Experimental modal analysis is widely used in many fields and many curve fitting techniques which estimate modal parameters have been developed. This paper presents a new multireference curve-fitting technique which is based on the maximum likelihood method. The proposed method utilizes the frequency response function (FRF) as input data and obtains modal parameters by iterative calculation. In this paper, initial values for the iteration are supposed to be obtained from the modal parameters estimated by the polyreference method. It turns out that modal parameters converge to the same value regardless of initial values. Therefore, the analyst can get rid of troublesome judgment on selecting physical roots in the polyreference method. The FRF constructed by the result of the proposed method agrees better with the experimental FRF than that by the polyreference method. Moreover, modal parameters estimated by the proposed method are more reliable from a statistical standpoint.

View full abstract

This paper discusses a methodology for estimating frequency response functions of structures. Firstly, causes of random errors associated with the frequency response function estimates are analyzed. Secondly, a method for improving the accuracy of the estimates is proposed. Finally, that technique is verified experimentally by multiple input testing for an actual structure. Consequently, the following results are obtained. A linear dependence between the input forces at the natural frequencies of the test structure increases the confidence intervals of the estimates. However, that estimating problem is improved by the proposed method in which the drive signals for exciters are considered to be the inputs for the testing system and both of the exciting forces and the response displacements are considered to be the outputs of the testing system.

View full abstract

This paper deals with the flexural vibration of the three-layered beam, the boundary condition of which is free-free. Three kinds of three-layered beams, core materials of which are all different and a normal steel beam, for reference, are tested. It was clarified that if the correct core material is chosen for the temperature, the vibration mode becomes a complex function and a high modal loss factor is achieved. Further, this paper presents the analysis method used to determine the damped normal mode (it is a complex function) described by Mead. The calculated modal loss factors have the same physical meaning as that in the one-degree of freedom. The accelerance at the driving point is calculated by the superposition of the obtained modes and values of the resonant frequency nearly equal to the result of the experiments are predicted.

View full abstract

This paper presents the results of experimental research on the dynamic behavior of a frame structure consisting of a concrete-filled square tubular steel column. It is said that the concrete-filled Square column, the so-called composite column, possesses good strength and ideal ductility. However, its characteristic of vibration under strong earthquakes is not clear enough at present. We conducted experiments to compare the effect of the composite frame with that of the noncomposite frame using a shaking table. In these experiments the following were observed. The composite structure was superior to the noncomposite one in terms of stiffness. The damping ratio increased remarkably outside of the elastic domain.

View full abstract

The authors apply the transfer influence coefficient method to the in-plane flexural free vibration analysis of a structure with double layers regarded as a discrete system with lumped mass, lumped inertia moment, and massless linear and rotational springs. They describe an effective method of reducing the computation time for the structure in which the lengths of layers are remarkably different from each other, and a systematic and generalized treatment of the releases and rolls in each layer at which the displacements are discontinuous. The results of the comparatively simple numerical computational examples on a personal computer also demonstrate the validity of the present algorithm for the particular structure, that is, the numerical high accuracy, the high speed and the flexibility for programming of the present method, compared with the transfer matrix method.

View full abstract

Oscillations caused by solid friction in the hydraulic driving system of the previous report are treated theoretically in the case where the critical value of static friction differs the value of kinetic friction withous slipping. The kinetic friction force-velocity characteristic considered is given by a polygon having two straight line segments. The system consists of a table lying on a rectilinear sliding surface, an actuator cylinder, a 4-way servo valve, a relief valve and an oil pump. The stick-slip motions of the table are analysed considering the 4-way valve pressure-flow characteristic, friction-velocity relation, oil compressibility and the sizes of the hydraulic driving elements. Several types of limit cycles and the regions in which they occur are shown in figures according to the parameters mentioned above. Lastly, the curves of amplitudes and their periods of the steady-state oscillations are given.

View full abstract

This paper deals with a self-excited vibrating system with a dynamic damper accompanied by Coulomb's friction. The direct numerical integral method presented in the previous reports is improved to reduce the calculation time and is applied to the analysis of the self-excited system subjected to Coulomb's friction with discontinuity. By using the solutions obtained from the direct numerical integral method and the averaging method, an optimum approach for quenching of the self-excited vibration is discussed. Moreover, it turns out from the methods that a small beat vibration accompanying the stick-slip occurs when the vibration amplitude of the self-excited system is minimized, and the resulting waveform with beat contains many frequency components from the frequency analysis.

View full abstract

The numerical simulation is a suitable method to get the practical riding comfort of railway vehicles by calculation. Since the running simulation of the vehicle needs realistic track irregularity shapes, artificial generation of them makes the simulation method more efficient tool for improvement of suspension and running gear systems. This paper sums up the artificial generation from track irregularity PSD by the inverse Fourier transform and describes the process for prediction of the riding comfort taking an example of the lateral vibration of a Shinkansen electric car caused by alignment irregularities. In this process, the simulation is executed by the Runge-Kutta-Gill method, the acceleration PSDs of the car body are analyzed from simulation results by the Fourier transform and the Riding Quality Level L_T is, calculated as the index for quantitative estimation of the riding comfort.

View full abstract

Our previous paper reported the fundamental dynamic characteristics of cracked rotors, crack detection methods by static tests, and a numerical vibration-analysis method of general cracked rotors. This paper describes an example of the diagnosis of a cracked rotor based on former results. A wind-tunnel fan showed a large 2N vibration at 465rpm and was inferred to have a crack from the excitation test, and a shaft crack was detected. The analysis of natural frequencies and responses of the cracked rotor shows the adequateness of the method.

View full abstract

The present report describes an application of the method (growth-strain method) previously proposed for analysis of uniform-strength shape with the finite-element method to the problems of steady-state vibration, especially free vibration with a normal-mode shape and response to a harmonic excitation. A difference in the implementation of the method for these vibration problems from the method for static problems is to analyze growth-measure (strength-parameter) distribution on the deformation with normal mode or amplitude of a response. Numerical experiments are made on simple beam structures. In the Case of normal-mode vibration, monotonous convergence is constantly observed. On the harmonic response problem, while the natural frequencies do not vary monotonously, the maximum value of the strength parameter, the mass and deflection at the excita-tion point converge monotonously.

View full abstract

An analysis of wave-motion problems in unbounded or infinite domains is presented. Analytical solutions that express outgoing waves from an artificial boundary such as a circle or a sphere can be obtained easily. These solutions may be transformed into discretized relations between values of velocity potential and volumetric flux on the artificial boundary. These relations are in the same forms as the FEM equations. They can be joined directly to finite elements within the artificial boundary in order to construct an equation system for the whole domain. Some problems of radiation and scattering are calculated numerically by this method. The so-called fictitious interior eigenvalue failure does not arise in this method.

View full abstract

Vibration analysis of thick rotating cylindrical shells subjected to an initial pressure is described. The basic equations based on the Timoshenko type shell theory, including the effects of the initial tensions and resultant shear stresses due to rotation, torque, external pressure and axial compression load, are employed. By the modified Galerkin's method, frequency analysis is carried out for four boundary conditions (i. e., both ends (A) clamped with axial constraint, (B) simply supported without axial constraint, (C) simply supported with axial constraint and (D) clamped without axial constraint). Numerical results are compared with those based on the Flugge type shell theory and the difference between both theories is examined.

View full abstract

Application of the pseudo-inverse method has now become available by replacing the inequality behavior constraint conditions with equality behavior constraint conditions. Application was made to a large scale model having as many as 1508 design variables. Necessary calculations were completed in seven cycles of repetition in the case of equality behavior constraint conditions, and eight cycles in the case of a combination of equality and inequality behavior constraint conditions. The CPU time used for optimization was 0.3 minutes while the CPU times for eigen-value analysis and sensitivity analysis were 6.4 and 3.4 minutes, respectively. This model is that of an engine cylinder block and has been applied for the improvement of torsional vibration. The test revealed new information. We manufactured a prototype of a cylinder block on the basis of the final calculation model and the test result proved quite satisfactory.

View full abstract

This paper proposes an optimum method of design of dynamic vibration absorbers for a system subject to stationary random excitation with ideal white noise. The optimization criterion is to minimize a performance measure defined by the mean square acceleration response of the primary system to unit input acceleration. Optimum design parameters of the dynamic absorbers can be formulated if the primary system has no damping. If there is damping in the primary system, the optimum parameters are obtained from the solution of simultaneous equations. The optimum parameters obtained here are different from the former ones evaluated by the so-called Den Hartog's theory. This comes from the difference in the performance measure in these theories. The optimum parameters are tested on a vibratory model. The experiment makes, it clear that our theory is very useful for protecting the system from vibration induced by random excitation.

View full abstract

Microtremor isolation is important for improving the yield rate of complete IC chips. For this purpose, passive isolation devices using air springs have been used so far. However, as the degree of integration in IC chips increases and patterns of circuits become more fine, more complete isolation is becoming necessary. Therefore an active isolation system using linear motors was developed, in which the table supported by air springs was controlled by 4 vertical linear motor actuators and 4 horizontal linear motor actuators. The isolation system used a full-state feedback control system designed by the optimal regulator theory. The controller, implemented by a 16-bit personal computer, calculated 8 control inputs to the actuators from 12 signals made by integration of absolute accelerations detected by sensors, 6 of which were absolute velocities and the other 6 of which were absolute displacements. Through tests, it was confirmed that the active system could achieve superior performance compared with a passive system with high performance.

View full abstract

For control of wind-induced vibration of tall buildings having long natural periods of 6s, or so, tuned mass dampers must respond to the floor motions with small accelerations. Therefore the tuned mass dampers must have mechanisms/structures which enable them to oscillate with the least friction. To satisfy this requirement, a tuned mass damper using multistage rubber bearings was developed, in which the moving mass was supported by the rubber bearings and could move with very small friction. To investigate dynamic characteristics of the mass damper, excitation tests were carried out for a scale model of the mass damper. The model was designed to have an effective mass of 5000kg with a horizontal natural period of 2.0s. In the model a single multistage rubber bearing was used to support the moving mass, which consisted of 12 stages (each stage comprising 4 rubber bearing elements and 8 oil dampers) stacked up with 11 stabilizing plates between the stages. Through the tests, it was revealed that the mass damper had some nonlinear vibration characteristics. Analytical models to simulate the nonlinear vibration were developed.

View full abstract

In the design of a flexible off-Shore riser, which is connected to a spring at the top and clamped at the sea bed, it is desirable to reduce both the riser deflection and the bending moment due to random wave excitation forces. Assuming that the motion of the riser is two dimensional in the vertical plane, these design requirements can be effectively satisfied if appropriate stiffness and damping are given to the riser. In this study, we employed the statistical equivalent linearization technique to linearize a nonlinear fluid drag force, and an approximate mode superposition method (Rayleigh Ritz's method) to transform the governing equation of motion into a set of discrete finite dimensional state equations. A sub-optimal control theory with control structural constraints is applied to determine appropriate stiffness and damping. The required optimal stiffness and damping are determined by a trade-off between the rms values of the lateral displacement and the bending moment of the riser.

View full abstract

A continuum can be controlled by a small number of sensors and actuators without causing observation and control spillover. The time response is determined by solving integral equations with respect to only the joint displacements of member beams. The integral equations can be solved with little computation time by transforming them into ordinary differential equations and by introducing additional mass corresponding to the inertial force due to the rigid-body axial translation and rotation of member beams. The proposed method is advantageous with regard to precision, computation time and resolution of the spillover problem compared with conventional methods resorting to FEM. Numerical examples deal with the control of the position, attitude and elastic deformation of a space structure and a manipulator arm, a vehicle frame supported by nonlinear springs and a truss with a dynamic absorber used for economizing the energy consumed for the active control.

View full abstract

This paper describes dynamic test results of high damping rubber bearings and discusses analytical models for the hysteretic restoring force. The dynamic tests were carried out for two types of rubber bearing which had a difference in hardness of rubber. The two types, the rubber bearing of 588kN rated load (the low hardness type) and that of 784kN (the high hardness type), had the same shape, and comprised of 23 rubber sheets of 6mm thickness and 450mm diameter bonded to steel plates. Each type of the rubber bearing was designed to provide the rated mass (60000kg or 80000kg) with a horizontal natural frequency of about 0.5Hz and an equivalent damping ratio of about 15%. For each type of the rubber bearings, the test results revealed various properties of the horizontal and vertical stiffnesses and equivalent damping ratios obtained from the hysteretic restoring force loops. Three analytical models were proposed to express the horizontal hysteretic restoring force, and appropriateness of each model was evaluated.

View full abstract

This paper describes the results of breaking tests of high damping rubber bearings. The breaking .tests were carried out for 0.5 scale models of two types of rubber bearing which had the same shape but differed in rubber hardness. The full-scale rubber bearings were tested to investigate their hySteretic restoring force characteristics, which were mentioned in the 1st report. Biaxial tests by compression and shear showed that the breaking shear strains gradually decreased with increased compressive stresses. The breaking shear strains under various compressive stresses were larger than 600% for the low hardness type and larger than 400% for the high hardness type. These breaking strains are large enough for reliable seismic isolation of buildings. Biaxial tests by tension and shear revealed that the breaking tensile stresses and strains were decreased with increased shear strains, and that the rubber bearings had considerably hard stiffnesses in small tensile strains. For simple tension without shear, the tensile stiffnesses were about 1/3 of the compressive one for the low hardness type and about 1/4 for the high hardness type.

View full abstract

The static and the dynamic characteristics for two types of damper seals with triangular hollows are experimentally investigated. The dynamic characteristics are measured for various parameters such as preswirl velocities. pressure differences between the inlet and outlet of seals, whirling amplitudes, and rotor rotating speeds. The results are also compared with those of the smooth seal obtained previously. As a result, the leakages of damper seals are only about 2/3 those of the smooth seals, and for a given surface pattern, there may be an optimum hollow depth which gives minimum leakage. The damper seals have smaller radial reaction forces and larger tangential reaction forces compared with the smooth seal. The effect of the preswirl velocities on the dynamic characteristics of damper seals is obviously smaller than that of the smooth seal.

View full abstract

This paper presents a measuring method of impact force and velocity at the top of a bar which has discontinuous cross-sectional areas and properties. These values are calculated from the stresses measured at a suitable point on a uniform part of the bar. This method can be used in the high-strain-rate deformation test of materials at high temperature and any condition which can be applied the theory of the propagation of longitudinal elastic stress waves.

View full abstract

The influence of layer arrangements of moving links upon the dynamic characteristics of mechanisms is one of the important problems which must be solved for the improvement of performances of mechanisms. In this paper, on planar mechanisms consideration is given to layer arrangements of moving links, and equations of motion of the mechanisms taking account of elastic deformations of links are formulated by dividing moving links into several components which consist of distributed mass. Applying the equations to planar four-bar mechanisms whose coupler and follower are elastic links, we discuss the influences of the profiles and elastic deformations of links and the differences of layer arrangements of moving links upon the link displacements from the ideal planes of motion of links together with the results of experiments.

View full abstract

In the author's previous research work it was left unsolved as to whether the linear theory predicts the hunting behavior at lower speeds. The present paper investigates the nonlinear behavior of hunting both on the basis of linearly calculated results and nonlinear computer simulation, the reason why the engine speed fluctuation diminishes at lower speeds where the linear theory gives the instability of an equilibrium state is that a stopper for the maximum value of rack movement constrains the higher frequency component of the control rack caused by the suction stroke of each piston and both the computer simulation and the experiment yield the occurrence of hunting even at extreme lower engine speeds when the maximum stroke of the control rack is increased, and the limit cycle amplitudes can be explained well qualitatively by the linearly calculated logarithmic increment of amplitude even at extreme lower speeds by noting the higher frequency component of the control rack and the nonlinearity of torque characteristics.

View full abstract

This paper deals with the expert system for the recognition and counting of cells using image processing. The biological cells are taken from animals and cultured on microcarriers, plastic balls/beads 150-200 micrometers in diameter floating in a culture solution. Cell images are obtained through a microscope attached to a TV camera. The number of cells in the culture vessel increases during a 6=day, and the expert system recognizes the state of cell growth during this time by means of image processing. In particular, the expert system is a more effective way to recognize and cluster cells on densely grown/distributed beads from the gray image data than simple image processing. In this study, we also applied the creature-image-segmentation method, which has previously been proposed by the authors. With this method, we can recognize the cell shape and location, avoiding individual differences on the plastic balls. The effectiveness of this method is shown by experimental results.

View full abstract

The dynamically reconfigurable robotic system (DRRS) is a new kind of robotic system which is able to reconfigurate itself to an optimal structure depending on the purpose and exvironment. To realize this concept, we proposed the CEBOT (cell-structured robot). Communication is needed in the CEBOT system as follows. When cells are separated, a communication master cell needs to know the other cell's function and position and determine the target cell for docking. Mobile cells should be able to coordinate with other mobile cell. When cells are docked, forming a cell structure/module, a master cell should control the bending joint cell and know which cells the construction is composed of. In this paper, we propose a communication protocol for both cases with optical sensor applicable to CEBOT. Some experimental results are shown by realizing the proposed communication method between cells.

View full abstract

This paper is concerned with the problem of determining the feedback gains of robust tracking control systems. In recent years, many papers have treated the robust tracking control problem, but the transient characteristics of the response have hardly been taken into consideration. In this paper, therefore, a numerical method for determining the feedback gains is proposed paying particular attention to the transient characteristics. Moreover, we make use of the output feedback control law since the only output may be measured in some real systems. The method is mainly based on the eigenvalue assignment, eigenvalue and eigenvector sensitivities. An example shows that the proposed method is effective and gives considerably good transient characteristics for all values of uncertain parameter.

View full abstract

The authors have developed a variable-structure-type locomotive robot which can undertake three variations of locomotion according to the working environment or application of the robot. The first variation is an arm-wheel-type, the second and the third are an arm-leg-type and a leg-leg-type, respectively. This paper reports the control methods and experimental results on the postural and driving control of the arm-wheel-type locomotive robot which has a twin wheel, an upper structure on its driving axis and double controlling arms suspended from the upper body. In order to stabilize the posture of the robot, feedback control based on the compound gravity center between the body and the arms has been proposed to control the arms, with which we have succeeded in establishing the postural stability of the vehicle. Furthermore, proportional and integral feedback controls concerning the arm angle or the inclination of the body together with the gravity feedback have enabled the body inclination to be controlled while following a prescribed path. Standing rest and servo-driving control of the vehicle were also attained by controlling the wheel independently from the arms.

View full abstract

Recently, adaptive control strategies have been developed theoretically. Meanwhile, rapid progress in micro processor techniques has enabled us to simulate complex and large-scale systems. We can see that unstable zeros may appear when a digital control system is designed for a continuous system. Clarke proposed a design for a self-tuning controller (STC) based on generalized minimum-variance adaptive control for such a system. However, this design does not guarantee that the output can track the reference signal. Therefore, we propose a new algorithm to improve its capability for tracking the output to the reference signal. In this paper, the STC algorithm is applied to the pressure control system. First, we show the generalized minimum-variance adaptive controller proposed by Clarke and an algorithm for improving the tracking property. Next, we construct a pressure control system and we show that the STC algorithm is more effective than the usual PID control algorithm. Finally, it is shown that pressure control by the proposed algorithm is more effective than Clarke's method in tracking the output to the reference signal.

View full abstract

In this paper, the method used to determine the control parameters concerned with a 3 d. o. f. robot arm which is actuated by hydraulic direct-drive motors equipped with servo valves is discussed. First, the mathematical model of this robot arm is formulated. Next, the design rule of SISO (single-input, single-output) 3-order system and Bode diagrams are utilized at the same time for the control design of one arm. Moreover, the stability and the sensitivity of this dynamic coupling system are discussed with generalized Gershgorin Bands of a MIMO (multi-input, multi-output) system. Finally, the numerical simulations and the experiments of this robot arm are executed; the effectiveness of this design is also verified.

View full abstract

Tooth bearing patterns for hypoid gears cannot be described by conventional theory. In gear manufacturing factories, tooth bearing development is still required in order to obtain good tooth bearing patterns. This paper presents a new method for the development of hypoid tooth bearing patterns which differs from the conventional and experimental gear cutting methods. This report makes it possible to obtain accurate hypoid gear dimensions which consist of the mating pressure angles, spiral angles and pitch angles, pitch cones, and principal directions of the gear tooth surfaces. These dimensions are required in the drawing of the tooth bearing patterns.

View full abstract

The influence of dynamic load on storing capacity was studied using gears having tooth profile errors on the recess sides. Scoring capacity in spur gears is more affected by specific sliding than by dynamic load and increases as specific sliding is decreased. In the mixed lubrication state, the scoring bulk temperature of the working tooth flank is barely influenced by dynamic load when the tip interference on the approach side is negligible, and it may be well estimated by the bulk temperature on the outer surface of the cylinder immediately before seizure in two-cylinder tests. In the calculation of the bulk temperature of the working tooth flank, it is essential to take account of dynamic load and tip interference as well as the heat transfer coefficient and friction coefficient.

View full abstract

The oil flow in the porous media is experimentally investigated to explicate the mechanism of lubrication in the porous journal bearings. For the test bearing, a simplified model of the porous bearing is used, whose matrices are composed of the packed glass spheres having small uniform diameter. To visualize the flow in the porous media, a dyed oil is fed into the porous media impregnated with uncolored oil and the fluorescent flow is rendered visible by the ultraviolet flash lamp. It is found that there exists the circulation of the Oil flow through the porous media, which the flow contributes to the lubrication in the porous bearings. In hydrodynamic lubrication conditions, the oil in the porous media flows away from the position of the load line towards the unloaded region. However, in boundary lubrication conditions, when the oil supply pressure is relatively small, the oil in the porous media flows from the loaded region towards the site where the oil film pressure would be negative.

View full abstract

A numerical analysis of the elastohydrodynamic lubrication under high speed and high load has been made for a line contact using a modified Houpert-Hamrock's method in order to examine the characteristics of the lubrication region which is called 'rigid-variable viscosity' in Hooke's chart. Further, the influence of inlet pressure on the lubrication characteristics has been investigated. Results show that the effect of elastic deformation on the film is significant and thus the film thickness in the region is further insensitive to the load than the film thickness formulated by Hhmrock-Jacobson from numerical solutions in the usual elastic variable viscosity (E-V) region. Also, the inlet pressure in the examined region exerts a marked influence on the lubrication characteristics in comparison to that in the typical E-V region and the pressure spike point moves to the inlet direction contrary to the case of the typical E-V region as the load increases. Therefore, the E-V region in Johnson's charts can be separated further into the two regions.

View full abstract

The elastohydrodynamic lubrication characteristics of an oil hydraulic vane tip contact with a cam ring have been analyzed numerically under various conditions using a modified Houpert-Hamrock's calculation method provided in the previous paper. Results show that the minimum film thickness in the suction port section is differs little from that in the small arc section and the minimum film thickness in the large arc section is greater than that in the small arc section. Nevertheless the contact load in the suction port section is 2.5 times that in the small arc section for the case of an intravane type and the load in the large arc section is greater than that in the small arc section. The combination of inlet pressure and nonlinear piezoviscous effect causes the increase in oil film thickness in the large arc section. Therefore, it is suggested that the elevating discharge pressure will not cause any lubrication trouble if mineral oil is used as the hydraulic fluid ; the water-based hydraulic fluids present an extreme disadvantage to the vane tip lubrication.

View full abstract

The results of friction and wear tests for choosing piston ring and cylinder liner materials for the Stirling engine are described. Polyimide composites and a PEEK composite for piston rings, and several surface hardened metals and ceramics for cylinder liners were tested in a helium atmosphere from room temperature to 200°C. The polyimide composite had the lowest wear rates for the counter surface roughness R_max=0.5∼1μm and the counter materials hardly affected these values. The PEEK composite seemed to decrease its wear rate when the counter surface roughness decreased. The coefficients of friction were lowest for the cylinder liner material surface roughness R_max=0.5∼1μm and they seemed to be affected by these materials. The polyimide composite which contained PTFE and graphite was chosen for the piston ring material and a Cr₂O₃ coating was chosen for cylinder liner surface hardening.

View full abstract

Assuming that the micro-asperities on semicylindrical hard asperities with spherical ends ploughing through a softer surface, analyses ware carried out on the effects of the surface topographical parameters, such as the mean heights of the micro-asperities, radius of curvature of the tips of the semicylindrical asperities etc., on the wear rate of a soft surface. The theoretical trends of the wear rates varying with a normal load and the topography of the worn surface agreed with the experimental trends of those of Si₃N₄ pins rubbed on four kinds of ceramic disks. The wear rates of the Si₃N₄ pins and disks increased with an increase in the mean depths of grooves and r. m. s. roughness on worn ceramic surfaces. Their trends, which varied with a normal load, also agreed with those of the truncated effects on a probability density of surface heights.

View full abstract

The authors constructed "slipping adhesion test bogie" designed upon a new idea of measuring the coefficient of adhesion continually while running, and have observed the state of adhesion on main lines since 1985. Recently, they placed a car equipped with this bogie on a steep line where wheel-spin and skidding of rolling stock had been frequently the cause of troubles. As the result of runs of the car under various weather conditions such as fine, fog, snow, sleet and frost, influence of the weather condition upon the adhesion between rails and wheels has been made clear and evaluated quantitatively.

View full abstract

This paper deals with a knowledge based design system of the machine structures for machining on the basis of kinematic functions. Design parameters of the machine structures are established by analyzing the constituent elements of the machines. The design processes of the machine structures are defined based on the dependency of the design parameters. Experiential knowledge needed in the design of machine structures is obtained, and a method is proposed which presents this knowledge in the combined form of the frames and rules. A prototype design system has been developed, and some case studies have been carried out to verify the proposed method.

View full abstract

The present paper deals with a precision inserting operation into the stepped hole without chamfer. Since there is a step in the hole, the peg-in-hole becomes a hard operation. To accomplish this complicated assembly, it is necessary to combine both active sensing and passive compliance. First, for active sensing, a force/torque sensor is used to recognize the contact state between parts. A SCARA (selective compliance assembly robot arm) corrects the lateral error from data of the sensor. Second, for passive compliance, the RCC (remote center compliance) device is used to avoid jamming and wedging. In this way, the algorithm of inserting a peg into the stepped hole without chamfer is constructed. This algorithm has three features : the use of the upper hole, the restriction on operational directions of the peg and the point-to-point operation between parts. The usefulness of this algorithm is shown by experimental results.

View full abstract

Two-step indentation of punches is attempted in cold lap welding of aluminum sheets ; cylindrical punches are indented after preindentation of ring-shaped punches. In this method, the threshold thickness reduction of the sheet where bonding commences is smaller and the maximum bond strength is higher than those obtained by indenting the cylindrical punches used in the conventional method. Distribution 6f the virgin surface ratio (the area ratio of the virgin surface to the bonding surface) and that of the bond strength are measured on the bonding surface. It is found that these distributions are fairly similar and that a high bond strength in the present method depends upon a high virgin surface ratio.

View full abstract

The preceding paper proposed a method for determining the nodes, the inflection points, and the maximum-and minimum-curvature points of the coupler curves. In this paper, each of the solutions of the relative angle of the compound three-link chain for determining the double points of the coupler curves of W23 mechanisms is expressed as a function of the angular displacement of the driving link. The conditions that the driving links of W22 and W23 mechanisms can make a complete rotation, are formulated in two inequalities respect to the kinematic constants. According to the numbers of the nodes, the inflection points, the maximum-curvature points and the number of rotations of the curve, the coupler curves of W22 and W23 mechanisms are automatically classified into 166 and 163 kinds. These curves are illustrated schematically together with their appearance frequencies.

View full abstract

In the present paper, kinematic structures of spatial 6 degree-of-freedom (dof) parallel manipulators are obtained by a number synthesis of multi-loop linkages Which consist of pairs with l, 2 and/or 3 dof and have no moving closed-loops. About a simple and practical manipulator among the obtained ones, that is, about a parallel manipulator consisting of six arms, each of which has three revolute pairs and a spheric pair, analytical equations for inverse and direct kinematics are derived and singular positions where the manipulator gains an extra degree of freedom are discussed. Moreover, taking account of allowable pressure angles and maximum swing angles of spheric pairs, the relation between the working space and parameters of the manipulator is investigated, revealing that the manipulator can operate with larger vertical motions than the Stewart platform can.

View full abstract

This paper proposes a scheduling method for multi-item assembly which will permit a constant increment of the operation time at each station and reduce the blocking time in relation to traffic intensity. This method has three steps: Division of the flow time into subintervals of equal length, arrangement of the product mix to equalize the sum of assigned job times of each station among all subintervals, assignment of products with shorter due dates to earlier subintervals. Determination of a target value of traffic intensity for each assembly station. Determination of an input order of products in such a manner as to minimize the difference between the target value and actual traffic intensity of each station. This method was applied to the mixed assembly line of seven kinds of products, and its advantages were evident.

View full abstract

The stability criteria based on the root locus method in the magnetically suspended positioning mechanism with six degrees of freedom were derived using characteristic matrices. These describe the linear characteristics of the mechanism including the driven tetrahedron armature with load mass, controller, displacement sensors and magnetic actuators. The nonlinear electromagnetic force was approximately linearized. From these were derived linear expressions of dynamic characteristics on every equilible position, and was shown how every real parameter affects the system characteristics of the mechanism. Eigenfrequency variation along with the static equilible position travel of the tetrahedron was computed using the dynamic matrix equations, and was compared with experimental data obtained from the tetrahedron model fabricated for this purpose. The results were used for the stability evaluation of the mechanism in a wide operational range.

View full abstract