Content area
Full text
Introduction
Improving car safety means building more crashworthy vehicle structures and providing them with more effective restraint systems. Hence, car manufacturers now incorporate a wide range of safety devices and features into their vehicles, including airbags, energy-absorbing steering columns, and side-door beams. As effective as these measures have been, the industry also needs more efficient methodologies for assessing vehicle safety. Crash tests are commonly used to evaluate vehicles for occupant protection. Although full-scale crash tests can accurately replicate actual accidents, they are complex and expensive. In a crash test, sensing devices and crash test dummies are mounted in the car to measure the displacement, velocity, and acceleration at different points on the dummy or on the car structure. The data from such tests are widely used to assess vehicle crashworthiness and occupant safety.
The crash pulse is one of the useful types of data for describing what occurs to a vehicle's structure in a crash. A crash pulse can be used in the sled testing of vehicles and in the design of airbag control unit. However, crash pulses can also be used as an input to a simulation model to determine how occupants react to a crash. The crash pulse measures energy variations during the impact and can be directly related to occupant injuries.
The correlation of pulse shapes to occupant experience has been the subject of several previous studies. Grimes and Lee[1] examined how the shape of the crash pulse correlates to predicted occupant reaction. They compared predicted occupant behaviors obtained using various approximations of crash pulses with predicted behaviors obtained using crash pulses from an actual crash test. Agaram et al.[2] compared acceleration pulses from vehicle-to-vehicle crash tests with those of single-vehicle crash tests. They also used shape and phase cross-correlation to compare acceleration pulses of single-vehicle crash tests with those of vehicle-to-vehicle crash tests. Linder et al.[3] surveyed a range of crash test results to calculate mean accelerations in different vehicles impacted in the same manner. This study also showed that real-world crashes involving the same vehicle model and producing similar delta-V can yield crash pulses that differ in both duration and shape. Varat and Husher[4] showed that a vehicle's pulse shape can be modeled using readily...