Dynamic and sensitivity analysis of blast-resistant aluminum foam sandwich panel

This paper presents a novel design of a blast-resistant aluminum foam sandwich panel for protective structures. The design incorporates multi-layered foam panels with varying densities to withstand a 1000 kg TNT explosion at a stand-off distance of 10 m. Nonlinear dynamic analyses were conducted to simulate the door response under free-air and surface blast scenarios, examining key parameters such as peak and permanent displacements, stress distribution, strain development, and energy absorption capacity. The layered-gradient design achieved up to 55% reduction in peak displacement and 67% reduction in permanent deformation compared to conventional concrete-filled and single-layer foam panels. The proposed design satisfies the UFC (Unified Facilities Criteria) operability criteria, with an edge rotation of 1.51° compared to 4.50° for conventional designs. Variance-based global sensitivity analysis using Sobol indices identified the front plate thickness and the third foam-layer density as dominant parameters, with sensitivity indices of 0.34 and 0.33, respectively. A multi-objective optimization framework utilizing Genetic Algorithms (GA) and the Non-dominated Sorting Genetic Algorithm-II (NSGA-II) was employed with Latin Hypercube Sampling, generating 80 design configurations and surrogate models achieving R2 = 0.97. The optimization yielded Pareto-optimal designs achieving a 46.1% reduction in peak displacement for maximum protection configurations and a 24.5% weight reduction.