ABSTRACT This study presents a deep‐learning framework to predict the ballistic performance and damage mechanisms of aramid‐fiber fabrics applied in spacecraft protective structures subjected to hypervelocity impacts (HVI) from space debris. A mesoscale finite element‐smoothed particle hydrodynamics (FE‐SPH) coupling algorithm is used to construct a numerical model at the yarn/fiber bundle level. A high‐fidelity dataset including five parameters: initial velocity, incidence angle, projectile diameter, number of plies, and pre‐tension is then generated. On this basis, a convolutional neural network (CNN) integrated with an adaptive attention mechanism is proposed to achieve high‐precision reconstruction of multi‐view damage images of front/warp/weft damage images of the aramid fabric. Meanwhile, a multi‐task collaborative multi‐layer perceptron (MLP) model is established to quantitatively predict the residual velocity of projectiles along with energy absorbed in fabrics. The results show that the CNN model can accurately capture the local details of the projectile‐hole contour and the global features of the fabric arrangement. The MLP model has high prediction accuracy for the residual velocity and the absorbed energy. Through SHAP interpretability analysis, the contribution weights of the five input parameters are further quantified. The deep‐learning surrogate model developed in this study substantially reduces the computational cost of time‐consuming numerical simulations and enables real‐time evaluation and optimal design of protective structures operating in extreme environments.

This study presents a representative volume element (RVE) based modeling approach to analyze the bending response and mesoscale damage evolution of 3D five-directional (3D5D) braided composites. These composites, recognized for their exceptional delamination resistance and impact toughness, are widely utilized in critical aerospace applications. In the proposed approach, the actual bending specimen is idealized as a homogeneous plate composed of periodically arranged RVEs, each subjected to a uniform bending moment. A shell-based periodic boundary condition (PBC) is developed, enabling simulation of the entire specimen using a single RVE with high efficiency and accuracy. A continuum damage model is integrated within the framework to capture the internal damage mechanisms and mesoscale fracture patterns. Experimental validation conducted across various braiding angles demonstrates strong agreement between the simulated and measured curvature-moment response, as well as matching full-strain fields with DIC data. The proposed model effectively captures both macroscopic bending response and the evolution of internal damage, providing valuable insights into the structural performance of braided composites under bending loads.

The increasing use of high-energy laser systems has raised concerns about the ablation resistance of aerospace carbon fiber-reinforced polymer (CFRP) structures. Predicting the response of such materials is challenging because laser irradiation triggers a strongly coupled interaction between heat transfer, thermomechanical deformation, and matrix pyrolysis. In this work, a multi-physics finite element (FE) framework is developed to investigate the ablation behavior of laminated CFRP composites subjected to continuous laser exposure. The model incorporates anisotropic heat conduction, thermally induced stresses, and temperature-dependent pyrolysis rate, enabling the evolution of temperature fields, stress distributions, and interlaminar damage to be examined under different laser power densities and fiber orientations. The simulations indicate that laser power density plays a dominant role in controlling the ablation process, whereas the influence of ply orientation is comparatively limited within the considered parameter range. To accelerate parametric evaluation, surrogate models based on Random Forest (RF) and Multi-Layer Perceptron (MLP) algorithms are constructed using datasets generated from Latin hypercube sampling of the simulations. Once trained, these models reproduce the numerical predictions with high accuracy while reducing the computational time. The proposed framework offers an efficient approach for rapid evaluation of laser-induced damage in CFRP structures and assists in the design of laser-resistant composite components.

3D angle-interlock woven fabrics (3DAWFs) have been extensively applied in aerospace and other industries because of their outstanding mechanical characteristics; thus, the investigation of their damage mechanism under impact loading has always received much attention. However, the high costs associated with experiments and the limited computational efficiency of finite element (FE) analysis have hindered the effective characterization of damage behaviors across varying parameter configurations. To address these challenges, this study presents two deep learning models. The first is a convolutional neural network (CNN) featuring an encoder-decoder architecture enhanced by a center bias mechanism, which consists of a spatial attention module with a Gaussian enhancement mask. This model, utilizing a dataset derived from FE simulations, accurately predicts damage images of 3DAWFs, as well as deformation images that illustrate the energy absorption mechanism through the input of projectile impact angle, initial velocity, and warp/weft direction, with its mean squared error (MSE) ranging from 6.73 x 10-3 to 9.86 x 10-3. The second model is a multi-layer perceptron (MLP) that quantitatively predicts the projectile residual velocity based on the aforementioned input features, achieving a coefficient of determination (R 2) of 0.9904 and a mean absolute percentage error (MAPE) of 3.72%, and explains the influence of each feature via SHapley Additive exPlanations (SHAP) interpretability analysis. Together, the CNN and MLP models successfully predict damage morphology and mechanical properties, enabling real-time monitoring of the impact damage behavior of 3DAWFs and offering a novel approach for a comprehensive study on the dynamic damage behavior.

High-altitude environments significantly influence the performance of flying-wing finned tube heat exchangers (FWFHEs); however, the specific variation patterns and the underlying mechanisms are poorly understood. To address this gap, this study systematically investigates the fluid flow and heat transfer characteristics, as well as the relevant mechanisms, of flying-wing finned tubes (FWFTs) under high-altitude conditions. The performance of the FWFHE was experimentally evaluated under simulated low-temperature ambient conditions corresponding to an altitude of 1000 m. The established simulation methodology was then validated using these experimental results. Following validation, numerical studies were conducted to examine the fluid flow and heat transfer characteristics of FWFTs over an altitude range of 0-5500 m, with inter-fin inlet air velocities from 6 to 12 m/s and Reynolds numbers (Re) between 1250 and 3000. The results indicate that, when the inter-fin inlet air velocity is held constant, the pressure drop, heat transfer rate, and convective heat transfer coefficient all decrease significantly with increasing altitude. Specifically, each parameter is reduced by approximately 5.5%, 5%, and 5% for every 500-m increase in altitude, respectively. Despite these dimensional changes, the primary dimensionless parameters-friction factor, corrected Nusselt number, and performance evaluation criterion-remain largely insensitive to altitude under constant Re conditions, indicating stable dimensionless fluid flow and heat transfer characteristics. This stability results from a compensatory mechanism: the increased airflow velocity needed to maintain constant Re counteracts the rise in kinematic viscosity at higher altitudes. To facilitate engineering calculations, altitude correction factors for pressure drop and heat transfer coefficient are proposed. These factors enable accurate prediction and optimization of the heat transfer area design for FWFHEs operating in high-altitude environments.

This paper presents a shell-based multiscale model for predicting the mechanical properties of through-thickness woven composites (TTWC) under in-plane and out-of-plane loadings, with capturing coupled damage evolution at the meso- and macroscale. A reduced shell scheme is employed for the macroscale, while full-thickness RVEs of TTWC are established at the mesoscale to account for inhomogeneous response along the thickness direction. The macro- and mesoscale simulation are coupled and executed simultaneously within a nonlinear concurrent framework. This approach enables the nonlinear macro-shell behavior, characterized by generalized strain and stress resultants, to be directly derived from the realistic mesoscale damage response. By incorporating scanned mesoscale geometries and damage constitutive relations, the shell-based multiscale model is applied to investigate damage and failure mechanisms in notched tensile and simply-supported bending plate samples, respectively. The predicted results are analyzed and compared with existing experimental data and direct numerical simulations (DNS), validating the efficacy and accuracy of the proposed model.

Crushing behavior analysis and energy absorption optimization are crucial for lightweight structures in automotive applications. The present paper aims to investigate the crushing behavior of thin-walled aluminum/CFRP hybrid tubes under axial loading using an explicit finite element (FE) simulation. The damage constitutive models of aluminum and CFRP are implemented by coding the user-defined subroutine VUMAT in ABAQUS/Explicit, which includes the damage initiation and evolution laws and element deletion scheme. Parametric studies are conducted to assess the effects of radius and aluminum layer thickness on the crushing performance of hybrid tubes. Additionally, a multi-objective optimization is performed on the Isight platform using a non-dominant sorting genetic algorithm (NSGA-II) and technique for order preference by similarity to ideal solution (TOPSIS) with entropy weight method. The optimization aims to maximize crashworthiness and increase energy absorption capacity, enabling designers to select an optimum size ratio.

A hierarchical coupled multiscale method is developed to capture the onset and propagation of damage within the three-dimensional four-directional (3D4D) braided composites subjected to bending loading. Using a directly two-scale coupled scheme, the macroscopic nonlinear behavior at the structural dangerous region could be iteratively solved, combined with the progressive damage response of the realistic mesoscale architecture. The continuum damage model is merely defined at the mesoscale and considers failure in each of constituents with the well-established Hashin failure criteria and the Bazant crack band damage model. To accelerate finite element computation in two scales, a parallel numerical implementation is presented alongside the commercial software ABAQUS/Standard. Besides the good agreement between the experimental and the predicted values, the results also show a significant effect of the braiding angle on the bending performance of 3D4D braided composites. With the increase of braiding angle, the bending stiffness and strength of 3D4D braided composites decreases, but the fracture toughness increases. This phenomenon was numerically investigated by identifying the different fundamental failure mechanism and damage development process at both the scales.

3D braided composites consolidated by heat-resistant resin are expected to become essential structural material of new generation aerospace vehicles. In this work, the thermo-mechanical behavior of 3D braided composites subjected to general off-axis tensile loadings at room and elevated temperatures is evaluated based on the meso-scale finite element (FE) model. A user-material subroutine UMAT is adopted to implement the thermal stress analysis, temperature-dependent failure initiation criteria and material properties degradation scheme of both fiber yarns and matrix. The meso-scale damage evolutions under thermal-mechanical coupling loads are simulated based on ABAQUS/Standard and the corresponding detailed failure mechanisms are revealed. The off-axial elastic moduli and strengths of 3D braided composites under elevated temperatures are also predicted. The numerical results indicate that high temperature affects the material properties mainly by weakening the matrix properties while off-axis loading affects the damage mechanisms by changing the load distribution of fiber yarn in each direction. The present work provides routine support for the numerical study of structural properties and damage behavior of other textile composites in service temperature environment.

3D orthogonal woven composites (3DOWCs) are increasingly utilized in various engineering applications due to their superior mechanical properties. However, their mechanical properties and damage mechanisms could be significantly influenced by the environmental factors such as temperature and moisture. This study aims to investigate the mechanical behavior of 3DOWCs under hygrothermal conditions through meso-scale finite element (FE) simulations. The proposed FE model is validated using existing experimental results under standard temperature and humidity conditions, and subsequently applied to predict the performance in hygrothermal environments. The effects of temperature, humidity, and loading conditions on the mechanical properties of 3DOWCs are analyzed and the corresponding failure mechanisms are revealed. This study offers significant insights into the hygrothermal behavior of 3DOWCs, thereby facilitating their optimal design and implementation in various engineering applications.

This paper presents a hierarchical coupled multiscale model for evaluating the progressive damage behavior of notched 3D woven composites subjected to the tensile loading. In the present work, the constitute behavior of structural dangerous area is iteratively solved, in conjunction with the finite element analysis (FEA) realized on realistic mesoscale architecture. This two-scale method is capable of simulating the propagation process of macro-meso coupled damage nearby the hole-edge using mesoscale continuum damage model without applying any macroscopic phenomenological failure criterion. Besides the good agreement between the experimental and the predicted stress–strain responses, the experimental phenomenon of the damage initiation location and branching pattern nearby the hole-edge are successfully tracked. The effect of apertures on the notched strength and failure mechanisms is numerically investigated. The present work provides routine support for the numerical study of structural damage behavior of composite materials with complex macro and meso geometries.

A parallel multiscale numerical framework based on the FE 2 method is established to perform the non-linear failure analysis of three-dimension (3D) composite materials with complicated architectures. Parallelization of the nested solution process is presented, as well as the strategy used to implement the FE 2 method alongside the finite element package ABAQUS. The emphasis is put on the detailed numerical implementation in both scales, using a set of python scripts and UMAT subroutines to allow the service of commercial finite element packages ABAQUS instead of the in-house code. The novel framework can be scalable to the distributed clusters and carries out the progressive damage analysis of 3D braided composite materials under uniaxial tensile loading and three-point bending loading, respectively. As a validation, the results obtained have been compared with experimental data from the literature. The parallel performances of numerical examples in runtime and acceleration effect are addressed to demonstrate the capabilities of the proposed framework.

The variational asymptotic homogenization (VAM) theory is extended to access freely to commercial finite element (FE) software to deal with periodic plate structures. In this work, the finite element format for periodic plate structures based on the variational asymptotic homogenization is developed, ensuring the commercial finite element software can be utilized to obtain the effective plate stiffness. A standard numerical framework and an integration algorithm are proposed for unifying the dimensional reduction analysis and the homogenization analysis in a formalized manner. As for model validation, the periodic plates composed of unit cells with three-dimension (3D) heterogeneous geometry are simulated by various elements and modeling techniques using the commercial FE software rather than programming in-house code. Compared to the results provided in the existing literature, the proposed approach shows excellent performance in terms of computational efficiency and time without compromising the VAM accuracy. It is preferable to enhance the application of the variational asymptotic homogenization theory for the more sophisticated heterogeneous plate structures.