国际机械系统动力学学报(英文)最新文献

This cover illustrates a data-driven approach to predicting 3D printing parameters using a combined reverse–forward model architecture, addressing the inverse problem of inferring optimal printing parameters from target consumption parameters and geometric constraints. A flexible training strategy with dynamically adjusted weighting coefficients across different training phases enables the model to outperform conventional fixed-weight approaches.

This study presents a novel neural network architecture called spectral integrated neural networks (SINNs), which combines physics-informed neural networks (PINNs) with time-spectral integration techniques to efficiently solve two- and three-dimensional dynamic piezoelectric problems. To avoid the numerical instability associated with time-differential operators, the coupled system of mechanical and electrical equilibrium equations is reformulated into a weak time-integral form. The temporal derivatives of displacement and voltage fields, treated as the primary unknown physical quantities, can be approximated utilizing fully connected neural networks (FCNNs). The displacements and electric potential are subsequently recovered through time-spectral integration of their respective derivatives. A physical-informed loss function is formulated by the weak time-integral type of the governing equations and boundary conditions, with the initial conditions embedded within the integral expressions. The proposed SINNs demonstrate superior stability and accuracy, even under large time steps conditions. Numerical verification is accomplished through three representative test cases of the method, and a comparison analysis is presented between the results obtained by the SINNs and those from the PINNs.

Cover Caption: Scoliosis Rehabilitation with a Robotic Brace Powered by RL-based Impedance Control and Digital Twin: Adolescent Idiopathic Scoliosis (AIS) is commonly treated with traditional braces that rely solely on passive strap tensioning, lacking intelligent control strategies. This study proposes a reinforcement learning-based position-based impedance control (RLPIC) method for robotic braces to enable active human–robot interaction. To safely simulate and train the control system, a novel five-dimensional, three-layer digital twin (DT) model is developed, integrating physical modeling, digital modeling, bidirectional interaction, and optimization, enhanced by a neural network-based parameter estimator. Both numerical simulations and real-time experiments validate the DT and RLPIC framework, demonstrating improved tracking and interaction performance in AIS treatment.

In recent years, the field of 3D printing has heavily relied on expert knowledge and complex trial-and-error procedures to determine appropriate printing parameters that meet desired consumption specifications. This study introduces a novel method for predicting 10 printing parameters based on 7 geometric features and 3 target consumption constraints (time, length, weight). Rather than using a traditional autoencoder model, we implement a variant that combines a reverse model with a forward-pretrained model. The forward model, pre-trained using XGBoost, predicts the 3 target consumption parameters from the 7 geometric features and 10 printing parameters. The reverse model then generates the 10 printing parameters from the 7 geometric features and the desired 3 consumption constraints. Through staged training and optimized loss function adjustments, our model achieves an R² of 0.9567, demonstrating its precise predictive capabilities and potential to optimize the 3D printing process while reducing reliance on expert intervention.

This paper analyzes the dynamic behavior of a two-degree-of-freedom system subjected to electromagnetic interaction modelled through a skew-symmetric coupling matrix. The system comprises two mechanically independent oscillators coupled by velocity-dependent electromagnetic forces. The equations of motion are formulated and analyzed in the modal domain, highlighting the effects of the antisymmetric interaction on natural frequencies and mode shapes. The classical orthogonality is broken, resulting in complex eigenvectors; nevertheless, the system remains conservative, as the interaction forces perform no work. The analysis is carried out using both configuration-space and state-space formulations, revealing modal frequency splitting and phase shifts induced by the skew-symmetric term. These modal features are further examined through time-domain simulations and frequency response functions. The main contribution of this study is the development and analysis of a deliberately simple yet general model that isolates the essential dynamic effects of skew-symmetric electromagnetic coupling. This minimal formulation, often hidden in more complex systems, reveals key phenomena such as modal frequency splitting, non-normal modes, and energy-conserving cross-effects. The model serves not only as a conceptual reference but also as a methodological framework applicable to a broad class of coupled electromechanical systems.

Accurate prediction of fatigue life under multiaxial loading conditions remains challenging due to complex stress–strain interactions. In this study, we integrate machine-learning (ML) regression with variance-based sensitivity analysis (SA) to predict multiaxial fatigue life in CuZn37 brass and to identify the dominant mechanical factors influencing fatigue damage. Several surrogate models were evaluated, with the Gaussian Process model achieving the highest accuracy (R² = 0.991) while maintaining robust generalization across loading paths. Gradient Boosting, Random Forest, and Penalized Spline Regression models also demonstrated strong predictive capabilities. Importantly, the SA explicitly accounted for statistical dependencies among input parameters, revealing that normal strain–stress interactions account for over 40% of the total variance in fatigue life. In contrast, shear-related parameters exhibited secondary, compensatory effects. These results highlight the importance of capturing parameter dependencies in fatigue modeling and demonstrate that ML-based surrogates can help provide both high-fidelity predictions and physical insights under complex multiaxial loading conditions.

Random loadings (RL) are prevalent in mechanical systems, yet their inherent stochasticity poses significant challenges to structural fatigue reliability assessment. In this study, a three-dimensional fatigue reliability model is developed under RL through amplitude modulating and Fourier transformation. The non-Gaussian RL characteristics are accurately characterized by employing power spectral density and loading kurtosis. The equivalent initial crack size distributions are evaluated through three-dimensional fatigue growth theory by joint use of the standard fatigue stress-life (S-N) data and the fatigue crack growth data of the materials. Fatigue life distributions in specimens made of different materials with different geometries and thicknesses are analyzed under RL. It is shown that fatigue life exhibits negative correlations with power spectral density, kurtosis, and initial crack size. Especially, it is found that fatigue life and kurtosis approximately follow a power–law relationship, with both mean and variance decreasing as kurtosis increases. Validations against the experimental data available in the literature show that the present model can provide an efficient prediction of the fatigue life of mechanical systems under RL with limited experiment data.

This study presents a comprehensive and standardized foundation for the mathematical modeling and control of flying-base-mounted manipulators, addressing several critical challenges in aerial robotics. The primary contributions of this study include: (1) the development of a unified framework for computing the system's generalized forces, incorporating both active motor inputs and passive constraint forces; (2) a trajectory planning method for the flying base that simultaneously accounts for both desired position and orientation; (3) an automatic and recursive methodology for deriving the system's equations of motion, ensuring that increasing the number of links in the manipulator or flying base does not introduce limitations; and (4) a motor configuration strategy that enables the flying base to achieve unrestricted motion in three-dimensional space. To address these challenges, the proposed approach systematically decomposes the robot structure—consisting of the flying base and the mounted manipulator—into a set of substructures. Each substructure, modeled as an open kinematic chain with a moving base, is analyzed using the recursive Gibbs-Appell algorithm to derive its equations of motion. These individual equations are then integrated to obtain the coupled dynamics of the complete system, capturing the mutual interactions between the flying base and the manipulator. Finally, a feedback linearization-based controller is designed to enable simultaneous trajectory tracking of both the flying base and the manipulator's end-effector. Simulation results validate the effectiveness of the proposed control strategy, demonstrating its ability to achieve precise positioning and accurate orientation tracking of the entire robotic system.

To address the complex coupling between aerodynamic characteristics and guidance control for morphing flight missiles, this study proposes a data-driven approach to integrated adaptive morphing and guidance. Firstly, an aerodynamic surrogate model is constructed using a fully connected neural network (FCNN), mapping the configuration parameters to aerodynamic parameters. Secondly, an adaptive physical parameters optimization network (PPON) is developed to optimize aerodynamic characteristics based on predictions from the aerodynamic surrogate model. Thirdly, an integrated morphing and guidance model is derived by applying the proximal policy optimization (PPO) algorithm from deep reinforcement learning (DRL), embedded with the adaptive aerodynamic optimization model. Eventually, the proposed integrated approach is applied to the guidance task of a morphing cruise missile with variable camber wings. Simulation results demonstrate that the integrated guidance model significantly enhances aerodynamic performance and generates more continuous guidance commands within approximately 4.3 s, outperforming the deep Q-Network (DQN) algorithm under morphing flight conditions. Moreover, compared to the PPO and DQN-based guidance laws without morphing flight conditions, the integrated model improves both the guidance accuracy and terminal kinetic energy. Furthermore, the integrated guidance model, trained on stationary targets, remains effective for engaging moving and maneuvering targets, showcasing its robust generalization capability.

The effects of barrel erosion on artillery firing performance have long been a subject of concern, but its effect on launch uncertainty has yet to be investigated. This article explores the influence of barrel erosion on the interior ballistic mechanical properties and launch disturbances. The interior ballistic mechanical properties under various barrel erosion conditions are tested, revealing a significant impact on the projectile lateral overload. Utilizing random matrix theory, a projectile-barrel coupled calculation model is developed, accounting for parameter-model uncertainties. Subsequently, a Bayesian posterior model uncertainty quantification method based on lateral overload root mean square (RMS) is proposed, and quantification and inversion are conducted based on the test results. The computational results confirm the accuracy of the quantification technique and highlight the effectiveness of the model uncertainty approach in addressing complex uncertainty issues, such as barrel erosion.