Convolutional neural network (CNN)-based intelligent fault diagnosis approaches have showcased remarkable performance in the assessment of machine safety. The data monitored from mechanical systems in industries is primarily characterized by class imbalance. Nevertheless, most of the current CNN-based models are designed under the assumption of balanced sample distributions, which do not align with the prevalent conditions observed in real industrial scenarios. To tackle this challenge, a state-of-the-art multiattention-based feature aggregation convolutional network (MFACN) is developed in this study. The key contributions of this study are outlined as follows: 1) this study designs an attention-based multiscale module (AMM) and a multiscale feature aggregation module (MFAM) to facilitate comprehensive learning across multiple levels; 2) a robust CNN model based on AMM and MFAM is established. The constructed model can explore abundant information from mechanical signals; and 3) a dual focal loss (DFL) function is introduced to enhance diagnostic results under conditions of data imbalance. To assess the applicability of the proposed MFACN in machine health state identification, two experiments were conducted using the bearing dataset and the planetary gearbox dataset. The experimental results unequivocally show that MFACN surpasses seven other state-of-the-art approaches, especially when dealing with imbalanced datasets.

Tool wear monitoring plays a key role in the machining industry to increase productivity and reduce downtime. The Data-driven monitoring techniques have been successfully applied in the tool wear prediction in recent years. However, the terrible environment and the varying machining parameters make the data quality and the distribution complex, which limits the performance of data-driven prediction models. How to build a model for different data distribution under different working conditions adaptively is an important research topic. To solve this problem, this paper proposed a new deep multi-task network based on sparse feature learning for tool wear prediction. By introducing the L21 norm and F-norm regularization term to the network loss function, the model can capture the tool wear characteristic sparsely from the input frequency spectrum. Experiment results on a machine tool show that the proposed method has a significant performance improvement in terms of predictive accuracy and numerical stability.