A Very Large Floating Structure (Mega-Float) has been developed as a new method for creating huge artificial floating land on the sea. Technological Research Association of Mega-Float was established in 1995 to perform the demonstrative and confirmation study by installing 300 m long experimental pontoon model (Phase-1) and 1, 000 m long floating airport model (Phase-2) in Yokosuka bay. Using the airport model, flight checks of landing aids and take-off/landing tests are being conducted to prove that airplanes can operate on Mega-float as same as on the land. By completing the corroborative research at the end of March 2001, Mega-Float would be recognized as the best solution to use ocean space.

View full abstract

A hydrodynamical mechanism occuring in near wake behind bluff bodies is considered together with historical reveiw of the drag theory of bluff bodies. Further, pressure drag on the body and its reduction is investigated on the basis of computer simulation. Numerical simulations of uniform flow of a viscous fluid around a bluff body like a circular cylinder or a D-shaped cylinder have been carried out at the Reynolds number 10⁴ under two-dimensional restiriction of the flow field. It is found that the drag on some cylindrical bodies such as the D-shape is lower than the drag on corresponding circular cylinder. Formation of vortex structure and subsequent vorticity dynamics in the wake have significant effect on the pressure drag. Experimental observations support this result.

View full abstract

In order to design next-generation reusable space transportation systems, their aerodynamic characteristics influenced severely by so-called real gas effects in their hypersonic regimes must be clarified. The effects are composed of disequilibria in and between intramolecular rate processes and chemical reactions. This paper features methodology for modeling disequilibrium in molecular vibration, i.e. the most dominant one in the rate processes. Methods for describing vibrational transition rates and relaxations, as well as those for coupling between vibration and dissociationrecombination, are outlined. Necessity of microscopic non-empirical modeling and its validation using prospective microscopic experiments is emphasized.

View full abstract

In this paper, it is described about a microstructural design method to create a new composite material. In this study, a good mechanical property of a composite material will be realized by a structural design of additional elements. It is assumed that a microscopic structure of reinforcement in a composite material can be modeled to beam structure. By the design system, a location of additional element, which is inserted into a base material of composite material, will be optimized. In order to decide a location of additional element, Convex Subspace Single Linkage Method (CSSL Method) is used. CSSL method is one of the global optimization algorithms. The design system is constructed by using FEM and CSSL method. As one example, the developed design method is applied to a design problem considering microscopic structural design, and a validity of the design system is investigated by a numerical example.

View full abstract

The mechanical behavior of a functionally graded plate with temperature dependent properties is studied in this paper. The development of a micromechanical model for the functionally graded material is presented, and its application to thermo-elasto-plastic analysis is discussed for the case of the W-Cu functionally graded divertor plate of the International Thermonuclear Experimental Reactor. The divertor plate is made of a graded layer bonded between a homogeneous substrate and a homogeneous coating, and it is subjected to a transient temperature change on the coating surface or to a uniform heat flow.

View full abstract

In this paper, the equations governing plane-stress behavior of a piezoelectric circular plate of crystal class 6 mm are presented. A general solution technique based upon potential functions is proposed. The uncoupled governing equations for those potential functions are derived from the equations of equilibrium for the plane-stress field as well as from the equation of electrostatics for the three-dimensional electric filed. The temperature field is considered to be governed by the three-dimensional heat conduction equation. An application is made to a plane-stress problem of a piezoceramic circular plate of crystal class 6 mm in the case where axisymmetric heating temperature acts on both flat surfaces. Numerical results for the radial elastic displacement, radial and hoop stresses, electric potential and radial electric displacement are compared with those obtained from a three-dimensional solution.

View full abstract

Characteristics of a singular stress field around an interface edge are studied for jointed dissimilar plates under uniform tension. Effects of yielding and strain-hardening on the stress field in bonded elastoplastic plates are considered. The order and intensity of singularity are shown to depend on the yield strength in the elastoplastic material. This is studied in detail for linear strain-hardening materials, and it is shown that the intensity decreases with increasing yield strength, while the order of singularity remains constant. This evolution of the singularity can, for a certain combination of material properties, lead to increase in the stress levels near the edge of the interface after yielding occurs.

View full abstract

In the present research, by using the extended modified equivalent inclusion method, micromechanical analysis is performed on a discontinuous fiber-reinforced ceramic-based composites showing interfacial sliding behavior between a fiber and a matrix, and analytical expressions for energy release rate and the dissipation energy release rate due to sliding are derived. In order to investigate the effect of temperature on the toughness of such a ceramic-based composite, the relationship between temperature and the interfacial friction stress is derived. By using this relationship, two energy release rates, and the surface energy of the matrix, the criteria for the propagation of the matrix crack is derived. In addition, fracture toughness of such a composite is evaluated in terms of the critical size of the crack which propagates across a composite in an unstable manner and the change in toughness with temperature can be obtained successfully. The results obtained are consistent with the Brennan’s experimental results.

View full abstract

In this study, the unfolding manner of corrugated hornbeam leaves was observed from buds to fully opened leaves. Based on the observation, a series of paper models was considered to investigate the effect of the vein angles and the growing triangular laminar elements near the midrib on the unfolding of leaves with corrugation folding pattern. By using vector analysis, a numerical simulation for the unfolding of corrugated leaves was carried out. The growth of the midrib and the growing triangle area during unfolding were paid attention because the gowth requires much energy for plants. The kinematic energy during unfolding of leaves was also examined to understand the effects of vein angles and unfolding rates on the total kinetic energy required for full unfolding of leaves and to estimate the deployability of the corrugation folding patterns.

View full abstract

Multiaxial stress state often occurs in trabecular bone under both physiological and pathological loads. Thus development of multiaxial strength criterion for trabecular bone is important for predicting bone failure in many clinical situations. In the present study, strength behavior of bovine trabecular bone was evaluated under multiaxial stress. Butterfly shaped trabecular bone specimens taken from bovine proximal tibias were glued to a specially designed test device to achieve uniform and plane stress state. Then they were loaded in on- or off-axis direction to produce pure shear or tensile-shear stress state. Experimental results were compared with the orthotropic Tsai-Wu and the isotropic Mises criteria to discuss their applicability for trabecular bone.

View full abstract

The relationship between elastic modulus and microhardness for ninety cortical bone specimens from nine bovine femora was determined from a Vickers microhardness testing and from measurements of local elastic modulus using a scanning acoustic microscope. The anterior and posterior specimens of bovine femoral cortical bone used here were plexiform and Haversian bone, respectively. These specimens were scanned with a peripheral quantitative computed tomography to measure the bone mineral density (BMD). The relationship between Vickers microhardness number H_v and elastic modulus E was found to be approximately linear. The Vickers indentation loads of 200, 500 and 1000 gf did not reveal statistically significant differences in H_v. Likewise, a linear correlation also existed between H_v and BMD.

View full abstract

This paper deals with the numerical treatment of thermal and mechanical waves in two-spatial dimensional thermoelastic solids. The development of the formulation is based on the generalized dynamic theory of thermoelasticity which predicts finite propagation velocities for thermal and mechanical disturbances. The numerical procedure is developed by using the method of characteristics, yielding the characteristics surfaces and the characteristic equations. The system of difference equations with second order accuracy for computing the solutions of thermal and mechanical disturbances are then derived. The method is applied to the generalized thermoelastic wave problems in finite plates subjected to the impulsive heating. The numerical results are shown in the graphical form.

View full abstract

The fundamental equations estimating the compressive load-carrying capacity of RC columns with tie and/or spiral reinforcements are used all over the world, based upon the ultimate limit state design, but the common equations include both the elastic term and the plastic one; so, there is no unification concept of the ultimate limit state. In recent years, the high-strength type reinforcement (SBPD type) has been used frequently in the RC column and beam in Japan. Now, the common equations can not apply to the case of the high-strength primary reinforcement of the RC column. The previous reports have already dealt with the concrete’s sharing capacity, the applicable range of the common equations and the generalized practical equation for the upper-bound load-carrying capacity considering the buckling effect of the primary rebars. Especially, this paper describes that the improvement of load-carrying capacity by virtue of the lateral confinement of tie bars depend on the buckling strength of primary rebars in assuming that those elemental buckling length equals the double pitch spacing.

View full abstract

Two main properties of the solid materials are strain rate dependence of stress and strain dependence of wave propagation speed. These phenomena are expressed by using a simple tensile elastic-plastic-viscoplastic constitutive equation in which the under-stress and over-stress are introduced. In addition, strain rate dependence of wave propagation speed is predicted by an equation derived from the above-mentioned one, describing speed of elastic-plastic-viscoplastic stress wave. Based on uni-axial elastic-plastic-viscoplastic consitutive equation, a generalized constitutive equation is proposed for the non-prestrained and non-prestressed solid materials. For the purpose of generalizing, not only under-stress and over-stress but also the third invariant of stress deviator is taken into account. Moreover, It is shown that the proposed equation includes a generalized elastic-plastic, elastic-viscoplastic and elastic constitutive equation.

View full abstract

This paper considers the axisymmetric problem for an infinite cylinder with multiple parallel circular cracks. Expressing the stress component along the crack plane as appropriate series, we reduce the problem to the solution of an infinite system of simultaneous equations. The stress intensity factors are shown graphically for various magnitudes of the crack diameter.

View full abstract

In particle reinforced composite materials, inclusions near at free surface cause locally high stress concentrations. To make clear the local concentrations is very important to estimate the strength of the whole material. In the present work, we present an analytic solution for a semi-infinite elastic plate containing an elliptic inclusion under uniaxial tension at infinity. Papcovich-Neuber displacement potentital functions are used for the analysis.

View full abstract

A formulation to evaluate the natural frequencies of a long, elastic cable that hangs down from the top suspension and swings due to its own-weight in a vertical plane is presented. A model of the serial assembly of rigid links interconnected by virtual revolute springs at their joints is employed. The Lagrange’s equations of motion are derived based on the generalized coordinates taken as the link angles. The eigenvalue problem is constituted by means of linearizing the equations. The validity of the present formulation is demonstrated by the numerical examples of a steel cable with flexural rigidity.

View full abstract

This paper presents the study on the free flexural vibration of a simply supported elliptical plate subjected to a uniform inplane force according to the ordinary thin plate theory. The analysis is rigorously made by the use of the Mathieu function and the modified Mathieu function, which are the solutions of the equation of motion expressed in terms of the elliptical coordinates. Applying the orthogonality of the Mathieu function leads to the frequency equations. The frequency parameters calculated numerically are given in tables and figures for various aspect ratios and inplane force parameters. The buckling of the system will be also discussed in detail.

View full abstract

Following the dynamic theory of linear piezoelectricity, we consider the scattering of horizontally polarized shear waves from a single piezoelectric fiber partially bonded to an elastic matrix. The debonding is assumed to be a curved interface crack with non-contacting faces. The crack opening displacement is represented by Chebyshev polynomials, and a system of equations is derived for the unknown coefficients. Numerical calculations for PZT5-epoxy and BaTiO₃-epoxy composites are carried out, and the effects of frequency, crack angle and piezoelectric material constants on the scattering cross section, dynamic stress intensity factor and dynamic energy release rate are shown graphically.

View full abstract

In the present paper, we examine the characterization for the class of all causal internally stabilizing repetitive controllers for linear minimum phase systems. Yamada and Okuyama gave the characterization of all causal stabilizing controllers for linear minimum phase systems. However, their repetitive compensator has limitations, such that the numerator of the repetitive compensator is a biproper rational function. In addition, their characterization includes controllers such that the control system is not internally stable; that is, the transfer function from any exogenous signals to any signal in the control system is not always causal. The purpose of the present paper is to extend the result of Yamada and Okuyama and give the characterization of all causal internally stabilizing controllers, for minimum phase systems, that have more general repetitive compensators.

View full abstract

In this paper, we examine a design method for low-sensitivity control with robust stability for minimum phase single-input/single-output systems. Yamada clarified that low-sensitivity control systems with robust stability can be designed under the assumption that the relative degree of the plant is equal to that of the nominal plant. In some cases, it is difficult to obtain an accurate relative degree of the plant. We expand Yamada’s result and propose a design method for low-sensitivity control systems with robust stability for systems with uncertain relative degree. Our method adapts a parallel compensation technique and a design method for a parallel compensator is given. A design procedure for low-sensitivity control systems with robust stability using a parallel compensation technique is presented.

View full abstract

This paper describes the numerical simulation approach for dynamical parameter estimation of deployable flexible structures for space application using the direct deployment analysis and three estimation algorithms. The elastic deformation under two-dimensional deployment due to structural flexibility is formulated by use of the finite element approach. Out of three estimation algorithms, the first is the sensitivity analysis that is based upon the sensitivity between the dynamical parameters and the deployment behavior. The second is the genetic algorithm by which evolution of the creature is imitated. The last is the artificial neural network in which the nonlinear mapping is formed between the dynamical parameters and the deployment behavior by learning. As a result of numerical simulations based on the proposed estimation algorithms, it was recognized that estimation of dynamical parameters is basically available and effective for the deployable structure which consists of a spacecraft main body and two deployable flexible links stowed on the side wall of the spacecraft. It is considered that, out of three estimation algorithms, the genetic algorithm is most effective in the category of this kind of estimation problem.

View full abstract

A statistical damage detection and assessment technique for an existing structure is proposed in this paper. Damage is characterized by a reduction in elastic modulus of parameterized finite-element model that is formulated by decomposing the stiffness matrix into constitutive parameters and kernel matrices for each element. Based on two kinds of error definitions between the real structure and analytical model we have derived two system identification algorithms, in which the constitutive parameters are estimated at the element level. Monte Carlo method is used to simulate a set of noise-polluted measured data in order to compare the behaviors of two methods in the face of noisy data and then to obtain the statistics distributions of the estimated parameters. As a statistical approach we introduce Hypothesis test to analyse the status of an existing structure by locating and assessing damage of elements. A numerical example is presented to demonstrate the validity of the proposed method.

View full abstract

We analyse a simple two-dimensional flapping model, two swinging rigid plates joined together on a hinge, to understand dynamical interaction between separation vortices and flapping wings. To make the problem clear, we assume a symmetric situation in which the flapping motion is completely symmetric with respect to up and down directions. By simulating this model using a discrete vortex method, we found a new type of symmetry-breaking mechanism that allows to achieve a mean speed in one direction. The most important factor in determining the behavior of the model is the nature of the flow following the second downstroke, in which the wing produces significant lift through its interaction with the separation vortices. The condition for the symmetry-breaking is also examined. At least two coherent vortices must coexist in fluid, and the intertial mass in a nondimensional form must be larger than a critical value. Two nondimensional parameters are introduced to describe the detail.

View full abstract

The application of the dynamics in organisms to the field of engineering is very instructive. We noted the utility of eukaryotic flagellar motion for the propulsion of micromachines in fluid, and proposed a micropropulsion mechanism modeled on the active sliding of microtubules in eukaryotic flagella. The bending movement and the propelling speed of the mechanism were simulated. For the modeling in the simulation, we took account of the elasticity of the micropropulsion mechanism. The influences of maximum sliding length, elasticity of the micropropulsion mechanism and viscosity of fluid on the bending movement and the propelling speed were discussed.

View full abstract

Micro-fluid dynamics of undulatory locomotion was studied numerically. We developed a time-dependent solution to the full Navier-Stokes equations at low Reynolds numbers (Res) ranging from 10^-5 to 10⁰. The method was validated through a study of flow around a circular cylinder and a conceptual model of an undulating finite sheet. Results are presented for a two-dimensional model of bacterial flagella undulatory locomotion. The geometry and kinematics of our model were based on the data of a bacterium. Our results show that neglecting the inertial effects due to rapid lateral undulation, as observed in bacterial flagellar movements, is inadequate when one tries to predict the thrust-generation, which can be achieved by utilization of both pressure and shear-stress components.

View full abstract

Optimal ways of locomotion in paddling are analyzed theoretically and verified by observing locomotion of webbed or broad feet of turtles swimming in a water channel. In the optimal locomotion, a paddle is placed perpendicular to a moving direction of turtles. And a direction of paddling motion may vary according to an angle of attack of the paddle whereas lift-to-drag ratio is the highest at maximum efficiency and fluid dynamic coefficient C_R(=√C_L²+C_D²) is maximum at maximum thrust.

View full abstract

An Experimental attempt to reduce the aerodynamic noise generated at the cavity by controlling the wavelengths of velocity fluctuation patterns along the spanwise direction is carried out. As a result of the flow control using piezo-ceramic actuator pieces attached at the upstream-edge of the cavity, the system is able to generate wavy structures which have finite spanwise wavelengths and the dominant peak of the cavity noise can be successfully reduced, for the case when the upstream boundary layer is laminar. Controlling the flow to have longer spanwise wavelength is found to be more effective for noise reduction. However, when the upstream boundary layer is turbulent, although noise reduction can be achieved, an attempt to form the wavy pattern is unsuccessful.

View full abstract

Magnets attract the air which is paramagnetic and give influences to air flows. In this study, air flows at an intake with magnetic fields are examined by means of three-dimensional numerical simulations. The flow is assumed to be compressible and inviscid. Under the magnetic fields, vortices are generated near the magnet by magnetic forces. Vortices disturb the air flow and relieve the flow congestion. The flow rate at the intake with magnets increases rapidly and it becomes almost three times as large as that without the magnetic fields.

View full abstract

Point sink flow of rotating stratified fluid of finite depth is studied as an initial-value problem. We first consider a small amplitude motion associated with the sink flow of an inviscid fluid. We then obtain a linear asymptotic solution for large t and r, where t is time after starting the discharge and r is a horizontal distance from the sink, of cylindrically propagating inertial gravity waves. The obtained solution represents the wave nature so convincingly and it is newly found that strength of wave front grows like t^1/6. Numerical calculation is also performed to investigate a thickening of the withdrawal layer near the sink caused by a strong swirling flow where the Coriolis and centrifugal forces are dominant. The numerical results show that, when viscous effects are negligible, the withdrawal layer grows more rapidly than was estimated from a prior work.

View full abstract

Recently, we proposed an integrated method for the problem of numerical conformal mappings of the unbounded multiply-connected domains onto the canonical slit domains of Nehari (1952). The problem is important in potential flow analysis. Our method which uses the charge simulation method offers not only highly accurate mapping functions, but also their analytical derivatives. This enables us to use various formulas for applications, which would be useful but lack easy numerical implementation. In this paper, as a simple example, we offer a Newton’s iteration method for computing the stagnation points around obstacles placed in the uniform potential flow. The algorithm is simple, and as shown, the numerical experiments give high accuracy.

View full abstract

In this study, the analysis of the turbulent flow separation on the wall of the main conduit in a pipe fitting with branching angleθ_p of 135° is presented. The flow in the pipe fitting is analyzed by the free-streamline theory. Then the separation point of the turbulent flow is analyzed using the boundary layer theory. The calculated results indicate that the separation point χ_s/h₁ (h₁ denotes the upstream section of the main conduit.) depends on the area ratio m, the length of the upstream section of the main conduit l_1d/h₁, the Reynolds number Re_1r and the ratio of lateral discharge to total discharge Q₂/Q₁. Also, the experiments were made on the same conditions for the analysis. The experimental reseults show that χ_s/h₁ is dependent on m and Q₂/Q₁. When l_1d/h₁ and Re_1r are large enough, the values of χ_s/h₁ obtained by the experiment indicate that the same tendency as the calculated ones. But the values of χ_s/h₁ obtained by experiment are usually smaller than those obtained by the calculation. Therefore, the separation point of the real flow exists farther upstream than that of the calculation.

View full abstract

The two-dimensional Karman vortex street behind a circular cylinder is known to become unstable and have three-dimensionality up to Re=200. In this paper, the global numerical stability analysis was performed for the flow and the stability characteristics were investigated in the range from Re=170 to 200. The analysis clearly showed the existence of the transition point. The computed critical Reynolds number was 190 and the critical spanwise wavenumber was 1.61. The stability spectra and the spatial form of the three-dimensional eigenvectors are presented. To study further, the energy budget was computed by making use of the data of eigenvectors. It was found that the increment of the production term is substantial within the investigated range of parameters and is considered the main cause of instability beyond the critical point.

View full abstract

Thermal convections in a square duct that is heated from below are studied. In the presence of a through flow, the roll aligning along the flow sets in because of the primary instability of the conduction state. The secondary instability of the longitudinal roll is analyzed to specify stability boundaries of the roll to three-dimensional disturbances. Three kinds of the instabilities are found. Among them, two instabilities with respect to symmetric disturbances make the longitudinal roll wavy. The other instability with respect to an anti-symmetric disturbance makes the roll thicker and thinner periodically. The breakdown occurring in relatively low Reynolds number is similar to the breakdown in available experimental results.

View full abstract

The flow around a circular cylinder was investigated, behind which another circular cylinder of equal diameter was placed as an interference element. Although this situation was the same as the flow around two circular cylinders in staggered arrangement that had been studied extensively, in the present study attention was focussed upon a critical nature of the base pressure of the upstream fixed cylinder. The downstream cylinder was traversed upstream horizontally; hence its relative position was defined by horizontal distance G (gap) and vertical distance Z (level) between two cylinders. As the downstream cylinder was shifted upstream, the base suction exhibited a critical fall at a certain gap for different levels off the wake centerline. The spatial trace of the critical gaps were compared with those of previous experiments, where splitter plates were used as interference elements.

View full abstract

We present a third-order upwind finite element method for solutions of a three-dimensional flow around two square cylinders in tandem arrangement and show validity of the method. The square cylinders are arranged in various distances between the surfaces. When the flow around two square cylinders is unsteady, aerodynamic forces acting on the leeward square cylinder, which locates in wake of the windward square cylinder, become very large depending on the distance between the surfaces of two square cylinders. The numerical results at Reynolds numbers 1000 are compared with the experimental results in terms of the aerodynamic characteristics of the windward and leeward square cylinders.

View full abstract

Through laboratory experiments and measurements, we confirmed that kaolin-water slurry and paraffin wax are very useful to simulate lava flows. Kaolin slurry is a good simulant for a slow moving isothermal lava flow which has a main rheological control. However, a theoretical analysis and some laboratory runs compared with real lava flows strongly suggest an intrinsic problem on this isothermal approach: the surface crust will be significant for a flow morphology after the transition time. To understand a development of a cooling crust and its complicated effect on a flow morphology, we have attempted laboratory experiments using paraffin wax which is especially useful for its relatively low solidus temperature. Although the movement of a flow with a surface crust is not fully understood, our simple analysis suggests that the maximum thickness attained by the inflation process before a breakout, can roughly be estimated by balancing the overpressure and tensile strength. Our model can also estimate the maximum lateral length between the original and the second flow unit after the breakout. These results will strongly help in understanding the complicated and unmodeled compound flow behaviour, which would greatly improve the accuracy of existing numerical models of lava flows.

View full abstract

When a high-speed train enters a tunnel, a pressure wave is generated in front of the train. This pressure wave propagates to the tunnel exit and spouts as an impulsive wave, which causes a bursting sound. It is important to estimate the waveform of the compression wave at the tunnel exit. In this paper we have experimentally studied the transformation of the propagating compression wave in a high-speed train tunnel simulator. The criterion for whether the wavefront steeps or spreads is discussed using the acoustic Reynolds number and the Burger’s equations.

View full abstract

Numerical and experimental studies of a compression wave generated by the reflection of an expansion wave at the open-end of a tube were carried out in this paper. The development of waveform of the compression wave propagated in a straight tube was numerically investigated. The equations of axisymmetric unsteady and compressible flow are solved by the TVD method. It is obtained that the transition to the shock wave of the compression wave occurs at the distance of 20 times of the wave-length of an incident expansion wave from the open-end of the tube.

View full abstract

A high-speed train entering a tunnel generates compression and expansion waves inside the tunnel, which are propagated through the tunnel. In addition, the entering train generates negative and positive impulsive waves outside the tunnel, which are radiated directly from the entrance portal to the surroundings (tunnel entry wave). Experiments have been conducted on the tunnel entry wave by using a scale model of simple geometry under the conditions that the maximum entry speed is over 400 km/h. The experimental results are compared with the theoretical results obtained by Howe and are found to be in good agreement in regard to the waveform, speed dependency and directivity.

View full abstract

The Finite Cover Method (FCM) is extended to handle non linear contact problems in three dimensional solids. We briefly introduce the FCM in a non linear context, describe the techniques used in the method to handle contact between two bodies, and finally show some numerical results obeained with the method.

View full abstract

We consider a numerical treatment of structural-acoustic coupling problem between a shell and two acoustic regions: one is a bounded inner region and the other is an unbounded outer region. The problem exhibits a formulation of the domain decomposition method with a generalized Lagrangian multiplier. The Lagrangian multiplier is the normal deformation of the shell which is coupled with the tangential deformation. The problem is approximated by using the finite element method. We apply the fictitious domain method with locally fitted mesh to discretize the inner and outer regions and to construct preconditioners for the resulting block matrix equation. The block matrix equation is solved in two algorithms with the Krylov subspace iteration: the Schur complement method and a direct method. The direct method performs very well in terms of iteration count and execution time.

View full abstract

The motion of the free surface of an incompressible dielectric liquid is numerically simulated. The dynamics are an electric field and the surface tension. The liquid is assumed to have no charge and is initially placed between a pair of metal electrodes, which yield an electric potential. The problem is described by the Laplace equation for the electric potential and the Navier-Stokes equations for the velocity and the pressure of the liquid with unknown position of the free surface. In order to solve the problem a Lagrangian finite element method is implemented in combination with a mesh refinement method. The shapes and the rising speeds of the free surfaces are calculated for several combinations of Reynolds and Weber numbers.

View full abstract

Until now, for the finite element analysis of magnetostatic problems, we have been demonstrating the effectiveness of the approach based on the mixed method. In this work, we try to extend the A-method to a nonlinear analysis, incorporating B-H curves. To show the validity of the approach, we take up two examples. They are an axi-symmetric problem and Problem 13 of the TEAM workshop.

View full abstract

We discuss an application of a multiple-precision system to numerical computation for ill-conditioned problems. We introduce the Fast Multiple-precision System (F-system) constructed by the first author and show some numerical results applied to numerical analysis for an integral equation of the 1st kind.

View full abstract

We present a unified treatment of the Laplace equation in two-dimensional domain enclosed by a smooth curve. The Dirichlet data can be prescribed on some part of the boundary, while the Neumann data can be prescribed on some other part of the boundary. This problem is reformulated as a variational problem, and it is recast into primary and adjoint boundary value problems of the Laplace equation. A non-iterative numerical method of solution using the BEM is presented. A numerical example is demonstrated for the Cauchy problem, showing viability of the treatment.

View full abstract

Present paper describes the use of a stochastic search procedure based on genetic algorithms (GAs), in developing near-optimal topologies of loadbearing frame structures. Practical structures are designed so as to bear various loads which are caused by earthquake, wind and so on. And a topology which is optimized under a certain loading condition can’t guarantee the effective performance toward the other loading conditions. In this paper, the mixed problem, that is, layout, sizing and material optimization is solved under the various loading conditions using the preferred optimization method. Numerical results are presented showing the efficiency of the proposed method.

View full abstract