Better Chemistry Through SoftwareBetter Chemistry Through Software

Most innovations in the chemical industry come out of the R&D labs, where scientists experiment with various composites under an array of conditions, looking for next-generation materials for making everything from paint to airplane parts. But research work can be tedious, time consuming, and labor intensive. New and evolving technology tools--including complex computer modeling and analysis tools--help make R&D endeavors in the chemical industry more efficient and effective in one sector in particular: plastics.

New software tools can help chemists reduce the number of experiments they conduct. Instead of testing combinations of polymers to measure how they react to temperature in 1-degree increments, scientists can test at 2-degree or 5-degree increments; the software uses that data to calculate the in-between temperatures. "You do the iterations on the computer to try different combinations," says Don Schomer, chairman of the American Plastics Council automotive group. "What took days to do now takes hours. It saves costs and time."

In addition, tools such as MoldFlow's flow-simulation software help plastics companies figure out the best injection-molding processes for specific materials, so that product flaws and weakness are minimized at the design phase. Then that information can be passed on to customers, such as auto parts makers, which use it to improve their designs.

CHEMICAL ROMANCE

Automakers have used plastics in the interior of vehicles for many years, but lighter, more cost-effective materials in a vehicle's engine area, for instance, could help make that megaton SUV more fuel efficient and therefore more attractive to cost-conscious consumers. Those new materials also could prove to have fewer failures in the field than traditional steel parts, lowering warranty costs as well, Schomer says.

That's why collaboration is under way among chemical companies to develop computer modeling and visualization techniques for predicting the performance of new polymer materials in automobiles. The use of next-generation tools for developing polymer products employing "long fiber" or "reinforced thermal" plastic materials for automotive parts is one of the hottest areas in the industry, says Guan Chow, senior manager of engineering for Magna Steyr, formerly Porsche Engineering Services. The evolution of computational tools for polymer-based parts "is like having a new drill bit to use with the drill," Chow says.

Such analysis tools fall under the banner of "virtual product development," says Scott Burr, Dow Automotive's engineering material science leader. Six or seven years ago, you'd have needed a supercomputer to run such tests, like fluid analysis in a mold, Burr says. Now most of these programs run on PCs. Most chemical companies, Dow included, use a combination of homegrown applications and third-party software tools, from vendors such as Altair Engineering, LS-Dyna, and McNeil Schwindler, for these design and analysis chores. Dow still uses its supercomputer for bigger projects, like crash-test analysis, to watch the interaction of various parts under stress.

Chemical companies also tap the computational resources and expertise of government research labs such as Argonne National Labs. Over the next four years, Argonne is building a center for massively parallel supercomputing, funded by the Department of Transportation, to perform extremely complex calculations related to fluid dynamic analysis, crash-simulation analysis and visualization, and traffic analysis.

Argonne's Technology Transfer program provides research findings to companies in the automotive, chemical, and manufacturing industries, as well as commercial software companies, from the lab's analysis using its massive computing power, says David Weber, research program director of Argonne's energy systems division. These companies can incorporate into their computational tools "energy analysis codes" developed by Argonne--software programs that perform complex calculations that, for example, analyze air flow around a vehicle. "These software codes are constantly evolving," Weber says.

Eventually, Argonne's new computing center might help chemical companies analyze the performance of polymer-based auto parts in crashes. "Next-generation computing will be powerful enough to look at these finer details," Weber says.

Says an Argonne spokeswoman: "It's a lot less expensive to do visual crashes than physical ones."