The Contribution of Vehicle, Occupant and Crash Factors to the Risk of Injury as a Result of Vehicle Rollover - New approaches to data and modeling analysis

Doctoral thesis, 2017

Injuries and deaths as a result of vehicle rollover remain a consistently large contributor of the overall

crash fatalities in the United States. While more new vehicles are becoming equipped with electronic

stability control and rollover ejection countermeasures as well as increased roof strength, it will take

several years of production before these newer vehicles permeate the fleet and the effectiveness of these

technologies can be fully assessed. In the interim, continued vigilance on assessment of rollover injury

causation is recommended. This can be done through systematic analysis of aggregate field crash data and

specific case studies. Also, mathematical modeling studies can be done to assess the contributions of

vehicle, occupant and crash factors to the injury risk of rollover involved occupants.

The major aims of the research in this thesis are to determine how rollover crash investigations and

crash field data analysis can determine the most frequent types of injuries and their mechanisms that occur

to belted, unejected occupants involved in rollover crashes and, once determined, identify the role of

vehicle, occupant and crash factors that can predict injury risk. The first aim can be achieved through case

studies and aggregate national crash data analysis while the second aim uses finite element and multi-body

modeling of rollover crashes.

Aggregate rollover field data was taken from the National Automotive Sampling System –

Crashworthiness Data System (NASS-CDS). Head, spine (cervical) and thoracic injuries dominated the

injury with specific injury types in each body region indicating areas of further interest to investigate

regarding injury causation. Analysis of specific case studies taken from the Crash Injury Research

Engineering Network (CIREN) indicated that single event, single vehicle (pure) rollovers were associated

with complex mechanisms of cervical spine injuries that were associated with vehicle roof strength

(strength to weight ratio), and the amount of vertical as well as lateral intrusion at the injured occupant

location. Occupant body mass index was a possible contributor to injury risk.

A finite element model of a contemporary sedan was used in a simulation of a Controlled Rollover

Impact System (CRIS) test to identify vehicle and crash parameters that were most associated with high

cervical neck forces in the Hybrid III dummy occupant model. The variables that contributed the most to

the occupant and vehicle structural response were pitch angle, roll angle, and drop height. These factors

determine where and with what force the vehicle roof impacts the ground. The analysis showed that proper

selection of a crash dummy model is also a critical step in the interpretation of effects of the factors used

in analysis. Subsequent MADYMO modeling of the CRIS test with the models of the Hybrid III and

THOR advanced frontal crash dummy and a facet model of the human body were performed with

imported finite element nodal vehicle model outputs representing vehicles with the strongest and weakest

roof. When coupled with a parameter analysis of advanced vehicle seat belt restraints, the analysis showed

that stronger roofs will reduce injury risk, and that restraint systems can provide additional protection to

reduce the potential for occupant head impact to the roof.

The analysis approach to both data and modeling in this thesis provided results through innovative

combined crash field data analysis, parametric computer modeling methods and statistical and human

body modeling techniques to arrive at the conclusions reached. As future vehicle design evolves with

respect to automation and other propulsion systems, designers and engineers need to be aware of the

iv

structural and occupant restraint requirements these vehicles will need as they interact with the fleet and

are exposed to potential rollover situations.