The nominal film thiekness in the elastohydrodynamic lubrication (EHL) regime is of the same order of magnitude as the surface roughness of contacting bodies. Since the breakdown of EHL films seems to be associated with asperity interactions, a detailed description of local fluctuations of EHL fims and pressures caused by asperity interactions provides an important basis for a clearer understanding of the working perforrnance of machine elements, the causes of surface failures and methods for their prevention. In this review, why and how EHL films are influenced by the surface kinematic conditions and the orientation and shape of surface irregularities are discussed based mainly on the experimental facts using model asperities. The results described will provide a clue to solving EHL problems in the mixed regime between full fluid film and boundary lubrication.
This paper presents a 3-dimensional seismic isolation floor developed to protect computer systems in nuclear power plants from strong earthquakes. The 3-dimensional seismic isolation floor consists of floor panels, a floor frame, horizontal coil springs, isolation-supporting devices and devices for controlling rocking motion. In order to examine the performance of the 3-dimensional isolation floor, vibration tests of an isolation floor model, on which an actual computer system was set up, were carried out using a 3-dimensional shaking table. Test results confirmed that the actual computer system was effectively isolated by the 3-dimensional seismic isolation floor.Moreover, an analytical model was validated against these test results. For the analytical simulation the floor frame was representdd by five lumped masses and the isolation device was modeled as 2-dimensional springs and dampers. Both the calculated acceleration time history and floor response spectrum agree well with test results.
To maintain the stability of a motorcycle running straight, the rider steers the handlebars with his hands and leans his body to the right or left on the seat in dead time (time lag) to the motion of the disturbances of the motorcycle's frame. Since these three movements relate to one another and it is difticult to analyze the motion of the motorcycl-rider system, we treat the motorcycle andthe rider as one rigid body. In this report, we studied the conditions necessaryfor stability and the rider's operation and control of the motorcycle running straight at low speed using the equation of motion, in which the movement of the center of gravity of the rider to the right or left, the reaction of the motorcycle to the rider's movement and the moment of inertia of the rider are Considered.
This paper investigates the positioning technology of a control system using an optimal control method. In this method, the control algorithm is based on minimization of a quadratic performance index, which consists of the position error and the weighted derivative of the error during a specific time interval. Minimization of this index allows us to obtain the optimal step control input command. It is shown that the control algorithm is efficient for positioning and simple for on-line implementation. The stability of the step control system is analyzed by the root locus method, and a means of selecting both the sampling period and the weighting matrix according to the specifications of the system is presented. Comparisons of the proposed control algo-rithm with bang-bang control and the optimum regulator problem are made. Also, a track-seeking servo control system of an optical disk driver is used to illustrate application of the proposed control method.
The robustness of pole assignment in a specified region for perturbed large-scale systems is investigated using the Lyapunov approach. Two theorems are proposed to guarantee that all the eigenvalues of the overall perturbed large-scale system are located in a specified region. The decentralized controllers from local state feedback are designed to satisfy the requirement of robust pole assignment. An example is given to illustrate the proposed method.
This paper proposes methodologies for analyzing the motions of a human arm during specific work activities. First, a musculoskeletal model of a human upper limb is constructed. Next, assuming that a muscle consists of a contractile element and an elastic element, the energy consumed by the muscles is formulated as a function of the muscular forces and the contraction velocities of the muscles. Then, the muscular force sharing is predicted by solving an optimization problemin which the summation of the energy consumed in all of the muscles of the upperlimb model is minimized. Finally, the optimal motion of a human body is defined as that where energy consumed in the muscles during working processes is minimized. Then a method to determine the optimal motion of the upper limb model is presented for fundamental work activities.
The use of hydraulic proportional valves in industry is becoming increasingly popular as they have better performance than conventional on-off valves and lower cost than servovalves. However, due to significant time delay in valve dynamics, they are currently used for low-accuracy and low-dynamics applications. In this study, an adaptive self-tuning controller is proposed to enable a hydraulic proportional valve to achieve accurate set-point flow rate control. A moving model which reduces the system order and hence reduces the identification effort was used for the controller design. The controller is based on the pole-placement technique which shifts the dominant pole to the desired location. In addition, a predictor has been proposed to overcome the effect of the time delay. Experimental results have shown that the proposed self-tuning controller is better than the conventional PI controller as both transient oscillation and steady-state error shown in the PI control have been significantly reduced. Also, the performance of the closed-loop system is very robust as the system response remains the same under Various operating conditions.
This paper describes an approach for estimating unsteady flow rate through oil hydraulic pipelines and components in real time. In this method, unsteady flow rate is estimated by using the hydraulic pipeline dynamics and the measured pressures at two distant locations along the pipeline. Unsteady flow rate in the time domain is obtained by convolutions between the weighting functions and the measured values of pressure.The weighting functions are obtained from the transfer functions by application of the inverse fast Fourier transform (IFFT). By scheming the high-speed operations of convolutions, data processing in a microcomputer is achieved in real time, using exclusive precalculated tables. Under unsteady laminar oil flow conditions, the good agreement between the estimated and measured flow rate waveforms illustrates the validity of the method proposed here.
We have developed a variable-structure-type locomotive robot which can employ three variations of locomotive style according to the working environment or application of the robot. In a previous study, we described the concept of the variable-structure-type locomotive robot, the controlling methods and the experimental results on the first variation. In this investigation, the leg-leg type, which is a type of quadruped walking robot, is reported. Two forms of walking are proposed. One is the "rub" gait in which each pair of legs is always in contact with the ground and walking is achieved by rubbing or slipping while moving the forelegs and hindlegs alternately. This locomotive style is advantageous in maintaining body stability. The other is the "bound" gait in whith two pairs of legs kick back against the ground simultaneously in order to propel the robot. It has been experimentally confirmed that both forms of locomotion are useful in the leg-leg type, and high-speed translation can be attained by the "bound" gait. The control method for the step width proposed here for the "bound" gait proved effective in providing the robot with a constant-width-step walk, independent of environmental conditions and slope of the ground.
This paper describes a new steering mechanism proposed by one of the authors(T.E.). This steering mechanism using noncircular gears has the capability of turning a carrier with a small radius. In order to facilitate the analysis of the steering mechanism, the concept of a virtual two-wheeled vehicle was introduced. As a result of analysis made by the other author (A.A.), it was proven that the noncircular gear for the steering mechanism is an elliptical gear of which the pitch curve can be represented by a simple equation. In addition, a method to apply this mechanism to wheeled robots, such as the carrier robots used in industrial production lines, was studied. It was confirmed that the concept of the virtual two-wheeled vehicle is very useful in the design of the steering mechanism.
This paper describes a system which enables fast 3-dimensional measurementof a human head by means of a slit ray projection method. The system developed here consists of a laser slit ray projector, two CCD TV cameras, a real-time video signal processor and a 32-bit computer. Due to the employment of this real-time video signal processor and a fast data-processing technique using a reference table, a 3-dimensional measurement of the human head can be Accomplished in 4.3 seconds, and all the image data are converted to 3-dimensional data in one second. Experimental data reveal the applicability of our system.
Robotic systems composed of identical units can be made flexible in size, degrees of freedom, and output motions in response to desired tasks. There are three basic characteristics which are desirable in such systems : (1) expandability, (2) perceptibility of unit's location, and (3) fault tolerance. These characteristics can be varied depending on the networking and information distribution scheme adopted. The present paper proposes a method of classifying and comparing networking schemes for such robotic systems based on the above-mentioned basic characteristics, and the optimal networking scheme is then clarified for a particular application. An experimental transfer robot composed of identical units, each of which has nineteen arms, was constructed, and the experimental results are shown to support the proposed method.
One specific feature of the aerodynamic sound produced at the face end region is that the radiation becomes equally weak by filling root spaces as by shortening the center distance. However, one can easily expect that such actions make the air flow faster, and consequently make the sound louder. This paperattempts to reveal the reason for such a feature. First, air flow induced by thepumping action of the gear pair was analyzed regarding a series of root spaces as volume varying cavities which have channels to adjacent cavities as well as the exit/inlet at the face ends. The numefical analysis was verified by the hot wire anemometer measurement. Next, from the obtained flow response, the sound source was estimated to be a combination of symmetrically distributed simple sources. Taking the effect of either the center distance or root filling into consideration, it became clear that the simplified model can explain such a feature rationally.
The effects of the combination of tooth flank hardness and tooth surface roughness on the surface durability of medium-hardness gears are investigated using a disk machine. The meaning of the difference in hardness between a pair of gears is discussed. In the case where the surface roughnesses differ between contact surfaces with the same initial hardness, severe wear and pitting occur on the smooth surface because the asperities on the rough surface are highly hardeneddue to the repetition of contact. Therefore, it is necessary that the initial hardness of the smooth surface is higher than that of the rough surface in a combination of different sufface roughnesses. This means that the hardness combination between a pair of gears should be determined in consideration of the final condition of the finished tooth surfaces of the gears.
The effect of an electric field applied between rubbing surfaces onfriction and wear characteristics was examined using a ball-on-disc testing apparatus under a mixed lubrication condition. The ball and disc were made of bearing steel with high hardness and plain carbon steel annealed with low hardness, respectively. The frictional characteristics were improved in an electric field where the disc was the anode, but with application of an electric field in which the disc was the cathode, the frictional characteristics were deteriorated. Improvement resulted from the formation of a passive-state film on the disc surface of the anode. The lowest limiting values of voltage and electric current possibleto form a passive-state surface film were found. Applying an electric field below the limiting values did not result in an effective passive-state film. The limiting values were also influenced by oxygen content in the oils tested.
The algorithm for the systematic detection of an embedded rigid structure in enumerating a basic kinematic chain has been developed. The main idea in thisalgorithm is the sequential removal of binary link strings from the given kinematic chain to simplify the remaining chain. The proposed algorithm is implemented in a program using the LISP language and applied to several kinds of kinematic chains, such as those with 1, 2, 3, and 4 degrees of freedom and having 2, 3, 4, and 5 independent loops, respectively. In most cases, the results are in good agreement with those presented by other investigators. However, there are some discrepancies for several cases in the number of kinematic chains. The proposed algorithm is fairly efficient in that it can be applied to some unexplored cases aslisted in the Results section. The structural data of each kinematic chain are also obtainable if needed.
Scoring-proof capability and frictional characteristics of oils containing solid lubricants such as molybdenum disulfide and graphite, a S-P-type EP additive and organomolybdenum compounds such as MoDTP and MoDTC were evaluated using atwo-roller machine under a step load condition. In the case of the addition of molybdenum disulfide or graphite, scoring load and frictional conditions, such a scoefficient of friction, surface temperature and the state of oil-film formation, were similar to those of the base oil in which those additives were not added. The addition of the S-P-type EP additive and the organomolybdenum compounds improved the scoring limits. These improvements can be attributed to the formation of substances which inhibit the occurrence of scoring. The substance is insulating and has a high coefficient of friction in the case of the S-P-type EP additivt, while it is conductive and has a low coefficient of friction in the case of organomolybdenum compounds. Two diagrams of the surface temperature-friction coefficient and insulating voltage-friction coefficient clarified the effects of these additives.
The frictional behavior of alumina ceramics was investigated at various temperatures up to 1200°C. The coefficient of friction decreased with increasing temperature and this dependence became more pronounced as higher contact pressure was applled. Wear loss at room temperature could be interpreted as beingcaused by one of two different behavior modes which have a rate difference of anorder of magnitude. At temperatures higher than 800°C, the wear loss was far less than that at lower temperatures and a smooth, flat friction surface, which is probably responsible for both the low friction coefficient and the small wear loss, was observed.
This paper describes the effect of crack-free laser processing of hot-pressed silicon nitride ceramics on fracture strength and residual stress. The fracture strength of the processed test pieces was determined by three-point bending tests. The residual stress in the processed surface was characterized by X-ray diffraction measurement. The strength of the laser-processed workpieces was reduced to 90-80% compared with that of the ground workpieces because of subsurface layer damage, which included the heat-affected layer and the residual stress layer. In particular, the latter greatly affected the fracture strength because the higher compressive residual stress layer generated during grinding was released by laser processing. It is concluded that the newly developed crack-free processing technique has great industrial possibilities.