Lifesaving Technologies

ORNL is finding innovative solutions for the Department of Defense.

Making drinking water from diesel exhaust could reduce the number of casualties associated with military supply operations.
Making drinking water from diesel exhaust could reduce the number of casualties associated with military supply operations.

George Fisher is focused. His single goal is ensuring that ORNL's substantial research portfolio is used to strengthen national and international security. Fisher, who manages the Global Security Directorate (GSD) Department of Defense (DoD) Programs, views one of GSD's primary jobs as engaging ORNL in research and development designed to address specific challenges for the DoD. "There's an urgency to this work because of real-world implications," he explains. "This country is in a fight. When we apply our technology to solve problems, we are impacting peoples' lives."

GSD also advises DoD of new laboratory developments that might be of future interest. With some 400 years of collective military experience, Fisher's staff has a unique ability to understand both needs and opportunities in the national security arena. "That experience means we can take DoD requirements and translate them into what scientists must know to determine whether they will be able to address the problem. Likewise, we spend a great deal of time with the scientists, so we in turn can explain to DoD what the science might deliver in terms of operational impact."

From diesel to water

A straightforward illustration of the DoD partnership can be found in research under way at ORNL to extract drinkable water from the exhaust of diesel-powered vehicles and generators. The technology's implications have particular value for U.S. troops deployed in remote regions. For example, in Afghanistan, the delivery of water accounts for approximately one-half of the military's logistical burden. Finding a technological solution that wrings drinking water from diesel exhaust could produce a dramatic reduction in the casualties and costs associated with this key aspect of military operations.

"It may sound funny to describe this kind of progress as just improving logistics," Fisher says. "We are talking about thousand-mile supply convoys. Sometimes they get ambushed. People get injured or killed, and there are impacts on families. The 'cost' of logistics, in every sense of the word, is what we are trying to reduce."

To date, the project shows promise. Using inorganic membrane technology to filter out impurities in the exhaust, researchers can extract more than a half gallon of water from the exhaust produced by a gallon of diesel fuel, with no effect on the performance of the vehicle or generator.

Life-extending research

Research designed to extend the life of the U.S. Marine Corps' Light Armored Vehicle (LAV) is another example of ORNL's expanding partnership with DoD. One of the Marines' primary fighting vehicles, the LAV was first fielded in 1983. The Marines would like to keep the LAV in the field for 20 to 30 more years.

New condition-based maintenance procedures could extend the life of the Marine Corps' Light Armored Vehicle and reduce the need for spare parts.
New condition-based maintenance procedures could extend the life of the Marine Corps' Light Armored Vehicle and reduce the need for spare parts.

ORNL is using two approaches to extend the LAV's longevity. Researchers are integrating sophisticated sensors into critical vehicle components, with the goal of detecting when these components need to be serviced or replaced. The ability to monitor the status of these components would enable the Marines to employ a sophisticated system of condition-based maintenance, rather than the current process of servicing components at arbitrary intervals. "We are experimenting with vibration and torque sensors to get a better idea of the mechanical stresses the equipment is under," says project scientist Steve McNeany of ORNL's Measurement Science and Systems Engineering Division. McNeany's group is considering installing slope indicators, cameras and GPS to correlate the terrain the vehicle is moving across with the rest of the data. Understanding how to predict maintenance needs would reduce the number of breakdowns as well as the costs and logistics involved in maintaining the vehicles. "If we can provide a condition-based maintenance system for this vehicle," Fisher says, "the result could influence the entire Marine Corps vehicle fleet."

Over the years, the Marine Corps has experimented with several basic sensor suites on the LAV, accumulating substantial data related to the performance of critical components and vehicle maintenance. ORNL researchers are taking advantage of this extensive cache of performance and maintenance data to uncover correlations between sensor readings and maintenance issues that will help the Marines avoid expensive breakdowns by recognizing when an LAV component needs to be serviced.

The Marines hope that a condition-based maintenance approach could also have a significant impact on the "logistics tail" that accompanies every LAV. McNeany says that a vehicle shipped to Afghanistan is literally followed by a ton of spare parts, what the Marines call the "iron mountain." Reducing the need for maintenance and the number of spare parts in the logistics tail would likely reduce the risks to troops involved in hauling the iron mountain back and forth across hostile territory.

Making sense of it

Yet another ORNL capability—intelligent agent technology—is being adapted to help DoD analyze large volumes of text data quickly and accurately. Intelligent agents are computer programs that rapidly extract relevant information from databases containing thousands or even millions of individual documents. The need for such technology is not as rare as one might assume. Within large military commands, the need to quickly sift through millions of documents is common. Fisher says that a U.S. Army brigade in Baghdad was processing three million reports an hour from its area of operations, with no ability to route the information to a central location and make sense of it manually. Effective analysis was possible only with the use of computational tools. A staff experienced with intelligent agents and data analysis, housed in the world's foremost center for high-performance computing, made ORNL the ideal place to find a solution to this problem. Laboratory researchers have demonstrated that their intelligent agents are capable of quickly traversing a computer network in search of very specific types of information, finding exactly what they are looking for and bringing the information back for people to analyze.

"We typically work with a military or intelligence organization that must analyze many thousands or even millions of pieces of information from a variety of sources," says Tom Potok of ORNL's Computational Sciences and Engineering Division. "Within such a massive collection of information, there may be a report that a terrorist is involved in a plot to blow up a plane, a document that provides the person's profile, another indicating the kind of explosives he might be using, and yet another that indicates where he obtained his funding." The challenge is how to search through this vast collection of data, identify the documents that highlight the plot, and share the information in time to respond. The challenge is enormous in its complexity, both for the government and for computer science, and is one that researchers have been working on for years. Potok says his group has developed some very clever statistical and machine learning algorithms that make it possible to glean important pieces of information previously lost in the mass of data.

Using intelligent agents to locate relevant data enables defense or intelligence personnel to spend less time reading and searching and more time analyzing the relevant information. The process is typically iterative, meaning that the intelligent agent provides data and a human analyzes the results, associates them with other significant information and if necessary initiates a new search. The process is repeated until the results are sufficiently relevant. The computer does not create the solution but instead sifts the available data to the point that a human can identify relationships among the various information threads. Incredibly, Potok believes ORNL will eventually be called upon to perform a similar analysis of information contained in more than a trillion documents.

An achievable vision

Building on these promising successes, George Fisher believes that future projects will draw on experts from an even broader suite of the laboratory's emerging technologies. "In the coming years, we can anticipate a continued emphasis on leveraging the capabilities of national laboratories to address DoD's research and development needs. ORNL is in an excellent position to meet those needs."

"What sets ORNL apart from many scientific organizations," Fisher observes, "is that when we provide a sponsor with a vision of what can be achieved, it is a technologically sound vision based on best-in-class technologies and a vast amount of operational experience. This means, simply, that the solution really can be built and that we can build it better than anyone else."

Scientists are applying high-performance computing to the challenge of quickly extracting relevant information from trillions of documents.
Scientists are applying high-performance computing to the challenge of quickly extracting relevant information from trillions of documents.