The view from the top of the FIU pedestrian bridge days before it collapsed
The Florida International University pedestrian overpass that collapsed catastrophically during construction last week was advertised as a “cable-stayed” bridge, but it actually wasn’t: Instead, it was a modern take on an old bridge design that experts say is vulnerable to failure when a single structural piece is damaged.
And now experts unaffiliated with the project increasingly are focusing on that as a possible cause for last week’s pedestrian bridge collapse, which killed six people on the roadway beneath it.
While it will likely be months before federal investigators lay out an official explanation, growing available evidence suggests that failure of a single diagonal support truss at one end was enough to cause the entire 174-foot concrete span to buckle under its own weight and crash onto Southwest Eighth Street near the entrance to Florida International University.
What now seems certain from video captures and enlarged photos of the collapse and its aftermath: Crews for a bridge subcontractor were working on top of the unfinished bridge to tighten steel rods inside a diagonal support truss at the north end of the span — possibly to address cracking that had appeared in the concrete — when the structure abruptly gave way.
The question now is, why would the concrete truss fail? Bridge engineers not involved with the project have said overtightening of rods or cables that provide tensile strength in concrete slabs and beams can cause the pieces to twist and shatter abruptly.
But an anonymous Canadian engineer or contractor has posted a crowd-sourced You Tube video exploring the cause of the collapse that has drawn more than half a million views.
The video, uploaded Friday under the handle AvE and spotlighted by Miami New Times on Monday, posits the structural piece was damaged while the prefabricated bridge was being moved into place on Saturday, March 10. When crews tried to tighten the damaged truss on March 15, it gave way suddenly as the support rod inside snapped, the video suggests.
Those angled, vertical support pieces were designed to tie the bottom pedestrian deck of the bridge to a canopy that ran along the top in what’s known as a truss design. In through-truss bridges, the top and bottom pieces — referred to as flanges — and the connecting struts, usually set up in an open cross-pattern or web, work together as a structural system to hold the bridge up.
Steel-truss bridges have been commonly used in roadway construction going back decades. But they have a well-known vulnerability: If a vehicle hits one of the horizontal support trusses, the entire span can collapse.
That’s what happened in the 2013 collapse of a 1955 steel-truss bridge over the Skagit River on Interstate 5 near Seattle: A truck carrying an oversized load struck supporting steel struts along one side of a bridge span, which split apart and fell into the river. That’s because there were no backup or redundant structural elements to support the span if one piece failed.
That’s also why truss bridges are mostly avoided in modern road construction, said Ralph Verrastro, a Cornell-trained bridge engineer with Bridging Solutions, a Tampa firm. “A through-truss bridge can catastrophically fail,” Verrastro said.
The FIU bridge was a truss bridge, its designers, the FIGG Bridge Group, confirmed after the collapse. Many have assumed it was a suspension bridge because renderings of the finished structure show a mast with pipes or cables connecting from its tip to the bridge in a sail-like pattern. Observers, including some engineers, have posited that had the mast been in place, the bridge might not have collapsed.
But in fact the mast would have provided no vertical support. FIGG advertised that as a “cable-stayed” bridge, and plans and other materials on the FIU website say the mast was there mostly to dampen vibration, provide some “stiffness” and create dramatic aesthetics.
After studying the engineering drawings on the website with colleagues, Verrastro confirmed the mast had no role in holding up the bridge.
“They definitely didn’t need it,” he said. “It’s there mostly for looks.”
Because the FIU bridge would see no motorized traffic that could strike a truss, the decision to go with that design by itself raises no red flags, Verrastro said.
But he did note something unusual in its approach: Truss bridges usually have vertical struts running along both edges of the deck. But the FIGG design had only a single row of trusses along the center of the bridge in a zig-zag pattern.
Verrastro said that single row might have made the design especially vulnerable. If one diagonal piece failed, it’s didn’t have a twin on the other side that might have provided some support in the same area.
“The unique part is, it’s a single truss,” he said. “When one member failed, there was no redundancy.”
Another decision by engineers or contractors that might have played a role, he said: The completed span would have included another bridge piece to the north over a canal that connected to the town of Sweetwater. That piece, to be built in place, would have provided additional strength and stability to the main span because the two would have been tied and cemented together, Verrastro said.
He also pointed to one possibly significant difference between the FIU bridge and a traditional truss bridge: It was made of heavy concrete, not much lighter steel. That decision is justifiable because concrete is far easier and less expensive to maintain than steel, looks better and lasts far longer. FIGG advertised the structure as “a 100-year bridge,” and Verrastro said that’s likely accurate.
He said that neither the length of the span nor its 950-ton weight should be an issue, because its design engineers would have ensured it had enough tensile strength to be self-supporting, with support pylons only at either end.
That would explain why there were no temporary supports while the bridge was finished, and could also explain the decision not to close the Tamiami Trail while workers tightened rods or cables inside the concrete, he said. Unless there was visible or measurable sagging of the structure, indicating damage, such tensioning strengthens the bridge.
In engineers’ minds, Verrastro said, “there should not be an issue, because you’re just making it stronger.”
But what the engineers may not have known, the faceless Canadian YouTuber suggests, is that the steel inside the struts could have been damaged. His evidence: photos showing a steel rod protruding from the top of the bridge canopy with a blue hydraulic jack — equipment used in tensioning support cables or rods inside concrete — still attached. A subcontractor crew member working on that spot fell to his death when the bridge collapsed.
In the video, the Canadian runs a demonstration showing how a steel rod undergoing tensioning will suddenly snap and shoot out of the jack if stressed beyond its capacity.
The video lays out further evidence: Drawings on the FIU website show how the bridge span, which was prefabricated by the side of the trail over a period of months, was to be moved into place using four powerful lift trucks. The technique, known as accelerated bridge construction, significantly reduces the time a road must be closed to traffic.
In the plans, two trucks would be placed side-by-side at either end, lined up under spots where the vertical trusses met in a joint, the Canadian noted. But video of the actual move shows one of the trucks at the north end, where the bridge appears to have failed, was moved farther towards the middle of the span, leaving the end unsupported. The Canadian says in the video that’s because a road divider or dip in the surface would have made it difficult for the truck to be positioned properly.
His source is not clear. But the Associated Press said in a story Tuesday that the Florida Department of Transportation ordered the northern support pylon be moved 11 feet to make room for future expansion of the trail. That required a design change that lengthened the span — and put the support pylon in the dirt well off the edge of the roadway, which could also explain why the northernmost truck could no longer follow its original planned route.
As the bridge was moved, stress points on the structure were constantly monitored with sensors.
But the weight of the unsupported end could have placed enough stress on the last diagonal strut — the one being worked on when the bridge fell — to damage the rod inside it and loosen the tension on it, the Canadian AvE says in his YouTube video. It would also explain cracking that appeared on the north end of the bridge that a FIGG engineer reported to FDOT on Tuesday, although he concluded that it posed no safety concerns. Cracking in new concrete is not uncommon, and could be superficial or a sign of deeper trouble.
On Thursday, the cracking was discussed at a meeting of the project team that included FDOT, and again no safety flags were raised. But NBC 6 investigative reporter Tony Pipitone reported Monday that the tensioning work being carried out was a result of the meeting, and that routing tensioning or tuning of the bridge had been concluded on Monday, before the Trail was reopened to traffic. The Miami Herald has requested meeting records from FDOT and FIU but the documents have not yet been supplied.
That night, after the bridge collapse, U.S. Sen. Marco Rubio tweeted that cables holding up the structure “had loosened” and engineers had ordered them tightened. On Friday, National Transportation Safety Board officials confirmed that tensioning work had been going on at the last diagonal strut at the span’s north end, but said it was too early to tell if that was related to the collapse.