Big Dig officials chose not to retest

Several ceiling bolts came loose in the Interstate 90 connector tunnel while it was under construction in 1999, even after they passed a safety inspection. But project documents show that officials overseeing the Big Dig chose not to retest most of the bolts in the tunnel, including those that would suddenly fail seven years later, causing last month's fatal ceiling collapse.

Instead, managers from Bechtel/Parsons Brinckerhoff ordered more rigorous safety testing only in the high-occupancy vehicle lane, where the loose bolts had been discovered. In a January 2000 memorandum, a Bechtel/Parsons Brinckerhoff engineer wrote that the new testing ``is for HOV ONLY."

That may help explain a curious finding investigators made after the death of Milena Del Valle under a pile of concrete in the tunnel's eastbound lanes: More than 200 bolts supporting the ceiling had come loose in most parts of the tunnel, but none was found in the HOV lane. In that section, each bolt was installed following strict guidelines to ensure the epoxy holding the bolts gripped as designed. The strength of each bolt was then tested in 2000, with each required to withstand almost twice as much weight as bolts in the rest of the tunnel.

``That's pretty good proof that testing does pay," said David W. Fowler, an engineering professor at the University of Texas at Austin and an authority on epoxy bolts. ``The key is to go back and make sure . . . that everything was done properly. Apparently they didn't do that."

The decision not to re test the epoxy bolts is one of several critical choices over a 14-year period in which Bechtel/Parsons Brinckerhoff -- a joint venture of engineering and construction management firms that the state picked to supervise the $14.6 billion Big Dig -- played a lead role in reducing the tunnel ceiling's safety margin. An investigation by the Boston Globe found that Bechtel/Parsons Brinckerhoff discouraged a proposal from designers to suspend the tunnel ceiling from a different kind of bolt they believed would be more reliable than epoxy. The Big Dig managers also turned down a pitch from the company that built the tunnel, Modern Continental Construction Co., to use lighter ceiling materials than concrete.

A review of correspondence between the companies involved in the ceiling project, as well as 1,800 pages of engineers' daily notes from 1999 and 2000, shows that Bechtel/Parsons Brinckerhoff allowed ceiling construction to proceed during the three months that managers debated what to do about the unexpected loose bolts. By the time they came up with a plan, in January 2000, workers had already hung concrete ceiling panels in the east and westbound lanes of the connector, making retesting of the bolts much more expensive and time consuming.

So Bechtel/Parsons Brinckerhoff decided to retest bolts only in the HOV lane and, ``if retests prove good," not to bother with the east and westbound lanes, according to a January 14, 2000, memo from Robert E. Steffy, Bechtel/Parsons Brinckerhoff's resident engineer on the tunnel. Seven months later, after only a few bolts in the HOV lane failed the tougher test, Bechtel/ Parsons Brinckerhoff officials declared the matter closed.

In the end, the connector tunnel got a ceiling like few others in the United States, counting on bolts that essentially have been super-glued into the roof to each suspend 2,600 pounds of concrete. Fowler and other outside experts could name only one other tunnel, in Virginia, that used a similar bolt system for such a heavy ceiling, and they say few are likely to be built in the future because of the Big Dig's problems. The connector tunnel's epoxy bolts, and the methods used to install and test them, are the major focus of investigators looking into the cause of the July 10 ceiling cave-in.

Still, engineers said the ceiling system used in the connector tunnel could have been safe if Bechtel/Parsons Brinckerhoff and the company that designed the ceiling had required enough precautions, but that's not what happened. The bolts in the east- and westbound lanes were tested to make sure they could carry 3,250 pounds, just 25 percent more than the weight of the ceiling they are supporting. That leaves little margin for vibrations, workers standing on the ceiling structure, or poor workmanship that weakens some bolts. By contrast, bolts in the HOV lane were tested to hold 6,350 pounds, and, after the accident, Governor Mitt Romney required new bolts, being installed to reinforce the tunnel ceiling, to be tested to bear 14,000 pounds.

Bechtel/Parsons Brinckerhoff declined to respond to questions, but defended its work in a statement issued Friday, stressing that no one has figured out exactly why the tunnel ceiling caved in.

``Investigators have said they are reviewing some 40,000 documents to figure out what happened. Everyone should be very cautious about rushing to judgment," said the statement from Andrew M. Paven, senior vice president of O'Neill and Associates who has served as Bechtel's spokesman. ``The full investigation will take time and, when all the facts are in, we will be here to stand by our work."

Bechtel/Parsons Brinckerhoff officials argue generally that it's unfair to single them out in Big Dig decisions, since virtually every important choice -- including any payments for additional bolt testing -- was made with the written support of the state as well as the Federal Highway Administration. By early 2000, with state officials fending off charges that they had concealed vast cost overruns on the Big Dig, proposals to spend more money were getting a cool reception.

Bechtel/Parsons Brinckerhoff officials also note that the joint venture did only ``conceptual design" for the Big Dig, leaving it to subcontractors to work out details such as the type of ceiling bolt to use and how much weight it should carry. In the end, they point out, Gannett Fleming, a design firm with offices in Braintree, put its professional stamp on the connector ceiling design, asserting that the engineering was sound.

However, engineers and architects who have worked on other parts of the Big Dig say Bechtel/ Parsons Brinckerhoff wielded enormous clout down to the smallest details. Though various state agencies oversaw Bechtel/ Parsons Brinckerhoff in the course of the project, relatively few state employees were assigned to the project and, often, they were nonetheless paid by Bechtel/ Parsons Brinckerhoff under a complex cost-saving program for the project.

In reality, design and construction of the connector tunnel ceiling came down to a dialogue between Bechtel/ Parsons Brinckerhoff, the designers at Gannett Fleming, and the builders at Modern Continental. More often than not, project documents show, Bechtel/ Parsons Brinckerhoff prevailed when there was disagreement.

The I-90 connector was built with a drop ceiling to create a chamber at the top of the tunnel that could be used for ventilation: In a fire, smoke could be sucked into the overhead area, then rapidly vented outside. The first drop ceiling in the Big Dig tunnels, located in the Ted Williams Tunnel that opened in 1996, was also held up by epoxy bolts. But the ceiling panels, composed of a thin layer of metal and concrete, weigh 60 percent less than those in the connector tunnel ceiling, so the bolts have far less stress on them.

Unfortunately, the Ted Williams ceiling was both costly and difficult to install, and Bechtel/ Parsons Brinckerhoff managers decided that future tunnel drop ceilings would be made of pre-cast concrete slabs. To make it easier to hang the drop ceilings, the project managers called for steel channels to be built into the tunnel roofs from which the ceiling hangers could be suspended without using epoxy bolts.

But the eastern end of the connector tunnel, site of the July 10 accident, had already been constructed without these steel channels. As a result, the last 200 feet of connector ceiling would have to be supported exclusively by epoxy bolts. Epoxy bolts would also be used elsewhere whenever a steel groove was not directly overhead to provide support.

Gannett Fleming and Modern Continental officials declined to be interviewed, but documents show they disliked the heavy concrete ceiling and both companies tried to develop lighter alternatives. Modern went so far as to hire another firm, Environmental Interiors of New Hampshire, to design a ceiling that would have been 60 percent lighter and, Modern argued, less expensive. But Bechtel/ Parsons Brinckerhoff rejected the idea, saying the lighter ceiling might not be rugged enough and, in any event, the contract to produce concrete ceiling panels had already been signed with a Vermont supplier.

Likewise, in a Sept. 24, 1997, letter to Gannett Fleming, a Bechtel/ Parsons Brinckerhoff engineer insisted that the designer abandon consideration of any alternative to the concrete panel design developed by Bechtel/Parsons Brinckerhoff. Gannett Fleming ``will proceed immediately with implementing this design," wrote engineer John P. Bartoletti.

Designers in Gannett Fleming's Braintree office didn't consider epoxy bolts a first choice either, noting in an August 25, 1997, memo to Bechtel/Parsons Brinckerhoff that the epoxy may not be strong enough in certain types of concrete and that it could fail in the heat of a fire. They told Bechtel/Parsons Brinckerhoff that they preferred bolts with a special tip that expands once inside the hole, creating a powerful grip. (After the accident last month, Romney called for bolts of this type to be installed in the connector.)

But Bechtel/Parsons Brinckerhoff officials balked at the idea of these undercut anchor bolts, noting that they require special equipment for installation. If these bolts are not installed correctly, warned Bechtel/ Parsons Brinckerhoff engineer Prabir Das, ``they can be pulled out by hand," according to a 1997 memo on his discussion with a Gannett Fleming official.

Ultimately, Gannett Fleming officials agreed to use epoxy bolts to hold up the ceiling, choosing bolt that could hold up to 31,000 pounds. That would make them strong enough to hold up the ceiling -- as long as they were installed correctly.

On October 7, 1999, Modern Continental notified Bechtel/Parsons Brinckerhoff in writing that five ceiling bolts in the connector's HOV lane had begun to pull out of the roof less than two months after the concrete ceiling panels had been suspended from them. None of the bolts had moved more than half an inch, project manager Robert Coutts reported, but it was disturbing because these bolts had already passed a strength test suggesting that they were safe.

Two weeks later, on October 21, officials from Bechtel/Parsons Brinckerhoff, Modern Continental, and the company that supplied the epoxy met in the tunnel to discuss the loose bolts, all located in the ``mock-up area" of the HOV lane, a place Modern had set aside to test ceiling installation techniques. Notes from the meeting show that the officials believed workers had been installing bolts incorrectly, and they were using the wrong type of drill bit, causing the surface inside the bolt hole to be smoother than would be ideal for the epoxy to bond.

Workers complained that the roof of the connector tunnel was so filled with steel reinforcement that they kept bumping into it whenever they drilled bolt holes. Every time that happened, they were supposed to drill a new hole and fill the one that was abandoned, a time consuming annoyance.

Bechtel/Parsons Brinckerhoff field engineer Jim Barrett warned in his notes from the October 21 meeting that epoxy bolts could lose up to two-thirds of their strength if workers used bolt holes that had been drilled through the rebar. And Barrett suspected that was happening: a few weeks earlier, he had discovered that 60 percent of the bolt holes in the HOV lane that he inspected had actually been cut through steel reinforcement bar despite Modern's claims to the contrary.

This was not the first time that epoxy bolts had begun to ``creep" out of their moorings in a Big Dig project. When it happened in 1994 during construction of the Ted Williams Tunnel, Bechtel/Parsons Brinckerhoff acted decisively. After he had been briefed that seven ceiling bolts were coming loose in the earlier tunnel, Bechtel/ Parsons Brinckerhoff's chief designer, Anthony Lancellotti, wrote that crews ``should halt any further installation until this is resolved."

By contrast, Bechtel/Parsons Brinckerhoff and Modern Continental took three months to come up with a plan to deal with the bolt ``creep" in the connector tunnel. During that time, crews installed most of the eastbound ceiling, including the section where the accident would later occur.

Under Bechtel's plan, described in a January 6, 2000, memo, Modern crews were supposed to follow stricter bolt installation procedures and to drill bolt holes with a carbide bit rather than the diamond bit they had been using. Once each bolt was installed, Bechtel wanted a testing company to nearly double the weight it had been using to check the bolts' strength, from 3,250 pounds to 6,350 pounds.

In addition, the Bechtel memo warned that Modern would need to retest bolts that had been tested at just 3,250 pounds because ``the capacity of all of the other anchor bolts is in question."

But Modern responded that testing all of the bolts would be expensive. By January 6, according to a company memo, workers had already hung the ceiling from more than 1,000 of the 1,146 bolts in the whole tunnel. Testing these would be time consuming, requiring a special lift to raise up the concrete ceiling panels around each bolt so that testers could suspend the test weight from them.

On January 14, Bechtel resident engineer Robert Steffy wrote back that the testing plan ``is for HOV ONLY." If not many bolts fail at the higher weight, he wrote, the other lanes ``probably will not require retesting."

Over the next eight months, 187 bolts in the HOV lane were tested at 6,350 pounds; 19 of them pulled out under the higher weight and had to be replaced. But 17 of those weak bolts were located in the so-called ``mock-up" area where the ceiling had been installed early in the project, supporting Bechtel's belief that worker inexperience with epoxy had caused the problem.

Bechtel's lead field engineer Phil Aikele said in a memo that the testing showed there was no need to retest bolts in the east- and westbound lanes. He suggested that, as workers got better at installing epoxy bolts , they made fewer mistakes.

Ceiling bolt work in the early days of the project was not confined to the HOV lane, however. In fact, on June 10, 1999, the day work began on the connector ceiling, bolt holes were drilled at the eastern end of the eastbound lanes, according to a field engineer's notes. That's the area where the ceiling would cave in seven years later.

But on August 31, 2000, Aikele recommended to Bechtel that the re testing issue be closed.

Scott Allen can be reached at allen@globe.com and Sean Murphy at smurphy@globe.com.