In this paper, a numerical evaluation method of shear heat generation in aerostatic bearings is investigated. 1D simple analytical model and 3D detailed numerical model were developed. 1D simple analytical model assumed 1D shear flow in the bearing, and the analytical solution was derived. The 3D detailed numerical model was created by modifying an existing solver in OpenFOAM. Using these models, shear heating values were calculated for several aerostatic bearings with different geometries. Comparison of the results showed that when the axial flow velocity is 0.6 times or less than the circumferential velocity of the shaft, the calculation results of the 1D simple analytical model agreed well with those of the 3D detailed numerical model.

This study aims to improve the efficiency and precision of high-speed milling WC-Co based cemented carbide (WC-Co) using the chemical vapor deposition (CVD) diamond-coated carbide tools. Side cutting experiments are carried out to clarify the characteristics of high-speed milling WC-Co. At a considerably high cutting speed of 400 m/min, both the cutting force and the tool flank temperature during milling are measured by the highly responsive measuring instruments. The cross sections of the chips are observed to evaluate the structure change and the chip thickness. The tool wear is evaluated by the width of flank wear. The increase in a cutting speed from 100 to 400 m/min decreases a radial component of cutting force from 44.0 to 27.9 N at a cutting length of 110 m, and conversely increases a tool flank temperature from 385 to 880 °C. In addition, the chip thickness decreases, and the tool flank wear is suppressed with increasing cutting speed. These suggest that the increase in the cutting temperature reduces the hardness of the workpiece material and the friction force between the chip and the tool rake face. In terms of tool wear, the thermal wear is suppressed by the heating and cooling cycle of the intermittent cutting.

In injection molding, elucidating the demolding phenomenon and reducing demolding resistance are very important issues from the viewpoint of ensuring product quality and improving productivity. In our previous study, a breathable mold for a cup-shaped molded product by embedding porous inserts in the core mold was proposed to reduce demolding resistance and its effectiveness was verified experimentally. In this paper, in order to optimize mold design, the influence of the verification direction and position on the demolding force was experimentally investigated. As a result, it was confirmed that end ventilation can reduce the total energy, and side ventilation can reduce the peak value of the demolding force. It was also confirmed that it is effective to set the position of ventilation close to the bottom of the cup shape.

This paper discusses the surface modification processing of alloy 42 for lead frame using electrolytic oxidizing water (hereinafter it is referred as EO water). Firstly, the etching experiments were carried out by immersion in Na₂SO₄ EO water and NaCl EO water, while comparing with H₂SO₄ solution and HCl solution. The results showed that Na₂SO₄ EO water has a stronger etching effect than H₂SO₄ solution, and NaCl EO water has a stronger etching effect than HCl solution. Furthermore, Na₂SO₄ EO water has almost the same etching effect as NaCl EO water. In addition, compared with Na₂SO₄ EO water, the change in etching rate for NaCl EO water is slightly larger as the electrolytic current increases. And compared to chemicals, EO water has a weak selective etching effect. As the immersion time increases, the etching rate for both Na₂SO₄ EO water and NaCl EO water decreases slightly. Next, an experiment was conducted to examine the roughening effect on the rolled surface of Alloy 42. The results showed that compared with H₂SO₄ solution and HCl solution, Na₂SO₄ EO water and NaCl EO water has a stronger roughening effect on the rolled surface of alloy 42. By this study, we received the suggestion that Na₂SO₄ EO water and NaCl EO water could be applied to the surface modification processing of alloy 42 for lead frame.

As part of efforts to address the issue of fairness in computer vision, image datasets are being reviewed, and inappropriate datasets are being temporarily suspended or withdrawn from public availability. What impact do pre-trained models using large-scale image datasets have on downstream tasks from the perspective of fairness? In this paper, we quantitatively evaluate how the pre-training methods, including manually supervised learning (MSL) and self-supervised learning (SSL) pre-training with ImageNet, affect downstream tasks in terms of fairness. We reached three findings from our experiments, summarized as follows: (i) SimCLRv2 and MoCov2 pre-trained models perform more fairly than MSL pre-trained models with ImageNet on IMDB-WIKI, FairFace, and CIFAR-10S. (ii) The MoCov2 pre-trained model achieves better performance in accuracy and fairness metrics than the SimCLRv2 pre-trained model. (iii) In MoCov2, color jitter tends to improve the fairness metrics for downstream tasks. By considering these findings, we demonstrate that SSL pre-trained models have the potential for fairer image recognition with similar accuracy compared to the MSL pre-trained model. By replacing human annotations with self-supervision, we can construct fair pre-trained models does not depend solely on human-annotated labels in the context of pre-training.

Reinforcement learning is an unsupervised learning method that enables an agent to learn its behavior via interaction with the environment. By maximizing the value that represents the expected reward over a certain period of time, the agent can learn to perform the required action. To obtain a high value, selecting the optimal action in an unknown future state is necessary. If an unknown future state can be predicted in advance, better actions can be performed. Therefore, obtaining a high value as a result is possible. In this study, we use a deep learning-based future image generation model to predict unknown future states in advance. By predicting the future state, selecting actions that lead to a higher value is possible. Thus, higher rewards can be expected at an early stage.

In our previous work, we developed a joint-independent kinematics method which is easy to solve an inverse kinematics for a variety of joint structures and verified it with robot arms of serial robot structure. However, when applying it to robot arms with redundant and complex structure, we need to solve a multiple-point position-control problem because we should satisfy multiple constraints at the same time, which we have not solved yet. In this paper, we develop a method to apply the joint-independent kinematics method to the multiple-point position-control problem and verify it with numerical simulations.