The Jeffcott Rotor used widely in the analysis of resonancies is a two-degree-of-freedom model with a disk at the midspan of a massless-shaft. Natural frequency of a forward whirling mode p_f and that of a backward whirling mode p_b have the relation of p_s:p_b=1:-1. Consequently, subharmonic resonancies of a forward and a backward whirling modes and a combination resonance occur at the same rotational speed of twice the major critical speed and internal resonance phenomena appear. In this paper, nonlinear phenomena in the neighborhood of this rotational speed are investigated. The following are clarified theoretically and experimentally. (a) The shape of resonance curves becomes far more complex than that of a single resonance. (b) A quasi-periodic motion occurs. (c) These phenomena are influenced greatly by the misalignment in assembly and misattachment of rotor position.

View full abstract

The advantages of applying sensorless control to mechanical systems are cost reduction, maintenance of stability of a control system by collocation, noise reduction, decrease of breakdown frequency, increase of reliability, and space saving. In order to realize such a control system, robust performances are required in the state estimation in addition to the control law, so we combined a VSS observer and sliding mode controller and constructed a sensorless nonlinear robust control system. We used a four-degree-of-freedom tower structure with a hybrid dynamic vibration absorber (hybrid DVA) using a linear motor as a control object. It is possible to estimate the state of the system by means of an observer from the coil current of the linear motor for the mutual interaction on the dynamics in the state equation of the system. Then we applied a VSS observer/sliding mode control combination type sensorless nonlinear robust control system to the vibration control system of the structure. Moreover, we investigated performances with regard to vibration suppression.

View full abstract

In order to isolate vibration transmitted to a machine from the floor and to reduce the vibration caused by shock force applied directly to the machine, a new method of ON-OFF control for the spring support is suggested in this work. The end of the spring is designed to control the clamping friction force by means of a brake mechanism. The clamping friction force is varied according to the switching law which is deduced using the variable structure systems (VSS) control theory. The switching law takes into account the energy dissipation due to friction. In the numerical simulations and experiments, the reduction of the impulse response and the displacement transmissibility are investigated. The results show that the method is effective for suppressing shock motion and isolating vibration.

View full abstract

This paper deals with a study on the active structure-borne sound control of a flexible plate by the feedback control method. It involves modeling a real structure as a reduced-order lumped parameter model on the basis of modal analysis, and then adopting modern control theory to control the vibration of selected multimodes of a plate that has high radiation efficiency. The modeling procedure and the LQ and suboptimal control design are presented; control simulation of the discrete model is carried out in order to demonstrate the effectiveness of the method. Experimental work with vibration control of the plate enables a final verification.

View full abstract

Recently in railways, lightweight vehicles have been used in order to save transport energy. This tendency is notable especially in high-speed railways. However, the flexibility of such vehicles sometimes induces deterioration of ride quality. Therefore, to obtain high ride quality, an H^∞ control method for an active suspension system in rail cars is proposed. In design of the H^∞ controller, it is necessary to determine suitable weighting functions so as to achieve a special frequency band at which high ride quality can be attained. Additionally, in this controller, robust stability against modeling errors such as changes in vehicle weight or disregard of high-frequency resonances must be pursued. The effects of H^∞ active suspensions are confirmed by digital simulations.

View full abstract

The motion of balls in a tumbling ball mill was simulated numerically using the discrete element method (DEM). Simulations are performed for the two-dimensional and three-dimensional cases. To study the difference between two-dimensional and three-dimensional powder milling analyses, the ball motion was analyzed. It was found that there is no significant difference between the flow patterns for the two-dimensional and three-dimensional cases, but some properties such as the trajectory of the center of the gravity of the balls, the profile of the average velocity vector, the average ball velocity and the fluctuation velocities of the balls show significant differences.

View full abstract

This paper concerns the optimum design of generally laminated rectangular plates having arbitrary boundary conditions. The optimum design criterion is maximization of the fundamental frequency, and the design variables are fiber orientation angles of the constituent laminae. A previously developed analytical method is used to calculate natural frequencies. A complex method, which is an extension of the simplex method to more complicated problems with constraints, is applied to the optimization procedure. The optimal stacking sequences of laminated plates for two composite materials with different degrees of orthotropy are obtained in numerical examples, and the applicability of the present design method is demonstrated.

View full abstract

The equation of motion of the real adaptive truss having nodal offsets is formulated to discuss the influence of the nodal offset on its dynamic behavior. It is shown that the equations of motion of the ideal and real adaptive truss models can be formulated in the unified description by paying attention to the variables which express the state of the dynamic motion. Simulation studies and the calculated natural frequencies demonstrate the significant influence of the nodal offset on the dynamic behavior of the adaptive truss. It is also shown that the magnitude of the offset must be small enough in order to eliminate the influence. The nodal offset design problem is formulated based on the natural frequency criterion, and a case study shows a possible nodal offset design of the adaptive truss.

View full abstract

This paper is concerned with an active suspension with preview of passenger cars using fuzzy reasoning. The model for motion of a vehicle is described by nonlinear differential equations with four degrees of freedom excited by irregular inputs from road surface. The active control for the suspension is constructed such that the mean squares of vertical and rotary accelerations of the vehicle body are minimized subject to the constraints of suspension and tire deflections. In the fuzzy control rules, the vertical accelerations and velocities of vehicle body at the front/rear suspension locations, and the preview information of road surface are denoted as input variables, and the linear combinations of input variables are denoted as output variables. The simulation results indicate that the proposed method is much improved in the suspension performance of the vehicle body.

View full abstract

In the latest passenger cars, 4 WS, a suspension system, and other vehicle systems are controlled using various sensors and actuators. Each control system is equipped with a fail-safe device to guarantee safety. However, the fail-safe device can detect only unequivocal failures such as a short circuit. This paper presents a method for estimating the ambiguous change in a 4 WS control system using generalized likelihood ratio (GLR). This method is designed to discriminate ambiguous failures from vehicle characteristic change due to running circumstance change by processing output errors between sensors and an observer. These output errors have a peculiar time series pattern corresponding to each ambiguous change. GLR can determine what the ambiguous change is by checking time series pattern of the output error. The result of the simulation using experimental data revealed that GLR can estimate the failure of the sensors, the actuator, and vehicle characteristic change. Also, failure detection performance between GLR and failure detection filter (BJDF) was compared.

View full abstract

We suggest an effective way of reducing drum-type brake noises. We noted that noises are most likely to occur when the eigenvectors and eigenvalues of the brake drum and brake shoes are close to the coupled vibration frequency. It was also noted that the eigenvectors of the leading and trailing brake shoes are independent from each other. Based on these findings, the two shoes were rendered asymmetric in rigidity, causing the eigenvectors and eigenvalues of either of the two shoes to exist apart from those of the drum in fundamental to high-order modes. This in turn effectively prevents increase in the magnitude of coupled vibration of drum and shoes, and significantly reduces brake noise level.

View full abstract

It is often necessary to utilize more than two probes in measuring three-dimensional (3-D) shapes by means of a 3-D coordinate measuring machine. Accuracy of measurement, especially of complicated shapes, can be improved by using multiple probes, because of the wide range of measured data. In the case that more than two probes are used, calibration of each probe must be performed to correct the zero position of the scale. In this paper, a new method for processing data measured by multiple probes is proposed. Two kinds of parameter related to shape and offset vector of a probe are involved in an observation equation. The parameters can be determined simultaneously by solving a set of observation equations for both standard probe and offset probe. The efficiency of the measurement by multiple probes is not affected because probe calibration and data correction are not essential in the proposed method. The simultaneous processing method is applied to experimental data. The most probable values and the standard deviations of the parameters are calculated by the least squares method. Results show that measurement accuracy of the method is almost the same as that of a conventional calibration method.

View full abstract

This paper deals with the modal control of a planar structure using distributed-parameter sensors designed with a view to extracting a particular vibration mode. First, this paper overviews the modal control based upon distributed-parameter sensors and actuators. It is found that the utilization of either distributed-parameter sensors or distributed-parameter actuators leads to the elimination of spillover destabilization of a feedback control system. In addition, the distributed-parameter sensors have the advantage of achieving the modal control of a targeted mode of the structure; however, a mere usage of the distributed-parameter actuators results in uncontrollability of the targeted mode. Finally, with one-dimensional PVDF film sensors attached to the planar structure, an experiment is conducted, demonstrating the capability of the distributed-parameter sensors for suppressing the mode of interest without causing any effect on the remaining modes of the structure.

View full abstract

In order to suppress structure borne sound, it is very important to analyze the vibration behavior of a machine. Accelerometers are commonly used in measuring the vibration behavior; however, the weight of those prevents the user from obtaining accurate results. In this paper, sound pressure distribution measured by microphones is utilized to predict the vibration behavior as noncontact measurement. The relation between the vibration and sound pressure can be expressed using boundary element method. Simple prediction of the vibration behavior can be achieved using measured sound pressure distribution caused by structure borne sound and a vibration-sound pressure transformation matrix. The proposed method is validated by numerical simulation and experiment at natural and off-natural frequencies.

View full abstract

This paper describes a formulation and stability of the modified LMS algorithm for active periodic noise control. A typical application of active noise control is the control of engine-induced noise inside cars, which synchronizes with the rotating speed of the engine crankshaft. This modified algorithm effectively decreases the operating time of the filter coefficients in a single computational step. We applied the active noise control system using the filtered-x adaptive notch filter algorithm. Numerical simulation results reveal the stability limits of the step size parameters for the successful implementation of this control system. An analysis of these results also show that a smaller control pulse number lead to faster convergence and consequently, larger step size parameters being able to be used. This, in essence, promotes the implementation of such algorithms using the more commonly used cheap microprocessors. The results also indicate that the filtered-x adaptive notch-filter algorithm has higher stability and faster response than other algorithms.

View full abstract

This paper is concerned with the positioning control of a single-link flexible arm using frequency-shaped sliding mode control theory. In the control object, a harmonic drive system for a reduction gear has a large friction torque, and thus we applied the sliding mode control theory to this positioning control problem in order to compensate for the nonlinear friction torque. We designed the control system using the output feedback method. The high-gain sliding mode controller, however, often causes instability of the control system due to spillover phenomena in the high-frequency region. Therefore, we also design a switching hyperplane using H_∞ control theory which can suppress such spillover phenomena markedly. We found that the frequency-shaped sliding mode control system exhibits good performance. However, the system performance is influenced by uncertainties which do not satisfy matching conditions.

View full abstract

Pneumatic actuators have many advantages. However, the difficulty of positioning control of such an actuator if a conventional linear control theory is adopted has been pointed out. An application of an optimal predictive and adaptive control method using a nonlinear model, i.e., a neural network, is proposed in this study. The suitability of this control method is demonstrated through experiments on a one-degree of freedom robot manipulator which uses a pneumatic actuator. Furthermore, a new control strategy, in which the control object is regarded as both the actuator and a PID control circuit, for improving the positioning performance of the actuator is proposed. The results of experiments demonstrate that precise positioning control can be realized by this strategy.

View full abstract

This report treats micro stick-slip vibrations occurring in a typical computer-controlled hydraulic servo system. The relevant system consists of a hydraulic cylinder, a servo valve and a personal computer. First, an experiment was performed for a conventional feedback control system to confirm the occurrence of two different types of stick-slip vibration of small amplitude. Detailed numerical analysis was then performed to identify the main factors contributing to the vibrations. The control signal generated by the D/A converter causes micro stick-slip vibration of small amplitude over a wide range of control feedback gain. Increasing the feedback gain results in another type of micro stick-slip vibration of large amplitude due to the locally increased flow gain within the underlap of the servo valve. Computational results agree well with the experimental ones.

View full abstract

We developed an electrostatic linear actuator, which consists of parallel plate capacitors and a slider, for use as an artificial muscle. For ease of manufacturing of the small slider, we used conductive material instead of dielectric material used in previous experiments. Computer simulation was performed to determine the effect of the structure of a slider that consists of both conductive and insulating material. Based on the results of simulation, a metal slider was designed and manufactured. The output force of 8.3 mN was generated by the actuator under the driving voltage of 300 V, which was 30% larger than that generated by the actuator under the same voltage in the previous experiment, where the slider was made of dielectric material.

View full abstract

This paper is concerned with an analysis and design method for a robust controller of magnetic bearing system with parameter perturbations. We apply the method of μ-synthesis based on the descriptor form representation to a magnetic bearing system. The unmodeled dynamics and the sensitivity performance are taken into cosideration. By using this method, physical uncertainties are treated independently and a controller is designed to achieve not only robust performance but also robust disturbance suppressions.

View full abstract

In order to design a highly stable and robust control system for a flexible rotor suspended by magnetic bearings, it is important to understand the actual system characteristics through experimental identification and assign appropriate eigenvalues based on the results. This paper describes new methods of system identification and pole assignment by output feedback, the Eigenvalues Move Method. The effectiveness of these methods were verified using the measured transfer functions of an experimental apparatus digitally controlled by DSP.

View full abstract

In this paper, the performance of alarm systems which are based on vibration monitoring is considered. Evaluation and comparison of alarm systems based on loss, e.g. missed prohibit due to false alarm and cost of repair due to missing alarm, are reported. As the high sensitivity and low false alarm rate contradict each other, loss is one of the important indices for alarm system evaluation. Two methods are compared: one is the threshold method (level-checking method) and the other is the Kalman-filter-based method combined with Bayes' theorem. Bench-size machinery is constructed, then its vibration is measured by an acceleration sensor attached on a bearing housing. In each method, parameters are adjusted to minimize the expected total loss. In this situation, the ranges in which the threshold method is applicable and the Kalman-filter-based method is applicable are shown on an ROC (receiver operating characteristic) chart. A generalized procedure for selecting an alarm system using ROC is also discussed.

View full abstract

Flexible control experiments have been performed as one type of on-orbit experiment for the Engineering Test Satellite VI (ETS-VI) which was launched in August, 1994. The aim of the flexible control experiments is the feasibility study of new attitude control schemes for future flexible space structures. We have designed two types of controllers. One is a controller designed according to the Low Authority Control (LAC)/High Authority Control (HAC) scheme. The purpose of LAC is damping augmentation of structural vibration modes, while that of HAC is achievement of high control performance. The other controller is a HAC-only controller: this HAC controller is designed for high-performance attitude control of a rigid body, as well as for damping augmentation of the vibration modes. The HAC controllers in both cases are designed using H_∞ control theory with the frequency-shaping technique. These control schemes have been tested on orbit by ETS-VI. The experimental results have been analyzed in both time and frequency domains. It is shown from the analysis that both control schemes can be accurately realized on orbit and perform well with respect to damping augmentation and attitude control.

View full abstract

In this paper, the free-flying space robot, comprising a satellite base body and articulated manipulator arms, is considered. We clarify the reduced effective inertia of this sophisticated mechanism is smaller or equal to the effective inertia associated with the articulated manipulator arm which is fixed on the earth. Furthermore, we develop the control system which decouples motions of satellite base body with the end-effector movements with application to control system for capturing a turget by a space robot. Finally numerical simulations demonstrate the usefulness of the proposed control system.

View full abstract

In order to utilize autonomous robots in the service fields such as in offices and hospitals, the robots should be equipped with functions for co-existing with humans. In this paper, initially, we discuss the functions which service robots require, such as (1) safety for humans, (2) adaptability to unknown environments and (3) maintenance of appropriate function. We propose the architec-ture based on SSA (Subsumption Architecture) to realize the above functions. We have developed a prototype system using a 5-axes robot and touch sensors. We have confirmed the efficiency of our system by experimentation.

View full abstract

This paper presents a trajectory planning method by which human cooperative mobile robots can avoid collision with human bodies. Indefinite human motion is predicted using a stochastic predictor. Evaluation of collision risk and efficiency of trajectory are taken in account in an evaluating function. Optimization using a genetic algorithm gives the robot trajectory. The mobile robots move so as to maximize safety taking into account every possible human motion. Simulation results showed that the robot could determine appropriate trajectories avoiding danger and that reasonable trajectories were generated even if human motion unexpectedly changed.

View full abstract

We describe a method of path tracking on 3D terrain using information of inclination of a robot. We modelled 3D terrain between two ridges based on measured shapes, and we made a winding desired path on the lowest portion of the field furrow. We have proposed a new method in which the rolling angle of the robot and its variation are used to track the desired path in real time. In this method, information of the desired path can be given to the robot as the rolling angle and its variation, and the robot can recognize deviations in its position and direction using rolling angle and its variation. Furthermore, the robot is controlled using a newly proposed algorithm in which path tracking is carried out using rolling angle. Finally, the newly proposed path tracking method is successfully simulated on a computer.

View full abstract

We report a method for teaching tasks based on human tasks in a virtual environment and making a database which is used for cases similar to the recorded environment. In order to describe the work environment, a method for representing configurational relationships is proposed. The representation method uses the qualitative states between two primitives. Since the representation is based on primitives, it can be combined with a parts connectivity graph, which can be used to analyze a machine from the topological point of view and represent the structur of a machine hierarchically. The tasks are hierarchically recorded as production rules using these representation methods, and when the structure of the work environment at any level agrees with the data, a new action can be created. Some learning case studies show that the method can be used to create new actions if the orientations and shapes of work objects are different from those in the teaching environment.

View full abstract

In the case of laser cutting machine of rocking mirror type, which is newly developed and a kind of optical manipulator, the positioning error of final laser spot on cutting plane, caused by assembling inaccuracies of optical system, would inevitably exceed the allowable limit without appropriate compensation. This research aims to establish the theory on the positioning error and its compensation for this new machine, with kinematics and inverse kinematics of manipulator. In this paper, a formula of actual position of final laser spot is derived as an implicit function of aimed position and the assembling inaccuracies.

View full abstract

First step is manipulator kinematics. The actual position of final laser spot [X^*, Y^*]^T, derived in 1st report as implicit function of the aimed position [X, Y]^T and assembling inaccuracies, is now transformed into explicit function of them, through approximate linearization. The explicit function shows the potential magnitude of positioning error and the existence of several pairs of assembling inaccuracies whose contributions to positioning error can be canceled each other with some appropriate hardware in assembly. Second step is inverse manipulator kinematics. Jacobian of [X^*, Y^*]^T to [X, Y]^T is derived. Since the Jacobian is proved nonsingular, the whole contributions of all kinds of assembling inaccuracies to positioning error can be collectively compensated, in operation, by purposefully shifting the reflector joint angles [θ₁, θ₂]^T from the honest value corresponding to [X, Y]^T. Practicable approaches are given to determine the purposeful shift [δθ₁, δθ₂]^T. Thus, the problem of positioning error in rocking mirror type laser cutting machine is solved by combined adjustment in assembly and bundle compensation in operation.

View full abstract

The authors developed a new apparatus that can measure lubrication film thickness accurately at the nanometer scale. Mica was used as a solid specimen and three kinds of liquids, i.e., octamethylcyclotetrasiloxane (OMCTS), cyclohexane and n-hexadecane were used as liquid specimens. From the ultrathin lubrication film thickness measurements using the apparatus, diseretization of the film thickness is observed for OMCTS and cyclohexane. The cause of the discretization is also discussed.

View full abstract

The effect of addition of liposomes of dipalmitoyl phosphatidylcholine (DPPC) and γ-globulin on the natural joint lubrication was evaluated through pendulum friction tests. The frictional characteristics of pig shoulder joints were confirmed to depend on the viscosity of lubricants only in the physiologically low-load condition and immediately after the loading. A detergent (polyoxyethylene p-t-octylphenyl ether) was successfully used to remove adsorbed films from the articular surfaces. The friction coefficient of natural joints was significantly increased in mixed lubrication mode with a saline solution of sodium hyaluronate (0.2 g/dl) by surface treatment with the detergent. The addition of liposomes of Lα-DPPC or γ-globulin significantly improved the boundary lubricating ability of the articular surfaces treated with the detergent, depending on the quantity of the additives. It appears that the liposomes of Lα-DPPC and γ-globulin can form protective films on the articular surfaces in severe frictional condition.

View full abstract

We investigated static and dynamic characteristics of an externally pressurized porous annular thrust gas bearing, which has a thin restricted surface layer, by numerical analysis and experiment. In the analysis, it is assumed that the fluid flow obeys Darcy's law in the porous material, restricted with Darcy's restrictor or orifice restrictor in the surface layer, and is governed by the conventional Reynolds equation in the bearing clearance. We also carried out an experimental investigation on static and dynamic characteristics of the porous annular thrust gas bearing. As a result, it is found that the theoretical results calculated using the Darcy model in the restricted surface layer agree with the experimental results better than the results calculated using the orifice restrictor model.

View full abstract

We numerically analyzed the dynamic characteristics of an externally pressurized porous annular thrust gas bearing which has a thin restricted layer on the porous surface. As a result, it was found that the surface-restricted layer had such an important effect in that it increases stiffness and shifts the positive maximum value of damping from a high-frequency region to a low-frequency region. It can also reduce the magnitude of negative damping in the low-frequency region. Under the conditions of allowable static stiffness and the local minimum value of dynamic stiffness, we derived the unique relationships between the design parameters and the squeeze number that can provide the maximum damping ratio. Considering the mass parameter of the body supported by the bearing, we propose an optimal design method for maximizing damping ratio at the natural frequency, while maintaining the required stiffness in the low frequency region.

View full abstract

The authors investigated the adhesion coefficient between rails and wheels using a model "Slip Adhesion Test Truck" and model rails, and quantitatively evaluated factors such as rail temperature and humidity which may affect adhesion. The adhesion coefficient was measured under various conditions. As a result, they find that adhesion is strongly influenced by the rail temperature and the humidity around the running surface of rails.

View full abstract

A method for inspecting a helical gear using a coordinate measuring machine (CMM) is proposed. In the method, the coordinates of a large number of points on the tooth surface of the gear are measured at random using a CMM. The method is based on the idea that the pitch error, profile error, pressure angle error, and spiral angle error can be calculated from the measured coordinates because the information on various errors of the gear is included in the measured coordinates. For calculation of each error, the tooth surface is described as a function of each error item and then the values for the items are estimated from all of the measured coordinates by the method of least squares. The difference between the estimated value for the item and the corresponding design value is the error of the item. From the results of experiments, it is clarified that each error of a helical gear can be estimated using a widespread CMM instead of a specially designed measuring machine for gears.

View full abstract

In the manufacture of straight bevel gears, there is usually a need for some "trial-and-error" process by which a good tooth bearing is obtained. In this paper, a method for cutting straight bevel gears in which a trial-and-error process is not necessary is proposed. The method can be applied without the necessity of cutter modification or generator alteration. Each gear and pinion is generated by a newly introduced "quasi-complementary crown gear" instead of a conventional complementary crown gear. The quasi-complementary crown gear is a bevel gear whose pitch cone angle is 90° minus the root angle of the workpiece and whose tooth surface is a plane. The pitch cones of the quasi-complementary crown gear and workpiece roll with each other. Therefore, the ratio of roll between the crown gear and workpiece is expressed as a function of the root angle. The quasi- complementary crown gear produces a profile modification of the generated gear tooth. Gears made on an experimental basis showed a good tooth bearing as expected.

View full abstract

A method for the analysis of load distribution and contact stress on tooth faces of a thin wall spur gear are proposed by combining a mathematical programming method with the three-dimensional finite element method (3D-FEM). A program that can be used to carry out tooth load calculation for a sophisticated gear structure with a personal computer is developed. The main procedures of the method are as follows: Firstly, the region on tooth faces surrounding the geometrical contact line which are fairly larger than the theoretical region calculated by the Hertzian's contact theory, is defined as a assumed contact face, secondly, reference points on the assumed contact face are taken on each gear, thirdly, influence coefficient on each reference point is calculated by 3D-FEM, lastly tooth load on each reference point is calculated by the mathematical programming method. The contact stress transformed from the tooth load is well agreed with the theoretical one for a thick gear. Furthermore, strains calculated on the compressive side of the tooth root of the thin wall gear are well agreed with the measured ones.

View full abstract

The fully automated running of machine tools requires early and automatic detection and prevention of abnormal cutting. Therefore, the purpose of this study is to determine a new method of detecting self-excited chatter vibration. Using this method, the occurrence of self-excited chatter vibration during a turning operation is evaluated by a neural network using a spectrum map of cutting tool vibration. The following conclusions were reached as a result of assessing this detection method during an actual turning operation. (i) The performance of the detection method is the most effective when the neural network uses five spectra as a spectrum map for the evaluation. (ii) The proportin of unsuccessful judgements was almost the same as that obtained with previous approaches. (iii) This proportion was almost unrelated to the noise level of the signal used for detection.

View full abstract

Recently, we have proposed an effective method for inspecting the delamination of thermal sprayed coating, which is called the ultrasonic bottom echo method. In this method, it is important to keep the position of an ultrasonic probe normal to the surface of the sprayed specimen during testing. However, the sprayed specimens are often distorted by a high-temperature heat source for spraying, such as electric arc or fuel flames. Furthermore, the shape of the sprayed specimen is not always flat. Thus, it is difficult to perform automatic inspection precisely using a conventional simple testing apparatus. A flexible scanning apparatus with five degrees of freedom was constructed that enables precise ultrasonic testing of curved thermal sprayed specimen. Control systems were fabricated to keep the UT probe normal to the specimen and to move it along its surface precisely. As a result, it was confirmed that the control systems can be applied to the ultrasonic testing of curved specimen by the bottom echo method.

View full abstract

Since in current stereo-lithography an object is fabricated in liquid resin, use of a "support" shape other than an objective shape is required so that isolated portions of the objective shape do not float away and overhang portions do not curve in the liquid resin. The necessity of designing supports before and removing them after fabrication restricts the range of fabricatable shapes as well as requires time and effort. To solve these problems, we propose use of refrigerative stereo-lithography. In its process, liquid resin supplied as a layer is cooled and solidified. Then a laser beam is used for selective photopolymerization of the resin layer. After iteration of this process for all layers, the fabricated object can be obtained by heating the entire solidified resin block to melt and remove only nonphotopolymerized portions. Since the object is fabricated in already solidified resin, there is no need for supports. We also describe the refrigerative stereo-lithography machine developed and a fabricated example.

View full abstract

Previously, dynamic characteristic charts were presented as the evaluation method of dynamic characteristics of the robotic parallel crank mechanisms. The dynamic characteristic charts reveal the change in dynamic torque which applied to each driving shaft of robots as they move the whole working space. The dynamic equations of the parallel crank mechanisms do not involve effects due to Coriolis' force so that the dynamic characteristic charts can be made easily. On the contrary, it is difficult to make the same charts for open loop mechanisms since their dynamic equations involve effects due to Coriolis' force. However, open loop mechanisms are the most commonly used so that an appropriate dynamic characteristic evaluation method is particularly needed for them. Therefore, we propose use of the dynamic characteristic charts for open loop mechanisms by transforming the dynamic equation. Furthermore, this paper describes a simple measurement method for dynamic torque applied to driving shafts of robotic mechanisms. Using the method, changes in the dynamic driving torque of robotic open loop mechanisms are measured practically, and the utility of the dynamic characteristic charts is confirmed based on those results.

View full abstract

Many trains were derailed and damaged by the earthquake in Kobe on January 17, 1995. At the same time, many railway girder bridges and viaducts were seriously damaged by the quake. It was postulated that serious accidents could occur as a result of trains running at high speeds into these girder bridges or viaducts. Fortunately, no fatal accidents occurred. On investigating such accidents, the authors found that they rarely occur, because a train derailed by the violent shock of a quake is so radically decelerated by the huge running resistance caused by the derailment that it stops just after the quake. On the other hand, it takes some time until the bridges or viaducts collapse after the quake.

View full abstract

When no slip is assumed in wheels, a steering mechanism based on Ackerman's theorem can be used to obtain smooth turning of a car. We propose a new steering mechanism which is composed of only spur gears. The steering angle of each driving wheel is controlled via the gear mechanism. The effectiveness of this steering mechanism is verified using a vehicle model. It is verified that the mechanism presented in this work is appropriate for vehicles. This mechanism can he applied to the attitude control of vehicles.

View full abstract