Computational modeling

Research projects:

  • Development of vehicle occupant models for vulnerable populations
  • Spine and lower extremity injuries due to high-velocity impacts
  • Rodent head model development
  • Pig head model development
  • Human head model development
  • Numerical simulations of vehicular safety

Completed projects

  • Characterizations of material properties
    • Magnesium
    • Rat skull
  • Bovine pia-arachnoid complex
  • Spontaneous hip fracture
  • Reduction of aorta ruptures in automotive crashes

Research projects

Development of vehicle occupant models for vulnerable populations

Current automotive safety designs are based on crash test dummies of three different sizes: 5th percentile female, 50th percentile male, and 95th percentile male. The goal of this project is to close the gap between current safety testing and the actual injuries suffered by 10-year-old child and 75-year-old female that have been historically neglected. This research will ultimately reduce injuries to all occupants -- regardless of how old they are.

Very little attention has been paid to the safety of child in automotive crashes. The present day pediatric crash test dummies are really just scaled-down versions of adult dummies, but this doesn’t give an accurate prediction of a child’s response to an accident.  For instance, a child’s head is proportionally larger than his or her body, which plays a significant role in the unique injuries he or she can suffer compared with an adult. The same goes for senior women, whose bones are typically more brittle than a senior man’s in addition to impaired visual function, slowed reflexes, and reduced flexibility.

Experimental investigations and numerical simulations of spine and lower extremity injuries due to high-velocity impacts

Much of the experimental fall- and automotive crash-induced lower extremity and spinal fractures involve loading rates that are much lower than those observed in anti-vehicle landmine explosions. As such, numerous finite element models previously developed to simulate such injuries have been focused on much lower loading rates. Additionally, none of these experimental and numerical investigations were specifically designed and validated for vertical loading conditions present in landmine blasts. The goal of this project is to improve warfighter survivability through rate-appropriate injury biomechanics research. Several dummy and whole human body model will be exercised in two seating positions, with loads initiating under the seat, as in landmine explosions. The results from these simulations will be compared to experimental data to ensure biofidelity, and analysis of injury mechanisms can be identified by the computational model.

Rodent head model development

An anatomically detailed finite element rat head model, consisting of 255,700 elements with an average element size of 200 microns, has been developed and validated against locally measured peak brain displacement obtained from dynamic cortical deformation tests. The model has been utilized to correlate the mechanical response parameters to contusion volume, regional injury variability in the hippocampus following controlled cortical impact (CCI) injury, and regional patterns of neuronal loss under and away from the impact site.  A finite element mouse brain model was further developed using morphing technique, consisting of t 258,428 brick elements with a typical spatial resolution of 150 microns.  The mouse brain model was used to help analyze CCI experiments on mouse.  The coupled experimental and finite element study on CCI-induced brain injury has been selected as cover art on Journal of Neurotrauma November 2011 Issue.

Pig head model development

A pig head model, including both brain and skull, is under development to incorporate with current primary blast injury research. Porcine specimens have been a common surrogate in blast testing, allowing for the measurement of both biomechanical and physiological parameters during experiments. The pig is considered an intermediate step between the more commonly used rat surrogate and humans, who cannot be tested. This model will have sufficient mesh quality for computational simulation of both free-field and shock tube exposure, using Arbitrary Langrangian-Eulerian methodology

Human head model development

Wayne State University has a long history of using computational models to investigate the dynamic response of the brain and skull in impact situations beginning in the 1990s. The original version of the model has been iteratively improved with advanced computer technology and more complete experimental validation data. The model includes the scalp, a three-layered skull, cerebrospinal fluid (CSF), dura mater, falx cerebri, and brain with differentiated white and gray matter. Brain motion, in particular has been extensively validated against experimental data, and a sliding interface was added between the pia and arachnoid to increase biofidelity.
 
The most advanced version of the model was developed with a much finer mesh and greater numerical stability, and is referred to as the Wayne State University Head Injury Model (WSUHIM), has been subjected to extensive validation using published cadaveric test data including the intracranial and ventricular pressure data reported by Nahum et al. and Trosseille et al., the relative displacement data between the brain and the skull by King et al. and Hardy et al., and the facial impact data by Nyquist et al. and Allsop et al.

Both thresholds for injury and injury mechanisms have been examined using this model. Impact biomechanics research has included automotive crashes, sports injuries, etc. The model is also currently being modified for use in primary blast injury research.

Numerical simulations of vehicular safety

The U.S. Federal Motor Safety Standards requires all original equipment manufacturers to pass a significant number of vehicular crash tests before they can sell cars in the market.  Meeting these requirements through testing alone is time consuming and very costly.  We routinely work with OEMs and their suppliers on projects related to improving crash safety through computer simulations.  Pictures below show simulations of vehicles in full frontal rigid barrier crash, frontal offset deformable barrier test, 30 degrees oblique impact, lateral impact, pole impact, roof crush, IIHS side impact, and read-end impact. 

Completed research projects

Characterization of material properties

Magnesium

In order to model vehicles involved in automotive crashes, the structural components of these vehicles may need to be modeled in detail. Current technology allows for the development of new alloy materials, such as super vacuum die cast magnesium AM60B. These materials are often used in energy-absorbing structural components which play a role in automotive safety. As these components are designed to experience large deformations and even failure as a part of their energy-absorbing functions, it is important to model such behavior in computational simulations for accurate vehicle intrusion predictions.
A reverse engineering optimization methodology was employed to determine parametric values for five LS-DYNA material models based on experimental testing, including four-point bending, quasi-static axial crushing and dynamic axial crushing. Each material model is controlled by distinct constitutive equations, some of the parameters of which cannot be measured directly during physical experiments. The goodness of fit was quantitatively measured using the Gross Correlation Index, which considers peak load, energy absorption, and displacement, with higher weight given to peak load and energy absorption. Similar correlations were found for all material models after optimization, but MAT99 (SIMPLIFIED_JOHNSON_COOK_ORTHOTROPIC_DAMAGE) predicted the most accurate damage patterns consistent with all three experimental test modes.


Rat skull

A reverse-engineering methodology was employed to determine the most accurate material property parameters of the rat skull by minimizing differences in three-point bending test responses in physical experiments and simulations. In general, material properties from bending tests use theoretical equations to calculate both elastic and plastic properties, but these equations involve many assumptions, including regular geometry and material homogeneity. This new computational approach uses sample-specific voxel-based meshes, allowing for the determination of material properties without the confounding variables of anatomical inconsistency in sample dimensions and internal porosity. The results of this study were used to improve the rat head finite element model developed at Wayne State University, simulating the biomechanics of the brain-skull interaction with a higher degree of biofidelity.

Bovine pia-arachnoid complex

Traumatic brain injury (TBI) has become a major public health and socioeconomic problem that affects 1.5 million Americans annually. Finite element methods have been widely used to investigate TBI mechanisms. The pia-arachnoid complex (PAC) covering the brain plays an important role in the mechanical response of the brain during impact or inertial loading. Existing finite element brain models have tended to oversimplify the response of the PAC due to a lack of accurately defined material properties of this structure, possibly resulting in a loss of accuracy in the model predictions. The objectives of this study were to experimentally determine the material properties of the PAC along its anatomical axes at different strain rates and to determine the material constants of constitutive equations derived for PAC from experimental data.

Bovine PAC was selected for experimental samples in this study. Three loading modes (in-plane tension, normal traction, and tangential shear) were conducted to investigate the biomechanical response of PAC at different strain rates ranging from 0.05 to 100 s-1. The strain-rate effects, as well as directional and regional dependencies, of PAC were analyzed. Based on observation of PAC material response, a set of constitutive equations was proposed to model the transversely isotropic, nonlinear viscoelastic characteristics of PAC. A curve fitting-based optimization algorithm was carried out to determine the material constants needed for the constitutive equations.

Results from this study provide essential information to properly model the PAC membrane, an important component in the skull/brain interface, in a computational model of the human head. Such an improved representation of the in vivo skull/brain interface will enhance future studies investigating brain injury mechanisms under various loading conditions.

(Taken from the dissertation of Dr. Xin Jin)

Spontaneous hip fracture

A laboratory experiment was designed to test the hypothesis of spontaneous hip fracture after medical literature reported hip fractures without a history of trauma. Although hip fracture as a result of falls has been well-documented, especially in the elderly, it has also been postulated that only five percent of the energy available during a fall is required to fracture bone. Therefore, from a biomechanical standpoint, it would be expected that more falls would result in fracture. As that is not the case, there may be another phenomenon in play. Some researchers have suggested that hip fracture is the cause, and not the effect, of falls. Abnormal motion, resulting in bending moments across the femoral neck due to musculoskeletal forces, may be the cause of such fractures, and this study examined hip fracture forces and patterns due to contraction of the iliopsoas and gluteus medius muscles.

The femoral head was fixed in an acetabular cup apparatus, allowing free rotation, and loaded at either the lesser or greater trochanter. Fractures occurred at an average of approximately 3,000 N and 57 J. For the lesser tronchanteric loads, fractures appeared in the cervical region, and for the greater trochanteric loads, the sub- or inter-trochanteric regions were affected. The findings of this study suggest that the rotator muscles are an important factor in the biomechanics of hip fracture. The average fracture force was 450% of body weight, well within the ranges producible by muscle contraction and only slightly higher than normal physiologic loads. With the combination of osteoporotic bone, hip fracture remains a valid concern in the medical community.


Reduction of aorta ruptures in automotive crashes

Traumatic rupture of the aorta (TRA) remains the second most common cause of death associated with motor vehicles crashes after brain injury. On average, nearly 8,000 people die annually in the United States due to blunt injury to the aorta. It is observed that more than 80% of occupants who suffer an aortic injury die at the scene due to exsanguinations into the chest. With the advent of more accurate and established human body finite element models, FE crash constructions have become a valuable tool when assessing crash scenarios and occupant injury mechanisms.

This study included near side left lateral real world finite element reconstructions, sensitivity study, thresholds for TRA, aorta mechanics in racing crashes and conceptual countermeasures. The crash reconstructions showed high strain in the isthmus region of the aorta, in the region of aortic failure. A parametric study was performed to identify key parameters in injury causation, showing sensitivity to impact direction and velocity. Further simulations indicated that acceleration alone is not sufficient to cause TRA. Thoracic deformation is essential and the shoulder-clavicle complex plays a crucial role in TRA by causing relative motion of the sternum and thoracic spine.

Through computational simulations, it was shown that the protective effects of an improved side-construction standard can be enhanced further by the more rational placement of an airbag-like structure, to buffer head and chest contact with the B-pillar along its entire length from car seat to roof. As a result of this modification, a reduction in mortality and injury related to lateral motor vehicle crashes could be expected.

(Modified from the dissertation of Dr. Aditya Belwadi)

Contacts

King-Hay Yang
aa0007@wayne.edu
313-577-0252

Liying Zhang
lzhang@wayne.edu
313-993-9434