A structural topology optimization technique which optimizes the frequency response characteristics is proposed in this paper. The natural frequency and the linear quadratic performance index with respect to the response are defined as the objective function. The density of each finite element is adopted as the design variable and the optimum distribution of the density is sought based on the quasi-Newton method. Minimizig the performance index against the white and the colored noise disturbances, frequency response shaping is achieved. In order to realize the actual structure, a method to separate the value of the density to zero and one, in which the penalty function to divide the density into the two values is introduced in the objective function, is also proposed.

抄録全体を表示

In this paper, the optimization of dynamic responses is presented using the hierarchical optimization method, which has been previously reported by the authors. The special features of the method are that the sensitivities of the design object to the eigenvalues of substructures are calculated, and the sensitive substructures are modified mainly to satisfy the design object. The hierarchical optimization method is applied to the minimization problem of dynamic responses of a rahmen structure with some constraints. The results from the presented method and the conventional method are compared in terms of accuracy and calculation time, and it was found that the present method is effective in practical applications.

抄録全体を表示

A computational scheme is presented for single or multiple eigenfrequency maximization of isotropic and composite plates. Eigenvalue extremization was achieved by means of an optimization process which sought to redistribute the material of the plate structure in an optimal way so that a bound on the total volume was satisfied. It was assumed that the plate structure possessed a repetitious microstructure and homogenization theory was used to obtain equivalent elastic moduli. The structural eigenvalues and modes were computed via a finite element analysis using a shear deformable laminated finite element which was also applied to discretize a single-layered isotropic plate. Sequential linear programming was employed to perform the optimization task. Numerical examples are presented for clamped and simply supported plates for which the natural frequencies were extremized independently as well as simultaneously. For isotropic clamped plates, the optimality of the obtained results was verified by discretizing the resultant topology with a set of finite elements, computing its eigenvalues and then comparing them with eigenvalues from a uniform plate having the same volume.

抄録全体を表示

This method is proposed for a FEM model which has a large number of elements. This minimizes the length of the modification vector in design variable space. In this paper, an expansion is introduced to satisfy unequal behaviour constraints and side constraints. Also, a new approach searching for the minimum norm of design variables is added. The calculations with a small model elucidate the effectiveness of this theory.

抄録全体を表示

We present a practical method of numerical analysis for optimization problems of domains in which natural vibration problems of linear elastic bodies are defined. In this paper, we apply the traction method that was proposed as a solution to the domain optimization problems to elliptic boundary value problems. The problems treated are those which determine the domain that minimizes a mass under constraints in specified vibration eigenvalues. Using the Lagrange multiplier method, we obtain the shape gradient functions for these domain optimization problems from the optimality criteria. A numerical analysis technique for the multiconstraint problems is also presented. We show the successful resolution of the problems of beamlike plates clamped at both ends.

抄録全体を表示

In the practical design process, we make up design models which can express the behaviors of the real design objects and we innovate the designs based on those models. However, making up accurate design models adequately sometimes becomes a hard burden to the designer and consequently such modeling process may consume the time for designers to have creative designs. In such circumstances, there exist some approaches which are making design models based on experimental data. Advantages of these approaches can reduce the burden of the designers. However, existed those approaches are not so robust to the errors made in experiments, and also they need some kinds of approximation method with some kinds of errors. To overcome these disadvantages, we need some methods which are robust to the experimental and approximation errors. We have been studying the optimum oriented method based on qualitative sensitivities. In this study, we will propose a method to derive qualitative sensitivities only by using model experimental data. And using those sensitivity, we will carry out qualitative optimization. We applied the proposed system to the design of shaft of the rotor as one of examples and show the effectiveness of the method.

抄録全体を表示

An approach for integrated optimum design of a structure and H^∞ control system using a genetic algorithm is proposed. The size and shape of the structure, and the sites at which the sensors and the actuators are set are adopted as the design variables. Since the sensors and the actuators are placed only on the nodes of the finite-element model in this study, the design variables are discrete. Therefore, in order to solve this problem, a genetic algorithm is used as the optimization technique. The approach is applied to vibration control of a planar plate. A frequency weighting function and mode weighting coefficient are used in the H^∞ control problem. The mode weighting coefficient enables control of multiple natural modes without increasing the order of the controller. Effectiveness and practicability of the approach are verified by both simulation and experiment.

抄録全体を表示

Optimum design of structures considering control performances is presented. This paper supposes that a structure is subjected to initial loads and time-varying disturbances and is controlled by an H_∞ controller to suppress the effect of disturbances. The purpose of the design is minimization of an objective function which represents the effect of the initial loads, under a constraint imposed on the weight of the structure. The optimum design using an H_∞ controller is compared with the design using an optimal regulator in order to demonstrate its effectiveness. A numerical example of a 3-D truss structure shows that the structural design using an H_∞ controller can attenuate vibrations caused by both initial loads and time-varying disturbances.

抄録全体を表示

In order to analyze frequency response problems in the higher frequency range, we have developed a component mode synthesis method, based on the mode superposition technique, which is capable of simulating a higher frequency range using fewer modal degrees of freedom. In this paper, it is expanded to the solution of a coupled structure-acoustic fluid interaction problem, and it is applied to the interior noise problems of the vehicle cabin. In the cases of analysis of higher frequency range, the required mesh size in the FEM models of the vehicle body becomes much finer than that in the acoustic model, in order to obtain sufficient accuracy. A technique based on the least square method is presented to connect the fine structural mesh with a coarse acoustic one on the interface having noncoincident nodal points of the structure and acoustic field. Several numerical examples are analyzed to verify the effectiveness of the present method.

抄録全体を表示

Linkages at high operating speed will produce high inertia loads which cause joint forces acting on the turning pairs of linkage to suddenly change. Because of the slight clearance in the pair, this change of joint force will cause the separation and collision of paired elements. This is the major cause of vibration and noise in the mechanism. Therefore, it is important to control the magnitude and the sudden changes of the joint forces acting on the pair. A nonlinear optimization technique is used in this paper to control the joint force. The objective function with the magnitude and the direction of joint force is considered and minimized by redistributing masses of links. A new method for design of planar linkage using a finite set of small masses is shown. Some kinetic characteristics accompanying by the optimization of joint force are also shown.

抄録全体を表示

The development of a dynamic structural optimization technique for largely deformed elastoplastic bodies can realize more rigorous structural design, such as an energy absorbing design for impact loading. In this study, an exact and direct sensitivity analysis technique of large dynamic deformation for the elastoplastic material governed by bilinear constitutive law is developed, based on the implicit differentiation method and Newmark method. A practical step-by-step calculation algorithm for the sensitivities of the large displacements and the elastoplastic stresses with respect to the design parameters is formulated. The validity of the sensitivity analysis technique suggested here is confirmed by implementing the numerical results.

抄録全体を表示

A method is proposed to estimate reliability of homology design in the form of sensitivity with respect to the uncertainties involved in the structural system under consideration. The uncertainties are supposed to be confined to convex domain of the uncertainty variables taken. The worst case of an error index devised to evaluate the degree of disturbed homologous deformation is estimated with respect to the convex model of the uncertainties, and is related to the expanse of the convex hull for the purpose of the sensitivity analysis of the design reliability. The validity of the proposed method is verified and exemplified by a numerical example concerned with bending vibration of a planar lattice frame analyzed by the finite element method, in which uncertain boundary conditions are taken into account.

抄録全体を表示

This paper considers the optimization of systems with uncertain parameters by using Min-Max programming. By adjusting the uncertain parameters which worsen a performance index and by selecting the design parameters which improve the same index, the problem can be formulated as a Min-Max problem. The system designed by the present formulation has robustness against fluctuations of the uncertain parameters. The sequential linear approximation method is applied to obtain nonlinear Min-Max problems. As numerical examples, we investigate a flight control system design for controlling an aircraft in the landing condition through a microburst which has certain velocity profiles and uncertainty in its size. The robustness of the control system against the uncertain microburst is demonstrated.

抄録全体を表示

Shape and topology optimization analysis using a homogenization method has received much attention because it can treat topological changes of a design domain. However, as far as dynamic problems are concerned, it has not been successfully applied until recently. Here we apply it to the reduction of interior noise in the coupled structural-acoustic system for the first time. We attempt to reduce the mean eigenvalue formulated by Ma et al. and the sound pressure level integral formulated by Kozukue et al. by using a homogenization method, and we demonstrate the validity of these combined methods.

抄録全体を表示

In the turbogenerator rotor set, the transient torsional vibration induced by short circuits in the generator and ground faults in the electric network by lightning strikes can lead to not only material fatigue but also occasional breakdowns of the turbogenerator rotor. In this paper, a dynamic damper is attached to the rotor in order to suppress the transient torsional vibration. Applying the traditional design approach, the shaft configuration is not modified to enhance the vibration suppression performance of the dynamic damper. This paper aims at proposing a simultaneous optimum design approach in which both shaft configuration and dynamic damper are designed simultaneously. In the numerical example, a model of the turbogenerator rotor set which is composed of HP, LP1, LP2 turbines and a generator is designed. The results indicate that the transient torsional vibration is well suppressed by dynamic damper, and vibration suppression performance is augmented by applying the simultaneous optimum design approach.

抄録全体を表示

Free vibration and buckling of rotating prestressed anisotropic cylindrical shells are investigated. Torque, external pressure and axial compression load are considered as the prestresses. The basic equations are derived from Hamilton's principle, depending on the Flugge-type shell theory. The equations of motion can be separated into the quasi-static and dynamic ones, i.e., the equations in the steady rotating state and those in the vibrating state. By the modified Galerkin's method, frequency analysis is carried out for four boundary conditions. Numerical results are compared with previous results for simple cases, and the validity of the present analysis is ascertained. Then the general cases with the influence of anisotropy and prestresses are analyzed, and those effects on the frequency and buckling characteristics are examined.

抄録全体を表示

A simple yet efficient shot impact damper consists of a bed of granular materials in a container mounted on a primary vibrating system. We have studied the performance of the shot impact damper which is applied to some kinds of vibrating systems subjected to harmonic excitations. This study clarifies the damping performance of the damper when applied to the self-induced vibration system with bilinear characteristics of friction force. The periodic motion of the self-excited vibration accompanied by the impact of the damper mass is determined by theoretical analysis as well as numerical simulation studies which utilize the Willson-θ method. The damping performances are obtained in terms of the damper parameters and self-excitations. The results show that heavier damper mass provides greater damping effect on the fully built-up vibrations, and that a lighter mass is effective at the onset of the vibration. Finally, we determined the optimum clearances under which the response amplitudes are reduced to a minimum for suppressing the vibrations.

抄録全体を表示

State equations of the natural vibration of a rotating blade are derived from Hamilton's principle. The state equations are based on the beam theory. Bending-shear deflection, torsion, longitudinal expansion-contraction, and warping are adopted as deformations. The following 6 factors are taken into consideration : (1) the discrepancy between the figure center and shear center in blade cross section, (2) rotary inertia, (3) centrifugal tension, (4) Coriolis force, (5) warping rigidity, (6) bending and shear rigidity variation due to twist-back of the blade body. Lateral-torsional coupled vibration is induced by the first and third factors, and lateral-longitudinal coupled vibration is induced by the fourth factor. The natural vibration analysis of a rotating blade should be divided into three steps : (1) the distribution of centrifugal tension, (2) twist-back of the blade body, (3) natural vibration. The second step determines the neutral position for vibration.

抄録全体を表示

This is a research on rigorous theory for the natural vibration analysis of blade, a geometrically and environmentally complicated beam. The reason why the term "rigorous" is emphasized here is that (1) all the factors related to beam deformation are taken into consideration, and (2) those factors are combined to form an organically unified theory of non-contradiction, which is fully confirmed. In the 1st report, such a unified theory has been formulated in the form of a set of state equations. In this 2nd report, they are converted into a set of transfer matrices, assuming lumped inertia. The reason of acceptability of this assumption is that the order of major blade vibrations of engineering importance is much lower than the number of lumped inertia. The reason of beneficialness of this assumption is that it separates inertia, Coriolis and centrifugal-force matrices from stiffness matrix to enable theoritical development of good perspective.

抄録全体を表示

In order to show the features of the method of analyzing blade natural vibration described in the first and second reports, it must be demonstrated that the inertia, stiffness, centrifugal force and Coriolis matrices, which are equivalent to the transfer matrix of the second report, are symmetric, symmetric, symmetric and antisymmetric matrices, respectively, and that the inertia matrix and the total stiffness matrix-the latter is the sum of the stiffness and centrifugal force matrices-are positive definite matrices. It also means that natural frequencies are positive real number, and that each natural mode shape is orthogonal to the others. In addition, the result is effective in ascertaining that the selection of state variables and the treatment of centrifugal tension resistance are appropriate. In the process of the above proof, an equivalent method for symmetry verification is proposed and used, in order to make symmetry verification easier.

抄録全体を表示

A new type of multi-tuned mass damper, which is designed to function well during both weak earthquake/wind and destructive earthquakes, has been investigated. This system uses impact forces from the primary mass and the added mass to reduce the displacement of the primary mass. From the results, it is confirmed that the performance of vibration control is about the same as that of optimally tuned mass dampers, although the displacement of the primary mass is markedly reduced to about half that generated by the TMD. In the practical design of vibration attenuation systems such as TMD or AMD, the stroke of the damper is one of essential problems to be considered to satisfy the seismic requirements for destructive earthquakes. This paper presents one of the "break-through-techniques" for such problem. In addition, using an actuator to control the added mass, it is confirmed that a smaller capacity of the actuator is required to obtain the same performance as that of the optimally controlled AMD.

抄録全体を表示

Vibration reduction of a flexible structure by controlling the stiffness of the supporting spring is described in this paper. The control is carried out based on the velocity feedback, and independent modal space control (IMSC) is used to calculate the spring stiffness taking into account the feedback of each mode. The spring with controllable stiffness is fabricated using a beam of rectangular cross section whose setting angle is controlled by the DC servomotor. The impulse response is investigated through numerical simulation and experiment, and the effectiveness of the control method is demonstrated.

抄録全体を表示

In this study, we apply digital control to a micro-vibration isolation system using both air springs and magnetic bearings. We design two isolation control systems : a disturbance cancellation control system consisting of both the feedback and the feedforward control loops, and an H_∞ control system designed by applying the solution of the H_∞ 4 block problem. We first compare the two isolation systems through by means of simulation. The calculated results show that both systems can achieve excellent isolation performance. Also it is known that the disturbance cancellation effect can be accomplished by H_∞ control, and the H_∞ control system is stable against variations of the loading weight on the isolation table. Moreover, experiments using DSP (digital signal processor) are performed on the actual isolation system. The experimental results are almost the same as the simulated ones, and the effectiveness of H_∞ control is confirmed.

抄録全体を表示

This paper proposes a vibration control method for piping systems using a feedback control system constructed with LQ-control and a neural network featuring feedback-error learning. The piping system is normally flexible. Therefore, natural frequencies of the system fluctuate when the density of the content varies. This paper shows that the piping system change the dynamics according to increase or decrease of the mass effects. In order to reduce the first vibration mode of the piping system without spillover instability, the control system is designed using LQ-control with feedback-error learning applied to an adapted nonlinear feedback controller. Effectiveness of this control method is confirmed by the neural network simulation program named NeuroLab and is experimented data measured by the control system constructed with the digital signal processing unit.

抄録全体を表示

This paper proposes an experimental study on the arrangement between setting points of an actuator and a sensor for vibration control of a flexible structure, when a vibration controller is mounted at an arbitrary position on the structure. The important vibration mode of the structure to be controlled is the first one, because it is excited most sensitively by strong wind. Therefore, it is necessary to make the reduced-order model represented by the one-degree-of-freedom system at the arbitrary position, considering prevention of the spillover instability. In this paper, the unobservability is used for producing the reduced-order model, and the LQ control theory is used for controller design. For controlling vibration, reduced-order model is made at the setting point of a hybrid dynamic absorber, and a displacement sensor is set at the vibration node of the second vibration mode. Then, the setting point of the sensor is moved to compare the control effect, using this model. It is demonstrated experimentally that the hybrid dynamic absorber designed by this method is able to control vibration well, without spillover instability. Furthermore, it is considered that the setting point of the sensor influences the robustness of the control system.

抄録全体を表示

It is difficult to satisfy the collocation requirement in the vibration control system of a cantilever beam with a piezoelectric actuator. Since not all modes of the cantilever beam can be stabilized with direct velocity feedback control, state-space methods must be used for the mechanical system. Though state-space methods are very useful, many sensors whose number is at least equal to the number of multiple eigenvalues concerned are needed to fulfill observability conditions. In this paper, we describe a control method with which eigenvalues whose number exceeds that of sensors are controlled using digital filters and inverse systems based on truncated mechanical models. From experimental results, it is found that five natural frequencies of the cantilever beam are almost independently damped with the above control method using one piezoelectric actuator and three strain gages.

抄録全体を表示

This paper describes the suppression of torsional vibrations of a rotor by a hollow cylinder partially filled with liquid. The hollow cylinder is divided into identical sections by radial walls. The numbers of sections of the hollow cylinders are 3, 6, and 8. Tuning conditions for fill ratios are examined. Surface waves are also obeserved for the cylinder with 3 sections. Accordingly, observed surface waves move backward, that is, they travel in the opposite direction of the rotation. Analysis including the Coriolis force is presented. Analytical results show good agreement with the experimental ones.

抄録全体を表示

For a heavy haul railway vehicle negotiating tight curves in a factory yard, it is beneficial to employ forced steering trucks to reduce the lateral force between the wheel and rail. The forced steering truck has a wheelset supporting mechanism composed of hanging links and compressing ones for easy steering. The truck should also have a short wheelbase to meet special requirements in the yard. This paper deals with numerical analysis of dynamic behavior of the forced steering truck, in which running stability and the lateral force are examined, considering the wheelbase, the running speed, the longitudinal stiffness caused by the link mechanism, steering link stiffness and the leverage ratio. Results show that it is disadvantageous for both the running stability and the curving behavior to use a short wheelbase, and that the decrease of the steering link stiffness and the increase of the running speed also have disadvantageous effects on the curving performance. It is important for trade-off between the running stability and the curving behavior to select proper values for the longitudinal stiffness of the link mechanism and the leverage ratio.

抄録全体を表示

In previous papers the authors proposed an experimental identification technique for nonlinear beams and two-dimensional structures. The proposed technique is based on the principle of harmonic balance, and can be classified as a frequency-domain technique. In this paper a new identification technique which can be classified as a time-domain technique is proposed for a nonlinear beam. This technique has less restriction on the type of excitation than the former technique. First, a case in which the data of the excitation contain noise is considered. The technique in this case requires a relatively simple algorithm. Then a case in which the data of both the excitation and the response contain noise is considered. The technique in this case has a wider range of application, although it requires a complicated algorithm. Numerical simulation is conducted to confirm the validity of the proposed technique in both cases.

抄録全体を表示

The purpose of this study is to propose a viscoelastic model of landing shock load during running at a steady pace. The vertical components of landing shock loads are measured by force platform under various running speeds in the range of 2.54 m/s to 5.68 m/s. A unique set of parameters of the model for each trial is computed by nonlinear programming, minimizing the sum of the relative standard error of landing shock load and the relative error of rebound velocity. The results are summarized as follows : (1) The viscoelastic model with variable mass proposed in this study can estimate accurately not only the landing shock load but also the rebound velocity. (2) Nonlinear unsymmetrical viscosity behaves as an active element because of its minus viscosity in the restore phase.

抄録全体を表示

In this paper, an elastoplastic-friction damper is proposed as a hysteresis damper which provides damping even in the case of small displacement. The damper is composed of an elastoplastic damper using V-shaped plates and a friction damper consisting of V-shaped springs, brake shoes and a brake plate. The relationship between the force and displacement of the damper was analyzed, and the resisting force characteristics and fatigue strength of the damper were measured. An experimental model of a machine-isolation device system using the damper was also made, and the seismic responses of the system were measured using an electrohydraulic-type shaking table. The experimental results were compared with the calculated results.

抄録全体を表示

This paper deals with a method to estimate unmeasured rotational degrees of freedom mode shapes of a test structure as a linear combination of those of the corresponding Finite Element (FE) model. In this method, weighting coefficients for each mode shape combined are determined by comparing experimentally measured translational degrees of freedom mode shapes with analytical mode shapes of the FE model. The accuracy of estimates is strongly dependent on both the number of mode shapes combined and measuring points for translational degrees of freedom. First, the number of mode shapes combined is discussed. It is found that the number is different for each mode shape to be estimated and is predicted by the Modal Assurance Criterion (MAC) value between measured and analytical translational degrees of freedom mode shapes. Second, a technique is proposed for overcoming a problem with regard to measuring points, which often causes difficulty in estimation because of the ill condition of the modal matrix of the FE model in an inverse problem. This technique incorporates the above method with the estimation method using the Modal Scale Factor (MSF) to compensate for the ill condition. A numerical example using a cantilever beam model is presented and discussed.

抄録全体を表示

The noise of a machine under operating conditions can be categorized as structure-borne, i.e., caused by vibration due to vibrational sources, or airborne, i.e., caused directly by sound sources. By applying the principal component analysis with acceleration pickups and microphones, the number of independent vibrational sources and sound sources can be identified. Based on the frequency response functions between the vibrational sources and the accelerations against the impact or exciter excitation measured in advance, and those between the vibrational sources and the sound pressure at the observation point, first the input forces can be estimated using accelerations under operating conditions and then the ratio of structure-borne noise can be evaluated ; furthermore the airborne noise can be obtained. Each contribution to the total noise can be identified every 1/3 octave bands. These results are confirmed by the boundary element method based on vibrational mode shapes and sound source.

抄録全体を表示

This paper focuses on the problem of identifying all individual rigid body modes and their modal masses for flexible structures under the free-free boundary condition with fewer multi-location excitations than the numbers of modes. The rigid body mass matrix of the structure can be identified with both the parameters of the intertia term determined previously by a modal parameter estimation method and the information of the coordinates of measurement points. As all rigid body properties can be obtained from the mass matrix, it becomes possible to simulate the FRFs between any two measurement points including the contribution of rigid body vibration with those identified parameters. First, the theory is explained. Then, two numerical simulations and two actural identifications for a plate structure and an automotive body component are carried out in order to verify the validity of the theory.

抄録全体を表示

It was concluded in the previous report that the body divergence phenomenon experienced during the flight test of a small test rocket was induced by the degradation of bending stiffness at the joint of a solid rocket motor chamber made from FRP and aluminum alloy vehicle structure. The cause for degradation was examined by testings and the results of the testings were analysed by introducing the effectivity factor of bending stiffness. It was found that the bearing type joint did not work as expected due to yielding of the FRP moter case. A friction type joint was introduced by providing slits at the joint and employing bolts with HI-TORQUE recess. Flight test results with the revised design have been successful.

抄録全体を表示

Recently, there are many studies on shape optimization and some of them have provided practical tools. Unexpectedly, there are not very many studies on determining the optimum topology, in comparison with those on shape optimization techniques. Therefore, in this study, we propose a method of finding optimum topologies of 2-dimensional structures using the genetic algorithm (GA). In this report, we propose a basic method, a method which comprises multistage finite-element partitioning and adopts the allocation of genes. We apply the proposed algorithm to some examples and examine their effectiveness.

抄録全体を表示

A finite-element-method-based formulation for flexible multibody dynamics, particularly targeting flexible link mechanisms, is proposed. The formulation is composed of two techniques : kinematics of largely displaced Euler beam element and direct coordinate partitioning. The kinematics is basically a corotational method. Direct coordinate partitioning, which is effective only in finite-element-based multibody dynamics, partitions the independent and dependent degrees of freedom hinge by hinge. A computational example on a slider crank with an assembly error shows the validity of the formulation and the developed program, SPADE.

抄録全体を表示

CDQ (coke dry quenching system) is a waste heat recovery system that recovers waste heat from hot coke with circulating gas and that generates high-pressure steam. In heat recovery systems, it is very important to obtain maximum efficiency. However, the waste heat quantity fluctuates because of operational changes in the waste heat source process. Therefore, it is necessary to establish the operating method to obtain maximum efficiency under these fluctuating conditions. For this purpose, the dynamic model of CDQ is proposed, and the validity of this model is confirmed by comparing the calculation result with the operation data. Next, it is shown that there exists an optimum circulating gas flow rate in relation to the coke flow rate using this model. By adopting this optimum circulating gas flow rate, 1 [t/h] more high-pressure steam can be generated, compared with the former operation.

抄録全体を表示

Recently, the absorption chiller/heater is attracting interest because it delivers airconditioning without using fluorocarbons. Practically, it requires periodical or non-periodical maintenance. It is difficult to detect failures except in summertime with large refrigerant load and, moreover, human expertise is needed for discrimination of their origin. This paper reports on a system that automatically predicts the possibility of failures and discriminates their origins, regardless of the driving situation.

抄録全体を表示

We propose a 3D object pose estimation method based on pattern matching of the potential fields generated from objects. In our recognition method, the potential fields of both an input object and a model object are generated at first. Secondly, the potential field of the model object is translated and rotated to match with that of the input object. The transformation parameters, i. e., the unknown parameters, are estimated by a nonlinear least squares method, and the position and orientation of the input object are estimated. Furthermore, we clarify some convergence characteristics quantitatively. We also show some ranges, where the solution converges to an optimum value, for two kinds of methods in order to generate the potential field.

抄録全体を表示

This paper shows an application of the already proposed final-state control method to a cart and pendulum system and demonstrates that the method is useful for positioning control of mechanical systems with a flexible mode. In numerical calculations and experiments, we assume that the effective length of the pendulum changes, and the settling characteristics and the robustness of the final-state control performance under this parameter variation are investigated. It is shown that the final-state control performance with the feedforward input is better than that without it within a certain range of parameter variation, and that the use of the feedforward input designed by considering the parameter variation is effective in improving the robustness.

抄録全体を表示

This paper concerns an in-process torque control system in drilling processes of machine tools. In drilling processes, it is necessary to increase the drilling speed without drill breakage. However, increase of the drilling speed often leads to breakage of the drill by the excessive torque arising from the undesiable conditions such as wear of the drill and the misadjustment of the feed rate of the drill. In order to overcome this problem, it is necessary to measure and control the cutting torque in real time. Our objective is to construct a system which can control the cutting torque irrespective of drill wear and the material of target workpieces. Paying attention to the characteristic that the cutting torque is almost proportional to the feed of the drill, a PID control system which varies the feed rate is constructed. From the experimental results, the proposed system can control the cutting torque and suppress the vibration of the cutting torque without breakage of the drill. Furthermore, some advantages of the torque control system are discussed.

抄録全体を表示

We deal with the problem of vibration control of a silicon wafer slicing blade cutting a crystal ingot. Transverse vibration is detrimental to the cutting performance of the slicing ID sawblade. A feedback control scheme is introduced to suppress the vibration of the blade under lateral dynamic disturbance. The rotating sawblade is clamped at the outer boundary and stressed initially in the radial direction, while the inner periphery is subjected to both stationary in-plane and dynamic lateral slicing loads from the workpiece. Furthermore, the normal pneumatic forces controlled by the PD algorithm are applied to the blade to control the vibration. Numerical results are presented for an actual SUS301 blade cutting a 6"-diameter silicon ingot at 1550 rpm. The optimal positions of the sensors and the applied control forces are studied in detail. Results obtained show that the PD controller effectively suppresses the vibration and improves the flatness of the rotating slicing blade.

抄録全体を表示

Lately, a great deal of attention has been paid to model predictive control in the field of process control. However, stability of a closed-loop system has not been guaranteed since the stability depends on the characteristics of a reference trajectory as well as the dynamics of a plant. Recovery methods for the stability have been proposed, but it seems that these methods cannot guarantee stability except for some special cases. In this paper, we propose a new design method for the model predictive control using a class of all stabilizing controllers, which is called Youla parametrization. In our method, if a free parameter is stable, the stability, which is not guaranteed using a conventional method, is guaranteed irrespective of reference trajectory. Control design is carried out while identifying plant uncertainty, which is unmodeled dynamics of an actual plant. A numerical example is given to illustrate the effectiveness of our proposed method.

抄録全体を表示

The aim of this paper is to propose a new method for the optimal "H_∞ norm" computation of time-varying systems. When we define the L₂-induced norm as "H- norm", the optimal "H_∞ norm" computation problem can be formulated as a minimax problem involving control and exogenous inputs. Our approach is as follows. First, we derive analytical solutions for the minimax problems with the exogenous inputs being frozen. Secondly, we present an algorithm for numerical solutions of the minimax problem with the control being frozen : Here, we prove that accumulation points generated by the algorithm satisfy the necessary conditions for optimality. Thirdly, combining the two above-mentioned techniques, we propose a relatively simple method for minimax solutions under assumptions that the saddle points exist. Lastly, we demonstrate the effectiveness of our approach by some simulations.

抄録全体を表示

We have developed a system for automatically sketching patterns on the surface of porcelain using a Slit-light method which is one type of three-dimensional image measurement. This system has two parts. One is a measurement system of the three-dimensional shape of porcelain piece, and the other is an automatical sketching system of patterns by using the result of three-dimensional measurement. Since we classified porcelain pieces into two types according to their shape, two three-dimensional measurement systems were developed. In these measurement systems, we proposed a calibration method using bilinear interpolation in order to calculate values of three-dimensional coordinates simply and speedily. This calibration could absorb errors such as lens distortion and mechanical errors. The automatical sketching system was realized by linking three-dimensional measurement systems and a three-dimensional CAD system.

抄録全体を表示

The repetition of the same key is important when playing the piano, especially since the fast repetition of the same key is difficult even for pianists. In this work, to determine the optimal input waveform for the repetition of the same key of a piano player, the authors first has studied the behavior of the piano-action before and after the string-striking of the hammer. The movement of the piano-action when the swing stroke is performed from the position that leaves the key down than its initial position has been measured, and analyzed using the dynamic model of the piano-action which the authors propose. Further, based on the results, the movements of the piano-action have been studied through experiments of repetition and fast repetition of the same key. As a result, the optimal input waveform for the repetition of the same key has been determined, and music which requires fast repetition can be played on the piano player.

抄録全体を表示

We have developed a productive-type NC gear grinding machine for automobile plants. A wheel spindle and a workspindle of sufficient rotational speed and torque were newly designed for this machine. An inertial damper provides damping torque to the servomechanism of the workspindle. Since silicon oil as viscous fluid is filled in the inertial damper, the increase of oil temperature decreases the viscosity of viscous fluid and reduces the damping coefficient of the inertial damper. We developed a new type of inertial damper which can withstand the effects of temperature rise. Brushless-servomotors which are built in the wheel spindle and the workspindle are driven with servoamplifiers. Since the servo-amplifier has a very high power, ripple current in PWM frequency is very large. This ripple current leads to heat generation and induces a strong noise. Therefore, we attempted to decrease the ripple current using a filter. In the experiments, we confirmed that the new type of inertial damper is useful to withstand the effects of temperature rise and the filter is very effective to suppress the generation of heat and noise.

抄録全体を表示