In turbomachinery, nonlinearity appears in restoring forces or damping forces due to various causes, such as clearances in bearings, squeeze film dampers, oil films in journal bearings, magnetic forces, seals, frictions and stiffening effect in elongation of a shaft center line. As a result, various kinds of nonlinear phenomena occur. For example, it is known that subharmonic oscillations, summed - and - differential harmonic oscillations, chaotic motions, jump phenomena, and limit cycles occur in rotor systems. In this review article, characteristics of these nonlinear phenomena are outlined and literature on these phenomena is introduced.

View full abstract

In the present study, a higher- and lower-order truncated modal and coupled structure-acoustic finite-element sensitivity analysis is performed to determine the rate of change of vibration responses of the body panels and sound pressure level at the position of the ear of a passenger/driver, and an inverse multi-objective optimization problem based on the structural synthesis method and pseudo-inverse method is solved to determine the required minimum changes of the design variables from the baseline design parameters to attain the selected design response in a shortest path. It is found that the structural synthesis method and the pseudo-inverse method converge to the solution when the step size at each iteration level of the optimization process is kept small.

View full abstract

In this paper we present a new controller design method that effectively suppresses vibration in flexible systems. The method allows integrated determination of sensor/actuator locations and feedback gain via minimization of an energy criterion which represents the integrated total energy stored in the system. This energy criterion is characterized by an effective representation of vibration, determined via an efficient solution of the Lyapunov equation and minimized with a quasi-Newton or recursive quadratic programming algorithm. The application of this method to a simple flexible structure confirms the direct relationship between our optimization criterion and effectiveness of vibration suppression.

View full abstract

An algorithm is proposed for determining the exact solutions for free vibration in an asymmetrical Duffing oscillator. The system is composed of a spring function of the Duffing type (with linear and cubic terms with respect to displacement) accompanied by a constant force. It is reduced from the most comprehensive 3rd-order polynomial having arbitrary terms from a constant to the 3rd power. By employing a bilinear transformation, the system is successfully converted into a regular Duffing equation whose exact solutions already exist and are abstracted here. The whole procedure of obtaining the present solution is summarized in a flow chart. Numerical values of conversion constants, skeleton curves and waveforms are computed and illustrated for the typical cases of asymmetrical (hard, soft and snap-through) spring systems. It is demonstrated that the skeleton curves and waveforms are asymmetrical and that some of them exhibit multivalued responses for a given frequency.

View full abstract

This paper presents a method for solving the dynamic stress concentration problem of inhomogeneous rods having two-dimensional arbitrary curvature, variable cross section and infinite length subjected to in-plane bending wave excitation. In this analysis, the exact solution of the equilibrium equations for curved rods has been obtained, and the transfer matrix, based on the exact solutions with consideration of the inertia forces has been derived. At discontinuous sections, solutions of curved and straight rods have been connected by adjusting the boundary conditions. As examples, stress concentration factors in circular, elliptical and parabolic arc rods with variable cross sections have been calculated.

View full abstract

Welding is widely used for construction of structures. This paper deals with the effect of welding on respnse characteristics and reliability of structures. First, the damping ratio and the natural frequency of a welded specimen are determined experimentally. Second, by using the obtained results, response characteristics of a mechanical model used to simulate a specimen subjected to stationary random excitations are examined. Third, the effect of welding on reliability is examined through estimation of the first passage probability. It is concluded that the effect of welding manifests as an increase in the damping ratio and as a decrease in response and failure probability. It is also concluded that the effect of welding on the natural frequency is very small. Finally, the applicability of the obtained results to actual structures is examined experimentally. In the experiment, the damping ratio increases when welding is used.

View full abstract

For continuums such as elastically vibrating substances, e. g., beams and plates, mathematical models are discussed with respect to the major problems involving the discretized formulations : (1) To decrease the order of spatial dimension. (Cantor's theorem of the infinite set theory and the continuous cardinality are referred to help imagining the idea.) The example of decreasing is given to show how the principle applies by means of the comparison of two cases which are formulations of modal controls of bending vibrations. (2) Finite freedom forms for infinite eigenmodes of distributed parameter systems. An approximating method using spectral decompositions and resolvent representations of the linear operator is shown. The necessities of the conditions for the state observability and the state controllability in the designs of finite freedom controls provided with the eigenmodes for continuously distributed objects are pointed out.

View full abstract

To reduce unbalance response of a flexible rotor suspended by electromagnetic bearings, two methods have mainly been investigated. One is the method of increasing damping of the mechanical system ; the other is the method of cancelling out unbalances. For the former method, PID control is generally used. Although PID control is very effective for reducing resonance vibration, in other ranges of rotational speed it yields little effect. On the other hand, with the latter method of open loop control, it is expected that unbalance response is greatly attenuated in all ranges of rotational speed. In this paper, we combine PID control and open loop control with a variable gain against rotational speed. PID control is needed to stabilize the rotor system and open loop control is used to reduce unbalance response. From calculated and experimental results, we find that the above method has greater capability of suppressing vibration of a flexible rotor than does ordinary PID control.

View full abstract

New measuring methods for the wheelset angle of railway vehicles on board and on the ground have been developed by the Railway Technical Research Institute in Japan. With these methods, the wheelset angle of attack has been measured on curves in field running tests of limited express trains on a narrow-gauge commercial line. The measured results are as follows. (1) The measuring accuracy of these methods attained the targetted levels. (2) The higher the velocity and the larger the curve radius, the smaller the steady angle of attack. The steady angle of attack of the first wheel is 0.28 degrees, and that of the second wheel is 0.02 degrees for a curve radius of 400 m at a velocity of 110 km/h. (3) The measurement and simulation results for the wheelset angle of attack are in agreement.

View full abstract

This paper presents a new differential method called the mixed method, which was developed to measure the profile accurately under the circumstances of on-machine measurement. In comparison with the 2-point method, this method can more remarkably reduce the datum errors due to scanning, vibration and thermal drift. This method also has the advantage of obtaining the high-frequency component of profiles accurately. An optical sensor was designed and constructed to realize both the 2-point method and the mixed method. With a modulation technique, the sensor has been made to have good thermal drift characteristics. It can measure the displacement and angle of two points on a mirror surface simultaneously with a resolution higher than 0.1μm and 1". Using this sensor ; an on-machine measurement system has been constructed, and measurements of the straightness of cylindrical mirrors have been attempted with the 2-point method and the mixed method. Comparing the results, the favorable characteristics of the mixed method have been confirmed.

View full abstract

A fuzzy expert system for abnormality diagnosis is developed for all component equipment at a refuse incineration plant. Since fuzzy logic is applied to the reasoning function, diagnosis can be made even in the case of abnormality under unstable conditions of the plant and under transitional conditions, and this system is more flexible than those of the past. At the same time, patterned diagnosis knowledge by membership functions provides a good effect to the human interface, compared with word expression. For example, easy acquisition and modification of the knowledge and easy understanding of the diagnosis process are attained. The results of diagnosis are output audiolly and visually and provide information for removal and correction of the cause of abnormality. This system is in actual operation at several plants.

View full abstract

In repetitive tasks of an automated handling mechanism, the access motion speed is often limited by the average temperature rise of actuators, especially when the mechanism is directly driven by electrical motors. Thus we developed a method to derive the minimum-time trajectory of a multiple-degree-of-freedom system under the average heat generation restriction. This method is divided into two steps. In the first step, the motion trajectory is determined so as to minimize the cost function including both traveling time and average heat generation of all motors weighted by Lagrangian multipliers. This step is performed based on the Hermite polynomial expression approximation. In the second step, the optimal Lagrangian multipliers are chosen so as to minimize the traveling time. Two typical results of a numerical analysis of a two-degree-of-freedom planar manipulator are presented.

View full abstract

This paper describes torque control of a geared DC motor. Torque sensory feedback is beneficial for fine torque generation using geared DC motors. However, it reduces the structural stiffness and results in low response in both torque and position controls. To improve these aspects, the proposed method employs positive feedback of the link angle acceleration. To enable high performance of this positive feedback, the concept of an acceleration controller, realized by a disturbance observer, is also employed. The proposed method was confirmed to be effective in constructing a fine torque control system while improving the structural stiffness reduced by the torque sensor. The paper also reports the experimental results from implementation of this controller.

View full abstract

In this paper we study model-following control of manipulators with elastic joints. In the dynamic model of manipulators, the function of the joint flexibility may be nonlinear. A simple design procedure based on the notion of the Lie derivative is given. Checking of the involutivity conditions is not required. The nonlinear system will be linearized by the nonlinear state feedback and nonlinear state transformation. It is shown that the output converges to the model asymptotically, and the state is bounded. It is also shown that the proposed control method can be applied to a broad class of nonlinear functions of the joint flexibility.

View full abstract

A learning control algorithm using neural networks is proposed for grasping and movement of an object by a pair of direct-drive (DD) robots with two degrees of freedom. The proposed algorithm has three feedback controllers and two neural networks. After the completion of learning, the outputs of the feedback controllers are nearly equal to zero, and the two neural networks play an important role in the control system. Therefore, the optimum setting of control parameters is unnecessary. In other words, the proposed algorithm does not necessitate any knowledge of the controlled system in advance. The effectiveness of the proposed algorithm is demonstrated by the experiment on the cooperative control of the parallelogram-type DD robots. It is also shown that the force of gravity can be compensated by this algorithm.

View full abstract

Organisms are quite functional and the application of dynamics in organisms to the field of engineering is very instructive. To create an artificial propulsion mechanism that functions by bending waves, a simulation study of a propulsion mechanism which imitated the bending motion of organisms in water was carried out. As the bending wave motion, flagellar motion, which depends on Stokes law, was employed. In the simulation, the bending wave of the organism was changed into a construction of the multilink mechanism. This simulation model is defined as the "multilink model". In addition, the simulation model based on the approximate theory of the organism's propulsion is defined as the "approximate organism model". The thrust force ratio and the swimming speed ratio of the multilink model to the approximate organism model were calculated. The relationships between the two ratios and both the number of links and the ratio wavelength/amplitude were clarified.

View full abstract

To avoid the influence of small displacement of gear axes under load, the authors propose that all contact normals of a tooth surface should be at right angles to the direction of the small displacement induced at each contact point under load. On the basis of this state of surface normals, a few methods of tooth surface generation have been devised for realizing the concept. In this research, an hourglass worm gear was made using one of the generating methods, and tested experimentally. As a result, the validity of this designing concept was substantiated.

View full abstract

The run-out characteristics of a rotor supported by an externally pressurized gas journal bearing are analyzed theoretically with special attention paid to the relationship with the machining errors of the bearing system. Both the static and dynamic run-out characteristics are discussed. The form deviation of the rotor cross section (i. e., out-of-roundness error) is attributable as the critical factor which may govern the run-out characteristics of the rotor. The effects of size deviation of the supply holes are also investigated. The obtained results may reveal a fundamental concept of the precision design of an externally pressurized gas bearing with respect to the rotating accuracy of the rotor.

View full abstract

In the course of this study, the run-out characteristics of the externally pressurized gas journal bearings will be analyzed theoretically to establish the fundamental concept of precision design of the gas journal bearing, with special attention paid to the relation between the run-out of the rotor and the machining errors of the bearing. For the first step of the study, the modified divergence formulation (DF) scheme to calculate the exact solution of pressure distribution in the bearing clearance is introduced, in which the conventional DF scheme is improved so that it is applicable to the case with point-source of supply holes. As examples of the application of this scheme, the effects of machining errors such as roundness error of the rotor on the bearing characteristics are discussed.

View full abstract

In this paper, the static characteristics of a new type of aerostatic thrust bearing with a self-controlled restrictor is proposed in order to achieve very high static stiffness (nearly infinite stiffness). The self-controlled restrictor used in this bearing makes use of the balance of fluid-film forces acting on both sides of the floating disk to control the gas flow entering the bearing clearance. Since various design parameters influence the static characteristics of the proposed bearing, the effects of these parameters are discussed theoretically and experimentally. Then the optimum design method to obtain higher stiffness is also discussed. It is consequently demonstrated that the proposed bearing has very high static stiffness, about ten times as high as that of conventional aerostatic thrust bearings.

View full abstract

Structuralism as a philosophical and theoretical orientation emerged in the 1950's and boomed in the 1960's. This study attempts to view machine design systems by using structuralist notions as a guide to system thoughts. The method using the concepts of duality, and the decomposition-composition technique is proposed for the creation of new designs.

View full abstract

The thermal inertia of a machine tool bed made from concrete material is generally much larger than that of conventional ones such as cast iron or steel. Hence, it is considered that the thermal response of a concrete bed depends strongly on the period of fluctuation of the ambient temperature. In this paper, the thermal behaviour of a concrete bed in fluctuating ambient temperature is discussed, comparing it with that of a conventional cast-iron bed. It is verified that a concrete bed has a great advantage from the viewpoint of thermal deformation when the period of ambient temperature fluctuation is relatively short.

View full abstract

To build an expert system, especially in the mechanical design field, a knowledge base which contains a wealth of knowledge is an important issue for improving the efficiency of the design process. However, it is very difficult to build a knowledge base in mechanical design because there is a great deal of complicated knowledge in this field. In this work we derive an approach to build large knowledge bases in mechanical design by using an extended daemon function. Using this method, we believe it becomes possible to realize the characteristics of data base systems (DBS) in knowledge base systems (KBS). As a result, KBS will be suitable for managing and representing knowledge, and it will become easier to build a knowledge base in the mechanical design field.

View full abstract

In this paper we propose a new facility layout procedure with two phases. In this floor plan placement problem, a placement unit is a rectangular block which can have various sizes. In the first phase, the initial block placement is obtained by an attractive and repulsive force method. In the second phase, a quad-tree is constructed by dividing a placement area into four rectangular areas recursively as long as the area contains two or more block centers. The quad-tree preserves information about topological characteristics contained in the solution of the first phase. Then, a block packing process is performed heuristically by using the quad-tree. The solutions of this procedure maintain global suboptimality. This procedure allows, from the beginning, consideration of two constraints of floor dimensions and certain blocks specified by their locations on a floor. It is easy to perform the algorithm of the second phase manually for a fairly large-scale problem. The methodology is presented and demonstrated using example problems.

View full abstract

Mechanical properties under uniaxial and equi-biaxial tensions are examined for electronic copper foils and sheets ranging from 5 μm to 1 mm in thickness. A hydraulic microbulge test applicable to very thin metal foils is proposed for the equi-biaxial tension, and its validity is confirmed by computer simulation based on the Bassani yield function. The yield stresses in both the uniaxial and equi-biaxial tensions are dependent on the grain size, but not on thickness. The conventional yield criteria are shown to be applicable to very thin foils as well as to sheets with usual thickness. However, the fracture behavior varies according to the thickness, and the limit strain and ductility become markedly low for very thin foils.

View full abstract