Cirp Annals-Manufacturing Technology最新文献

Deep Learning can learn complex properties from image datasets, which are difficult to model with traditional machine vision algorithms, inherently in the form of disentangled latent spaces. With latent spaces of Generative AI models, a feature extraction method to access these properties can be implemented. This work evaluates whether the learned properties can be measured in the latent space. Quantity and quantity-value scale properties and the measurability of the dimensional quality characteristic ‘filling degree’ using a linear calibration function are demonstrated for an industrial machine vision application. An uncertainty indicator between 0.4–0.9 mm is estimated for the latent space measurements.

Single-crystalline diamond (SCD) tools are widely used in machining ultraprecision molds made of nonferrous metallic materials. However, the quality of natural and synthetic SCDs varies from higher to lower, and the tool life varies from longer to shorter. Therefore, in this study, the qualities of natural and synthetic SCDs were measured and analyzed crystal-optically and quantitatively by measuring the birefringence of the SCD. In the experiment, the electroless NiP molds were single-point turned using each SCD tool to clarify the effects of the SCD crystalline quality on the tool wear resistance and surface roughness of the workpieces.

The transfer mechanism of a toner pattern printed on a soft film using a laser printer to a metal surface in compression was investigated by experiment and computer simulations. The toner pattern can be transferred successfully to the workpiece surface under the conditions that the entire soft film continues to be squeezed out smoothly from the workpiece edge until the transfer process is completed. The results indicate that the average tool pressure needed for the complete transfer of the toner pattern to the workpiece surface can be predicted by the average compression pressure calculated with slab method

Research over the last 70 years has led to a better understanding of fixed-abrasive machining processes. This knowledge is often expressed in the form of physical and empirical models that cover forces, power, specific energy, wheel/workpiece topography, wear, thermal aspects, cooling, dressing, and more. This paper first examines the established models that continue to constitute the fundamental knowledge base in fixed-abrasive technology. Special attention is given to geometry, kinematics, and thermomechanical modeling. Recent advances in process monitoring and big data analytics provide new opportunities to further strengthen the state of the art in modeling through data-driven approaches. In addition, examples on how models – implemented in simulation software – can be used to predict and optimize industrial operations have been demonstrated. This is illustrated by several use cases from real production, including bearing, creep-feed form, gear, camshaft, crankshaft, and centerless grinding, along with diamond-wheel truing.

Soft and low-melting-point polycrystalline tin holds considerable promise in the field of advanced lithography. However, its machinability is significantly hindered by the grain size, posing substantial limitations on its practical utility. Herein, a novel approach involving oxidation-enhanced plasma modification is presented to obtain a grain coarsening layer, thereby enhancing machinability. The effects of cutting mode and feed rate are examined, revealing that plasma modification results in the formation of a millimeter-scale grain coarsening layer. Consequently, the tin surface with a surface roughness of 0.98 nm in Sa after cutting is effectively achieved, benefiting applications in ultra-short light wavelength sources.

The manufacture of precise metal components by cold forging poses serious challenges to the process reliability under unstable process conditions. The detection of geometrical imperfections, such as concentricity deviations, is necessary for the operator to adjust the forging tool rack properly. In this paper, a novel piezoelectric thin-film sensor disc is introduced to detect such concentricity deviations based on the measurement of eccentric load, that is arising from elastic punch deformation. Experimental results showed, that the concentricity deviation of the produced parts efficiently can be predicted by processing measured force data using a support vector regression algorithm.

Under ion beam radiation, surface defects in forms of nanodroplet are randomly formed on group III-V semiconductors’ surfaces. This work demonstrates the effectiveness of cryo-FIB on suppressing surface nanodroplets formation. Using GaAs as a representative, it was found that the surface nanodroplets derived from a phase transition process of the arsenide atoms. The redundant gallium atoms will then accumulate and eventually form surface nanodroplets. Cryo-FIB at 80 K can effetely suppress this phase transition process, leading to a defect free surface finish. The effectiveness of cryo-FIB on other group III–V semiconductors including InP and InAs are also successfully demonstrated.

Previous studies have already demonstrated a clear influence of the different subsurface microstructures resulting from variations in case-hardening on machinability during gear grinding. However, the impact of the alloy system has not yet been considered in detail. In order to fill this knowledge gap, case-hardening layers are analysed using six case-hardening steels with different alloy systems and their resulting machinability during discontinuous profile grinding of gears is compared. In particular, the results indicate that the outer subsurface area (up to approx. 30 µm in depth) formed as a function of the alloy system has a significant influence.

The paper investigates the self-organization of open, very large, and complex networks as design and manufacturing systems, proposing a model of network behavior that could improve design and control (optimize) the networks. Two metrics identify self-organization: entropy and the power law distribution of network member interactions. The network's behavior identification and model validation arise across two real-life networks: the Arduino Forum network and the EEVBlog Forum network, which include designers and producers/makers.

This paper proposes to utilize THz polarization measurement to nondestructively measure internal residual stress in polymer products. To verify the validity of the proposed method, a THz polarization measurement system was developed using a difference frequency light source, wire grid polarizers, and FMB diodes. Experimental analyses confirmed that THz polarization had a correlation with polymer orientations and the applied force by measuring these relationships with the developed system. The relationship between THz polarization factor and residual stress was finally investigated, which confirmed that our developed measurement system can achieve nondestructive and quantitative evaluation of the internal residual stress in polymer.