A Microtunneling Machine

When Mike Garver rolled into Houston in 1964 he was the classic first chapter: a fresh degree in mechanical engineering in his pocket, his possessions loaded in the trailer behind, ready to take off after anything that looked legal.

At the time the city was growing frantically, throwing buildings up all over town. Garver knew that all this action above the surface had to have its reciprocal under the surface, in terms of water and sewer projects. (Not everyone appreciates the scale of this industry -- out of sight, out of mind -- but it is handsome; American municipalities spend more than $5 billion/year on capital improvements for sewerage and treatment alone.)  After spending a few years selling draglines (Texas for cranes), he started his own contracting company, intending to specialize in the construction of municipal utility systems.

Even in Houston this business had been spoken for decades, and Garver, like any newcomer in an established sector, had to carve his own path into the business. In his case this meant keeping an eye open for jobs with something novel or complicated about them, since it would be hard for a newcomer to outbid companies on work they had done a dozen times before. It was from this perspective that he contemplated a city job posted in the fall of 1986 that endorsed the use of a piece of machinery called a "laser- guided microtunneler", a device neither Garver nor any other contractor in Houston had worked with before. The job was worth $10 million --- potentially the biggest piece of business yet in the short history of Garver's company.

He remembers calling up a vendor and asking what the learning curve was like with this equipment. "You know how it is when you talk to sales," he says now. "'Just slip Tab A into in Slot B'". Sales sent over a video, which Garver watched. Thus reassured, he made a low bid, got the job, and ordered one of the first microtunnelers to be sold in the United States.

Microtunnelers, aka robot moles, had been developed by the Japanese in the early 70's as a way of laying pipe without first digging long trenches. The idea was to dig shafts at successive street intersections, lower pipe down into the shafts, and push (or jack) it into the next shaft, leaving the surface above undisturbed. As the pipe was jacked a steerable digging or cutting mechanism, like an auger or toothed wheel, removed the spoils or "muck" from in front and passed it back through the machine, through the pipe, and out into the shaft, where it was lifted to the surface.

The technology turned out to have many advantages: It reduced surface disruption, accelerated production, and allowed contractors to dive right under the tangle of pipes that clutter the near surface of any city, avoiding service interruptions. By the early 80's microtunneling was established in most of the industrialized world.

Except the United States. Americans like to think of themselves as receptive to new technologies, but this enthusiasm is not much in evidence in the construction of public facilities like roads, sewers, aquaducts, and treatment plants. While there is no consensus as to the underlying cause (observers have suggested reflexively risk-averse bureaucracies, the worship of the low bid, contractor cartels, regulators, and lawyers, among others) the phenomenon is clear enough. Until well into the 80's the city engineers of Houston actually forbade the use of a backhoe, insisting instead that trenches be dug with an ancient device that worked by scraping digging buckets hung between chains up against the working face of a site.

"Trenching ladders" were clumsy and broke down constantly and required huge crews to run, but they had been used forever and the city was confident that they would not detonate some public relations disaster. (The oldline contractors, who had amortized their trenchers decades ago, provided the political muscle behind this policy.) Sometimes a contractor would do a job with a backhoe and then stick a trencher down in the cut for a few hours while the city engineer made his tour. But most used the prescribed tools in the prescribed way, as their fathers had before them.

Over the late 70's and early 80's it was gradually borne in on the city engineers that most of the sewer work they had bought over the last twenty years needed to be done over. For one point there were far too many pumps. Pumps are the weak points in any sewer system: the machines have to work 24 hours a day under difficult operating conditions and are forever breaking down, backing up, and overflowing. The right way to build a sewer between points A and B is to let gravity do as much of the work as possible, which means allowing the pipes running between the pumps to fall as deep as possible. (The outlet of a thousand yard pipe has to be 30 feet lower than its inlet).

Cutting trenches to depths like those is expensive in soft, water-saturated soils like Houston's, which are prone to cave-ins. The city had tried to build a cheaper sewer system, with lots of short, shallow, connections linked by hundreds of pumping stations, but the overflows were turning into a serious public nuisance. Clearly the city needed something deeper.

One of these new lines had to go through a politically well-connected neighborhood (River Oaks) with fairly narrow streets. Excavation on the scale required would have meant shutting a lot of rich people up into their homes for weeks. (Tony Cresci, who was the city's wastewater engineer on the job, says that acres of expensive, exotic, and presumably cherished plantings would also have been at risk.) Houston had looked at this job in the 70's, winced, and shoved it back on the bottom of their to-do pile. By the mid-80's pressure from the EPA was making it unavoidable.

By coincidence an engineer consulting for the city, Calvin Morgan, had family in the British contracting culture, and knew that in Europe this job would have been microtunneled without a second thought. Morgan began advocating the technology from inside, and the combination of his enthusiasm and the political constraints of the project opened the door. When the bid went out microtunneling was written right into the specifications, in the same place where the trenching ladder requirement used to go.

Garver took delivery of his machine, lowered it into a shaft, aimed it, and started digging. The tunneler ran off into the ground and stalled, wedged tight. He dug a shaft down from the surface, loosened the machine, pointed it again, and again it jammed. The technical consultants from the vendor were baffled. The machine had worked fine in Japan. They offered to give Garver back his money and take their machine home.

While this was an honorable gesture from their point of view, Garver was in no position to accept, since giving the machine back would have left the contractor sitting around watching a $10,000/day late penalty clause eat up his business. He had a hunch that the water-saturated silts, sands, and clays of Houston (known as "gumbo" soil) were imposing a different regime of forces on the operating cycle of the machine than the rocky heterogeneous soils of Japan, so he picked up a cutting torch and reshaped the throat of the mole.

Gradually he got the machine going, but clearly there had been some significant miscommunication with the vendor on this Tab A and Slot B business. "River Oaks was a real baptism by fire," he says now. Microtunnelers take a mix of systems -- hydraulics, electrical, electronics, mechanical, optics, even video -- down into an environment that is dirty, wet, unpredictable, characterized by extreme pressures, inaccessible and largely unobservable. If anything goes wrong with any of those subsystems, the whole cycle stops and it doesn't start until that failure is fixed.

One wrong touch on the steering knobs (the machines are guided remotely from a control console set up on the surface) and the tunneler might start "porpoising": oscillating back and forth across an axis instead of running straight down it. Too little pressure on the jacks and the machine will stick (and the longer it stays stuck the more tightly the surrounding ground will settle around it); too much pressure and the pipes will crack, though not necessarily in such a way that will be obvious until later. The exact value separating too much pressure from too little varies constantly with soil type, which is unobservable until the tailings start coming up out of the shaft.

The pipes are jacked through lubricating fluid, the exact recipe for which also differs with soil type. These and other constraints mean that jobs have to be planned to a far greater level of resolution than with cut-and-cover trenching. Hundreds of connections need to be monitored and maintained. (If a cable breaks when the machine is underground a day's production can be lost.) A large parts inventory has to be kept onsite and a versatile equipment repair shop built up in the home office. The work force needs to be trained to high skill levels and selected for adaptability and an obsessive attention to detail. How fast an employee could learn became a critical factor in planning work assignments. "A worker who doesn't like to learn isn't going to be very happy doing this work," he says.

Amazingly, the River Oaks project came in on time, though just barely, and Garver even made a small profit. From the point of view of many of the local contractors this was not good news. The Houston rebuild, which is coming in at well over a billion dollars, could have kept the whole community of cut-and-cover contractors in steak for years; now Garver was threatening to grab most of it with this Buck Rogers technology.

The contractors lobbied the City Council, arguing that it was unfair to the taxpayer for city engineers to get involved with specifying tools. They should just define the job and let lowest bidder win. (In other words, surface disruption was no business of the city's.) Several jobs were in fact announced that way, but even within these ground rules Garver kept winning contracts, since while trenching gets more expensive with depth, the cost of microtunneling stays pretty much the same. Today his company does roughly $30 million dollars of business a year, about half of which comes from microtunneling projects.

River Oaks made believers out of Houston Public Works, but the rest of the country stayed unconverted. This lack of acceptance presented a long-term problem for Garver, who knew the Houston work would come to an end in a few years. So, in the phrase of his estimator, Clifford Tubbs, "we decided to let our light out of the basket." Going against all the instincts of contractor culture, he began inviting competition into the business, knowing that in the long run that was the only way to build market.

In 1990, in cooperation with a trade magazine, he sponsored the first US conference on the technology. He set up a microtunneling field demonstration (and barbecue) on his company grounds and ran in busloads of engineers, contractors, and clients from the conference for working demonstrations. In 1992 he helped organize the Gulf Coast Trenchless Association, a promotional group that supports lectures, technology demonstrations, and conferences in underground construction. "He's the leader," says Prof. Tom Eisley of the Purdue School of Engineering Technology. "There's been nobody else with that kind of vision."

The last job Garver bid on eight other microtunnelers were bidding against him (he lost), which is one measure of his success. Another is that the number of feet dug with microtunneling has begun to edge up: 300K feet in 1994, compared with 90K in 1989. The director of the new microtunneling center at the Colorado School of Mines, Tim Coss, estimates there are about 40 machines in the country now, which is impressive compared to the half dozen we might have had five years ago, though less so compared to the thousands of machines he thinks are working in Japan.