Recently, graph convolution network (GCN) has been the focus in fault diagnosis for its powerful representational ability in relationship mining. However, with the difficulty in extracting the weak features of the signal under variable load conditions, GCN is not suitable for deep neural network (DNN), and the receptive scale of GCN is unknown that limits the application of GCN in machine fault diagnosis. To address these issues, a multi-scale cluster-graph convolution neural network with multi-channel residual network (MR-MCGCN) is proposed for machine fault diagnosis in this article. First, multi-channel residual network (MCRN) is proposed for extracting the weak feature in the signal. Then, the finite graph data of signal and different scales are generated by the autoencoder (AE) graph generation layer. Finally, a multi-scale cluster-graph convolution neural network is proposed for achieving intelligent fault diagnosis. Also, the three different datasets are used for verifying the effectiveness of the proposed MR-MCGCN. The experimental results show that the proposed MR-MCGNN can achieve the highest diagnosis results than other methods even under variable load conditions.

For cross-domain fault diagnosis of rotating machinery, how to reduce the discrepancy between the source and target data distributions is still a key problem. To this end, this study proposes a novel cross-domain fault diagnosis method based on improved multi-scale fuzzy measure entropy and enhanced joint distribution adaptation, aiming to address inconsistent data distribution between the source and target domains. Specifically, improved multi-scale fuzzy measure entropies are firstly developed to generate discriminative and similar features from original vibration signals. Subsequently, through embedding maximum covariance discrepancy into the existing joint distribution adaptation, an enhanced joint distribution adaptation model is utilized to reduce the distribution discrepancy between two domains. As such, more discriminative and similar features are obtained in a new subspace. The unlabeled samples in the target domain are classified by a simple statistical classifier. Finally, three public and private datasets are used to verify the effectiveness and superiority of the proposed method. Experimental results demonstrate that improved multi-scale fuzzy measure entropies have better distinguishing ability and transferable ability than several existing entropy methods, and the enhanced joint distribution adaptation is more generalized to transfer scenarios with complex data distributions.

The hydropower units have a complex structure, complicated and changing working conditions, complexity and a diversity of faults. Effectively evaluating the healthy operation status and accurately predicting the failure for the hydropower units using the real-time monitoring data is still a difficult problem. To this end, this paper proposes a prediction method for the early failure of hydropower units based on Gaussian process regression (GPR). Firstly, by studying the correlation between different monitoring data, nine state parameters closely related to the operation of hydropower units are mined from the massive data. Secondly, a health evaluation model is established based on GPR using the historical multi-dimensional monitoring information and fault-free monitoring data at the initial stage of unit operation. Finally, a condition monitoring directive based on the Mahalanobis distance (MD) is designed. The effectiveness of the proposed method is verified by several typical examples of monitoring data of a hydropower station in Guangxi, China. The results show that, in three cases, the abnormal conditions of the unit are found 2 days, 4 days and 43 days earlier than those of regular maintenances respectively. Therefore, the method can effectively track the change process of the operation state of hydropower units, and detect the abnormal operation state of hydropower units in advance.

In most engineering scenarios, the collected datasets typically have small and imbalanced characteristics, and usually possess different data distributions between variable working conditions. Thus, several existing diagnosis models trained by small and imbalanced samples under one working condition may suffer from low accuracy for fault data under another working condition. To effectively alleviate such problem, a novel domain adaptation method, referred to as generalized transfer extreme learning machine (G-TELM), is proposed in this study. First, in a random feature space, marginal and conditional distribution discrepancy between source and target domains is minimized based on projected maximum mean discrepancy. Second, in order to give varied attention to different faulty samples, a cost-sensitive source domain classifier is utilized to reduce the difference between types of misclassification loss for small and imbalanced source data. Third, misclassification loss of target data is used to minimize the difference between the network’s outputs and prediction labels of target data. Last, L2-norm regularization is leveraged to constrain network output weights, and they are iteratively updated to optimize the transferability of model G-TELM. Experimental results of 168 transfer tasks with varying difficulty verify that the proposed model G-TELM can achieve an accurate diagnosis and outperform several state-of-art domain adaptation methods.

An ultrasonic coupling agent, as an acoustic medium between the ultrasonic probe and the surface of the specimens, is indispensable in Nondestructive Testing (NDT). Whether it is liquid, air, or solid coupling agent, the problem of improving the efficiency of ultrasonic propagation in a coupling agent is one worth studying. Glycerol and hydrogels are two common liquid coupling agents in NDT. This study intended to investigate the effect of graphene addition on the performance of these coupling agents in NDT. Firstly, based on the theory of acoustic impedance matching, the authors established an index system to evaluate the performance of ultrasonic coupling agent by experiments. Secondly, hydrogel–graphene and glycerol–graphene composite coupling agents were prepared by adding three-dimensional graphene structure powders with mass fraction of 0.25%, 0.5%, 0.75%, and 1% to CG-98 hydrogel coupling agent and HG-99 glycerol coupling agent, respectively. Corresponding experiments were conducted on these composite coupling agents. Peak-to-peak value, attenuation coefficient, and energy value of first echo are calculated at different frequencies. The experimental results showed that graphene can significantly improve the ultrasonic propagation performance of hydrogel and glycerin coupling agents. In addition, when the mass fraction of graphene added was 0.75%, the coupling agent had the best performance. Finally, we measured the acoustic impedance values of the composite couplings with different graphene contents to demonstrate the reliability of the experimental results.

The early faults of rolling bearings are the common causes of rotating machinery failures. Rolling bearings with local faults usually generate periodic shocks during operation, but the pulse information is easily masked by a large number of random shocks and noise. To effectively diagnose the early fault information of rolling bearings, a dual-dimensional sparse low-rank (DDSLR) model is proposed in this paper, which can simultaneously extract the sparsity within and across groups and periodic self-similarity of fault signal. In the DDSLR model, a newly developed dimension transformation operator is used to transform the fault signal between one-dimensional vector and low-rank matrix, and the periodic overlapping group shrinkage and truncated nuclear norm are used to improve the traditional sparse low-rank model. In addition, the setting rules of periodic prior and parameters in the DDSLR model are discussed, so that the DDSLR model has certain adaptive ability. Finally, the DDSLR model is proved to be a multi-convex optimization problem, and its solution algorithm is derived by using soft threshold operator and majorization-minimization algorithm under the framework of block coordinate descent method. The results of simulation analysis and experiments show that the proposed DDSLR model has higher fault signal estimation accuracy and better fault feature extraction performance than some classical sparse noise reduction models.