This paper focuses on the output-constrained tracking control problem of magnetic drive transmission systems subject to modeling uncertainties. Specifically, a tracking error-based time-varying transformation function is introduced to convert the constrained system into an unconstrained framework. And radial basis function-based neural networks (RBFNN) will be employed to approximate the unknown nonlinear dynamics. Meanwhile, the extended state observer will be incorporated to estimate and compensate for external disturbances. The simulation results demonstrate the effectiveness of the proposed neuroadaptive learning algorithm in the presence of uncertainties.

To address the limited diagnostic efficiency caused by fault data scarcity in industrial scenarios, this paper proposes a fault dataset expansion method based on a Feature-Enhanced Cycle-Consistent Generative Adversarial Network (FE-CycleGAN). Using bearings as the experimental subject, the study utilizes the Case Western Reserve University (CWRU) dataset, which originates from real physical experiments and provides standardized samples of inner race, rolling element, and outer race faults. By constructing an enhanced generator architecture, adaptive training strategies, and a comprehensive loss function system, normal state data is enabled to learn the characteristics of fault states to achieve cross-domain data generation. Finally, fault diagnosis experiments are conducted using a lightweight two-dimensional Convolutional Neural Network (2D-CNN). Experimental results demonstrate that when using fault datasets generated by training the FE-CycleGAN model with 10%, 50%, and 100% of the original fault data, the final diagnostic accuracy improves from a baseline of 86.97% (using only original data) to 94.34%, 98.04%, and 100%, respectively. Furthermore, the time-frequency domain feature retention rate of the generated data exceeds 85%, and the training time is maintained within the highly efficient range of 10.2 to 11.2 seconds. This method provides an efficient and reliable technical solution for industrial intelligent operation and maintenance under small-sample conditions.

In practical applications of mechanical fault diagnosis, vibration data typically exhibit the characteristics of small sample sizes and class imbalance. The minority class often fails to cover diverse operating conditions and noise perturbations, which leads to an unstable decision boundary of learning models for the minority class. To this end, this paper proposes an interpretable data augmentation method for vibration signals. First, the original vibration segments are mapped into low-dimensional, readable, and parameterized intermediate variables, where class discrepancies and noise characteristics are characterized in a feature space with clear statistical meaning, thereby constructing an 'Interpretable Space'. Subsequently, within this space, a minimalist generative backbone based on the JiT architecture is introduced for conditional sampling, producing synthetic samples of the interpretable variables. These variables are then translated back into time-domain simulated signals via an interpretable decoder to supplement the minority-class samples, enabling controllable augmentation for imbalanced datasets and parameter-level traceable interpretation. The proposed method is validated on three representative bearing datasets with different scales, and achieves a significant performance improvement compared with conventional data augmentation methods.

The application of deep learning in constructing data-driven remaining useful life prediction models through historical bearing degradation data has demonstrated significant potential. However, accurate prognostics for extra-large-scale bearings remain constrained by limited operational lifespan data. While extensive degradation datasets exist for standard-sized bearings, inherent mechanistic disparities between different bearing scales create cross-domain transfer challenges. To address this limitation, this study proposes an innovative dual-model fusion framework that synergizes small-bearing full life-cycle data with mechanical principles for extra-large-scale bearing remaining useful life prediction. Our methodology comprises three core innovations: Development of an attention mechanism-enhanced bidirectional gated recurrent unit network integrated with transfer learning; Construction of a physics-informed degradation model based on ISO281 standards; and a novel threshold continuous triggering algorithm for precise degradation phase segmentation. The framework implements a progressive model updating strategy through coordinated utilization of cross-scale bearing data at different degradation stages, establishing an adaptive "data + mechanism" dual-model fusion prognostic system. Experimental validation confirms significant enhancement in prediction accuracy through iterative updating, ultimately achieving reliable RUL estimation for extra-large-scale bearings.

The health status of mechanical equipment bearings is highly variable under complex environment, with fault progression often accompanied by substantial noise interference. Additionally, the timing of fault occurrence is unpredictable, significantly compromising the accuracy of fault diagnosis. Therefore, this article proposes a two-distribution approximate evaluation multiwavelet cascade feature method for addressing the problem of noise-unbalanced bearing fault diagnosis, and multiple wavelet filters to adapt to the filtering requirements of various faults are introduced in this method. Additionally, network branches are cascaded to enhance the characterization capability of network. The two-distribution approximate evaluation method is employed to assign differential weights to the cascaded network features, thereby achieving high-accuracy fault diagnosis. Specifically, measuring the feature approximation degree of the branch cascade and dynamically selecting the best result of the cascade before the decision level is conducive to accurately locating the differential expression of the multiwavelet kernel function, which expands the expected feature information during the network training process and can effectively improve the fault diagnosis performance of the network. The experimental platform signal is used for performance testing and fault diagnosis of rotary bearing in real operation, which can accurately identify the signals of mechanical equipment loading, normal operation, performance decline, and other stages. The experimental diagnostic results from both the public datasets and the slewing bearing working datasets concurrently validate the reliability of the diagnostic method.

3D free-bending technology is an innovative technology in tube forming that has attracted significant attention owing to the characteristics of dieless forming and the diverse shapes of the product. The arc section of the tube was separated into controlled and uncontrolled areas based on free-bending technology. During the free-bending process, the moving mode of the bearings in controlled areas plays a significant role in the quality and accuracy of the tube. In this study, the movement of the bearings in the controlled areas was categorized into three types: uniformly accelerated, uniform, and uniformly decelerated. To study the deformation behavior of the tube under different bearings moving modes, a finite element model was developed and validated based on experimental results. Subsequently, the effect of each bearings moving mode on the deformation section was investigated based on FEM simulation. The results indicate that the ovality of the tube is minimal when the bearings uniformly decelerate in the area before the bending zone and are uniform in the area after the bending zone, compared with other modes. Furthermore, two moving modes were proposed, and bending experiments were conducted. A new moving mode, T1, was presented. The bearings movement was set as uniformly decelerated in the area before the bending zone and considered uniform in the area after the bending zone. The ovality was reduced by 1.59% compared with the conventional moving mode. Another moving mode, T3, was proposed: the bearings movement decelerated uniformly in the area before the bending zone and uniformly accelerated in the area after the bending zone. The deviation in the bending angle between the design and T3 was 1.75%, which is an effective way to achieve the design bending angle.

In the area of bearing fault diagnosis, deep learning (DL) methods have been widely used recently. However, due to the high cost or privacy concerns, high-quality labeled data are scarce in real world scenarios. While few-shot learning has shown promise in addressing data scarcity, existing methods still face significant limitations in this domain. Traditional data augmentation techniques often suffer from mode collapse and generate low-quality samples that fail to capture the diversity of bearing fault patterns. Moreover, conventional convolutional neural networks (CNNs) with local receptive fields makes them inadequate for extracting global features from complex vibration signals. Additionally, existing methods fail to model the intricate relationships between limited training samples. To solve these problems, we propose an advanced data augmentation and contrastive fourier convolution framework (DAC-FCF) for bearing fault diagnosis under limited data. Firstly, a novel conditional consistent latent representation and reconstruction generative adversarial network (CCLR-GAN) is proposed to generate more diverse data. Secondly, a contrastive learning based joint optimization mechanism is utilized to better model the relations between the available training data. Finally, we propose a 1D fourier convolution neural network (1D-FCNN) to achieve a global-aware of the input data. Experiments demonstrate that DAC-FCF achieves significant improvements, outperforming baselines by up to 32% on case western reserve university (CWRU) dataset and 10% on a self-collected test bench. Extensive ablation experiments prove the effectiveness of the proposed components. Thus, the proposed DAC-FCF offers a promising solution for bearing fault diagnosis under limited data.

The finite element simulation is an effective way for tube formability research based on free bending technology. To resolve the controversial issues regarding key parameters in the finite element model (FEM) establishment process and to significantly improve the simulation accuracy of tube-free bending, the study focuses on selecting and unifying key parameters to balance accuracy and efficiency in model development. An anisotropic constitutive model for the metal tube was established based on the Hill48 yield criterion, and its validity was verified by experiments. Then, a whole-tube model and a half-tube model were developed based on the anisotropic constitutive model. The results showed that the half-tube model demonstrated a 51.4

Cross-section distortion and axial instability are inherent challenges in the tube bending process. This paper investigates the use of a floating ball within the tube to mitigate ovality. Initially, the tube bending process with the floating ball is analyzed to elucidate the floating ball’s role during tube bending. Subsequently, a finite element model without a floating ball was developed and validated through simulations. Finally, the influence of the floating ball’s position and the gap between the floating ball-tube on the formability of the tube was analyzed by using finite element analysis (FEA). Optimized parameters for the floating ball were identified, enhancing tube bending quality and offering practical guidelines for implementing the floating ball in tube free bending.

The fault signals of rotating machinery exhibit complex characteristics, including nonlinearity, high noise, and dynamic changes, etc. These issues pose significant challenges to fault diagnosis. Traditional multimodal fusion technologies often fail to fully capture the unique features of different data modalities, resulting in low diagnostic accuracy and robustness under complex working conditions. This paper proposes a FuseCT multimodal feature fusion network, which integrates one-dimensional vibration signals and two-dimensional time-frequency images obtained via Generalized Linear Chirplet Transform (GLCT), and adopts a two-branch structure for processing. The network extracts vibration signal features through the CNN-BiLSTM-1DCBAM module, utilizes the optimized SpectraFocus module to extract GLCT time-frequency features, and finally completes multimodal feature fusion via the self-attention mechanism. This method has been experimentally validated on the gearbox dataset from Southeast University and the extra-large bearing dataset from Nanjing Tech University. The results demonstrate that FuseCT achieves higher accuracy and better generalization ability in rotating machinery fault diagnosis, outperforming traditional single-modal and multi-modal methods.

Traditional physical parameter identification methods struggle to analysis the modern multi-degree-of-freedom time-varying (TV) structures. Aiming at this problem, an adaptive reconfiguration algorithm based on polynomial chirplet transform (PCT) is proposed. First, the wavelet basis function of PCT is extracted, and its one-level and tow-level derivative basis functions can be obtained through integral operation. Second, the optimized PCT is applied to the acceleration signal, enabling reconstruction of velocity and displacement signals while establishing their conversion rules. Finally, the vibration differential equation is reformulated, allowing parameter identification. Unlike traditional identification methods, PCT can better track the TV parameters due to the introduction of the frequency modulation slope parameter. Numerical simulations on a three-degree-of-freedom TV structure under varying conditions demonstrate the superior identification accuracy of the proposed method.

This article focuses on the motion control issue for uncertain robotic manipulators with unknown nonlinear dynamics and exogenous disturbances simultaneously via the output position measurement only. In detail, the extended state observer is employed to obtain the estimations of immeasurable system state information and lumped disturbances simultaneously. Meanwhile, the multilayer neural network with good approximation performance is incorporated into the observer-controller scheme to identify the unknown nonlinear dynamics. As a result, both large nonlinear dynamics and strong lumped disturbances can be compensated feedforwardly. Significantly, prescribed tracking performance and asymmetric time-varying state constraints can be realized simultaneously.

Outsourcing machine maintenance to third parties has been a trend due to the increasing complex of machines and the benefits of maintenance outsourcing. However, this new phenomenon is ignored in previous studies pertaining to the integrated optimization of production and maintenance scheduling (IOPMS) problems, resulting the lack of theoretical guidance for production managers to formulate optimal scheduling schemes under this new trend. In this study, we investigate an IOPMS problem considering maintenance outsourcing in a distributed parallel machine environment, referred as IOPMSTW, which requires the use of third-party worker resources to perform preventive maintenance. Makespan and total cost are two optimization objectives. We first formulate the problem by developing a mixed integer linear programming. Then a memetic algorithm incorporating iterated greedy method (IG) is proposed to solve the IOPMSTW, in which an improved decoding method and a problem-dependent local search operator based on IG are designed to respectively ensure the feasibility of new generated individuals and improve the searching efficiency. The validity of proposed mathematical model is verified by CPLEX based on eight small instances. Based on 240 constructed instances, a set of comprehensive experiments are conducted. The results demonstrate that the local search operator improved the searching performance of the proposed algorithm by 100%. Comparison results with other well-known algorithms show that the proposed algorithm achieved the best results on more than 85% of the tested instances.

The intelligent transfer diagnosis model is used to address the issue of feature drift caused by the changing working conditions of rotating parts in engineering. However, few models can perform transfer diagnosis on multiple unbalanced samples of rotating parts simultaneously, and even fewer models can visually enhance the domain-invariant features, making them more interpretable. To address these issues, we propose a novel interpretable Siamese dual attention enhancement transfer compound diagnosis model for unbalanced samples. The model can diagnose multiple rotating parts simultaneously and consists of a channel feature attention enhancement (CFAE) network, a fragment feature attention enhancement (FFAE) network, and a Siamese feature fusion (SFF) network. The CFAE network enhances features of different convolutional channels, the FFAE network improves segment features in various frequency domains, and the SFF network extracts domain-invariant features of diverse rotating components under varying working conditions. The model is validated using bearing fault data collected under different loads and planetary gear fault data obtained at varying speeds. Its diagnostic accuracy remains above 96.4%, and the diagnostic variance is controlled within 1.0%. The model has good interpretability for imbalanced sample domain-invariant features, providing an effective tool for interpretable transfer diagnosis in this compound engineering situation.

Slewing bearing is a critical transmission component in large-size construction machinery due to its low-speed and heavy-load conditions. Fault prognostics and health management of slewing bearing are crucial for ensuring their high availability and profitable operation. However, the presence of background noise in construction machinery signals restricts the applicability of existing signal processing approaches in prognostics and health management. To address this challenge, a novel signal de-noising method is proposed based on adaptive decomposition, along with a new strategy for recognizing fault components using statistic detection through kernel principal component analysis (KPCA). First, robust local mean decomposition is utilized to adaptively decompose the fault and normal vibration signal over the entire service life. Then, product functions (PFs) decomposed by fault and normal vibration signal are used for KPCA anomaly detection. Finally, the fault PFs are reconstructed to obtain the de-noised signal. The effectiveness of the proposed method is validated through the use of both simulated and experimental vibration signals obtained from a slewing-bearing life-cycle test. The results illustrate that the proposed method has superior de-noising capability and decomposition efficiency, making it an effective signal preprocessing technique for prognostics and health management.

For servo electromechanical systems, the existing modeling uncertainties, signal measurement noises and so on always make it difficult to design high-performance closed-loop controllers. In this paper, a novel intelligent controller will be designed to deal with these uncertainties. Specifically, two multilayer neuroadaptive disturbance observers will be proposed to estimate the uncertain nonlinear dynamics and exogenous disturbances simultaneously. And these uncertainties will be compensated feedforwardly. Moreover, in order to reduce the influence of signal measurement noises, the desired-command compensation technique will be incorporated. Additionally, different validation examples will be proposed to demonstrate the advantages of the designed controller.

The pulsation frequency of inclined jet fire was systematically studied in this study. Ex-periments of syngas jet fire with a nozzle diameter of 15 mm at different fuel flow rates (2.5 L/min-20 L/min) and inclination angles (0 degrees-90 degrees) were conducted. The pulsation fre-quency of inclined jet flame is quantified by applying FFT to the time variations of image correlation coefficient, flame height and flame width, corresponding to the global, vertical and spatial pulsation frequencies, respectively. Experimental results indicate that the vertical pulsation frequency can be derived only for jet flames with significant clip-off. The effect of inclination angle on the pulsation frequency depends on the fuel flow rate. For Qf < 7:5 L/min, the natural frequency is dominant and increases slightly at small inclina-tion angles. For larger fuel flow rates, a transition from subharmonic frequency to natural frequency occurs at a critical inclination angle qcri, which is sensitive to fuel flow rate. The pulsation behaviors of inclined jet flame are driven by three different instability modes, R-T instability at small fuel flow rates, and Extended R-T and K-H instabilities at larger fuel flow rates. The increase of global pulsation frequency with an increasing inclination angle for Qf >= 10 L/min results from the increased natural frequency at the nozzle exit. By defining a characteristic diameter D*, a dimensionless model is developed to predict the global pul-sation frequency of inclined jet fires with large fuel flow rates, St = 0.284(1/Fr)0.473 for natural frequency and St = 0.516(1/Fr)0.473 for subharmonic frequency.(c) 2022 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.

Slewing bearing is one of critical transmission in wind turbine and shield machine withstanding low-speed and heavy-load working condition. Fault recognition is crucial to their high reliability operation. Many studies have been conducted using traditional shallow networks for fault recognition. However, they suffer from inherent disadvantages, such as low learning ability under high-dimensional nonlinear features, which make them unsuitable for fault recognition of slewing bearing. To solve these shortcomings, a novel fault recognition method is proposed based on improved deep belief network (DBN) using sampling method of free energy in persistent contrastive divergence (FEPCD). A systematic methodology based on multi-domain feature extraction is proposed to describe the fault characteristic information. After that, improved DBN optimized by FEPCD is employed to capture the fault features and recognize the fault condition of slewing bearing. The application and superiority of proposed methodology are validated using a slewing bearing life-cycle test dataset. Meanwhile, a comparison is conducted between traditional sampling methods contrastive divergence (CD) and persistent contrastive divergence (PCD). The results illustrate that improved FEPCD gets better result in training sampling. Compared with other deep learning methods such as deep Boltzmann machine (DBM) and stacked auto-encoder (SAE), and shallow intelligent algorithms like back propagation (BP) neural network and support vector machine (SVM), the fault recognition accuracy of slewing bearing is improved by using the improved DBN with FEPCD.

In order to improve the processing accuracy of free bending forming technology for tubes, the precise prediction work was conducted on the forming parameters for the processing accuracy of planar bending tubes, and the experimental values of bending radius and bending angle were consistent with the designed values by establishing the prediction model of forming parameters. Firstly, the finite element simulation model was established and modified by tube processing experiments, and the optimized simulation model was used to establish the predicted sample database. Then, taking the bending radius and bending angle obtained by finite element simulation as input, and the designed values of bending radius and bending angle as output, combined with BP neural network and grey wolf optimizer algorithm, the forming parameter prediction model was built. The results show that the improved PGWO-BP neural network predicts the bending radius and bending angle with the maximum error of no more than 2%. At the same time, the prediction model is used to develop the process parameter determination software of tube precision forming.