Wei Sun

Addressing the critical challenges of limited labeled data and operational variability in bearing fault diagnosis, a Multiscale Recalibration Attention Network (MRAN) integrated with adaptive grouped convolution is proposed in this paper. Firstly, a statistical-aware recalibration module that dynamically fuses mean and standard deviation information to suppress noise and amplify discriminative features under data scarcity. Secondly, a kernel-adaptive grouped convolution mechanism that intelligently assigns computational resources based on kernel size to optimize efficiency; Thirdly, a hierarchical fine-tuning strategy that selectively adapts deep layers to resolve cross-domain feature misalignment. Extensive experiments on BJTU-RAO and cross-device datasets demonstrate the superior performance of MRAN: it achieves an accuracy of 98.67% without transfer learning, which is significantly better than the state-of-the-art CNNs and Vision Transformers. Crucially, using only 30% of the target domain data, MRAN has an accuracy of 94.57% in cross-conditional tasks and an accuracy of 92% in demanding cross-device diagnostics, while greatly reducing training time and computational burden. This work establishes MRAN as a groundbreaking, efficient, and highly accurate solution for practical industrial bearing fault diagnosis under data and operational constraints.

The main bearing of a tunnel boring machine (TBM) is a critical component of the main driving system that enables continuous excavation, and its performance is crucial for ensuring the safe operation of the TBM. Currently, there are few testing technologies for TBM main bearings, and a comprehensive testing and evaluation system has yet to be established. This study presents an experimental investigation using a self-developed, full-scale TBM main bearing test bench. Based on a representative load spectrum, both operational condition tests and life cycle tests are conducted alternately, during which the signals of the main bearing are collected. The observed vibration signals are weak, with significant vibration attenuation occurring in the large structural components. Compared with the test bearing, which reaches a vibration amplitude of 10 g in scale tests, the difference is several orders of magnitude smaller. To effectively utilize the selected evaluation indicators, the entropy weight method is employed to assign weights to the indicators, and a comprehensive analysis is conducted using grey relational analysis. This strategy results in the development of a comprehensive evaluation method based on entropy weighting and grey relational analysis. The main bearing performance is evaluated under various working conditions and the same working conditions in different time periods. The results show that the greater the bearing load, the lower the comprehensive evaluation coefficient of bearing performance. A multistage evaluation method is adopted to evaluate the performance and condition of the main bearing across multiple working scenarios. With the increase of the test duration, the bearing performance exhibits gradual degradation, aligning with the expected outcomes. The findings demonstrate that the proposed performance evaluation method can effectively and accurately evaluate the performance of TBM main bearings, providing theoretical and technical support for the safe operation of TBMs.

For the guide vane adjusting mechanism, precision represents the primary design requirement. Meanwhile, due to the presence of aerodynamic loads under actual operating conditions, stagnation forces emerge that affect the mechanism motion characteristics, including the response speed and precision. This paper establishes kinematic and static analysis models of the guide vane adjusting mechanism through analytical modeling methods, investigates analytical approaches for mechanism adjustment precision and stagnation force, and conducts error and sensitivity analyses of the mechanism parameters based on these analytical models. Building upon this foundation, an optimization design method integrating adjustment precision and force transmission performance is proposed using a multi-objective genetic algorithm. Optimizing the critical design parameters, such as the mechanism dimensions and positions, can enhance both the adjustment precision and force transmission performance. Through case studies, significant reductions in motion precision errors and the peak stagnation force and maximum differences in stagnation force were achieved, validating the feasibility of this optimization design approach.

The development of digital twins and smart bearings in bearing industry demands efficient computational methods for factors such as strain analysis. Symbolic regression offers a promising solution by efficiently deriving analytical expressions from limited data, offering low computational cost, rapid speed, high accuracy, and strong interpretability. In this paper, symbolic regression is applied to the strain calculation problem of radially loaded hollow cylindrical rollers. The resulting expressions are evaluated and selected based on physical insights obtained from theoretical analysis and physical experiments. Compared to other methods, the selected expression retains interpretability and generalization capabilities while achieving superior computational accuracy.

As one of the key components of an aeroengine, turbine blisk endures complex coupling loads under a harsh operational environment so that the reliability of turbine blisk directly influences the safe operation of aeroengine. It is urgent to precisely perform the reliability estimation of a complex blisk structure. To address this issue, an enhanced Moving Neural Network Framework (MNNF) is proposed by integrating compact support region theory, improve sooty tern optimization algorithm (ISTOA), and Bayesian regularization strategy into artificial neural network. The compact support region theory is applied to select the efficient samples for modeling from the training samples set, the ISTOA is to determine the optimal compact support region, and Bayesian regularization thought is utilized to improve the generalization ability of neural network model. The operational reliability assessment of aeroengine blisk is performed with the consideration of transient loads to verify the proposed MNNF method. It is shown that the reliability degree of turbine blisk stain is 0.9984 when the allowable value is 5.2862 × 10⁻³ m. In line with the comparison of methods, the developed MNNF approach has 0.99738 in root means square error, 3.1634 × 10⁻⁴ m in goodness of fit, 0.423 s in modeling time, 99.99% in simulation precision, and 0.496 s in simulation time under 10,000 simulations, which are superior to all other methods (i.e., 99.96%, 99.91%, 99.93%, 99.97%, and 99.97% in simulation precision and 16.27%, 4.82%, 30.07%, 39.87%, and 23.59% in simulation efficiency, for the response surface method (RSM), Kriging, support vector machine (SVM), back propagation-artificial neural network (BP-NN), and BP-NN based on particle swarm optimization (BP-PSO) methods, respectively). It is demonstrated that the MNNF method holds excellent modeling and simulation performances. The efforts of this study provide promising tools and insights into the reliability design of complex structures, and enrich and develop reliability theory.

Sensors are essential for the prognosis and health management of equipment. The fusion of multiple sensors data could improve the accuracy of fault diagnosis. Current data fusion methods focus on high-quality sensor data. However, in practical applications, constraints such as cost or space may limit the use of multiple high-quality sensors. Therefore, it is necessary to explore how to effectively use all available sensor data, even those containing some useful information but less effective when used individually. This paper presents a novel method for sensor data fusion, utilizing the highest-performing sensor data as a back bone and complementing it with preliminary data-level fusion data through multi-stage adaptive feature complementation. Comparative results validate the effectiveness and superiority of this method. The proposed method also demonstrates strong predictive capability by using data from a single period as training samples to forecast future data states.

An optimal design method for the main bearing of a tunnel boring machine is proposed. In this method, the fatigue life is used as the objective function. Structural parameters, including clearance, are considered as design variables. First, a quasi-static model of the main bearing and a calculation model of the fatigue life are established. The correctness of the theoretical method is verified by comparing it with the calculation results of the finite element method. Next, the influence of clearance on the load-carrying performance under external loads is analyzed. There is an optimal negative clearance for the axial loaded and radial rows. With the increase in the external loads, the optimal negative clearance gradually decreases. The variation laws of the load-carrying performance for the axial loaded and supporting rows affected by axial clearance mainly depend on the bias load degree of the main bearing. Finally, based on the optimal design model of the main bearing, the optimal internal structure is obtained using the genetic algorithm. The optimized fatigue life is improved by 92.2%. The load-carrying performance of the optimal main bearing has also been significantly enhanced compared to the initial design. Therefore, the proposed optimization method provides a practical approach to the main bearing design.

Bearings are vital components of a rotary machine. Real-time fault diagnosis of bearings has great significance in the maintenance of equipment. Current vibration-based fault diagnosis methods rely on the usage of long time series data to reduce the influence of noise but often suffer from imbalances in datasets, significantly affecting diagnosis accuracy. In this paper, a fault diagnosis method based on Intertemporal Return Plot and data augmentation is proposed. The Intertemporal Return Plot is employed to transform one-dimensional time series data into two-dimensional images. The Wasserstein Generative Adversarial Network is employed to generate synthetic images for data argumentation. This approach helps reduce the influence of data imbalance, thereby improving the accuracy and convergence speed of the fault diagnosis. Comparisons with existing well-known methods are performed to demonstrate the effectiveness of the proposed method. The results show that the proposed method achieves a high accuracy of fault diagnosis for short time series data. An inverse transformation is also proposed to convert the images to time series data.

As the different distribution between source and target domain, rolling bearing fault diagnosis based on domain adaptation has achieved a good result in unlabeled sample fault diagnosis of target domain. However, most transfer strategies focus mainly on the domain invariant feature between the source and target domains, with little attention paid to the intrinsic feature of the target domain. This limitation causes poor knowledge transfer from the source domain due to domain confusion. Therefore, this paper propose a novel method for unlabeled rolling bearing fault diagnosis in the target domain that leverages both the intrinsic features of the target domain and the shared label space between the two domains. Firstly, a new deep clustering network is constructed to extract the intrinsic features of target domain, where these features are clustered several central areas based on their same attributive feature. Secondly, the novel adversarial domain adaptation network is designed to extract the invariant features in order to construct the sharing label space between source and target domain. Thirdly, the new pseudo label technology is introduced to mark the target unlabeled data by fusing the invariant feature of target domain with the sharing label space, and eliminate the negative knowledge transfer. Finally, these marked labels combining the intrinsic feature of target domain are used to train the fault diagnosis classifier, where this realizes a well-prediction model for the target domain. Several transfer tasks on two rolling bearing datasets demonstrate the efficiency of proposed method. The proposed transfer technology realizes the accuracy prediction of rolling bearing state on target unlabeled samples. Moreover, the efficiency of method is better than the contrasting method, and this new diagnostic technology offers a promising fault diagnosis tool under some practical industrial problem.