Addressing Safety, Head On

Addressing Safety, Head On


Before the late 1980s, when seat belts and airbags became common safety features in cars, head-on collisions tended to result in serious trauma to the head, neck, chest, and abdomen. In fact, a study by the National Highway Traffic Safety Administration (NHTSA) found that airbags alone have cut fatalities from frontal crashes by more than 30 percent, principally by markedly reducing upper body injuries.

That’s the good news. The bad news is that drivers and front-seat passengers who survive head-on collisions often suffer serious injuries to the lower extremities— particularly the knees, hips, and thighs. While rarely fatal, these injuries often result in costly, lifelong disabilities (one study put the annual price tag at $4 billion, along with 60,000 life-years lost to injury—a measure of lost productivity).

The solution, most experts agree, is to redesign car interiors so they afford the lower body the same protection that seat belts and air bags now provide the upper body. But to do that, designers will need a new generation of tools that help them understand the kinds of insults that knees, hips, and thighs endure in a crash, and to cost-effectively test how well new designs can combat them.

With more than $1 million in support from NHTSA, a WPI research team led by Malcolm Ray, professor of civil and environmental engineering and a widely sought-after authority on highway safety, has spent the last four years developing those tools. The result is a detailed three-dimensional finite element model of the bones, muscles, ligaments, tendons, and skin that make up the lower body.

In essence, the model is a mathematical representation of the body—its structure, its mechanics, and its strengths and weaknesses—that can be placed in the driver’s seat of a virtual car. When the digital car is put through a simulated crash test, the virtual human will react just as a real person would. Impacts with the car interior will create the same forces as real impacts, producing the same fractures and other injuries.

To create the model, Ray and his team—which includes Chiara Silvestri ’08 (PhD), research instructor in civil and environmental engineering—worked from the inside out. They first built digital bones, a surprisingly difficult task. "The material properties in bones change with direction," Ray says. "Different bones will break in different ways for that reason."

In addition, bones are not uniform: most have a hard shell that encloses a spongy interior. To the finished bones, the team added muscles. "Mechanically, muscles are like springs and dampers," Ray says. "They are one-degree-of-freedom elements that work only in tension." Next came tendons (which link muscles to bones) and ligaments (which connect bones to one another). There was little in the literature about how these tissues perform under stress.

To fill this knowledge gap, the team placed bone-ligament units in a drop tower and subjected them to forces consistent with those experienced in automobile crashes. They found that ligaments behave like ropes, Ray says. "If you add weight to a rope slowly, it can hold more than if you drop the weight on quickly. Similarly, ligaments and tendons tend to get more brittle the faster you put tension on them."

As the model took form, Ray and his team sought to validate it by comparing its predictions with data from actual NHTSA crash tests that used cadavers. The model predicted some aspects of the test results almost perfectly, but certain results simply didn’t add up.

"As we looked into it, we found that the tests had not been thoroughly documented and that some of the documentation didn’t agree with what we could see in the test photos," Ray says.

Over the past four years, Malcolm Ray has led a team that has built a finite element model of the lower body that designers can use to create safer car interiors. The model includes realistic simulations of the structure and function of bones, muscles, tendons, ligaments—even flesh, which is represented by the yellow and brown areas in the illustration above of the model’s workings.

Like detectives, Ray and Silvestri pored over the evidence and reconstructed the tests in detail, then adjusted the parameters they fed into the model accordingly. 'We found that very small changes—perhaps just a few degrees in how the leg was positioned—could make a big difference in the outcome,' Silvestri says. After the adjustments, the model’s predictions and the actual test results came into close—though not perfect—alignment.

The difference, it turned out, was only skin-deep. While the model accounted for the weight of the virtual human’s skin and fat tissue, the mass of the flesh had been lumped at the joints, rather than being spread out more naturally.

'The flesh and the bones are not rigidly coupled,' Ray says. 'So on impact, some of the force will go into the bone and some will be absorbed by the flesh. In our model, that wasn’t happening.' When an accurate representation of the flesh was added to the model, the discrepancy disappeared.

'As an engineer,' Ray says, 'you are taught—and you believe with every fiber of your being—that physics is correct. So when physics and your results don’t agree, you try to find out what it is in the evidence or the physics that you don’t understand. Of course, that’s what makes a model better.'

As the model development enters its final stages, Ray and his team are looking forward to seeing their work put into practice in the design of future cars, which, thanks to their diligence, will be able, finally, to protect occupants from top to bottom. The model will likely find other applications, such as helping prevent lower-extremity injuries in skiers or football players.

The team is also finding time to reflect on other lessons they’ve learned. 'I am an old-fashioned structural engineer, 'Ray says. 'I deal with wood, concrete, steel. This project has been a lot of fun because I’ve found that while the geometry of the human body is very complicated and much different from most engineered structures, creating this model has been, at its core, a structures problem. Mechanics, it seems, is still just mechanics.'

Maintained by
Last modified: March 27, 2009 15:49:17