Customer Interview
Defining Success at the Scientific Level
 |
Dr. Peter W. Chung Team Leader Interdisciplinary Computational Sciences Team, Computational & Information Sciences Directorate (CISD), Army Research Laboratory (ARL) |
Project summary:
I work in the general area of computational sciences with specific emphasis in computational mechanics. We look, in particular, at multi-scale simulation techniques and visualization methods, ranging from the nanometer scale up to and including the meter scale. In my personal work, I emphasize engineering science enriched and, in a sense, merged with computational physics and chemistry.
Who else is involved in your projects?
The Computational Sciences & Engineering Branch of the High Performance Computing Division/Computational & Information Sciences Directorate; the California Institute of Technology; University of Southern California; the United States Military Academy at West Point; the University of Maine; and the University of Minnesota.
Impact on DoD:
The main function of the Interdisciplinary Computational Sciences Team is to perform basic and applied research and develop computational tools to improve software and its capabilities for Army engineers and researchers to study any type of problem, whether it's materials-based problems, soldier specific-based problems, or weapons-based problems. The motivation for us is scientific advancement. It's exciting to think that a lot of the things that we do will have a direct implementation in the war fighting.
For example, one of the things that we strive very hard to understand is the impact of projectiles on our targets. This type of research is very difficult for a university to work on, mainly because of the large amount of data it cannot get. We can do a lot of interesting problems with expertise available only from DoD and address particular concerns that DoD has that no outside entity would be interested in. A lot of the research that gets inspired around the country, if not the world, is actually motivated by us and some of the things that we look at. There are thousands of graduate students and hundreds of researchers all around the world trying to deal with problems and dedicating their careers to issues we sort of identified in some way.
Research is a community; there is always some sense of camaraderie and collaboration that has to occur in order before research and progress can be gained.
Objectives:
What were your initial goals?
To develop new computational methodologies of the behavior of Army-specific systems. I try to use computations to predict things that you'd normally have a hard time to predict. My area of expertise is computational sciences and computational mechanics. To put it another way, I try to learn enough science, enough mathematics, and enough engineering to bring them all together into a computer. At the moment I have five or six projects going on right now, all at different stages, and I work with five or six people.
We use a lot of different software, some we write ourselves, and some of it we buy commercially off the shelf. Others we might modify.
 |
| Graphite in a three-fold symmetric configuration at equilibrium and a reduced symmetry configuration under mechanical load. |
How do you measure a successful outcome?
Success at the scientific level comes in many forms and it's never overnight. There's a reason why Nobel Prizes are awarded to people in their sixties and seventies. You have to be involved in the scientific community, you have to be publishing, getting peer reviews, talking at conferences, and discussing problems with the community. Over time, people will take your work and try to do it themselves. There is a repeatability issue that always comes into science. There has to be independent observerations of the same discoveries that you've made. And when that happens, that is when success begins to show its face. And it builds up over time.
Maybe that's why this job is so hard. Nobody is going to pat you on the back and say, "Good job." If you're a scientist you really should be measuring yourself against the entire world. You have to be doing research that is really important. The last thing you want to be is irrelevant.
All the more inspiring is when you are doing research that has a specific purpose that can be used by the Army to protect its soldiers. That by itself is a rewarding experience. If I can save one life in my lifetime, then I'll die a happy man.
Methodology:
We take a combination of theory, mathematics, and computations and combine them in the appropriate way to address specific problems. In the past we have used abstract mathematical formalisms to develop actual computer software that incorporates theoretical notions that are difficult to prove by themselves through, for example, experiments alone.
Results:
We have had numerous successful outcomes from the myriad of projects we are currently working on. For instance, I'm currently working on a project where we are bringing in the principles of quantum chemistry and combining that with principles of mechanical engineering to predict new kinds of materials, all done on a computer. The new kind of materials comes in many different ways.
Table salt, for instance, has a particular configuration where sodium and chlorite have these boxed orientations and the boxes sort of interweave together. Normally at room temperature, most materials have a sort of configuration. You can do certain things to them - apply an electric field, apply high voltage, apply high temperatures - and when you do that, those molecules and those atoms find new configurations to fit into. And when they fit into those new configurations, they may have a totally different type of property based on those new configurations. These new kinds of properties can be done for a lot of different things. This is the hallmark idea of that setup and nanotechnology, which is building systems, materials or devices at the atomic level.
One of the things that we find is that there have been, historically, a lot of new types of material discovered just by looking at materials you'd find in your basement or something. You apply voltage or high temperatures to the material and, all of a sudden, it behaves in a new way.
The project I'm really excited about deals with carbon nanotubes. Nanotubes are one of the basic building blocks for what a lot of people envision as the future of nanotechnology. Purely on the computer, we have been able to do a lot of calculations and show that nanotubes in and of themselves have many interesting features that can be used.
What we are finding is that with these nanotubes, we can mechanically activate different types of behavior. It's like a transformer; all of sudden it finds a new configuration. All of a sudden it changes its behavior, the material looks different, the mechanical properties look different, the electrical properties look different. The reasons why we do these things is twofold: first, so we can extract some of the behavior from the system. Secondly, we don't want these things happening unexpectedly.
I always say that the greatest physicists and greatest engineers of our history - Galileo, Albert Einstein, DaVinci - if they had supercomputers there is no telling what they could have done. I think, in a way, Albert Einstein died trying to do the math for these problems he could have done in a day if he had a supercomputer. So, in a way, we're going back and revisiting a lot of the things they did with this incredible tool. That in itself is very exciting.
 |
| Dislocation structure under a nanoindenter. |
Significance:
Different aspects of similar problems are being considered in various organizations outside of the government. The one key element that makes our research unique is that we take into consideration scientific research of hostile environments, where we often need to know how things work in high temperature, large loading rate situations. This is especially important for new futuristic technology where we need to make and break things on the computer before they happen in the real world (where they might end in tragic outcomes).
Hardware:
Which HPC systems did you use?
IBMs, SGIs, PCs
How many processors did you use?
1 to 1000.
Number of computer hours used?
More than 100,000 hours
How would you have accomplished your objectives if HPCs were not available?
It would be impossible. HPC is what makes our research possible today.
Did you use any other ARL MSRC hardware?
The Scientific Visualization Lab.
Software:
What software was used?
The software was self-written and needed to be parallelized.
Personnel:
Which ARL MSRC team members helped you with your project?
Tom Kendall, Phil Matthews, and Steve Thompson have helped us run our codes on the HPC platforms.
 |
| Atomically resolved penetration of a rigid penetrator through a ceramic plate. |
Other projects:
What other projects have you worked on using ARL MSRC resources?
Director's Research Initiative projects for 01-03.
What career achievements make you feel proud?
Scientific advancement makes me proud, to be able to work on a problem that leads to results that have never been known before, is accepted by a community of peers, and transitions to useful technology and practical engineering tools.
What would you say has been the highlight of your career?
Working for ARL.
What was your last research project?
Work on using computational quantum mechanics to understand mechanical behavior of materials.
What will you be working on next?
Merging quantum mechanics with engineering science to discover new methods, materials, and phenomena that are important for Army applications. One of the things that deters us from moving forward and designing these new devices and new systems is that we still have one foot in the past, in a sense that we rely on software tools that we have now openly available, but those software tools are inherently limited. They only incorporate so much science, so much mathematics, so much theory, so much engineering.
So one of the things that we really need to do is put into the hands of designers and researchers better software tools that can capture a lot of the things that we really want to look at for these new kinds of devices and new systems. The software available now for nanotechnology is inherently restricted. The computational tools we've developed are specifically addressing those kinds of issues like, i.e., How do you bring new science into these problems, develop new software and make sure that the software solves the problem the correct way with the right science and with the right theory?