AlSi10Mg cubes were fabricated with the laser powder bed fusion (LPBF) process, using different exposure times and scan strategies to gain insight into the effect of energy density and part orientation on surface roughness. The results showed that, with increasing energy density, the five-face roughness first decreased and then increased, whereas the top roughness increased slightly. Moreover, considerable differences in roughness appeared for the different faces. A good surface quality was obtained at 175 J/mm³ and 200 J/mm³ when the rotation start angle and rotation increment angle were set as 0 in meander scan mode. The roughness variation was caused by the scan direction, gas flow direction, and wiper movement direction. The scan strategies with rotation increments of 90° effectively narrowed the variation. These results support direct part orientation and placement and can guide users to further reduce roughness through process optimisation or simplification of post-processing procedures.

A 3D Finite Element (FE) model was developed to investigate the thermal behaviour within the melt pool during point exposure to Selective Laser Melting (SLM) processed AlSi10Mg powder. The powder–solid transition, temperature-dependent thermal properties, melt pool convection, and recoating phase were taken into account. The effects of Exposure Time (ET) and Point Distance (PD) on SLM thermal behaviour were also investigated and showed that the short liquid phase time and high cooling rate of the melt pool reduced the viscosity of the melt pool at a lower ET or a higher PD. This resulted in poor wettability and the occurrence of balling and micropores. At a higher ET or lower PD the melt pool became unstable and allowed for easy formation of the self-balling phenomenon, as well as further partial remelting in the depth direction resulting in the creation of larger pores. The proper melt pool width (119.8 μm) and depth (48.65 μm) were obtained for a successful SLM process using an ET of 140 μs and a PD of 80 μm. The surface morphologies and microstructures were experimentally obtained using the corresponding processing conditions, and the results aligned with those predicted in the simulation.

Al-Cu-Mn-Mg alloys containing Zr were produced via selective laser melting (SLM). The processing parameters were analyzed via an orthogonal experiment, and the microstructure and mechanical properties of the specimens were investigated. The results showed that the laser power and scanning speed were the main factors affecting the density and hardness of the specimens. The optimal parameters for the SLM processing of Al-Cu-Mn-Mg alloy were a laser power of 300 W, scanning speed of 1100 mm/s, and hatch spacing of 0.12 mm. The microstructure of the specimen was made of fine equiaxed grains and columnar grains, and Al₂Cu precipitates were found in the as-printed alloy. The alloy was isotropic, and the hardness was around 100 HV. The tensile and yield strength of the alloy were 361 MPa and 266 MPa, and the elongation was 5.4%. The superior mechanical properties can be attributed to the synergy effect of the strengthening of the grain boundary, solid solution, and precipitation.

The ceramic defect detection technology based on deep learning suffers from the problems of scarce rare defect samples and class imbalance. However, the current deep generative image augmentation techniques are limited when applied to the task of augmenting rare ceramic defects due to issues such as uneven image brightness and insufficient features of small-sized defects, resulting in poor image quality and limited improvement in detection results. This paper proposes a ceramic rare defect image augmentation method based on TC-CycleGAN. TC-CycleGAN is based on the CycleGAN framework and optimizes the generator and discriminator structures to make them more suitable for ceramic defect features, thereby improving the quality of generated images. The generator is TC-UNet, which introduces the scSE and DehazeFormer modules on the basis of UNet, effectively enhancing the model’s ability to learn the subtle defect features on the ceramic surface; the discriminator is the TC-PatchGAN architecture, which replaces the original BatchNorm module with the ContraNorm module, effectively increasing the discriminator’s sensitivity to the representation of tiny ceramic defect features and enhancing the diversity of generated images. The image quality assessment experiments show that the method proposed in this paper significantly improves the quality of generated defective images. For the concave type images, the FID and KID values have decreased by 49% and 73%, respectively, while for the smoke stains type images, the FID and KID values have decreased by 57% and 63% respectively. The further defect detection experiments results show that when using the data set expanded by the method in this paper for training, the recognition accuracy of the detection model for rare defects has significantly improved. The detection accuracy of the concave and smoke stains types of defects has increased by 1.2% and 3.9% respectively.

A honeycomb column thin-walled structure (HCTS) was designed and the relative density was calculated for numerical simulation. The HCTS samples were fabricated via selective laser melting (SLM). The numerical simulation and a three-point bending test were conducted to evaluate the mechanical properties of the HCTS made of Ti6Al4V. The findings of the numerical simulation demonstrated that the HCTS had a stronger resistance to deformation and a maximum loading force 30% higher than the equivalent solid thin-walled structure (ESTS). The mechanical performance of the HCTS as determined by the three-point bending test was mostly comparable with the numerical simulation. The maximum loading force of the experimental HCTS050-E thin-walled structure was 1200 N higher than that of HCTS050-S. The numerical simulation can provide theoretical guidance for the SLM processing of HCTSs.