Suspended in Midair
January 1, 2010
On Labor Day weekend 2009, the California Department of Transportation (Caltrans) and construction engineering firm C.C. Myers Inc. successfully completed an amazing “high-wire act” involving thousands of tons of concrete suspended 150 feet above the ground. It was an aerial ballet performed on the San Francisco-Oakland Bay Bridge--a feat requiring exacting precision and perfect execution. Failure would have crippled Bay Area transportation infrastructure for months. “It’s not the first time this kind of project has been done, but it may be the most-complicated one yet,” says Jim Brainard, PLS, of R.E.Y. Engineers Inc., the firm subcontracted by C.C. Myers to handle construction layout and monitoring.
The project involved replacing a massive section of the bridge with a seismically retrofitted segment built onsite on temporary falsework and slid into place on rails. A similar retrofit had been done for another segment of the Bay Bridge on Labor Day weekend in 2007. Brainard and R.E.Y. were involved then, too; they measured the incremental progress of the new section as it was jacked into place. But the 2009 slide was more complicated for several reasons. At 45 meters above Treasure Island, it was the highest slide yet attempted. Also, the section being replaced was not demolished; instead, it was lifted and slid out of the way. Consequently, the measuring work done by R.E.Y. was also more complex and precise. The firm relied on technology that hadn’t even been considered for previous slides.
It’s All in the Details
About 270,000 vehicles per day cross the Bay Bridge. Closures are extremely rare and expensive--just publicizing closure dates costs about $1 million. In this case, Caltrans officials closed the bridge on Thursday, Sept. 3, the first workday closure since the 1989 Loma Prieta earthquake, which damaged a section of the upper deck. From the moment the bridge was closed, a meticulously constructed flowchart dictated every move with responses planned for delays.
The new double-deck section, commonly known as the “S-curve,” was built literally alongside the 1930s era 3,300-ton section being replaced. Both decks--measuring 283 feet long by 65 feet wide (five lanes per deck) and weighing 3,600 tons--were lifted 150 millimeters (5.9 inches) by eight jacks capable of lifting up to 500 tons each and then slid along the same set of Teflon-lined steel rails lubricated with Dawn dishwashing detergent. The lift-and-slide operation was handled by Mammoet, a Dutch firm renowned for moving extremely massive objects.
On the 2007 slide, R.E.Y. set and tracked four prisms with one instrument and calculated positions after incremental 1-meter moves. This time, Mammoet planned to move the bridge sections in pushes that were as long and steady as possible. To do so, the operators required continuously updated information on position and deformation. If the sections were flexing or twisting during the moves, the Mammoet operators wanted to know by how much, and they wanted to know immediately. So for this slide, R.E.Y. monitored eight points with eight Trimble autotracking total stations (four Trimble SPS930s and four Trimble S6s) working simultaneously. “We knew that tracking any two points would tell us position,” Brainard explains, “but we needed a lot more information to be sure about deformation.” Ultimately, prisms were set at the corners of the section, at midpoints along the long sides, and at two points on the deck.
Getting New Use Out of Existing Software
To monitor all points simultaneously and to continuously synthesize the information coming from eight sophisticated instruments into a comprehensible display--ideally on just one screen--R.E.Y. chose the Trimble HYDROpro Navigation Software. HYDROpro, as the name implies, is typically used in marine construction projects to monitor the location of large ships and construction platforms. To get the most out of the software, R.E.Y. turned to Lou Nash, owner and president of Measutronics, a Florida firm specializing in marine positioning and guidance systems and an expert in HYDROpro. “The first thing we did was add a decimal place to our system,” Nash says with a laugh. “There’s not much call for millimeters working on the water!”
The Trimble SPS930 Total Station was chosen as a main instrument because Nash has tested it often with HYDROpro. “We didn’t want to experiment on this job,” Brainard explains. Using a combination of eight total stations provided high data rate measurements, minimized lag, and enabled R.E.Y. to record horizontal and vertical measurements simultaneously. Haselbach Surveying Instruments and California Surveying & Drafting Supply Inc. worked behind the scenes to assemble the suite of instruments and optics needed for the project.
Each total station was fitted with a Trimble TSC2 Controller. The controllers were connected by Bluetooth to a computer running HYDROpro and specialized presentation software (SOTI Inc.’s Pocket Controller Pro) to synthesize the data and present it on one screen in useful displays. “We really had to drill down deep into the software to make sure we were getting all we could from it,” Nash says.
“I usually use HYDROpro to guide and track marine construction,” he continues, “and in those applications, the dynamics of moving platforms are usually a challenge. Here, the dynamics--the rate of movement--wasn’t such a challenge, but the tolerances were much tighter. I learned a lot more about the software as I worked with it to handle this level of accuracy (Mammoet and R.E.Y. were striving for 5 millimeters or better at all phases of the slide) with real-time reporting.”
Large corrections would have been no good. Mammoet needed to be sure that left and right skid beam travel was continually aligned within 5 centimeters to prevent binding of the lift jacks. So Nash set up a system that gave Mammoet’s hydraulic system operators continual course and guidance input to within 15 millimeters, which they used to make continual, gradual corrections. This was a substantial contrast to the previous Bay Bridge slide, when shots were taken, positions were calculated, and hydraulic system pushes were readjusted at 1-meter intervals.
Two weeks before the slide, Nash met with Brainard and R.E.Y. Crew Chief Tom Cade, PLS. “I helped them to get up to speed on HYDROpro and on interfacing with the total stations,” Nash says, “and we made our best estimate of what Mammoet would want to see.” With those decisions made, the team set up a dry run in R.E.Y.’s parking lot to verify their approach; then, Nash says, “We had two weeks to sleep on it.”
Or not. As the main person in charge of monitoring on the day of the slide, Cade had a couple of nervous weeks. At night, he says, “I’d lie down and think about this and wouldn’t sleep.” But when the big day came, Cade, along with crew members Jeff Cook, Tim Pringle and Brett Brusatori, LSIT, were as ready as they could be, given the novelty of what they were attempting.
The Big Day
Lou Nash and the R.E.Y. crew were onsite soon after the closure on Thursday evening and remained onsite for 36 hours with a seven-hour break between the roll-out and roll-in. Since the “break” was filled with data processing, reports, battery charging and other necessary chores, everyone got by with naps. “We kept the caffeine flowing,” Brainard says.
The value of precise monitoring was shown during the roll-out. “One primary purpose of monitoring deformation was to let Mammoet know when the old section was actually released,” Cade explains. Since the old section had been in place since the mid-1930s, it was possible that parts were fused in ways that couldn’t be detected.
This proved to be the case. By following all points being monitored and comparing to ideal positions at each point in the move, Cade could tell there was a problem soon after the initial lift of the old section commenced. “I had a spreadsheet going and could watch deflection in real time,” he says. “At midspan, there was little or no deflection when lift started, and we were at plus 7 millimeters. But at the ends of the span, we were seeing minus 8 millimeters.” On both the roll-out and the roll-in, measurements like this were occasionally physically checked by Mammoet workers underneath the sliding sections--these onsite measurements always confirmed the monitoring setup.
After inspection, it turned out that the southeast corner was stuck tight, and pneumatic hammers and torches were brought in. This added unexpected time to the roll-out, which took 13 hours instead of five. Not only did the additional time threaten the carefully constructed schedule, it gave Cade something new to worry about: Unexpectedly, battery life could become an issue. He monitored more or less continuously during the 13-hour procedure (a long period for any battery) and was relieved when the first slide was completed. “We finished up at 10:30 p.m. on Friday, and we had to be back onsite and be ready for the roll-in at 5:30 a.m. the next morning,” he says. “That gave us seven hours to get to the hotel, get the batteries on, do our processing, and get back to the site. I didn’t sleep--I did the processing and wished I had more chargers.”
By the next day, just 24 of 27 batteries were recharged, and Cade was definitely hoping for the roll-in to go with fewer hitches than the roll-out.
Meanwhile, another drama was unfolding. Although it didn’t affect R.E.Y. or Mammoet, the rest of the Bay Area was breathlessly following the tale of a cracked eyebar on a different section of the bridge three-quarters of a mile away that threatened to shut down the Bay Bridge for weeks, even if the roll-in was successful.
The eyebar was part of a network of eight eyebars joined up in a critical support structure. Caltrans official Bart Ney said that the crack was serious enough to warrant a bridge closure, even if one weren’t already taking place. So while one of the world’s best examples of “extreme infrastructure” was taking place, Caltrans was also faced with the necessity of doing a major repair on short notice--and they had the opportunity to do it during a preplanned closure. C.C. Myers, lead contractor for the new alignment project, jumped into the emergency repair with the brio for which the firm is noted. A repair strategy was crafted, parts were fabricated at plants in Oakland, Calif., and Arizona, and the eyebar repair was completed with less than a day added to the overall closure time. “There could not have been a better time to discover the cracked eyebar,” says Steve Heminger, executive director of the Bay Area Toll Authority. Though it didn’t slow down the slide project, it did supercharge the atmosphere in an already tense environment.
While Cade looked after the battery chargers, Cook and Pringle donned double-strapped safety gear and moved prism setups from the old span to the new span. With prisms set, the lift of the new section began at 8:30 a.m. Deformation was less than expected. “On the roll-in, we were expecting the ends to deflect about 2 to 3 centimeters down,” Brainard says, “but, in fact, it was less--about 1 centimeter. This was definitely good news, and there were no issues during the 150-millimeter lift.”
At maximum slide rate, the new section could have covered the 30-meter gap in two hours, and the first half of the roll-in did go quickly. Then an unexpected obstacle appeared. “About halfway through the slide,” Cade says, “Mammoet noticed that some bracing, which had been added to reduce wind vibration, was now interfering with the skidding operation.” Brainard adds, “It was supposed to be tack welded, but it turned out to be fully welded, which meant that torches had to be brought in.” Removing the obstacle caused a four-hour delay, and once again, Cade was getting nervous about battery life.
But after the obstruction was cleared, the slide continued very smoothly. After relatively fast travel over most of the 30 meters, the last four centimeters were covered in 20 minutes. Final location of all corners was within 4 millimeters of design--impressive precision considering that a 3,600-ton, double-decker bridge section built 45 meters above the ground had just been moved 30 meters into position.
There is one moment on the day of the slide that Cade won’t soon forget. “There was a trailer onsite where we had our little 20-inch monitor set up,” he says, “and in the same trailer, Mammoet had a big 44-inch monitor. And I realized that the trailer was full of people all watching our little monitor. It was memorable to be at the center of all that.”
Precision, Not Sheer Power
Even with the eyebar repair and slide delays, the Bay Bridge reopened on Tuesday morning as originally planned. By harnessing and measuring massive force, R.E.Y. Engineers and Mammoet were able to place a massive piece of infrastructure precisely on target and keep traffic flowing on one of America’s greatest bridges. As the San Francisco Chronicle noted in a follow-up article, “That effort was one of precision--not sheer power.”