welding to cut end of prestressing strand

[Prestressed Guy]

Is is permissible to weld to the cut end of a strand in a prestressed member? I have a DT that was cast without the stem bearing reinforcement and the unreinforced bearing stress is right at the limit with very little margin. I need to add a plate to the end of the stem that has reinforcement developed into the stem, but they are so narrow that this is a problem. One though was to slide a plate with a beveled hole over the end of the strand and weld it to the strand.
The concept is something like this.

 

[Tomfh]

You’re not supposed to weld (or heat up) tendons. They are connected mechanically. Plus I’m not sure how much that plate is really going to help in terms of bearing stress.

I’m not a bridge guy but I’d be wanting a steel boot if the bearing stress was excessive

 

[SwinnyGG]

I wouldn't suggest it.. you're effectively going to anneal the last 12" or so of that tendon. I'm not a bridge guy either but that can't be good. More importantly, the true affect on the capacity of that tee is a complete unknown.

 

[Ingenuity]

I too would NOT conduct any welding adjacent or to the strand.

An alterative to achieve a similar result - develop stem bearing reinforcement - would be to use the same plate with a 9/16" dia hole and install a strand barrel chuck over the strand that bears on the new plate. Essentially a passive anchorage.

Could you amend the holed plate to a steel 'shoe' with a 9/16" hole, that is epoxied to the end/underside/web of the DT to provide increased bearing stress capacity, with the barrel chuck engaged over the strand, bearing on the plate/shoe?

 

[KootK]

I agree with the objections regarding welding to the strand. I object to a bit more than that however.

If I recall correctly, the end anchorage method of bearing improvement is, effectively, employing shear friction across a steep, assumed crack at the end of the member. I'm not sure that it's even legitimate to utilize pre-stressing strand in this way. I feel as though a predictable result would be difficult to achieve with restressing tendons given that:

1) Tendons are slow to develop relative to rebar.

2) Tendons relax and induce creep.

3) In engaging the tendons in tension, you'll be undoing an in place Hoyer effect at the tendon ends.

To the extent that one might prestress the end anchor, that will yield a shear friction benefit similar to the use of reliable dead load in more conventional shear friction applications. Still, there would be a lot of uncertainty in that I think.

 

[Prestressed Guy]

Thanks for the comments. Just to be clear, this is fully bonded prestressed double tee. For some real numbers, the design reaction on this stem is 29.5k. The unreinforced capacity of the stem assuming 3¾" wide by 4" bearing pad contact comes out to 29.7k so nominally "OK". This is in a parking garage so nothing is level and if this was the high side bearing, I would not be concerned with those numbers because the slope of the stem relative to the flat bearing would make the stress distribution such that the min stress would be at the end of the stem and the max 4" inboard. In this case, this is the low side stem, so the stresses hit max right at the very end of the stem. The reinforced bearing seat is there to anchor the corner of the stem to prevent it from spawling off.

Please to no consider my comments below to be argumentative. I am just trying to get a clear picture of the pros and cons of this concept. I had already been debating many of these points with myself prior to my initial post. When I received the photo, my first though was "Wouldn't it be great if I could just weld a plate to that bottom bar and be done with it but that is simply a stupid ideal". The problem with simply stupid ideas is that sometimes I can't come up with concrete reasons why it should necessarily be bad.

@Tomfh The strand is developed by bond over a length of approximately 24" from the end of the member for cut strand. This is bonded prestressed double tee not unbonded mechanically anchored post tension.

@SwinnyGG This is fully bonded prestress. As such, by default the strand is annealed at the ends with 0 tension because the strands are always burned back flush to the end of concrete so the annealing due to the welding does not change anything. The annealing will most likely not be anywhere near 12" because it is fully encased in the concrete which makes a great heat sink. It should not really matter how much of it is annealed because that will not affect the E of the strand which stays at 2.9E6 psi. Annealing will reduce the Fy below the 270ksi but not below the 60ksi of rebar The bearing seat angle that should have been cast into the stem was designed with (2) #4 but a single #4 will provide the required As for the reinforced bearing.

@Ingenuity A chuck could not be applied to the projecting strand for two reasons.
1. there is only approximately 1" of space between the end of the stem and the bearing wall which does not provide enough room to install a chuck.
2. There needs to be a minimum of 1/4" of elongation in the strand for chuck take-up. Without that take-up distance of elongation, the chuck jaw assembly would just be loose and do nothing.

@KootK The bearing reinforcement is a shear friction-based design but my proposed design is no different than any shear friction design for steel cast against concrete. I would mechanically roughen the mating concrete and bed the steel into a mortar bed. The predicted cracking plane is assumed to be somewhere in the range of mid-bearing and beyond so at that point, the reinf. would be fully developed on the steel end by direct bearing of the plate anchored to the strand which is fully developed by the full length of the stem. In this design, there is no intention to retention. The weld is to the untensioned end of the strand to anchor the plate against the end of the stem.

1. It is true that strand is slow to develop relative to rebar but the end I need to develop is fully anchored by the 59' long stem. The other end would be anchored by the weld to the plate.
2. relax and creep - This strand is not tensioned at the end of stem, but the tension increases to 31k about 24" from the end so if there is any relaxing, it would tend to tension the free end if anything.
3. There will be not retensioning of the strand end so it will not affect the Hoyer effect and if anything will provide a better end region development because the strand will be mechanically anchored rather than bond anchored.
This will only affect the bottom strand and only in the few feet from the end which is no way is a critical section for strand
Again, the intent is not to have a prestressed end anchor, just to use the prestressing strand as a replacement for a fully developed #4 gr-60 bar. The more conventional design would be to drill two holes a minimum of 18" into the end of the stem on either side of the strand and epoxy a couple of #4 bars into those holes. Because of the fitment challenges of drilling those two holes, I would need to have the plate welded to the rebar after they were bonded in place so that side of the equation would be unchanged.​

 

[JStephen]

If the strands have high carbon content, they may be unweldable in the first place, regardless of annealing effects. Welding procedure and qualification would be problematic.
I'm also not sure how welding to that geometry would work, I would think the net result would be having half the strands melted right in half.

 

[Prestressed Guy]

@JStephen That is a great point and the one that was my greatest concern in my initial post. I am not sure how to go about answering it. One thought was to make a mockup with a piece of 1/2" strand weld to a 3/8" plate in a 9/16" hole with 1/4" bevel and pull it to destruction and see what happens.

 

[Prestressed Guy]

I did some internet searches for welding of strand and found this from Caltrans or CA DOT

https://dot.ca.gov/-/media/dot-medi...econstruction/bcrp-vol2/bcm50103aatt1a11y.pdf

Surprisingly, it does not say "Oh hell NO!" like I expected from that body.

 

[bugbus]

I was reading this article yesterday, which is interesting but doesn't seem to touch on the heat issue with prestressing steel:

https://www.structuremag.org/?p=22381

It gave me the impression that flame cutting is standard practice in North America? I don't imagine that welding would introduce significantly more heat to the strand than flame cutting, but maybe I'm wrong on that. Either way it seems wrong in my mind.

Our local transport authority forbids flame cutting of prestressing strands (Transport for New South Wales B113):
Cut strand only with high speed carborundum disc cutters or hydraulic guillotines. Do not use flame cutting under any circumstances.

 

[Prestressed Guy]

@ gusmurr Thanks for the linked article. This article was dealing specifically with unbonded mechanically anchored post tensioned strand which is a completely different animal than the than the fully bonded prestressing strand that 100% of my work is with. I have no experience with the unbonded post tensioned strand and was surprised to read that they flame cut the strand ends. I would be uncomfortable with that based on my knowledge of the mechanics of the anchoring chucks. I would be concerned about changing the metallurgic properties of the section of strand in the jaws that does all of the real work.
In the prestressed industry flame cutting of strands is the normal and safest way to destress the bed unless the bed is gang pulled which is not common except on large girder beds. I have seen beds that have needed to be destressed after casting and it can be scary because you need to pull the strand at least 1/2" to get the jaws to unseat and given the short length of strand between the abutment and the concrete, this will often take the strand which is at 0.75Fy to greater than Fy which will lead the strand to break unpredictably with terrifying effect.
The plants I work with ussually destress with a 6' torch with a support on the end to hold it at the hight of the strand. They put the flame on the strand and after a few second, Pop, Pop, Pop and it is done, and no one is near the action. Then the strand checked for slippage before being further burned back about 1/4" inside the concrete. This hole is then patched with epoxy grout. this causes high localized temperature for a few inches into the end of the member, but this is in the 0-tension region so does not affect the end capacity.

 

[bugbus]

Thanks, that's good info to know, I did not properly read your original post so I missed that you were talking about prestressed and not post-tensioned. I agree that the flame cutting / welding would only be an issue where you have that extreme heat applied directly next to the mechanical anchorage as you said, especially if it is left unbonded as it seems to be done in North America a lot of the time. For prestressed concrete, I wouldn't see an issue with the first few inches of the strand having undergone some serious heating. Our transport authority states (TfNSW B110) that flame cutting is allowed but not within 5 mm from the end of the member - the rest to be removed mechanically.

 

[rapt]

While I do not think it will affect the performance of the beam flexurally as the end 6 - 12 diameters length of the strand is assumed permanently at zero stress anyway.

How much force are you expecting to be able to generate at this strand/plate connection?

I doubt that full temperature welding will work. I understand the strand becomes very brittle at those temperatures. We have previously used a lower temperature braise to connect pulling strands to a group of strands, but the heated ends were removed.

Flame cutting is done sometimes with pre-tensioned ends as it gives a more gentle release than saw cutting, but again, the end is expected to develop no force anyway.

I would not have heat anywhere near the ends of post-tensioned strands, even outside the anchorage.

PS, while you are checking it, have you considered how the beam force actually gets to the end of the beam as required by all logical design models. A minimum tension force is required at the bearing point. With a pre-tensioned member with bearing pads like that, you will effectively have zero tension force at the ends. Welding the strands to an end plate will still provide none.

 

[Tomfh]

How much force are you expecting to be able to generate at this strand/plate connection?

That’s the real question here. If you have genuine concerns about bearing capacity I suspect you’d want something more decisive.

 

[BridgeSmith]

If I understand the issue correctly, there's not enough reinforcing in the local anchorage zone to confine the concrete around the strand. If that's what we're talking about, I suggest 2 possible solutions:

1) Test the strength of the concrete, and see if using the actual strength of the concrete gets you to a capacity/demand ratio you're comfortable with.

2) Add a steel jacket around the bottom and up the sides of the stem, with carefully placed bolts through the stem, to provide confinement to the concrete.

Rod Smith, P.E., The artist formerly known as HotRod10

 

[Prestressed Guy]

If I understand the issue correctly, there's not enough reinforcing in the local anchorage zone to confine the concrete around the strand. If that's what we're talking about, I suggest 2 possible solutions:

1) Test the strength of the concrete, and see if using the actual strength of the concrete gets you to a capacity/demand ratio you're comfortable with.

2) Add a steel jacket around the bottom and up the sides of the stem, with carefully placed bolts through the stem, to provide confinement to the concrete.

The issue is not anchorage zone confinment of the strand. Bursting reinforcement is provided by strirrups in the stem which is seperate from the stem end embed

This issue is a bearing issue per ACI-318 22.8 and PCI 5.5. The issue has nothing to do with the strand except that the untensioned strand happens to be in a good place to provide shear friction reinforcement to increase the concrete bearing to greater than the unreinforced concrete bearing capacity.

I have attached PCI 5.5 which includes the plain concrete and reinforced bearing equations. I should note that my 3-year-old grandson (who is a budding engineer) was helping me with understanding the details.

 

[Prestressed Guy]

I have attached the design detail of the stem bearing reinforcement. In the casting of this DT, the P6 embed was left out and there were (4) of the [S7-4] stirrups at 6" beginning at 1" from the end of the stem. (Note: the strand is not shown in the detail)

 

[BridgeSmith]

Is it too late to extend the bearing seat length?

Rod Smith, P.E., The artist formerly known as HotRod10

 

[Prestressed Guy]

extending the length of the bearing is possible but difficult to insure. This happens on two stems out of about 100 so would require a special detail for this one location out of the total project. In the past, I have found this to be difficult to get correct in actual field conditions. The bearing is on a precast corbel that is then cast into the wall. The corbel is cast in a steel form so would require a custom corbel. When you add to that the fact that this DT is one of a dozen identical members it becomes a problem to get this one of 12 at this one of 12 locations.

 

[BridgeSmith]

Could you put a bearing plate on the bottom of the stem and anchor it at a point farther out on the stem?

Rod Smith, P.E., The artist formerly known as HotRod10