Home Home Home

Experimental validation of airbag abrasion test procedure

The validity of a laboratory test procedure for predicting the skin abrasion potential of driver-side airbags was investigated by comparing test results with abrasion injuries resulting from airbag deployment tests with human volunteers. A series of thirty-two airbag deployment tests was conducted using the laboratory test procedure, and a second identical series was conducted using the legs of volunteer subjects. In each series, the tests included deployments with two levels of inflator capacity, distance, module cover retention, airbag fold, and gender. In the volunteer tests, abrasions resulted from all but four deployments. Statistical analyses were conducted to investigate the effects of the test variables on abrasion severity. Abrasions were found to be significantly smaller in area for a deployment distance of 275 mm than for 200 mm. The two folds produced significantly different abrasion severities. Gender and module cover retention did not affect abrasion severity. Results for inflator effect were compromised by a misclassification of airbag modules.

The abrasion prediction procedure is based on measurement of the pressure exerted by the deploying airbag on a rigid target cylinder using pressure-sensitive Prescale film. Previous research had shown a correlation between patterns of peak pressure on the Prescale film and patterns of skin abrasion. In the current study, a correlation between the area of Prescale film exceeding 90 kg/cm2 and the area of injury was observed in tests with the target surface located 200 mm from the module, but not at 275 mm. This threshold level is substantially lower than the 175-kg/cm2-threshold level previously proposed. High-speed films of airbag fabric kinematics suggest that skin abrasion can occur with fabric actions that produce between 90 and 175 kg/cm2 in the Prescale-film test if lateral fabric motion is present. Guidelines for applying the test procedure using the injury tolerance data obtained in human-subject testing are presented.

Research Group: