Dim light at the end of the Brightwater tunnel

Metro King County's Brightwater sewage treatment plant tunnel is a project in a very awkward fix. There is a big problem with a contractor and its crippled tunnel boring machine on one of its essential tunnel contracts. All this carries lessons and warnings for the future waterfront tunnel to replace the Alaskan Way Viaduct, and is a critical early test for the new county executive, Dow Constantine.

On February 18, Constantine wrote the King County Council that it was time for decisive action. His fix: Take out of the picture the existing contractor with the failed tunnel-boring machine. Hand over the problem in mid-project to another contractor. Stand up to protect Brightwater ratepayers from costs and delay. Since then, this seemingly simple course correction ran into further problems and plot twists.

Last Wednesday, the Metro King County Council was briefed on the details of the new arrangements. To their credit, the county's Brightwater program managers were brutally honest at the briefing, where the watchword was 'ꀜmaking the best of a bad situation.'ꀝ Councilmembers saw a disturbingly more costly and risky course than Constantine had bullishly anticipated in February. It was expected that Constantine would quickly sign off on the new contract, which only happened over the past weekend. What's up?

The main problem is that the county had little bargaining leverage in switching contractors, and that meant a no-bid and no-guarantees deal to do the remedial work of finishing the excavation of the troubled tunnel section — no fixed price, no firm completion date, not even a firm promise that the work can be done at all. The bottom line for the public: Hold your breath to see how this all comes out. Plan B may be the best way out, but it's not at all a comfortable place for the project or the politicians to be.

So how did the county get itself in this doo-doo? Time for a short history and engineering lesson.

The 13-mile tunnel for the pipes to carry after-treatment wastewater from the new plant near Woodinville to Puget Sound offshore near Point Wells is an indispensable part of the cheerfully named Brightwater project. Three separate contracts were set up to build more or less simultaneously four separate tunnel sections eventually to be linked end-to-end for the run of what is in effect the pipe chase from the plant to the Sound.

Work has gone reasonably well on the two easier sections, the westernmost one inshore from the Sound, and the easternmost section extending from the treatment plant.

However, the third section, boring a tunnel west from Kenmore to Shoreline (Ballinger Way) for a distance of about 3.8 miles, is in deep trouble. For this section King County tapped one of the world'ꀙs biggest and most experienced contractors, France'ꀙs Vinci Construction. Vinci and its three partners were awarded the job on a competitive bid of $210 million in July 2006. The bid also won for Vinci and its team the second of the tunnel sections, running underground for 2.2 miles immediately east from Kenmore to Bothell.

In each of its two segments Vinci would excavate a 14-foot-diameter tunnel, reaming through the hard-to-mine glacial soils and high-pressure groundwater. The tunnels would be as much as hundreds of feet beneath neighborhoods and streets.

The Vinci segment driven eastward from Kenmore would end at a shaft in Bothell — meeting the first tunnel segment built by another contractor and today completed and connecting with the treatment plant near Woodinville. The Vinci segment driven westward from Kenmore would end at a shaft in Shoreline, meeting the tunnel arriving there from its mining start point on Puget Sound'ꀙs shore at Point Wells. That latter segment has been driven successfully by another contractor team headed by the big Michigan contractor, Jay Dee Contractors, and including Seattle'ꀙs own Frank Coluccio Construction Co. That section has also gone well, and the Jay Dee/Coluccio boring machine has only 150 feet left to mine to reach its destination in Shoreline.

Both Vinci segments were troubled almost from the outset, two years ago, by a host of obstacles and erratic mining progress. Underground construction characteristically has its challenges, so this was neither surprising nor a reason for panic. The hope in the early going was that progress would improve. But then things turned much worse in May 2009 when evidence emerged that the mining process was causing serious damage to the machines themselves. Both Vinci machines stopped mining for problem assessment and critical repairs.

Vinci'ꀙs machine on the Kenmore-to-Bothell segment appears to have been repairable. After months of downtime, Vinci got it back to work in February 2010 with just 4,000 feet left to mine. It'ꀙs successfully chewed through a quarter of that in the intervening few weeks.

But the other machine operated by Vinci has now been judged by King County to be best given up for dead. Immobile for almost a year, 300 or so feet below Lake Forest Park, it is badly stuck 10,000 feet from where it started in Kenmore in 2008 and 10,000 feet from where it was supposed to arrive in Shoreline by last month.

All agree that it would be a tricky, expensive, disruptive, and uncertain proposition to try to get this crippled Vinci tunnel-boring machine back into operating condition. The repairs would be technically challenging in any event, but with the machine trapped in a ground-water zone, a work space must be pressurized like a deep sea pressure chamber. That's a very challenging environment for worker safety.

Vinci'ꀙs threshold proposal to King County to fix the problem and finish the job for the third segment was that it would cost $98 million and an extra year on its contract completion date. That was a very tough outline of Plan A to recover the job. The county'ꀙs frustrated supervisory staff concluded that the true costs in time and money might be even higher and that Vinci couldn'ꀙt get back on track quickly enough to avoid significant delays to the entire project.

What was not clear in February, when Constantine first made his case for such a Plan B, is just how tough a bargain would be driven. That came out at last week's briefing.

Enter the new County Executive Constantine, who promptly jumped into the mess. Embracing the staff's negative view of Vinci'ꀙs prospects and seizing on bullish assumptions about attractive costs and schedule outcomes by bringing in another contractor to finish the undone mining on the Vinci problem segment, Constantine sent a letter to the council stating that another contractor must be found to complete the excavation of the troubled tunnel segment. He secured the council'ꀙs cooperation in paving the way to do so by loosening procurement rules.

So in February the course was set. Full speed ahead to Plan B. Details still to come.

From the council briefing last Wednesday, the exact contract terms and conditions for the Plan B program have become clear, and rather daunting. It is simple in outline. Jay Dee Constructors and Coluccio, the contractors teamed up on the fourth segment Point Wells to Shoreline section, would finish the expected drive of their machine into the shaft at Shoreline. Then they would re-launch their machine pushing it eastward and on past Shoreline into and through the 10,000 un-mined feet in the third segment project. In effect the Jay Dee/Coluccio machine would mine the rest of the third segment from the back door until it reached the location of the crippled westbound Vinci at the halfway point of the segment.

The key all along to Plan B has been the availability and willingness of Jay Dee/Coluccio to do the job with its on-hand, veteran machine as soon as possible after it has finished its tunneling task from Point Wells to Shoreline. If a different new contractor had to be enlisted and mobilized for the program, months and months inevitably would be lost to the schedule, probably altogether defeating the Plan B purpose.

That creates the rub. Even though Jay Dee/Coluccio is credited with having done excellent work on its share to date of the Brightwater project, Plan B has put King County unavoidably at the mercy of the Jay Dee team in negotiating the terms of its participation in the remedial work in segment three. It's a sole-bidder deal, unavoidably.

What was not clear in February, when Constantine first made his case for such a Plan B, is just how tough a bargain would be driven. That came out at last week's briefing. Against Vinci'ꀙs proposal to finish the work for $98 million, the target set for Plan B (in what is not precisely an apples-to-apples comparison) is $73.5 million, plus possibly a performance bonus of almost $4 million. That'ꀙs considerably less attractive than the $50-60 million anticipated in Constantine'ꀙs February 18 letter to the council.

But more importantly, it is not a contract price at all! It was described to the council as just a placeholder figure That'ꀙs because Jay Dee/Coluccio will be paid for its work on Plan B not an agreed fixed price, but by dollar-for-dollar reimbursement for all of its expenses and costs on the work. These are only estimated, not capped in the contract. On top of the costs would be $4.5 million for profit and also perhaps a $3.7 million bonus above whatever else it should earn if it should actually succeed in the work.

Then there'ꀙs the target date, not a fixed date, for the schedule. Vinci'ꀙs Plan A schedule for resuming its tunnel portion was a big part of the problem with its proposal, since Vinci suggested that the mining could not be finished until December 2012. In February, it looked as though Plan B might project getting the work done by June 2011. At Wednesday'ꀙs briefing, that hope had already slipped to December 2011. But no firm date was committed to or promised. So the schedule differential of Plan B over Plan A is actually speculative at this point.

These two key problems are so important that the following sentence is typed in bold-face in the construction document described to the council that Jay Dee/Coluccio for its part already had signed:

"The Contractor does not guarantee or represent a maximum cost or time to complete all or any section of the Work, and the Contractor does not guarantee or represent it can be completed."

On top of this, the briefing disclosed that the engineers still have not finalized a plan for a daunting and tricky part of the remedial task: When the Jay Dee/Coluccio machine indeed meets the front of the crippled Vinci machine, how will the two tunnel segments actually be joined together, in an exercise that will require Jay Dee/Coluccio and Vinci to coordinate their efforts?

There are still other loose ends. What will Vinci eventually be paid on its large disputed claims for costs outside the contract price to which it may already have entitlement? What about incidental implications of Plan B like the significantly extended duration of materials removal and barging at Point Wells, long after operations at that location were to have been completed? And what about the further cost for downtime for the Jay Dee/Coluccio machine currently stopped 150 feet from the Shoreline shaft awaiting further orders?

Sounds bad. But there is some saving grace to be found in Plan B. One is the strong performance to date on its own contract by Jay Dee/Coluccio. That's a critically important piece of support for Plan B. Without that, Plan B would not even be in the running. Another point in favor of Plan B is the professional directness with which the entire mess has been handled by Brightwater staff.

At the last minute, at least one new player appeared on the scene, pushing a possible Plan C.

As painful as the package they and County Executive Constantine have shown the county council, they have at least have held true to the important proposition I was taught in the 1990s superintending three big sewer and water tunneling projects in Boston: If no path is free of risk, uncertainty and contingency, it'ꀙs better to choose a course thought to be the best bet for success, even if uncertain. Dithering, which this Plan B certainly is not, in the long run is always likely to be the highest-cost, highest-pain option. It is better to step up to risk and get on with the job.

Plan B also shows real concern for practical details. Brightwater staff have not shirked adjustments that make the plan stronger, even if less immediately appealing. For example, the decision that time and good money must be spent on first-class attention to the condition of the Jay Dee/Coluccio machine before it launches into 10,000 feet of mining beyond what was originally intended for it on the project. Accepting this necessity has also been one of he reasons that the schedule targets for Plan B in April are more cautious than the bullish predictions made in February. Hard to step up to. Smart management.

For another, with almost no negotiating leverage with Jay Dee/Coluccio, the program staff seems to have won one small round in closing the door to a vexing game around equipment rental charges that can play enormously into contractors'ꀙ hands in the kind of situation on which the program now embarks. Inside baseball. Another good call.

In the days before the county signed off on Plan B, there was one last plot twist. At the last minute, at least one new player appeared on the scene, pushing a possible Plan C. According to a well-placed source, this plan was associated with a the big marine-facilities contractor, SMIT, a Dutch company tracing back to 1842 and currently in advanced merger talks with another Dutch company, Royal Boskalis Westminster N.V., a leading hydraulics and dredging concern.

According to the source, the eleventh-hour entrants made suggestions for easily repairing the immobilized Vinci machine by freezing the surrounding soils so that work in an unpressurized underground cavity would avoid the difficulties and dangers of a pressurized environment. Concerns for propriety commercial interests in an internationally competitive arena complicated discussions of a detailed pan. Neither Vinci nor its machine manufacturer, Herrenknect, welcomed the outside suggestions, the source reported. Ultimately, according to another source, the problem with the Dutch bid was that it only addressed fixing the Vinci machine, not completing the troubled segment.

Finally, after a weekend of review, the staff recommended, and Constantine's representative Christie True, director of the Wastewater Treatment Division, signed the contract for Plan B. Credit goes to the county for inclining toward a contractor who has earned their trust, and for being frank about the bind the failed machine has put them in.

And are there lessons here for the coming Viaduct-replacement tunnel? A Vinci-led team was initially designated as one of the four competitors for the deep-bore tunnel contract. In late March, it withdrew its interest in the Viaduct project. Meanwhile, sources close to the Viaduct program report that they are paying close heed to Brightwater's lessons. Brightwater should warn us all of the challenges of big-time underground construction.