Step-out Technology for Virtual Airbag Deployment

NOVA Chemicals initiated a virtual airbag deployment project to advance the predictive capabilities of computer-aided engineering (CAE), as it applies to instrument panel (IP) substrate material performance. Understanding IP substrate material performance using computer simulation lowers developmental costs by reducing the number of tests required during validation. This proprietary technology is also used to proactively direct the design of the airbag system early in the design cycle, which reduces development time and produces a more successful airbag architecture and substrate performance.

To accurately simulate a virtual airbag deployment, the DYLARK CAE development team must account for several variables. These variables include bag fold, airbag and inflator characteristics, airbag design interfaces, contacts, different materials, temperature effects and airbag architecture.

NOVA Chemicals’ approach to each variable requires a four-phase development program. The first phase identifies and understands the variables and inputs for the analysis. An airbag deployment incorporates a multitude of variables that may affect the final result. For example, variables such as the instrument panel design, materials being used in the components and the propellant for the airbag and airbag fold can greatly affect the dynamics of the event. Inputs such as the material properties and airbag pressures also affect the results. Correctly modeling the variables, inputs and their dynamic interactions is critical to achieve accurate results.

During the second phase, the existing technology is accessed and reviewed for software capabilities. Several different software packages for modeling airbag folds and dynamic analysis are investigated to determine the most efficient and accurate code for the analysis.

The third phase creates a baseline analysis to validate the technology and compile a database of lessons learned. This phase reduces the analysis to its simplest form to understand the fundamental problem. Rather than modeling an entire instrument panel, the analysis uses a simple test fixture consisting of an airbag can in a test cell with a simplified airbag door. The smaller model reduces the analysis computation time, complexity of the problem and more readily identifies potential debugging issues in the analysis. From this simplified model a list of lessons learned can be generated and applied to a more complex problem as seen in phase four.

The fourth phase incorporates the airbag can, full instrument panel assembly, supporting brackets and all significant structures to simulate a static airbag deployment. This is the most complex phase and requires the knowledge base acquired from phases one through three. The final phase applies the lessons learned in phase three to a more complex series of events occurring in a full instrument panel static airbag deployment and correlates the event with the virtual airbag deployment.

If the variables and inputs are modeled correctly, the results produce an accurate simulation of an airbag deployment to highlight potential areas of concern in a virtual prototype. Overall, the virtual airbag deployment analysis provides significant cost-saving results by reducing the number of airbag deployment tests for validation.