Flexure hinges are widely used for very high precision mechanisms, such as a high precision linear guide, a nano positioning stage, a robot hand and etc., because of no gap and no friction. Regarding circular hinge stiffnesses, Paros and Weisbord derived exact solutions, but they are unwieldly because of the complexity. Simpler equations are preferred to use in analyses of the assembled mechanisms composed of flexure hinges. Therefore, in this paper, approximate stiffness equations for the semicircular hinge are discussed. The approximate equations were also derived by Paros and Weisbord, but their tensile stiffness equation does not give a good result. And, many studies were performed for the circular hinge, but their stiffness equations are apart from the exact solutions when the hinge thickness is large. In this paper, new approximate stiffness equations are derived by substituting approximate functions of series for the small terms of the exact solutions. Errors between the delivered approximate solutions and the exact solutions are less than 3 percentages in the wide range of the hinge thickness. The derived approximate stiffness equations give better results than other approximate equations in the wide range of the hinge thickness. And, they are useful for the actual design because it is easily recognized how design parameters contribute to the stiffnesses.

In this study, best-fit fatigue curves are constructed considering material properties and loading modes. Three common fatigue test methods (rotating bending, axial load-controlled and strain-controlled) are compared, and their scopes of applications are clarified. The best-fit fatigue curves are given as a function of tensile strength and have been constructed based on the strain-controlled fatigue test data. The comparison of the best-fit fatigue curves with the axial load-controlled test data reveals that both deviated from each other for the carbon and low-alloy steels and the austenitic stainless steel. Therefore, the best-fit fatigue curves are reconstructed considering the data by the rotating bending fatigue test and the axial load-controlled fatigue test. In the case of the carbon and low-alloy steels, the best-fit fatigue curves are classified by two kinds of microstructures. In the case of the austenitic stainless steel, the best-fit fatigue curves are classified by two kinds of loading modes. Finally, design factors based on the reliability of best-fit fatigue curves are specifically presented.

This paper reports the relationship between the length of panel side and the deflection of the sandwich panels with corrugated core. The deflections of the sandwich panels having the lengths of panel side from 100 to 1300 mm and the height of the panel of 13.6 mm were determined by finite element method, where the four sides of the panel were rigidly fixed and a uniformly distributed load was applied to the top surface of the panel. The maximum deflection at the center of the panel increases with increasing the length of panel side, confirming that a power law is established between the length of panel side and the deflection. The scaling exponent of the sandwich panel was about 3.3 and this value was smaller than that of the single plate, i.e., 4. The ratio of the weight of the sandwich panel to that of the single plate having the same length of the panel side and the different thickness, which becomes same deflection compared to the sandwich panel, was found to be smaller as the length of the panel side becomes longer. This effect of weight reduction was possible to be estimated analytically. Smaller value of scaling exponent of the sandwich panel compared to that of single plate was found be related with that the normal force applied to the panel was effectively converted to the shear force at the bonding interface between the core and the bottom plate.

Electromagnetic acoustic resonance method (EMAR) is promising for online monitoring of pipes wall thinning with corrosion during operation. However, the improvement of measurement accuracy is one of issues to be solved, although some signal processing methods such as superposition of nth compression method (SNC) have been proposed in order to improve the accuracy of wall thickness estimation. In this study, for the purpose of highly accurate evaluation of the wall thickness, the thickness of 51 corrosion specimens was measured and compared with the true thickness obtained by 3D laser scanner. The fundamental resonance frequency of EMAR signal spectrum was evaluated by discrete Fourier transform method (DFT) and SNC method, and the wall thickness was calculated. As a result, evaluated thickness by EMAR almost corresponded to the average of true thickness. The accuracy of the evaluated thickness by the SNC method depends on the method to set the postulated thickness for determining the number of compressions. It was found that using the evaluated thickness by the DFT method for determining the number of compressions improved the estimation accuracy of the evaluated thickness by the SNC method. In addition, it was confirmed the estimation accuracy of this evaluated thickness was higher than that of the evaluated thickness by the DFT method.

Dimensionless numbers, which are frequently used in fluid dynamics and thermal engineering, are important to generalize transport phenomena. Although dielectric heating of microwave has been attractive to enhance effectiveness of various processes in last decades, dimensionless number has not been introduced. Therefore, the difference between thermal effects and non-thermal effects has not been clear. In this study, by using our dimensionless number (As number), which was proposed as an index of non-equilibrium local heating of microwave in previous study, and the usefulness of our dimensionless number was verified through boiling and quick responses of interfacial tension. Experiments of interfacial tension and boiling were carried out when water phase includes surfactant and different NaCl concentration, and it was found that higher As number is required for surfactant desorption from the interface in the case of the higher NaCl concentration. Finally, boiling behavior by microwave local heating can be predicted from the double-logarithmic plot of the As number and the accumulation energy.

To investigate the dynamic behavior and flame acceleration of spherically expanding hydrogen flames under the low-pressure conditions, the deflagration experiments were conducted in a closed combustion chamber. The dynamic behavior of hydrogen flames was observed by high-speed Schlieren imaging, and the flame radius and propagation velocity were measured by analyzing the Schlieren photography. When the flame radius was sufficiently small, smooth flame surface was observed, where flame stretching affected strongly the propagation velocity. From the correlation between the flame stretch rate and the propagation velocity, we estimated the Markstein length and Markstein number. When the flame radius was large, on the other hand, cellular surface was observed and the flame acceleration was confirmed. As a result, the critical flame radius corresponding to the occurrence of flame acceleration and the increment coefficient of propagation velocity were obtained. Under the low-pressure conditions, the dynamic behavior of hydrogen flames became weaker and the increment coefficient of propagation velocity became smaller. This was because the intrinsic instability was weakened by increasing the preheat zone thickness. It was also found that the increment coefficient became larger at lower equivalence ratios, which was because the diffusive-thermal instability became stronger. Moreover, the effects of preheat zone thickness on the flame acceleration was elucidated. We obtained the correlation between the normalized increment coefficient of propagation velocity and the preheat zone thickness. Based on the characteristics of dynamic behavior of hydrogen flames, the parameters of flame acceleration model depending on the initial pressure was obtained, and then the flame propagation velocity under the low-pressure conditions were predicted.

Automated driving is attracting attention as a technology that can improve the convenience and comfort of mobility. Research on automated driving continues to develop not only in the approach of vehicle automation but also in the approach of driver assistance. As an example of driver assistance technologies, a system has been proposed in which multiple light sources are placed at equal intervals in a tunnel and blink to encourage drivers to adjust their speed. This system is called a Pace Maker Light (PML). In this paper, we focus on the PML and consider applying it to automatic driving. Therefore, this paper aims to realize the velocity control of an autonomous mobile robot using the PML. First, we modeled the PML and the mobile robot and formulated the design problem of the velocity controller of the robot. Then, we proposed a control law that can achieve velocity control based on the light intensity information obtained from the PML. Finally, we conducted velocity control simulations and experiments with a miniature-scale experimental setup. As a result, it was confirmed that the proposed method could be used to achieve the target speed on a level road and a slope.

This paper describes a body weight reducing walking support system using pneumatic artificial muscles. The user of the system walks while sitting on a saddle. The saddle supports the weight of the upper body and reduces the load on the lower limbs. The saddle moves up and down according to the user's gait by force control using pneumatic artificial muscles, and always performs a constant body weight load reducing action while walking. To confirm the effect of this body weight load reducing motion on walking, a walking support system was constructed, and floor reaction force, gait, and muscle activity were measured through experiments. As a result, it was found that the body weight load reducing motion contributed to the reduction of floor reaction force and some muscle activities. It was also found that varying the amount of the assist force did not significantly affect gait.

Virtual sensing is a method to estimate unobservable or unmeasurable state variables by combining exsisting measurement data with model information. When FE model is applied to the virtual sensing, vibration response data estimated at all nodes. When it enables to estimate the vibration response data at all nodes, for example, it will be possible to simulate sound radiation from mechanical structure with the vibration response data. In this paper, it is illustrated how to estimate high accuracy vibration response data. It then examines the effect of excitation force on the vibration estimation results. Based on the results of this verification, a method to improve the accuracy of vibration response estimation is proposed. It also tests the relationship with accuracy of estimated vibration data and the difference between real and estimated location of excitation force. Finally, it tests how the quality of estimation for sound radiation.

The geometry of truss structure is expressed in terms of its nodal positions; that is, the nodal positions vector is the fundamental design variable for the optimal geometry design of a truss. In order to solve an optimal design problem, there are various approaches that uses differentiation of the objective function in terms of the design variable, such as the steepest descent method and the gradient projection method. Many of the objective functions of structural optimal design problems are formulated in terms of the stiffness matrix; the gradient or differentiation approach requires differentiation of the stiffness matrix in terms of the nodal positions vector. It is, however, not impossible but still impractical in many cases since we have to deal with the third-order differential coefficient tensor. In this paper, we deal with the differentiation of the product of the stiffness matrix and a vector in terms of the nodal positions vector. We develop a formulation in general form that can be applied to trusses of any configuration. Examples of optimal geometry designs of 2D and 3D trusses are demonstrated.

An experiment was conducted to proove that a single-point cutting tool made of nano-polycrystalline diamond with a cutting edge radius of 0.3 nm can fabricate a cut face without damaged layer by decreasing the depth of cut. A face turning test of oxygen-free copper showed that an actual surface roughness of 2.4 nmRz which is close to the theoretical value of 2.0 nmRz could be achieved when the depth of cut and the feed rate were set at 0.2 μm and 2.5 μm/rev, respectively. In addition, it was observed that surfaces of the removed chip at a thin part of a free face were smooth and showed no lamellar waviness. The minimum thickness of cut, from which a comma shape chip was generated, was calculated from the measured chip width. This minimum thickness of cut was calculated to be 0.31 nm, which is equivalent to the lattice constant of copper. The kernel average misorientation map of the cut face obtained by electron beam backscatter diffraction clarified that almost no lattice defects were formed in crystal grains when the depth of cut was set at less than 1 nm. This means that the dislocated area generated in front of the cutting edge was negligibly small and the cut face without damaged layer could be fabricated.

Anomaly detection for predictive maintenance in the cutting process is one of the challenging problems in shop-floor management. A modern machine learning approach, including deep learning, has been widely studied for the last decade. This study focuses on the multimodality of various cutting time-series data for extracting features of cutting status and proposes a multimodal variational autoencoder (MVAE) method. We collect a time series of vibration acceleration of a cutting tool and a main spindle motor load. Various cutting data is collected by conducting cutting experiments under diverse cutting conditions. Normal and abnormal data are collected, and only normal data is used to train MVAE. MVAE learns a so-called generative model, which is implicit but stochastic, capable of reproducing original time series data. Because MVAE is an unsupervised learning method, it does not require abnormal data during training. Therefore, it is considered suitable for tools management where it is difficult to collect abnormal data. Euclidean distance is employed to evaluate the normality of a given cutting status on the latent space acquired by MVAE. We demonstrate the applicability of the proposed MVAE method in anomaly detection for end-milling by comparing it with conventional machine learning methods such as autoencoder.

In the railway sector, efforts are being made to improve the efficiency of operations. One of these is the automatic operation of trains. To keep the train running safely, it is necessary to make sure that there are no obstacles in front of the railway vehicle. Therefore, automatic train operation requires obstacle detection system in front of the railway vehicle. In recent years, some railway operators have introduced onboard cameras as drive recorders in order to understand the situation at the time of an accident and to investigate the cause. In addition, efforts are being made in the railway sector to improve operational efficiency through the use of onboard cameras. In this paper, we propose an obstacle detection method using an onboard stereo camera. The proposed system can be divided into three parts. First, the center position of the track is estimated using RGB images, and the search area for obstacle detection is set. Second, proposed system checks for the presence of objects within the set search area and creates a candidate area of obstacle. Finally, the area of each neighboring candidate area of obstacle is calculated, and if the area is greater than a certain value, the object is determined to be an obstacle. Furthermore, the proposed system is verified by using a 1/10 scale model experimental vehicle. Validation results show that the proposed system is capable of detecting obstacles in front of railway vehicles.

In recent years, research on occupant behavior models has been promoted for vehicle motion control that includes occupant comfort as a control target. In previous research, modeling and verification focusing mainly on behavior, such as occupant head motion, has been conducted. But in order to accurately predict occupant comfort with a model, it is necessary to model and verify not only the behavior but also the joint torque required to maintain posture. When the occupants maintain their posture against the inertial force associated with the motion of the vehicle, they utilize the seat reaction force with the joint torque. So, in order to accurately model the joint torque, it is necessary to measure and model the joint torque and the seat reaction force separately. Therefore, in this research, we tried to make a occupants behavior model and parameter identification method that predicts joint torque and seat reaction force separately. We focused on the occupant’s head and torso motion among the whole body, and these are modeled in the form of an inverted double pendulum. And the seatback reaction force is set as an input in order to accurately predict the joint torque exerted by the occupants. Then, by utilizing existing occupant behavior analysis technology and we measured joint torque and seat reaction force in addition to the behavior of seated occupants and identified each parameter of the model. The model is experimentally verified, and as an example of the use of the proposed method, experiments and parameter identification are performed on occupants with different gaze points, and the changes in occupant’s joint torque control when view conditions are different were quantitatively clarified.

The critical wind speed of overturning is one of indices for the safety evaluations of railway vehicles against crosswind. The critical wind speed of overturning is defined as the wind speed when wheel loads on the upwind side becomes zero, and is often evaluated by the “RTRI’s detailed equation”. In the RTRI’s detailed equation, the lateral vibration inertia force of car bodies is considered as one of the acting forces affecting overturning. The inertia force can be obtained from the product of the lateral vibration acceleration and the mass, and the lateral vibration acceleration is assumed from past running test results. However, there are differences in position, direction and situation between the lateral vibration acceleration measured in the running tests and the lateral vibration acceleration in the calculation model of the RTRI’s detailed equation. In this paper, model experiments in a wind tunnel were conducted to clarify the effects of the differences with respect to the lateral vibration acceleration. Next, vehicle dynamics simulation was implemented to calculate the lateral vibration acceleration that occurs at the center of gravity of a car body under strong crosswinds. The results showed that the critical wind speed of overturning can be calculated accurately by using the simulation results of the lateral vibration acceleration that occurs at the center of gravity of the car body under strong crosswinds. In addition, a new method is proposed to estimate the lateral vibration acceleration using the track displacement data. Finally, when the critical wind speed of overturning was calculated assuming typical commuter vehicles, the wind speed obtained from the proposed new method was up to 2m/s higher than the wind speed obtained from the assumed equation in the RTRI’s detailed equation.