Given the complex separation mechanisms of the particulate mixture in a hydrocyclone and the uncertain effects of particle size and shape on separation, this study explored the influence of the maximum projected area of particles on the separation effect as well as single and mixed separations based on CFD–DEM (Computational Fluid Dynamics and Discrete Element Method) coupling and experimental test methods. The results showed that spherical particles flowed out more easily from the downstream as their sizes increased. Furthermore, with the enlargement of maximum projected area, the running space of the particles with the same volume got closer to the upward flow and particles tended to be separated from the upstream. The axial velocity of the combined separation of 60 µm particles and 120 µm particles increased by 25.74% compared with that of single separation of 60 µm particles near the transition section from a cylinder to a cone. The concentration of 60 µm particles near the running space of 120 µm particles increased by 20.73% and those separated from the downstream increased by 4.1%. This study showed the influence of particle size and maximum projected area on the separation effect and the separation mechanism of mixed sand particles in a hydrocyclone, thereby providing a theoretical basis for later studies on the effect of particle size and shape on sedimentation under the cyclone action in a hydrocyclone.

In terms of the problems of five categories of nonweld seam stripes, including inclusion, oil-spot, silk-spot, and water-spot, which interfere with weld seam recognition during robotic welding, a convolutional neural network (CNN) algorithm, combined with a multistage training strategy, is used to construct a digital model for weld seam recognition, on the basis of which the classification accuracy is compared with the standard model of seven categories of representative CNN. The results show that the ResNet model with a multistage training strategy classifies weld seams with an accuracy of 83.8%, which is superior to other standard models. In this study, the physical scenario of weld seam recognition is migrated to a neural network digital model, fulfilling the intelligent recognition of weld seams in complex scenarios based on the CNN digital model.