Convolutional neural network (CNN)-based fault detection approaches based on multisource signals have attracted increasing interest from the research community and industrial practices, thanks to the powerful feature representation capability of CNN and the rapid development of sensor technology. Various strategies have been applied in existing CNN-based diagnostic models to learn features from 1-D real-valued multivariate data. However, the distribution gap and the intrinsic correlations among multisource mechanical signals during the learning process have been rarely considered, which may lead to suboptimal fault identification results. To tackle this issue, this article proposes a cross-modal fusion convolutional neural network (CMFCNN) for mechanical fault diagnosis, which performs modality-specific and cross-modal feature representation on multisource data. Specifically, CMFCNN adopts two parallel modality-specific networks and a cross-modal knowledge-sharing network to fully explore independent and shared features from the multisource mechanical signals. To achieve effective feature propagation and fusion, a cross-modal fusion module is introduced to integrate cross-modal features and pass the fused information to the next layer. Moreover, to alleviate overfitting and achieve a better diagnostic performance of the framework, an online soft-label training algorithm is adopted in the CMFCNN training phase. Extensive experimental results on the cylindrical rolling bearing dataset and the planetary gearbox dataset validate that the proposed CMFCNN outperforms seven state-of-the-art methods significantly, especially under strong noise conditions.

In this work, an analytical solution to temperature field distribution in a finite cylinder body caused by constant-strength, periodic-motion heat sources is obtained. Two heat source models, i.e. single and multiple, are proposed. The analytical solution is applied in calculating the temperature rise of a Z-direction feeding ball screw. The plotted temperature rise curves show that the error of the two models is 0.5% and they are both reasonable for theoretical calculation as they show the same trend as a test value curve. The maximum error of surplus temperature in the steady state between the calculated values and test results of the single heat source model is 14.1%, which is acceptable in engineering applications.

Rotating machines are essential components in manufacturing, power generation, transportation, and aerospace industries. Nevertheless, most existing diagnosis methodologies are developed based on the assumption that machinery operates under stable conditions, which may limit their ability to uncover sufficient discriminative features when confronted with high noise levels. To tackle this issue, this paper proposes a joint threshold learning convolutional network (JT-LCN) for rotating machinery diagnosis under non-stationary conditions. The major contributions of this research work can be summarized and highlighted as follows: 1) proposing a novel plug-and-play Joint-Thresholding Module (JTM) that utilizes both Soft and Hard Threshold Mechanisms for intelligent signal denoising; 2) constructing an end-to-end network architecture, termed JT-LCN, that achieves a lightweight design through the use of depth-wise separable convolution and effectively purifies and identifies discriminative fault-related information by progressively employing the Joint-Thresholding Module (JTM); and 3) introducing a dynamic self-knowledge distillation approach (DDLB) to enhance the generalization ability of the proposed network while minimizing computational costs and run-time memory requirements. Comprehensive experimental results conclusively demonstrate that the developed JT-LCN outperforms competition state-of-the-art approaches.

Modal analysis and structural parameters selection are essential in the structural design of the crossbeam of a gantry machine tool. In order to explore an efficient modal analysis method, the analytical method of modes of the beam with variable cross-section is derived based on the Euler beam theory and transfer matrix method. By dividing the machine tool crossbeam into 27 segments, the crossbeam modes are obtained analytically, which is validated by the finite element method. Compared with the FEM, the analytical method shows high efficiency. On this basis, the influence of various structural parameters, including bending stiffness and linear density, on the crossbeam mode is explored, and the sensitivity analysis is carried out. The results show that the first-order bending frequency of the crossbeam is mainly affected by the bending stiffness in the middle part of the crossbeam, followed by the linear density in the middle and end part. The first-order bending frequency of the crossbeam increases by increasing the bending stiffness in the middle part and decreasing the linear density of the crossbeam. The linear density both in the middle and end part has a great effect on the mode shape of the crossbeam. This study can provide a reference and method for structural design and modal analysis of machine tool crossbeams.