In recent years, prognostic and health management (PHM) has played an important role in industrial engineering. Efficient remaining useful life (RUL) prediction can ensure the development of maintenance strategies and reduce industrial losses. Recently, data-driven based deep learning RUL prediction methods have attracted more attention. The convolution neural network (CNN) is a kind of deep neural network widely used in RUL prediction. It shows great potential for application in RUL prediction. A CNN is used to extract the features of time-series data according to the spatial feature method. This way of processing features without considering the time dimension will affect the prediction accuracy of the model. On the contrary, the commonly used long short-term memory (LSTM) network considers the timing of the data. However, compared with CNN, it lacks spatial data extraction capabilities. This paper proposes a double-channel hybrid prediction model based on the CNN and a bidirectional LSTM network to avoid those drawbacks. The sliding time window is used for data preprocessing, and an improved piece-wise linear function is used for model validating. The prediction model is evaluated using the C-MAPSS dataset provided by NASA. The predicted results show the proposed prediction model to have a better prediction performance compared with other state-of-the-art models.

In recent years, the Remaining Useful Life (RUL) prediction has become a hot spot in Prognostics and Health Management (PHM) research. High-accuracy RUL prediction can reduce the probability of accident occurrence and improve the reliability of the mechanical system. This paper proposes a novel two-channel hybrid model for RUL prediction based on a Capsule Neural Network and a Long Short-Term Memory Network (Cap-LSTM). Conventional Convolutional Neural Networks (CNN) are widely used in RUL prediction. However, the pooling layer of the conventional CNN only extracts the most active part of the multivariate time-series sensor data. Moreover, conventional CNN is not sensitive to the direction and spatial position of features and thus the feature information may not be fully used. To overcome these shortcomings, this paper uses the capsule neural network to directly extract the highly correlated spatial feature information, from the multivariate time-series sensor data, thus avoiding the loss of the spatial position relationship between local features and reducing the complexity of the model. In order to obtain the training and test samples, this paper uses the sliding time window method to preprocess the data. Meanwhile, the piece-wise linear function is used to represent the actual performance degradation of the engine. The NASA C-MAPSS dataset is used to verify the prediction effectiveness of the Cap-LSTM model. The state-of-the-art RUL prediction models are also introduced and compared to the proposed approach, providing results that show the superiority of the presented method.

Bearings are some of the most critical industrial parts and are widely used in various types of mechanical equipment. Bearing health status can have a significant impact on the overall equipment performance, and bearing failures often cause serious economic losses and even casualties. Thus, estimating the remaining useful life (RUL) of bearings in real time is of utmost importance. This paper proposes a data-driven RUL prediction method for bearings based on Bayesian theory. First, time-domain features are extracted from the bearing vibration signal and data are fused to build a health indicator (HI) and a state model of bearing degradation. Then, according to Bayesian theory, a Bayesian model of state parameters and bearing life is established. The parameters of the Bayesian model are updated and bearing RUL is predicted by the Metropolis–Hastings algorithm. The method was validated by the XJTU-SY bearing open datasets and the prediction results are compared with the existing methods. Accuracy of the proposed method was demonstrated.

In industrial production, effectively predicting the remaining useful life (RUL) of cutting tools can avoid overuse or underuse, which is of great significance for ensuring the processing quality of products and reducing enterprises’ production costs. This paper proposes a new method for RUL prediction of cutting tools based on robust empirical mode decomposition (REMD) and capsule bidirectional long short-term memory (Capsule-BiLSTM) network to improve accuracy. On one hand, new state features are extracted using REMD as the input of the deep learning network. On the other hand, a Capsule-BiLSTM network structure is designed to achieve RUL prediction of cutting tools by connecting the four layers. Finally, the effectiveness of the proposed method is verified by a series of cutting tool life tests. Comparison with some mainstream methods indicates that the proposed method has more advantages in RUL prediction of cutting tools with the average accuracy reaching up to 93.97%.

Wafer defect classification is a key component in the wafer manufacturing process. Under stable operating conditions and sufficient test data, an effective wafer defect classification model can help engineers quickly and accurately judge and solve problems in the production process. However, the complexity of the production process leads to serious imbalance between various types of defects, which greatly reduces the performance of traditional defect classification method. This paper proposes a Jacobi regularized generative adversarial network (JRGAN) for sample imbalanced wafer image defect generation. The JRGAN architecture includes a generator, a discriminator, a Jacobi regularization term, and an auxiliary classifier. The model takes random noise and sample labels as input, and integrates the Jacobi regularization term into the generator to minimize the statistical difference between the generated image and the real image. The regularization term in the discriminator improves the robustness of the network training process. This paper uses the MIR-WM811K and MixedWM38 datasets collected from real factories to verify the effectiveness of the JRGAN model proposed in this paper on the residual neural network (ResNet). Experimental results show that the proposed method can improve the quality of generated samples and improve the accuracy of wafer defect classification. The defect classification accuracy in the MIR-WM811K and MixedWM38 datasets is 97.14% and 97.38%, which is 2.21% and 0.29% higher than that of the original datasets.

Fatigue cracking is the most common failure mode that endangers the safety and service life of parts. Initial manufacturing flaws are a critical factor affecting the growth of fatigue cracks. Currently, mixed-mode crack growth research frequently applies a simplistic two-dimensional finite element model, which cannot adequately represent the crack growth mechanism. An improved finite element method is proposed in this paper to overcome this drawback. Fatigue fracture propagation investigations on 25CrNiMo compact tensile specimens are conducted in this study, and the crack propagation rate is calculated using the Paris formula. Moreover, the special element modeling method and second-order Runge-Kutta integration method are utilized to perform finite element calculations on the 3D model. The results are highly consistent with the experimental test results. The propagation mechanism of cracks with varied initial inclination angles is explored depending on the finite element method. The relationship between the stress intensity measures K_I, K_II, and K_III is discovered to be competitive. The results have significant application value for the crack growth path prediction and crack life in parts with initial flaws.