Load-testing of post-tensioned beams

[jmy]

I am evaluating an existing structure that is approximately 16 years old. Minimal documentation pertaining to the original design and construction is available; however, I do know what the original design loads are.

The slabs and supporting beams are post-tensioned. The columns appear to be a combination of precast and cast-in-place.

Due to the existence of what appear to be flexural stress cracks in two of the post-tensioned beams, I am considering load-testing the beams to verify their ability to support their design loads. Over the years I have load-tested several different types of structures such as pre-stressed structural tees, concrete mezzanines, etc., but I have not previously load-tested a post-tensioned structure.

I envision performing the load test per ASTM E 196 and ACI 318.

Can anyone tell me if there are pecularities with post-tensioned structures that preclude their being load-tested? Is there anything different about load-testing a post-tensioned structure compared to others?

Thanks in advance for your thoughts.

 

[Ron]

Nothing in particular, other than to monitor the anchor points as closely as you would monitor the deflections. Watch the load vs. deflection response and make sure they are tracking immediately. If you load the beam and you get excessive corresponding deflection on lower loads, you might have loose cables.

As with any full scale load test, safety is paramount. Make sure "catch" shoring is in place and that you can monitor remotely. Make sure that you can remove the load quickly...that is difficult with dead load unless you are using water with a quick release capability.

 

[jmy]

Thanks, Ron. Yes, we will have "catch" shoring for the dreaded "just in case" scenario. We plan to load with water via the use of a liner with built-in quick release capability. I always require remote monitoring via binoculars, walkie talkies, etc.

The two beams I will load test are helping support a large, elevated post-tensioned concrete slab in a parking deck (as luck would have it the slab is the top/6th level of the deck!). I will load test one beam. The second beam will be load tested after the first load test is complete.

Do you mind giving me your thoughts on the size of the load test area? The beams supporting the slab are spaced at 30 feet on center, span approximately 60 feet, and frame into a column at each end. To make sure I isolate and get deflections of just the test beam, I plan to load a slab area that is 60 ft. x 60 ft. (i.e. 30 ft. to the beam on either side of test beam and 60 ft. along the beam length). I plan to install shoring tight to the underside of each of the adjacent beams that are 30 feet away to keep them from deflecting and affecting my test beam deflection measurements.
To keep from having to load such a large area, I thought about calculating the maximum design load moment to the beam based on the worst case loading pattern across the slab and then backing into an equivalent uniformly distributed load to apply along the beam; however, if I apply a uniform load just across the beam, I am concerned about the the slab on each side of the beam deflecting also, adversely affecting my beam deflection measurements.
Any ideas?

 

[VoyageofDiscovery]

It is inevitable that the system will spread the load out as it shares the loading or as adjacent beams attract load. I don't know what you will gather from this narrow a test. Not knowing the configuration of the beams and barring isolating the subject beam, I think the entire floor would need to be loaded to design load. Any thoughts? Obviously if you had all the post-tensioning information, forces and losses, you could model this in 3d and see the actual response from any size test load.

I would load the subject beam along with a series of adjacent similar beams, monitor the existing crack widths and if cracks are not occurring on adjacent beams, you have a definite flexural problem with the subject beam. That alone would mean it to be substandard to the adjacent beams, and most probably lacking from a design load perspective.

HTH

VoD

 

[Ron]

jmy...I think your approach is good. It is impractical to load the entire structure, even though that would obviously give you a better idea of load distribution. Load tests are not cheap. They can easily run upwards of $80k to $100k.

Isolating portions of the structure is sufficient for load testing as long as you understand the limits of the isolation, which I think you do.

As for remote monitoring, I was thinking of using either wired LVDT's or electronic dial gauges with wired or wireless transmission. That prevents having to have line of sight monitoring for difficult geometry.

I have done load tests of similar scope using strain gauges, LVDT's and electronic dial gauges. I've also done them like you are proposing....and in my more stupid days, physically walking under the unshored structure to take readings.

 

[jmy]

Thanks, Ron and VOD, for the feedback.
Ron, do you have a vendor or website for the various measuring/data acquisition system(s) you reference?

 

[Ron]

jmy...a few to get you started! Keep in mind that you will not likely find everything for your application conveniently packaged. You'll have to mix and match pieces and parts to develop everything you need. Once you have everything together, you can then test and calibrate it before putting it into the application.

http://www.vishaypg.com/micro-measurements/

http://www.omega.com/

http://www.humboldtmfg.com/c-2-p-548-id-2.html

Also, here is a link to a paper on of my partners and I wrote quite some time ago. It is a copyrighted article so can't post it here, but the abstract gives the gist.

http://cedb.asce.org/cgi/

http://WWWdisplay.cgi?89841#

Good luck

 

[Ron]

jmy...a few to get you started! Keep in mind that you will not likely find everything for your application conveniently packaged. You'll have to mix and match pieces and parts to develop everything you need. Once you have everything together, you can then test and calibrate it before putting it into the application.

http://www.vishaypg.com/micro-measurements/

http://www.omega.com/

http://www.humboldtmfg.com/c-2-p-548-id-2.html

Also, here is a link to a paper on of my partners and I wrote quite some time ago. It is a copyrighted article so can't post it here, but the abstract gives the gist.

http://cedb.asce.org/cgi/

http://WWWdisplay.cgi?89841#

Good luck

 

[Skeezix]

jmy,
In addition to good advice from Ron and VoD, here are a few more ideas for your consideration:
1. You may be able to find out more about the original post-tensioning by calling local subcontractors and suppliers. Even if the p-t was furnished and installed by a contractor from out of town, the local guys often know who their competition was.
2. Consider subdividing your water enclosure with dams or baffles to reduce the effect of existing camber, deflection or slope. Parking structure finish floor elevations are seldom dead level over 60-feet spans. As soon as you start your load test, slopes change and the water flows to lower areas making your load less uniform and complicating your analysis.
3. If you know the original design loads, do you know if they reduced the live load for the beam design? If they did, loading the beam to “design load” may be as much as 40-percent less than the load indicated on the original drawings.
4. If the post-tensioned beams are stacked vertically, shoring adjacent beams will transfer some of the load to the floor below causing those beams to deflect, too. If the beams below the adjacent beams deflect more than the unloaded beam below the test beam, it may influence your deflection readings. Another option would be to backshore the adjacent beams all the way to the ground.

Good luck. Let us know how it goes.

 

[jmy]

Ron, thanks for the links and again for all of your other thoughts.

Skeezix, thanks for the ideas. Yes, I plan to shore to ground level to help prevent the potential for overloading the underlying layer and causing progressive collapse. Great point about the dams/baffles to help negate the "ponding" effect.

 

[Ron]

Skeezix...sounds like this isn't your first rodeo either! Agree with the baffles...necessary to prevent progressive deflection accumulation. Have used in all of our "pooled water" tests.

It's interesting to look at the water from one baffled section to the next to see the differences on either side of the baffles...makes you know they were needed.

 

[jmy]

Anyone have any thoughts in regards to calculation of beam deflection? To calculate the anticipated maximum beam deflection resulting solely from the test load, I initially (conservatively) looked at the beam as a simple span condition, but that approach yields a large (almost 5 inches) deflection.
When I calculate the deflection based on a fixed end condition, I get less than 1 inch of maximum deflection at mid-span.

Intuitively the fixed end value is more realistic (besides, I am do not think the clearance requirements inside a parking deck can accomodate 4 inches of beam deflection); however, based on the size of the beam (24 in. wide x 34 in. deep) compared to the size of the columns (28 in. wide x 32 in. deep) into which it frames, I am concerned that the ends of the beam are not truly fixed but rather experience some degree of rotation. If in reality the ends of the beams undergo some roation, their mid-span deflections will be greater than that of a beam with fixed ends.

I don't want to "fail" the beams based on excessive deflection if they are actually designed to have more than fixed-end deflection, but without adequate drawings, I do not know of any other approach to take.

Any ideas?

 

[dik]

Any chance you can load the beams with hydraulic load cells? rather then ponded water? They are more forgiving for failure. but may be impractical due to the loading beams required.

Dik

 

[jmy]

Thanks, Dik. I had not considered that possibility; however, I do not believe I can make it work here, especially given the owner's desire to resolve this matter quickly. Water is readily available, and we have a contractor lined up that is experienced with this type of load test.

I am concerned about calculating my bench-mark defelctions for the test beams using the correct beam end conditions. A pinned end condition is not realistic, but a fixed end condition seems too conservative and somewhat unrealistic as well. Given the size of the beams compared to the size of the columns, I believe the column stiffenss is insufficient to truly fix the beam ends and prevent their rotation; however, I know of no way to determine if the original engineer designed for some degree of rotation.
I may have no choice but to go with fixed beam ends in my calculations.

 

[dik]

A load deflection curve can be 'accurately' determined by using hydraulic loading. This can give an approximation of the end conditions.
It may not be as costly as you think... and a degree of confidence is there.

Dik