Industrial robotic systems operate under complex spatial and temporal interactions, making reliable fault diagnosis critical for safety, productivity, and maintenance efficiency. Existing data-driven fault diagnosis approaches primarily rely on deterministic deep learning models, which often neglect inter-joint dependency modeling and lack mechanisms to quantify prediction uncertainty, leading to unreliable decisions under ambiguous conditions. To address these limitations, this study proposes an Uncertainty-Aware Evidential Spatio–Temporal Graph Learning framework for industrial robot fault diagnosis. The objective is to achieve accurate, interpretable, and reliable fault identification by jointly learning spatial relationships, temporal dynamics, and predictive confidence. The proposed method models the robotic system as a spatio–temporal graph, where joint interactions are learned using a Graph Attention Network, temporal fault evolution is captured through a Temporal Convolution Network, and uncertainty is quantified via Evidential Deep Learning with Dirichlet outputs. The framework is implemented using Python with PyTorch and PyTorch Geometric, enabling efficient graph-based learning. Experiments conducted on the UR3 CobotOps dataset demonstrate that the proposed Evidential GAT–TCN model achieves an accuracy of 93.4%, representing an improvement of approximately 3–5% over existing CNN-LSTM and graph-recurrent baselines. In addition, the model significantly reduces calibration error and predictive uncertainty, enhancing decision reliability. The results confirm that integrating evidential uncertainty estimation with Spatio–temporal graph learning improves both diagnostic performance and trustworthiness. This study concludes that the proposed framework provides a robust and deployable solution for real-time fault diagnosis in industrial robotic systems, supporting safer and more intelligent maintenance strategies.
Regenerative braking energy recovery is essential for electric vehicles due to their limited driving range. A fuzzy regenerative braking control strategy based on an improved differential evolution algorithm and segmented braking force distribution is proposed. Firstly, a segmented braking force distribution strategy for the front and rear axles is designed by integrating the ideal braking force distribution curve (I-curve) with ECE regulations. This approach ensures a balance between energy recovery efficiency and driving safety under varying braking intensities. Secondly, a fuzzy controller is developed to regulate the proportion of regenerative braking force. The controller takes vehicle speed, battery state of charge (SOC), and braking intensity as inputs based on the recognition of the driver’s braking intention. Furthermore, an improved differential evolution algorithm with adaptive control parameters is introduced to optimize the membership functions of the fuzzy controller in an offline manner. Ultimately, co-simulation using MATLAB/Simulink and CarSim is conducted to validate the proposed strategy. The results demonstrate that, compared to a baseline fuzzy control strategy, the proposed method increases the energy recovery rate by 15.8% and 14.9% under the New European Driving Cycle (NEDC) and the Worldwide Harmonized Light Vehicles Test Procedure (WLTP), respectively. Moreover, when compared to a fuzzy control strategy optimized by the particle swarm optimization (PSO) algorithm, the proposed strategy achieves additional improvements of 1.04% and 2.90% under the same test cycles, thereby effectively enhancing the energy recovery performance of the electro-hydraulic composite braking system.
Under complex operating conditions, a three-module and six-axle virtual track train (VTT) with full-axle steering exhibits severe multibody dynamic coupling, making it challenging to simultaneously balance stability and tracking accuracy. To address this issue, this paper proposes an LQR-Fuzzy PID hierarchical cooperative control strategy. First, a 4-DOF vehicle dynamics model is established, and the Dynamic Relative Gain Array (DRGA) is introduced to analytically decouple multi-axle dynamic interactions, thereby establishing a control allocation architecture where axles 3 and 5 are dedicated to governing vehicle stability. Second, an LQR controller is designed to minimize tracking errors and optimize overall vehicle posture stability, computing the optimal steering angles for axles 1, 3, and 5. Concurrently, distributed Fuzzy PID controllers are introduced to adaptively compensate for nonlinear disturbances, achieving high-precision tracking for axles 2, 4, and 6. TruckSim/Simulink co-simulation results demonstrate that, under a composite urban trajectory scenario at 18 km/h, the proposed strategy reduces the maximum tracking error by 48.6% and 30.7%, and the peak hinge angular velocity by 55.5% and 37.5%, compared to the full-axle LQR and full-axle MPC baselines, respectively. Furthermore, the strategy exhibits excellent robustness to structural parameter mismatches, such as a 5% center-of-mass offset and a 3% degradation in tire stiffness.
The precision of gyroscopes, as core components of inertial navigation systems, critically determines system accuracy and stability. Precision threaded connections, widely used for assembly, are highly susceptible to preload relaxation during repeated start-stop operations, affecting overall performance. However, the short-term relaxation behavior of preload under coupled external loads remains insufficiently understood. This study investigates the preload relaxation of precision threaded connections under thermo-vibrational loading. An improved Iwan-based tangential contact stiffness model is developed to characterize interface slip behavior and stiffness weakening. By combining numerical simulations with experimental validation, the relationship between tangential stiffness degradation and preload loss is quantified through an interface contact state parameter. Experimental results confirm preload relaxation and contact surface hysteresis. Analysis of the gyroscope rotor assembly reveals that a 9.5% increase in stiffness weakening rate under thermal-vibration loading correlates with a 23.7% rise in axial deviation. A strong positive correlation (up to 0.9691) between stiffness weakening and preload relaxation is established. The findings provide critical insights into the dynamic degradation mechanisms of threaded connections and offer theoretical guidance for improving the stability and signal accuracy of precision aerospace systems such as gyroscopes.
Elastic contact of layered elastic plates is common in sports equipment and engineering systems, yet most impact-contact analyses of elastic spheres still rely on classical Hertzian theory based on homogeneous half-spaces. For a table-tennis racket, however, the impact occurs on a strongly heterogeneous layered plate composed of a rubber top-sheet, a sponge layer, and a stiff wooden blade, for which the applicability of Hertzian theory is limited. In this study, a closed-loop framework is established by coupling free-fall rebound experiments, finite element (FE) simulation, Hertzian baseline analysis, and a semi-analytical discrete-convolution fast Fourier transform (DC-FFT) solver for functionally graded-equivalent layered plates. The measured rebound height is used for baseline FE calibration, whereas the contact duration and the contact radius at the maximum compression instant are reserved for validation of the transient contact response. Hertzian theory is then used as a homogeneous-reference estimate, so that the magnitude and direction of its deviation from the calibrated and experimentally checked FE response can be discussed. The DC-FFT framework is used for response-level forward reconstruction of the peak-compression contact state and for generating thickness-dependent maps of maximum compression, peak contact pressure, and contact radius by independently varying the rubber and sponge thicknesses. The results show that sponge thickness is the dominant structural variable governing the global compliance of the layered covering, while rubber thickness mainly provides a secondary near-surface tuning effect. The proposed framework also enables thickness-allocation screening and identifies a favorable rubber–sponge thickness combination for the racket considered. The experiment–FE–DC-FFT framework provides a practical and mechanically interpretable route for thickness design of layered plates under elastic impact. The optimized thickness pair is therefore reported as a model-based design recommendation.
Assembly eccentricities of the stator and rotor disturb the motor air-gap distribution and induce unbalanced magnetic pull (UMP), which subsequently affects the rotational accuracy of aerostatic motorized spindles. To clarify the underlying mechanism, this study establishes a gas–magnetic coupled dynamic model incorporating motor eccentricity. This model employs VBS scripts to realize bidirectional co-simulation between the finite-difference gas-bearing model in MATLAB and the finite-element electromagnetic-field model in ANSYS Maxwell. The developed framework enables real-time feedback of UMP into the rotor dynamic equations. The results indicate that the dominant effect of stator eccentricity is a shift of the whirl center, with a regular 11-lobed waviness superimposed on the shaft-center orbit. By contrast, rotor eccentricity produces a substantially more pronounced enlargement of the whirl radius than stator eccentricity, indicating an amplitude-amplification-dominated response similar to that induced by an additional unbalance mass. Mixed eccentricity simultaneously causes both whirl-center offset and whirl-radius enlargement, while producing more complex local petal-like waviness in the shaft-center orbit, thereby exerting the strongest adverse effect on spindle rotational accuracy. Frequency-domain analysis indicates that the eccentric orbit is formed by the combined action of the fundamental-frequency response and higher-order harmonics, suggesting that its local response contains richer higher-order disturbance components. Experimental results confirm that the synchronous radial error increases with increasing stator and rotor eccentricities. The experimental trends are in good agreement with the corresponding simulation results. The study reveals the distinct mechanisms of UMP under different eccentricity forms and provides a theoretical basis for assembly-error control and high-precision operation of aerostatic motorized spindles.
A multi-scale finite element (FE) framework was developed to analyse the bending, buckling, and free-vibration behaviours of carbon nanotube (CNT)-reinforced aluminium (Al) nanocomposite beams while explicitly accounting for the interphase between CNTs and the metal matrix. A three-phase representative volume element (RVE) comprising CNTs, Al matrix, and an interphase region was employed to determine the effective elastic modulus and density for CNT volume fractions of 2%, 5%, and 10%. The homogenized properties were then incorporated into macroscale FE models of cantilever beams using 2D (beam, plane-stress) and 3D (shell, brick) elements. The predicted elastic moduli showed excellent agreement with experimental data (deviation < 5%), confirming the accuracy of the homogenization scheme. CNT reinforcement substantially improved structural performance: maximum deflection decreased by ∼21%, critical buckling loads increased by 5%–27%, and the first-mode natural frequency rose by ∼16% with 10% CNT. All FE formulations produced consistent results, with 3D models additionally capturing torsional and out-of-plane effects. Overall, the proposed framework successfully links nanoscale reinforcement to macroscale mechanical behaviour, demonstrating that CNT–Al nanocomposites offer outstanding stiffness, stability, and dynamic performance for lightweight structural and aerospace applications.
Data-driven methods have been widely applied to fault diagnosis of high-speed train gearbox gears; however, their performance heavily depends on high-quality training data. Due to the inevitable degradation of gears during long-term service, significant distribution discrepancies arise between real-time operational data and historical data, which degrade the generalization capability and diagnostic accuracy of models. To address this issue, this paper proposes a digital twin-based generalized fault diagnosis method for high-speed train gearbox gears. First, a gearbox dynamic model is established using multibody dynamics and Hertzian contact theory,capable of characterizing operational states and generating vibration data under different health conditions. Second, an interactive generative adversarial network–based data fusion framework is developed. Through an interactive training strategy, the distribution gap between data in the digital and physical spaces is reduced, enabling collaborative learning of fault mechanism information and environmental characteristics. Finally, a deep convolutional neural network is trained using the fused data to achieve high-accuracy generalized fault diagnosis. Experimental results demonstrate that the proposed method achieves an average diagnostic accuracy of 96.70%, outperforming the best comparative method by 2.20% and conventional methods by 4.67%, thereby validating its effectiveness and superiority in improving generalization performance for fault diagnosis.
This study investigates the natural characteristics of carbon fiber reinforced polymer (CFRP) laminates by considering the effects of stacking sequence, layer thickness sequence (LTS), and material anisotropy on the constitutive relationship. A simply supported laminate model is established using the state-space method. The analytical solution is derived through the transfer matrix method. The proposed model has been validated against literature data and has been further verified by finite element analysis and vibration experiments. The main innovations and contributions of this study are as follows. First, the coupled effects of LTSs and stacking sequences on natural frequencies, mode shapes, elastic displacement fields, and interlaminar stress distributions are systematically revealed. Second, an analytical-simulation-experimental validation framework has been established to improve the reliability of the proposed method. Third, the state-space method has been extended to laminates with different layer material combinations to evaluate the influence of material distribution on natural characteristics. Interlaminar modal order variations are interpreted using the ABD stiffness matrix. The results show that the natural characteristics of CFRP laminates can be effectively regulated by adjusting LTSs, stacking sequences, and material distributions. This study provides a useful reference for vibration reduction and structural design of CFRP laminates.
This paper introduces a comprehensive leakage-free methodology for bearing fault diagnosis by three of the most representative feature learning approaches: Raw Signal Convolutional Neural Network (Raw CNN), Fast Fourier Transform (FFT) based hybrid MLP-CNN, and handcrafted statistical descriptor based Multilayer Perceptron (MLP). The vibration data collected from the Case Western Reserve University (CWRU) bearing set for four operating conditions (0, 1, 2, and 3 HP) was used. To eliminate information leakage, all models were trained under the base-line condition 0 and tested under the unseen conditions 1, 2, and 3 HP, which were cross loads to assess the robustness of the models objectively. The Raw CNN achieved accuracies of 59.18%, 59.84%, and 61.92%, while the FFT-based MLP-CNN improved performance to 79.92%, 77.01%, and 65.22%. The handcrafted feature based MLP gave the best accuracy of 99.47%, 89.43%, and 78.92% for each of the three datasets. In order to support the evaluation, computational complexity, external validation with the Paderborn University bearing dataset, comparison with alternative machine learning classifiers, SHAP, and Permutation Feature Importance (PFI) analyses were performed. The external validation yielded 98.76% accuracy for classification, further confirming the generalization capability of the proposed framework. The results show that the handcrafted statistical descriptors, along with a simple MLP, is an interpretable, computationally efficient, and robust method for cross-load bearing fault diagnosis.
Accurate parameter identification of permanent magnet synchronous machines (PMSMs), including stator resistance, inductances, and permanent magnet flux linkage, is crucial for advanced controller design in robotic systems. Conventional identification methods, utilizing polynomial-based magnetic models to capture magnetic saturation and cross-coupling effects, suffer from degraded inductance identification accuracy under nonideal factors such as inverter nonlinearity and core loss effects. To address this issue, this paper proposes an offline parameter identification method for PMSMs based on a physically feasible magnetic model. This model explicitly incorporates the magnetic constraints of PMSMs, effectively preventing overfitting and enhancing inductance identification accuracy. A sensitivity analysis via an ablation study shows that the symmetric and mutual inductance constraints mainly suppress overfitting, while the magnetic saturation constraint mainly reduces model complexity. Based on this model, a high-frequency (HF) current injection method is developed for inductance identification. Crucially, this method eliminates the need for numerical differentiation of the magnetic model, thereby substantially improving the robustness of inductance identification against nonideal factors. Additionally, stator resistance and inverter nonlinearity are estimated simultaneously by solving a constrained nonlinear least-squares problem. Finally, the permanent magnet flux linkage is identified based on the steady-state voltage equation when rotor rotation is permitted. The effectiveness of the proposed physically feasible magnetic model and offline parameter identification method is experimentally validated on a 0.21-kW surface-mounted PMSM. Compared with conventional polynomial-based methods, it avoids overfitting and achieves higher inductance identification accuracy with lower cross-validation errors. Furthermore, the identified parameters improve the accuracy of MTPA control and remain robust under different thermal conditions.
Pulse wave propagation is intrinsically linked to both cardiovascular physiology and pathology. Although zero-dimensional (0D) and one-dimensional (1D) numerical models are widely used, they involve trade-offs among geometric fidelity, parameter uncertainty, and computational cost. Analytical approaches offer high computational efficiency but have traditionally been limited to uniform and elastic tubes, neglecting the critical effects of vessel tapering and wall viscoelasticity. In this study, a generalized frequency-domain analytical framework is proposed for geometrically heterogeneous arterial networks comprising both tapered and uniform vessels through a unified junction formulation. The framework is systematically validated against high-precision 1D numerical simulations using single-vessel and artery tree network geometries. The results demonstrate that tapered and uniform vessels play fundamentally different roles in arterial wave propagation. Geometric tapering introduces distributed wave reflections and substantially modifies input impedance, whereas wave propagation in uniform vessels is governed primarily by terminal impedance and blood viscosity. In complex arterial trees, proximal tapered arteries predominantly regulate central wave transmission, while distal uniform arteries mainly contribute to peripheral impedance and cumulative viscous dissipation. Under the conditions considered in this study, wall viscoelasticity primarily attenuates high-frequency waves but has little influence on the wave dynamics of uniform vessels. Furthermore, incorporating a resistance-based ( R 0 D ) correction effectively reduces errors in distal vessels. Overall, the proposed framework provides a computationally efficient platform for global wave propagation analysis and demonstrates particular utility for proximal pressure and impedance analysis. Following validation against patient-specific data, it may also provide a useful computational tool for future investigations of central aortic hemodynamics.
In this study, the hybrid AA6082 composites reinforced with boron carbide (B 4 C) and carbonized pistachio shell powder were successfully fabricated using the stir-casting technique. Microstructural characterization by EDS, XRD, and SEM confirmed that the reinforcement particles were successfully incorporated and uniformly distributed without any undesirable phase formation. The composite with 8 wt.% B 4 C and 2 wt.% carbonized pistachio shell powder exhibited the best overall performance with a maximum hardness of 92 BHN and tensile strength of 384 MPa, equal to an improvement of 26.02% and 28.85%, respectively, compared to the base alloy. The optimized composite showed wear loss between 4.3 and 9.6 × 10 −6 kg m −1 compared to 4.6 and 10.5 × 10 −6 kg m −1 for the unreinforced AA6082 alloy, which indicates better wear resistance. These results demonstrate that the synergistic combination of B 4 C and carbonized pistachio shell powder is an effective and sustainable approach for improving the mechanical and tribological performance of AA6082 composites.
Condition-number-based optimization of hybrid cable-driven parallel robots (HCDPRs) generally favors isotropic designs and therefore does not explicitly address task-dependent directional requirements. To overcome this limitation, this paper proposes a directional performance optimization method for HCDPRs based on directional manipulability and directional stiffness indices. In the present study, the optimization is restricted to the cable-driven subsystem, in which the cable anchor layout is taken as the design variable, while the manipulator is treated as a standard component and modeled as an external load. Accordingly, the proposed method should be understood as a direction-oriented geometric optimization of the CDPR subsystem rather than a holistic coordinated redesign of both HCDPR subsystems. A case study is carried out along a prescribed trajectory by comparing the proposed method with conventional condition-number optimization. The results show that the proposed method improves the target-direction performance, yielding about 20% enhancement in directional manipulability and about 25% enhancement in directional stiffness along the prescribed task direction. At the same time, the comparison also reveals measurable performance trade-offs in some non-target directions. These results indicate that the proposed method is particularly suitable for task scenarios in which the directional performance in prescribed directions is of primary concern.
This paper develops a phase field model for simulating crack propagation in hyperelastic solids subjected to coupled thermo-mechanical loading under finite deformation. The model integrates the Neo-Hookean constitutive law with a thermodynamically consistent framework that captures key multiphysics couplings: finite thermoelastic deformation, degradation of thermal conductivity with progressive damage, and heat convection across diffusely represented crack surfaces. A robust staggered solution algorithm is employed to solve the coupled system of displacement, temperature, and phase field equations. Numerical simulation of a notched tension specimen demonstrates the model’s capability to predict crack initiation and growth, while revealing the interdependent evolution of mechanical damage and thermal response. Results show that damage-induced thermal conductivity reduction leads to localized thermal gradients and “hot-spot” formation near the crack tip, while crack-surface convection significantly moderates the temperature field within the crack. Parametric studies elucidate the influence of thermal expansion and convection coefficients on the fracture process. This work provides a computational tool for failure analysis and design of rubber-like components in thermal environments.
In this paper, the dynamic response of a vibro-propelled capsule robot within a fluid-filled small intestine environment is studied, and the influences of various capsule parameters on different performance metrics are investigated. To achieve this, the dynamic model of the capsule robot moving along a flat inner surface of the small intestine is established and solved using the fourth-order Runge-Kutta method. Combining bifurcation theory, the motion states of the capsule robot under different parameter values are determined. The capsule’s parameter values determine its average speed, and by selecting appropriate parameter values, the forward and backward movement of the capsule robot can be controlled. Multistability behaviors are also been discovered and explored. To optimize the energy consumption of the capsule robot during operation, the energy efficiency under two sets of parameters is calculated. The genetic optimization algorithm successfully further optimized the performance of the capsule robot. This study provides theoretical guidance and data support for the parameter design of capsule robots in fluid-filled small intestine environments.
The impact of tread patterns on the road surface is one of the main sources of automobile tire noise. Reducing such impact noise has become an important challenge in the field of tire technology. To address this issue, a method based on contact area spectrum analysis is developed to optimize the tire pitch arrangement and reduce tire noise. First, digital matrices are generated by scanning the images of tread pitches and footprint profile. Then, the digital matrices are superimposed to extract the contact area during tire rolling, and its spectral signal is adopted to establish the evaluation index of tire noise. On this basis, to minimize the peak of the contact area spectrum, a hybrid optimization strategy combining genetic algorithm and simulated annealing algorithm is developed to optimize the pitch sequence. The optimization result based on the contact area spectrum analysis is further improved by adjusting the misalignment between the upper and lower molds of the tread blocks. Finally, tire noise analysis data is used to verify the evaluation results based on contact area spectrum analysis. The results show that, compared with the initial pitch arrangement, the optimized pitch arrangement achieves a 64.01% reduction in the peak of the contact area spectrum, a 67.38% decrease in the peak sound pressure of tire noise, a 9.73 dB reduction in the maximum sound pressure level, and a 0.1052 dB reduction in the overall sound pressure level. The proposed method can effectively evaluate the tread pattern noise and optimize pitch arrangement in the early design stages of tires.
To address the problems of low accuracy and slow speed in the recognition and positioning of composite material tee fittings in industrial environments, this paper proposes the YOLOv8-CCS algorithm. The algorithm designs a CGF module to enhance local and contextual feature extraction, introduces a CPS attention mechanism that decomposes 2D global pooling into 1D feature encoding to preserve positional information, and replaces CIoU with the SIoU loss function to accelerate convergence and improve localization accuracy. The 2D coordinates of the fitting center are determined using detection bounding boxes, and 3D coordinates are obtained by combining depth information from a depth camera, with the introduction of a lens distortion correction model. Experimental results show that YOLOv8-CCS achieves an accuracy of 95.5% and 125 FPS on a self-constructed dataset, outperforming mainstream algorithms such as YOLOv8n and Faster-RCNN, with additional comparisons against Transformer-based architectures. Ablation studies validate the effectiveness of each module and report standard deviations. Furthermore, the integration scheme and cost advantages of the vision system with robots and PLCs are analyzed, and the deep integration with the grinding robot’s motion control and force feedback system is explored, envisioning a “perception-decision-execution” integrated intelligent workstation. Photothermal material-based self-cleaning coatings are expected to maintain long-term stable imaging of cameras in dusty environments. In summary, the proposed algorithm meets the positioning accuracy and real-time requirements for automated grinding of composite material tee fittings.
Combined clearance, manufacturing errors, and misalignment make tooth engagement in involute spline couplings strongly non-uniform and induce pronounced nonlinear stiffness evolution. Existing slice-based formulations usually activate contact once the nominal approach becomes positive and evaluate the meshing geometry near the nominal pitch radius, which obscures the actual gap-closure sequence and weakens robustness near contact-state switching. This work develops an analytical stiffness framework with gap-dependent tooth engagement for involute spline couplings under combined errors. Tooth-dependent initial gaps are introduced to describe staggered contact onset, while the real meshing radius is used to update the working pressure angle, tooth compliance, and moment-arm distribution during deformation. A differentiable contact regularization is further adopted to stabilize Newton iterations in repeated equilibrium solves. Three-dimensional finite-element models generated from a parameterized structured-mesh workflow are used for verification. The analytical predictions agree well with the FE results for both lateral and angular stiffness, with mean relative deviations below about 2.2% for the representative cases considered. The results show that the tolerance-induced gap distribution governs the early-stage nonlinearity and load-sharing heterogeneity, whereas transmitted torque and engagement length increase the global stiffness level and misalignment softens the response. The proposed framework provides an efficient mechanics-based tool for evaluating stiffness evolution and error sensitivity in spline couplings.
Adhesive joints with drilled holes, which received considerable attention in the last few years, improve the failure load of adhesive joints by mechanical interlocking, offering great promise for the future to overcome the limitations of conventional non-drilled adhesive joints. In the first stage of this research, the influence of different hole configurations on the shear strength of PMMA/PMMA single-lap adhesive joints was examined. Moreover, the geometrical parameters of the drilled holes were measured by an optical microscope and profilometry analysis. The experimental results showed that the performance of all adhesive joints with drilled holes was improved significantly relative to non-drilled adhesive joints. Of these, Configuration 4 produced the maximum load of 1308.42 N, reflecting an improvement of 271.62%. In addition, FEM analysis was carried out, showing excellent agreement with the experimental results. In the second stage, Analysis of Variance (ANOVA) was used to examine the most effective parameters for the designs. The ANOVA showed that the hole configuration was the most effective parameter for designs. The results obtained in this research clearly show that the geometrical parameters of the holes play essential role in the mechanical performance of adhesive joints, offering an essential basis for future designs to be optimized.
