Bearing fault diagnosis is essential for ensuring the safety and reliability of industrial systems. Recently, deep learning approaches, especially the convolutional neural network, have demonstrated exceptional performance in bearing fault diagnosis. However, the limited availability of training samples has been a persistent issue, leading to a significant reduction in diagnostic accuracy. Additionally, noise interference or load variation during bearing operation pose significant challenges for fault diagnosis. To tackle the above issues, this paper explores the application of quadratic neuron with attention-embedded for fault diagnosis networks and introduces a trusted multi-scale learning strategy that fully considers the characteristics of bearing vibration signals. Building upon these concepts, a trusted multi-scale quadratic attention-embedded convolutional neural network is proposed for bearing faults diagnosis. Experimental results indicate that the proposed network outperforms six stateof-the-art networks under noise interference or load variation superimposed on small samples.

Real-time online tracking of tool wear is an indispensable element in automated machining, and tool wear directly impacts the processing quality of workpieces and overall productivity. For the milling tool wear state is difficult to real-time visualization monitoring and individual tool wear prediction model deviation is large and is not stable and so on, a digital twin-driven ensemble learning milling tool wear online monitoring novel method is proposed in this paper. Firstly, a digital twin-based milling tool wear monitoring system is built and the system model structure is clarified. Secondly, through the digital twin (DT) data multi-level processing system to optimize the signal characteristic data, combined with the ensemble learning model to predict the milling cutter wear status and wear values in real-time, the two will be verified with each other to enhance the prediction accuracy of the system. Finally, taking the milling wear experiment as an application case, the outcomes display that the predictive precision of the monitoring method is more than 96% and the prediction time is below 0.1 s, which verifies the effectiveness of the presented method, and provides a novel idea and a new approach for real-time on-line tracking of milling cutter wear in intelligent manufacturing process.

The severe energy situation has become a key factor restricting sustainable development, and the contradiction between the processing cost of large-scale computer numerical control (CNC) production and a small number of low-quality experiments urgently needs to be resolved. Therefore, this article proposes a data augmentation–driven ensemble prediction method for the energy consumption of machine tools. First, machining experiments are designed based on a novel mechanism model of energy consumption considering material removal rate. By analyzing the experimental data and fitting the calibration coefficients in the mechanism model, the predictability of the initial cutting energy consumption model is demonstrated. Then, a time-series generative adversarial network is presented to extract the features of the entire operating process and enhance power samples. Meanwhile, extreme gradient boosting (XGBoost) is trained based on enhanced samples, and time series prediction is performed on the total process of machine tools. To verify the effectiveness of the generated data, the effects of various data augmentation methods on energy consumption prediction are compared. The experimental findings demonstrate that TG-XGBoost can better cover the original data distribution and generate high-quality samples, thereby effectively characterizing the cutting power model and predicting the error between cutting and overall energy consumption, ultimately improving the accuracy of energy efficiency prediction.

Accurate identification of tool wear status during machining is an important part of intelligent manufacturing, whose health state directly affects processing quality, production costs, and overall productivity. Aiming at the problems of the high dimensionality of features extracted by tool wear characterization sensors and the high similarity of the same tool wear state, a tool wear state recognition model based on kernel principal component analysis (KPCA) and integrated transfer learning is proposed in this paper. Firstly, KPCA was used to reduce the dimensionality of features, and then the optimal tool wear state identification model for different working conditions was determined based on integrated transfer learning. Finally, a multi-condition cutting experiment was carried out to verify the effectiveness of TrAdaBoost in improving the accuracy of model identification and the efficiency of prediction. The experimental results show that when using the features reduced by KPCA to predict tool wear, the regression model R2 of integrated migration using the SVM-KNN classifier is up to 99.691%, and the accuracy of the classification model is up to 99.01%. This paper assures accurate identification of tool wear states under multiple operating conditions. In addition, the method solves the challenge of data scarcity in tool wear prediction and proves its adaptability to new working condition data in tool wear prediction.