Downloaded from SAE International by University of Edinburgh, Monday, August 27, 2018

2005-01-1299
SAE TECHNICAL
PAPER SERIES

A Study of Driver Airbag Shape Design

with Process Integration
Seongjin Kim, Soongu Hong and Jeongkeun Lee
Hyundai MOBIS

Reprinted From: Occupant Safety, Safety-Critical Systems,

and Crashworthiness
(SP-1923)

2005 SAE World Congress

Detroit, Michigan
April 11-14, 2005

400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760 Web: www.sae.org
 Downloaded from SAE International by University of Edinburgh, Monday, August 27, 2018

The Engineering Meetings Board has approved this paper for publication. It has successfully completed
SAE’s peer review process under the supervision of the session organizer. This process requires a
minimum of three (3) reviews by industry experts.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise,
without the prior written permission of SAE.

For permission and licensing requests contact:

SAE Permissions
400 Commonwealth Drive
Warrendale, PA 15096-0001-USA
Email: [email protected]
Tel: 724-772-4028
Fax: 724-772-4891

For multiple print copies contact:

SAE Customer Service

Tel: 877-606-7323 (inside USA and Canada)
Tel: 724-776-4970 (outside USA)
Fax: 724-776-1615
Email: [email protected]

ISSN 0148-7191
Copyright © 2005 SAE International

Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE.
The author is solely responsible for the content of the paper. A process is available by which discussions
will be printed with the paper if it is published in SAE Transactions.

Persons wishing to submit papers to be considered for presentation or publication by SAE should send the
manuscript or a 300 word abstract to Secretary, Engineering Meetings Board, SAE.

Printed in USA
 Downloaded from SAE International by University of Edinburgh, Monday, August 27, 2018

2005-01-1299

A Study of Driver Airbag Shape Design

with Process Integration
Seongjin Kim, Soongu Hong and Jeongkeun Lee
Hyundai MOBIS

Copyright © 2005 SAE International

ABSTRACT the changes and MADYMO ver.5.4.1 (occupant analysis

software) utility program called BAGGEN was used to
In this study, an integrated process was constructed to model the driver airbag [5,6,7].
perform DOE analysis and find out effects of factors
related to the driver airbag shape in NCAP frontal crash. For automated driver airbag modeling process, the
A finite element driver airbag was modeled using occupant analysis model was integrated into iSIGHT.
BAGGEN, one of the MADYMO ver.5.4.1 utilities. The Also, XML’s format of MADYMO ver.6.1 is compatible
airbag modeling process was integrated with a driver with iSIGHT by having a converting process feature that
NCAP frontal occupant crash model. An additional reintegrates the process [8]. Overall, series of integrated
translation process was also integrated to the processes process were used in iSIGHT and DOE on driver airbag
for the XML-based crash model that was modeled using design factors and additional factors was applied in
MADYMO ver.6.1. iSIGHT was employed to the whole order to find out effective driver airbag design
integration process parsing and the entire processes specifications required for improvement of occupant
from driver airbag modeling to Design of Experiment crash safety.
(DOE) were performed automatically exchanging files
between MADYMO server and iSIGHT server. From DRIVER AIRBAG MODELING
DOE analyses, improved results were achieved
compared to the base design specification and factors To complete the analysis of driver NCAP simulation, at
that dominantly affect passive safety performance were first, driver airbag modeling needs to be done. In this
found out. research, MADYMO BAGGEN utility was used to model
the driver airbag.
INTRODUCTION
The input files for BAGGEN consist of geometry and
In order to achieve a certain level of vehicle crash safety material properties that determines airbag configuration,
performance, not only the barrier test is performed as shown in Figure 1. The geometrical center for airbag
during the vehicle development phase, but also the was fixed at (0, 0, 0) and vent hole center, tether
occupant simulation analysis is carried out to reduce locations and outer radius of airbag were defined from
costs and examine various design specifications. this point. The major material properties such as
Young’s Modulus, Poisson’s ratio, density, and
In the previous vehicle development researches, optimal permeability were defined.
design methodology, considering an uncertain design
factor, has been applied [1,2,3,4]. The most interesting
topic for this research is the occupant safety standard
from NCAP (New Car Assessment Program).
Considering the occupant’s safety and need for
achieving optimal crash, research for various design
factors were performed. However, the design factors for
airbag volume and shape are not being considered
enough during research.

In this research, NCAP was chosen as a crash test

standard to evaluate the simulation model and find driver
airbag design factors improving Pcomb (Overall crash
injury performance). In order to generate the driver
airbag models every time its configuration needs to be
changed, iSIGHT was used for automatically processing Figure 1. Driver Airbag Geometry Definition
 Downloaded from SAE International by University of Edinburgh, Monday, August 27, 2018

Input file that contains data with definitions needed for airbag model generated by BAGGEN were created by
airbag generation was used by BAGGEN to produce MADYMO ver.6.1. The simulation model was correlated
output file called MIN (MADYMO INput). MIN file is used with an NCAP barrier test, and Figure 3 shows the
as part of FEM MODEL block, which contains airbag correlation results of head, chest and pelvis acceleration.
configuration and material properties in MADYMO input In this research, this simulation model was used to
deck. MIN file also contains mesh nodes and elements. perform optimal process integration and DOE.

After creating MADYMO input deck by using MIN file, PROCESS INTEGRATION
tether characteristics and contact nodes were redefined.
Tether length was defined in STRAP block and tether The integration process for this research is shown in
stiffness was calculated by the following equation: Figure 4. The optimization software, iSIGHT ver.7.0, was
used for parsing. Overall process of airbag generation
K = (A × E ) 2L = (W × t × E ) 2L (1) was integrated in MADYMO ver.5.4.1 and XML’s
translation step was included in MADYMO ver.6.1 for the
Here, K is tether stiffness, A is tether cross sectional final completion of the integration process.
area, W is tether width, t is tether thickness, and L is
tether length. PROCESS INTEGRATION USING ISIGHT

Support nodes were also defined in MIN file. In order to MADYMO ver.5.4.1 and iSIGHT were used to integrate
define contacts between the occupant, airbag, and the processes, as shown in Figure 5. This integrated
interior components, support nodes were set apart from process is divided into 3 parts, as shown in Figure 5(b).
all the nodes and the other nodes that were defined as
contacts were renumbered. Simcode0 performs a process of generating a MIN file
by using BAGGEN, as shown in Figure 5(c). The
process of renumbering the contact nodes was done by
NCAP SIMULATION MODEL
coding using Tcl inside the iSIGHT’s “description file.”
Also,
The driver NCAP simulation model in Figure 2 and the

Figure 2. NCAP Simulation Model for Driver

Figure 4. Diagram of Process Integration

(a) (b) (c)

Figure 3. Test and Simulation Correlation Figure 5. Process Integration using MADYMO ver.5.4.1
 Downloaded from SAE International by University of Edinburgh, Monday, August 27, 2018

similar to Figure 5(a), Simcode1 is a process that Finally, STEP 4 is the process that acquires PEAK files
analyzes the driver model. From MIN file, parts that used to analyze occupant injury. “DabShpOpt.xml” is the
include MADYMO input deck were stored with different file that models and analyzes the crash simulation in
file name and were included in the MADYMO input deck. MADYMO ver.6.1, and the file that defines airbag’s node,
element contact nodes, and tether was set to be
Finally, out of all the output files, Pcomb was calculated in included. Thus, in this case of optimizing the driver
the PAEK file by using the reference values of HIC and airbag configuration with the driver analysis model in
Chest g. MADYMO ver.6.1, cushion’s geometry related
specifications such as the outer diameter, vent hole
XML FORMAT EXPANSION diameter, and tether were parsed in BAGGEN input file.
Material properties and other parameters require parsing
In MADYMO ver.6.1 with XML format foundation, in in the occupant analysis model.
order to use the airbag generated in BAGGEN and
include translation steps, integration process described DOE ANALYSIS AND RESULTS
in section 4.1 was categorized into three parts. The
overall process is composed of four steps, as shown in DOE FOR AIRBAG SHAPE DESIGN FACTORS
Figure 6.
DOE was performed for the analysis of design factor’s
STEP 1 is also constructed in ver.5.4.1 for the input file effects on the driver airbag. Chosen design factors were
of BAGGEN. After the translation from ver.5.4.1 to airbag’s outer radius, vent hole radius, and tether length.
ver.6.1, additional parsing of some parameters was HIC, Chest g and Pcomb, that determine NCAP Star
needed in ver.6.1. Rating, and Pcomb was selected as the objectives for
DOE. Table 1 shows each design factor’s base, highest,
STEP 2 shows the process that executes translation and lowest values. Latin Hypercube algorithm was
step from a MIN file that has a compatible format as applied to this DOE analysis.
MADYMO ver.6.1. Because the translation in MADYMO
has ver.6.0 format, translation is executed once more Pareto charts and Main Effect graphs of HIC, Chest g
into ver.6.1 in STEP 3. Here, final generation of XML file and Pcomb are illustrated in Figure 7 and 8. Each design
was used for parsing the input data used in the final factor’s effect on HIC, Chest g and Pcomb is in the same
analysis. order of vent hole radius > outer radius > tether length.
From this result, the vent hole radius was the most
dominant factor in improving the occupant’s injury for the
NCAP analysis model applied to this research.
Compared to the base design specification as shown in
Table 2, DOE analysis results gave 12.0 % improvement
of Pcomb.

Table 1. Design Variables and Values

Lower Base Upper

Variables
Limits Value Limits
Outer 0.3175 m 0.3429 m 0.3683 m
Radius (25 in.) (27 in.) (29 in.)
Vent
0.0125 m 0.0175 m 0.0225 m
Radius
Tether 0.2413 m 0.2667 m 0.2921 m
Length (9.5 in.) (10.5 in.) (11.5 in)

Table 2. DOE Results and Pcomb (A)

Factors Base Model DOE

Outer Radius 0.3429 mm 0.3289 mm

Vent Hole Radius 0.0175 mm 0.0194 mm

Tether Length 0.2667 mm 0.2579 mm

Pcomb 0.0931 0.0819

Figure 6. Process Integration using MADYMO ver.6.1
 Downloaded from SAE International by University of Edinburgh, Monday, August 27, 2018

ADDITIONAL FACTORS APPLICATION different from the manufacturing companies and their
products, two more different types of inflators were
To research other design factors that affect safety chosen for DOE analysis, and they were manufactured
performance, inflator type and steering column axial from different companies. Figure 9 shows the dynamic
stiffness were included as design factors. Table 3 shows pendulum test acceleration for those inflators. This
each additional factor’s value applied in this research. simulation was done on the same pendulum test
Because inflator pressure and mass flow rate are condition just by changing inflators, and the order of type

(a) HIC (a) HIC

(b) Chest g (b) Chest g

(c) Pcomb (c) Pcomb

Figure 7. Pareto Plots (A) Figure 8. Main Effects (A)

 Downloaded from SAE International by University of Edinburgh, Monday, August 27, 2018

was determined by the stiffness of airbag. Inflator type 2 illustrated in Figure 10 and 11. In the Pcomb results,
was applied in simulation model correlation. Steering Column axial stiffness was the most dominant
factor of all, and it was remarkable that steering column
The other added factor was steering column axial axial stiffness and inflator were significant design factors
stiffness. In most cases of driver frontal crash situation, to reduce occupants’ head and chest injury and improve
steering column collapsing load and displacement are NCAP star rating. Table 4 shows the DOE result
important design variables. In MADYMO simulation summary.
model, those parameters are employed as a force-
displacement function in axial direction. Therefore, in
this additional study, steering column force scaling factor
was chosen as a factor to reflect different steering
column collapsing conditions.

Pareto charts and Main Effect graphs of HIC, Chest g

and Pcomb including additional two design factors are

(a) HIC

Figure 9. Pendulum Acceleration of Inflators

Table 3. Additional Design Variables and Values

Lower Base Upper

Variables (b) Chest g
Limits Value Limits
Inflator Type Type 1, 2, 3
S/Column
0.7 1.0 1.3
Scaling Factor

Table 4. DOE Results and Pcomb (B)

Factors Base Model DOE

Outer Radius 0.3429 mm 0.3212 mm
Vent Hole Radius 0.0175 mm 0.0189 mm
Tether Length 0.2667 mm 0.2722 mm
Inflator Type 2 Type 3
(c) Pcomb
S/C scale factor 1.0 0.7
Pcomb 0.0931 0.1103 Figure 10. Pareto Plots (B)
 Downloaded from SAE International by University of Edinburgh, Monday, August 27, 2018

of driver airbag were determined from Pareto and Main

Effect results of HIC, Chest g and Pcomb.

1. In this research, DOE analysis was carried out using

the integrated process that minimizes repetitive
airbag modeling steps due to the change in driver
airbag configuration design factors. Furthermore,
The integrated process can be extended to
applications of optimization as well as parametric
studies regarding a driver airbag shape design
factors.

2. From the first DOE analysis result regarding driver

airbag shape design factors, the most effective
factor in reducing occupant’s injuries was found to
(a) HIC be the vent hole radius. From DOE analysis result
including additional factors, steering column axial
stiffness and inflator, which is not related to the
airbag shape design, were effective design factors.
This result shows that design factors that are not
related to driver airbag shape also need to be
considered to improve passive crash safety
performance.

3. Design factors’ effects from Pareto and Main Effect

charts in this research could be different from
package layouts of vehicles, crash body pulses and
other crash conditions. However, the integration
process would be useful to find out effective design
factors to improve passive crash safety performance
by adapting simulation models that reflect required
crash conditions.
(b) Chest g

REFERENCES

1. Yih-Charng Deng and J. T. Wang, “Computer

Experiments for Optimal FMVSS 208 and NCAP
Per-formance”, SAE Paper 942230.
2. Hideoki Yajima, Yasuhiro Dokko, Shigeo Ito, Keiichi
Motoyama, QiangYu and Masaki Shiratori, “The
Application of the Statistical Desing Support System
Toward Optimization of Vehicle Safety Equipment”,
SAE paper 1999-01-3209.
3. Honglu Zhang, Madana M. Gopal, Roopesh Saxena
and Xavier J.Avula, “An Integrated Optimi-zation
System for Airbag Design and Modeling by Finite
Element Analysis”, SAE paper 2003-01-0506.
(c) Pcomb 4. Kwon-Hee Lee, Byung-Hoon Lee, Ju-Young Lee,
Won-Sik Joo, “Robust Design of a Driver-Side
Figure 11. Main Effects (B) Airbag Considering Uncertainties”, KSAE Spring
Conference Proceedings, pp. 798~804, 2003.
5. iSIGHT User’s Guide ver.7.0, Engineous Soft-ware,
Inc., 2002.
CONCLUSION 6. MADYMO User’s Manual ver.5.4.1, TNO, 1999.
7. MADYMO Utility Manual ver.5.4.1, TNO, 1999.
In this research, series of processes, including from 8. MADYMO User’s Manual ver.6.1, TNO, 2003.
modeling the driver airbag to performing DOE analysis,
were integrated to automate parametric studies of driver
airbags. Also, by using integrated processes, out of all
the occupant injury criteria, the dominant design factors