Search

A new paper sheet feeding mechanism using a friction force owing to electrostatic adhesion is proposed to realize long-term reliability and reduction in maintenance work. In this mechanism, friction force owing to electrostatic adhesion of an electrostatic pad acts as a resistance force to make paper sheets separate successively instead of a friction force of rubber parts used in the conventional feeding mechanism. The electrostatic pad with comb-type electrodes and whose upper surface is coated with an insulator with high wear resistance generates the resistance force. First, the width of the electrode and the spacing between electrodes of the electrostatic pad that generate a large electrostatic adhesion at the lowest possible voltage are determined from an experiment using the evaluation electrostatic pad. Then, the feed force acting on the first and the second paper sheets, and the resistance force owing to the electrostatic pad are measured by pulling the paper sheets in the separation mechanism. Furthermore, the performance of the paper sheet feeding mechanism is measured while changing the value of the voltage applied to the electrostatic pad. Consequently, it was found that the paper sheets could be accurately fed successively by adjusting the applied voltage to an appropriate value.

On the flaming combustion of highly densified cypress cylindrical briquette of which diameter/length ratio is 1/2 in the high temperature air, it was reported that the flaming combustion duration is proportional to the product of the inverse square of the specific surface area and the briquette density. However, the influence of biomass raw material on the combustion behavior is not clarified yet. To utilize various biomass as raw material of briquette fuel, the knowledge on this should be obtained. The objective of the present study is to confirm the influence of biomass raw material on the flaming combustion behavior, especially, the flaming combustion duration. Cypress, bamboo, rice husk and paper are used as raw material of the briquette. From the results of the thermogravimetric analysis and the estimation of lignin content in the biomass samples, it was observed that the activation energy decreases with increasing the lignin content. From the results of combustion experiments, it was found that the flaming combustion durations for each biomass raw material are proportional to the product of the inverse square of the specific surface area and the briquette density. It is suggested that these proportional constants have the correlation with the mass percentage of the cellulose in the biomass.

Argon ion sputter-etching of SUS420J2 and SUS316 stainless steels was carried out at a power of 250W for 10.8ks to form cone-shaped sharp protrusions with bottom diameter of 10-30 μm and fine quasi-column-shaped protrusions with diameter smaller than 500 nm respectively by using a radio-frequency magnetron sputter-apparatus. Effects of two types of plasma-nitriding on gripping ability were examined; one is the nitriding of both steel specimens using nitrogen gas of 0.53 Pa mixed with argon gas of 0.67 Pa at a power of 50W for 1.8 ks to maintain the original sharpness of protrusions, and another is the nitriding of SUS420J2 steel specimen using only nitrogen gas of 1.2 Pa at a power of 200 W for 7.2 ks to obtain the cone-shaped protrusions with round top. The coefficient of static friction, or gripping ability, of the plasma nitrided sharp protrusions of SUS420J2 steel specimen to polyethylene, polyvinyl-chloride, polyethylene-naphthalate sheets and a copy paper was about 1.8, 2.4, 1.1 and 1.5, respectively, at a large nominal compressive stress of 18 kPa. The reason for such large frictional coefficients is due to localized deformation of sheet under the sharp protrusions or piercing of the protrusions into the sheet. Although the sharp protrusions were partly broken, the frictional coefficients of these sheets were still about 1.6, 1.8, 1.1 and 1.4 after the second series of tests, which are more than twice as large as that of the specimen ground with #100 emery paper. The plasma-nitrided round protrusions of SUS420J2 steel and the fine protrusions of SUS316 steel were not broken during the friction tests of the polyethylene sheet and the copy paper but the frictional coefficients are smaller than those of the sharp protrusions. The protrusions with large frictional coefficients can be applied to the surface of a frictional conveying roll.

Ball bearings are used for many mechanisms. For the precious linear motion mechanism, it is very important to know the ball bearing deformation because the driving load is supported by the bearing. Especially, for the focus control mechanism of the optical telescope, even a very small deformation is not negligible because a sub-micron precision is required for the driving stroke. The Hertz’s formula is very familiar to the relation between the load and the displacement in a small contact area. Hertz derived the formula by using a potential of the electricity distribution in the ellipsoid body. However, it is reported that an error becomes large when the contact area is large. It is a reason that the elliptic paraboloid is used for the ellipsoid body in the Hertzian contact theory. Some modified equations are reported by Nishihara et.al, Tatara and Villaggio. However, the modified equations by Nishihara et al. and Villaggio are a little different from the FEM result. Therefore, in this paper, a contact deformation for the 2 equivalent spheres is analyzed by the indirect-fictitious boundary integral method. For the relation between the load and the displacement, a new modified equation with the fractional expression is derived in this paper. And, for the relation between the load and the contact radius, it is confirmed that the Hertz’s formula is correct.

In this paper, we present a void topology identification analysis in structures based on the level-set-based topology optimization using a weighted sensitivity method. The void topologies are identified by using the observed hammering response data and the gradient of the Lagrange function with respect to the level-set function. The numerical simulation of structural oscillation is performed by the finite element method, and the void topology identification analysis is formulated by the adjoint variable method. An improvement of how to give numerical parameters in the weighted sensitivity method is conducted, and numerical tests are carried out for minimization problem of the Rosenbrock function. The improved procedure is applied for the void topology identification analysis, and applicability of the proposed method is shown in this paper.

The authors developed a vibration suppression device (VSD) that has variable moment of inertia using magneto-rheological (MR) fluid in order to get series inertia mass effect for vibration reduction. In previous paper, the proto-type VSD was manufactured and the variable series inertia mass effect was confirmed. From the experimental results, the series inertia mass effect can be switched by the current of electromagnets. And theoretical model of solidified form of the fluid was proposed by formula and the trend were closed the experimental results. Then the vibration tests of 1DOF system with the VSD were conducted by a shaking table. Vibration reduction of the VSD was confirmed. The current of the electromagnet was controlled by using the reinforcement learning. But the fluid behavior inside the flywheel has not investigated before. In this paper, to investigate the fluid behavior inside the flywheel, it is analyzed by computational fluid dynamics (CFD). Three types of the flywheels are manufactured, and 2 types of electromagnets are used here. Equivalent inertia mass in totally 20 cases of the flywheel and the electromagnets are numerically calculated by using the OpenFOAM, and also measured in experiments. From the analytical results, convection is initially occurred in the direction of the magnetic poles when the magnetic field applies, and then the particles are eventually solidified at locations other than near the magnetic poles. Finally, the analytical results are compared with the theoretical and the experimental results.

In general, the life of a rolling bearing is calculated as the total number of rotations until rolling fatigue occurs on the raceway. However, when rolling bearings are used under light loads and high rotational speeds, they are damaged not by rolling fatigue on the raceways but by cage wear due to lubrication life. In such cases, the conventional method for calculating the rolling bearing life is not applicable. Rolling bearings for traction motors of railway vehicles (hereafter referred to as traction motor bearings) are one example of this and are replaced primarily due to lubrication life. However, there is no practical way to calculate lubrication life and these rolling bearings are replaced empirically. The purpose of this paper is to obtain basic data to create a method for calculating lubrication life. Therefore, traction motor bearings were rotated under conditions like actual operating conditions to reproduce damage due to lubrication life. As a result, damage due to lubrication life could be reproduced, so the next step was to investigate the effect on lubrication life when the operating conditions of traction motor bearings were varied as parameters. The parameters considered were “bearing temperature”, “vibration acceleration”, and “grease mass”. As a result of these verifications, lubrication life decreased exponentially when the bearing temperature was increased. In addition, lubrication life decreased exponentially when vibration acceleration was increased. Furthermore, lubrication life increased logarithmically when the grease mass was increased. The effects of these verified parameters on lubrication life could all be formulated within this paper.

This paper presents a data augmentation method for generating a surrogate model of numerical analysis results. The proposed method focuses on the relative frequency of learning data for generating a learning model using deep learning techniques. Generally, data augmentation techniques are known to be useful for improving prediction accuracy. Adding noise and data duplication are commonly used for predicting numerical simulation results, but it is essential to carefully consider the amount of noise or choose a duplication target. However, these techniques are not appropriate for generating surrogate models. The reason is that the numerical analysis results mostly have high data imbalance, and no specific solution has been presented. The method proposed in this paper solves this problem and aims to be a simple and highly versatile data augmentation method. This paper describes the application of the proposed method to predict two-dimensional stress fields. It was confirmed that by increasing the number of data augmentations using the proposed method, the prediction errors were reduced for three different stress components stably. Additionally, it was confirmed that the prediction accuracy improved 5.81 to 27.0% compared to that of the data augmentation by simple duplication.

This paper describes a method for continuously measuring the contact position between the wheel and rail using strain gauges. The authors investigated the relationship between the bridge output sensitivity rate and the change in the placement of strain gauges using the result of finite element analysis for the instrumented wheelset. It is noted that the bridge output sensitivity rate is defined as the bridge output change per unit distance of the contact position (load position). Based on the result, strain gauges position suitable for measuring contact position was identified. On the other hand, static load tests were conducted on the instrumented wheelset with strain gauges attached to measure contact forces (wheel load, lateral force, and tangential force) and positions, respectively. The test results were found to be generally consistent with the finite element analysis results, and it was also observed that the load position of each contact force influenced each bridge output. Finally, this paper proposes a method for calculating the contact position by solving a nonlinear equation. The calculation uses the results of the load tests and also wheel rotation angle, bridge outputs, wheel load, lateral force and tangential force which measured in running railway vehicle. Simulations were performed for various cases, and the contact positions were obtained stably and continuously. This suggests that the proposed method is reasonable and applicable to various situations.

This paper presents an analysis of the motion of a motorcycle at high-speed using a motorcycle dynamics model that considers the compliance of the frame components. Several motorcycle nonlinear dynamics models have been proposed in previous studies and in the industry. However, these models do not consider the compliance of the frame components, which affects the maneuverability and stability of motorcycles. Moreover, there is not any researches that analyzes the effect of such compliance on motorcycle dynamics in a domain that includes turning. In this paper, we derive a nonlinear dynamics model of a motorcycle that takes into account the compliance of the frame components and analyze the effect of such compliance on the vehicle body motion during high-speed straight riding and turning. For the dynamics model, the Kane’s equation is constructed based on multi-body dynamics analysis. In conclusion, we demonstrate that the compliance corresponding to the lateral torsion of the front forks is the most dominant factor in improving stability in the high-speed region, both in straight riding and in turning.

Instrumented wheelsets are widely used in the railway industry for the purpose of measurement of the wheel-rail interaction forces, which are crucial factors in a running safety assessment. Information on lateral contact position between wheel and rail is also an important factor when discussing the wheel-rail contact conditions, such as the friction coefficient at the contact point. However, it is difficult to measure the contact position by using ordinary instrumented wheelsets since they are specialized only for the measurement of interaction forces. Therefore, the authors have proposed a signal processing procedure to extract information of contact position from bridge circuits constructed for the measurement of the lateral forces between wheel and rail by using bending and shear strains of wheel webs in a previous work. The purpose of this paper is to develop an online processing system which implements the proposed signal processing procedure. The original procedure requires long computational time and is not suitable for an online processing. As the first step to develop the processing system, this paper proposes a faster algorithm of signal processing procedure based on frequency analysis for extracting information of contact position from instrumented wheelset at first. Using the faster algorithm, a multi-rate processing system was developed to acquire the strain data at a high frequency and to execute the signal processing at a low frequency. The performance of the developed system was evaluated with a single-wheelset roller-rig test.

Vibration performance is the one of the important performances in automotive product because of related to not only ride comfort and control performance of vehicle but also durability and noise performance. However, it is difficult to design efficiently because the performance is determined by the influence of all components constituting of a whole structure. In previous paper, we presented an effective way to design vibration performance based on an inverse method for structural modification that keeps one specified eigenfrequency and its eigen vector of the whole structure the same. In this paper, we propose a new method to keep the specified multiple eigenfrequencies and their eigen vectors the same. Firstly, the matrices for structural modification are calculated as dynamic stiffness matrix variation from the zero-divisors of the specified multiple eigen vectors. The design of dynamic stiffness matrix variation is simplified by representing the matrix variation as just stiffness matrix variation which is independence on eigen frequency. Furthermore, it is possible to represent the calculated matrices redundantly by using an arbitrary non-zero weighting matrix. However, the matrices are generally calculated as fully populated and non-symmetric matrices. Therefore, secondly, we transform the matrices calculated as the zero-divisors into the sparse matrix of reduced row echelon form in advance. This transformation simplifies building symmetric and sparse matrices for realizable structural modification. This sparse matrix, redundantly represented by a weighting matrix, allows a value analysis to select among alternative structural changes to simplify structural complexity while keeping several performances originated from vibration performance. Finally, the proposed method was applied to a numerical case study.

This paper proposes a novel path planning algorithm designed for efficient lunar water exploration. The proposed algorithm leverages a Bayesian network-based sensor model in combination with the RRT* algorithm. It is well-known that water is highly likely to be present in permanently shadowed regions (PSRs). However, during exploration within these regions, the rover must escape to illuminated areas within a limited time to recharge its battery. Therefore, it is crucial to develop a path planning algorithm that ensures both timely battery recharging and efficient water exploration. To address this challenge, we first introduce a sensor model based on a Bayesian network for the neutron spectrometer. Subsequently, we propose an algorithm that utilizes this sensor model and the RRT* algorithm to effectively explore water on the lunar surface. The paper provides a numerical simulation, using real data about time-variant illumination areas and time-invariant PSRs on the lunar surface, to illustrate the effectiveness of the proposed algorithm. The results reveal that the algorithm accurately and efficiently detects water under realistic illumination conditions.

To support engineers in solving engineering problems utilizing digital twin, we are proposing ‘digital triplet’ concept. We are also developing a problem-solving process support method by using the Log Level Description (LLD) that describes an executed problem-solving process and the Generic Process Model (GPM) that describes a generalized and prescriptive process. This paper illustrates a support method for describing LLDs to put the proposed method into practical use. Through an attempt to describe LLDs, three time-consuming operations were identified: viz., interviewing, structuring the information, and relating each action with the GPM. To save the labor of these operations, we propose the method that the engineer who executes the process sequentially describes the LLD by oneself based on the GPM. This paper also illustrates an LLD description support system based on the proposed method. Finally, a case study was conducted to describe the actual fault detection system design processes with the developed support system. By comparing the LLDs created with the support system in the case study with the ones created manually, the effectiveness of the proposed method was discussed. In addition, by interviewing engineers who created the LLDs in the case study, the usefulness and feasibility of the proposed method were discussed.

In recent years, there has been active development of robots that play a role in people's lives. These robots require actuators that have high responsiveness and high output force in addition to being small. Furthermore, safety and contact affinity are also required for these actuators. An actuator that can balance high responsiveness, high output, and contact affinity is developed in this study. The proposed system is a hybrid system of hydraulic and pneumatic systems. Hydraulic systems are generally used in machines that require large forces, such as construction machinery, because they have high responsiveness and high output force. However, hydraulic systems have poor contact affinity. On the other hand, pneumatic systems are used in machines that come into contact with people, such as power assist suits and massage chairs, because they have high contact affinity, although their output is inferior to that of hydraulic systems. Generally, the mixing of air bubbles into hydraulic oil is a problem. However, as the pressure increases, the rigidity of the working hydraulic oil mixed with bubbles approaches that of pure working hydraulic oil. If this characteristic is appropriately modeled and controlled, bubbles mixed into the working hydraulic oil can create a cushioning effect. Firstly, the relationship between the pressure and rigidity of hydraulic oil mixed with bubbles is demonstrated in this paper. Secondly, the mechanism of the hybrid system and the prototype model are explained. Lastly, the results of performance evaluation through analysis and experiments are presented. From these results, this paper shows that the proposed system has the potential to balance high responsiveness, high output, and contact affinity.

All-terrain wheelchairs are assumed to be used under such conditions that are on the roads along mountain topography, steep slopes, bumpy roads, muddy roads, etc. It is often used for people with disabilities to enjoy outdoor activities in nature with friends and family. In order to deal with such conditions and usages, all-terrain wheelchairs have different structures and features from normal wheelchairs. In this paper, we focus on the type of all-terrain wheelchair that needs to keep the pitch posture during turning. Here, the pitch posture means the state that the front wheel is lifted. It is hard to keep a stable pitch posture on a steep slope and a bumpy road even though the wheelchair is manipulated by an experienced operator. This paper proposes an electric power assist control that keeps this pitch posture stable and reduces the manipulating load of an operator. The most important feature of the control is to control the kinetic energy of the wheelchair, the sum of the energies of longitudinal and pitch motion. This control is implemented on an electrically power assisted wheelchair, and the effectiveness of the control is demonstrated by the experimental results.

In this paper, the large deflection response of a circular mesh tube under pure bending was investigated by using FE analysis. In the previous study[Nakamura, T., Ushijima, K., Evaluation of large deflection response of circular mesh tubes, Transaction of the JSME (in Japanese), Vol. 88, No. 915 (2022)], the effect of the ovalization across the cross-section on the bending response was neglected, but in this study, the influence of cross-sectional flattening on the nonlinear bending moment versus curvature was taken into consideration. There are three reasons for arising the maximum bending moment, but in this paper, we mainly focused on the two mechanisms (the large flattening response in elastic deformation and the plastic yielding behaviour) to the maximum bending moment. Based on our FE analysis, we proposed theoretical equations for predicting the maximum bending moment as a function of tube geometries and material parameters such as Poisson’s ratio and the plastic yielding stress.

There are several studies focusing on rapid tapping motion for a robotic finger, including not only a theoretical approach of musculoskeletal modeling of a human finger but also an empirical challenge with robotic finger. Shourijeh had made a success of 6Hz-periodic tapping motion by an index finger model for forward dynamic simulation, however the contribution is restricted to analyzing the validity of a musculoskeletal model of finger and is based only on simulation of not a force control but an angular control of the finger. This paper, therefore, presents a novel three-DOF robotic finger that enables fast tapping motion applicable in piano playing and keyboard typing, which consists of metacarpophalangeal(MCP), proximal interphalangeal(PIP), and distal interphalangeal(DIP) joints. In previous study, we developed Twisted Round-belt Actuator(TbA) capable of producing contraction force by twisting a small-diameter round-belt having high elasticity, and empirically derived a contraction force model applicable only to two-DOF robotic finger. This study, therefore, derives such force model available for an extended three-DOF finger mechanism that is containing Variable-pitch Screw Module(VpSM) newly designed in this paper. This module is able to eliminate irregular twist phenomenon resulting in unexpected discontinuous movement in finger joints. Thus, we first describe the detail mechanism of the three-DOF robotic finger and indicate great advantage of the finger, which is obtained by combining the TbA and VpSM. Finally, we demonstrate a new force feedback method based on fingertip force estimation without any force/pressure sensors, in which the fingertip force can be accurately estimated by using mathematical model of the contraction force by TbA with VpSM. In addition, this paper shows successful experimental results of fast tapping motion by the robotic finger, and clearly reveals that the robot is able to move against sinusoidal force reference up to 5Hz interval.

In this paper, we propose a spatial measurement method based on object geometry and velocity information using past measurement information from 3D-LiDAR. If geometric information is obtained from 3D-LiDAR, it can be applied to global localization, object identification, etc., and robots will be able to run autonomously with greater stability. However, 3D-LiDAR cannot always measure point densities high enough to acquire sufficient geometric information. Conventional methods have been used to control point density by interpolating areas of low point density or by varying the rotational scanning speed of 3D-LiDAR. However, these methods have problems such as the shape of the measured object differing from the real object and the loss of real-time measurement. In this paper, therefore, a remaining time is set for each point based on the shape information of the object, and each point is allowed to remain until the remaining time is reached, depending on the robot’s movement. The remaining time is set so that the point density is preferentially increased in areas where shape information is estimated to be abundant and in distant areas where the point density tends to decrease due to the LiDAR mechanism. For dynamic point clouds, velocity information is added, and the point cloud is moved to the estimated position the next time and superimposed on the latest dynamic point cloud while considering the distance between points, thereby improving point density while suppressing shape collapse. We demonstrate the usefulness of this method through experiments on actual equipment.

Overhead cranes are widely used to transport heavy loads, and their efficient operation requires fast and accurate positioning. However, residual vibration which prevents positioning accuracy generally tends to be induced by fast transportation. In previous paper, we proposed a control method which is based on the characteristic that the residual vibration is completely suppressed in a linear undamped system when excited by an external force which does not contain the natural frequency component of the system. To apply this characteristic to nonlinear damped systems, we defined the apparent external force which includes the influence of system nonlinearity and damping. It was confirmed that the residual vibration of a double pendulum type system model of crane can be suppressed by eliminating two natural frequency components from each the apparent external forces. In this paper, in addition, to improve the robustness for a double pendulum type system against the estimation error, new conditions are added to reduce the component around the natural frequencies from each of the two apparent external forces. The effectiveness of these conditions for robustness is verified through numerical simulations. Furthermore, changes in the robustness are examined when the position of the mass point near the support point of the double pendulum is changed. The results showed that the robustness can be improved in many cases except when the control time is short and the mass point is close to the support point.

Press forming is a mass production method that is used to manufacture many machine components such as automobiles, industrial machinery, and various plants. However, there is a limit to the degree of shape freedom that can be applied, so countermeasures against cracks, wrinkles, spring-back, etc., must be taken in the development process. It is well known that it is difficult to apply it economically in the case of small-lot production because of its high capital investment. In a previous report, the author reported that the folding line processing by press method / bending has a high degree of shape freedom, requires little capital investment, and can be expected to be applied for experiments and prototypes from small to mass production. This paper shows that press folding line processing using V-shaped punch can be applied as deformation induction based on experiments and FEM results. To improve the impact energy absorption performance in the event of a collision, notches and uneven surfaces called beads have been formed in the crush boxes and front side members as automobile parts manufactured by press forming. The bead has the function of inducing deformation so as to improve the impact energy absorption performance during the axial crushing process. However, bead setting for inducing deformation by the press forming, in addition to the difficulty of mold design/manufacturing, needs to correct the bead design after impact experiments using prototypes. This paper shows press folding line processing as deformation induction is very convenient and effective.

This paper proposes a method for a robot to recall multiple action candidates for an object by learning object manipulations based on observation of human actions. When learning, multiple answers to a single input in supervised regression manner, it is usually necessary to map all correct answers to the same input. However, only one action can be observed for an object at a time in observing object manipulations, and other possible actions are not always observed for the identical object. It is, therefore, important to automatically share various observed actions between similar-shaped objects by recognizing common shape cues among individual objects. The proposed method learns the code descriptions of object shapes by a variational auto-encoder (VAE) with an object image as input data, and the code descriptions of actions by a conditional VAE with object shape as a condition and an action as input data. Since the action is unknown recall target, it is desirable to obtain the code description of the action from only the object shape during recalling. The distribution of the code description of actions conditioned by input object shape on the obtained code description space is obtained by marginalization of the distribution learned by the encoder part of CVAE. However, since this marginalization is difficult to analytically and numerically operate, a deep regression model that “imitates” this marginal distribution is trained by using a maximum likelihood method based on sampling. Common actions of similar-shaped objects are shared among the similar objects in this “marginalization by imitation” process. Various possible actions for the input object shape can be recalled by repeatedly sampling from the imitated marginal distribution. This paper describes the results of experiment using actual object images and manipulation actions, and demonstrates the effectiveness of the proposed method.

Thin high-strength steel sheets are increasingly being applied to vehicle frames in order to increase the strength and reduce the weight. The thinner the plate, the more likely it is to occur buckling. So it is important to understand the structural properties based on buckling theory in order to efficiently design for buckling in the early stages of vehicle development. The authors have addressed the issue of theoretically determining the shear buckling stresses in torsion of thin-walled square cross-section beams. In the previous paper (Budiansky, et al., 1948) for determining shear buckling stresses in torsion of infinitely long thin-walled square beams, two adjacent plates were treated collectively, the buckling modes were expressed as an infinite series in the width direction and a single sinusoidal wave in the infinite length direction, then, the method of Lagrange multiplier was used to consider the constraint condition in the center line, finally the shear buckling stress was obtained by the energy method. However, this process cannot be applied to finite-length beams because a single sinusoidal wave cannot satisfy the boundary conditions at both ends in the longitudinal direction. So, the purpose of this paper is to theoretically determine the shear buckling stress in torsion of a thin-walled square beam of finite length. First, based on the previous process, two adjacent plates are considered to be a single plate, and the buckling mode is represented by an infinite Fourier series as in the shear buckling of a single plate. The shear buckling stress in torsion is obtained by the energy method, taking into account the constraint condition in the center line of the width direction of the plate by the method of Lagrange multiplier. The validity of the derived values is compared with the FEM results, and the results are in good agreement.

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 paper, we propose a simulation method to evaluate the packed state of catalyst pellets in a pipe reactor for gas phase reaction. This simulation method is based on dynamic explicit FEM (finite element method) and consists of three steps. Specifically, first, the catalyst pellets freely fall under gravity and flow into a packing hose with a hopper-like inlet. Then, after the catalyst pellets reach the bottom of the pipe reactor, the packing hose is pulled up to create a random packing of the catalyst pellets in the pipe reactor. Finally, the bulk density and stress distribution of the catalyst pellets randomly packed in the pipe reactor are calculated, and the packed state of the catalyst pellets is evaluated. Using this simulation method, we simulated the packing of cylindrical catalyst pellets into the pipe reactor and evaluated the packed state of the catalyst pellets in the pipe reactor. As a result, it is clarified that the maximum equivalent stress generated at the bottom of the catalyst pellets packed in the pipe reactor approaches a constant value as the ratio of the packed height of the catalyst pellets to the diameter of the pipe reactor increases. In addition, the effects of the shape of the catalyst pellets on the bulk density and average equivalent stress of the catalyst pellets in the pipe reactor are elucidated, and the optimal shape of the catalyst pellets that can increase the bulk density and reduces the stress is found. The optimally shaped catalyst pellets are expected to improve their packed state in the pipe reactor, extend the durability of the catalyst pellets and increase the efficiency of production. In this paper, it is assumed that the packed state and durability of the catalyst pellets are evaluated based on the bulk density and stress distribution of the catalyst pellets in the pipe reactor.

This paper deals with applications of data-driven control, which is a controller tuning method that directly utilizes the data, for mechanical systems. Particularly, we focus on the valve opening control of the motor-operated valve as one of the representative mechanical control units which are used in much industrial equipment and are needed to be precisely controlled. If we have a mathematical model of such a system, a model-based controller design is the most rational approach. Otherwise, data-driven controller tuning is also expected to be a rational approach. However, there exists a nonlinear characteristic, such as the limit of a speed implemented in the nonlinear compensator. Since conventional data-driven controller design can not be used because data-driven control is generally derived for only linear time-invariant systems, we have to modify the method of driven controller tuning method so as to be applicable for such systems. This paper proposes a controller tuning method for the valve-opening control system implemented as a cascade control system with a limit of the speed. Here, we apply Virtual Internal Model Tuning, which is abbreviated as VIMT, to tune the inner linear controller. Then, we also apply data-driven prediction to compute the outer loop controller including the limit of the speed. Finally, we verified the usefulness of the proposed method using experimental verification.

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 discusses explanatory variables that constitute the driver’s subjective risk level when encountering a pedestrian who suddenly pops out from the roadside. In this research, near-miss incident data, called “Hiyari-hatto data,” against pedestrian recorded by drive recorders equipped on taxis are used. First, we evaluate the subjective risk level of 90 near-miss incidents quantitatively in 6 levels and select the candidate of eight explanatory variables. Second, Principal Component Analysis (PCA) is applied to determine the principal components of the subjective risk level. Moreover, Multiple Regression Analysis (MRA) is used to compare the contribution of explanatory variables. From these analyses, some explanatory variables have become effective candidates. Time to collision (TTC) and its related variables might be the main explanatory variable and collision timing might be another explanatory variable. However, it is difficult to determine that collision timing is a sufficiently influential variable by PCA and MRA. Therefore, the latter of this paper examines the possibilities of collision timing (in our experiment, qualitative collision timing was evaluated using quantitative value as “collision gap” which is the difference between TTC and TTV) as a candidate explanatory variable by an actual vehicle test. To reveal the relation between subjective risk level and explanatory variables, in particular the collision gap, the driver’s several reactions are collected by an actual vehicle test in which a pedestrian suddenly pops out from the roadside to in front of the vehicle. Based on the about 900 case experimental data, we examined the possibilities of “collision gap” as an explanatory variable in subjective near-miss incident level.

In previous paper the authors successfully developed a vibration suppression device (VSD) that has variable moment of inertia using Magneto-Rheological (MR) fluid. The MR fluid is filled into a flywheel. Ferrite particles of the MR fluid are clustered when magnetic field is applied to the flywheel by electromagnets, series inertia mass of the VSD will be switchable. In this paper seismic tests during switching series inertia mass by reinforcement learning are focused in order to get more vibration reduction while random excitations. Firstly, agents are trained by the reinforcement learning using Deep Deterministic Policy Gradient (DDPG) algorithm while a one-degree-of-freedom system with the VSD, which has the switchable series inertia mass, is excited under earthquakes. Secondly, the VSD is improved for having 8 electromagnets, magnetic field is analyzed by FEM, resisting force characteristics and response performance of the VSD are investigated. Then, to confirm the vibration control under some earthquakes, seismic tests of one degree-of-freedom system with the VSD under the reinforcement learning using DDPG are carried out by a shaking table, and seismic responses are simulated numerically. From the experimental results, the peaks of response acceleration and displacement decrease about 1/2 in case with the VSD using DDPG, and also cut 20 % by switching series inertia mass. Finally, validity of vibration reduction during switching series inertia mass by the reinforcement learning is confirmed.

To clarify the behavior of railway vehicles under crosswind, a new method using the left-right balance of air spring pressure is presented in this paper. The rate of off-loading, which is an evaluation index for vehicle behavior, can always be estimated by the new method without conventional special wheel-set. The method is theoretically derived and the accuracy is experimentally verified comparing the new method with the conventional method. In the experiment, lateral forces equivalent to crosswind were loaded statically or dynamically, and a running test was conducted. The results show that the new method and the conventional method are in good agreement and that the new method is useful when analyzing vehicle behavior under crosswind. By applying the practical method to the Shinkansen, it is possible to obtain statistical data on vehicles running at higher speeds or under higher wind speeds easily than with conventional methods. Much data was analyzed on the relationship between wind speed and the rate of off-loading, and the validity of the Kunieda’s equation used for vehicle design under crosswind is verified. This method can monitor effects of crosswind on vehicles not only at specific points but also along railway lines. Therefore, it is possible to identify areas where vehicles are likely to be affected by strong winds. Some examples further improve safety against strong winds are shown in this paper.

Computer simulations using the combination of finite element method and modal analysis are now common methods for vibration prediction. However, when it comes to detailed models with large degrees-of-freedom, it is time consuming to evaluate and treat each characteristic mode one by one. Therefore, an efficient vibration analysis method is strongly desired. Macroscopically, characteristic modes of vehicle bodies, which mainly consist of frame structures and attached panels, are regarded as combinations of global and local deformations. Global deformation is attributable to frame structure which is defined as main component of the model, and local deformation is attributable to each panel component defined as sub component. As these deformations differ in priorities and countermeasures in design phase, it is essential to clarify the relation of global and local deformations, especially when they combine and compose coupled vibration. The purpose of this study is to present an analytical method to efficiently evaluate characteristic modes based on global and local deformations. In this paper, component mode decomposition is proposed to automatically decompose global and local deformation from characteristic modes of a frame-panel structure model. In this method, characteristic mode is first transformed into volume data and then three-dimensional wavelet transform is applied to separate global and local deformations. Then, from these separated deformations, component mode which consists of each single global and local deformation is derived using orthogonal transformation. By means of this orthogonal transformation, characteristic modes are effortlessly analyzed how much they contain each global and local deformation. In this paper, the proposed method is demonstrated on a simple frame-panel structure model.

This paper presents an improved calculation method of fatigue damage in spacecraft equipment under wideband random vibration. The method is based on principle of linear fatigue damage theory (Miner’s rule) to calculate accumulated fatigue damage, where the calculation of stress cycles is improved by more general Rice distribution instead of conventional Rayleigh distribution, which has been commonly used for decades. Examples of numerical simulation and test data of spacecraft equipment are used to demonstrate the accuracy of these two different statistical distributions and show that the generated stress of the equipment subject to wideband random vibration excitation would be filtered into a narrowband random stress response due to resonance of equipment. However, the statistical distribution of peak value of the narrowband random stress response can vary depending on the spectrum of excitation and resonance frequency and differs from the conventional Rayleigh distribution assumption. This paper, through the numerical and test examples, demonstrates that the improved method is a more feasible method which can create more accurate and less conservative accumulated fatigue damage than the conventional Rayleigh distribution.

We propose a method for solving differential equations implicitly using iterative calculations in spreadsheet software. This method uses difference equations by discretizing the coordinates of the independent variables with cells of spreadsheet for analysis. Thus, the equations are embedded in the cells performing substitution of grid point. In addition, the embedded equations are solved by Newton’s method, and the iterative function of the spreadsheet software is used in the calculation. This method can be applied even when the differential equation is discretized by an implicit solution method. Therefore, it can be applied to ordinary and partial differential equations of different types. The advantages of discretization using the implicit method (high computational accuracy, large time increments) are also presented in this paper. The analysis procedure is almost the same regardless of the type of equation. In this paper, as an example, a simple differential equation was analyzed using Microsoft Excel, a typical spreadsheet software, to illustrate the method. The results are also shown by the analysis of the gas lubrication equation.

This paper addresses a method of controlling the nodal positions of mode shapes as desired. The method is developed to calculate reflection and transmission coefficients at discontinuities and ends required for shaping mode shapes. The method is based on the one-dimensional Ray-Tracing method, which is a method that has been used to obtain eigenfrequencies, mode shapes, frequency response functions of a structure, based on the analysis of wave propagation on three-dimensional beam structures. Natural vibration in an undamped system refers to a state in which waves propagating on a structure are superimposed and become standing waves everywhere on the structure. Therefore, the mode shape indicates the shape of this standing wave. In the paper, a method is derived to estimate the reflection and transmission coefficients at discontinuities and ends, which is necessary to change the nodal points of the mode shape of interest. This method requires a matrix operation that changes only the values of the reflection and transmission coefficients at discontinuities and ends without changing the shape of the structure. Therefore, a rank-one reduction method via similarity transformation of matrices is developed. Finally, the effectiveness of the proposed method is confirmed through numerical examples.

Previously, the authors developed a strain gage (K-value gage) that can analyze the open mode stress intensity factor exclusively. This K-value gage was designed as a one-element strain gage with a gage width of 1 mm and a grid angle of 104° curvature. This paper presents an analysis method that can extract only the opening mode stress intensity factor from mixed mode (opening + shearing) cracks using the developed K-value gage. The extraction analysis method using this K-value gage was derived for the case of Poisson's ratio of the material to be measured 0.3. At the same time, Poisson's ratio led to use other than 0.3. Furthermore, this paper verified the validity of the proposed analysis method from experiments using a K-value gage. In the comparative verification experiment with the theoretical value, a tensile test was performed by inserting simulated cracks of the same length with only the inclination angle changed in a flat plate test piece. As a result, the experimental values of the aperture mode stress intensity factors were obtained within ±10% of the theoretical values.

Strip walking is an unstable lateral movement in a hot strip finishing mill that causes the strip to crash into the side guides, resulting in a tail pinch where the strip is rolled in a folded state. Strip walking should be appropriately controlled to prevent tail pinches, and an adequate mathematical model is essential to design the strip steering control. In this paper, the traditional model and a recently proposed model representing strip walking are numerically compared, and the latter is shown to characterize the details of the actual phenomenon. Such a mathematical model derivation makes it possible to use model predictive control to optimally control the strip walking, considering the rate limit constraint of the actuator. However, model predictive control requires time-consuming optimization computation. Thus, control performance deteriorates with longer control periods. In this paper, we propose a new model prediction control method to reduce optimization computation time with discretized control input that enables fast discrete optimization instead of slow continuous optimization. It is also shown that fixing the misalignment between the roll and steel strip centers is insufficient to prevent the tail pinch. Moreover, it is vital to suppress the rotational motion of the steel strip by using numerical simulations.

Recently, an automated driving of four-wheeled vehicle such as an automobile have been expanding in terms of various viewpoints such as a safety and an economy driving. Therefore, this paper investigated about the application of the control system for automated driving in order to run on the target turning trajectory. A feedback control is required to compensate a disturbance effect to a vehicle motion for automated turning driving to coinside with the target trajectory. However, comparing the control systems are little investigated because the designing a control system for the motion control of four-wheeled vehicle is complicated. From view point of evaluating a suitable control for automated turning drive, two control system were investigated in simulation study and experimentally. At first, this paper investigated a control system by using a backstepping method after converting the non-holonomic system to the chained system. Also, as a comparison target, this paper designed a control system using a kanayama’s method, which is well known control algorithm to the underactuated mechanism, adding a compensation control. In addition, in order to evaluate the robustness of the control system when the trajectory changes, this paper used an algorithm proposed by authors that make easy setting of the target turning trajectory consisting of clothoid curve. In these way, the comparison between the feedback control system using the backstepping method and the integrated control system using the kanayama’s algorithm was conducted, then, the characteristics of each system was shown by means of simuation study and experimental results.

In the conceptual design phase, generating novel concepts is a cornerstone for producing innovative products. This paper proposes a framework for design concept generation with an extensional conceptual space (ECS). A conceptual space is a geometric space of concepts constructed with concept distance. ECS focuses on the characteristics, i.e., function and physical attribute, each of which is a set of entities that fall under its definition. The concept distant measurement method based on word2vec with Wikipedia corpus is adopted to extract concepts from documents of design cases and to build a concept space. The framework facilitates a designer to explore a concept space containing text and image information with iterative operations, such as function development, and create a new design. This paper demonstrates a case study of chair design to verify the effectiveness of the proposed framework.

With the increasing sophistication of systems and complexity of social problems, co-design, which diverse designers with expertise in various fields participate in design process, is becoming general design practice. In this process, Designers drive design activities to reduce uncertainty (i.e., ambiguities in design requirement) to elaborate a design solution that fulfill the requirement by sharing and adapting each knowledge. Meanwhile, where heterogeneous designers with different backgrounds, beliefs and knowledge participate, uncertainty and its reduction direction are perceived uniquely for each designer. These discrepancies in the uncertainty perception and their reduction directions are barriers to the rational elaboration of design requirement and consensus building among designers. Therefore, it is necessary to proceed with design activities under a converges the discrepancies in their uncertainty perception through sharing the contents of uncertainty and its rationale that explains why designer perceives that uncertainty. This paper develops a typology of discrepancies in uncertainty perception among designers in co-design process. This paper firstly defines uncertainty and its relationship with elaboration of design requirement under the context of co-design through literature review of extant research related to uncertainty in design. Then, we developed a typology of the discrepancies through a set-theoretic formalization of the designers' thinking. The developed typology was applied to design experiment of toilet cleaning machine for identifying the discrepancies in the design process. The all discrepancies identified during its process could be classified into the typology. This result thus demonstrated certain validity and comprehensiveness of the developed typology.

This paper discusses the conditions for the undershoot responses in a circular path-following vehicle control. As a method to follow the circular path, it is known to combine the rotational and expansionary coordinate transformation to the vehicle kinematic model. However, the undershoot responses in which the output moves in the opposite direction to the reference value occurred under certain initial conditions of the vehicle. It is necessary to clarify the cause of the undershoot responses because the undershoot responses delay the time for following the vehicle to the reference value. The occurrence of the undershoot responses is related to the zeros of the transfer function. To derive the initial conditions of the vehicle that occurs the undershoot responses, in this paper, we analyze the zeros of the transfer function including the zero input response. Furthermore, we confirm the conditions for the occurrence of the undershoot responses through the numerical simulations.

Experienced drivers sometimes explain vehicle transient behavior to steering as "after the front wheels start to turn, the rear wheels start to turn with a delay" or "the rear wheels swing around the front wheels like a pendulum. In this paper, the former is called two-stage motion and the latter is called pendulum motion. In order to realize more fun handling performance, it would be effective to physically understand these motions and to reflect these understanding into vehicle developments. Thus, these motions were studied using equations of motion. In these studies, however, since the yaw moment of inertia has been limited to a special case, the valid ranges to which these studies results can be applied are unknown. Therefore, this paper studies these motions for general cases of yaw moments of inertia. Firstly, this paper yields the equations of motion for two-stage motion by algebraically manipulating the equations for, so-called, bicycle model whose yaw moment of inertia is not limited. These yielded equations mean that the front wheels begin to turn immediately after steering, and then the point a, where one of impact centers locates, begins to turn with a delay. Secondly, these equations are transformed into form of the true pendulum equation, whose one of the variables is the attitude angle at position e, where the attitude angle is zero on steady turn. By interpreting these equations, finally, it is pointed out that the pendulum motion is a phenomenon in which the point a accelerates toward the extension of the speed vector at the point e.

In this paper, we propose a control system design method for a human-machine cooperative system in which a human and machine cooperate to propel a cart to a reference trajectory. In the previously proposed method, the human plays the role of recognizing the environment and determining the propulsive force, while the machine determines the steering angle based on the position and posture of the cart, using a vector field with the reference trajectory as the orbit attractor. However, since this method cannot avoid obstacles on the trajectory, we extend it to a region attractor. By setting an index variable that changes depending on human input, the vector field is designed in the extended dimensional space including the index variable, and the cart can avoid obstacles by moving within the region defined by the reference trajectory with its parallel trajectory depending on human input. In this paper, we describe (1) the design method of the reference trajectory and the parallel trajectory considering the workspace of the steering and the distance between the cart and the wall, (2) the design method of the region attractor in the extended dimension including the index variable, and (3) the experimental verification of the proposed method.

This paper discusses a multiscale stochastic stress analysis of a particle reinforced composite material with a successive approximation based on the local sensitivity analysis of microscopic stresses with respect to a random location variation of particles. A microscopic geometrical random variation will have a significant influence on the microscopic stress fields in a heterogeneous material, and probabilistic analysis of the stresses should be encouraged for estimation of probabilistic properties of the stresses for more reliable structural design. Further, a more complicated microstructure reflecting an actual material considering wider analysis region, for example, including a larger number of inclusions in composite materials will be required for a practical application. This numerical analysis will be very expensive, and therefore a successive local sensitivity analysis-based approximate multiscale stochastic analysis method has been proposed for unidirectional fiber reinforced composite material. In this research, this approach is extended to a three-dimensional problem, and effectiveness of the approach for the multiscale stochastic analysis of a particle reinforced composite material is investigated. In this paper, the problem setting and outline of the methodology are provided, and the effectiveness and accuracy of the presented method are discussed with the numerical results.

This paper addresses a method of classifying modal clusters to realize resonance control. Especially in this paper, this method is applied to the simultaneous placement of resonance frequencies of the multiple modes belonging to a specified modal cluster through a structural modification to a limited subsystem. The method is useful for separating the resonance frequency of the modal cluster to be designed from the peak frequency of the excitation spectrum all at once. Therefore, the paper proposes a method to consider the multiple modes of a whole structure formed from the same mode of the subsystem belonging to the same modal cluster by using the modal contribution analysis method which is proposed by the authors. Then, under the above definition of a modal cluster, we introduce a method of moving the resonance frequencies of multiple modes belonging to the same modal cluster together without significantly changing the resonance frequencies of multiple modes belonging to other modal clusters. Finally, its usefulness is shown by a numerical example.

It is important to estimate the strength of the vehicle structure in the initial design stage of the vehicle development. The buckling of the plate is known to occurs in thin-walled structures before yielding, however, the structures can withstand loads even after buckling. The load acting on the thin-walled beam constituting the vehicle frameworks may be not only a compressive force and a bending moment but also a torsional torque. The purpose of this paper is to derive the equation for the torsional post-buckling behavior of a thin-walled square cross-section beam. When a torsional torque acts on a square cross-section beam, in-plane shear stress acts on the four plates constituting the cross section, which may cause shear buckling. Assuming that the same buckling deformation occurs on the four plates, Timoshenko's approximate expression can be applied as an out-of-plane displacement after shear buckling. The post-buckling stress distributions are calculated as a function of the out-of-plane displacement amplitude. Also, the post-buckling torsional constant and relationships between the torsional torque and the stresses are derived on the basis of Karman's effective width theory. The validity of derived expressions is examined by comparing with the computation results by the finite element method (FEM). As a result, in the target dimensional range of the plate (width-to-length ratio 1/3 or less), the absolute difference values of the derived relationship between the torsional torque and the maximum Mises stress are within 10% for FEM results. However, it is found that instable phenomena such as secondary buckling may occur as the width-to-length ratio of the plate becomes smaller. From the above results, it is found that the relational expression derived in this paper is sufficiently useful for calculating the torsional strength of a thin-walled square cross-section beam.

In this paper, we present a new approach to data-driven update of table-typed controller gain in tracing control system. This type of control system is used in various precision processing systems for industrial products. It is important for a tracing control system to be maintained so as to achieve the accuracy of positioning for reducing the gap between the workpiece and the machining tool during the operation. In addition, it is also important to update the elements of the table in the controller as faciley from the practical points of view. The update of the controller by directly using the data without mathemaical models is one of the rational ways to realize them. From these backgrounds, we have previously applied FRIT (Fictitious Reference Iterative Tuning), which is a data-driven controller tuning method with only one-shot experiment, to update the table-typed controller in the tracing control system. However, this method requires calculations using a fictitious inverse system of an implemented controller, which leads to computational errors. In this paper, we propose a method utlizing VIMT (Virtual Internal Model Tuning) to update the table-typed controller without using its inverse system. Moerover, we update not only each gain in the table but also each corresponding range to improve the tracking accuracy of tracing control system.

To improve robot manipulability, force sensing for scissors is expected to be significant. In general scissors, the moment and the crossing angle of the blade depends on the cutting position, resulting in the difference of the required input force for cutting. Though logarithmic spiral scissors enables to equalize crossing angle of scissors blade, the scissors also causes the moment difference. This paper proposes a mechanical motion conversion mechanism which equalizes the cutting input force of logarithmic spiral scissors. For the mechanical motion conversion mechanism, we suggest a curved slider mechanism which is combined with a linear slider. Because the curved shape can be calculated by analytic method, the linear slider stroke design procedure is simple compared with 4 linkage mechanism. Furthermore, the proposed mechanism enables to control the cutting speed easily compared with the former studies because the cutting speed is proportional to the linear slider speed. In the geometric and mechanical experiments using dummy scissors(40 mm in blade length), we confirmed that the cutting input force was independent of the cutting position if the constant load is applied. In the kent paper experiments using usual logarithmic spiral scissors (65 mm in blade length), the proposed mechanism could convert the constant load excepting friction effect.

Generating novel design concepts in the conceptual design phase is a cornerstone for producing innovative products. This paper proposes a method to support the generation of new design concepts from patent documents. The proposed method is based on the theory of novelty potential that the combination of abstract concepts leads to the generation of the novel design concept and that the distance between the combined abstract concepts correlates to the possibility of generating it. This research adopts the distributed word representation model of word2vec to extract abstract concepts from patent documents and to measure the concept distance. The two matrices are introduced to visualize the relationships of the extracted abstract concepts; the novelty potential matrix which shows the value of the novelty potential between the abstract concepts, and the void matrix which shows the numbers of existing patents of the abstract concepts. They help a designer to identify the combinations of abstract concepts that are high novelty potential and no existing patent. This paper demonstrates a case study of generating novel ideas of sports business with blockchain using the proposed method. The results show its uniqueness and effectiveness.

Lower limb exoskeletons with actuators aid physically challenged individuals in walking. We had developed a novel, lighter than conventional, magnetic strain wave gear for exoskeletons. However, output shaft was not rotated when overloaded even through input shaft was rotated (step out phenomenon). Then, the transmission torque was limited by a step-out phenomenon. In the paper, we aimed to elucidate a step-out phenomenon for future design. We observed deformation shape of two types of magnetic strain wave gears, through multi flexible dynamics analysis with magnetic force database and measurement using displacement sensors. Basically, the deformation shape was ellipse. The phase difference between rotational angle of input shaft and tilt angle of ellipse deformation shape was enlarged when the load was increased. The phase different became excessive and the deformation shape was changed to true circle when overloaded showed. Compared a gear with small 4-pole magnets, a gear with large 16-pole magnets could keep deforming shape and it had higher limitation torque.

Low temperature difference Stirling engine is an excellent teaching material for mechanical design because (1) students can experience all of four dynamics (material mechanics, fluid dynamics, thermodynamics, and dynamics of machinery) by designing and constructing it, (2) it is safe even if the design is not appropriate, so it allows students to freely try their own ideas, (3) there are many variations of engine configurations, including types and mechanisms, so there is much room for creativity and ingenuity, and (4) as the structure is simple, it is easy for students to understand if their ideas make sense. On the other hand, it is not easy to construct a working engine because of its low power. In order to make low temperature difference Stirling engines more effective teaching materials, this paper describes the design and fabrication of a low temperature difference Stirling engine with high power per swept volume that can run at higher speed. The results of analyzing the working fluid temperatures in the expansion and compression spaces indicate that the temperature ratio increases with increasing speed even in a low temperature difference Stirling engine. We found that an arrayed resin regenerator can provide high performance, and realized that a Stirling engine with a power piston swept volume of 4.5 cc and using hot water and ice as heat sources can run at a maximum power of 171 mW and a maximum speed of 31.3 Hz. Compared with other low temperature difference Stirling engines, the power per swept volume of test engine is more than 10 times higher and the rotational speed is nearly three times higher than those of the low temperature difference Stirling engines already reported. The results show that low temperature difference Stirling engines with high power per swept volume can be designed and fabricated by using 3D printers, and the value of the low temperature difference Stirling engine as a teaching material has been enhanced.