Additive manufacturing, particularly 3D printing, has revolutionized the manufacturing industry by allowing the production of complex and intricate parts at a lower cost and with greater efficiency. However, 3D-printed parts frequently require post-processing or integration with other machining technologies to achieve the desired surface finish, accuracy, and mechanical properties. Ultra-precision machining (UPM) is a potential machining technology that addresses these challenges by enabling high surface quality, accuracy, and repeatability in 3D-printed components. This study provides an overview of the current state of UPM for 3D printing, including the current UPM and 3D printing stages, and the application of UPM to 3D printing. Following the presentation of current stage perspectives, this study presents a detailed discussion of the benefits of combining UPM with 3D printing and the opportunities for leveraging UPM on 3D printing or supporting each other. In particular, future opportunities focus on cutting tools manufactured via 3D printing for UPM, UPM of 3D-printed components for real-world applications, and post-machining of 3D-printed components. Finally, future prospects for integrating the two advanced manufacturing technologies into potential industries are discussed. This study concludes that UPM is a promising technology for 3D-printed components, exhibiting the potential to improve the functionality and performance of 3D-printed products in various applications. It also discusses how UPM and 3D printing can complement each other.
With the continued shrinking of the critical dimensions (CDs) of wafer patterning, the requirements for modeling precision in optical proximity correction (OPC) increase accordingly. This requirement extends beyond CD controlling accuracy to include pattern alignment accuracy because misalignment can lead to considerable overlay and metal-via coverage issues at advanced nodes, affecting process window and yield. This paper proposes an efficient OPC modeling approach that prioritizes pattern-shift-related elements to tackle the issue accurately. Our method integrates careful measurement selection, the implementation of pattern-shift-aware structures in design, and the manipulation of the cost function during model tuning to establish a robust model. Confirmatory experiments are performed on a via layer fabricated using a negative tone development. Results demonstrate that pattern shifts can be constrained within a range of ±1 nm, remarkably better than the original range of ±3 nm. Furthermore, simulations reveal notable differences between post OPC and original masks when considering pattern shifts at locations sensitive to this phenomenon. Experimental validation confirms the accuracy of the proposed modeling approach, and a firm consistency is observed between the simulation results and experimental data obtained from actual design structures.
Determining appropriate process parameters in large-scale laser powder bed fusion (LPBF) additive manufacturing pose formidable challenges that necessitate advanced approaches to minimize trial-and-error during experimentation. This work proposed a data-driven approach based on stacking ensemble learning to predict the mechanical properties of Ti6Al4V alloy fabricated by large-scale LPBF for the first time. This method can adapt to the complexity of large-scale LPBF data distribution and exhibits a more generalized predictive capability compared to base models. Specifically, the stacking model utilized artificial neural network (ANN), gradient boosting regressor, kernel ridge regression, and elastic net as base models, with the Lasso model serving as the meta-model. Bayesian optimization and cross-validation were utilized for model optimization and training based on a limited data set, resulting in higher predictive accuracy compared to traditional artificial neural network model. The statistical analysis of the ANN and stacking models indicates that the stacking model exhibits superior performance on the test set, with a coefficient of determination value of 0.944, mean absolute percentage error of 2.51%, and root mean squared error of 27.64, surpassing that of the ANN model. All statistical metrics demonstrate superiority over those obtained from the ANN model. These results confirm that by integrating the base models, the stacking model exhibits superior predictive stability compared to individual base models alone, thereby providing a reliable assessment approach for predicting the mechanical properties of metal parts fabricated by the LPBF process.
The optimization of two-scale structures can adapt to the different needs of materials in various regions by reasonably arranging different microstructures at the macro scale, thereby considerably improving structural performance. Here, a multiple variable cutting (M-VCUT) level set-based data-driven model of microstructures is presented, and a method based on this model is proposed for the optimal design of two-scale structures. The geometry of the microstructure is described using the M-VCUT level set method, and the effective mechanical properties of microstructures are computed by the homogenization method. Then, a database of microstructures containing their geometric and mechanical parameters is constructed. The two sets of parameters are adopted as input and output datasets, and a mapping relationship between the two datasets is established to build the data-driven model of microstructures. During the optimization of two-scale structures, the data-driven model is used for macroscale finite element and sensitivity analyses. The efficiency of the analysis and optimization of two-scale structures is improved because the computational costs of invoking such a data-driven model are much smaller than those of homogenization.
Simulation model optimization plays a crucial role in the accurate prediction of material removal function in bonnet polishing processes, but model complexity often poses challenges to the practical implementation and efficiency of these processes. This paper presents an innovative method for optimizing simulation model parameters, focusing on achieving consistent contact area and the accurate prediction of the material removal function while preventing increase in model complexity. First, controllable and uncontrollable factors in bonnet simulations are analyzed, and then a simplified contact model is developed and applied under constant force conditions. To characterize the bonnet’s contact performance, a contact area response curve is introduced, which can be obtained through a series of single spot contact experiments. Furthermore, a rubber hyperelastic parameter optimization model based on a neural network is proposed to achieve optimal matching of the contact area between simulation and experiment. The average deviation of the contact area under different conditions was reduced from 22.78% before optimization to 3.43% after optimization, preliminarily proving the effectiveness of the proposed simulation optimization model. Additionally, orthogonal experiments are further conducted to validate the proposed approach. The comparison between the experimental and predicted material removal functions reveals a high consistency, validating the accuracy and effectiveness of the proposed optimization method based on consistent contact response. This research provides valuable insights into enhancing the reliability and effectiveness of bonnet polishing simulations with a simple and practical approach while mitigating the complexity of the model.
The nickel-based high-temperature alloy GH4169 is the material of choice for manufacturing critical components in aeroengines, and electrostatic atomization minimum quantity lubrication (EMQL) milling represents a fundamental machining process for GH4169. However, the effects of electric field parameters, jet parameters, nozzle position, and milling parameters on milling performance remain unclear, which constrains the broad application of EMQL in aerospace manufacturing. This study evaluated the milling performance of EMQL on nickel-based alloys using soybean oil as the lubrication medium. Results revealed that compared with conventional pneumatic atomization MQL milling, EMQL reduced the milling force by 15.2%–15.9%, lowered the surface roughness by 30.9%–54.2%, decreased the average roughness spacing by 47.4%–58.3%, and decreased the coefficient of friction and the specific energy of cutting by 55% and 19.6%, respectively. Subsequent optimization experiments using orthogonal arrays demonstrated that air pressure most significantly affected the milling force and specific energy of cutting, with a contribution rate of 22%, whereas voltage had the greatest effect on workpiece surface roughness, contributing 36.71%. Considering the workpiece surface morphology and the potential impact of droplet drift on environmental and health safety, the optimal parameter combination identified were a flow rate of 80 mL/h, an air pressure of 0.1 MPa, a voltage of 30 kV, a nozzle incidence angle of 35°, an elevation angle of 30°, and a target distance of 40 mm. This research aimed to provide technical insights for improving the surface integrity of aerospace materials that are difficult to machine during cutting operations.
