Two-way PT slab, EFM method questions
Two-way PT slab, EFM method questions
(OP)
I'd like to preface my questions with a brief introduction. I am a structural designer with good experience in wood and steel structures, I am currently working on a mixed use project that will involve my first pt podium slab. I did not take prestressed concrete courses in undergrad or graduate school. Because of this I have taken a very long and slow approach to understanding Prestressed principles, starting with the common text books, then moving to practical hand examples. My studies have led me to PTI and their manual as well as there tech. Doc for pt floors. Finally I have literally searched eng-tips for any instances of: pt slab, podium slab, and post tension, I have read many (nearly all) discussions that I found to be relevant to my learning, as usual the incredible contributions to this forum have been a marvelous resource.
Project Description:
(2) residential floors over a concrete slab and commercial space.
IBC 2012 code, seismic design category D in us west coast.
The wood system is similar for he roof and frames floors, light joists spanning to bearing walls, the joists span in the N-S direction.
After managing to convince the architect to align the concrete columns and walls in a grid, I decided to begin the undertaking of the pt slab design by the Equivalent frame method.
Working with the contracting team I have selected a Banded-Distributed layout.
Immediately after beginning my preliminary layout two big questions came to mind...
1) banded strips align with the bearing walls from above, I used line loads to apply the super imposed dead load and live load from above this was fairly straightforward. Now turning 90degrees to design the distributed equivalent frames, I am puzzled what the best way to model the loads are. The predominant loads are perpicular to the design strip and would effectively act on the support line. I understand that the intent of the EFM is to design for the full applied load in both directions. The only way that I can see to accomplish this would be to 'smear' the wood structure loads and use uniform distributed loads. This is a huge stretch from the actual behavior, and I'm sure that Rapt will writhe in pain simply at he mentioning of this method (for more than one reason I.e. Average moment over entire strip). But I understand that the EFM , and banded-distributed is the accepted approach here and I would to stay with it. Can anyone provide some guidance? I have to believe that other podium structures involve the same issue.
2) seismic anchorage of shearwalls from above. I have seen several articles on this topic here, but I have a particular question that I do not believe has been addressed...
Thanks to an abundance of Windows and wall offsets, I have shearwalls hold downs scattered throughout the pt slab. I have seen a few responses that suggest how to model the point load over an area, but I am thinking more about the PT slab spans supporting the discontinuous chords. I have seen Some responses that suggest the addition of 'a few' tendons at point loads, but this mehod strata from the EFM, and frankly I have seen no such detail on the several podium structural plans that I have for reference. How have you podium guys reconciled the hold down loads (w/ over strength) using the EFM? Even if I was to place tendons for this reason how would I select the required drape and effective force?
I checked one of the spans with a hold down near the center of the free span (and almost on the column line) I found that the load did not control the PT design over DL, LL, and SL. It isn't too much of a task to check the rest of the anchors in this fashion except that using Adapt-PT I have to set the load as X combination thus I cannot envelope the design in a single run. The design obtained this way would not include any detail for additional tendons at he hold down.
Gravity direction point loads are trickey enough for the EFM but reversible loads!? I am stumped here and cannot seem to find any reference that addresses this. This communities wise opinions and experience would be much appreciated here.
And to Rapt, I have read at length your views on the banded-distributed approach, as many users have pointed out there are many structures in place that use this method (albeit at a loss of serviceability as you have pointed out), my thought is that I have already started down this path and I would like to stay with it because my mentor engineer will be most familiar with it, and I have familiarized myself with as many of the nuances as I can. If there is a compelling reason to modify my approach (ideally one that I can use the same software) I may be willing to change. That being said I do respect your point of view and I have often questioned why this method is for two-way plates but effectively produces a one-way system. Your responses are refreshing because I find that often times we blindly follow the code even when the basis is not the most rational.
Project Description:
(2) residential floors over a concrete slab and commercial space.
IBC 2012 code, seismic design category D in us west coast.
The wood system is similar for he roof and frames floors, light joists spanning to bearing walls, the joists span in the N-S direction.
After managing to convince the architect to align the concrete columns and walls in a grid, I decided to begin the undertaking of the pt slab design by the Equivalent frame method.
Working with the contracting team I have selected a Banded-Distributed layout.
Immediately after beginning my preliminary layout two big questions came to mind...
1) banded strips align with the bearing walls from above, I used line loads to apply the super imposed dead load and live load from above this was fairly straightforward. Now turning 90degrees to design the distributed equivalent frames, I am puzzled what the best way to model the loads are. The predominant loads are perpicular to the design strip and would effectively act on the support line. I understand that the intent of the EFM is to design for the full applied load in both directions. The only way that I can see to accomplish this would be to 'smear' the wood structure loads and use uniform distributed loads. This is a huge stretch from the actual behavior, and I'm sure that Rapt will writhe in pain simply at he mentioning of this method (for more than one reason I.e. Average moment over entire strip). But I understand that the EFM , and banded-distributed is the accepted approach here and I would to stay with it. Can anyone provide some guidance? I have to believe that other podium structures involve the same issue.
2) seismic anchorage of shearwalls from above. I have seen several articles on this topic here, but I have a particular question that I do not believe has been addressed...
Thanks to an abundance of Windows and wall offsets, I have shearwalls hold downs scattered throughout the pt slab. I have seen a few responses that suggest how to model the point load over an area, but I am thinking more about the PT slab spans supporting the discontinuous chords. I have seen Some responses that suggest the addition of 'a few' tendons at point loads, but this mehod strata from the EFM, and frankly I have seen no such detail on the several podium structural plans that I have for reference. How have you podium guys reconciled the hold down loads (w/ over strength) using the EFM? Even if I was to place tendons for this reason how would I select the required drape and effective force?
I checked one of the spans with a hold down near the center of the free span (and almost on the column line) I found that the load did not control the PT design over DL, LL, and SL. It isn't too much of a task to check the rest of the anchors in this fashion except that using Adapt-PT I have to set the load as X combination thus I cannot envelope the design in a single run. The design obtained this way would not include any detail for additional tendons at he hold down.
Gravity direction point loads are trickey enough for the EFM but reversible loads!? I am stumped here and cannot seem to find any reference that addresses this. This communities wise opinions and experience would be much appreciated here.
And to Rapt, I have read at length your views on the banded-distributed approach, as many users have pointed out there are many structures in place that use this method (albeit at a loss of serviceability as you have pointed out), my thought is that I have already started down this path and I would like to stay with it because my mentor engineer will be most familiar with it, and I have familiarized myself with as many of the nuances as I can. If there is a compelling reason to modify my approach (ideally one that I can use the same software) I may be willing to change. That being said I do respect your point of view and I have often questioned why this method is for two-way plates but effectively produces a one-way system. Your responses are refreshing because I find that often times we blindly follow the code even when the basis is not the most rational.






RE: Two-way PT slab, EFM method questions
I believe that you can simply ignore the loads for the bending and one way shear design of the distributed strips. It shall be your much deserved reward for getting those walls and columns aligned.
I really see no way around this other than to explicitly design for those forces, either via judgement for small forces or calculation for large. I've seen some clever anchorage details but there's nothing about that that's really specific to PT. I try like the devil to keep my hold down forces small or non-existent in these situations. Granted, I'm not in high seismic country.
I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.
RE: Two-way PT slab, EFM method questions
Thanks for the reply. I thought I might be hearing from you on this
So in the direction transverse to bearing walls, I could design for self weight of slab + live load to slab level only.
I was thinking maybe I could use the distributed PT panels to carry all the point loads from the hold downs, following the methods listed in Design Fundamentals of PT Concrete FLoors Section 3.2.6 B. THis could be useful on my end because I can do seismic in the one direction, and snow loads in the other direction, and not have to use (2) analysis (again because of Adapt-PT load combinations). I have not the experience really know if this would be sufficient. I suppose one could argue that the banded tendon frames need to be checked for the load so that they could support the reactions from the distributed tendons... which leads to checking the point loads in both directions....
Further Complicating the method laid out in 3.2.6 B, the anchors will be a combination of uplift and downward forces and reversible. I have a few instances were there are enough hold downs in the span to use the smearing method with the amplification factor... And my thought is that I could just take all the hold downs as acting up, then again all of them acting down. Again I am not really sure if this would be conservative or not. My guess is that the all down or all up method would produce a higher moment in the span than would the couple moment applied by hold downs with opposite signs...
So what do you guys think about the (2) proposals above?? :
1. Design only the distributed frames for the seismic point loads.
2. Assuming hold downs act either all up or all down over one equivalent frame (ignoring actual overturning effects where applies).
RE: Two-way PT slab, EFM method questions
Here is my logic....
PT design for point loads I am concerned with punching shear, and the effect of the design moment for the PT slab equiv. frame due to any point load situated over a free span. Adapt-PT allows for the input of Lateral moments at the face of the support. I could check the PT slab reinforcing for the additional fixed end moment in the seismic load combo without having to use the concentrated load feature.
Then I would check punching shear and slab anchorage separate.
And nicely enough the lateral moments can be reversed
Has anyone done this before, its kind of a round-about approach but it could greatly simply things for me and if the results are good enough then... well they are good enough!
Thoughts?
RE: Two-way PT slab, EFM method questions
Banded /distributed and total load over the panel width on total panel should only be used for UDL loads (if ever). I know that this logic has gradually been removed from the thinking of designers in the US over the years, but think about it.
If a load is applied in a concentrated area, reinforcement and tendons 5m away will have no effect on carrying it to the supports!
RE the rest of your post, I would prefer to see a drawing of the support/wall layouts and banded/distributed arrangements before commenting! But if the wall above is on the line of the supports below, the slab at mid-span parallel to it will not even know it is there, so cannot contribute to supporting it. A thin strip either side of the supports will support that wall!
If you are going to analyse for this on FEM, you will have to create extra design strips to pick up the concentrated load strips.
Remember you have to provide a load path for any load to supports. Draw out your orthogonal load paths for the concentrated loads using sensible widths of slabs to carry those loads.
RE: Two-way PT slab, EFM method questions
I would certainly be arguing for this. In general, both the distributed and banded design strips need to be designed to help carry all point loads from the point of origin back to your columns etc. The one exception is the one that we've been discussing where the point loads fall directly on the banded design strip center line and the distributed strip can be expected to not participate.
While my gut feel is that this would be conservative in most cases, it a tough thing to guarantee with confidence. You'll be dealing with continuous frames which means that you might have to deal with load patterning of the various spans to guarantee conservatism. And that would surely be more laborious than just considering the reversible loads explicitly.
I'm not a fan of this proposal. I simply don't think that captures enough of the important behavior to be considered adequate. For one thing, mid-span flexure would not be captured properly.
Here's a rough, perhaps questionable method that I've employed in the past. Keep in mind that I'm in a low seismic zone and my shear wall T/C loads are usually pretty modest.
1) Design the slab ignoring the shear wall T/C loads.
2) Run some quick hand calcs to prove to myself that the shear wall loads don't have a big impact on the macroscopic aspects of the design strips (whole panel flexure, one way shear, column punching shear). In exactly one instance I've had a problem where moment applied to a design strip was difficult to handle at a column because the slab was tending to uplift and the column/podium slab connections tend to be poor for uplift. Consider that a minor thing to keep an eye out for.
3) Make sure that the slab works for punching shear and anchorage where the shear wall T/C loads are delivered.
4) Consider only downward point loads / distributed strips. Add mild bottom steel solely for the purpose of providing the flexural capacity needed to deal with the point loads alone. For the purpose of this bit, assume that the the slab panel is simply supported at each end.
5) Consider only upward point load / distributed strips. Add mild top steel until you're comfortable that your uplifting load would engage enough of the slab self weight to hold the anchorage down. Distribution steel essentially. If this gets ridiculous, use a strategy like #4 but with top steel instead of bottom.
6) Consider the banded strips. These tend to work out well with no additional reinforcement because the bands effectively act as girders that collect load from the distributed tensions. As such, small downward point loads aren't a big deal and it's pretty easy to engage a lot of weight for your upwards loads. If the load are still a problem, add reinforcement similar to #4 & #5.
No doubt, my method has some "by the book" technical holes in it. I can think of a few myself. Regardless, I feel that it is fairly rational and computationally tractable, which is the crux of most things engineering.
Additional caution: while there are many benefits associated with putting you wall on top of the column/banded PT lines, construct ability may not be one of them. Give some thought to your sill plate anchor bolt situation and how things may be remedied if cast-in anchors wind up out of place. No doubt you're already on top of this.
I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.