PT Pan Stressing Tendon Elongation 1

[BacBac]

Hi All,

Structural engineer from Australia here.
I'm currently working on the construction phase of a project where the floor system is PT suspended slab, and our office is detailing the PT slab.
The typical tendon length on this project is >24m.

The design was completed using popular software in Australia which are RAPT and RAM Concept, with its default tendons anchorage loss values and default initial jacking stress (85% minimum breaking load).
Due to site constraint, most of the PT anchorage are done by pan stressing (refer snapshot below).

However, when the elongation stressing record were sent by the contractor, we noticed that the actual elongation on these pans are much lower (up to approx. 15-20%) than the theoretical elongation from RAPT or RAM Concept.
We requested the PT subcontractor to restress those tendons, however, what they did was to apply a jacking force >85% of the minimum breaking load of the tendons to achieve longer elongation.
I'm not too sure if this is the correct approach as most of the pans tendons have low elongation and the risk of tendons snapping.
Could you advise on the typical approach in this situation?
Any help would be greatly appreciated.

 

[Tomfh]

Are they consistently out? Is the jack calibrated correctly?

What's an example of the extension you're getting vs what you think you should be getting?

 

[ANE91]

Prestress losses are a “wild ass guess rounded to 3 sig figs,” per my old professor. If you compare the ACI (423.10R) and AASHTO models for creep, steel relaxation, anchor seating, friction, shrinkage, and elastic shortening, the numbers are all over the place. Further, elastic shortening losses very much depend on sequence of post-tensioning.

I wouldn’t be too worried about snapping. The structure likely never sees any load that stresses the strands in excess of that jacking stress, up to about 20% of which is lost by service. I’m used to 75%; >85% is indeed high, but hey, you’ve got more precompression now. If the early-age concrete can handle it at the ends without cracking, then I don’t see a problem. Also, losing a strand or two is undesirable but not all that rare.

This thread is great:
https://www.eng-tips.com/threads/pt-cable-strands-look-stripped-after-stressing.410514/post-8157854

 

[Just Some Nerd]

Encounter this a lot, I have noticed that pans usually end up an extra 5mm or so shorter than expected when it comes to extensions compared to edge stressing. Elongations seem to be 50% black magic, 50% random nonsense.

Personal experience has also taught me that many contractors here in Aus are also just making shit up by measuring tendon lengths off plan and multiplying by an expected strain, but typically that would be to hide their extensions being short. Even funnier, once a contractor was just doing it to avoid having to go around measuring it properly even when an extension was actually correct on site, but they'd given a number based on the wrong tendon length.

Being consistently 15-20% below expected elongations is enough that I'd even start running models a strand less to simulate the reduced prestress to see if the design still works. Usually it's only finding a few tendons coming in too low, and after communicating the tendon is understressed (without giving the expected elongation) they'll come back with an answer more within tolerance.

I wouldn’t be too worried about snapping.

Conventional wisdom from those I've learned from is to not try and re-stress a tendon more than once or twice to avoid snapping. Was a funny story at work recently of a tendon getting stressed just above the jacking stress to test the elongation and it immediately snapped, had a laugh about that one.

 

[dik]

Is it the Polypropylene type of sheath. For the stressing lengths, I would assume no loss.

 

[Ingenuity]

…however, what they did was to apply a jacking force >85% of the minimum breaking load of the tendons to achieve longer elongation.
I'm not too sure if this is the correct approach…

Stressing to elongation is definitely NOT the correct approach.

Always stress to force using a calibrated hydraulic system and account for the elongation shortcomings as best as possible.

For pan stressing a series of curved stressing barrels is used and the friction of these acute angles changes need to be accounted for in the friction calcs to better estimate elongations.


Also, any chance that some of the tendons are double-end pan stressed? If so, you could place a load cell at one end, stress from the other end and measure the force at the load cell - effectively giving real values of force at the tendon ends and therefore friction.

I have ‘heard’ recently that double-live end stressed tendons are seldom done in AU due to safety concerns. It was fine in my days but for some reason today it is not acceptable.

Good luck.

 

[Ingenuity]

Is it the Polypropylene type of sheath. For the stressing lengths, I would assume no loss.

dik:

AU does not permit unbonded PT - except for SOG - so the tendon sheath in this situation is galvanized flat ducting with up to 5 strands to form a bonded (grouted) tendon. Each strand is stressed with a mono strand jack.

 

[BacBac]

@Tomfh
Yes, they are constantly out for pan stressing. We have asked the contractor to ensure that the hydraulic jack is calibrated correctly.
For instance, we have 27m long tendons adjacent to each other with pan stressing on one and edge stressing for the adjacent, the elongation result we get is:
From RAM Concept: 186mm
Pan Stressing: 158mm
Edge stressing: 185mm
We'd accept if the elongation is about 10% lower than RAM Concept's results. 15% seems very low to be accepted.


@ANE91

I wouldn’t be too worried about snapping. The structure likely never sees any load that stresses the strands in excess of that jacking stress, up to about 20% of which is lost by service. I’m used to 75%; >85% is indeed high, but hey, you’ve got more precompression now. If the early-age concrete can handle it at the ends without cracking, then I don’t see a problem. Also, losing a strand or two is undesirable but not all that rare.

The contractor kinds of saying that if we keep asking them to restress, they won't be liable if any of the tendon snaps.

@Just Some Nerd

Elongations seem to be 50% black magic, 50% random nonsense.

True, seems like this elongation stuffs are very random.

Being consistently 15-20% below expected elongations is enough that I'd even start running models a strand less to simulate the reduced prestress to see if the design still works. Usually it's only finding a few tendons coming in too low, and after communicating the tendon is understressed (without giving the expected elongation) they'll come back with an answer more within tolerance.

The thing is, if we review the slab with less PT, we'll definitely have crack control issue as the P/A drops to less than 1.4MPa (minimum P/A for moderate degree of crack control as per AS3600). There might even be a deflection issue if we reduce the stress on the tendons.
There's not much rectification you can do to the poured slab I think?
Not sure if the typical practice in the industry is to always review the model once we get the tendon elongation from the contractor, don't think the fees are allowed for this exercise?

@Ingenuity:

For pan stressing a series of curved stressing barrels is used and the friction of these acute angles changes need to be accounted for in the friction calcs to better estimate elongations.

We have tried to increase the anchor friction coefficient to 5% to simulate the extra loss from the pan, but we still didn't get as short as what we found on site.

I have ‘heard’ recently that double-live end stressed tendons are seldom done in AU due to safety concerns. It was fine in my days but for some reason today it is not acceptable.

Yea, this is not a typical practice anymore.
We don't have double end stress on our job.

Thanks all for the insight.