City Hall

Calatrava Bridge Test Failure Prompts a Round of Finger-Pointing and Blame-Casting

The hike and bike portions of the Margaret McDermott Bridge remain closed because three elements cracked after construction was completed in 2016. The paths, which are separate from the highway, were scheduled to open during the summer.
The hike and bike portions of the Margaret McDermott Bridge remain closed because three elements cracked after construction was completed in 2016. The paths, which are separate from the highway, were scheduled to open during the summer. Hannah Ridings

Two cable-stayed structures across the Trinity River
, collectively called the Margaret McDermott Bridge, remain closed to traffic for one reason and one reason alone: No one is willing to certify the structures intended for use by hikers and bicyclists are safe for human beings. The difficulty is that no one involved wants to admit  that human safety is the problem.

The city even says safety is not the problem, that the bridge has passed key tests and will be open soon. Last week after the Dallas Observer reported on structural problems, Sarah Standifer, the city official overseeing the bridge project, told television cameras the only problem was a long-range maintenance issue.

Maybe, if she meant long-range maintenance of human life.

The Observer reported that the bridge remains closed months after its scheduled opening because three key elements of the bridge had cracked in high winds soon after basic construction was completed in 2016. In response to that story, a spokesperson for Dallas City Manager T.C. Broadnax sent the Observer an upbeat memo citing "field wind tests, field strength tests, lab simulated testing, addition of dampeners [weights] to cables to essentially stop wind induced vibrations ... vibration study with positive results" and so on. Broadnax sent council members a memo saying the city has spent $312,000 on various tests and remediation to fix problems with the bridge.

All dodgeball. One test counts. That test was not done before the bridge was built, even though the designer of the structures, Spanish architect Santiago Calatrava, pleaded insistently with the city to get it done after the city cut corners on construction costs. The test still wasn't done months after Calatrava offered to lend the city the money to get it done.

click to enlarge
Spanish architect Santiago Calatrava offered to lend the city of Dallas money to test the bridge.
Cati Cladera/EFE/Newscom
There seems to be an obvious reason it was not done sooner. On Jan. 11, 2017, Standifer wrote to Calatrava's chief engineer on the project, Iain Rowe in Zurich, Switzerland, to say, "Based on preliminary data, the city is concerned that the current design will fail the tests that you insist are necessary."

A trove of correspondence uncovered by Dallas City Council member Scott Griggs shows the Calatrava people over a two-year period refusing to back off demands that the basic structural soundness of the bridge be tested before the bridge was opened to pedestrians and cyclists. Finally, at the end of last year, the key bridge parts were subjected to the required test in Kirkel, Germany. They failed.

In a structure that invites the public to walk and bike on a platform suspended high in the air from cables, human safety obviously is the first consideration. But in this case, safety is not the only concern. In a stubborn drumbeat in almost two years of correspondence, the city of Dallas insists that the Texas Department of Transportation keep its construction contractor on the job until the cable problem is solved, and each time in response, TxDOT issues the same somber warning to the city, almost word for word in every letter:

"It is currently estimated that the monthly contractor overhead for the project is approximately $1 million per month. ... Per the Advanced Funding Agreement signed by the city on July 16, 2012, the city of Dallas is responsible for any project cost overruns associated with the Margaret McDermott Bridge and therefore will be responsible for the cost of the schedule impacts and remedial work."

The bridge project is now approximately eight months beyond the original announced opening. The problems with the cables occurred because the city saved $30,000 by scratching a planned fatigue test called for by the architect. Now what all of the parties want is a way out. 

The Dallas engineering firm serving as the city's chief engineer, Huitt-Zollars, has proposed just that. Huitt-Zollars, which must certify the safety of the bridge before it can be opened, suggested in a letter to the city that maybe this one bridge should no longer be required to pass the same universal safety test that every other similar bridge must pass before being declared safe for human traffic.

"We understand that meeting these requirements could have negative consequences for many parties," Huitt-Zollars Vice President Charles E. Quade says in a Dec. 22 letter to Standifer. He tells Standifer his firm has searched for other tests the bridge could pass, but he says the firm has "found no established criteria" other than the standard test.

But Quade makes Standifer a startling offer anyway. He proposes a way around the fact that the Margaret McDermott Bridge cannot pass the standard test.

"We could consider alternative probabilistic methods be utilized to determine a project specific fatigue criteria," Quade says.

A "project specific" test would be one created for the Margaret McDermott Bridge alone. A probabilistic test would be a mathematical computation, as opposed to what industry standards call for — putting parts of the bridge in a stretching machine and tugging on them millions of times over a period of weeks.

"We understand that meeting these requirements could have negative consequences for many parties." – Huitt-Zollars Vice President Charles E. Quade

tweet this

On hearing of the Huitt-Zollars offer to do a one-off probabalistic test for the McDermott, Griggs, in whose district the bridge stands, asked, "What will they do after it fails the probabilistic test? A possibilistic test? If it fails that one, will they do an improbabilistic test?"

I wrote to Huitt-Zollars and the city Jan. 31 and asked the following questions about the one-off written test for the bridge: "Why should the city's citizens not be concerned that this may amount to a one-off fudging of the standard of safety for the bridge to avoid 'negative consequences for many parties?' Can you tell me why the bridge, having failed in fact and then failed again in testing, should be held to what may be a lower standard rather than be rebuilt to something known to be a higher standard?"

Quade at Huitt-Zollars wrote me back: "Our understanding is our client and the City is preparing a response to your request. We have no further comments."

Monica Cordova, a spokesperson for Broadnax, wrote me: "The City is relying on the expertise of the engineer of record [Huitt-Zollars].

"Based on their professional opinion, it is prudent to perform the additional study that is being recommended," Cordova said. "The engineer of record has provided information related to the lack of specific testing protocols for this type of pedestrian bridge, has performed additional field work that informs their analyses, and awaits the conclusion of the study. This study will determine the long-term maintenance activities related to the cable system."

She concluded, "We are taking all steps to make sure the bridge is safe and functional."

That problem — safety and function — is not complicated. The McDermott is a signature Calatrava bridge design in which the deck hangs from arches on cables. If the cables or the anchors attaching them to the deck fail, the deck falls.

The McDermott consists of two independent giant arches standing on either side of, but not connected to, two standard concrete highway bridges. Each arch carries a narrow pedestrian and bicycle deck.

Last year, soon after the decks were completed, elements of the cables began to crack in high winds. The specific part of the cables that cracked, called an adjustment rod, is a kind of turnbuckle for tuning the tension on the cables.

When the city put out the bridge design for bids six years ago, the offer from the construction contractor used by TxDOT for all downtown freeway projects came in $12 million higher than available city funds. In all, the Calatrava elements at the sides of the state highway bridge had been expected to add just over $90 million to the overall cost of replacing the Interstate 30 bridge across the Trinity. When the bids came in higher than that, the city asked bidders to propose ways to cut costs.

In a long list of cost-cutting measures amounting to $3.8 million in savings, the contractor included dropping fatigue tests for the cable arrays for the pedestrian and cycling bridges, which would have cost $30,000. The contractor said tests already had been done for equivalent arrays elsewhere and that data from those tests could be used to qualify the McDermott as safe without doing a fresh test.

The city agreed to dropping the fatigue tests with the provision that the contractor would come up with the prior data to prove up the bridge without a new test, but that data never arrived.

The question why the previous test data was never provided is one that lawyers probably will have to settle. Knowledgeable parties, who agreed to discuss this project with the Observer on the condition they weren't named, did not contradict a strong suggestion running all through Griggs' trove of correspondence: No previous data was available for this bridge because this design was altered too much to be equivalent to previous bridges.

Fault for not getting the testing done and for the lack of data is bandied back and forth bitterly between all parties in the stack of correspondence — Calatrava, the city, Huitt-Zollars, TxDOT and the TxDOT construction contractor. At one point or another, each accuses another of being at fault. The tense recriminations are a stark contrast with briefings given to the City Council by city staff in the same period, always upbeat and full of broad assurances that the project was flying along without a hitch.

The one thing all parties agreed on was the test that the cable assemblies were supposed to be able to pass — a metal fatigue test prescribed in a document from the Post-Tensioning Institute called Recommendations for Stay Cable Design, Testing and Installation.

That test, which the cables ultimately failed, involves two things — one called fatigue, like what happens when you bend a paper clip enough times to make it break, and the other called angular deviation, which is more like a waggling or twisting of the metal rather than straight bending. The turnbuckles on the McDermott cables, called adjustment rods, are shorter and of a thinner diameter than what was called for in the original Calatrava specifications and may move differently than the size Calatrava had called for.

"The city does not understand how any related construction was allowed to proceed without the submittal and Design Team [Calatrava] approval of previous testing data, and/or actual test results." – Sarah Standifer, Dallas project overseer

tweet this
In the test that the adjustment rods failed in Germany at the end of last year, the rods were clamped at both ends and hung vertically, then tugged hard from both ends and also twisted slightly millions of times over a period of weeks. The test rods did not hold up.

Presumably, that's what happened to the three failed rods on the bridge last year. They bent. They waggled. They cracked.

An important thing to remember is that the rods on the bridge failed at a time when there was no traffic on the hike and bike decks. Those elements of the McDermott have not yet been opened to traffic. In the correspondence, all parties seem to agree that the single powerful force that wore out the rods so quickly was high wind, buffeting the decks and twisting the cables.

Yet in proposing a new kind of one-off test for the McDermott, Huitt-Zollars makes no mention of wind. The Huitt-Zollars letter to Standifer says the existing PTI test may be too tough on the McDermott because the PTI standard is designed for bridges with trucks and cars going across them, not walkers and cyclists:

click to enlarge
The turnbuckles on the McDermott cables, called adjustment rods, are shorter and of a thinner diameter than what was called for in the original Calatrava specifications.
Hannah Ridings
"The PTI standard was developed for highway bridges and stay cable assemblies, not specifically pedestrian bridges and structural bridge strand assembles," Huitt-Zollars tells the city in its letter.

The knowledgeable sources who spoke to the Observer suggested Huitt-Zollars may not be all the way out in left field, however, for suggesting that some new testing regimen might be appropriate. Since the failure of the three rods and at the instigation of Calatrava and Huitt-Zollars, TxDOT's contractor has embarked on a campaign of installing weights, called dampers, on the cables on the bridge to calm them down in high winds.

That fix appears to be working. The theory and the hope are that the dampers will keep the cables from jumping around so much, less twisting in the cables will contribute less stress to the rods, the rods will not experience metal fatigue so soon and the rods will not crack.

It's a great theory. But that's all it is. If the bridge opens under that theory, then the real test will be time. If all goes well and the bridge passes the test of time, everyone will be happy. If it fails, the outcome could be catastrophic.

One measure of the seriousness of the testing issue is the rising tone of urgency in the correspondence as the various parties realize over time that the rods, which are failing, were never tested.

On Sept. 8, 2016, the project manager for TxDOT writes Quade at Huitt-Zollars: "As stated in my letter dated August 24, 2016, the final solution to the cable adjustment rod issue is still unknown."

In a second letter dated the same day, TxDOT says, "TxDOT has determined the contractor could not have anticipated including angular deviation in the cable fatigue testing based on the original bid documents."

In a Jan. 11, 2017, letter to the Calatrava firm, Standifer tries to put the blame on it: "The city does not understand how any related construction was allowed to proceed without the submittal and Design Team [Calatrava] approval of previous testing data, and/or actual test results."

But in the same letter, Standifer complains that the city fears the bridge will flunk the tests that Calatrava wants to see done. The suggestions seem to be that the Calatrava people should have come up with something that passed some test in the beginning. The problem there is that the bridge, as built, is not the bridge as Calatrava designed and engineered it.

Iain Rowe, Calatrava's chief engineer, fires back a letter a week later: "We were very disappointed to receive this letter, particularly as you hadn't responded to our letter dated 1st November 2016, or my request for a project progress discussion on 14th December 2016."

"Regardless of the TxDOT's opinion on the fatigue testing, the requirements were stated clearly in the original project Specification from 2012." – Iain Rowe, Calatrava's chief engineer

tweet this
Rowe provides in his letter a flowchart depicting nearly three years of inactivity on the city's part concerning the fatigue test. He calls other statements in Standifer's letter unfounded.

"Many of the accusations of your letter are without basis," he says. "Regardless of the TxDOT's opinion on the fatigue testing, the requirements were stated clearly in the original project Specification from 2012."

A month later, Rowe writes Standifer, "We did not understand, and we do not understand, why the Contractor has not yet conducted the specified testing and/or why the contractor is not proceeding with the testing."

In September, Rowe writes Standifer, "This is obstructing the resolution of the problem and necessarily threatens the completion schedule of the bridge."

Finally, in December 2017, Huitt-Ziolars writes Standifer with its new offer of a unique written test to replace the standard empirical tests that the bridge elements recently failed. The paper trail ends there.

The paper trail is public because Griggs maintains a spreadsheet of weekly progress on more than 120 city construction projects in his North Oak Cliff district with the announced completion dates. The McDermott hike and bike lanes were supposed to open at midsummer last year.

As Griggs began to see the opening of the bridge slipping week after week and month after month, he insisted that Standifer explain what was holding it up. Eventually, Standifer referred him to the city attorney. Griggs asked City Attorney Larry Casto for all public record correspondence on the bridge, and Casto complied.

In the unresolved McDermott mess, Griggs sees a theme at play larger than metal stress and fatigue tests. "The overarching theme is trying to be what we're not," he says, explaining that, in his view, the bridge's problems stem from its genesis as a decorative addition to the standard freeway bridge built by TxDOT.

It is a one-off and an anomaly, he suggests, because it is not truly what it purports to be — a kind of grand suspension bridge over a great body of water — and it is not in a place where it belongs. It was a fake and a pretension from conception, Griggs says, and its physical failures are born of those sins.

click to enlarge
The Margaret McDermott Bridge encompasses four structures: eastbound and westbound highway bridges and hike and bike suspension arches at their flanks.
Hannah Ridings
Drive across the Margaret McDermott Bridge westward on I-30 from downtown to Oak Cliff or on the other side in the opposite direction, and you can almost feel what Griggs is saying. Out in the huge, roaring, 11-lane center of the bridge, the delicate arches on both sides — structures that seemed so impressive while approaching from a distance — suddenly feel too small, less like wings of an eagle than flippers on a seal.

If the arches feel like unlikely supports for such a massive structure, it's because they are not reinforcing it. The two Calatrava arches bookending the broad pier-and-beam expressway bridges, one eastbound and the other westbound, play no role whatsoever in supporting them. They stand off from the highway bridges, so that the McDermott really consists of four independent structures — the east and west highway bridges and the two hike and bike suspension arches at their flanks.

The McDermott was a compromise reached when an earlier regime at City Hall realized its original vision for the river was impossibly expensive. A group of wealthy women, returning from a vacation in Spain where they had seen Calatrava's imposing cable-stayed bridges flying across mountain gorges and canyons, decided that Dallas needed a brace of Calatrava bridges up and down the Trinity River where it flows between downtown and Oak Cliff.

At first, there was talk of as many as six Calatrava bridges up and down the muddy Trinity. That idea quickly was trimmed to three, then to two, which exist now — a short span across the river that's part of a longer pier-and-beam bridge, called the Margaret Hunt Hill Bridge, and the two McDonald's arches at the sides of the I-30 bridge called Margaret McDermott.

Hunt Hill was an heiress and philanthropist. McDermott was a wealthy widow and philanthropist. Both of their families chipped in to the cost of the bridges in exchange for naming rights, but taxpayers bore the lion's share.

Defenders of the two arched bridges refute claims that they are fake or pretentious, arguing that the two bridges are — or someday will be — functional and brought drama and sophistication to a bleak urban wasteland in the center of the city. But Griggs and others suggest the problems at McDermott are direct outgrowths of the structure's specious nature — two massive, plain-vanilla concrete highway bridges masquerading as one suspension bridge. The reason there is no test data for it from earlier projects is that no one else ever wanted to stick arches on the sides of bridges that didn't need them.

In that sense, the McDermott feels like it is of a piece with another project of the same City Hall regime just down the river from it — fake concrete kayak rapids built in the river at a cost of $4 million, a project foisted on City Hall by the same group of wealthy vacationers who had seen kayak rapids somewhere in the Rocky Mountain West and wanted to bring one home with them.

The feature was closed the day it opened because it was ill designed and dangerous. It must be torn out this year on orders from the U.S. Army Corps of Engineers, which deemed it a menace to navigation.

Griggs is one of a group of new young leaders in the city who think the most unique and dramatic feature Dallas could create in its center would be a nature preserve up and down the Trinity River corridor through downtown. The way to make that happen, they believe, is to let the river have its peace.
KEEP THE DALLAS OBSERVER FREE... Since we started the Dallas Observer, it has been defined as the free, independent voice of Dallas, and we'd like to keep it that way. With local media under siege, it's more important than ever for us to rally support behind funding our local journalism. You can help by participating in our "I Support" program, allowing us to keep offering readers access to our incisive coverage of local news, food and culture with no paywalls.
Jim Schutze has been the city columnist for the Dallas Observer since 1998. He has been a recipient of the Association of Alternative Newsweeklies’ national award for best commentary and Lincoln University’s national Unity Award for writing on civil rights and racial issues. In 2011 he was admitted to the Texas Institute of Letters.
Contact: Jim Schutze