Virtual safety tests save money on product research

June 9, 2013 at 12:00a.m.

The Wichita Eagle

WICHITA, Kan.

Gerardo Olivares likes to crash test dummies on a crash-test sled at the National Institute for Aviation Research in Wichita.

He cranks them back, shoots them forward. Heads, arms and legs snap forward, and their waistlines collapse, crushed in painful-looking imitations of what happens when a body folds around a seat belt at high velocity.

He uses dummies, he said with a grin, “because it’s really hard to get live volunteers to ride the crash-test dummy sled.”

Dummies do valuable work, as his boss, John Tomblin, said. But it is not the work that Olivares will be known for, Tomblin said.

Instead it is possible, Tomblin said, that Olivares will go down in aerospace history as the guy who showed how to do aerospace safety tests in three-dimensional virtual models on a computer screen — making dummies much less important than they are now and making costly physical tests less important.

Olivares is getting the dummies out of aerospace. It will not only save lives, Tomblin said; it will save the industry hundreds of millions of dollars.

The most important work aerospace people have done since the Wright brothers is physically test every part that goes into a plane, Tomblin said.

This has become incredibly expensive. A first-class airplane seat, with all its engineering, can cost $120,000 to $150,000, Olivares said. Break a prototype of one of those, and you’ve spent real money.

Aerospace engineers ruin lots of prototypes in tests. At NIAR, they crank up their giant wind tunnel. They stick electronic parts in a giant microwave oven. They bend, twist, pound, drill, smash, shake, burn and freeze things.

Beechcraft Corp. engineers, when they do their own testing, even cover planes and parts with dust and dirt to see how they perform in a desert environment, said John Kraft, the manager of advanced technology for the company.

“Smashing things from every conceivable angle and every temperature is what shows us how things might break and how to prevent them from breaking,” Tomblin said.

Wrecking a costly seat prototype in a test is one thing. Wrecking them in repeat tests is still another thing.

“But now imagine doing a drop test on an air frame section made from composite materials,” Olivares said.

Drop tests of frame sections — in which a section is lifted with a crane and then dropped — have been standard for decades. But those airframes cost millions.

The cost of getting a plane safety-certified became astronomical years ago, Tomblin said.

“A clean-sheet airplane, a new airplane design that starts with a blank sheet of paper, can now cost half a billion dollars to develop,” he said.

Olivares has thought for years that aerospace needed to convert much of its testing to virtual models, in programs that look like some video games. (“They are not video games,” Olivares said with a grin. “They are mathematical models.”)

Other aerospace people are working with virtual programs, too, Tomblin and Kraft said, but Olivares has pioneered how to do “crash-worthiness testing” and has made NIAR the premier place where that testing is quickly moving to what might be called reality.

Kraft said Olivares has been “incredibly innovative” in pioneering virtual testing. He’s not only reshaping how the industry can do it but is also training people. About half of the 40 people who work for Olivares at NIAR are master’s or doctoral students who, after training, will get jobs in the aerospace industry.

“He’s trained dozens, probably more than a hundred people, who go on to get jobs in aerospace and can pass this knowledge along,” Kraft said. “Everywhere I go now, I run into someone trained by him.”

The Federal Aviation Administration is so confident about his work that it has given him more than $700,000 to do it, Olivares said. The industry has matched that.

Since 2005, he and his team have worked on more than 60 crash-worthiness projects paid for by clients from 12 countries.

Olivares now supervises 40 people as director of the NIAR’s Crash Dynamics & Computational Mechanics Laboratory. Much of what he’s done in his pioneering work occurred while working shoulder to shoulder with other NIAR researchers, he said.

“It wasn’t just me.”

He started by creating virtual testing for that most mundane of parts: the passenger seat. Then he moved on to airplane interiors, creating programs so realistic that you can see reflections on walls and windows, how light plays inside a cabin.

Eventually, he said, his researchers will create virtual reality for structures so realistic that no one will build expensive airplane mock-ups anymore.