Supercomputing & Modeling Innovations

Liquid Kevlar

Scenario: The use of Kevlar in current body armor designs requires multiple layers of fabric to provide sufficient protection, making the resulting composite material too heavy and inflexible for complete coverage. This design is especially inadequate for protecting a soldier's extremities—where a significant number of critical wounds now occur.

Advanced lightweight body armor can provide our soldiers with significant technological advantages in combat situations. Working with the Department of Defense (DoD), HPTi scientists in the DoD's Programming Environment and Training (PET) program have participated in providing an initial baseline toolset that enables new materials to be modeled and simulated. With these simulations of how various fabrics molecularly bond with Kevlar, scientists are envisioning new materials that combine the flexibility and weight of fabric with the structural resistance of Kevlar. These new fabrics can be used for advanced lightweight body armor. From these preliminary results, HPTi personnel are working in partnership with DoD to design, optimize, and accelerate the insertion of these protective fabrics into DoD systems.

Modeling Chemical Weapons Dispersion in Cities

Scenario: Intelligence uncovers a plot to detonate a chemical weapon in downtown Chicago. Thousands could be killed or injured. Government officials and first-responders need to know how the toxins will disperse. How long will they remain toxic? Where might survivors be clustered?

With the threat of terrorist chemical attacks a clear and present danger to the United States and its armed forces, military planners need tools to accurately indicate how toxic plumes will travel and degrade as they move through urban environments. HPTi is playing a key role in adapting the use of reactive-flow modeling to predict the flow and degradation of toxic chemicals as they disperse through a city. This work provides the first substantial planning tool to military planners as they seek to understand both the flow of a toxin as well as the concentrations of its degraded products over time. Further development of this tool could help authorities pinpoint areas for search and rescue efforts, allowing for a better use of resources in finding and rescuing chemical-attack victims. Ultimately, the better we can predict and plan, the more lives we can save in the event of an attack.

Protecting Against Biological Warfare

Scenario: American troops are stationed within range of an enemy known to have biological weapons. Leaders fear that the weapons may be used, and they need to be able to make effective counteragents quickly. Military leaders turn to DoD researchers to develop promising small molecular inhibitors to protect against these potentially lethal biological agents.

Creating defenses against chemical and biological warfare requires work at the molecular level. Chemical and biological agents and the protection systems against such agents often involve molecules binding to protein receptors. In order to save time and money, researchers use virtual screening through computer programs to identify the most promising molecules for binding. However, previous versions of a screening program could analyze only a single molecule at a time. They also required the program to load a large amount of data at each startup, creating a backlog. These factors could delay the formulation and deployment of new counteragents in the field—with potentially deadly consequences.

By studying the screening program, HPTi's computational scientists have created a way to speed up this process. Chemists and biologists are now able to screen multiple molecules at a time—significantly decreasing the time it takes to identify possible agents or drugs for use in DoD applications—and potentially saving lives.

Weather Prediction Supercomputers

Scenario: The National Oceanic and Atmospheric Administration's (NOAA) Forecast Systems Laboratory (FSL) affects millions of people every day, but few of them realize the influence FSL has on ordinary things like their flight or local newscast. FSL researches and develops new observing and forecasting systems, then transfers those technologies to operational users such as the National Weather Service (NWS), other government agencies (e.g., U.S. Air Force and the Federal Aviation Administration), commercial and general aviation communities, foreign weather forecasting services, and other private interests. FSL depends on the latest supercomputer technology to provide the extraordinary amount of computer processing necessary to support the technology transfer process.

The newest supercomputer in FSL's list of weather prediction tools is the JET upgrade. Designed and integrated by HPTi, JET uses the latest Intel 64-bit processors. It has less than half the processors of the previous system, but it can do the same amount of work in less time, processing over 2 billion operations per second. JET is used by NOAA researchers to conduct real-time weather prediction, weather modeling, and to research new technologies in weather analysis. Faster, more accurate weather prediction will greatly benefit society by protecting human life and enabling commerce.

Modeling the Battlefield of the Future

Scenario: On the battlefield of the future, a commander coordinates an attack from a central command post. Each soldier and vehicle in the field sends back images and data in a barrage of information that—properly networked—provides an accurate and robust view of the situation. The scale of the network is daunting, and the architecture of the network is constantly changing, but the commander relies on its smooth and continuous operation for success.

The Army envisions network-centric warfare where every soldier is a sensor node on a network—a network that must properly capture, route, process, and represent vast amounts of data. To achieve this vision, the Army will deploy thousands of mobile ad hoc networks (MANET). Sheer logistics prohibit the Army from conducting MANET field experiments with thousands of soldiers and laboratory emulations, where each MANET node is emulated by a physical processor. However, HPTi's computational scientists are working with government and academic researchers to create new HPC modeling and simulation techniques for supercomputers that will allow researchers to run practical, large-scale networking scenarios on the order of 10,000 nodes and beyond. The resulting models will enable scientists to test current MANET components and system designs in realistic scenarios-tests essential to developing situational awareness for soldiers in harm's way. The Army depends upon this type of research to develop the networking capabilities of the future.