This thesis deals with numerical flow simulations of a cold gas inflator to enhance the predictive capability of deployment simulations for curtain airbags. Previous methods such as tank tests do not provide insight into the flow field. A deeper understanding of flow dynamics is required to improve the simulation quality. This study contributes to a better understanding of cold gas inflators and the numerical method chosen for airbag simulations. The study is supported by LES and URANS simulations using real gas models. The thesis is divided into two parts.
The first part focuses on the development of a numerical method for quantifying the discharge process of a cold gas inflator. The main focus is on determining the relevant quantities: pressure, temperature and mass flow. The simulations reveal a complex and highly turbulent flow field with supersonic and subsonic flows. An influential longitudinal vortex forms in the cold gas inflator, causing a highly dynamic outflow behavior. This vortex cannot be identified with tank tests. The k-ω -SST model simulates the flow field with sufficient accuracy compared to the LES and the provided experiments. The real gas effects are important for quantifying the discharge process. They originate from the initially 660 bar helium reservoir. The Peng-Robinson real gas model deviates from measured pressures by approximately 5-10 %, while the ideal gas assumption gives deviations of about 20-25 %. A simplified simulation model with low computational cost and high-quality inflow data is developed and proves to be a practical approach for airbag simulations. The numerically validated method from the first part is evaluated in the second part of the thesis.
For the second part, the evolving flow field of the cold gas inflator into confined duct systems is investigated through numerical flow simulations. Different generic duct systems typical in curtain airbag deployment are used. A distinctive characteristic of the flow field of cold gas inflators is a turbulent underexpanded jet. The duct systems provide various forms of the underexpanded jet typical of curtain airbags. The larger ducts generate a flow field without the wall interaction of the underexpanded jet, satisfactorily captured by LES and the k-ω -SST model compared to provided experiments. In the simulations of the smaller ducts, a shock train with Ma>10 is formed, demonstrating a sensitive behavior. With the k-ω -SST model, it is challenging to capture the unsteady shock train and achieve a grid-independent solution with the available computational resources. The time-resolved simulations provide deep insights and reveal that the flow field and shock train are highly three-dimensional. Previous studies focused on evaluating the kinematic deployment of the airbag fabric without investigating fluid dynamic influences. The simulations conducted here demonstrate that a highly accelerating supersonic flow – whether in the form of the shock train or the turbulent underexpanded jet – immediately induces a backflow at a branching point, significantly affecting the flow distribution. Branching points occur multiple times in the complex geometry of a curtain airbag, which can significantly affect its deployment behavior. The validated numerical simulations provide a better understanding of the fluid dynamic processes in enclosed environments such as curtain airbags and serve as a basis
Dennis Schütte
Airbag Kaltgasgenerator Unterexpandierter Strahl Shock train Helium Hochdrucktank