Research
Metamaterials
Controlled effective wave propagation characteristics
Architectured materials can be coined to exhibit static properties that well differ from the ones typically encountered in common engineering materials. Certain unit-cell arrangements of high material anisotropy can be used to tune the wave propagation attributes of the effective material in a systematic manner.
Lightweight designs with tailored static properties
The inner material design allows for effective static properties properties beyond the bounds of common engineering materials. Lightweight, shear soft and shear stiff inner material architectures, along with high and low volumetric resistance designs are analyzed. Yield limits can be tuned and strain localizations can be controlled.
Architected Interpenetrating Phase Composites
Nature-inspired architected materials premise effective property combinations that are well beyond the limits of classical engineering materials. Interpenetrating phase composites allow for effective attributes beyond the performance limits of regular cellular materials, opening new frontiers in the design of advanced structural applications.
Advanced Metastructures
Micro-architected beam structures: tuning deflections through inner strain couplings
Beam structures made of metamaterial with inner normal to shear strain couplings can balance bending loads through tension. Tetrachiral unit-cell designs are representative lattice patterns with an inherent coupled normal to shear strain behavior. The structure deflects when axially loaded, with the force required to equilibrate the bending to be analytically computed, through closed-form parametric expressions.
Mechanically joined Aluminum and Steel Sheet Structures: Self-Piercing-Riveting Feasibility and strength
The mechanical performance of self-piercing rivet (SPR) joints connecting aluminum and high strength steel sheets is investigated, numerically and experimentally. The joint strength and failure mechanisms are characterized for different loading modes, including lap-shear, cross-tension, inclined cross tension and coach peeling. The underlying influential parameters are analyzed, associating structural performance with inner design.
Static flexural and dynamic modal response of architected beams
Architected materials and structures have garnered significant interest out of their potential to furnish mechanical performances beyond the bounds of customary designs. TPMS-based architected beam structures, allows for enhanced specific flexular moduli, along with a control over the dynamic response of the architected structure.
Machine Learning & Explainability
Neural network modeling of the stress-strain response of polymers
Photopolymerization is the governing chemical mechanism in two-photon lithography, a multi-step additive manufacturing process. The relationship among the process parameters, the degree of polymerization, and the nonlinear stress-strain response of polymer structures can be well captured by Neural Network models, allowing for a rigorous assessment of their constitutive performance.
ML Modeling, feature importance and Explainability analysis
The response of composite beams depends on a series of parameters which include the base material orthotropy, its gradation, as well as structural design parameters, such as the element΄s slenderness. Which parameter affects the most the effective performance? How do the individual features interact? Do the boundary conditions matter and which type of ML modeling best captures the effective performance?
ML Modeling, feature importance and Explainability analysis
Automated fracture identification requires the development of appropriate deep-learning models. The transfer learning potential of a wide range of well-known CNN architectures is evaluated, while low-cost CNN models are developed. CNN models capable of automatically identifying fracture in different base material testing experiments, including Uniaxial Tension and Shear testing are developed.
Helical Structures
Helical structures: Numerical modeling and simplified analytical expressions
The computational modeling of helical structures commonly comes along with considerable numerical analysis costs, for their geometry to be adequately described. This can result in simulation limitations. Low order, two-dimensional finite element models allow for the computation of the mechanical behavior of helical constructions under axial, torsional, as well as radial and thermal loads in a rigorous manner, exploiting the geometric symmetry of the helical arrangement. Full scale 3D and simplified 2D models are developed for the analysis of the effective structural performance under different loading conditions.
Strength, ductility and cyclic loading performance of helical, fiber-based structures
Helically-braided, load-bearing structures crafted from plant and animal-based natural materials, including goat hair, coconut, palm, manila (abaca) fibers, and palm leaves. The constituent fibers are subjected to chemical analysis through spectroscopy, while geometric and material density attributes are assessed. Their static and cyclic mechanical attributes are analyzed, identifying primal differences among the fiber types for each loading type case. Manila fiber-based structures yield the highest effective static properties, with peak stress values above 70 MP a and energy absorptions up to 30 J/mm3, followed by coconut-fiber structures with corresponding values of 33 MP a and 7.3 J/mm3. These values are approximately an order of magnitude higher that those recorded for animal-fiber-based, helically-braided structures. The associated cyclic loading characteristics differ significantly from the corresponding static properties. Moderate static strength coconut fiber structures exhibit substantially higher hysteresis loss values (above 0.04 J/MP a after sesveral loading cycles) in comparison to the stiffer, manila-based designs. Furthermore, the ratio of the hysteretic loss energy to the total energy differs, with low-strength palm leaf fiber structures to yield a comparable performance with the high static strength, manila-fiber designs. In all cases, the evolution of the hysteretic energy loss magnitude as a function of the loading cycle is characterized, deriving correlations among the structural composition and the recorded cyclic response performance. Moreover, Ashby-type classifications are conducted with respect to a wide range of natural materials, highlighting the distinctive energy absorption capacity of manila and coconut fiber designs, approaching 1000 kJ/kg, values that are several times higher than those recorded for metallic helically-braided structures.
