Current injury criteria for the leg in axial loading fail to consider the effect of duration on force at fracture, which has been shown to be significant for body regions such as the femur and spine.^197,208 Consequently, application of current injury risk functions based solely on peak tibia force is limited to specific boundary and input conditions. The primary objective of this dissertation was to develop an injury criterion for use in both underbody blast (UBB) and automotive intrusion (AI) loading environments in order to encompass the loading characteristics anticipated for UBB scenarios where floor mats or other injury mitigation devices are employed. The secondary goal was to provide design criteria for expanding the range of loading durations (or frequencies) over which current anthropomorphic test device (ATD) legs provide a biofidelic force response.

Injury and response data was collected from 54 post-mortem human surrogate (PMHS) tests and from previous axial impact studies from the literature, with footplate force durations ranging from 5.1 to 89.9 ms. This test data was also used to generate force and leg compression corridors which were utilized for benchmarking a human leg finite element (FE) model. A parametric study was performed using the human leg FE model to improve understanding of the trends in fracture force observed in the PMHS experimental data and to characterize the effect of duration and pulse shape on force at fracture.

The leg injury risk function was developed using PMHS data and human leg finite element models to establish a relationship between fracture force and duration. An eventual optimization approach was used to estimate the survival model parameters for an injury risk function utilizing an injury predictor which combined peak force and impulse at peak force measured at the plantar surface of the foot. The injury risk function demonstrated a high level of injury prediction accuracy (81.8%) for load durations ranging from 5.1 to 89.9 ms, and improved injury prediction accuracy compared to previous injury risk functions which used peak tibia force as the injury predictor by up to 18%.

To improve the utility of the injury risk function, an ATD that is biofidelic for larger range of input frequencies was needed. Recommendations for ATD leg improvements were based on the design of the Mil-Lx. Mil-Lx experimental testing was performed to provide data to benchmark an FE model of the Mil-Lx, which was then use to establish the range of frequencies for which the Mil-Lx provides biofidelic force responses at both the upper tibia and the footplate. Optimizations were performed using the finite element model and a lumped mass model (LMM) of the Mil-Lx to estimate the characteristics of the compliant materials necessary to expand the range of loading frequencies for which Mil-Lx forces match forces in the human leg. Further, several additional LMMs were developed based on the human FE model response to characterize the mass and compliance distribution which should be used in future ATD design.

Recommendations for short-term, cost-effective Mil-Lx design improvements include replacing the heel pad and tibia compliant element with 70-durometer neoprene and 90- durometer polyurethane, respectively, adding angular rate sensors on the foot and tibia, and adding an ankle rotational potentiometer. These modifications would enable the Mil-Lx to be used with the proposed force and impulse injury risk function by yielding biofidelic plantar forces for impact acceleration durations ranging from 4.5 to 100 ms. Long-term recommendations for ATD leg development are more expensive and time-consuming, and involve designing an ATD to match the characteristics of a 3-mass LMM of the human leg and including a foot or heel load cell;​ in theory, this design could expand this range to 5 to 156 Hz.

In summary, the duration-dependent leg injury risk function described in this dissertation increases the injury prediction accuracy by up to 18% compared to previous injury risk functions, and expands the range of load durations for which a single leg injury risk function may be used, encompassing load durations of 5.1 to 89.9 ms. Modifications to the Mil-Lx design produced improvements in the range of loading frequencies for which the Mil-Lx provided biofidelic forces, but further improvement to the design is necessary to encompass the range of load durations applicable to both the underbody blast and automotive environments and to encompass the range of durations included in the injury risk function. Nonetheless, Mil-Lx modifications outlined in the dissertation provide a temporary, cost-effective solution for providing an ATD which can utilize the proposed injury criterion for a subset of the desired frequency range. These contributions are the first step toward establishing more accurate assessment of the effectiveness of injury mitigation schemes for underbody blast, and have the potential to inform the safety design process for both automotive and military vehicles.