Ballistic performance of spaced multi-layer aluminum armor under fixed dimensional envelope constraints: FEM simulations with ANN-based design analysis

Multi-layered armor is conventionally assumed to outperform monolithic plates of equivalent mass. This study challenges this paradigm for dimensionally constrained systems by investigating AA2024-T3 aluminum configurations where a fixed 1.74 mm envelope is progressively subdivided into two to five layers, necessarily reducing total material mass from 1.64 mm to 1.04 mm as interface count increases. Using validated Abaqus/Explicit simulations with three double-nosed projectiles, we demonstrate that the two-layer configuration (L1) consistently delivers superior ballistic resistance, achieving limits of 45.7–61.4 m/s depending on projectile geometry. Performance systematically degrades with increasing layer count: the five-layer configuration (L4) exhibits 32% lower ballistic limits despite possessing four interfaces. A critical thickness threshold of 0.34–0.51 mm was identified, below which layers fail through low-energy shear rather than energy-intensive plugging. An artificial neural network trained on 249 simulations ( R 2 = 0.98, 0.92, 0.90 for residual velocity, absorbed energy, and impact time) independently confirmed through feature importance analysis that total material mass and individual layer thickness dominate performance over interface count. For thickness-constrained applications, maintaining fewer, thicker layers proves more effective than maximizing interfaces—a finding with direct implications for lightweight armor design in aerospace and defense sectors.