The diagrams below show how an actual collision at typical city speeds (section 1) is simulated in a crash lab (section 2), and then how the data from that test (section 3) is used at another lab to simulate the accident again with a moving sled (section 4). Engineers often use a sled to test child seats because it allows them to run repeated trials without smashing cars. But both methods have limitations. The drawings below are simplified to make the physics involved easier to follow.

1. THE REAL WORLD When one car hits another in the side, it can crush one or both doors and intrude into the passenger compartment. The merged wreckage rotates and slides off in a diagonal direction determined, among other things, by relative weights and speeds. People in the struck vehicle can be injured by the intrusion, the sudden change in motion, or both. Speeds and angles in our diagram are typical of a collision that NHTSA's test might simulate, except that we assume both vehicles are of equal weight and hit each other right in their centers of mass. Click here for larger image.








2. AT THE CRASH LAB Test engineers emulate the above collision by ramming a so-called moving deformable barrier (MDB) into a stationary car that is to be tested. The MDB's nose is designed to crumple in nearly the same way as that of a real vehicle; its wheels are angled so that its front hits the test car squarely. The angled impact and the slightly higher 38-mph speed of the striking vehicle make up for the fact that the other car doesn't start out moving. As a result, the damage to the test car and its crash-dummy occupants will be similar to that of the real accident. Click here for larger image.










3. THE CRASH PULSE The crash lab records a graphical signature of the impact in the "crash pulse," a moment-by-moment plot of acceleration measured by sensors in the struck vehicle. Its exact pattern depends on the model tested, the sensor location, and so forth. But the pulse is important because it influences things like how an occupant's head might be whipsawed by the collision. Click here for larger image.










4. IN THE SLED LAB The "crash pulse" signature is key to simulating the same type of accident at a different lab that uses a moving sled. Engineers mount the seat and dummy on the sled, then accelerate it in a pattern that matches the pulse. The test--the important part of which is over in a fraction of a second--estimates how a seat will react to the sudden acceleration. But, unlike a crash-lab test, it generally doesn't gauge risk from factors like rotation and structural intrusion. The latter is a major cause of injury in side impacts. Click here for larger image.