This thesis presents a computational, three-dimensional biodynamic model of a vehicle-occupant, crash-victim during rollover accident. The model represents the human body by a series of connected bodies simulating the limbs, torso, head and neck of the human frame. The computer algorithms are developed using DYNACOMBS – a three-dimensional multibody dynamics computer simulation model. DYNACOMBS provides a dynamic analysis of arbitrary collections of bodies allowing for both translation and rotation between adjacent bodies. The relative orientation of adjacent bodies is defined by Euler parameters to avoid computational problems of singularities, which occur with other orientation angles (such as Bryan angles or Euler angles). Thus there are distinct computational advantages of the presented model over other crash-victim simulators.
The governing dynamical equations for the model are based upon Kane’s equations and their associated kinematical quantities (partial velocities, partial angular velocities, generalized speeds, and lower body arrays). Formulations based upon Kane’s equations are believed to be the most efficient and most reliable of the various methods available for studying large multibody systems. Consequently they are ideally suited for studying the dynamics of crash victims.
The developed vehicle and victim model allows for arbitrary vehicle motion. Specifically, given the vehicle rollover motion, as would occur in an accident environment, the simulation predicts the vehicle occupant movement and the forces exerted on the occupant by the seat and seat belts. The model is validated by a series of recently recorded sets of experimental data from dummies and cadavers in a variety of vehicle crash conditions.