Abstract
Keywords: Occupant safety, Airbag deployment, Simulation, CFD, Real gas flow phenomena
The present approach describes an integrated procedure for the analysis of an airbag deployment process. It is based on a balanced combination of numerical and experimental investigations in order to overcome the existing uncertainties.
The first step of the procedure focuses on simulations of flow phenomena inside a gas generator. A next prerequisite of an airbag deployment simulation is the simulation of the fabric folding process.
At the end, the deployment behaviour of a knee airbag could be investigated.
An enhanced physical model was used to characterize the output of the gas generator of an airbag module: The Redlich Kwong equation of state was programmed into the CFD code STAR-CD together with a simple mixing rule for the calculation of the underlying critical gas mixture properties.
The outflow of a pure Helium filled gas generator of a knee airbag module was simulated using two CFD codes, being STAR-CD and CFX. Both of them used a real gas equation of state. As a quick validity check of the applied, enhanced physical models the static temperature distribution was analyzed: It features higher temperatures than results based on the use of the ideal gas law which is due to the inverse Joule-Thomson effect of Helium.
The computed mass flux profiles of both codes showed a nearly identical behaviour which is, not surprisingly, quite different compared to the somewhat questionable data available from can tests.
Next, the complex airbag folding pattern was simulated based on a mostly automated process called ESI SimFolder. Special attention was paid to the compatibility of the resulting model with the requirements of the flow solver FPM inside the crash code PAMCRASH.
The subsequent deployment simulation used a newly developed PAMCRASH FPM option taking the complete CFD output into account, especially flow direction and flow momentum.
As a result of the new approach, the pressure forces exhibited on the fabric walls are much higher than seen in preliminary simulations based on the standard procedure to apply inflow boundary conditions at the inflators of a gas generator.
The correct representation of the flow resistance of internal straps leads to a quite different deployment behavior than seen before.
An overall good correlation with experimental findings was observed.