This work describes, through experimental and numerical investigations, the mechanical behavior and energy absorption characteristics of 3D-printed sandwich panels with cellular cores subjected to low-velocity impact. Using fused deposition modeling techniques (FDM), three different sandwich panels, one with a regular hexagonal core and two with re-entrant cores at 0 and 90 degrees, were fabricated. The sandwich panels were subjected to low-velocity impact, at impact energies of 10 J and 15 J. A comprehensive investigation of the panels’ behavior through experimental testing and numerical simulation was conducted. The results indicate that the sandwich panel with a 90 degrees re-entrant core is stiffer and absorbs the largest amount of impact energy but, at the same time, suffers significant damage to the upper facesheet. The 0 degrees re-entrant core is compliant and provides both impact resistance and good energy absorption characteristics. Such a sandwich panel finds its application in the construction of personal protective equipment, where the aim is to minimize the forces transmitted during low-velocity impacts and maximize the total absorbed energy. Re-entrant core sandwich panels prove to be very good candidates for replacing the honeycomb core sandwich, depending on the desired engineering application.

Sandwich structures with triply periodic minimal surface (TPMS) cores have garnered research attention due to their potential to address challenges in lightweight solutions, high-strength designs, and energy absorption capabilities. This study focuses on performing finite element analyses (FEAs) on eight novel TPMS cores and one stochastic topology. It presents a method of analysis obtained through implicit modeling in Ansys simulations and examines whether the results obtained differ from a conventional method that uses a non-uniform rational B-spline (NURBS) approach. The study further presents a sensitivity analysis and a qualitative analysis of the meshes and four material models are evaluated to find the best candidate for polymeric parts created by additive manufacturing (AM) using a stereolithography (SLA) method. The FEA results from static and explicit simulations are compared with experimental data and while discrepancies are identified in some of the specimens, the failure mechanism of the proposed topologies can generally be estimated without the need for an empirical investigation. Results suggest that implicit modeling, while more computationally expensive, is as accurate as traditional methods. Additionally, insights into numerical simulations and optimal input parameters are provided to effectively validate structural designs for sandwich-type engineering applications.

This study explores the mechanical performance of triply periodic minimal surface (TPMS) and stochastic lattice structures subjected to low-velocity impact. Two structurally promising geometries—one TPMS-based and one stochastic—were tested and compared with the well-established gyroid. Specimens were fabricated using stereolithography (SLA) and subjected to impact energies of 30 J and 40 J to assess the structural response and energy absorption capabilities. Experimental results show that the proposed TPMS structure exhibits higher impact forces compared with the gyroid, which are associated with significant impactor displacement and deep indentation. These samples demonstrated extensive damage, with cracking propagating through the entire core at higher energies, highlighting their susceptibility to structural failure despite their high initial strength. On the contrary, the stochastic structures allowed localized deformation in the impacted region, thus successfully avoiding catastrophic failure. The impact force efficiency was higher for both gyroid and stochastic geometries, with values ranging between 0.6 and 0.7, indicating effective energy absorption with reduced internal stress gradients. Furthermore, the evaluation of damping performance showed that most structures displayed high damping, as minimal energy was transferred back to the impactor. This work highlights the feasibility and functional versatility of TPMS and stochastic geometries for use in impact mitigation, vibration control, and related engineering applications.

Sandwich panels are widely used where high specific stiffness and impact tolerance are required. However, published comparisons of core systems are often confounded by differences in base material, relative density, specimen size, interface condition, manufacturing route, and test protocol. This exploratory study compares three complete sandwich panel core systems under common nominal low-velocity impact conditions: an integrated FDM-printed PLA hexagonal honeycomb, an integrated FDM-printed PLA re-entrant lattice, and an extruded polystyrene (XPS) foam core adhesively bonded between printed PLA face sheets. All panels had identical nominal external dimensions (108 × 100 × 30 mm) and the same nominal PLA face-sheet geometry. The two cellular configurations were produced by fused deposition modelling (FDM) using identical reported printing parameters, whereas the XPS core was adhesively bonded between the printed PLA face sheets. Instrumented drop-weight tests were performed on an Instron CEAST 9450 drop tower, following the principles of ASTM D3763 with modifications for the sandwich panel geometry and available instrumentation, at nominal impact energies of 2.8 J and 11 J. Among the individual specimens tested, the honeycomb configuration produced the highest representative peak contact forces (2008 N at 2.8 J and 3387 N at 11 J). The re-entrant configuration produced the highest mass-normalized criterion-limited work at 11 J (0.0553 J/g), while the bonded XPS configuration produced the lowest and smoothest recorded force response. The work metric was evaluated only up to a representative force criterion and is not interpreted as complete irreversible energy absorption. The comparison provides a preliminary engineering reference for the tested panel systems, but it does not isolate core topology independently from core material, relative density, manufacturing route, or interface condition. Because the integration endpoint is specimen-specific, the work values are descriptive metrics of the individual records rather than directly comparable measures of total energy-absorption performance; repeatability, uncertainty, confidence intervals, and statistical significance are not established.

A convolutional neural network (CNN) was developed to predict the Poisson’s ratio of representative volume elements (RVEs) composed of a bi-material system with soft and hard phases. The CNN was trained on a dataset of binary microstructure configurations, learning to approximate the effective Poisson’s ratio based on spatial material distribution. Once trained, the network was integrated into a greedy optimization algorithm to identify microstructures with auxetic behavior. The algorithm iteratively modified material arrangements, leveraging the CNN’s rapid inference to explore and refine configurations efficiently. The results demonstrate the feasibility of using deep learning for microstructure evaluation and optimization, offering a computationally efficient alternative to traditional finite element simulations. This approach provides a promising tool for the design of advanced metamaterials with tailored mechanical properties.