Large-Scale Parametric Evaluation of Child Booster Seats

Introduction: Motor vehicle crashes (MVCs) are the second-leading cause of death for US children between the ages of one and 14 years old. Child restraint systems have been shown to reduce injury and mortality risk for children in this age group. Specifically, booster seats reduce the risk of serious injury by 45% for children between the ages of four and eight. However, the link between booster seat characteristics and the booster’s safety performance is not well understood. This problem is exacerbated by the wide range of booster seat designs available on the market, and physical testing of booster seats is limited by anthropometric test devices with poor biofidelity. Finite element (FE) simulations with human body models (HBM) represent a solution to improve the understanding of booster seat performance.
Goals of Study: The goal of this thesis was to examine how booster seat parameters affect occupant response and risk of submarining. In order to achieve this goal, the following research questions were addressed: (1) Does the occupant’s posture affect the likelihood of submarining? (2) Which characteristics of booster seats influence submarining and overall performance? (3) What are the robust regions of booster design space that are capable of providing safe performance regardless of occupant posture and external belt anchor geometry?
Methods of Study: A large-scale, parametric study of 714 individual booster seats was executed in order to examine the link between booster seat parameters and occupant response. A study of this size required development of an automated simulation cycle, as well as evaluating the performance of the PIPER human body model, developing an FE model of the new FMVSS 213 bench, and characterizing key parameters in the booster design space from a sample of 44 physical booster seats. The automation cycle was made up of the following steps: parameter sampling, booster FE model generation, human body model positioning, running the settling simulation, running the seatbelt routing simulation, and running the final simulation. The results from these simulations were then used to develop a neural network metamodel, capable of describing the effects of booster design across the entire design space rather than just the individually simulated data points.
Results of Study: The results of the simulations allowed the three research questions to be addressed, as well as additional lessons about the relationship between booster seats and occupant response. A ranked list of the booster and environmental parameters that had the largest impact on submarining risk and a combined metric of multiple safety outcomes was generated using clustering methods. Booster stiffness, occupant posture, total cushion length, and the lateral location of the shoulder belt guide were some of the most influential parameters. Finally, the following characteristics had the greatest effect on predicted booster performance: high stiffness, high back, high cushion depth, long total cushion length, inboard shoulder belt guide position, and low lap belt guide position.
Impact of Thesis: The insights gained from this thesis directly address the field’s limited understanding as to how booster seat characteristics affect occupant response. In addition, the automation framework developed in this thesis will allow for future large-scale, parametric studies regarding booster seats, and may also be adapted to other applications. Finally, this study introduced metamodels to the field of pediatric injury biomechanics, which has traditionally lacked the necessary study size to utilize these powerful analysis tools.