Derek Warner studies how materials fail.
“If we can understand how things break and bend, we can make them better,” he says.
He does this by modeling what happens at the atomic scale. By doing so, he hopes to predict the behavior of macroscopic structures. “Right now, there’s a disconnect between modeling things at the nanoscale and reality,” says Warner.
Expanding structural analysis and design optimization to the nanoscale is necessary, says Warner, for solving pressing problems, such as how to transition to a hydrogen fuel cell economy. To speed that transition, some have suggested using existing natural gas pipelines to transport hydrogen. But over time, hydrogen makes steel brittle and brittle steel breaks easily—with potentially catastrophic consequences if that steel contains explosive hydrogen.
While it’s clear that hydrogen makes steel brittle, scientists aren’t sure how. “We really need to understand these processes,” says Warner, noting that radiation has a similar effect on the metals used in reactor vessels. “We need to be able to predict when these materials are going to break.”
Material failure isn’t always unwanted. Take the ceramic used in body armor. By breaking into pieces, the ceramic absorbs the energy of a bullet. “I’m also looking at the dynamic fragmentation of ceramics—how many pieces does it break into and the distribution of sizes,” says Warner. “When you drop a plate from a table, it may break into five large pieces, but when you throw it out the window, it breaks into many small pieces.”
Warner uses supercomputers to simulate the deformation and fracture of solids with atomic resolution at times-scales approaching one millionth of a second. “I use the biggest computers I can get my hands on, but you still have to be clever with your models or you’ll end up with just a billionth of a second instead of a millionth.”
He’s also looking at the untapped computing power of networked desktops to model material behavior. Such grid computing, famously used by SETI in its search for extraterrestrial intelligence, has been applied to other problems such as finding a cure for AIDS. “No one does that yet for the classes of problems that I work on,” says Warner.
In his course on scientific supercomputing, Warner hopes to share his outside-the-box approach with students. “It’s basically how to solve problems on big computers,” he says, “but I want to be a little unconventional, like solving problems on PlayStations in parallel because that’s cheaper to do.”
Though drawn to the intellectual challenges of basic research, Warner’s motivation is to solve real world problems. “I knew I wanted to be an engineer, but I just thought I wanted to design mountain bikes and weld stuff,” he says. “That all changed when I started doing research. It’s just a lot of fun.
“Now, I’m really more of a scientist,” he says. “I’m asking ‘Why?’ But as soon as you understand why, you can go to design and kick butt.”
Prof. Warner's Web page
