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

 

[saikee119]

Got a feeling some of us are getting confused when it comes to modelling RC structures.

To start with in global analysis of the full structure, using 3D finite element or 2D frame analysis, it is sufficiently accurate using the linear model even a concrete section cracks under load and so its second moment of area changes. This is because the equilibrium computation is based on the relative stiffness of the members joining at a node. The old engineers would know that we could calculate all the moments and shears of a continuous beam by using the ratios of the I-vales. The actual I-values are needed only when we have to estimate actual quantities like deflections.

All international RC design codes allow us to use gross sections for each member in the structure analysis because the relative stiffness ratios of the cracked concrete sections, with concrete resisting only in compression but no tension on the other part of the neutral axis, will not be materially different from those based on the full sections when the structure has no load. It is an approximation but has been accepted as good engineering practice and able to produce acceptable and safe design.

An other example is we all know a structure must deflect under load so the equilibrium condition can only exist in its deflected geometry. However everybody uses un-deflected geometry when the structure has no load for the analysis. This is again an approximation and accepted as good engineering practice for the majority of the application. People would bother non-linear analysis with flexible structure known to produce large deflection like a a suspension bridge. Last time when I looked at one the difference between geometrical nonlinear and linear solutions was only 4%.

In element or node analysis we put in rebar to carry their structural duties. This is done according to the code so if we need the rebar to give its maximum steel stress the bar has to be embedded to adequate length to give the full stress development.

As long as the rebar is sufficient and arranged in manner according to good engineering practice the design assumption that at each joint the moment, shear, torsion and axial force can be distributed according to their stiffness will be realized.

What we have here is a joint failure because it was rigid enough. Therefore the deficiency should be either the rebar was inadequate or some bar ineffective or some bars placed in a manner at variant with the good engineering practice or concealed construction defect or a combination of several flaws.

Finding fault with the structure modelling or the analysis techniques is barking on the wrong tree.

 

[Sym P. le]

TITLE: 11/12 vs. Deck/Diaphragm

My revised drawing and images from OSHA for comparison of left/right side diaphragm and 12 cracking as well as indicating damage to underside of diaphragm.

saikee119's B shear plane is the only significant shear zone as the deck surface is not in contact with 11 and there is no indication that the cold joint shears through 12.

 

[hpl575]

It has bothered me, as others have stated, on how the decision to restress the PT bar in Member 11 was made. I would like to think it went beyond, seems like it was working better with the force in the rods, lets put it back on.

A thought I heard many years ago on the safety of a structure that has undergone distress is that whatever happened, the structure that remains not only can resolve the static forces, it was able to stop the movement after the event occurred, and should be relatively safe until another event makes it mad again. Further it is generally better to do nothing than do something wrong and creates a situation you do not fully understand.

A scenario, you get a call late one night. My roof trusses are failing due to very heavy snow fall, don't worry we've already installed shores, please come out tomorrow and take a look. You arrive the next morning to find they installed shores right in the middle of panel points of the bottom chords. Now the stiffest axial support path creates bending in members that were primarily designed for tension. Now you not only have to design a fix, you have to design something to fix new situation before the snow begins to fall again and only hope that they didn't send someone in while the snow was still falling to install the shores.

Most of the time you have a good idea what happened, heavy snow the structure has never seen before, someone drives a fork truck over metal roof deck, someone removes a load bearing wall that they thought was just a partition. Side note, it always amazes me on the home fix-it shows that they suddenly find a load bearing wall. Because all homes built in the 1920s had floor joists that span 32 feet.

If the event occurs quickly, then most prudent people would "red tape" the area to restrict access until the situation is fully understood. Events occurring over time are much more complex. It can be as simple as, yes the cracks in your sidewalk are going to get bigger over time, to determining what is happening in a very complex structure. It would still seem prudent to me if you do not understand the situation, "red tape" the area and do not do something that could make the structure madder than it was before.

Also remember that construction workers want to go home at the end of the day too. Never assign a task that you wouldn't stand beside them while they do it.

 

[human909]

"Weather is fundamentally a chaotic macro system and one that doesn't readily partition."

So is concrete. At least fluid compression and expansion is linear; not so with concrete. As I said non-linear and non-homogeneous makes for large variability, especially on the micro scale (and I'm talking inches, not microns).

I don't want to sidetrack this too much with a tangent debate but lets have some real context here. Weather is a dynamic system, a concrete structure isn't (well at least it shouldn't be. Like saikee119 said, "finding fault with the structure modelling or the analysis techniques is barking on the wrong tree". FEA is used to design and test much more complex structures than this bridge including structures involving multiple materials and non homogeneous behaviour. Sure you can get it completely wrong if you don't model it correctly but that is the case with all analysis.

 

[hpl575]

The construction joint, Egad, I still can't believe anyone would have a construction joint at such an angle to the primary axial force in Member 11. Still thinking peer review, ASI, RFI, shop drawing markup, field observation, or something would have changed this from flat across the deck. Although as others as posted, it would appear that FDOT did raise concerns and was politely told to sit down and shut up.

Why would you even consider creating such a situation. ACI 318 Commentary, "Construction or other joints should be located where they will cause the least weakness in the structure." Below the deck, now you have to actually resolve the forces in a thin section of a strut and tie model, above the deck, "ugly construction joint" in a structure that is supposed to meet the "Bridges As Art" criteria.

 

[Tomfh]

Finding fault with the structure modelling or the analysis techniques is barking on the wrong tree.

They used 3D finite element analysis to reassure themselves and/or others that the node was safe. Such software cannot properly analyse a joint like this from first principles, so they should have been doubly cautious when the software said it’s ok.

 

[saikee119]

I have prepared a break down of the failure surfaces and explain them with the following sketches.

The center line through the truss/bridge should look like the sketch below.

Fig-1 Deck center line not meeting center line of 11/12

An interesting feature is the deck does not join the centers of 11/12. It eventually broke off from Member 11 leaving 11/12 substantially as one entity despite serious spalling at the joint.

I then looked at how the deck was bonded to 11/12. This is shown in the sketch below.

Fig-2 slid CJ altered the internal resistance pattern of the deck/11/12 joint

Area C is deck only. One can say the deck could be bonded to 11/12 by one vertical interface between Area A and Area C plus two side interfaces of Area A+B with 11/12.

FIGG is his Mar 15 presentation had assumed the Area A to be triangular but I have assumed it to be rectangular here. The shape doesn't matter because we all know now the construction joint (CJ) slid and so any concrete bonded to the two sides of 11/12 would have been irrelevant and so is the shape of Area A. The resistance would then be transferred to the surface D, the interface between the north edge of the deck (in AREA A) with the south face of Member 12 (Area B). Area A therefore had no shear failure and remained part of the deck even after the collapse.

The concrete bond from the deck with the two sides of Member 12 is therefore left to Area B which is the same full cross section as Diaphragm II.

However as I have pointed out FIGG's designer had chosen to put 2 No. of 4" ID vertical sleeves and one 8" ID horizontal pipe at this critical location to weaken the joint dramatically. The consequence is shown in the next sketch.

Fig-3 side shear resistance of 11/12 significantly compromised by the embedded sleeves and pipe

The Area A was divided into three vertical strips of 3", 2.25" and 9.75" wide. The first two were too flimsy and would break quickly when they were pressed against the flexible sleeves.

The 8" horizontal pipe occupies 40% the concrete sectional area. Thus the section with remaining 60% area became the weakest plane against the horizontal force. I denote this failure surface as Area E. It has two strips each with 30% of the original area on either side of the pipe.

These two strips now form Area B1 which has 17.3% of the original section area of 24" wide by 40.5".

I summarize each available failure areas or surface, including the easily broke area B2 and B3, by the last sketch plus a spread sheet showing their dimensions and the likely resistance by concrete, rebar and clamping force.

I have labeled the failing surface starting from CJ as a, b, c, d and e. The two sides area now have B1, B2 and B2 for completeness.

Fig-4 Details of failure surfaces and areas

The above sketch is relevant if we want to assess why the two PT rods could have different strains when the same tension force was use in tension and de-tension.

Fig.5 A tabulation of the failure surfaces/areas

The use of this spreadsheet is for us to examine what rebar crossing each failed area or surface and sum the total shearing resistances to prove if the joint was strong enough against the collapse.

 

[saikee119]

it would appear that FDOT did raise concerns and was politely told to sit down and shut up.

I comment on the above as I don't want people misunderstand the good will gave by FDOT.

The bridge was never FDOT responsibility. FIU owns this bridge and so it paid a consultant like BPA to do check the construction. The peer review was done by Louis Berger who remains the sole authority to determine the structural integrity of this bridge. That peer reviewer wasn't properly engaged to do a full review to include the temporary conditions during construction is another story.

My interpretation of FDOT's responsibility is that it oversees the prject to ensure the right code, loads and materials were used. It liaises with the design consortium on all matter interfering with usage of the public highway. FDOT has to ensure the finished product would not have maintenance problem jeopardizing the function of Highway 41. FDOT did intervene to revise the lanes arrangement to suit the future expansion of the traffic.

FDOT has considerable experience in bridge design and so everyone was making use of its expertise. FDOT did put down a lot of very useful comments. They were valid, fair and not pessimistic or nitpicking. FDOT even marked down comments to remind FIGG's omissions.

What FDOT gave in design matter has no regulatory or contractual power and FIGG could dismiss them all. FIGG can design its bridge differently to address the concerns. Not many realize FDOT advice was free and represented no more than a good will to the FIU project.

To build this bridge FIU should have its own engineering representative in verifying the design and the same or different representative to oversee the construction. It appears FIU had been over reliant on FDOT to save the cost/need to hire a professional representative to check every stage of the bridge design.

In conclusion without contractual obligation or regulatory power FDOT would not fight back if its advice were disregarded. By giving out the advice FDOT can consider its duty discharged.

The lawyers may have a go at FDOT citing bad advice. However I am confident the expert witnesses will be on FDOT side. FDOT did issue a fact sheet to clarify its position but FIU thought it was a fake news! Such was the depth of misunderstanding.

Had FDOT owned the project the game would have been played differently.

 

[FortyYearsExperience]

I have prepared a break down of the failure surfaces and explain them with the following sketches.

Saikee119. Appreciate the effort you have gone to in preparing these sketches.

However, there is a major flaw in your analysis. Your sketches show only a vertical view of the failure zone: there is no identification of the failure zone as seen from above. No proper analysis can be performed without this aspect being taken into account. This is because failure was caused by excessive stress. Without the shear zone from above being known, we cannot calculate stress which is a two dimensional unit (force per area).

I believe that if you take into account the failure zone as seen from above, your analysis of the failure zone you identify as seen from vertical aspect will substantially change.

 

[Tomfh]

The above sketch is relevant if we want to assess why the two PT rods could have different strains

Saikee, how come your failure surfaces are horizontal and vertical? Only limited portion of the actual surface was horizontal and vertical. Most were classic cone shaped cracks...

Maybe I’m missing something. Not sure. The previous green crack sketch seems a more accurate reflection of actual failure planes.

 

[saikee119]

FortyYearsExperience (Structural),

My spreadsheet describes the area of each failure surface. This is the detachment of the deck from 11/12.

The information is based on the cracked on condition prior to the collapse and the failure mode afterward.


Tomfh (Structural)

I notice some cone failure at the rear of diaphragm but that is likely tensile failure due to the principle stress at play.

The failure B1 did not occur as a perfect vertical shear but spreading 45 degree out in the field. The purpose of the sketch is to show what amount of concrete left to resist shear. How the principle stresses changes the failure pattern is just the result of the shear failure. It is the same when crushing a concrete cylinder it can often fail by shear at 45 degree and not by compressing the concrete to disintegration.

The failure surfaces are important for us to count the rebar participating the resistance. FIGG in his presentation had disregarded the contribution from the concrete and concentrated on the contribution from the rebar. FIGG's assumed total failed surfaces is much larger than mine.

In the RC design against shear we used the concrete surface parallel to the shear and count the cross sectional areas of the concrete and rebar across the shearing face. The actual shearing pattern can be different and may not resemble the design.

 

[Settingsun]

Saikee, a curious phenomenon with expert witnesses is that they tend to be on the side of the person paying them. If someone wants money from FDOT, they'll find an expert who'll say there was a problem in what FDOT did.

I assume USA has a duty of care for engineers, and that FDOT has authority over road opening/closure. If you can show FDOT had structural concerns but allowed the road to be opened to traffic anyway, the water is muddied. Then the question of settlement cost vs litigation cost trumps 'right' and 'wrong'.

 

[Tomfh]

I notice some cone failure at the rear of diaphragm but that is likely tensile failure due to the principle stress at play.

Oh, I see. You are treating it as a shear cross section akin to shear in a beam, and not really looking at the actual failure surface.


To my mind the cone failure is the critical feature. The cone snuck around all the critical PT reo, so I dont' believe shear theory can be applied. It's more akin to an unreinforced punching/cone failure.

 

[jrs_87]

We need saikee119's 2D drawings converted to 3D interactive drawing.

 

[Sym P. le]

TITLE: Door Closers vs. Door Stops

Consider there is no effective resistance to shear at saikee119's plane B. The fracture merely jumps over to the longitudinal deck PT cable interface against the back of the 4" sleeves (OSHA pg 97, Figs, 64/65). The result is that there are no shear planes of any capacity in play, the model becomes one of a giant torque load between the base of 12 and the diaphragm/deck duo, and 11 becomes a giant point load on 12 just above the torque application.

This model anticiptes movement from the outset and builds load on 11 through to failure. Any rebar in the system merely acts as a slow release dampening device like a poorly adjusted door closer with big kick in the ass on the clasp setting. There is no door stop. The only good news is that in this instance you have at least five days to figure it out.

 

[saikee119]

steveh49 (Structural),

I agree if FDOT had a safety concern or knew about the risk but did not act then it will be culpable.

FIU bridge isn't a FDOT project so its involvement is on the as-needed basis.

The cracks reported by FIGG to FDOT by the well known voicemail on Mar 13 however assure the recipient party no safety issue.

There is no evident that the bridge was in imminent collapse until FIGG instructed MCM to re-tension the bridge. On record the bridge collapsed just before the re-tensioning operation could complete the very last step.

I am not sure how much FDOT know about the problem until it sent Afredo Reyna to attend FIGG's presentation on Mar 15. The OSHA photos were essentially BPA's work. BPA was employed by FIU and may not even send out its photo collection to every party. Mar 15 would be the key date for others to learn about the crack problems.

FIGG's presentation was designed to assure everything was under control and no need to worry. Under such atmosphere I doubt if FIGG would allow BPA to distribute the crack photos to undermine its assertion. SO it would be useful if we know exactly what information the attendees got on Mar 15. OSHA report mentioned only three photos of Fig 24, 25 & 26 were used by FIGG's presentation.

The curious feature of the Mar 15 meeting both FIGG and FDOT were not interested in seeing the re-tensioning work, apparently ordered by FIGG on Mar 13 and due to commence after the site meeting had finished. OSHA described the FDOT engineer wasn't a structural guy who did seek advice from Tallahassee FDOT office on the re-tensioning operation as well as on FIGG's presentation but didn't get the right person, who did FIU bridge reviews, for help.

From the Mar 15 attendees BPA's JOse Morales and Carlo Chapman and MCM's Pedro Cortes were the only ones recorded by OSHA went to witness the re-tensioning operation.

The story could be entirely different had Tom Andres, who was the one predicting the shearing cracks, attended the Mar 15, 2019 meeting.

 

[jrs_87]

Does this concept make sense? > When concrete is under increased compression (PT bar), can we think of it has being harder? So with taut PT bars member 11 between dead and live plates was similar in hardness to deck and forces penetrated to dead end. When bars went slack, the forces from the softer 11 concentrated above the harder deck interface. With this in mind, what would re-tension do?

Also, unrelated, please explain tributary area (

https://www.eng-tips.com/search.cfm?q=tributary&action=search)

Tributary Area (

http://www.engineersdaily.com/2014/05/how-loads-flow-through-building.html)

The tributary area is related to the load path, and is used to determine the loads that beams, girders, columns, and walls carry. The reader is expected to be familiar with the concept of tributary area from other design courses, as it also applies to design of timber and steel structures; however, a brief overview is presented in this section. The tributary area for a beam or a girder supporting a portion of the floor is the area enclosing the member and bounded by the lines located approximately halfway between the lines of support (columns or walls), as shown in Figure 4. For example, a tributary area for the reinforced concrete beam AB that is a part of the one-way floor system is shown hatched in Figure 4a. A typical column has a tributary area bounded by the lines located halfway from the line of support in both directions (shown hatched in Figure 4b). In the case of uniformly loaded floors, tributary areas are approximately bounded by the lines of zero shear, that is, the lines corresponding to zero shear forces in the slabs, beams, or girders supported by the element for which the tributary area is determined. Zero-shear locations are generally determined by the analysis. For buildings with a fairly regular column spacing, the zero-shear locations may be approximated to be halfway between the lines of support.
Figure :4 Tributary area for reinforced concrete members: a) beams; b) columns.
Figure :4 Tributary area for reinforced concrete members: a) beams; b) columns.
​

 

[saikee119]

Sym P. le (Mechanical),

I know you have been preoccupied with the joint rotation at the support.

Even the concrete has sheared off the joint could be kept going by the rebar crossing the shearing face. The rebar could be acting as dowels then.

The collapse was likely by the failure of the rebar. Some could have sheared off completely. Other may lost the concrete surrounding the steel bar if there was insufficient concrete around or the embedded length was too short.

 

[jrs_87]

saikee119 (Structural)30 Jun 19 01:26

FDOT authority on bridge can be inferred by the fact FDOT vest wearing personnel were on the bridge immediately after collapse. (Possible lawyer talking point)

http://bridgecollapselawyer.com/

https://www.gettyimages.com/detail/...he-florida-news-photo/934372508?adppopup=true

 

[saikee119]

jrs_87 (Mechanical),

The way I explain one PT rod may experience different hardness in the tensioning is the lower rod was secured to the deck already highly compressed. The upper rod on the other hand was physically outside the deck and diaphragm if you take a look at the first sketch of my 28 Jun 19 16:58 post.

Both rods compressed axially Member 11 but the lower rod was additional deck to compress. Therefore by applying the same tension to both rods the axial shortening in the lower rod has to be less than the upper rod. The difference in strains means the two rod anchors distance were progressively changed.