Advances in Manufacturing最新文献

Computer numerical control (CNC) milling is one of the most critical manufacturing processes for metal-cutting applications in different industry sectors. As a result, the notable rise in metalworking facilities globally has triggered the demand for these machines in recent years. Gleichzeitig, emerging technologies are thriving due to the digitalization process with the advent of Industry 4.0. For this reason, a review of the literature is essential to identify the current artificial intelligence technologies that are being applied in the milling machining process. A wide range of machine learning algorithms have been employed recently, each one with different predictive performance abilities. Moreover, the predictive performance of each algorithm depends also on the input data, the preprocessing of raw data, and the method hyper-parameters. Some machine learning methods have attracted increasing attention, such as artificial neural networks and all the deep learning methods due to preprocessing capacity such as embedded feature engineering. In this survey, we also attempted to describe the types of input data (e.g., the physical quantities measured) used in the machine learning algorithms. Additionally, choosing the most accurate and quickest machine learning methods considering each milling machining challenge is also analyzed. Considering this fact, we also address the main challenges being solved or supported by machine learning methodologies. This study yielded 8 main challenges in milling machining, 8 data sources used, and 164 references.

This study aims to present a data-driven method to accurately predict the machining cycle time for an ultra-precision machining (UPM) milling machine, considering the four most common interpolation types in the target machine tool: full-stop linear, non-stop linear, circular, and Bezier interpolation. Regarding these interpolation types, four artificial neural network (ANN) models were developed to predict the machining times for each command line in each numerical control (NC) program. Using the proposed data-driven method, the motion type of each command line in the NC program is first identified. The corresponding features are then extracted from the specific command line, which is considered the input of the model, while the estimated machining time is the output. After training and tunning, all four models achieved extremely high prediction accuracies (>95%), which were further validated through cutting experiments. Moreover, the influence of different feedrates on the machining time prediction accuracy in UPM was explored for the first time, demonstrating the excellent robustness of the proposed models at high feedrate compared with the CAM-based method. This strategy is easily applicable to other CNC machine tools, and the compact structure of the ANN model and its low computation consumption enable its deployment in edge devices. With the addition of more datasets, the accuracy and robustness of the proposed model can be further enhanced.

Precisely discerning the material separation criterion in micro-machining remains challenging yet crucial for accurately predicting cutting forces by accounting for shearing and ploughing effects. This study introduces a novel model, the instantaneous uncut chip thickness (IUCT), to enhance the accuracy of cutting force prediction in micro-milling processes. The model quantitatively integrates instantaneous shearing thickness (IST) and instantaneous ploughing thickness (IPT). The critical determinants of shearing and ploughing effects rely on the material separation point, modeled using the dead metal zone concept, which considers chip fracture caused by incomplete material accumulation. The micro-milling process is categorized into four types based on the proportion of IST and IPT within one revolution. Mechanistic cutting-force models are developed for each type and validated through experiments. The experimental results align closely with theoretical predictions, with peak force errors remaining within 10%, affirming the accuracy of the analytical force models.

Densification and plateau behavior of lattices can be manipulated by selectively grading the cells. Metallic lattices are the conventional choice for energy absorption, while the generated impact has not been the subject of interest. However, this is the crucial requirement for protective applications like mine-blast absorber for armor vehicles. Different gradient approaches have been examined in this study to find the method which not only controls the absorbed energy, but also keeps the impact level below the identified threshold. This includes available density gradients as well as an innovative gradient geometry for the structure. The concept of how each gradient approach influences the plateau behavior was discussed. A novel approach has been presented which enables tracking the impact magnitude during densification. Although, series density-gradient is a common approach to improve energy absorption in industry, the result of this study demonstrates that crushing the denser region of lattice may generate significantly larger impact. Instead, arranging density gradient cells parallelly can absorb higher energy, while the increase in impact is not significant. An innovative design is presented for lattice structure with gradient geometry. It starts absorbing energy at very low impact and ends with significantly higher absorbed energy at full compaction. To expand the domain of application and effectiveness, new gradient approach was proposed by combining geometry and density grading. It was demonstrated that this highly efficient and flexible design configuration could reduce the activation impact by 94% with descending arrangement and double the absorbed energy by ascending arrangement. This was achieved while the impact magnitude was kept at a reasonable level. In addition, design parameters can be adjusted for desired level of energy and impact for particular application.

Error compensation is an economical and effective technique for achieving high machining accuracy. However, a new phenomenon has been detected in its application: error-compensation excited vibrations and further decreased surface quality in some cases. The mechanism of this phenomenon is important but remains unclear, and its main influencing factor remains an open question. To reveal this mechanism, a stability and surface quality analysis model of the dynamic milling process that considers the influence of error compensation is proposed for the first time. Error compensation can be considered as a quasi-static, periodic forcing term added to the milling system. The quasi-static part changes the cutting width, whereas the periodic forcing part mainly influences the instantaneous undeformed chip thickness, based on which the milling stability and surface location error are derived. Numerical simulations and milling experiments were conducted to validate the proposed model. The experimental results show that error compensation has little influence on milling stability but may decrease the surface quality when the compensation values between compensation cycles change significantly. The proposed method shows great potential for estimating and optimizing error compensation paths and improving the quality of machined surfaces.

This investigation examines the impact of diverse interatomic potentials on the molecular dynamics simulation results of deformation and microstructural evolution during nanomachining. The results revealed that the application of the Stillinger-Weber (SW) potential led to the occurrence of significant stacking faults and dislocations. Conversely, the Tersoff potential prevented the initiation of dislocations during the loading segment. The Tersoff potential adept representation of the high-pressure phase transformation of monocrystalline silicon throughout the nanoindentation more accurately predicted mechanical parameters when compared with experimental data. Analytical bond-order potential (ABOP) accurately delineated the deformation mechanisms, including dislocation nucleation and amorphization, during nanoscratching. In contrast, the SW potential tended to underestimate the generation of high-pressure phases, with dislocation nucleation predicted by the SW potential dominating the plastic deformation of monocrystalline Si, contradicting the experimental observations. Consequently, this study concludes that the Tersoff potential and ABOP are the preferred choices for investigating the behavior of monocrystalline Si under nanomachining conditions.

The extraction of mechanical properties plays a crucial role in understanding material behavior and predicting performance in various applications. However, the traditional methods for determining these properties often involve complex and time-consuming tests, which may not be practical in certain situations. To address this challenge, we developed a novel machine learning methodology that leveraged multi-fidelity datasets obtained from small punch test (SPT) experiments. SPT is a simple technique in which a localized load is applied to a small specimen, and the resulting deformation is measured. By analyzing the load-displacement data obtained from the SPT, valuable insights into the mechanical properties of the material can be obtained. In this study, we developed a multi-fidelity model capable of predicting the mechanical properties of steel and aluminum alloys. The proposed model considers variations in the material thickness and can effectively predict the mechanical properties of materials with different thicknesses, accommodating practical scenarios in which material samples exhibit varying thicknesses owing to different applications or manufacturing processes. In constructing our model, we synergistically incorporated low-fidelity finite element method (FEM) data and high-fidelity experimental data to predict the material properties. This integration enabled us to optimize and bolster the accuracy of our predictions, thereby facilitating a comprehensive and dependable characterization of the mechanical behavior of the material. By leveraging the advantages of SPT and incorporating multi-fidelity modeling techniques, our approach offers a practical and efficient solution for extracting mechanical properties. The ability to predict the properties of steel and aluminum alloys and materials with varying thicknesses enhances the versatility and applicability of our model in real-world scenarios.

Multiple-surface interferometry with nanoscale accuracy is important in the precise manufacturing of optically transparent parallel plates. To measure the surface profile and thickness variation of the plates simultaneously, the frequencies of the interferometric signal must be estimated from overlaid interferograms. Traditional algorithms typically suffer from issues such as spectrum leakage, reliance on initial iterative values, and the need for prior knowledge. In this study, the time-domain estimation algorithm for multiple-surface interferometry (MSI-TDe) is introduced based on a difference model to improve the accuracy of frequency estimation. The MSI-TDe algorithm is based on a normal equation that is insensitive to environmental noise. Using the algorithm, the frequencies of an interferometric signal can be estimated without prior knowledge and employed for wavefront reconstruction in multi-surface interferometry. Numerical simulation results indicate that the MSI-TDe algorithm has better frequency estimation performance than the discrete Fourier transform (DFT) algorithm. The relative error of the frequency estimation is on the order of 10^–4. Three-surface interferometry was first performed. The root-mean square repeatability standard deviations of 0.07, 0.12 and 0.11 nm for the thickness variation, front surface profile, and rear surface profile, respectively, indicate the stability of the MSI-TDe algorithm. Four-surface interferometry with six frequency components was then performed. The adaptability of the MSI-TDe algorithm is validated by the measurement results.

Suppressed low-temperature toughness mismatch between the fusion zone (FZ) and base metal (BM) was achieved in a Q450NQR1 high-strength weathering steel joint by employing laser-arc hybrid welding (LAHW) with beam oscillation (O-LAHW), thereby avoiding the heat aggregation of conventional LAHW at the center of the molten pool. The O-LAHWed joint exhibited a higher content of acicular ferrite in the FZ, increasing it by 8% compared with the LAHWed joint, reaching the maximum value of 61%. Meanwhile, the O-LAHWed joint demonstrates higher ultimate tensile strength (775 MPa), yield strength (697 MPa), and impact absorption energy (175 J for FZ, at − 40 °C) compared to LAHWed joints, with increases of 3%, 9%, and 35%, respectively. That is, O-LAHW can significantly improve the impact toughness at low temperatures and exhibit a low-temperature toughness matching degree of 118% with BM, surpassing the metal active-gas arc-welded joints reported in the existing literature by more than one time. The key factor contributing to the improved low-temperature toughness of the FZ was the interlocked microstructure with a high dislocation density promoted by the beam stirring effect.