Data-driven optimal design of foam-padded vests for female protection against air blast

Abstract

Blast-induced thoracic injuries pose a significant threat to military personnel, particularly during training and combat operations involving high-explosive weapons. While conventional body armor is primarily designed for ballistic protection, it is often inadequate for mitigating internal injuries caused by air blast waves. This study presents a data-driven approach for the optimal design of foam-padded protective vests tailored for female users exposed to air blast loading. A finite element (FE) model of the VIVA+ 50th percentile female torso was integrated with a multilayer foam vest model and subjected to simulated blast pressures ranging from 140 kPa to 1.4 MPa. The densities of three foam layers (inner, middle, outer) and the explosive charge mass were varied parametrically, and a surrogate model was developed to map these inputs to a biomechanical injury metric- peak chest wall velocity. The surrogate, based on second-order nonlinear regression, was validated against simulation data (R² = 0.98) and used in conjunction with a multi-objective genetic algorithm (NSGA-II) to identify optimal foam configurations. Results revealed that high-density inner and middle layers minimize local deformation, while the outer layer should have lower density to dissipate energy efficiently. Energy absorption analysis confirmed a shift in layer contributions with increasing blast intensity. Additionally, a simplified regression tree model (M5P) was developed to partition the design space and enhance interpretability without compromising accuracy. This study offers new insight into blast-specific armor design and demonstrates the value of surrogate-based optimization in protective equipment engineering.