Miami Pedestrian Bridge, Part XI 32

[JAE]

A continuation of our discussion of this failure. Best to read the other threads first to avoid rehashing things already discussed.

Part I
thread815-436595

Part II
thread815-436699

Part III
thread815-436802

Part IV
thread815-436924

Part V
thread815-437029

Part VI
thread815-438451

Part VII
thread815-438966

Part VIII
thread815-440072

Part IX
thread815-451175

Part X
thread815-454618


Check out Eng-Tips Forum's Policies here:
faq731-376

 

[hpaircraft]

The order to de-tension and then re-tension the rods in #11 shows how completely confused the design team was. These rods could add and/or subtract ZERO force to the region of failure, which was outside of the #11 envelope. The rods would have had no effect on the failure mode whatsoever. What caused the failure was a total absence of any steel tying #11 to the deck.

I can't disagree with the conclusion, but I have to say that there is a certain internally consistent logic to the re-tensioning. When you take an action and then something bad happens, most often the best thing to do is undo the action and see if that fixes it. For example, you deploy the wing flaps and the airplane starts rolling uncontrollably left. What to do? Put the flaps back where they were, of course! That might buy you some time to go do some troubleshooting and save the day.

Of course, the logic at work assumes that a) correlation means causation and b) the action is in fact reversible. It appears that the cracks started widening precipitously when the PT in 11 was de-tensioned, so they ordered the tension restored, hoping for both. However, it appears now that the widening cracks were not necessarily caused by the de-tensioning, but rather symptoms of deeper problems and merely exacerbated by the de-tensioning. And also that the action was in fact not reversible.

But in the heat of the moment, with engineers and contractors suffering from the narrowing tunnel vision that too often comes from a stressful situation, it probably seemed like a good idea at the time.

--Bob K.

 

[Earth314159]

The order to de-tension and then re-tension the rods in #11 shows how completely confused the design team was. These rods could add and/or subtract ZERO force to the region of failure, which was outside of the #11 envelope. The rods would have had no effect on the failure mode whatsoever. What caused the failure was a total absence of any steel tying #11 to the deck.

I agree the design team was confused. They detensioned the rods and cracks opened up. They made an association with a correlation and confused it with a cause. However, the retensioning of the lower rod added a substantial shear stress across the concrete joint. The member force was unchanged (equal and opposite reaction) but the shear stress on the concrete was substantially higher on an already over-stressed surface.

 

[Tomfh]

The detensioning did cause the cracks to open up further (it wasn't mere correlation), which put the structure into a state that retensioning caused the structure to collapse. Like a ratchet, or a wedging mechanism.

 

[Earth314159]

The detensioning did cause the cracks to open up further (it wasn't mere correlation), which put the structure into a state that retensioning caused the structure to collapse. Like a ratchet, or a wedging mechanism.

I don't think a definitive statement like that can be made. It may or may not have help initiate the movement but there was already signs of movement before the transportation. The release of the PT did reduce the shear friction clamping force but it also reduce the shear stresses on the concrete. Based on calculation, the release of the tension decreased the shear on the pour joint more than it decreased the shear friction capacity. These is not to say the theory is 100% correct/accurate but it is all we have to go off of. The only definitive way to confirm this is through a test. Regardless, we can say the joint was over stressed in shear friction. I think we can also say that it was a significant error to re-tighten the PT.

 

[Sym P. le]

I'll throw in this mark up for thought. The 11 lower PT cable had the transverse deck PT cable to pull against (cyan). I think the node is more apt to be described as the 11/deck/diaphragm node. 11 had various longitudinal cracks and a number of longitudinal rebar had seemingly little deformation.

 

[Tomfh]

Based on calculation, the release of the tension decreased the shear on the pour joint more than it decreased the shear friction capacity.

The cracks widened significantly when it was detensioned. If calculations say detensioning improves the situation (by reducing shear stress more than it reduced shear friction capacity) then the calculations are wrong.

 

[saikee119]

I think the 2 vertical plastic sleeves on each side of 12 have played a significant role in the demise of the 11/12 joint.

In Sym P. le (Mechanical) 26 Jun 19 01:22 sketch it can be seen the two vertical sleeves occupy about half the area the deck bonding with the Member 12. We now know the front and the rear of these sleeves have cracked through on both east and west of Member 12. Thus the deck was likely to have debonded from the two vertical sides with Member 12 after the tension in Member 11 had been released. The upper PT rod was therefore unable to distribute the stress into the deck and became more effective in compressing Member 12.

Lower PT rod on the other hand was able to anchor more effective into the deck as it is wholly inside the deck and close to the transverse PT tendons. Since the deck was likely to have debonded from Member 12 and so the lower PT tension would not be effectively felt by Member 12.

The above sketch shows if the deck were not monolithic with 11/12 then the upper PT rod will not be effective on the deck as it is nearly outside the deck area.

The photos after Mar 13 have confirmed serious movement in the construction joint and so the the two haves of either of the CJ would have been debonded too.

Thus the re-tension scheme was operating on the condition totally different from that in the casting yard. It is also instructive to note 11/12 had cracked CJ, at least partially, even before leaving the casting yard according to OSHA Fig 18.

The condition of the CJ prior to re-tensioning can be seen in OSHA Fig 32 to 38 inclusively.

 

[Earth314159]

The cracks widened significantly when it was detensioned. If calculations say detensioning improves the situation (by reducing shear stress more than it reduced shear friction capacity) then the calculations are wrong.

I don't think that is true. It is easy to confuse correlation with causation. This is human nature. The cracks happened to get worse after then tension was released. You can not say that was the cause of the cracks getting worse. It does not conclusively say the calculations are wrong or that the release of the tension initiated the movement. We know the movement started before the transportation but we can not conclude one way or the other that the cracks got worse because of the release of the tension.

 

[human909]

The cracks widened significantly when it was detensioned. If calculations say detensioning improves the situation (by reducing shear stress more than it reduced shear friction capacity) then the calculations are wrong.

Exactly. And there is more going on than simple shear friction. For example a tense PT member will take some shear. A slack one won't. Likewise in a perfect theory world shear friction should be pretty linear. In reality you can get non linear behaviour.

I don't think that is true. It is easy to confuse correlation with causation.

I think everybody here is aware of the difference. However the immediacy of the occurrence is an extremely big hint. If I accidentally hit my thumb with a hammer and it starts hurting I don't think it makes much sense to argue correlation vs causation.

 

[Earth314159]

I think everybody here is aware of the difference. However the immediacy of the occurrence is an extremely big hint. If I accidentally hit my thumb with a hammer and it starts hurting I don't think it makes much sense to argue correlation vs causation.

There is also reason to believe it had nothing to do with the initiation of the shear friction failure. We know that the pour joint was over stressed and there was some small amount of sliding before the transportation. Calculation indicated the detensioning helped the condition. If the failure was going to occur anyways and the plan was to detension, then it is just a correlation. For all we know, the detensioing may have improved the condition. I think you need more evidence to back up the statement. I also think it is irrelevant in the sense that there was a problem with this joint regardless of whether there was detensioning or not. There were technical, judgement and ethical errors made.

Your example is misleading since you pick an example where there is causation. Look at another example. My child was vaccinated. My child was diagnosed with autism. Vaccinations cause autism. It is the same thinking and it is human nature.

 

[Tomfh]

We know that the pour joint was over stressed and there was some small amount of sliding

Yes it was already cracking. The joint appears to have been struggling from the outset. The point is that the situation became much worse once it was detensioned. The members shifted and the cracks opened up to nearly an inch wide.

For all we know, the detensioing may have improved the condition.

How can it have improved the condition if it "cracked like hell" once they loosened the bars?

 

[Earth314159]

How can it have improved the condition if it "cracked like hell" once they loosened the bars?

How do you know it wouldn't have "cracked like hell" if it wasn't loosened?

 

[Tomfh]

How do you know it wouldn't have "cracked like hell" if it wasn't loosened?

I don't know for sure. But it makes sense. There are two possibilities:

A- releasing the PT made things worse, causing it to "crack like hell"

B- releasing the PT actually made things better, and by coincidence the bridge "cracked like hell" at that moment for a different (unknown) reason.

Barring good evidence for option B, option A makes a lot more sense to me.

 

[Earth314159]

I don't know for sure. But it makes sense. There are two possibilities:

A- releasing the PT made things worse, causing it to "crack like hell"

B- releasing the PT actually made things better, and by coincidence the bridge "cracked like hell" at that moment for a different (unknown) reason.

Barring good evidence for option B, option A makes a lot more sense to me.

The movement took days. It is not reasonable to say "at that moment". If it was instantaneous, the course of action would have been much different.

 

[Tomfh]

The movement took days. It is not reasonable to say "at that moment".

The OSHA report describes the appearance of the severe cracking following the detensioning as follows:

As they began to de-stress the PT bars of diagonal 11, cracks began to appear at multiple locations, most prominently at the construction joint of diagonal 11 and the deck and at the top of the diaphragm II. There were three VSL employees performing the de-stressing – Kevin Hanson (supervisor), Navarro Brown and Chester Ashley. Kevin is regarded as one of the most knowledgeable PT field personnel in South Florida. After observing the cracks, Kevin became visibly disturbed and informed other VSL employees of the situation. Kevin took pictures of the cracks, and sent them to his supervisor, Sam Nunez, stating that “it cracked like hell”

 

[Vance Wiley]

I agree. A tear on the dotted line.
Consider that the force from each rod is applied to the top end of member 11 and it is 27 feet away. The forces have aligned themselves wherever they want to in 11 so both rods can apply their force to the cold joint from the top. The location of each is critical to events at the deck zone, and the upper PT may have influenced the splitting in 11.
So if we consider member 11 as extending into the deck and including the bottom anchor plates, adding PT force affects only 11 and nothing else if we ignore the integral nature of the deck. Likewise releasing the PT in 11 also affects only 11. But 11 is not isolated from the deck - it is intended to be a part of the deck. How well it is integrated into the deck is the thing that is/was undefined and improperly evaluated.
By failing and blowing out the end it has shown us where it needed reinforcing and where attention was lacking.
Consider please my previous post posing the idea that loosening of PT and subsequent increased cracking caused loose pieces in the joint and the re-tensioning crushing those now mis-aligned pieces, leaving crumbles and sand in the joint with even less capacity to transmit loads to the deck.
In reviewing FIGG calcs in the March 15 PowerPoint Party, they did not address the failure plane/surfaces that actually developed. It appears that the contact surface of the deck was basically smooth and the factor Mu should be 0.6 instead of 1.4 or 1.0. Applying that to both the shear friction steel across the joint and the perpendicular force of 950 kips reduces those contribution significantly. Consider also that the early cracking and movements had broken any bond and the cohesion was lost and movements were underway. At what point can this zone at the deck surface no longer support any horizontal load?
With the remaining blow out being almost diagonal tension with little reinforcing, is it possible to now calculate the contribution of the breakout zone at failure? Considering reductions for sleeves, etc? This was step 2 of the two part failure, as I see it, and step 1 had already given its all.

 

[Vance Wiley]

To tension or not to tension - - -
An added thought to tensioning of member 11 -
After releasing the PT loads and evoking the resulting "it cracked like hell" remark, there could have been new cracks or just larger existing cracks, but the failure plane was likely defined. If so, the upper PT rod in 11 was above the actual failure plane and was effectively isolated from the deck, therefore the effect of retensioning of the upper PT rod was primarily internal to 11.
The lower PT rod remained anchored in the deck, even after the collapse, and any force in the lower rod created more shear force along the sloping portion of the failure plane, which was basically the slope of member 11 and this lower PT rod. If this sequence has merit, the tensioning of the lower PT hastened the collapse.
EDIT ADD "I suggest there is more certainty that the condition just described existed by the time node 10/11 had dropped a foot. It makes a good triggering scenario."
I think collapse was inevitable before retensioning began.

 

[Vance Wiley]

Thank you for the primer/refresher on surface energy.

 

[Tomfh]

Vance, I think you’re generally correct.

My impression is the CJ was already starting to shear, but the node was still hanging on. detensioning reduced confinement (or otherwise weakened) the node allowing the prism punch out to occur. It isn’t clear to me whether that failure prism was already a hairline crack, a latent crack, or completely uncracked, but in any case after detensioning the failure surface was a complete crack.

Apparantly it was continuing to slip bit by bit over the next day or two, but was still hanging on by its fingernails. Retensioning then gave it a final push (by adding shear force at the crack), causing it to let go.

 

[Earth314159]

The OSHA report describes the appearance of the severe cracking following the detensioning as follows:

I am not too sure why it even maters. In this particular case, what recommendation can come of determining when or why the an over stressed shear friction crack started to opened up. Calculations imply the rod tension was worse than no tension. The coincidence of the timeline implies the detensioing made it worse. The best we can say is the smaller cracks started before the move, larger cracks were first notice shortly after detensioing, and the cracks grew larger. At least we have some kind of timeline but to say what exactly was the perturbation that started the cracks is somewhat irrelevant because without a physical rationale, you can't really based any recommendations on this. Why not just use it as a timeline and say the detensioing may have been the likely disturbance that initiated the cracks on an already over-stressed joint. We know retensioning was a big mistake since that is backed up with substantial evidence and numbers.