As more and more active restraint devices are added by vehicle manufacturers for occupant protection, the history of driver frontal airbags illustrates that the design performance of such devices for in-position (IP) occupants often have to be limited in order to reduce their aggressiveness for out-of-position (OOP) situations.
As of today, a limited number of publications dealing with FE simulation of airbag deployment for OOP are available. The objective of our study was to evaluate the feasibility of airbag deployment simulations based on an extensive set of well-defined physical test matrix.
A driver frontal airbag was chosen (European mid-size car sample) for this study. It was deployed against a force plate (14 tests in a total of 6 configurations), and used with Hybrid III 50th percentile dummy (HIII) in OOP tests (6 tests, 4 configurations).
Special attention was paid to control the boundary conditions used in experiments in order to improve the modelling process.
The initial positioning of the dummy (chin against the top of the steering wheel rim, and back of the torso parallel to the plane of the rim) for both physical and numerical dummies was maintained from 23 targets digitized using a 3D Faro arm. Specific test position/conditions that were deemed important were repeated to understand the sensitivity and variation.
The software used for the FEM simulations was Radioss, using uniform pressure method.
The bag was meshed and folded using Excel and Matlab routines. The inflator characteristics were adapted from data provided by the inflator manufacturer. The body-block test conducted at 7 mm was used to tune the different model parameters and the remaining body-block, 50th%le HIII OOP and plate tests were used for validation.
The results show comparison of simulation and tests records.
The simulations show a satisfactory matching of the test results within the first 60 ms and capture the key events of the bag deployment in a promising manner.
The major study limitation remains the confirmation/validation aspect of the study as opposed to prediction. Indeed, the model was developed based on the physical tests. The airbag deployment pattern too was found to be very sensitive to physical and numerical input parameters in the test and models respectively. It is not yet possible to transfer the methodology in order to design an airbag a priori. However, such a model is very useful for the understanding of the loading patterns, injury mechanisms and sensitivity studies.