David Poulard

Development of a computational framework to adjust the pre-impact spine posture of a whole-body model based on cadaver tests data

A method was developed to adjust the posture of a human numerical model to match the pre-impact posture of a human subject. The method involves pulling cables to prescribe the position and orientation of the head, spine and pelvis during a simulation. Six postured models matching the pre-impact posture measured on subjects tested in previous studies were created from a human numerical model. Posture scalars were measured on pre- and after applying the method to evaluate its efficiency. The lateral leaning angle θL defined between T1 and the pelvis in the coronal plane was found to be significantly improved after application with an average difference of 0.1±0.1° with the PMHS (4.6±2.7° before application). This method will be applied in further studies to analyze independently the contribution of pre-impact posture on impact response using human numerical models.

Understanding how pre-impact posture can affect injury outcome in side impact sled tests using a new tool for visualization of cadaver kinematics

The effect of posture and subject-specific factors on injury outcome is an active field of research in injury biomechanics, in particular in automotive safety research where post-mortem human subjects (PMHS) are used as surrogates. Current PMHS tests routinely include acquisition of the subjects׳ geometry and kinematics. However, combining these two datasets to better understand the injury mechanism is still a challenge. This study investigated the connection between pre-impact posture and resulting injuries in six previously published side impact sled tests (three with a rigid wall and three with an airbag) by creating three-dimensional kinematic animations (3DKA) of the tests. The 3DKA allow qualitative assessment of parameters related to posture and their possible effect on injury outcome. The orientation of the struck scapula and the lateral leaning of the torso were identified as potentially significant parameters. The ranges of variation in these parameters were quantified and compared to the number of rib fractures for each subject: the data suggested a correlation, but there was insufficient data for a probabilistic analysis. The 3DKA were published with this study and are freely available.

Evaluation of biofidelity of THUMS pedestrian model under a whole-body impact conditions with a generic sedan buck

ABSTRACT Objective: The goal of this study was to evaluate the biofidelity of the Total Human Model for Safety (THUMS; Ver. 4.01) pedestrian finite element models (PFEM) in a whole-body pedestrian impact condition using a well-characterized generic pedestrian buck model. Methods: The biofidelity of THUMS PFEM was evaluated with respect to data from 3 full-scale postmortem human subject (PMHS) pedestrian impact tests, in which a pedestrian buck laterally struck the subjects using a pedestrian buck at 40 km/h. The pedestrian model was scaled to match the anthropometry of the target subjects and then positioned to match the pre-impact postures of the target subjects based on the 3-dimensional motion tracking data obtained during the experiments. An objective rating method was employed to quantitatively evaluate the correlation between the responses of the models and the PMHS. Injuries in the models were predicted both probabilistically and deterministically using empirical injury risk functions and strain measures, respectively, and compared with those of the target PMHS. Results: In general, the model exhibited biofidelic kinematic responses (in the Y–Z plane) regarding trajectories (International Organization for Standardization [ISO] ratings: Y = 0.90 ± 0.11, Z = 0.89 ± 0.09), linear resultant velocities (ISO ratings: 0.83 ± 0.07), accelerations (ISO ratings: Y = 0.58 ± 0.11, Z = 0.52 ± 0.12), and angular velocities (ISO ratings: X = 0.48 ± 0.13) but exhibited stiffer leg responses and delayed head responses compared to those of the PMHS. This indicates potential biofidelity issues with the PFEM for regions below the knee and in the neck. The model also demonstrated comparable reaction forces at the buck front-end regions to those from the PMHS tests. The PFEM generally predicted the injuries that the PMHS sustained but overestimated injuries in the ankle and leg regions. Conclusions: Based on the data considered, the THUMS PFEM was considered to be biofidelic for this pedestrian impact condition and vehicle. Given the capability of the model to reproduce biomechanical responses, it shows potential as a valuable tool for developing novel pedestrian safety systems.

Evaluation of Geometrically Personalized THUMS Pedestrian Model Response against Sedan-Pedestrian PMHS Impact Test Data

ABSTRACT Objective: Evaluating the biofidelity of pedestrian finite element models (PFEM) using postmortem human subjects (PMHS) is a challenge because differences in anthropometry between PMHS and PFEM could limit a models capability to accurately capture cadaveric responses. Geometrical personalization via morphing can modify the PFEM geometry to match the specific PMHS anthropometry, which could alleviate this issue. In this study, the Total Human Model for Safety (THUMS) PFEM (Ver 4.01) was compared to the cadaveric response in vehicle–pedestrian impacts using geometrically personalized models. Methods: The AM50 THUMS PFEM was used as the baseline model, and 2 morphed PFEM were created to the anthropometric specifications of 2 obese PMHS used in a previous pedestrian impact study with a mid-size sedan. The same measurements as those obtained during the PMHS tests were calculated from the simulations (kinematics, accelerations, strains), and biofidelity metrics based on signals correlation (correlation and analysis, CORA) were established to compare the response of the models to the experiments. Injury outcomes were predicted deterministically (through strain-based threshold) and probabilistically (with injury risk functions) and compared with the injuries reported in the necropsy. Results: The baseline model could not accurately capture all aspects of the PMHS kinematics, strain, and injury risks, whereas the morphed models reproduced biofidelic response in terms of trajectory (CORA score = 0.927 ± 0.092), velocities (0.975 ± 0.027), accelerations (0.862 ± 0.072), and strains (0.707 ± 0.143). The personalized THUMS models also generally predicted injuries consistent with those identified during posttest autopsy. Conclusions: The study highlights the need to control for pedestrian anthropometry when validating pedestrian human body models against PMHS data. The information provided in the current study could be useful for improving model biofidelity for vehicle–pedestrian impact scenarios.

Experimental Investigation of the Effect of Occupant Characteristics on Contemporary Seat Belt Payout Behavior in Frontal Impacts

ABSTRACT Objective: The goal of this study was to investigate the influence of the occupant characteristics on seat belt force vs. payout behavior based on experiment data from different configurations in frontal impacts. Methods: The data set reviewed consists of 58 frontal sled tests using several anthropomorphic test devices (ATDs) and postmortem human subjects (PMHS), restrained by different belt systems (standard belt, SB; force-limiting belt, FLB) at 2 impact severities (48 and 29 km/h). The seat belt behavior was characterized in terms of the shoulder belt force vs. belt payout behavior. A univariate linear regression was used to assess the factor significance of the occupant body mass or stature on the peak tension force and gross belt payout. Results: With the SB, the seat belt behavior obtained by the ATDs exhibited similar force slopes regardless of the occupant size and impact severities, whereas those obtained by the PMHS were varied. Under the 48 km/h impact, the peak tension force and gross belt payout obtained by ATDs was highly correlated to the occupant stature (P =.03, P =.02) and body mass (P =.05, P =.04), though no statistical difference with the stature or body mass were noticed for the PMHS (peak force: P =.09, P =.42; gross payout: P =.40, P =.48). With the FLB under the 48 km/h impact, highly linear relationships were noticed between the occupant body mass and the peak tension force (R2 = 0.9782) and between the gross payout and stature (R2 = 0.9232) regardless of the occupant types. Conclusions: The analysis indicated that the PMHS characteristics showed a significant influence on the belt response, whereas the belt response obtained with the ATDs was more reproducible. The potential cause included the occupant anthropometry, body mass distribution, and relative motion among body segments specific to the population variance. This study provided a primary data source to understand the biomechanical interaction of the occupant with the restraint system. Further research is necessary to consider these effects in the computational studies and optimized design of the restraint system in a more realistic manner.

The contribution of pre-impact posture on restrained occupant finite element model response in frontal impact

Objective: The objective of this study was to discuss the influence of the pre-impact posture to the response of a finite element human body model (HBM) in frontal impacts. Methods: This study uses previously published cadaveric tests (PMHS), which measured six realistic pre-impact postures. Seven postured models were created from the THUMS occupant model (v4.0): one matching the standard UMTRI driving posture as it was the target posture in the experiments, and six matching the measured pre-impact postures. The same measurements as those obtained during the cadaveric tests were calculated from the simulations, and biofidelity metrics based on signals correlation (CORA) were established to compare the response of the seven models to the experiments. Results: The HBM responses showed good agreement with the PMHS responses for the reaction forces (CORA = 0.80 ± 0.05) and the kinematics of the lower part of the torso but only fair correlation was found with the head, the upper spine, rib strains (CORA= 0.50 ± 0.05) and chest deflections (CORA = 0.67 ± 0.08). All models sustained rib fractures, sternal fracture and clavicle fracture. The average number of rib fractures for all the models was 5.3 ± 1.0, lower than in the experiments (10.8 ± 9.0). Variation in pre-impact posture greatly altered the time histories of the reaction forces, deflections and the rib strains, mainly in terms of time delay, but no definite improvement in HBM response or injury prediction was observed. By modifying only the posture of the HBM, the variability in the impact response was found to be equivalent to that observed in the experiments. The postured HBM sustained from 4 to 8 rib fractures, confirming that the pre-impact posture influenced the injury outcome predicted by the simulation. Conclusions: This study tries to answer an important question: what is the effect of occupant posture on kinematics and kinetics. Significant differences in kinematics observed between HBM and PMHS suggesting more coupling between the pelvis and the spine for the models which makes the model response very sensitive to any variation in the spine posture. Consequently, the findings observed for the HBM cannot be extended to PMHS. Besides, pre-impact posture should be carefully quantified during experiments and the evaluation of HBM should take into account the variation in the predicted impact response due to the variation in the model posture.

A Cortical Thickness Mapping Method for the Coxal Bone Using Morphing

As human body finite element models become more integrated with the design of safety countermeasures and regulations, novel models need to be developed that reflect the variation in the populations anthropometry. However, these new models may be missing information which will need to be translated from existing models. During the development of a 5th percentile female occupant model (F05), cortical thickness information of the coxal bone was unavailable due to resolution limits in the computed tomography (CT) scans. In this study, a method for transferring cortical thickness information from a source to a target model with entirely different geometry and architecture is presented. The source and target models were the Global Human Body Models Consortium (GHBMC) 50th percentile male (M50) and F05 coxal bones, respectively. To project the coxal bone cortical thickness from the M50 to the F05, the M50 model was first morphed using a Kriging method with 132 optimized control points to the F05 anthropometry. This technique was found to be accurate with a mean nodal discrepancy of 1.27 mm between the F05 and morphed M50 (mM50) coxal bones. Cortical thickness at each F05 node was determined by taking the average cortical thickness of every mM50 node, non-linearly weighted by its distance to the F05 nodes. The non-linear weighting coefficient, β, had a large effect on the accuracy and smoothness of the projected cortical bone thickness. The optimal projection had β = 4 and was defined when the tradeoff between projection accuracy and smoothness was equal. Finally, a quasi-static pelvis compression was simulated to examine to effect of β. As β, increased from 0 to 4, the failure force decreased by ~100 N, whereas the failure displacement increased by 0.9 mm. Results from quasi-static compression tests of the F05 pelvis were comparable to experimental results. This method could be applied to other anatomical regions where cortical thickness variation is important, such as the femur and ribs and is not limited to GHBMC-family models. Furthermore, this process will aid the development of subject-specific finite element models where accurate cortical bone thickness measurements cannot be obtained.