Hybrid Experimental–Computational Study on the Energy Absorption of Graphene Nanoplatelet-Reinforced Sandwich Structures

Background: Sandwich composites are widely used in aerospace, automotive, and marine applications because of their lightweight, stiffness, and strength, but they remain prone to out-of-plane impacts and barely visible impact damage (BVID). Graphene Nanoplatelets (GNPs) have shown strong potential to improve fracture toughness and impact resistance. However, hybrid experimental–computational studies applying such reinforcement are still lacking in developing contexts like Iraq, where practical and accessible solutions are essential. Aims: The study investigates the effect of GNP reinforcement on the low-velocity impact behavior and energy absorption of glass fiber/epoxy sandwich panels with PVC foam and balsa wood cores. Objectives include fabricating panels with varying GNP content, performing standardized impact and compression-after-impact tests, applying Digital Image Correlation (DIC) for strain analysis, and validating finite element simulations against experimental data. Methods: Specimens were prepared with E-glass/epoxy face sheets containing 0.0–1.0 wt% GNPs and two core types (PVC foam, balsa wood). Low-velocity impact tests (ASTM D7136) and CAI tests (ASTM D7137) were conducted. Damage areas were measured using ImageJ, while DIC provided strain distribution maps. A finite element model using Abaqus/Explicit with Hashin’s criteria and a Cohesive Zone Model (CZM) simulated failure modes. Model accuracy was validated by comparing force–time histories and energy absorption with experimental results. Results: The addition of 0.5 wt% GNPs yielded optimal performance, improving absorbed energy by up to 28% and peak force by 24% compared to unreinforced panels. Balsa wood cores outperformed PVC foam in resistance and energy absorption. DIC analysis confirmed that GNPs reduced crack propagation and delamination through better stress distribution. The finite element model closely matched experimental data, with errors below 5%, confirming its reliability as a predictive design tool.