Virtual dummies become smart move in aviation crash tests
/ The Wichita Eagle
WICHITA, Kan. — Gerardo Olivares likes to crash-test dummies on a sled at the National Institute for Aviation Research (NIAR) in Wichita.
He cranks them back, shoots them forward. Heads, arms and legs snap forward, and 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 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 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 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 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 company’s manager of advanced technology.
“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.
“But now imagine doing a drop test on an airframe section made from composite materials,” Olivares said.
Drop tests of frame sections — in which a section is lifted with a crane, 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,” he said. “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.
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.
Aerospace people could learn from car designers, Olivares said. Name a new safety innovation, he said, and automotive people 12 years ago were more open to it than aerospace people.
“Virtual testing, advanced crashworthy structural design, air bags, sensing technology, advanced seat-belt technology,” Olivares said, ticking off the achievements by the automotive industry.
Olivares 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.
He started by creating virtual testing for that most mundane of parts: the passenger seat.
Then he moved on to plane 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.
Tomblin said aerospace engineers now understand how valuable this work is.
“Some of them would still say, ‘I still want to see the (physical) test,’ ” Tomblin said.
“But in the past, most simulation testing has always followed physical testing. Gerardo figured out how to show them virtual tests that showed them exactly how their physical tests were going to turn out. They repeatedly saw how the test part always broke in the exact place that Gerardo’s simulation showed.
“They got real excited, too. They even said simulation testing was going to drive more than future testing. Some said, ‘Do you realize what we can do with future aircraft designs?’ ”