Restraint System Evaluation - A Comparison Between Barrier Crash Tests, Sled Tests and Computer Simulation

Injury criteria measured in belted dummies in frontal barrier crash tests were compared with results from two alternative evaluation methods: sled tests and mathematical simulation. Five identical cars, the 1975 Volvo 245, were loaded with maximum cargo weight and were occupied by two 50th percentile male dummies in the front seat. Cars underwent frontal impacts perpendicular to a fixed barrier at 30 mph. Dummies were restrained by a three-point retractor, with the retractor mounted in the B-pillar (Volvo production belt). Cars were equipped with accelerometers on the inside of door sills; dummies were equipped with accelerometers in the head and chest and with load cells in the femur. Webbing force was measured between the shoulder and the webbing guide (referred to as shoulder belt force) and between the hip and the outer attachment (referred to as lap belt force). Sled tests were performed with a Bendix sled. The computer model involved the simulation of an occupant with three mass elements (head, torso, and legs) allowed to move in a vertical plane in the vehicle's longitudinal direction. The recorded values from the barrier crash tests and sled tests were compared by two different statistical inference tests in order to establish a possible significant difference between the methods. The two models gave about the same results. When they did not agree, it was due to the confidence interval barely including zero, and, in these cases, the ambiguity is resolved by an additional examination of the physical events of the crashes. The confidence intervals were rather wide due to the statistically small number of crashes. Results of barrier crash tests compared well with results from sled tests, except for the driver's chest severity index and head injury criteria. Only the passenger's situation was simulated mathematically, and results compared well with results from barrier crash testing except for the head injury criterion number. Some differences in belt loads were observed. Supplementary information on test results are appended.