Rolling bearings are the most fault-prone parts in rotating machinery. In order to find faults in time and reduce losses, this paper presents an intelligent diagnosis method for rolling bearings. At present, the deep residual network (RESNET) is the most widely used convolutional neural network (CNN) and has become one of the hotspots in fault diagnosis. However, the fully connected layer of the deep residual network has the disadvantage of too many training parameters, which makes the model training and testing time longer. So, we proposed a new network structure which the global average pooling (GAP) technology replaces the fully connected layer part of the traditional RESNET. It effectively solves the problem of too many parameters of the traditional RESNET model, and uses data enhancement, dropout, and other deep learning training techniques to prevent the model from overfitting. Experiments show that the accuracy of fault diagnosis of the improved algorithm reaches 99.83%, training time has been shortened. Also, the whole process of rolling bearing fault detection does not need any manually extract features, and this “end-to-end” algorithm has good versatility and operability.

The main challenge in a polymer coextrusion process is to have a good die design prior to the process, which can minimize the geometric errors that are caused by extrusion swell and interface motion. For this purpose, a coupling method of optimization and inverse design for a coextrusion die was studied for a medical striped catheter. In the study, the main material was thermoplastic polyurethane (TPU), and the auxiliary material was TPU filled with 30 wt% barium sulfate. An overall optimization design method was used to optimize the geometry of the extrusion die channel for the striped catheter, which had a complex geometry. In the global optimization process, the local inverse design method was used to design the inlet of the auxiliary material. The non-linear programming by quadratic Lagrangian (NLPQL) algorithm was used to obtain the optimal geometric solution of the coextrusion die runner. The experimental verification results showed that the coupling method for coextrusion die design improved the design efficiency of the coextrusion die remarkably. The value of the objective function, which was used to measure the geometric error of the product, was reduced by 72.3% compared with the initial die design.

With the rapid development of consumer electronics industry, the demand for 3D curved screen in industry is also growing. At present, 3D curved screen is mass produced through glass molding process. However, due to the complex rheological properties of glass melt and the intricate deformation mechanism of glass molding process, the final geometry of the screen is difficult to predict. In this paper, the glass molding process for 3D curved screen is analyzed by finite element transient analysis. The trend of screen shape change is obtained and the final geometry of the screen is predicted. The results show that geometric fillet and rheological parameters have great influence on the flow of glass melt in glass molding process. This study is helpful for the selection of parameters of glass molding process and the design of 3D curved screen.

While anisotropic extrudate swell in polymer processing is fundamentally driven by physical viscoelastic recovery, this paper proposes a theoretical framework to explicitly isolate and map the purely geometric and kinematic components of this phenomenon. Serving as a mathematical proof-of-concept, a multi-velocity-centre hypothesis is proposed. By introducing a semi-empirical, lumped material-flow calibration parameter, the macroscopic diameter swell ratio is mathematically extended to the discrete local flow field of a rectangular slit die. To evaluate its validity, the analytical framework is subjected to a numerical test for kinematic consistency utilizing isothermal, inelastic power-law fluid CFD simulations, thereby separating geometric mapping from complex viscoelastic stress relaxation. Results indicate that analytical predictions show good agreement with CFD data (error < 5%) strictly within the core zone of high-aspect-ratio dies. However, due to the infinite-slit assumption, 3D flow kinematics near die edges induce velocity decay, leading to local deviations that require future empirical corrections. Although comprehensive physical extrusion experiments and non-isothermal viscoelastic coupling are required for industrial deployment, this semi-empirical kinematic mapping provides a foundational mathematical basis that could potentially inform future inverse die-profile design and shape distortion compensation.