Florida International Bridge Collapse

[hardalee]

Last Thursday, a bridge under construction collapsed west of Miami killing 6 and severely engineering others.

The why is a big question with no answer so far.

I have been studying the original proposal, photographs and videos and have come up with the following conclusions so far:

First diagonal at north was in compression. It can be seen failing looking up:" fiu bridge video slow motion".

You can see it start to buckle downward in video. When this member stopped carrying compression load, the next member to the south which was a tension member became a compression member, dropping the vertical gravity load that would have been carried in the failed compression member into the bottom cord of the truss (the walk way) as a downward load. The system at this time was no longer a full truss, but a beam in the area of failure.

This same conclusion can be reached by looking at failure modes of the system and comparing them with the photos after the collapse.

The above assumptions predict the same mode of failure that the videos show.

Why did the first compression member fail? That would take plans to determine the stresses, etc as well as knowing about the state of the tendons and how the bridge was built, but it may have had something to do with the reported loosening of tendons which were said to have been in the process of being re-tightened. The compression load of the diagonal under gravity from the unshored span may have loosened the tendons and required the re-stressing of the tendons.

Any structural engineers here who would care to discuss it?

 

[Peter Dow]

There's an article in New Civil Engineer

Experts cite explosive joint failure as cause of Florida bridge collapse

with some nice diagrams.

Any structural engineers here who would care to discuss it?

I'm looking for a wider discussion, of the social science aspects - public safety, management, legal and political.

I think there are wider lessons for society here than just how not to build a bridge.

So I am thinking of starting another thread on the FIU Bridge Collapse, in another forum - most likely World Events.

Here is the link to the topic I started
Florida International Bridge Collapse - World Event


I wouldn't want to branch out into a wider discussion here in hardalee's topic, considering matters which are off-topic to structural engineering and then be accused of "trolling" with off-topic remarks.

 

[hardalee]

Very nice drawings. I agree with the findings.

Thank you.

 

[Peter Dow]

Very nice drawings. I agree with the findings.

Thank you.

It seems to me however that this article's findings somewhat underplays the culpability of the bridge design engineers for their poor design.

It looks to me like a severely botched design

• 175 feet x 18 feet "mainspan" - ratio of length to height 9.7
• thick heavy concrete deck
• too weak vertical truss
• weak reinforced-concrete poor material choice for truss members and joints
• thicker or steel girder / pipe truss members and effectively revolute joints would have been a lot stronger and better able to cope with varying tensile and compression forces as the mainspan was moved around and into position
• an all-metal bridge - steel or aluminium - would be a better choice for an "accelerated bridge construction"

This a 173 page PDF proposal with full details of what was proposed.
http://facilities.fiu.edu/projects/BT_904/MCM_FIGG_Proposal_for_FIU_Pedestrian_Bridge_9-30-2015.pdf

 

[hardalee]

It seems to me however that this article's findings somewhat underplays the culpability of the bridge design engineers for their poor design.

It looks to me like a severely botched design

• 175 feet x 18 feet "mainspan" - ratio of length to height 9.7
• thick heavy concrete deck
• too weak vertical truss
• weak reinforced-concrete poor material choice for truss members and joints
• thicker or steel girder / pipe truss members and effectively revolute joints would have been a lot stronger and better able to cope with varying tensile and compression forces as the mainspan was moved around and into position
• an all-metal bridge - steel or aluminium - would be a better choice for an "accelerated bridge construction"

This a 173 page PDF proposal with full details of what was proposed.
http://facilities.fiu.edu/projects/BT_904/MCM_FIGG_Proposal_for_FIU_Pedestrian_Bridge_9-30-2015.pdf

 

[hardalee]

Not sure what caused it, but know why it fell. A .6 “ diameter tendon was being stressed (witness the stressing ram still in the tendon) it broke and the first diagonal, that heading south from bottom to top, broke. This member was a compression member, likely stressed to a minimum psi or for the placement where it would have carried tension while being moved only.

See original post for move detail.

A broken tendon would not caused failure in this compression member if it was properly reinforced with mild steel reinforcement.

So, something else was wrong, what I don’t know yet. I look forward to learning how this compression member was reinforced and more about the reported cracking in the area.

Time will tell.

 

[Gawdzilla Sama]

Floridaman is involved in some fashion, I just know it.

 

[Peter Dow]

A .6 “ diameter tendon was being stressed (witness the stressing ram still in the tendon) it broke

According to the engineer's drawings in the proposals the Post Tensioning Bars are specified with a diameter of 1.75".

What makes you think they used bar of only .6" diameter?

 

[Peter Dow]

tendon was being stressed (witness the stressing ram still in the tendon) it broke

tensioning a tendon / bar in truss member "9" in the diagram from New Civil Engineer (NCE) posted above.

and the first diagonal, that heading south from bottom to top, broke. This member was a compression member, likely stressed to a minimum psi or for the placement where it would have carried tension while being moved only.

That's member number "10" in the NCE diagrams. It does NOT have ANY tendons / tensioning bars in it, according to page 115 of the MCM-FIGG proposal PDF.

Note that the numbering of the members is different (number 10 in NCE = number 11 in the MCM-FIGG proposal page 115).

You can see from the detail of the joint that it appears to be reinforced concrete throughout, with zero flexibility and no rotation at the joint possible whatsoever (unless the concrete cracks and crumbles) so strictly speaking such rigid concrete joints do not create a true "truss" because the members are not connected by revolute joints in any way, shape or form, are rigid and therefore any sag or movement in the structure can generate bending moments or torques at these rigid joints leading to shear forces in addition to normal truss tension - compression forces, possibly generating forces which can exceed the material strength of the concrete or P.T. bars.

https://en.wikipedia.org/wiki/Truss

In engineering, a truss is a structure that "consists of two-force members only, where the members are organized so that the assemblage as a whole behaves as a single object".[1] A "two-force member" is a structural component where force is applied to only two points. Although this rigorous definition allows the members to have any shape connected in any stable configuration, trusses typically comprise five or more triangular units constructed with straight members whose ends are connected at joints referred to as nodes.

In this typical context, external forces and reactions to those forces are considered to act only at the nodes and result in forces in the members that are either tensile or compressive. For straight members, moments (torques) are explicitly excluded because, and only because, all the joints in a truss are treated as revolutes, as is necessary for the links to be two-force members.​

 

[hardalee]

tensioning a tendon / bar in truss member "9" in the diagram from New Civil Engineer (NCE) posted above.


That's member number "10" in the NCE diagrams. It does NOT have ANY tendons / tensioning bars in it, according to page 115 of the MCM-FIGG proposal PDF.

Note that the numbering of the members is different (number 10 in NCE = number 11 in the MCM-FIGG proposal page 115).

You can see from the detail of the joint that it appears to be reinforced concrete throughout, with zero flexibility and no rotation at the joint possible whatsoever (unless the concrete cracks and crumbles) so strictly speaking such rigid concrete joints do not create a true "truss" because the members are not connected by revolute joints in any way, shape or form, are rigid and therefore any sag or movement in the structure can generate bending moments or torques at these rigid joints leading to shear forces in addition to normal truss tension - compression forces, possibly generating forces which can exceed the material strength of the concrete or P.T. bars.

https://en.wikipedia.org/wiki/Truss

In engineering, a truss is a structure that "consists of two-force members only, where the members are organized so that the assemblage as a whole behaves as a single object".[1] A "two-force member" is a structural component where force is applied to only two points. Although this rigorous definition allows the members to have any shape connected in any stable configuration, trusses typically comprise five or more triangular units constructed with straight members whose ends are connected at joints referred to as nodes.

In this typical context, external forces and reactions to those forces are considered to act only at the nodes and result in forces in the members that are either tensile or compressive. For straight members, moments (torques) are explicitly excluded because, and only because, all the joints in a truss are treated as revolutes, as is necessary for the links to be two-force members.​

.6 inches diameter is what was seen at the site as well as specified in the proposal drawings. 1 3/4 bars are not correct though noted, though I cannot exclude .5 inch diameter bars. Also anchor type seen is for strands, not large bars. I assume the compression member was tensioned though it does not show as such on the proposal drawings. Either for nominal PT stress in a member or possibly for transportation which would have reversed the failed member to tension until is was placed on the columns.

We know the diagonal was stressed since we can see the failed PT tendon coming out of the blister on top of the canopy in photos where it was stressed and the PT jack is still attached. Man doing the stressing died in the collapse. He can be seen in the video. Crane was likely there as a man box to lift him in place.

 

[hardalee]

Floridaman is involved in some fashion, I just know it.

Good call.

 

[Peter Dow]

According to the engineer's drawings in the proposals the Post Tensioning Bars are specified with a diameter of 1.75".

What makes you think they used bar of only .6" diameter?

.6 inches diameter is what was seen at the site as well as specified in the proposal drawings. 1 3/4 bars are not correct though noted, though I cannot exclude .5 inch diameter bars. Also anchor type seen is for strands, not large bars.

Can you show me ".6 inches" because I can show you 1.75 inches or 1 3/4" on the proposal drawings?

The following images are taken from page 115 of pdf file at the link I already gave but here it is again.
http://facilities.fiu.edu/projects/BT_904/MCM_FIGG_Proposal_for_FIU_Pedestrian_Bridge_9-30-2015.pdf

Here's the whole of page 115.

I've added a graphic, page number and the file URL but that's it.

Now where is your

".6 inches diameter is what was seen at the site as well as specified in the proposal drawings. "​

Because I am showing you 1.75" but you are not showing me 0.6".

You talk a good game there hardalee but it is cards-on-the-table time.

 

[hardalee]

Can you show me ".6 inches" because I can show you 1.75 inches or 1 3/4" on the proposal drawings?

The following images are taken from page 115 of pdf file at the link I already gave but here it is again.
http://facilities.fiu.edu/projects/BT_904/MCM_FIGG_Proposal_for_FIU_Pedestrian_Bridge_9-30-2015.pdf

Here's the whole of page 115.

I've added a graphic, page number and the file URL but that's it.

Now where is your

".6 inches diameter is what was seen at the site as well as specified in the proposal drawings. "​

Because I am showing you 1.75" but you are not showing me 0.6".

You talk a good game there hardalee but it is cards-on-the-table time.

Try drawing b-2 in what you have quoted.

Thought you were real. This place is a waste of time.

Best to all.

I’m gone.

 

[Peter Dow]

I assume the compression member was tensioned though it does not show as such on the proposal drawings. Either for nominal PT stress in a member or possibly for transportation which would have reversed the failed member to tension until is was placed on the columns.

We know the diagonal was stressed since we can see the failed PT tendon coming out of the blister on top of the canopy in photos where it was stressed and the PT jack is still attached. Man doing the stressing died in the collapse. He can be seen in the video. Crane was likely there as a man box to lift him in place.

Well here is what was said in a thread in news.slashdot.org

The Ordinary Engineering Behind the Horrifying Florida Bridge Collapse

McGruber on Saturday March 17, 2018 @11:48AM
replied to
Anonymous Coward on Saturday March 17, 2018 @09:28AM

"Mod AC up!

I'm a Professional Engineer, though not licensed in Florida nor am I an expert in concrete bridges. Based upon the pictures of the debris I've seen, the bridge was made using prestressed concrete. Prestressed concrete is an amazing material - the steel reinforcement inside the concrete is used to compress the concrete, which causes the concrete member to act like high strength steel when it is loaded with tension forces.

AC's comment above explains what happened:

In preliminary drawings, #11 is shown with no post-tensioning bars, but the actual construction shows it with two. While those bars in #11 may have been necessary due to the move, since the ends of the bridge were cantilevered (which is different than shown in the preliminary drawings), they likely weren't needed after placement; not needed to be post-tensioned, since #11 would be in high compression.

It appears workers were post-tensioning #11 using a crane and other equipment attached to one of the post-tension rods. It appears tensioner (blue) and part of the bar is sticking out several feet in photos of the collapse. According to some, this likely lead to the collapse.
​

In its final placement as a bridge, truss member #11 would be in compression, so there was no reason to prestress it. Per AC, the preliminary design drawings did not show the two post-tensioning rods in this member.

My guess, based upon AC's post, is that when the builders decided to assemble the bridge using the accelerated technique, the designers realized that there would be tension forces on truss member #11 during the move.... so the design drawings were changed to add the two post tensioning rods to truss member #11.

Once the bridge was in place, the construction workers evidently began tightening the two post tensioning members even more. Member #11 was already being loaded with compression forces, from the dead weight loading of the bridge.... and tightening the post tensioners would have placed more compressive loading onto member #11. Once the combined compressive loadings (from the dead weight and the tensioning) exceeded the compressive strength of the concrete, the concrete would fail (it makes a loud popping sound when it fails in lab tests) and the single, non-redundant truss would fail.

If my speculation is correct, it will be interesting to see whether the Figg, the engineering firm that designed the bridge, or the construction contractor gets blamed."
​

- which sounds plausible as a sequence of events but I have to say that this case proves this assumption wrong -

"Prestressed concrete is an amazing material - the steel reinforcement inside the concrete is used to compress the concrete, which causes the the concrete member to act like high strength steel when it is loaded with tension forces." ​

Concrete truss members DO NOT IN ANY WAY WHATSOEVER "act like high strength steel when it is loaded with tension forces".

A high strength steel truss member (say a heavy steel pipe) would NEVER have had to be tensioned with PT bar but if for some bizarre reason someone tried to do it (why would they though?) then either the tensioning bar would break but the steel pipe would be fine without it so who cares, (or I suppose maybe the pipe might buckle if it had a too thin pipe wall.)

So high strength steel truss members are superior in every way except cost to prestressed concrete truss members (and maybe fire resistance too admittedly).

That said you could easily have a high strength steel truss and embedded it within concrete as per sky-scraper super structures to get both high strength and fire resistance.

Cost-cutting and opting for cheaper prestressed concrete when high strength steel should be used, will be the death of us.

 

[Peter Dow]

Try drawing b-2 in what you have quoted.

Thank you kindly for showing me what you meant.

On Sheet B-2, page 100, the prestressing strand diameter is specified as "0.6 INCH".

Whereas on Sheet B-17, on page 115, the post-tensioning bar diameter is "All P.T. BARS ARE 1 3/4" DIA BARS"

I have illustrated this information in this graphic

Now I trust that we both can see where the other got his figure from.

Thought you were real.

I am real and I really didn't mean to offend you. Sorry.

This place is a waste of time.

Well you didn't think so for 350 messages.

Best to all.

I’m gone.

Well maybe when you cool off you will come back? I hope so.

 

[DaveC426913]

Not an engineer. Use small words plz.

Was this a single point of failure that cascaded? Was there no redundancy?

Surely there must be multiple design flaws and insufficient redundancy for it to fail so rapidly and catastrophically?

 

[hardalee]

Thank you kindly for showing me what you meant.

On Sheet B-2, page 100, the prestressing strand diameter is specified as "0.6 INCH".

Whereas on Sheet B-17, on page 115, the post-tensioning bar diameter is "All P.T. BARS ARE 1 3/4" DIA BARS"

I have illustrated this information in this graphic

Now I trust that we both can see where the other got his figure from.


I am real and I really didn't mean to offend you. Sorry.


Well you didn't think so for 350 messages.


Well maybe when you cool off you will come back? I hope so.

 

[hardalee]

Ok, sorry. Been fighting since it went down. Folks are just starting to tnderstand down here. Bad times.

We still need to know why it failed. We know how, but not why.

Civil discussion is always welcome. Maybe we could learn.

I’ve already calculated tendon losses due to elastic shortening after loading, 10% nothing. Looking at buckling but meaningless unless I have messed up the units. Tomorrow will check.

The cracking in the area may mean something. Need more info. Mainly the mild longitudinal reinforcement and ties. They are are not yet providing that info.

To sleep now. Perchance to dream.

Sorry but I’m sometimes just a bad dog.

 

[Peter Dow]

Well here is what was said in a thread in news.slashdot.org

The Ordinary Engineering Behind the Horrifying Florida Bridge Collapse

McGruber on Saturday March 17, 2018 @11:48AM
replied to
Anonymous Coward on Saturday March 17, 2018 @09:28AM

"Mod AC up!

I'm a Professional Engineer, though not licensed in Florida nor am I an expert in concrete bridges. Based upon the pictures of the debris I've seen, the bridge was made using prestressed concrete. Prestressed concrete is an amazing material - the steel reinforcement inside the concrete is used to compress the concrete, which causes the concrete member to act like high strength steel when it is loaded with tension forces.

AC's comment above explains what happened:

In preliminary drawings, #11 is shown with no post-tensioning bars, but the actual construction shows it with two. While those bars in #11 may have been necessary due to the move, since the ends of the bridge were cantilevered (which is different than shown in the preliminary drawings), they likely weren't needed after placement; not needed to be post-tensioned, since #11 would be in high compression.

It appears workers were post-tensioning #11 using a crane and other equipment attached to one of the post-tension rods. It appears tensioner (blue) and part of the bar is sticking out several feet in photos of the collapse. According to some, this likely lead to the collapse.
​

In its final placement as a bridge, truss member #11 would be in compression, so there was no reason to prestress it. Per AC, the preliminary design drawings did not show the two post-tensioning rods in this member.

My guess, based upon AC's post, is that when the builders decided to assemble the bridge using the accelerated technique, the designers realized that there would be tension forces on truss member #11 during the move.... so the design drawings were changed to add the two post tensioning rods to truss member #11.

Once the bridge was in place, the construction workers evidently began tightening the two post tensioning members even more. Member #11 was already being loaded with compression forces, from the dead weight loading of the bridge.... and tightening the post tensioners would have placed more compressive loading onto member #11. Once the combined compressive loadings (from the dead weight and the tensioning) exceeded the compressive strength of the concrete, the concrete would fail (it makes a loud popping sound when it fails in lab tests) and the single, non-redundant truss would fail.

If my speculation is correct, it will be interesting to see whether the Figg, the engineering firm that designed the bridge, or the construction contractor gets blamed."​

- which sounds plausible as a sequence of events but I have to say that this case proves this assumption wrong -

"Prestressed concrete is an amazing material - the steel reinforcement inside the concrete is used to compress the concrete, which causes the the concrete member to act like high strength steel when it is loaded with tension forces." ​

Concrete truss members DO NOT IN ANY WAY WHATSOEVER "act like high strength steel when it is loaded with tension forces".

A high strength steel truss member (say a heavy steel pipe) would NEVER have had to be tensioned with PT bar but if for some bizarre reason someone tried to do it (why would they though?) then either the tensioning bar would break but the steel pipe would be fine without it so who cares, (or I suppose maybe the pipe might buckle if it had a too thin pipe wall.)

So high strength steel truss members are superior in every way except cost to prestressed concrete truss members (and maybe fire resistance too admittedly).

That said you could easily have a high strength steel truss and embedded it within concrete as per sky-scraper super structures to get both high strength and fire resistance.

Cost-cutting and opting for cheaper prestressed concrete when high strength steel should be used, will be the death of us.

Actually what may have happened is that the 0.6" prestressing strands (which were never intended by the design to be used as 1.75" post tensioning bars) which were originally specified to be in member #11 (number 10 in NCE = number 11 in the MCM-FIGG proposal page 115) as per the proposal, were used inappropriately to post-tension member #11 to the point of failure of the strand, member #11 and the entire mainspan?

In which case there was NO change to the design drawings and NO addition of post-tensioning bars to member #11.

Presumably, someone thought that the prestressing strands on member #11 needed post-tensioning for some reason.

So someone said - "Hey let's post-tension these 0.6 inch strands / bars in member #11 - what could go wrong?"

 

[Peter Dow]

Not an engineer. Use small words plz.

Was this a single point of failure that cascaded? Was there no redundancy?

Looks like it. By simply over-jacking 0.6 inch prestressing strands, perhaps mistaking them for post-tensioning bars (1.75 inch), the strand failed, the truss member failed and the mainspan failed.

Surely there must be multiple design flaws and insufficient redundancy for it to fail so rapidly and catastrophically?

Agreed.