Fatigue life prediction is crucial for the safety of metallic materials and structures. Machine learning (ML) models have demonstrated strong predictive capabilities in this field, but their "black-box" nature limits their reliability, trustworthiness, and application in engineering practice. Explainable Artificial Intelligence (XAI) provides key techniques to open the ML "black box". This paper aims to systematically review the current applications of XAI in the field of metal fatigue life prediction. Addressing the issue that current research applies various interpretation methods without systematic categorization, this paper proposes classifying existing methods into two main categories: "post-hoc explanations" and "interpretable by-design". This paper outlines the key techniques of these two categories and their specific application examples in fatigue life prediction, discusses the limitations of model interpretation methods and current technical challenges, and prospects for opportunities and challenges in subsequent research.

Invertebrates like earthworms and slugs exhibit continuum body structures and undulatory propulsion, making them ideal references for designing worm-inspired soft robots. Developing dynamic models of such biomimetic systems is essential for guiding robot design. This study investigates axial motion in a worm-like system using a slender circular hyperelastic rod as the mechanical model. We integrate axial and radial coupling effects by incorporating the Yeoh hyperelastic strain energy, viscous friction, and internal driving forces. A sinusoidal squared strain wave drives the system to study how retrograde peristaltic waves affect its motion. The derived nonlinear continuum equations are analytically challenging. To address this, we propose a hybrid analytical-numerical method: First, geometric approximations link radial deformation to axial displacement, capturing nonlinear coupling effects. Second, peristaltic wave propagation replaces direct force calculations in simulations. Results show the model works for both small and large deformations. Through a numerical case, the correlation between average locomotion speed and waveform parameters in the axial-radial coupling model is investigated. Under constant strain energy coefficient, friction coefficient, and slenderness ratio, increasing the sinusoidal squared strain wave amplitude or decreasing the wave-width-to-body-length ratio enhances the worm-like system's average axial propulsion velocity.

Large-scale aerospace structures are mostly supported by truss. After being launched into orbit, the truss will unfold through deployable mechanisms to form the final structure. Therefore, the smooth deployment of the truss is great significance for ensuring the normal operation of the spacecraft. This paper studies the mobility problem of space flexible truss by using the multi-body dynamics theory. Firstly, the deployable dynamic model of the truss is given. Then, the contact dynamic model of joint in truss is established and the calculation method of joint friction force is presented. Subsequently, we propose a criterion for the mobility of the truss structure. Finally, numerical simulations are carried out. The simulation results show that smaller joint clearance, larger joint friction force, and smaller deploy driving force will reduce the deployable mobility of the truss structure.

Previous reports have described a new phenomenon of interstitial fluid (ISF) flow in the perivascular and adventitial clearances (PAC) within neurovascular bundles, characterized by a flood-like, waterfall-like, high-speed, continuous centripetal flow. This paper provides more conclusive evidence to confirm that PAC is the primary conduit for ISF flow, demonstrating the objectivity, universality, and fundamental characteristics of the heart-oriented microflow of ISF within this main channel. The specific content includes the following: (a) further elaboration of the concept of PAC; (b) revealing the intrinsic correlation between the diversity of PAC's topological structure and the diversity of the band-like flow patterns within PAC; (c) estimating the spatial location of PAC through flow imaging near nourishing small blood vessels; (d) confirming the objectivity of PAC through flow patterns around bifurcation points of branching blood vessels; (e) revealing the variability of the main PAC flow and the perturbability of the microflow through externally stimulated wave-like microflows; (f) showing the plasticity of the PAC main channel and the controllability of the microflow through flow imaging around the "charred islands" of connective tissue membranes. Once the main PAC channel is confirmed, it will pave the way for verifying ISF circulation imaging and lay the foundation for uncovering the driving mechanisms of ISF circulation.

Previous reports have described the main conduit of interstitial fluid (ISF) flow—the perivascular and adventitial clearances (PAC) around neurovascular bundles—demonstrating a waterfall-like and flood-like centripetal flow of ISF within PAC. This paper continues the previous work and addresses the question: Where does the ISF within the main conduit, PAC, originate? The answer is that it comes from the secondary conduit, the perivascular space (PVS) around branching blood vessels. We confirm the following: (1) At the bifurcation points of main and branching blood vessels, the structure of the "branching deflection point" and the "deflection/parallel" topology are formed. (2) A continuous, smooth, and transparent connective tissue membrane simultaneously covers the main and branching blood vessels, creating two types of interstitial spaces. The first type is the gap between the connective tissue membrane and the main blood vessel, referred to in previous work as the perivascular and adventitial clearances around neurovascular bundles (main conduit PAC). The second type is the gap between the connective tissue membrane and the branching blood vessels, known as the perivascular space (secondary conduit PVS). (3) The main conduit PAC and secondary conduit PVS form a connected topology. ISF flows along the secondary conduit PVS towards the branching deflection point, eventually converging into the main conduit PAC, forming aggregated flow. (4) The combination of "PAC+PVS" constitutes the interstitial spaces surrounding the vascular tree, where the ISF within the secondary conduit PVS flows into the main conduit PAC, merging with the existing ISF in PAC and moving collectively toward the heart, forming coordinated flow. Thus, this paper clarifies the flow dynamics of ISF within the vascular tree interstitial spaces (PAC+PVS) and lays the foundation for confirming ISF circulation.

This paper reports two key findings: (1) the directional flow patterns and fundamental kinematics of interstitial fluid (ISF) within the perivascular spaces surrounding the coronary arteries and veins, and (2) the directional flow patterns and fundamental kinematics of ISF in the cardiopulmonary interstitial space. Using fluorescent tracer technology, we observed the following novel phenomena: (a) In the cardiac vascular tree, ISF flows from the perivascular space surrounding branching vessels (Perivascular Space, secondary conduit PVS) toward the perivascular and adventitial clearances surrounding the main blood vessels (Perivascular and Adventitial Clearances, main conduit PAC), forming the aggregated flow; (b) Within the main blood vessel perivascular and adventitial clearances (main conduit PAC), ISF flows from the apex of the heart toward the base, forming the ISF circulation within the coronary vascular clearances; (c) Within both the main conduit PAC and secondary conduit PVS, the flow of ISF follows three fundamental laws: the existence law, the co-directional law, and the counter-directional law; (d) ISF flows from the heart to the lungs, forming the cardiopulmonary interstitial fluid circulation; (e) Based on the flow patterns of ISF in the peripheral, cardiac, and cardiopulmonary interstitial spaces, we completed the "puzzle" of systemic ISF flow, constructing the image of ISF circulation within living organisms; (f) The ISF circulation is coupled with the blood circulation, forming a coupled dual circulation model of internal and external circulations.

Variable-stiffness composite laminates with variable-angle layups expand the design space and enhance the mechanical properties such as buckling strength and stiffness. However, commonly used single-objective optimization methods often failed to achieve simultaneous enhancement of multiple mechanical properties. To address this issue, this paper proposed an active learning Kriging model optimization method based on the Mean Squared Error and Expected Improvement criterion for the multi-objective optimization of variable-stiffness laminates. Its effectiveness was validated through an optimization case study of variable-stiffness laminates with fiber paths represented by B-spline curves. Compared to the gradient descent-based optimization method, the Kriging model-based variable-stiffness optimization method achieved a significant improvement in computational efficiency. Finally, a multi-objective optimization framework integrating the Kriging model and a multi-objective particle swarm optimization (MOPSO) algorithm was proposed. The resulted Pareto solution set included a design with a buckling factor comparable to that of ±45° laminates but an 87 % reduction in compliance.

Based on the technique of physics-informed kernel function neural network, a neural network method is proposed to solve the forward/inverse bending problems of beams and plates. This model includes two independent neural networks with different physics-informed kernel functions, and can achieve adaptive collocation by the neural network itself, which enhances the computational adaptability. Since the activation function includes the physical information of the differential governing equation, it is not necessary to consider the collocation nodes in the considered domain, which may effectively decrease the computational cost. A Galerkin equivalent method is presented to solve the particular solution for the discontinuous source term, which can reduce the difficulty in the training process. Numerical examples are given to verify the effectiveness in solving the the forward and inverse bending problems of beams and plates. This neural network has simple structure and flexible solution form, which can be applied to solve the problems of mechanics and (partial) differential equation in other fields.

Mortise-tenon structure is the essence of Chinese traditional culture, which is widely used in ancient wooden buildings and furniture. However, the existing restoration schemes and research of ancient wood furniture mortise-tenon structure mostly rely on artisan experience of craftsmen, experiments or numerical simulations, lacking simple theoretical guidance for restoration. This paper fabricated T-shaped mortise-tenon specimens using 3D printing technology, and obtained the experimental curve of the pull-out force versus the pull-out distance of these specimens. By analyzing the changes in the mortise-tenon response curve and the stress characteristics of the tenon head, a two-stage analytical model for the mortise-tenon pull-out force is established. Based on this, the influence of the interference fit on the structural pull-out limit is studied using finite element method. The predicted results not only reproduce the corresponding pull-out experimental and finite element numerical results, but also reveal the decisive mechanism of the static-dynamic friction conversion at the contact interface in determining the ultimate bearing capacity of the mortise-tenon structure.

In response to increasing demands for enhanced safety and weight reduction in modern aircraft, the development of high-strength, lightweight materials are still highly demanded. This study focuses on an ultra-high-strength titanium alloy with a tensile strength of 1 300 MPa, intended for use in critical lug. It is important to determine the tensile strength of the titanium alloy lug for its use in aerospace components. However, the applicability of the tensile strength formulas from the "Aircraft Design Manual" to this material remains uncertain, as does the tensile failure mechanism. Therefore, in this paper we design multiple titanium alloy lug configurations and perform tensile tests to determine the corresponding failure loads. Through the test results, the tensile strength formulas in the "Aircraft Design Manual" for titanium alloy lugs are updated. Combined with elastoplastic finite element analysis and experimental data, the tensile failure mechanism of the titanium alloy lugs is revealed.

The construction of lunar bases is a core objective of the fourth phase of the lunar exploration program. Control of the thermal environment poses significant challenges due to extreme temperature fluctuations on the lunar surface. This study focuses on analysing temperature field response within lunar lava tubes under base energy supply conditions. A nonlinear heat conduction model for basalt with temperature-dependent properties is developed, and a two-dimensional numerical model is established using finite element software CODE_BRIGHT. The model incorporates periodic temperature boundary conditions on the lunar surface and heat sources from the tube walls. The spatiotemporal evolution of the temperature field is systematically analyzed, with a focus on the effects of regolith thickness, burial depth, cross-sectional shape, and heating power. Some conclusions are obtained: (1) The temperature-dependent nature of the conductive model significantly affects prediction accuracy, as a constant thermal conductivity model overestimates the conducted temperature by approximately 30 %; (2) Lunar rock environment parameters exhibit threshold effects: a regolith thickness greater than 1 m can fully shield surface temperature fluctuations. When the burial depth is greater than five times the diameter of the cavern, the deep-layer constant temperature characteristics will dominate the thermal evolution, with a thermal impact zone approximately three times the tube diameter; (3) An elliptical cross-section (with a long-axis ratio of 4∶1) reduces the temperature at the tube wall vertex by around 80 % compared to a circular cross-section through directional heat dissipation. These findings provide critical parameter thresholds and numerical theoretical support for lunar base site selection, energy configuration, and thermal control system design.

To simplify the geometric modeling of three-dimensional complex cross-sectional beams and reduce computational scale while ensuring the accuracy of three-dimensional structural mechanics analysis, this paper proposes a one-dimensional elastic straight beam model that accommodates arbitrary cross-sectional geometries. To achieve efficient dynamic and static analysis of complex cross-sectional beams, a local differential quadrature method (LDQM) is developed based on the global differential quadrature method. The proposed method is used to efficiently discretize the established one-dimensional beam. Extensive numerical examples validate the accuracy of the proposed model and method in seismic time-history response, free vibration characteristics, and cross-sectional stress analysis. Through comparative convergence analysis with 3D solid finite element models and 3D plate finite element models, the numerical convergence performance of the LDQM-based 1D complex cross-section beam was validated in terms of displacement, stress, and free vibration. Numerical results indicate that the proposed one-dimensional model, utilizing LDQM, achieves significantly faster convergence in structural dynamic and static response analysis compared with traditional three-dimensional solid finite element methods, achieving higher accuracy than the three-dimensional plate model with a smaller system degree of freedom size, while yielding results consistent with the three-dimensional solid model. This study provides a new approach for high-efficiency structural analysis.

The study focuses on the curing process and mechanical behavior of E44 epoxy resin in unconstrained structures. A Fiber Bragg Grating (FBG) sensor is used to monitor the temperature and strain of the resin in the thin-walled encapsulation structure in real-time. Based on curing kinetics, heat conduction, and viscoelastic models, the curing performance parameters and stress relaxation functions are derived, providing essential input data for the numerical simulation of epoxy resin curing behavior. The objective of the study is to develop a finite element simulation model for the epoxy resin encapsulation structure, taking into account residual stress and temperature effects, and to validate the model's accuracy by comparing it with experimentally measured temperature and strain data. The results show that the maximum temperature error does not exceed 10 %, and the maximum strain error is within 30 %, demonstrating the reliability of the analytical method and finite element model. This research provides testing and characterization methods, as well as theoretical support, for the cracking and interface debonding of epoxy resin under high and low-temperature cycling conditions.

Fractional-order dynamical systems with multiple delays exhibit rich dynamic behaviors. Investigating their conserved quantities can lead to a deeper understanding of the intrinsic properties of these behaviors. Based on the combined Riemann-Liouville fractional derivative model, this paper establishes the multi-delay fractional Pfaff-Birkhoff principle and derives the differential equations of motion (along with other formulations) for multi-delay fractional Birkhoffian systems. By applying the variational principle to the action of multi-delay fractional Birkhoffian systems, we derive the Noether theorem for such systems. Finally, illustrative examples are provided to demonstrate the application of the results.

Ensuring the safe and stable operation of floating platforms in marine environments necessitates a scientifically rigorous design of mooring systems, as well as accounting for the risk of mooring line failures. This study focuses on the OC4-DeepCwind (OC4) semisubmersible platform. By employing three-dimensional potential flow theory and radiation-diffraction theory, a multi-field coupled dynamic model of the six-mooring-line semisubmersible platform is developed. The platform's motion responses and the mechanical behavior of the mooring lines under various operational conditions are then analyzed. The study shows that when no mooring line breakage occurs, the platform remains safe under the monsoon case. However, under the typhoon case, both platform displacement and mooring line stress will increase, making the platforms unsafe. In addition, under the monsoon case, when a single mooring breaks, the displacement of the platform and the force on the mooring cable are both within a safe range. However, when mooring lines 1 and 5 break simultaneously, the platform sway will exceed the safe range. The research results can provide reference for the appropriate design of mooring systems.

This study investigates the dynamics of a single squirmer within a two-dimensional lid-driven and double-lid-driven square cavity flow using Smooth Particle Hydrodynamics (SPH) method. Through a comprehensive investigation of the effects of initial position, squirmer size, swimming parameter  β and particle Reynolds number (Res) on the squirmer motion, the study reveals the confined behavior of squirmers within steady-state cavity flows. In lid-driven square cavity flows, the initial position of the squirmer has a minimal effect on its swimming behavior and the region of steady-state confinement. Small-sized neutral squirmers are readily confined by the flow to the low-pressure region in the upper left corner of the cavity, while large-sized neutral squirmers engage in periodic circular motion in the cavity's central region. squirmers with different  β exhibit different confined behaviors. Neutral squirmers and pusher with small |β| are generally confined to the low-pressure region in the upper left corner, while puller and pusher with larger |β| are confined to the lower right corner. The Res significantly and intricately affects the confined behavior of the three types of squirmers. In double-lid driven square cavity flows, all types of squirmers demonstrate analogous confined behaviors at low Res, primarily performing periodic circular motions near the cavity's center. However, as Res increases, their motion area expands outward and eventually becomes confined to the lower right corner of the cavity. This study offers novel perspectives on the fluid dynamics of microorganisms within confined environment.

This study focuses on an enhanced folded core sandwich structure, proposing a novel hyperbolic carbon fiber sandwich configuration. Given the structural complexity, a separated male-female mold hot-pressing process was adopted to enable efficient fabrication. Out-of-plane compression tests and Abaqus finite element simulations were conducted to systematically investigate the failure modes, mechanisms, and mechanical properties of three layup configurations:[0/90]3,[0/0/90]S, and [45/-45]3. Corresponding stress-strain curves and failure nephograms were obtained. The results demonstrate that the compressive strengths of the three layup configurations are 6.38 MPa, 6.72 MPa, and 3.94 MPa, respectively. The fiber orientation angle exhibits negligible influence on the failure modes of the core, with all three configurations showing combined failure mechanisms dominated by buckling and crushing.

This paper proposes an isogeometric multi-scale topology optimization method based on reduced Gauss integration for the concurrent optimization design of macro and micro structures. Multi-scale concurrent topology optimization effectively expands the design space and achieves lightweight structures with more ideal mechanical performance. Meanwhile, this paper adopts a reduced Gauss integration method, which reduces the computational complexity and improves the computational efficiency by reducing the number of integration points. Finally, it is shown through several standard arithmetic examples that the computational efficiency using the reduced Gaussian integration method is improved by more than 77 % compared to the traditional method, and the effectiveness and smoothness of the isogeometric multiscale topology optimization are verified.

To enhance the computational efficiency and accuracy of determining the maximum shear stress in the surface layer of urban asphalt pavement structures, this study calculated shear stress values for multiple pavement configurations using an elastic layered system analysis program. These results served as training data for a neural network model. A BP (Back Propagation) neural network fitting model was subsequently constructed in MATLAB, with optimization via the Levenberg-Marquardt (LM) algorithm. Global sensitivity analysis was then performed using the Sobol method to identify key parameters influencing the maximum shear stress in the surface layer. The results indicate that the proposed BP neural network model exhibits a strong correlation with calculations from the elastic layered system program, achieving a correlation coefficient R exceeding 0.999 9. Sensitivity analysis revealed that the surface layer thickness predominantly affects the maximum shear stress under low horizontal force coefficients, whereas the shear modulus of the surface layer becomes more influential at higher coefficients. The developed shear stress calculation model significantly improves both efficiency and accuracy, offering a valuable reference for optimizing urban asphalt pavement structural design.

In marine environments, the corrosive effects of chloride and sulfate ions on concrete structures are significantly more severe than those of other ions. However, there remains no consensus regarding the evolution of concrete strength under their synergistic attack. This study conducted coupled chemo-mechanical experiments on concrete specimens immersed in sodium sulfate solution, sodium chloride solution, and their mixed solutions with varying concentrations. By incorporating chemical reaction equations and kinetic rate equations governing continued hydration and aggressive ion transport, a chemo-mechanical conversion methodology was employed to elucidate the competitive mechanism between chloride ingress and sulfate attack in influencing concrete strength. Furthermore, a chemo-mechanical constitutive model was developed to characterize the strength evolution of concrete subjected to coupled sulfate-chloride corrosion.

Self-circulating microchannel gas generators have garnered significant attention owing to their zero parasitic power consumption. Elucidating the bubble dynamics mechanism is critical for optimizing microchannel design in high-efficiency gas generators. In this study, a numerical simulation of bubble motion in self-circulating micro gas generators was conducted using Ansys Workbench, based on fluid dynamics principles. The effects of key parameters, i.e., microchannel height ratio h ̅, wall contact angle θ_t, and hydrophobic layer contact angle θ_f were systematically investigated. The results demonstrate that: (1) In hydrophobic layer-free channels, when h ̅≥3 and 0<cosθ_t≤0.5, the bubble transit time can be accurately predicted by 1/Δp. In other cases, shifts in vortex positions and quantities enhance energy dissipation, causing the bubble transit time reduction rate to decelerate and deviate from theoretical predictions. (2) For channels without hydrophobic layers, the bubble transit time decreases monotonically with increasing h ̅. This trend stabilizes when h ̅≥5. Additionally, the transit time decreases with higher cosθ_t, though the rate of decrease diminishes significantly when θ_t≤45°. Notably, θ_t exhibits greater sensitivity to transit time compared to h ̅. (3) Introducing a hydrophobic layer in the main channel triggers damping oscillations at the bubble interface due to abrupt contact angle transitions. Larger |cosθ_f  | values amplify oscillation amplitudes, thereby increasing energy dissipation and invalidating theoretical time predictions. Furthermore, the transit time decreases linearly with increasing |cosθ_f  |.

This study investigates the dynamic response characteristics of soil-tank-liquid (STL) systems under various seismic excitations through shaking table model tests. Based on the geotechnical conditions of a soft soil site in Shanghai, scaled soil models were fabricated for experimental simulations. Three representative seismic waves were selected as input motions: conventional-frequency El Centro wave, near-fault pulse-type TCU075 wave, and far-field long-period HKD095 wave. A series of comparative experiments were conducted on both free-field (FF) soil models and STL systems under progressively increasing peak ground accelerations (PGAs). Key findings include: (1) The HKD095 wave, containing rich low-frequency components (0.1-0.5 Hz), induces pronounced amplification effects across these frequencies. This results in the maximum soil surface acceleration amplification factor of 2.64 at PGA = 0.2 g, exceeding those observed under El Centro and TCU075 excitations. (2) Tank wall acceleration responses demonstrate height-dependent amplification, with HKD095 excitation producing larger peak accelerations compared to other waveforms. (3) Hydrodynamic pressure distributions reveal maximum values near the tank base, corresponding to the vulnerable "elephant's foot" buckling zone observed in seismic damage patterns. (4) Liquid sloshing amplitudes under HKD095 excitation reach critical levels at PGA = 0.4 g, approaching the tank's freeboard height and substantially increasing spillage risk.

Current research on vibration analysis of multilayered lattice sandwich beams, particularly multi-span configurations, remains scarce. This study proposes an analytical formulation where the modal shapes of multi-span multilayered lattice sandwich beams are approximated using interpolation-corrected uniform beam modes through the Assumed Modes Method (AMM). The structural dynamics framework incorporates Reddy's Third-order Shear Deformation Theory (TSDT) to establish deformation relationships, with the governing equations derived via Hamilton's principle. The calculated natural frequencies of the multi-span multi-layer pyramidal sandwich beams show good agreement with those obtained from finite element analysis, indicating that the method presented in this paper applies to the problem of free vibration of the multi-span multi-layer lattice sandwich beams. In particular, the influences of changes in span length ratio and span number on the vibration characteristics of multi span and multi-layer lattice sandwich beams are analyzed.

Time-dependent rheological behavior of soft rock in deep-buried tunnels poses significant engineering risks, particularly through accelerated liner deterioration that compromises structural safety and long-term durability. To quantify creep mechanisms and mitigate associated damage, this study develops a 3D viscoplastic numerical model implementing a time-dependent Bingham constitutive formulation. The influence of radial depth, burial depth, and lateral pressure coefficient on the rheological behavior of the surrounding rock of the tunnel are examined. The research results indicate that the deformation of the surrounding rock gradually increases with time, but the rate of increase gradually decreases and eventually stabilizes. Moreover, the greater the rheological deformation of the surrounding rock, the longer it takes to reach stability. In addition, the 15 day period after the completion of construction is a period of rapid development of rheology. By comparing the deformations at various locations, it was found that the arch crown had the largest deformation, followed by the arch foot, and the arch shoulder had the smallest deformation. Moreover, the farther away from the excavation surface around the tunnel, the less the influence of the tunnel on the surrounding rock. The rheological behavior at each location increases linearly with the increase of burial depth. In actual construction, safety verification and monitoring should be emphasized for sections with larger burial depths. When the lateral force coefficient is greater than 1.4, the most unfavorable position of the tunnel shifts from the arch crown to the arch foot. In actual construction, safety monitoring at the arch foot should be strengthened to provide reference for actual engineering construction and design.