Reliable detection of barely visible impact damage is critical to ensure the structural integrity of composite components in service, particularly in safety-critical applications such as pressure vessels and transportation systems. This study presents a solution for detecting such damage in woven glass fiber-reinforced thermoplastic composites using terahertz (THz) time-of-flight tomography and convolutional neural networks. THz provides non-contact, non-ionizing, high-axial-resolution imaging of subsurface and back-surface damage, addressing key limitations of surface-based inspection methods. While THz imaging alone may not always permit conclusive damage identification, we bridge this gap by training neural network classifiers on depth-resolved THz B-scan images using ground truth from co-located X-ray micro-computed tomography. Among several pretrained architectures tested via transfer learning, DenseNet-121 exhibits the highest accuracy. The model remains robust even when trained on truncated B-scans excluding surface indentation features, confirming its ability to detect structural anomalies located internally or on the back surface. This is particularly relevant for applications where back-side access is not feasible. Experimental validation is performed on impacted glass-fiber-reinforced thermoplastic coupons prepared in accordance with ASTM D7136, with damage severity quantified through force-displacement data and micro-tomographic analysis. Labeling for supervised learning conforms to acceptance criteria from industrial standards for composite pressure vessels (ASME BPVC Section X, CGA C-6.2), ensuring regulatory alignment and enabling deployment in quality control workflows. The proposed method minimizes the need for expert interpretation or secondary validation and offers direct applicability to in-service inspection and manufacturing quality control.

This study examines alloys in the pseudobinary system. The objective is to gain a deeper understanding of the thermodynamic limits of the ferromagnetic Al-Mn phase within the broader context of high-entropy alloys (HEAs) and CompositionallyComplex Alloys (CCAs). Although this phase shows promise as a rare-earth-free permanent magnet, its industrial application remains challenging due to its intrinsic instability. The present work investigates nine distinct compositions located along a line in a four-dimensional compositional space, combining an alloy with in varying proportions. The alloys have been produced using a cold-crucible levitation furnace, heat-treated, and studied in their as-cast condition. Characterization methods included X-ray diffraction (XRD) and magnetic propertymeasurements. The experimental findings highlight a complex interplay between composition, phase stability, and magnetic response, demonstrating that conventional HEA design strategies are insufficient for this alloy system. The study concludes that a more robust, thermodynamically informed approach is required to efficiently explore the design space of HEAs and CCAs based on metastable phases. Furthermore, the magnetic properties evolve significantly with composition. This evolution is not reflected in the XRD patterns, indicating that it arises not only from crystal structure changes but also from factors such as local atomic ordering.

In the aerospace, nuclear, and machining industries, alloys require high hardness and excellent wear rate. Most high‐performance alloys rely on cobalt for high‐temperature properties, despite its political, ethical, and health concerns. High‐entropy alloys (HEAs), enabled by structural hardening, lattice distortion, and sluggish diffusion, offer pathways to eliminate this critical element. This study examines cobalt substitution in HEAs to optimize hardness and wear rate. New alloys based on the Cantor system (CoCrFeMnNi) are produced by individually replacing cobalt with copper, aluminum, vanadium, or molybdenum. Four equiatomic HEAs (AlCrFeMnNi, CrFeMnNiV, CrCuFeMnNi, and CrFeMnMoNi) are compared with two literature alloys (Al 0.2 Co 1.5 CrFeNi 1.5 Ti and CoCrFeMnNi) and with the pure substituent elements, all evaluated in the same metallurgical state. All synthesized HEAs except CoCrFeMnNi are multiphased and do not mimic the structure of their corresponding pure element; CrCuFeMnNi also departs from valence electron concentration predictions. Pure cobalt shows the lowest wear rate, while the Cantor alloy exhibits a higher one. Aluminum, vanadium, and molybdenum strengthen HEAs despite limited performance in their pure state. Ultimately, pure cobalt, CrFeMnMoNi, and AlCrFeMnNi display similar and superior wear rate compared with the optimized reference alloy Al 0.2 Co 1.5 CrFeNi 1.5 Ti.

The aim of this article is to present a consistent higher-order homogenization theoretical framework, fully compatible with the laws of thermodynamics, for analyzing composites under coupled thermomechanical conditions. The constituents of these composites exhibit nonlinear dissipative behavior and fall within the class of generalized standard materials. Free energy potentials are formulated at both the microscale and macroscale levels, enabling the derivation of constitutive relations between stress, entropy, strain, and temperature, while ensuring that dissipation remains strictly non-negative. For the first time, the strong theoretical foundations of the general framework are outlined, with particular emphasis on the special case of second-order nonlinear homogenization.

A micromechanics-informed neural network framework is developed for homogenization of periodic unidirectional thermoconductive composites with cylindrically orthotropic fibers. The framework hard-imposes the steady-state governing heat conduction equations within the network architecture, enabling accurate capture of singular heat flux fields at the fiber center that are challenging for conventional numerical approaches. In contrast, continuity and periodicity conditions are enforced via boundary collocation points in the loss function. Validation against finite element simulations across a wide range of fiber volume fractions shows that accurate and converged temperature distributions can be achieved after 9000 training epochs using 8-16 harmonic terms. Additional higher-order harmonics are difficult to train reliably and may degrade predictions. While strong agreement is observed in the matrix heat flux distributions, noticeable discrepancies persist in the fiber phase due to varying ability to capture the singular heat flux fields. Furthermore, uniform collocation points converge faster than random points during solution refinement. Finally, transfer learning is employed to accelerate training for new configurations, allowing the network to achieve comparable accuracy after only 2000 training epochs, which is substantially fewer than the 9000 epochs required when training from scratch.

We propose a thermodynamics-informed multi-head attention neural network (TMANN) framework for predicting elastoplastic behavior under arbitrary loading paths. In contrast to earlier thermodynamics-informed networks that rely solely on internal state variables to encode loading history, the TMANN incorporates a multi-head attention mechanism that explicitly captures the material history sequence, thereby enhancing predictive accuracy and stability. The architecture comprises an attention network for predicting increments of internal state variables and a companion neural network for estimating the Helmholtz free energy at each time step. To ensure physical consistency and strengthen the model’s generalization capability, the loss function explicitly enforces thermodynamic constraints, including non-negative free energy, non-negative dissipation rate, and monotonic accumulation of effective plastic strain. Furthermore, a rolling iterative prediction strategy is implemented to ensure the model’s compatibility with the stepwise nature of arbitrary loading paths, as only the initial stress and strain states are known a priori. The integration of TMANN into ABAQUS through a user material subroutine verifies its practical applicability to structural simulations. The effectiveness of the proposed TMANN is validated through comparisons with classical numerical methods at both the material point level and in structural simulations. New results showcase the TMANN’s robust generalization performance, maintaining high prediction accuracy under incremental loading/unloading and complex random loading scenarios.

Glass fiber reinforced composites exhibit highly nonlinear and path-dependent thermomechanical responses under coupled loading. Accurately predicting this behavior using traditional multiscale frameworks such as FE remains computationally prohibitive, particularly for complex geometries involving constituent-level interactions. To overcome these limitations, this study investigates the Thermomechanical Elasto-Plastic Artificial Neural Network (ThEP-ANN), a novel data-driven multiscale surrogate designed to accelerate fully coupled thermomechanical simulations of such materials. The ThEP-ANN framework circumvents the complexity of explicit constitutive modeling by directly predicting macroscopic quantities (e.g., stresses and energy rates), rather than relying on free-energy potentials. This strategy significantly reduces the computational cost associated with high-dimensional automatic differentiation. Crucially, the surrogate explicitly incorporates the evolution of internal variables, enabling an accurate representation of history-dependent and path-integral material responses. The ThEP-ANN is implemented non-intrusively into the commercial finite element code Abaqus via a Meta-UMAT subroutine, allowing seamless integration into macroscopic simulations. The framework is assessed through a three-stage, fully numerical validation strategy, comprising generalization to unseen loading paths, mesoscopic validation at the Representative Volume Element (RVE) level corresponding to a single macroscopic Gauss point, and macroscopic structural-level simulations on a finite element model. The results demonstrate excellent agreement with high-fidelity FE solutions while achieving orders-of-magnitude reductions in computational cost. This work establishes a practical pathway for efficient large-scale numerical analyses of nonlinear thermomechanical composite structures.

Interest in High Entropy Alloy metallurgical development has grown and become a major topic in the field of innovative high-temperature metallurgy. CrFeNiNbTa is a new class of refractory alloy that conserves its hardness properties at high temperature, competing with alloy grades such as nickel-based superalloys. Inconel 718 is well known for its limited machinability due to its precipitation-hardened and multiphase microstructure and high strength. Also, a disruptive approach is to improve machinability through High Entropy (HEA) and Medium Entropy Alloy (MEA) strategy development. Dry ball nose end milling of HEA Cr33Fe26Ni23Nb9.5Ta9.5 and an alternative MEA Cr35Fe25Ni30Nb5Ta5 is performed. Results were compared with those of an Inconel 718. Cutting force was measured to analyze the effect of cutting speed and feed rate on the process phenomenon. Tool flank wear evolution and rake face wear mechanisms are clarified. The change in surface roughness, surface morphology, and the subsurface modification was comprehensively investigated and discussed. Results show that machinability is improved with HEA and MEA. Cutting forces evolution during machining of Inconel 718, MEA, and HEA present similarities. However, HEA and MEA reduce the ploughing effect at the trailing edge. Tool wear is limited when machining HEA, reducing the Build-up Edge phenomenon. Surface roughness is improved with MEA and HEA, and evolution is limited despite the tool wear. The subsurface of the HEA exhibits phase deformation and cracks in the Ta-rich phase, which can lead to partial tearing and the formation of cavities. MEA strategy preserves subsurface integrity thanks to the ductile phases.

Composite overwrapped pressure vessels (COPVs) are central to high-pressure hydrogen storage, where low mass and structural integrity under burst conditions are simultaneously critical. Optimising their stacking sequences is difficult because the governing failure mechanism changes through the wall thickness, and because established failure criteria diverge appreciably under the multiaxial stress states produced by internal pressure. Designing to the most critical predicted mode is standard engineering practice; what is not established is how to embed that practice inside an automated optimisation loop, and what it delivers when it is. This study integrates the Tsai–Wu interactive quadratic, Hashin mode-separated fibre/matrix, and Puck inter-fibre failure (IFF) criteria into a triadic envelope the point-wise maximum failure index over a discretised radial profile and places that envelope directly in the fitness function of a genetic algorithm (GA), so that the governing criterion is resolved radially and per layer rather than assumed in advance. Ply stresses are obtained from a Lekhnitskii-type thick-walled multilayered anisotropic cylinder solution under generalized plane strain with closed-end axial equilibrium. Applied to a benchmark linerless (Type V) cylindrical COPV of 174 mm internal radius and 175 MPa burst pressure, the triadic formulation yields a 36-layer configuration of 54.24 mm wall thickness, with a maximum combined failure index of 1.12 and a layer exceedance rate of 5.6%. Evaluated under identical design constraints, this halves the cross-criterion exceedance rate relative to the Hashin-only (1.18, 11.1%) and Puck-only (1.07, 11.1%) designs, at a wall thickness 5.6% greater than the least conservative Tsai–Wu-only design (0.97, 0%). The balance of ply orientations across the low-, mid- and high-angle ranges is an explicit constraint of the fitness function rather than an emergent property of the envelope; holding it fixed across all four runs isolates the effect of the failure measure itself. The analytical solver is verified against the closed-form Lamé solution in the isotropic limit and against the boundary, interface and axial-equilibrium conditions of the multilayer problem.

This study investigates fuzzy fiber composites, characterized by a viscoplastic matrix and fuzzy fibers, i.e. fibers coated with radially aligned carbon nanotubes (CNTs). A comprehensive micromechanical framework is developed to model and optimize these composites, with a particular emphasis on interfacial damage mechanisms introduced through microvoids growth in the region between the fuzzy fibers and the matrix. By developing an equivalent fiber model, the complexity of the multi-phase structure is effectively reduced, facilitating efficient parametric analyses. Various homogenization techniques, including Composite Cylinder Assemblage (CCA), Transformation Field Analysis (TFA), and periodic homogenization, are combined to predict the overall stress-strain responses of the equivalent fiber approach and then the full fuzzy fiber composite. The identification of the framework and model parameters enabled a parametric/sensitivity analysis to study the effect of varying key parameters, including the volume fraction. The results of this paper contribute to a deeper understanding of unidirectional fuzzy fiber composites and establish a foundation for future parametric investigations and fuzzy fiber composite applications accounting for nonlinear regimes.

In this short review, the multiscale modeling of dissipative composites undergoing fully coupled thermomechanical processes is outlined through the models presented in a collection of recent works. The aim is to demonstrate the challenges and limitations of: (1) the multiscale approaches (full-field or mean-field techniques), (2) the computational approaches dealing with complex material systems, (3) the alternative methodologies dedicated to the analysis of composite structures, such as those founded on the data-driven modeling and the model order reduction techniques.

Medical implants are a common treatment for orthopedic injuries. Their apparent stiffness can be reduced by using architected internal lattices to match the gradient stiffness of the bone, thereby avoiding postoperative biomechanical problems such as stress shielding. The use of TPMS-based lattice structures with smooth junctions offers the potential to tailor the apparent modulus of an implant while minimizing stress concentration throughout the microstructure. In this study, four TPMS-based unit cells are investigated, namely: Schoen's Gyroid-like (sheet and skeletal), Schwartz's Primitive, and Schoen's IWP topologies. The objective of the investigation is to numerically replace a small region of a femoral bone, hereafter referred to as the area of interest (AoI). Multiscale approach is proposed for the 3D model of the femur. The latter consists of the global model (femur bone) and the local model (TPMS unit cell). The unit cells are selected to satisfy the elastic and mechanical loading requirements and are compared according to the von Mises stress distribution after applying periodic boundary conditions. A statistical analysis is performed and a function factor is proposed to facilitate the comparison. The developed methodology allows the design of customized and patient-specific implants when a large medical database is used due to the varying size and shape of patients' bones.

A new physics-informed deep homogenization neural network (DHN) framework is proposed to identify the homogenized and local behaviors in periodic heterogeneous microstructures. To achieve this, the displacement field is decomposed into averaged and fluctuating contributions, with the local unit cell solution obtained via neural networks subject to periodic boundary conditions. The periodic microstructures are divided into sub-domains representing the fiber and matrix phases, respectively. A key contribution of the proposed method is the marriage of elasticity solution and physics-informed neural network to each phase of the composite, namely, the fiber phase as a mesh-free component whose fluctuating displacements are expanded using a discrete Fourier transform, and the matrix phase using material points with fluctuating displacements handled through fully connected neural network layers. The interfacial continuity conditions are enforced by minimizing the traction and displacement differences at separate material points along the interface. Transfer learning is exploited further to facilitate training new microstructures from pre-trained geometry. This hybrid formulation inherently satisfies stress equilibrium equations within the fiber, while efficiently handling the periodic boundary conditions of hexagonal and square unit cells via a series of trainable sinusoidal functions. The innovative use of distinct neural network architectures enables accurate and efficient predictions of displacement and stress when discontinuities are present in the solution fields across the interface. We validate the proposed DHN with the finite-element predictions for unidirectional composites comprised of elastic fiber significantly stiffer than the matrix, under various volume fractions and loading conditions.

Inconel 718 alloy is used for high-temperature industrial applications in its optimized multiphase metallurgical state. Nevertheless, the machining of Inconel 718 alloy becomes problematic and challenging. One alternative consists of developing a new material design strategy based on the metallurgy of high-entropy alloys (HEAs). These alloys have become a hotspot in the field of innovative high-temperature metallurgy toward the improvement of the alloy's manufacturability and thermomechanical properties. This study aims at designing, elaborating, and characterizing a new class of alloys with increased entropy, referred to as: "Inco-like." The mechanical responses of the alloys, in terms of hardness, have been analyzed using an indentation test at a wide range of temperatures. The dry machinability of the developed alloys has been performed and compared with that characterizing the Inconel 718 in terms of several machining features. Finally, the phases of the studied alloys have been analyzed using metallurgical investigations. The experimental findings and comparisons underscore the advantages of the high-entropy strategy in terms of tool wear reduction and cutting tool durability. The results demonstrate that the Inco-like HEA retains a significantly higher hardness of 291 Hv at 800 degrees C, compared to 160 Hv for Inconel 718 at the same temperature.

We present a novel elasticity-inspired data-driven Fourier homogenization network (FHN) theory for periodic heterogeneous microstructures with square or hexagonal arrays of cylindrical fibers. Towards this end, two custom-tailored networks are harnessed to construct microscopic displacement functions in each phase of composite materials, based on the exact Fourier series solutions of Navier's displacement differential equations. The fiber and matrix networks are seamlessly connected through a common loss function by enforcing the continuity conditions, in conjunction with periodicity boundary conditions, of both tractions and displacements. These conditions are evaluated on a set of weighted collocation points located on the fiber/matrix interface and the exterior faces of the unit cell, respectively. The partial derivatives of displacements are computed effortlessly through the automatic differentiation functionality. During the training of the FHN model, the total loss function is minimized with respect to the Fourier series parameters using gradient descent and concurrently maximized with respect to the adaptive weights using gradient ascent. The transfer learning technique is employed to speed up the training of new geometries by leveraging a pre-trained model. Comparison with finite-element/volumebased unit cell solutions under various loading scenarios showcases the computational capability of the proposed method. The utility of the proposed technique is further demonstrated by capturing the interfacial debonding in unidirectional composites via a cohesive interface model.

Fiber-reinforced thermoplastic composites are valued for their strength-to-weight ratio, cost-effectiveness, and recyclability, highlighting the need for efficient recycling technologies amid environmental concerns. This study addresses these challenges by examining the mechanical response of recycled glass fiber reinforced polyamide 6 composites and modeling their nonlinear, time-dependent behavior under complex loading conditions. Advanced nonlinear constitutive and multiscale models, initially developed for conventional fiber composites, are adapted to capture the stochastic response of recycled materials. These models integrate viscoelasticity, viscoplasticity and damage in the polymer matrix and account for anisotropic damage in the strands, addressing the heterogeneity introduced by the recycling process. A modified random sequential adsorption technique replicates the microstructures for nonlinear response modeling. Hypotheses based on microstructural investigations consider processing effects that disrupt the initial chip woven structure and create matrix-rich areas. The model captures anisotropy and variability observed in experimental data, providing a reliable framework for predicting the performance of recycled thermoplastic composites and improving the understanding of the relationship between microstructure and mechanical properties, with a focus on inelastic nonlinear behavior.

This contribution presents a new physics-informed deep homogenization neural network model for identifying local displacement and stress fields, as well as homogenized moduli, of nanocomposites with periodic arrays of porosities under general loading conditions. Notably, it accounts for the surface elasticity effect, utilizing the Gurtin-Murdoch interface theory. First of all, a fully connected neural network model is established that maps the spatial coordinates, passing first through several sinusoidal functions, to the microscopic displacements. The loss function is formulated as the weighted sum of residuals of Navier-Cauchy equations in the bulk domains and the Young-Laplace equations on the energetic surfaces, evaluated on separate sets of collocation points. To more effectively predict stress concentrations inside the microstructures, we introduce fully trainable weights to each collocation point. The capacity and effectiveness of the new homogenization technique for capturing the size-dependent local and global response of nanocomposites with distinct pore sizes and shapes are verified upon extensive comparisons with the finite-element benchmark results, under various loading conditions. New results showcase the proposed theory's ability to model random distributions of nano-porosities with a high degree of accuracy, a task not easily achievable with alternative techniques except for the specialized finite-element method.