The Underground Cutting Edge

IN THE MIDDLE OF THE NINETEENTH CENTURY, WESTERN civilization was at the peak of its early intoxication with railroads. Everywhere dreamers were bent over maps, drawing lines. Walt Whitman wrote:

Lo, soul! seest thou not God’s purpose from the first? The earth to be spann’d, connected by net-work, The people to become brothers and sisters, The races, neighbors, to marry and be given in marriage, The oceans to be cross’d, the distant brought near The lands to be welded together.

Alas, time and time again these lines, and therefore these dreams, would bump into a mountain, or worse, a chain of them. Building over or around was expensive and time-consuming, so engineers yearned to poke straight through. But that meant tunneling, and the bill for tunneling was even higher. Tunneling by its nature offered only a tiny working area, with a rock face not much larger than a bed sheet and about the same amount of room to stand in. Only a handful of people could labor within it at any one time, and much of every shift was wasted wrestling tools and work product out of one another’s way. The Hoosac Tunnel, in Massachusetts, took more than 20 years to go five miles, and while that was worse than usual, the ancient formula “time equals money” made tunneling costs distressingly high everywhere.

As railroads spread, people on two continents had the same thought. This was the nineteenth century, and engineering and science could be counted on to triumph over every sort of obstacle. The fix was obvious: Just build a big machine to dig the tunnels. Amplify the power of labor, and bring the Industrial Revolution underground. How hard could it be?

The first to try was a Belgian engineer named Henri-Joseph Maus. In 1845 he got the King of Sardinia, whose domain included the Savoy and Piedmont regions on the mainland, to approve construction of the first railroad connecting France and Italy. Maus had an international reputation in mining engineering and the self-confidence to match. He shrugged off the idea of running a line up and over a pass, insisting that the right idea was to go straight through—specifically, through Mont Fréjus, near the famous pass at Mont Cenis.

Even given the expectations of the day, this was bold. A tunnel following the route Maus had in mind would have stretched for 40,000 feet, a highly implausible distance. In that era the tunneling cycle ran as follows: Drill holes in the rock face, pack them with gunpowder, light the fuse, run around a corner, wait for the explosion, run back carrying bracing timbers, hope you can hammer them in place before you get killed in a roof collapse, and shovel or toss the rock fragments into carts for removal.

The problem was that detonating gunpowder in a confined space saturated it with toxic fumes, so all this activity depended upon sucking the combustion products from each blast out of the tunnel in a reasonable amount of time. Maus’s tunnel was far too long for the ventilating technology of the era to deal with, but of course he had thought of this. He planned to dispense with blasting altogether by building the world’s first tunnel-boring machine.

Maus’s “mountain slicer,” as it was dubbed, took shape in an arms factory near Turin in 1846. It was big and complex, larger than a locomotive, bristling with more than 100 percussion drills, all set in a forest of cams and shafts and gears and springs. Functional or not, it was great to look at, and tourists came to admire it as a monument to the age, more a piece of art than a tool.

Of course, other visitors appraised the machine more as a tool than a piece of art, and some of them left with doubts. The enormous levels of power required to drive the mountain slicer would be generated outside the tunnel and carried to the work face via mechanical linkages. The farther the tunnel proceeded, the more of these linkages would be needed, and the more power would be lost in transmission. It seemed inevitable that the sheer would eventually stall out. The everconfident Maus felt he could fix on the fly whatever problems cropped up, but the doubters were not convinced. Then, after the political convulsions of 1848, which left Europe feeling less optimistic and expansive in every sphere, Maus’s funding was pulled. A little more than ten years later a tunnel was built close to Maus’s route, but it was done with drill and blast, and it relied on vastly improved ventilating technology and drills that ran on compressed air (created with waterpower) rather than steam. (Today’s drills run on electricity from outside.)

With variations, the story was repeated around the globe, year after year, for the next century: an enormously challenging project, a brilliant engineer, an impressive piece of machinery, admiring visitors, enthusiastic speeches about human progress, and, finally, disappointment. In 1851 Richard Munn & Company of South Boston built a huge machine (for its time), weighing 75 tons, to tunnel through the Hoosac Mountain in northwestern Massachusetts. It jammed before it had gone 10 feet. The entire Hoosac railroad project, which had been sold in part on the promise of the machine, bogged down and became something of a scandal. In 1856 one of the most famous engineers in the country, Herman Haupt, announced his intention of rescuing the Hoosac project with another tunneling machine. Haupt was so confident that he funded its development out of his own pocket. His machine died before it had penetrated even a foot, leaving Haupt bankrupt.

Yet the engineers kept at it, ransacking the repertoire of mechanical design: drums, arms, pistons, steam, compressed air. Nothing worked. Several reasons for this sorry record can be found in subterranean working conditions, which are extremely hostile to machines. One of the most important, however, was conceptual. Many of the designs were based on the idea of taking a rock drill and scaling it up. There is some logic to that: A hole made by a drill does look like a little tunnel. It seems to follow that all you need to do to make a tunneling machine is to build a big rock drill.

But the fact is that the amount of energy needed to cut rock is partly a function of how small the resulting fragments are. The smaller the pieces, the more energy is required. Drills grind small, which means that a drill that could hollow out a cylinder 10 or 20 feet in diameter would be a real power hog. Even if you could generate so high a level of power and carry it to where it was needed, the forces involved would almost certainly overstress your equipment. And if you tried to protect your machine by exerting a little less power, it would seize up against the rock. In the locution of the trade, it would become “muckbound.”

By the early 1930s the message had sunk in, helped along by the arrival of the Depression. The engineers gave up. According to Barbara Stack, the preeminent historian of this technology, “For the next twenty years … few, if any, patents for rock machines were submitted by engineers, nor were any units built.” A new method was needed, but the times were not right for the huge investment of money and resources that would be necessary for its development.

A few years earlier, in 1927, James S. Robbins had graduated from the Michigan College of Mines (now Michigan Technological University). According to his son Richard, he knew nothing about tunnel-boring machines, and of course he had no reason to, since by then they were generally assumed to be a dead end. For nearly two decades he punched his ticket around the industry, doing hard-rock mining in California and gold placer mining (washing or dredging river sediments) in Alaska, among other jobs. After the war he set up in Illinois as a consulting engineer for the coal industry.

In 1952 Robbins welcomed a tunneling contractor named F. K. Mittry to his office. It is an integral part of this story that the people who dig tunnels were and are a different breed from the contractors who put up hotels and office buildings. First, tunnel construction is highly stressful, both physically and mentally. The working area is tiny, the lighting is usually terrible, and the constrained dimensions mean you are generally about two inches from having your arm torn off by some enormously powerful machine. The medium, the ground, may switch instantly to anything from tough limestone to water-saturated gravel to sand to mud, and any of those changes can plunge an excavation into disaster, either by flooding or by collapse. While the money is excellent, the risks of bankruptcy or injury or death are higher than most people like to bear (which is why the money is excellent). Tunneling attracts only those who like long odds and big bets and can stand very hard work for a very long time.

MITTRY WAS JUST SUCH A CHARACTER (AS WAS ROBBINS). He had just won a bid to dig a water diversion tunnel for a dam outside Pierre, South Dakota. The bedrock around Pierre was so riddled with cracks that geologists had given it a specific label: Pierre shale. The fragility of Pierre shale made it scary stuff to blast in, since you never knew what would end up falling on your head when the charges went off. In any other profession, people figure out solutions to problems like these before they sign the papers, but this was tunneling. Mittry had just gone ahead and bid, figuring there would be time to worry about Pierre shale if he won. Now he had won, and he was visiting consultants, shopping for ideas. It is a measure of how intently he was scanning the ground that he had come knocking on the door of an expert in coal-mining machinery for help in digging a tunnel.

Robbins had an idea for him. The mining industry had just begun working with a technique for cutting coal that used no blasting. The idea was to push a group of metal fingers or picks, like the tines of a fork, into the coal face and then rotate the group, scoring deep circular cuts. Freely rotating “wedging wheels” or “bursting wheels” were suspended between the tines; these shattered the weakened mineral off the face. The head carrying this pick and wheel assembly would rotate once and then retract, the coal would be shoveled up by hand, and the process would repeat.

The situation, of course, would be very different in tunneling. To begin with, tunnels have to be dug to a much higher degree of precision than mines, and the dimensions of the work face are much different too. However, Mittry was low on alternatives, so he commissioned Robbins to design a machine based on the pick-and-wheel idea. He took delivery in 1953.

Like all such machines, Mittry’s Mole (as it was called) was impressive to look at: 125 tons, 90 feet long, with a diameter of almost 26 feet. Unlike its predecessors, it was impressive in performance as well. The rotating plate shattered the rock like so much peanut brittle, pushing the tunnel forward at rates of up to 160 feet in 24 hours. This was a breathtaking number, almost 10 times as fast as most contemporary drill-and-blast projects. Robbins might not have built the world’s first tunnel-boring machine, but he had done better: He had built the first one that worked. He had beaten a century of the profession’s most famous minds.

Such a feat calls for a moral. Perhaps the lesson is that important innovation is not just a matter of ambitious vision or engineering research in academic institutions; perhaps sometimes advances are made when good engineers are tightly focused on the spécifie problems of a specific client. That’s a plausible inference, and it was soon tested. The success of the machines on the Pierre project (Mittry ordered several more, and they all performed wonders) led to a small flurry of contracts. And when those machines went into the field, they died in all but the softest rock.

It turned out that Robbins had been lucky. The same properties of Pierre shale that had made Mittry reluctant to blast, that had made him so eager for alternatives, also made the rock a perfect medium for Robbins’s technology. The rotating head had shattered the rock so effortlessly because from a geomechanical point of view, it might as well have been glass. Few contractors were lucky enough to find such accommodating material on their jobs.

In a sense, the curse of the tunnel-boring machine had returned, but this time there was a big difference. A number of engineers—Robbins, his crew, Mittry’s crew, and visitors to the Pierre site—had now seen a tunnel-boring machine work the way it was supposed to. It was one thing to look at blueprints or listen to speeches or gaze at a machine sitting in quiet magnificence on a factory floor, and another to see a radical new idea, basically a prototype, go buzzing through the ground like a terrier after a rabbit. Robbins now knew in his bones that this technology was a very big deal, despite all the problems that no doubt lay ahead. He started James S. Robbins and Associates (later the Robbins Company), the first operation dedicated exclusively to the manufacture of tunnel-boring machines.

One of his early machines was built for a sewer tunnel in Toronto. The headache here (there was always a headache) was that the drag picks kept hitting hard rock and snapping off, which meant shutting down the machine for maintenance over and over, wasting time (not to mention that the drag picks themselves were not cheap). Robbins kept searching for the open door his gut told him was there.

One day he decided to strip the picks off the rotating head altogether. That was counterintuitive, since the theory was that the picks were primarily responsible for the cutting, with the disks basically just cleaning up. But the engineer’s intuition proved out, as the machine ran just as fast as before, only without the pit stops. It turned out that the bursting wheels, which now started to be called cutter disks or wheels, had been doing the real work all along.

In retrospect, it is easy to see why. Every natural rock is riddled with cracks and flaws on several scales. When the cutter wheel pushed down on the rock, the compressive force it introduced concentrated around these weaknesses, with most of the compression organized around the worst flaws. Exert enough pressure, and the cracks will extend into the medium. As the wheels roll on, the cracks spring open, splitting the rock still further. All this happened in order of weakness, making big flaws even bigger.

Cutter disks enormously improved the efficiency with which energy could be channeled into cutting rock. In today’s tunnelboring machines, the cutter disks are typically one to two feet in diameter and are pressed into the rock under 30 tons or so of pressure apiece. A tunnel-boring machine’s cutter head may contain several dozen cutter disks and rotate at around 10 rpm.

The Toronto project was important to the history of the device for a second reason. The most taxing and time-consuming part of tunneling is not breaking the rock, but shoveling its fragments into railroad carts and pulling the carts away from the face. For the Toronto machine, Robbins set up an ingenious system of buckets that rotated with the cutter plate, scooping the “muck” up off the floor and dropping it onto a conveyor that carried it back for disposal. With this bucket-and-conveyor system, Robbins had extended the automation challenge another step: The tunnel-boring machine had become a tunnelboring-and-muck-extraction machine.

By the late 1950s tunnel-boring machines had developed to a point where Maus and Haupt would have recognized them as the embodiment of their dreams. When they worked—which, granted, was by no means all the time—they tunneled at two to three times the rate of drill and blast through the same ground. Such speed meant huge savings and fewer injuries and fatalities. In theory, at this point contractors and clients would have recognized that the day of revolution was at hand. In reality, however, nothing of the sort happened. Most of the contractors stuck with drill and blast. They had several reasons for holding back. Until then tunneling had been a pay-asyou-go operation, with low capital costs and high operating costs. No great investment in equipment was needed to get a job under way. But these big machines had to be paid for up front, and they were not cheap (more than a million dollars apiece). In the terms of the trade, they imposed very high mobilization costs.

MOST IMPORTANT, THEY WERE STILL NOT RELIABLE. Humans may have been slow, but if you put a crew underground, you could pretty much bank on getting some footage every day. A tunnel-boring machine went faster when it was going, but when it broke down, you might have to waste days waiting for Robbins to fly in a part or an engineer. Or you might need to disassemble it and pull it out of the tunnel to perform the repairs. If it hit the wrong ground, it could literally shake itself to pieces. In the very worst case, in which the boring machine had to be abandoned altogether, the tunnel would have to be completely rebuilt to accommodate human crews. This would be disastrous for a contractor working under a serious penalty clause. True, tunnelers are gamblers, but even gamblers have their preferred ways of risking their money.

Then in December 1958 James Robbins died in an airplane accident. This was the worst possible time for the struggling new technology to lose its Moses. Contractors might not have known anything about tunnel-boring machines, but they knew, liked, and trusted Robbins, who was skilled and ingenious and understood the business. When he died, they had no one to turn to. His son Richard took over the company, but he was a young man, only a couple of years out of Michigan Tech, and an unknown face in a community that valued experience above all.

The company swung from client to client until the late 1960s, when the technology received another boost. At that time the city of Chicago announced an enormous tunneling project, the biggest in history, amounting to billions of dollars and decades of steady work. (As of today, the project is scheduled to wrap up around 2012.) Every tunneling contractor in the country scrambled for a piece. And in the fine print they found something hair-raising: No contractor would be allowed to bid on the project unless he brought a tunnel-boring machine to work with him.

Chicago’s problem was that every time a rainstorm passed over (which, if you know Chicago, is not rare), its wastewater treatment plants would overflow, dumping Chicago-size volumes of untreated sewage into Lake Michigan. That effluent would then travel through the rest of the Great Lakes, afflicting communities in two countries. (See “Chicago’s War With Water,” Invention & Technology , Spring 2003.)

Eventually the outrage and litigation mounted to such a pitch that the city agreed to build a vast system of holding tanks, 15 billion gallons in all, large enough to hold the rain runoff until the treatment plants could catch up. More than a hundred miles of tunnels would be needed to connect the sewers to these tanks. With drill and blast, a project this size would take forever, and it could not be overlooked that some of those tunnels would be located under politically sensitive communities, which tend not to like having their china rattled.

On the one hand, the tough bedrock that runs under the city was right at the limit of what the tunnel-boring machines of the day could handle. On the other, a technology that held out any chance of getting the work done within a lifetime was irresistible. Public works is a famously riskaverse profession (as opposed to the contractors that public works people hire), but this time the city’s engineers decided to throw the dice. To overstate only slightly, from a contractor’s point of view, they imposed a requirement that no one could afford to meet for a job that no one could afford to miss.

All across the world the big construction-equipment companies, like Hughes Tool and Krupp of Germany and IngersollRand, set up their own R & D operations, competing with one another to make the biggest improvements fastest. “When the project began, the hard-rock tunneling record was about 600 feet per month,” recalls Howard Handewith, who is now a tunneling consultant based in Seattle. But in Chicago, “whenever we had a month of more than a thousand feet, we had a party. We had some humdinger parties.” (Tunnelers are famous for their parties.) Gradually the machines grew tougher and faster. Soon 1,500 feet per month became routine, then 2,000. “This is where the tunnel-boring machine industry grew up and cut its boring teeth,” Handewith says.

Today a serious boring machine often turns in a production rate of 4,000 feet per month. The newer models come with automatic lining installers, in which a robot arm picks up precast lining segments and clicks them into place around the freshly created interior walls like so many Lego bricks. Such machines integrate all the functions of tunneling into a single device. They can cut through almost any kind of rock and often carry high-tech imaging devices that allow them to look ahead to see what kind of rock they will hit. Some are more than 40 feet in diameter.

There are now probably around 120 of these splendid machines working at any one time around the world. This number reflects not only the domination of tunneling by tunnel-boring machines (drill and blast is now used mainly on short tunnels) but the world’s increasing reliance on the underground. Almost every conflict between surface uses (for instance, between widening a highway and open space) can be solved by putting one of the uses underground.

The development of the technology is nowhere near a plateau. Right now all tunnel-boring machines are custom-built for specific project specifications and geologies. At some point in the near future the industry expects to see a universal machine, powerful and versatile enough to handle any job. This will lower costs both by standardizing design and by improving the market for used machines. Improvements in materials science will soon allow cutter wheels and their bearings to be built of a supertough and microscopically flawless material that will allow the machines to run for hundreds of miles between pit stops. Finally, it should soon be possible to control the machines from the surface, eliminating all the expensive features and procedures now needed to keep humans safe. Small tunnels are already being dug this way.

COULD A MACHINE LIKE TODAY’S HAVE BEEN BUILT IN THE 1840s? It’s not inconceivable. It would have been much smaller, of course, and the problems of ventilation and power would have remained insuperable. But a modern-style tunnel-boring machine could probably have been built long before the first Robbins model if anyone had come up with the right concept. As the industry dreams of a tunnel between Japan and Korea, or even a world-spanning subway that runs on magnetic lévitation, it’s worth remembering how one contractor with a tough job and one consultant with a great idea did something that had baffled the greatest minds in engineering for more than a century.