RAPT - Flat Plate - No. of Tendons 2

[arik79]

Hello,

Can anyone provide any guidance/recommendations of correctly assigning the number of tendons in column and middle strips using RAPT

In particular I am struggling to find the best way to assign no. of tendons for situations where the column strips varies in a run

For example, a 3 span run, with constant 5000 slab panel width and with tendons spaced at 1500 centers

Span 1: 2000 long

Column strip width = 1000 (Max column strip width = Span/2 as defined in AS3600)​

Span 2: 7000 long

Column strip width = 2500 (not limited by span)​

Span 3: 2000 long

Column strip = 1000 (As above)​

Now, what is the best way to assign a 1500 tendon spacing in RAPT? The software only allows a number of tendons for flat plate, no spacing

Should I just conservatively assume the column strip is 1000 for all spans and assign a 0.67 number of tendons for all spans? I think this is too conservative

It would also be incorrect to define multiple tendons and stop and start them at the columns/changes in column strip width as in reality this is not actually happening.

Essentially I want to be able to assign the correct P/A for all three spans with the varying column strip widths. But its very difficult when the number of tendons is fixed for the defined tendon for all spans and there is no option for spacing of tendons.

Thanks in advance for any help

 

[hokie66]

Contact your software provider direct to discuss your situation. You will find them very helpful.

 

[arik79]

hokie66

I had asked the question about adding a feature in rapt to specify tendon spacing rather than number of tendons

They responded by saying they'll think about adding the feature in.

Hence, I'm asking here for people's opinions

 

[rapt]

arik79,

In the situation you have described, with 2000 7000 2000m spans, the centre span and the support regions of the 2000m end spans all have a column strip width of 2500 based on half of the width. Those areas are the critical design areas. The end spans are basically in negative moment over the full length, pretty close to being cantilevers, so the "nominal" column strip width logic of the code which a designer has to interpret in cases with such varying span lengths would suggest that the end span length ruling the column strip width logic may not be logical for that span in this case.

If you do the same run as 2 2000 cantilevers with a 7000 span, you will get similar bending moment diagrams. And your column strip width would be a constant 2500 over the full length of the frame.

So applying some sensible engineering logic, I would change the column strip widths full length of that design to 2500 and use a constant number of tendons in the column strip.

My advice to designers when they ask this question with details of the actual case has always been to rationalize the column/middle strip widths to be more logical. Sudden steps and sloping column strip widths need rationalization for logical design.

 

[arik79]

rapt,

Thanks for your reply.

The case I gave was just an example, not an actual design case

I understand it's unreasonable to have sudden sharp steps and sloping widths in the column strip like the code may imply.

So let's say we have a situation like this, an actual design case this time.

So RAPT has automatically defined the column strips by sloping them from the supports.

Now if I have a nominal 1500 tendon spacing, the actual amount of tendons I would have for each span would be 1.53, 1.53 -> 2.03, 2.03 -> 2.33, 2.33 for each span respectively ( The " -> " means the number of tendons changes from x -> y in the span)

Now in a situation like this would you simply perform three RAPT runs to obtain the correct reo at the columns and midspans? Say, a run for 1.53, 2.03 and 2.33 tendons?

Or would you fix the column strip widths to a certain width, say 3000, which is sort of the average strip here and assign 2 tendons. I don't agree with this method as the strip width is greater than what is allowed in the code at the 2nd column.

The run gets even more complex when you start adding horizontal steps

How about deflection, what would be the appropriate values for the number of tendons and strip widths? I suppose we can just be conservative here, and assume 1.53 for the full width, but if I'm getting deflection issues then, what is the best course of action?

Can I please get your opinion on the above?

 

[rapt]

First, I do not run RAPT support online in public forums. You have obviously contacted me before. if you have questions, send me an email.

Are you suggesting that you have tendons at 1500 over the full slab width, column and middle strips? In both directions? That is not something I would recommend as a logical prestress layout.

Using some common sense on deflections, obviously by inspection the left 2 spans are not going to be controlling deflections. In fact, if the loading is uniform, they could be reinforced only with spans 3 and the cantilever prestressed and spans 1 and 2 still would not control deflection. Same for crack control in spans 1 and 2 except at support 3 and the prestress from span 3 would extend over the support into span 2 so that is covered anyway, so would it really affect the design adversely if you assumed a column strip width of 3050 everywhere? The width of the column strip does not have much effect on ultimate strength!

If you wanted to have varying prestress in the column strip in each span, you could have a full length tendon type, another type starting before column 2 to the right end and another starting before column 3 to the right end. or any combination of that. Or RAPT allows you to terminate strands in a tendon at any point you want. I would not normally do that in construction, but you could use it as a tool to model reducing prestress along the frame.

In fact for an economical design, you would be stressing from the right end and 50% of your prestress could terminate about 2m to the left of column 3.

 

[arik79]

rapt

Thanks for your reply.

Okay, I will send an email if I wish to contact you directly. The purpose of this thread was to get other designers opinions on this topic as well.

Are you suggesting that you have tendons at 1500 over the full slab width, column and middle strips? In both directions? That is not something I would recommend as a logical prestress layout.

The maximum spacing for tendons specified in the code is 1500, unless calculations can prove otherwise (Which in itself is very vague). We have opted for 1500 spacing everywhere and have added extra tendons where column strips are working hard. Is that illogical?

Using some common sense on deflections, obviously by inspection the left 2 spans are not going to be controlling deflections. In fact, if the loading is uniform, they could be reinforced only with spans 3 and the cantilever prestressed and spans 1 and 2 still would not control deflection. Same for crack control in spans 1 and 2 except at support 3 and the prestress from span 3 would extend over the support into span 2 so that is covered anyway, so would it really affect the design adversely if you assumed a column strip width of 3050 everywhere? The width of the column strip does not have much effect on ultimate strength!

That is right, the deflections won't govern, nor would changing the width affect the strength too much. The width, however, does affect the number of tendons we can have in the column strip, which in turn affects the calculations RAPT performs. I know we can easily make a judgement call on this and just use 3050 + 3 tendons everywhere because it won't make a difference at those columns in this scenario. But, in a scenario where it does govern, say, there's a transfer column in the middle of that first span (or a change in slab thickness) then surely its important that the strip widths and number of tendons are modelled correctly.

If you wanted to have varying prestress in the column strip in each span, you could have a full length tendon type, another type starting before column 2 to the right end and another starting before column 3 to the right end.

If we use this method then wouldn't the losses calculated in the tendon be incorrect since we are modelling it as 3 tendons when it is actually 1 long tendon? I feel this would underestimate the reinforcement required, albeit by not that much.

RAPT allows you to terminate strands in a tendon at any point you want. I would not normally do that in construction, but you could use it as a tool to model reducing prestress along the frame..

For the case above this seems like the most logical way.

However for a situation where there were 2 4600 spans on the right hand side as well (So overall it would be 4600, 4600, 7600, 4600, 4600) what would be the approach here? We can't add strands into tendons after all, or at least not yet Also, I don't think modelling two tendons stressed from either ends would capture the losses correctly because in reality the tendons are just continuous and stressed from one end. I suppose in this scenario doing two runs is the answer.

Coming back to my previous point. I feel like this process could be significantly more streamlined if there was just an option for tendon spacing rather than just number of tendons for flat plates as well. This way, it would be just like RC where you specify a nominal mat and add extras where need be. Hope you can consider this in a future update.

Really appreciate your assistance

 

[rapt]

- Maximum 10D or 1500 spacing is because the unreinforced slab between the tendons has to span that distance. At that spacing, it is assumed that the stresses are low enough for the unreinforced slab to transfer the load to the tendons. A wider spacing requires reinforcement transverse to the tendons. If there are concentrated loads, the 10D does not apply and calculation is required in all cases. The code assumes that an engineer understands the design area he is practicing in. It does not give design solutions for every aspect of the calculations, just sets out requirements and limits!

- The tendon layout should provide a "load balanced" layout to transfer loads to the supports. Equal spaced tendons in both directions of a flat slab does not provide that.

Plus extra tendons need to be concentrated over the support within D of the support face, so 1500mm spaced tendons do not work in most cases.

If you were analysing in a 3D finite element program (and still dividing into design strips for design, not the US full panel logic which is not accepted in Australia and many parts of the world) which will account for the discrepancy in not using a load balance layout then you could justify equal spaced tendons as correctly calculated Secondary Moments would show the difference. That is if the people who programmed the prestress routines in the FEM design part of the program know what they were doing. I have no idea how the current FEM design programs do this internally so I cannot comment on which does it correctly.

RAPT is 2D. There is an inherent assumption that the designer is using a load balance arrangement of tendons and that load balance arrangement is consistent in both directions. This will affect secondary prestress effects on strength and crack control and deflection calculations especially.

- Yes, in some cases it is essential that it is modeled correctly. But if there is a transfer column in the middle of the 1st span, it is going to control the design and probably require extra tendons and reinforcement anyway. You will not be relying on equally spaced 1500mm spaced tendons in that span if it is that critical.

You have to look at each case, as I have done for the ones you have presented.

RAPT requires that you define the tendon pattern in the column and middle strips so that it knows how many are in each strip at each location. It is up to you to provide a logical distribution for the overall slab. In earlier RAPT versions, Load Balancing was used in the calculations to help calculate the prestress uplift effects which in turn was used for Secondary Moment calculations. If you have ever tried to do these calculations, they do not work logically for varying prestress along the span in the way you are trying to do it. For that reason, RAPT never allowed it.

Several years ago, we changed RAPT to not use the Load Balancing approximation for Secondary moment calculation. This was because we introduced calculations of the stress loss in the tendons at dead ends and in pretensioned strands due to development and bond effects. This resulted in very large stress variation in the strands which resulted in unacceptable inaccuracy in the Secondary Prestress effects from the Balanced Load calculations. So we went back to first principle calculations for secondary prestress to remove the inaccuracy. I have no idea how you could do these calculations in 3D FEM and that is the reason for my doubts in the earlier item! They do not allow for the prestress force profiles that we account for either as far as I know.

Using this new calculation method, it is possibly possible for me to change to allow a spacing of tendons in flat slabs. As I said, I will investigate this. But it is complicated and not a matter of simply flicking a switch to allow spacing to be defined. Considering equally spaced tendons is not allowed (unless the profiles in the different strips are adjusted to give a proper load balanced distribution of tendons), it is not something I think is a good idea so is not at the top of my list of priorities. And offering it might encourage other people who do not really understand what they are doing to use equally spaced tendons on the whole slab.

I have been designing for 45 years and been able to get around it using design logic, as have designers and RAPT users for the last 35 years.

- You have mis-interpreted my suggestion. All tendons start at and are stressed at the right end for this case.
1st tendon type is full length
2nd tendon type finishes about 1.5m into span 1 past column 2.
3rd tendon type finishes about 1.5m into span 2 past column 3.
etc

So span 1 has 1 tendon type, span 2 has 2, span 3 has 3 etc.

- same solution with a 3050 column strip width works for the 5 span arrangement.

The varying tendon arrangements above do not work in this case, possibly might if stressed from both ends.

 

[arik79]

- Maximum 10D or 1500 spacing is because the unreinforced slab between the tendons has to span that distance

I was not inferring that this statement is vague, I think the standard explains the reason for the clause sufficiently. Unless calculations show a wider spacing is possible This is the statement I was referring to. How would you even perform this calculation? Would you just assume the tendons are supports and the unreinforced slab is spanning in between the tendons? Just doesn't seem like an appropriate method. The standard provides clear calculation methods/guidance/requirements for most RC and PT elements, so it's unusual that it wouldn't provide any sort of requirements on performing this calculation. I might be overthinking this one.

- The tendon layout should provide a "load balanced" layout to transfer loads to the supports. Equal spaced tendons in both directions of a flat slab does not provide that.

Does this imply that if we space out the middle strip tendons (or change their profiles) and keep the column strip tendons as they are then we would obtain a load balanced layout? Sorry, I am a little confused about this. I didn't think over balancing the middle strips would have a significant effect on the secondary prestress effects in the column strips in the other direction? I understand load balancing as a means of economic design, think I am missing the point here.

Plus extra tendons need to be concentrated over the support within D of the support face, so 1500mm spaced tendons do not work in most case

Assuming you are talking about clause 9.1.2 here. We have specified reinforcement in lieu of tendons to resist the 25% moment.

Using this new calculation method, it is possibly possible for me to change to allow a spacing of tendons in flat slabs. As I said, I will investigate this. But it is complicated and not a matter of simply flicking a switch to allow spacing to be defined

Yes, I didn't mean for this to come across as being easy, apologies if it came out that way.

- You have mis-interpreted my suggestion. All tendons start at and are stressed at the right end for this case.
1st tendon type is full length
2nd tendon type finishes about 1.5m into span 1 past column 2.
3rd tendon type finishes about 1.5m into span 2 past column 3.
etc

Yes, this makes sense, thanks for the clarification. This sounds like the most logical way.

same solution with a 3050 column strip width works for the 5 span arrangement.

Depends on the loading arrangement. As you mentioned above, you have to look at it from a case by case basis. So I would need to focus on where the critical points are and ensure at the very least those zones are modelled correctly. Whether I am adding extra tendons, or doing multiple runs to capture each of the critical locations.

Thanks again for your help.

 

[rapt]

1

There are many code clauses that are basically "in lieu of more accurate calculations, the following shall be adopted"​

They hardly ever define the "more accurate calculations".

Unfortunately some designers use these to circumvent the rule without really proving anything. But that seems to be how things work these days, until something collapses and kills enough people.

2

For a flat plate, a perfect two-way load balance would have 50% of the tendons in each direction equally spaced and the other 50% concentrated on the support line. This is normally relaxed to allow the second 50% to be at most spread over the column strip. My preference is to have a concentration on the column line and gradually increasing spacing. basically the tendon spacing matches the 3D moment profile!

Flat slab with drop panels complicates this. It is really 50% of the load is balanced over the whole panel in each direction and the other 50% by a concentrated group of tendons on the support line in each direction.

A perfect one-way load balance would have equally spaced tendons in one direction and all tendons concentrated on the support line in the other direction. This is the USA/PTI method that ignores column/middle strips and is used for unbonded prestress where weaving of tendons is difficult so they took the easy way out but in doing so require enormous amounts of unchecked and uncontrolled redistribution to make it work. If this is used in Australia, all strength and serviceability calculations still have to be done based on column/middle strip logic, not as the PTI method allows.

Other load balance arrangements in between these also work as long as they are consistent in both directions in providing a load path to the supports.

Even the USA/PTI full panel method is only accepted as long as the tendon layout uses a tendon layout consistent in both directions that provides a correct load balance. Unfortunately software does not check for this, it is left to the designer to make sure they know what they are doing. If a non-load balance layout is used in FEM, then the full panel method is not allowed. Individual strips need to be checked. If the loading is not uniform, individual strips need to be checked. Unfortunately in the days of computer software, designers are now experts without ever having understood the requirements for design in cases where they are outside the scope of the design concept being used and the need for them to use the software correctly to achieve a good design.

No one should be designing PT if they do not understand this and detail their tendon layouts accordingly.​

3

​

And 9.2.1!

6

​

Yes, every case is different. In your first case, the area where the wider column strip is adopted and would then include more tendons is move than 100% over capacity with the extra tendons. So in that case, there is no problem. You would need to confirm this sort of thing in each case. So yes, if you cannot decide this from the main run, multiple runs might be needed, but that only takes a minute or so for each to check with smaller strip width and less tendons.