Ceramic Body Armor Single Impact Force Identification on a Compliant Torso Using Acceleration Response Mapping

Abstract

This research proposes and experimentally implements a new method to identify the location and magnitude of a single impulsive excitation to body armor. The technique could easily be extended to other components that undergo rigid body dynamics. Impact loads are identified in two steps. First, the location of the impact force is determined from time domain acceleration responses by comparing them to attributes of either reference acceleration time histories or the responses generated by a proposed Newtonian analytical model. Then based on that location, appropriate reference frequency response functions are used to reconstruct the input force in the frequency domain through a least squares inverse problem. Both the analytical model and experiments incorporate the mechanical properties of a deformable dummy’s torso. Experimental results demonstrate the validity of this method both at low energy excitations, which are produced by a medium modally tuned impact hammer, and at high energy excitations, which are produced by dropping 0.2-0.6 kg rods from a height of 2 m. The maximum error in the estimated location or magnitude for the low energy excitations on the 10 cm square ceramic body armor was 7.07 mm with an average error of 1.09 mm. For the high energy excitations, which produced accelerations at the measurement locations up to 50 times greater than that of the low energy excitations, the maximum error in the predicted location of the input force was 11.18 mm with an average error of 6.64 mm. There was no force transducer to capture the input force on the body armor from the rod but, from non-energy-dissipative projectile motion equations, the validity of the solutions was confirmed by comparing the impulses. The frequency response functions (FRFs) used to reconstruct the input force from the rods relied on frequencies that were poorly excited by the metal-tipped modal impact hammer used to generate the FRFs. Despite this limitation, responses from impact points near the center of the tile, which were predominantly comprised of frequencies in a regime where the FRFs were adequately excited, were successfully used to reconstruct the input forces with impulses close to theoretical predictions. Optimal sensor locations for predicting the impact location are also determined with the use of a genetic algorithm that utilizes a fitness function derived from the proposed physics-based analytical model.