The success of intelligent fault diagnosis of machines relies on two conditions. 1) Labeled data with fault information are available. 2) The training and testing data are drawn from the same probability distribution. However, for some machines, it is difficult to obtain massive labeled data. Moreover, even though labeled data can be obtained from some machines, the intelligent fault diagnosis method trained with such labeled data possibly fails in classifying unlabeled data acquired from the other machines due to data distribution discrepancy. These problems limit the successful applications of intelligent fault diagnosis of machines with unlabeled data. As a potential tool, transfer learning adapts a model trained in a source domain to its application in a target domain. Based on transfer learning, we propose a new intelligent method named deep convolutional transfer learning network (DCTLN). DCTLN consists of two modules: condition recognition and domain adaptation. The condition recognition module is constructed by a one-dimensional convolutional neural network (CNN) to automatically learn features and recognize health conditions of machines. The domain adaptation module facilitates the one-dimensional CNN to learn domain invariant features by maximizing domain recognition errors and minimizing probability distribution distance. The effectiveness of the proposed method is verified using six transfer fault diagnosis experiments.

Deep transfer-learning-based diagnosis models are promising to apply diagnosis knowledge across related machines, but from which the collected data follow different distribution. To reduce the distribution discrepancy, Gaussian kernel induced maximum mean discrepancy (GK-MMD) is a widely used distance metric to impose constraints on the training of diagnosis models. However, the models using GK-MMD have three weaknesses: 1) GK-MMD may not accurately estimate distribution discrepancy because it ignores the high-order moment distances of data; 2) the time complexity of GK-MMD is high to require much computation cost; 3) the transfer performance of GK-MMD-based diagnosis models is sensitive to the selected kernel parameters. In order to overcome the weaknesses, a distance metric named polynomial kernel induced MMD (PK-MMD) is proposed in this paper. Combined with PK-MMD, a diagnosis model is constructed to reuse diagnosis knowledge from one machine to the other. The proposed methods are verified by two transfer learning cases, in which the health states of locomotive bearings are identified with the help of data respectively from motor bearings and gearbox bearings in laboratories. The results show that PK-MMD enables to improve the inefficient computation of GK-MMD, and the PK-MMD-based diagnosis model presents better transfer results than other methods.

To integrate the complete degradation information of machinery, deep learning-based prognostics approaches usually use monitoring data acquired by different sensors as the inputs of networks. These approaches, however, lack an explicit learning mechanism to effectively identify the distinctions of different sensor data and highlight the important degradation information, thereby affecting the accuracy of deep prognostics networks and limiting their generalization. To overcome the aforementioned weaknesses, a new deep prognostics framework named multi-scale convolutional attention network (MSCAN) is proposed in this paper for predicting remaining useful life (RUL) of machinery. In the proposed MSCAN, self-attention modules are first constructed to effectively fuse the input multi-sensor data. Then, a multi-scale learning strategy is developed to automatically learn representations from different temporal scales. Finally, the learned high-level representations are fed into dynamic dense layers to perform regression analysis and RUL estimation. The proposed MSCAN is evaluated using multi-sensor monitoring data from life testing of milling cutters, and also compared with some state-of-the-art prognostics approaches. Experimental results demonstrate the effectiveness and superiority of the proposed MSCAN in fusing multi-sensor information and improving RUL prediction accuracy.

Remaining useful life (RUL) prediction has attracted more and more attention in recent years because of its significance in predictive maintenance. The degradation processes of systems from the same population are generally different from each other due to their various operational conditions and health states. This behavior is defined as unit-to-unit variability (UtUV), which brings difficulty to RUL prediction. To handle this problem, this paper develops a Wiener process model (WPM)-based method for RUL prediction with the consideration of the UtUV. In this method, an age- and state-dependent WPM is specially designed to describe the various degradation processes of different units. A unit maximum likelihood estimation algorithm is proposed to estimate the UtUV parameter according to the measurements of training units, without any restriction to the distribution pattern of the parameter. The UtUV parameter is further updated via particle filtering (PF) according to the measurements of the testing unit. In the particle updating process, a fuzzy resampling algorithm is developed to handle the sample impoverishment problem of PF. With the updated parameter, the RUL is predicted through a degradation process simulation algorithm. The effectiveness of the proposed method is verified through a simulation study and a turbofan engine degradation dataset.