MORGANTOWN, W.Va. -- With a few dull thuds, the one-ton bag of high-strength fabric tumbled from the wall of the mock subway tunnel and onto the floor. Then it began to grow.
As air flowed into it through a hose, the bundle inflated until it was crammed tight inside the 16-foot-diameter tunnel, looking like the filling in a giant concrete-and-steel cannoli.
The three-minute procedure, conducted on a chilly morning this month in an airport hangar not far from West Virginia University, was the latest test of a device that may someday help guard real tunnels during disasters -- whether a terrorist strike or a storm like Hurricane Sandy, whose wind-driven surge of water overwhelmed New York City's subway system, shutting it down for days.
"The goal is to provide flooding protection for transportation tunnels," said John Fortune, who is managing the project for the federal Department of Homeland Security's Science and Technology Directorate.
The idea is a simple one: rather than retrofitting tunnels with metal floodgates or other expensive structures, the project aims to use a relatively cheap inflatable plug to hold back floodwaters.
In theory, it would be like blowing up a balloon inside a tube. But in practice, developing a plug that is strong, durable, quick to install and foolproof to deploy is a difficult engineering task, one made even more challenging because of the pliable, relatively lightweight materials required.
"Water is heavy, there's a lot of pressure," said Greg Holter, an engineer with Pacific Northwest National Laboratory who helps manage the project. "So it's not as simple as just inflating and filling the space. The plug has to be able to withstand the pressure of the water behind it."
The idea has been in development for more than five years -- this test was the 21st -- and Dr. Fortune says there are at least a few more years of testing and design work ahead. If the plugs are shown to be effective, they will be made available to transit systems around the country; at least initially, they are expected to cost about $400,000 each.
"We have frequent conversations with folks in mass transit agencies, the true experts in the field, on how this would be deployed," Dr. Fortune said, although he declined to name specific agencies, citing the antiterrorism aspect of the project.
Adam Lisberg, the chief spokesman for the Metropolitan Transportation Authority, which runs New York City's transit systems, said the damage caused by the storm "is certainly going to focus our attention" on ways to protect the subways. "But it's way too early to talk about any particular technology that's been proposed," he said.
In all, seven of the city's 14 under-river subway tunnels were flooded during the storm, as were several major highway tunnels. Dr. Fortune said that plugs might work for highway tunnels, too, but that the larger size of those tunnels created additional technical challenges that would have to be overcome.
Work on the plug began in 2007, after Ever J. Barbero, a West Virginia professor whose specialty is the use of advanced materials in engineering, was contacted by a Homeland Security official looking for outside-the-box ideas on ways to keep a subway system from flooding if an underwater tunnel were breached -- by a terrorist bomb, for example.
"I didn't know anything about this," he said. "Then I found out what happened in Chicago."
Dr. Barbero was referring to a 1992 episode in which an abandoned freight tunnel under the Chicago River was breached by a crew sinking bridge pilings. That led to a flood that caused close to $2 billion in damage to downtown buildings as the water spread underground.
Dr. Barbero came up with an idea and shared it with Homeland Security officials. "I said, 'We'll put an air bag in a tunnel,' " he recalled. The department was intrigued -- officials were familiar with European efforts to develop plugs to seal tunnels from gas and smoke after a series of deadly tunnel fires -- and decided to finance the project. About $8 million has been spent so far.
Dr. Barbero realized that the forces exerted on the pressurized plug, and the need to rely on friction against the tunnel walls to keep it in place under the onslaught of floodwaters, meant that it had to be made from very tough materials. Experts from ILC Dover, a company in Delaware that makes high-strength soft structures like spacesuits and the force-absorbing air bags used for some of the Mars rover landings, suggested fabric made from Vectran, a strong but lightweight yarn spun from a liquid-crystal polymer.
But the first full-scale plug, made with a single layer of Vectran, failed during a pressure test in 2010. "It ripped right down the middle, from back to front," Dr. Fortune said.
So Dr. Barbero and ILC came up with a three-layer plug, with the outer layer consisting of woven Vectran belts. It is designed so that the tearing of one belt will not cause a catastrophic failure.
"It took time to figure out how to make it work with such huge pressure," Dr. Barbero said. "But it also took time for the team to understand how a subway works, what are the constraints in the tunnels."
A subway tunnel is hardly a pristine environment; it is full of grease and grime -- and, often, rats.
"That's something we've talked about," Dr. Fortune said. "We've actually put Vectran samples in tunnels, to see if rats ate it. They didn't."
There are also obstructions like tracks, as well an electrified third rail, pipes and safety walkways, all of which could cause gaps between the plug and the tunnel walls. Most of the obstructions can be dealt with by modifying a short section of the tunnel to accommodate the plug, which is 32 feet long when inflated. Sharp corners can be curved, flush tracks of the type used at grade crossings can be installed, the third rail can be discontinued for a stretch, and pipes can be made to swing against the ceiling.
Those modifications will reduce potential gaps but not eliminate them. In the most recent test, when Dr. Barbero and a colleague, Eduardo M. Sosa, inspected the front of the plug, they discovered a two-inch gap in one corner. The procedure called for filling the plug with water to pressurize it further, and then introducing water behind it to simulate a flood. But a plumbing failure, unrelated to the plug, ended the test prematurely. It was repeated successfully several days later, Dr. Fortune said, and the plug held back all but a small amount of water.
Dr. Fortune said that was the goal of the design: to keep the inevitable gaps small enough that any water that made it through could be easily pumped out. "The idea is that in a real-life situation, you reduce unmanageable to manageable," he said.
Placement of the plugs would vary depending on a transit agency's particular needs, Dr. Fortune said. Since water pressure increases with depth, shallower placement would make it easier for a plug to hold back floodwaters. And while a plug at either end of an underwater tube would protect against a tunnel breach, Hurricane Sandy showed that flooding can occur in many other ways -- through ventilation grates and station entrances, for example. So in a system like New York's, there may be other strategic places to install plugs.
"The positioning of the plug and potential sources of leakage all come into play in considering how to use these things," Dr. Fortune said.
Dr. Barbero's initial idea was to put plugs on rail cars so that they could be sent to any location, as needed. "That sounded really far-fetched" to Homeland Security officials, he recalled. He came up with a more permanent installation: the plug sits, folded up, inside a long container snug against a tunnel wall, out of the way of passing trains.
Pumps and other equipment are even farther out of the way. On a signal, the container opens up like a parachute bag, and the plug unfolds and inflates.
As with a parachute, proper folding is critical. "If you fold it wrong, it won't come out right," Dr. Barbero said. The procedure involves laying the plug flat and folding it in on itself, to avoid wrinkles that might cause a very large gap.
"It's a pancake, and then it becomes a burrito," he said.
One of the reasons it has been tested so often is to ensure that it will unfold properly every time. "This is something like the air bag in your car," Dr. Barbero said. "You depend on it. You don't think about it. You never use it, probably. But if you need it, it's got to be there, and it better work."
This article originally appeared in The New York Times.