is wrong and I need to revisit plate theory really does nothing to sway me. Not that you need to sway me, but it's not an effective way to make an argument.

Also


Rowing-
I agree that Ieff=0.7Ig results in deflections that are less than a mild slab, but not by enough to justify the bump the span/20 to a span/40 for slab thickness.

 

[rowingengineer]

Lion06,
you are losing me again, why the span/20? are we talking two way slabs still?

here is a previous discussion in regards to the full width assumption were rapt expands on his reasons for calling the ACI and PTI design assumption incorrect. thread588-202462

http://www.nceng.com.au/

"A safe structure will be the one whose weakest link is never overloaded by the greatest force to which the structure is subjected” Petroski 1992

 

[rapt]

Lion06,

What FE analysis are you using to get all of the shortening effects?

If the time at pout is 35C (very possible depening on the location) and airconditioning is set to 20C, there is a 15C degree temperature difference that needs to be allowed for.

I have made this arguement several times on this site (thanks for ther reference RE) and am not inclined to go through it all again. Though any engineer who understands flat slabs should be able to work it out for himself, which is what I was trying to encourage you to do!

 

[Lion06]

I'm using RAM Concept.

 

[rapt]

Lion06,

Unless it has changed recently, RAM Concept only considers shortening due to prestress. It does not consider shortening restraint effects due to shrinkage etc. Shrinkage/creep effects are only considered in the prestress loss calculations and in long term deflection calculations if you select the top tier deflection design method. Restraint to shrinkage effects etc are NOT included!

 

[Lion06]

Why are the restraints not included. If you have the shearwalls modeled and the boundary conditions in and it's doing a true FE analysis, how is that restraint not being accounted for?

Also, why does it matter if the actual prestress force accounts shrinkage and creep with the loss calculations or if the program physically takes it into account? What is the difference?

 

[rapt]

Lion06,

Have you ever done in depth (complete) design calculations by hand? You obviously do not understand computer software and FEA.

A computer program does what the programmer tells it to do. FEA can be programmed to do what the programmers wants it to do. If the programmers have included a shrinkage model/loading into the program and added it to the other load effects, it's effects will be included. If they have not, it will not be included.

RAM Concept does not include a shrinkage loading into the analysis so there is no allowance for shrinkage restraint in the design. They do give the option of allowing for restraint to axial shortening from the prestress force if you want to include it.

They have allowed for a reduction in the prestress force due to shrinkage, but this is not an FEA analysis, it is a pre-processor calculation of the prestress force which is then used for analysis.