Catenery Action in Unbonded PT 1

[PSR_1]

A paper by kenneth B. Bondy titled "Two way post-tensioned slabs with bonded tendons" has some interesting description about catenary effect associated with unbonded PT( see the highlighted part in the picture.

I want to see if this claim is sound and logical.

 

[rapt]

efr

Yes, as the strand is not bonded, it does not have strain compatibility with the concrete surrounding it, so instead of increasing the strain in the steel at a local point, it increases the strain in the entire length of the tendon, varying a little depending on friction.

This is an advantage during extreme events such as earthquakes at points with high stress concentrations such as support plastic hinge locations in an earthquake where very large plastic rotations occur. The strand is not a class E steel (very high ductility), it is much closer to Class N (normal ductility) so bonded strand could be a problem in these areas is an extreme earthquake event such as you might experience in California or Japan or NZ.

But this ability is also a disadvantage in normal design as the full capacity of the strand cannot be utilized so you need extra bonded steel to make up for the reduced capacity compared to bonded tendons. Also, because the natural failure condition of unbonded strand is catnary action, this is not the way you want a normal unbonded member to act under service or ultimate load conditions. There is no ductile failure mechanism as once you get a crack at a hinge location, there is nothing to tie the concrete together across the crack as you would get with bonded steel so technically there is no flexural capacity (catenary action is not technically flexure) from the tendons at that location unless a bonded tie is provided. Hence a minimum area of bonded reinforcement is required to force the member to act like a flexural member rather than going into catenary as plastic hinges develop under normal gravity load design conditions.

While the ability to go into catenary action might help when a column is lost, allowing the loads to distribute back to surrounding columns and the slab to go into catenary in the spans around the lost column, this will not guarantee against progressive collapse as the punching shear will then be much worse at surrounding columns trying to support the extra loads and rotations from the spans in catenary action and could result in further failures resulting eventually in a pancaking failure of the complete floor and possibly structure. Hence the need for bottom reinforcment in the slab to improve punching shear failure capacity, something that has been watered down in the latest ACI318 which now assumes that catenary action will save this situation.

retired13,
Huh!

 

[PSR_1]

Rapt, as usual useful and enlightening! thanks!

 

[r13]

I don't want to get into the argument of bonded vs unbonded tendons, as each has its advantages and disadvantages. My question is on the validity of the authors statement that was quoted on my previous response. The catenary action starts at the moment it was set to the profile.

 

[BAretired]

To go into catenary mode, tendons must deflect considerably more than their initial sag. I'm visualizing tendons and slab, under severe overload, draped downward, perhaps a foot or more below initial elevation. The horizontal reaction at each anchor, having no slab capable of resisting it, fails by pulling the exterior column inward. The deflected tendons may have three to four times demand at factored load when suitably anchored, but what resists the horizontal component of the reaction?

BA

 

[rapt]

BAretired,

Probably even much more than a foot. Headroom would be a problem for tall people. It is a collapse condition, you are basically hoping the designer has tied everything together sufficiently to redistribute the horizontal forces back through the rest of the structure and overall bracing system to hold it up until the building can be evacuated. And hoping it only happens locally in one span!

It is a good reason why you should have a reasonable portion of your capacity provided by bonded reinforcement in a building with unbonded tendons. This used to be a requirement of the UBC (Mu = DL + .25LL carried by bonded steel alone) but as I understand it, this has been removed in later versions and you now only have the ACI minimums, with none required in the bottom of a flat slab if the service stress is low enough.

 

[BAretired]

rapt,

I suppose, if the catenary behaviour occurred at an interior span, the intact exterior bay could likely pass the horizontal force over to adjacent bays, but if it occurred at an exterior span, it would appear to be game over. Considering this discussion so far, I am having difficulty seeing "post flexural catenary capacity" as being any advantage, let alone a significant advantage of unbonded over bonded tendons. It might be interesting to see the tests which show the high capacity of the catenary according to Ken Bondy.

BA

 

[hokie66]

Just in my opinion, Mr Bondy is grasping at straws to show that an inferior system is better in some way than a superior system. If that is the most significant advantage, the unbonded system would not be used. I suggest the most significant advantage is cost.

 

[BAretired]

Mr. Google found the Bondy article as noted below. I have not had time to study it in any detail.



BA

 

[rapt]

Yes, if an end span you would be relying on the external column taking the horizontal force and transferring it to the floors above and below. So it could possibly be ok for a local failure of 1 span, but not anything more complex as would happen in a seismic event.

Hokie,

Even the cost part is questionable. We found it no cheaper if providing a full site service PT operation. Based on our costings, It only worked in a factory PT operation as the USA system is set up for where the PT company manufactures the complete tendons and delivers them to site for the contractor to install. But we were completely set up for full site service, so that may have biased the results.

 

[rapt]

BAretired

The actual benefit is that the steel does not over stress at a point of stress concentration in the concrete as there is no strain compatibility. So if the member is well designed with adequate high ductility bonded reinforcement as well, then the section will maintain strength as a flexural member during an earthquake and the strand will not break at a plastic hinge at a support as may occur with bonded strand.

It will only go into catenary if there is insufficient bonded reinforcement to make it work as a flexural member.

The other situation mentioned was Robustness due to a punching shear or supporting column failure. That is when catenary action is required and you get out quick or hope the designer tied everything together well to take the forces induced!

 

[Ingenuity]

...I say you "get out quick"!

Hurricane Katrina from 2005:

 

[PSR_1]

Ingenuity, Is the picture showing unbonded PT building failure due to the hurricane?

 

[Ingenuity]

efr: Unbonded PT tendons are the red spaghetti.

There were contributing detailing issues that assisted in the failure.

 

[PSR_1]

Rapt, it is out of topic of the thread, but I want to hear more on what contributed to the economic questionablity u mentioned ?

 

[PSR_1]

Ingenuity, can I ask for more on the "detailing issues" ?

 

[rapt]

Ingenuity,

That joint in the bottom picture looks like there was not a lot of care taken in placing the top reinforcement well!

And being old fashioned, I do not believe in punching shear reinforcement in thin slabs.

Where is the bonded reinforcement in the failed beams in the top picture? And the failed face looks like there was no concrete bond across it.

efr,

We simply costed both alternatives, a little over 40 years ago.
Australia was always a bonded PT country from the early 1960's so everything was set up for bonded PT as a full service sub-contracting operation with the PT company supplying all materials and doing all installation, stressing and grouting. So we knew what that cost. We did a comparison costing to do full service unbonded and found it did not compete on pricing. Good quality coated strand was a major cost that killed it. And well trained site staff. Then you had the protection problems at the anchorages.
USA was always an unbonded PT country where the PT companies then basically was a factory manufacturing operation supplying the tendons. Most of the on-site work was done by the builder and presumably the steel fixers.
I will let Ingenuity tell of the problems with a lot of the lack of quality control in USA PT operations in the earlier years because of this approach if he wants to say anything. I only have the stories about quality of strand coating, anchorage protection and everything else that has caused problems second hand. I did once see unbonded PT is Singapore in the 1970's that used current USA technology that was basically paper wrapped strand with a smearing of grease. That is not the coated strand option that we priced!