Search Results (239)

The feasibility of employing the dynamic substructure approach for seismic response analysis of complex structures has been widely recognized. However, the analytical accuracy of this method is affected by several factors, including the element type, the structural configuration, and the analysis method. To address these issues, four types of reticulated shell structures were designed and analyzed using the mode superposition response spectrum method (MSRSM) and the time history analysis method (THAM). The displacements of the key nodes and all member stresses were extracted to compare the simplified finite element models with the original models. The relative errors of nodal displacements calculated by the models with reduced degree of freedom (DOF) were within 1.62%. For the member stresses of the single-layer reticulated shells, the relative errors of the simplified models were within 14.35%. In the simplified models of double-layer reticulated shells, several members exhibited a relative error greater than 30%; however, these members were mainly located near the substructure boundaries and accounted for less than 0.62% of the entire structure. Three strategies are proposed to mitigate the influence of the member stress errors on the structural analysis conclusions for double-layer reticulated shell structures. In addition, the dynamic substructure method was extended to the coupled system of large-span spatial structures and point-supported glass facades. The seismic response results confirmed that this method effectively reduces computational costs while maintaining satisfactory accuracy, indicating that it is a useful tool for simplifying large-span spatial structures in extensive numerical analyses. Full article

The thermal stability of the shot-peened gradient microstructure in LPBF AlSi10Mg during annealing at 500 °C for up to 8 h was investigated. EBSD boundary maps were analyzed using the box-counting method to determine fractal dimension, D, as a quantitative descriptor of grain-boundary geometrical complexity. It was found that D decreased with depth from the dynamically recrystallized surface layer (0–10 µm; D = 1.73; ECD = 0.8 ± 0.2 µm) through the transition layer (10–40 µm; D = 1.56; ECD = 2.4 ± 0.7 µm) to the matrix (>40 µm; D = 1.16; ECD = 3.0 ± 0.7 µm). After 5 min, the surface network simplified (D = 1.63; ECD = 2.4 µm), whereas the transition layer exhibited increased complexity (D = 1.72; ECD = 3.7 µm), suggesting a strong contribution of near-surface particle pinning and extensive recovery/polygonization within the subsurface. The matrix showed a transient increase in D (1.16 → 1.71), associated with fragmentation of the cellular Si network. Continued annealing reduced D in the surface and transition layers to 1.50 and 1.64 after 1 h due to progressive boundary smoothing and consumption of deformation substructure. Prolonged exposure triggered sparse discontinuous recrystallization exclusively within the transition layer, producing abnormally large grains that migrated bi-directionally into both the pinned surface layer and the bulk matrix. After 8 h, the gradient microstructure collapsed and the boundary trace became disconnected, yielding an apparent exponent D ≈ 0.97 at the map scale. Full article

To address the complex dynamic behavior of four-axis precision motion platforms under high-speed and high-acceleration conditions, as well as the difficulty of traditional modeling methods in balancing accuracy and efficiency, this paper proposes a data/model-driven dynamic modeling and analysis method that integrates generalized receptance coupling substructure analysis (GRCSA) with artificial intelligence (AI) algorithms. Based on the GRCSA theory, the initial analytical framework of the dynamic model of the precision motion platform is established, and the frequency response functions (FRFs) of the substructure and interface are preliminarily obtained. On this basis, the nonlinear prediction model of the dynamic parameters of the interface driving direction is established by using the AI algorithm, enabling fast and accurate prediction of the dynamic characteristics of the interface under different servo control parameters in the guide rail driving direction. Finally, based on the data/model-driven dynamic modeling and analysis method, the interface control parameters are optimized. The interface and substructure parameters are modified to reduce the prediction error of the FRFs from 3.50% to 2.47%. This method can achieve the prediction error of the dynamic characteristics of the interface under different control parameters of about 2.5%. Full article

Gantry-type dual-drive feed systems are widely used in high-precision CNC machine tools, and their synchronization performance directly affects machining accuracy and operational stability. To reduce synchronization errors caused by load-position variation, nonuniform stiffness distribution, and inertia mismatch, this study proposes a structural parameter optimization method for a gantry-type dual-drive feed system. The novelty of this work lies in integrating position-dependent dynamic modeling, critical-position identification, sensitive structural-parameter selection, and response-surface-based optimization into a unified framework for synchronization-error reduction. First, a position-dependent dynamic model is established using modal reduction, spline interpolation, and substructure synthesis. The dynamic model is then coupled with a servo control model to construct an electromechanical coupling model, which is validated experimentally on a gantry-type dual-drive feed system. Next, the synchronization-error distribution over the entire workspace is evaluated, and the critical position with the poorest synchronization performance is identified. Based on sensitivity analysis, the key structural parameters affecting synchronization error are selected as design variables. A response surface surrogate model is then constructed, and particle swarm optimization is used to obtain the optimal structural-parameter combination. The results show that the synchronization error at the critical position is reduced by 20.5%, while the average synchronization error at the validation positions is reduced by 17.3%. These results demonstrate that the proposed method can effectively improve the synchronization accuracy of gantry-type dual-drive feed systems and provide practical guidance for the structural design of high-precision dual-drive machine tools. Full article

The application of robotic milling manipulators demonstrates a promising method for the efficient manufacturing of large-scale structures. However, the cutting accuracy and efficiency of milling robots are predominantly subjected to their low stiffness, which may easily cause chatter during machining. Accurate prediction of chatter stability for robots is of practical importance and is challenging. This paper develops a dynamic model of flexible link elements by considering link flexibility and joint torsional deformation and then constructs a multi-link flexible coupled dynamic model using the receptance coupling substructure analysis (RCSA) method. Subsequently, the equivalent dynamic parameters are identified via the particle swarm optimization (PSO) algorithm. On this basis, the end-effector frequency response functions (FRFs) of the robot under different poses are predicted, and the stability lobe diagram (SLD) for milling is generated based on chatter theory. Finally, the predicted FRFs and stability regions are validated through modal tests and milling experiments. Experimental results demonstrate that the proposed model can predict the end-effector dynamic characteristics and chatter occurrence conditions under different poses, confirming its effectiveness in the analysis of milling chatter stability. Quantitative validation yields a maximum error of 3% for predicted first-order modal frequencies and relative modal amplitude errors below 10%, with experimentally confirmed critical depths of cut of 0.1–0.2 mm at 3000 rev/min and 0.5–0.6 mm at 5000 rev/min. Full article

Pumping stations serve as the foundation platform for large-scale vertical fluid machinery, and their structural dynamics directly govern the vibration levels and long-term reliability of the installed pump units. In low-head vertical pumping stations, the interaction among the massive underwater substructure, flexible above-ground powerhouse, and surrounding backfill soil creates a complex dynamic system whose behavior remains insufficiently characterized. This study presents a comprehensive dynamic analysis of a large-scale vertical pumping station using a high-fidelity three-dimensional finite element model that incorporates the powerhouse superstructure, submerged concrete substructure, and backfill soil. Modal analysis under four boundary condition scenarios—varying in soil participation and interface contact conditions—systematically quantifies the influence of soil–structure interaction on natural frequencies and mode shapes. Resonance verification against three primary excitation sources—rotational frequency (4.917 Hz), blade passage frequency (24.583 Hz), and rotor–stator interaction frequency (196.667 Hz)—is extended from the first 50 modes to the 400th mode to assess potential high-order resonance risks. Results show that the roof slab, with its large span and low stiffness, exhibits the highest vibration susceptibility. For the rotational frequency, modes 4–12 fall below the 20% code-specified safety margin but rapidly exceed the threshold thereafter. For the blade passage frequency, the separation ratio decreases progressively with increasing mode order within the first 50 modes, and the extended analysis up to the 400th mode shows that the separation ratio remains well above 20% throughout modes 51–400. Consequently, no substantial resonance risk exists for the blade passage frequency within the entire computed range. The rotor–stator interaction frequency remains safely separated with margins exceeding 95%. These findings demonstrate the profound influence of soil–structure interaction and confirm that, despite a decreasing trend in frequency separation at higher orders, the blade passage frequency poses no substantial resonance risk up to the 400th mode. This work provides a rigorous analytical framework for vibration-informed design and optimization of pump foundation systems, with direct implications for the reliability and operational safety of large-scale vertical fluid machinery. Full article

An efficient hybrid modeling framework is developed for the dynamic analysis of offshore wind turbines (OWTs) by coupling an aero–servo–elastic rotor–nacelle superelement with a hydroelastic substructure. The complex rotor–nacelle dynamics are condensed into a reduced-order 14-DOF representation through a modal-based multibody formulation, while retaining blade deformation, spinning effects, nonlinear aerodynamic loading, and active servo controls. Its interface compatibility at the nacelle enables the coupling with either numerical or physical substructures, establishing a unified basis for system hybrid formulation, co-simulations, and real-time hybrid simulations. The validity of the superelement is verified by comparing the resulting fully coupled modal model against OpenFAST, demonstrating high consistency in time-domain responses. As a demonstration, the verified superelement is further coupled with a 1D finite element model of the supporting structure (tower–monopile substructure) to form a hybrid model, enabling accurate force analysis of the OWT structure. Dynamic analyses of the IEA 10 MW OWT reveal that while the blade flapwise responses and the operation-related edgewise responses are 1P-dominated, tower side–side responses and idling-related tower fore–aft and blade edgewise responses manifest at their corresponding resonance frequencies. The maximum displacement and maximum bending moment envelopes vary monotonically with height. Instead, the maximum stress envelope possesses high values in the mid-lower sections of the tower. This high-stress region undergoes a spatial shift driven by the blade feathering mechanism. Full article

Session-based recommendation seeks to deliver personalized suggestions by decoding transient interaction sequences generated by anonymous users. Although graph neural networks have advanced this field by modeling pairwise item transitions, they fundamentally struggle to capture the complex, high-order dependencies inherent in real-world user behavior modeling. Consequently, while hypergraphs offer a natural mathematical solution for representing these multi-item relationships, existing approaches frequently overlook the localized structural semantics necessary to ground these abstract relations in physical browsing logic. To address these limitations, we introduce MoHyNet, a novel motif-guided hypergraph framework explicitly designed to capture both inter- and intra-session dependencies. By extracting localized hypergraph motifs, MoHyNet effectively decodes the recurring topological sub-structures and latent intentions behind anonymous interactions. Rather than treating hypergraphs merely as static representations of item co-occurrence, our approach utilizes these motifs as dynamic semantic filters to extract stable behavioral signatures from pseudo-sequential noise. To complement this intra-session modeling, we construct an augmented line graph that maps multi-hop dependencies across different sessions, employing neighborhood-aware convolutions to aggregate global collaborative context. A dual-view contrastive learning optimization is subsequently integrated to semantically align these intra-session structural signatures with the inter-session global context, ensuring a robust and holistic understanding of user intent. Extensive empirical evaluations on three real-world e-commerce datasets demonstrate that MoHyNet consistently outperforms state-of-the-art methods in session-based recommendation performance. Full article

Laser powder bed fusion (LPBF) of aluminum matrix tribological composites holds high potential for advanced bearing applications, yet its widespread implementation is often constrained by high porosity and severe residual stresses. In this work, the influence of hot pressing (HP) temperature (100–400 °C) on the microstructure, substructural evolution, mechanical properties, and fracture mechanisms of the LPBF Al-10Sn-10Pb alloy was investigated to achieve simultaneous densification and matrix optimization. Processing was carried out at 300 MPa with a 30 min holding time. It was established that at temperatures >200 °C, near-full consolidation is achieved through liquid-assisted pore closure. Increasing the temperature leads to the coarsening of Sn and Pb inclusions and the disruption of the initial dispersed network of soft phases. Williamson–Hall analysis revealed a transition from dislocation accumulation at 100 °C (~15 × 10¹³ m⁻²) to dynamic recovery at 200 °C, followed by matrix recrystallization at higher temperatures. A combination of strength (up to 127 MPa) and ductility (~11%) is realized at 200 °C due to the synergy between remaining substructural strengthening and pore healing. At 300–400 °C, the strength decreases to 108–113 MPa with a concomitant increase in ductility to 34–44%. A shift in fracture mechanisms from quasi-brittle to ductile is shown; at 400 °C, the development of intergranular fracture associated with the influence of liquid phases is possible. Full article

To address the longitudinal vibration issues in high-speed elevators induced by external excitations, this study constructs a high-precision multi-degree-of-freedom (MDOF) dynamic model to systematically analyze vertical dynamic response characteristics. Utilizing the substructure method, the complex traction system is decomposed into several subsystems, including the traction device, tensioning device, car and car frame, counterweight system, and segmented wire ropes. By integrating Lagrange’s equations with Newton’s second law, the governing differential equations of motion for each component are derived, establishing an adaptable global dynamic model. The forced vibration analysis focuses on the impacts of periodic excitation from traction sheave eccentricity, piecewise reverse braking torque, and vertical impacts from guide rail joints on car vibration response and wire rope dynamic stress. The results indicate that: traction sheave eccentricity leads to periodic fluctuations in car acceleration, with vibration peaks decreasing as the payload increases; reverse braking torque triggers impulsive acceleration overshoots, where the peak value under full-load conditions increases by approximately 15% compared to the no-load condition, accompanied by a longer duration of low-frequency vibrations; guide rail joint impacts produce instantaneous acceleration spikes, which increase by about 18% under high-speed operating conditions; and the wire rope stress exhibits significantly higher sensitivity to load variations within the low-load range of 0–0.2. Full article

Social media engagement drives both individual behavior and content dissemination, yet traditional analytics often reduce interactions to simple counts, obscuring the complex structures underlying user activity. In the highly competitive digital landscape, understanding how users interact with content is crucial for businesses aiming to optimize social media campaigns and maximize return on investment (ROI). Traditional engagement metrics, such as likes and shares, fail to capture the underlying structure and dynamics of user behavior. This study investigates the latent patterns of engagement by combining topological data analysis (TDA) with behavioral clustering across 100,000 posts on multiple platforms. Using persistent homology and k-nearest neighbour graphs, we reveal a primary bifurcation between Active (validation-focused) and Passive (consumption/propagation) users, nested four-strain substructures, and over 650 significant $H_1$ loops indicating recurring feedback cycles. Active users exhibit strong cluster cohesion and high engagement rates, while Passive users contribute broadly to content diffusion with slightly higher loop counts, highlighting distinct functional roles in social media dynamics. These findings provide a principled framework for targeting content, reinforcing feedback loops, and leveraging hub posts to amplify engagement. By linking topological structure to behavioral patterns, this work advances both the theoretical understanding of digital interaction and the practical design of more effective social media campaigns. Full article

Graph neural networks (GNNs) have achieved great success in graph classification, with graph pooling methods being widely adopted for related tasks. Existing approaches typically rely on node ranking or clustering to coarsen graphs, but often fail to effectively leverage global structural information, leading to loss of critical substructures and limited interpretability—key limitations in molecular analysis and social network mining. To address these issues, we propose SparsePool, a graph pooling method that integrates node features and structural patterns through atomic decomposition. By dynamically decomposing graphs into interpretable atomic units via Boolean matrix factorization, SparsePool preserves semantically meaningful substructures while providing transparent evidence of retained patterns. We further introduce an Atomic Pooling Neural Network (APNN) for graph representation learning. Extensive experiments on relevant benchmarks including biochemical and social network datasets demonstrate that SparsePool outperforms state-of-the-art pooling methods, achieving an average classification accuracy improvement of 1.03% over baseline models while reducing structural information loss. We also discuss its compatibility with emerging quantum computing paradigms, such as quantum-accelerated sparse decomposition, as a promising direction for scaling graph processing in industrial contexts. Full article

Background/Objectives: Operational DNA databases traditionally rely on static locus-count thresholds to determine search eligibility and report matches. While computationally straightforward, these rigid criteria routinely discard high-value investigative leads from degraded forensic profiles while simultaneously permitting adventitious matches when common alleles are involved. To overcome the limitations of static rules, this study introduces an automated framework for dynamic likelihood ratio (LR) thresholding. Methods: Utilizing a Fast Fourier Transform (FFT) algorithm, the system calculates the Probability Mass Function (PMF) for any specific combination of shared loci in real-time, natively incorporating the Balding–Nichols model to account for population substructure. Instead of applying an arbitrary locus count or fixed LR cutoff, the framework defines admissibility based on a user-defined maximum upper bound of acceptable false positives at a specified confidence (probability) level (e.g., 95%). Results: This empowers database custodians to precisely predict and adapt their search criteria to match an acceptable administrative workload, dynamically adjusting the required LR threshold to the exact size of the searched database. This approach was validated through massive-scale empirical simulations across five reference population groups. Receiver Operating Characteristic (ROC) and Poisson distribution analyses reveal that static thresholds inevitably collapse under the multiplicity effect of large-scale comparisons; for instance, a static locus rule that maintains safety within a small DNA database yields an unmanageable false positive risk when scaled to larger DNA databases or international networks like the Prüm DNA Exchange. Conclusions: By explicitly coupling the decision threshold to the database size and the genetic rarity of the evidence, this dynamic framework provides a mathematically rigorous and scalable solution. Most notably, it identifies rare, low-locus matches that static rules typically discard, offering a method to maintain a predefined expected false positive rate. Full article

Shell structures with built-in plates are widely used in engineering. This paper presents a unified analytical method for the dynamic stiffness model of coupled plate–shell structures, considering the effect of internal plates on the vibration characteristics of the assembled system. The coupled structure is decomposed along the plate–shell interface. Using Gorman’s superposition method and structural symmetry, the boundary displacement solution of the full structure is simplified to a quarter-structure problem. The dynamic stiffness matrices of substructures are derived and assembled to establish the analytical model. Numerical examples are conducted to investigate the dynamic behaviors of the coupled system, and the convergence and accuracy of the proposed method are verified against numerical simulations. Furthermore, a test rig is established for a rectangular plate–cylindrical shell structure, and modal experiments are carried out. The measured natural frequencies and mode shapes agree well with theoretical predictions. The proposed method provides a general theoretical approach for the vibration analysis of plate–cylindrical shell coupled structures. Full article

Precise segmentation of cardiac substructures in magnetic resonance imaging is pivotal for diagnosis and treatment planning but remains impeded by anatomical scale heterogeneity and the scarcity of high-quality pixel-level annotations. Existing deep learning paradigms often struggle to simultaneously resolve the global geometry of ventricular cavities and the fine-grained boundaries of the myocardium, particularly in low-data regimes. To address these challenges, we propose SAS-SemiUNet++, a holistic semi-supervised segmentation framework. This architecture incorporates two novel mechanisms: (1) The Scale-Aware Semantic Recalibration (SASR) unit, which functions as a dynamic semantic gate to adaptively adjust receptive fields, mimicking a radiologist’s variable-focus mechanism to capture multi-scale anatomical details, and (2) Stochastic Consistency Regularization (SCR), a dual-path perturbation strategy that enforces geometric invariance on unlabeled data, thereby mitigating overfitting to noisy pseudo-labels. Comprehensive evaluations on the ACDC benchmark demonstrate that SAS-SemiUNet++ significantly outperforms state-of-the-art methods, achieving superior segmentation accuracy and boundary fidelity, particularly in reducing the 95% Hausdorff distance. This study presents a data-efficient and robust solution for cardiac image analysis, offering potential for scalable clinical deployment. Full article