VEXTEC In IBM: Making The World Work Better

Predictive models are also emerging as powerful tools to anticipate breakage. Manufacturers dole out tens of billions of dollars in warranty payments a year.[i][i] Knowing when and how a product will fail would not only save money but also improve the safety of workers and customers.

This is what a Tennessee startup called VEXTEC is attempting to do. Funded in part by the Defense Advanced Research Projects Agency and using a database of the world’s known metals, VEXTEC has developed a way to predict the durability, performance and lifetime cost of machine parts by simulating the behavior of their component materials. Working for clients ranging from the US Navy to medical device manufacturers, the company simulates the workings of turbine blades, automotive axles and other machine components to reveal how various metals behave under differing levels of stress. “The problem that all industries face is that products come off the assembly line looking good, but they fail prematurely,” says Loren Nasser, the company’s CEO and cofounder.[ii][ii] “Product development has always been a trial-and-error process. Failure incurs bad press, warranty costs, recall costs and the loss of consumer confidence. That’s what we’re trying to change.”

CAD/CAM software has made usage simulation possible on personal computers for years. But rather than just virtually swinging a tennis racket 250,000 times, VEXTEC breaks it down into 250 million particles and simulates stress on each of the 250,000 rackets a manufacturer will make in a given year. The company’s proprietary simulations are written in FORTRAN, the classic programming language invented at IBM in the 1950s, because, as VEXTEC vice president Frank Priscaro said, “Nothing handles math quite like it—as old-time IBMers will know.”[iii][iii]

VEXTEC’s business is based on a materials library that was digitized thanks to recent advances in computational processing, such as cloud computing. Combine that information with a client’s usage data, and insights begin to emerge. “As manufacturers gather more data about what they’ve made—from sensors, from statistical analysis, from usage profiles—our software gives them a way to make sense of it,” Priscaro said. “If you know when a component is about to go, you can take steps to minimize the impact. So things like airline maintenance will be far more efficient because they won’t be guessing about how long their parts will last, or taking them out of service prematurely. They’ll know.”