The proceedings of the JSME annual meeting
Online ISSN : 2433-1325
2003.1
Displaying 1-50 of 245 articles from this issue
  • Article type: Cover
    Pages Cover1-
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Download PDF (266K)
  • Article type: Appendix
    Pages App1-
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Download PDF (130K)
  • Article type: Index
    Pages Toc1-
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Download PDF (1001K)
  • Feng XIAO
    Article type: Article
    Pages 1-2
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Download PDF (304K)
  • Kei ITO, Takaaki SAKAI, Akira YAMAGUCHI
    Article type: Article
    Pages 3-4
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Free surface vortex is known as the most significant cause of gas entrainment at coolant free surface in sodium-cooled fast breeder reactors (FBR). In this study, numerical simulation for the free surface vortex was carried out for Moriya's free surface vortex experiment. In order to enhance the analysis accuracy, free surface dynamics were calculated by the PLIC type of the VOF method with the CSF model for surface tension. As a result, it is revealed that the simulation accuracy considerably depends on the horizontal cell size in the simulation model. The influence of the surface tension on the simulation accuracy is not significant. The numerical diffusivity has a significant effect on the free surface configuration. Therefore, higher order accuracy scheme is needed for the advection term of the momentum equation. The numerical simulation can predict the experimental data with reasonable accuracy at the low flow rate condition. The simulation accuracy, however, decreases with the increase of the flow rate.
    Download PDF (354K)
  • Changhong HU, Masashi KASHIWAGI
    Article type: Article
    Pages 5-6
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In this paper a CIP based finite difference method, which is under development at RIAM, Kyushu University, is proposed for the extreme wave-body interaction problems. Our research interest is particularly focused on the development of the method that can handle the interaction of largely deformed or broken free surface and a violently moving floating body. The basic idea of this method is summarized at first, and a couple of validation computation results are then presented.
    Download PDF (330K)
  • Tsutomu KITAYAMA, Toshiteru YAMASAKI, Nobuatsu TANAKA
    Article type: Article
    Pages 7-8
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In this study, we present a numerical method simulating multiphase flow having complicated geometries based on the flexible and highly accurate interpolation method of CIVA. discretisation is performed by the finite volume method combined with C-CUP method for universal capability to handling compressible and incompressible fluid flows. In order to check the validity of the method, we solve some benchmark problems of compressible flow, incompressible flow and multiphase flow having complicated geometry.
    Download PDF (339K)
  • Kenji Takizawa, Minoru Chino, Takashi YABE
    Article type: Article
    Pages 9-10
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    The purpose of this paper is to develop a new three-dimensional hydrodynamic simulation algorithm for strongly-coupled fluid-structure interaction and demonstrate its ability to capture the complex surface of solid and liquid. For such an example, we use the skimmer phenomena with rotational motion. This result is compared with the experimental result.
    Download PDF (260K)
  • Masashi SAKATA, Takashi YABE, Ryo NAKAGAWA, Masashi YAMAGUCHI, Choijil ...
    Article type: Article
    Pages 11-12
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In previous papers, we reported the laser propulsion using water for micro airplane. In order to apply this concept in high altitude like Stratosphere, we need to insulate the water layer from low pressure, otherwise water begins to evaporate and finally freezes because of latent heat. For this purpose, we here propose an "air curtain" using a high-speed air flow to insulate. We have done simulation and experiments on this concept. Especially, the result of simulations based on C-CUP Method show that both density and speed of the flow influences the ability to sustain the pressure of the cell. The result shows that Air-Curtain is effective for the laser propulsion in Stratosphere.
    Download PDF (325K)
  • Kenji ONO
    Article type: Article
    Pages 13-14
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Cartesian grid method has higher availability for actual problems of design as First Order Analysis. Especially, simple block approximation is of great stability at the grid generation stage. In this case, arbitrary patterns of wall appear in the computational domain where adequate pressure boundary condition should be employed. In this paper, an implementation of boundary condition that is build into the discretized scheme of pressure Poisson equation is proposed. This method eliminates the difficulties such as the reference points that have multiple value of pressure effectively.
    Download PDF (355K)
  • Takanobu OGAWA
    Article type: Article
    Pages 15-16
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    An mesh generation method for an adaptive Cartesian mesh is developed based on the bottom-up approach in which the smallest meshes near a solid surface are generated first, then the surrounding coarser meshes are generated. The Bresenham's method, one of the rendering techniques in computer graphics, is used to detect intersection between a mesh and a solid surface. With this method, the number of the meshes to be detected can be minimized. The generated meshes are organized with the tree data structure. Utilizing the hierarchical property of the tree data structure, the surrounding meshes can be generated with a simple algorithm. Efficiency of the mesh generation method is presented.
    Download PDF (351K)
  • Shoichi FURUYAMA, Teruo MATSUZAWA
    Article type: Article
    Pages 17-18
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Elastic Domain Decomposition Method (DDM) for a parallel Adaptive Mesh Refinement (AMR) Solver is described. To use this method saving a message passing cost, a good load balancing, and a high parallel performances (73% on 32PEs) were achieved.
    Download PDF (298K)
  • Masako IWATA, Kangbin LEI, Ryutaro HIMENO
    Article type: Article
    Pages 19-20
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We developed quad-tree mesh system to simulate two-dimensional unsteady, viscous and incompressible flows with heat transfer. The mesh is adaptively refined according to object geometries or the gradient of physical value. Furthermore, the mesh is adaptively coarsened where the gradient of physical value is smaller than the threshold value. The fractional volume of fluid approach based on the finite volume method is employed to express object boundaries. Calculation results of coefficient of dragging force on a cylinder is in good accordance with experimental results.
    Download PDF (320K)
  • Kangbin LEI, Masako IWATA, Ryutaro HIMENO
    Article type: Article
    Pages 21-22
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We have proposed a voxel Cartesian grid method combined with cut cells for simulating unsteady, viscous, and incompressible flows with arbitrary boundaries. In this study, the low load and robust fluid analysis software V-FLOW developed using the proposed approach is introduced. By integrating completely the V-CAD data system, which is expected as the next generation CAD, V-FLOW shows the practical usefulness of the fluid analysis technique based on the Cartesian Grid with Cut Cells. Especially, it could mitigate the cost concerning grid generation in practical use.
    Download PDF (562K)
  • Yasuo HATTORI, Shuji ISHIHARA, Nobukazu TANAKA
    Article type: Article
    Pages 23-24
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    The numerical simulation code with a finite difference method for unsteady three-dimensional wind flows was developed in a Cartesian grid system. The immersed boundary method, which enables us to represent the complex geometry in the Cartesian grid system, was used. In addition, the fluid characteristics of complex turbulent flows were predicted with a large eddy simulation, which is based on a mixed time-scale SGS model. The present code was applied to the prediction of flow fields over a steep hill, and it was confirmed that the present results agreed well with those obtained by BFC code, indicating that the present code is suitable to predict the fluid characteristics in the boundary layer over a steep hill.
    Download PDF (470K)
  • Seiichi Ohtaki, Takahiro Yamamoto
    Article type: Article
    Pages 25-26
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    The birefringent coating method is one of the effective experimental method for the analysis two dimensional elastic-plastic problems. In this method, using epoxy polymer which is adhered to the surface of the metal specimen, stress distribution of the metal can be calculated with the isochromatic fringe patterns by the reflection of polalized light. On the contrary, numerical simulation is also important and the finite element method is now familiar to the elastic-plastic problem. In this study, the numerical calculation of the thin plate with a central thin defect subjected to uniaxial tension is conducted using FEM software LUSAS. Quasi-isochromatic fringes obtained by FEM are compared with experimental results.
    Download PDF (329K)
  • Takashi YABE, Kenji Takizawa
    Article type: Article
    Pages 27-28
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We propose a scheme that has the third-order accuracy both in time and space, and can be used in curvilinear system. The method is based on the CIP method and Soroban grid. Although the mesh is used tentatively, it does not need the connectivity of the grid and is a class of mesh-less scheme. Some benchmark test programs are proposed to check the accuracy in curvilinear coordinate.
    Download PDF (251K)
  • Yasunori MOTOYAMA, Nobuatsu TANAKA, Tsutomu KITAYAMA, Toshiteru YAMASA ...
    Article type: Article
    Pages 29-30
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In this study, we present a new numerical method to quantitatively analyze the error of numerical solutions by using the adjoint sensitivity analysis. If a reference case of typical parameter is once calculated with the method, no additional calculation is required to estimate the results of the other numerical parameters such as more detailed solutions. Furthermore, we can estimate the exact solution from the results and can quantitatively evaluate the numerical error. To confirm the effectiveness of the method, we consider here a two-dimensional pure advection problem.
    Download PDF (294K)
  • Nobuatsu TANAKA, Tsutomu KITAYAMA, Toshiteru Yamasaki
    Article type: Article
    Pages 31-32
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We present here a good benchmark problem for three-dimensional passive scalar advection. In the problem, the Eulerian scalar transportation in the Arnold-Belrtami-Childress (ABC) flow field is solved for checking the accuracy, conservation properties, etc of a numerical scheme. Using the benchmark problem, we investigate the effectiveness of newly developed upwind scheme combined with the volume preserving integrator for three-dimensional problems.
    Download PDF (303K)
  • Susumu SHIRAYAMA
    Article type: Article
    Pages 33-34
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Computational methods in CFD have been highly sophisticated, and ways to deal with numerical uncertainties have been proposed. In this case, the benchmark test problems to the equations modeled from the Navier-Stokes equations play important roles in verification. However, many of these are designed for the effect of the advection term, and there is little benchmark problem concerning the equation of continuity. In this paper, we propose several benchmark test problems for the continuity equation, and issues of existing techniques and a future direction in this field are described.
    Download PDF (380K)
  • Takayuki Aoki, Yohsuke IMAI, Koichi SAKURAI, Kaori KATO
    Article type: Article
    Pages 35-36
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A benchmark test for one-dimensional continuum equation is proposed to evaluate numerical schemes to solve hydrodynamic equations. An analytic solution with the profile including discontinuities is derived by means of Lagrangian transformation. The numerical results of various schemes are compared from the view points of mass conservation, sharpness of the discontinuity, and overshoot and undershoot of the profiles after expanding and compressing phases.
    Download PDF (266K)
  • Daigoro ISOBE, Daisaku IMAIZUMI, Youichi CHIKUGO
    Article type: Article
    Pages 37-38
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A new era of computational mechanics, where the Finite Element Method (FEM) is applied to the field of control, has arrived. A parallel solution scheme of inverse dynamics developed by using the FEM, can cope with most of link systems as a unified scheme, regardless of their boundary conditions. The torque curves can be obtained seamlessly without changing the numerical algorithm, even in such cases where the dynamics of the link systems would continuously change. A feed-forward control experiment of a continuously transforming link mechanism is carried out to verify the validity of the scheme, and some numerical results when applied to flexible manipulators are also shown. It is one of the cases where the application of the FEM can be well demonstrated.
    Download PDF (382K)
  • Atsuya OISHI, Takeshi OHYA, Shinobu YOSHIMURA, Genki YAGAWA
    Article type: Article
    Pages 39-40
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    This paper describes an application of a multi-layer neural network to the local contact searching method for higher-order elements. Generally, contact searching process consists of two phases : a global searching phase for finding nearest node-segment pair and a local searching phase for finding exact local coordinates of the contact point in the segment found. In the present method, a neural network is utilized for finding local coordinates of the contact point. In this paper, the basic performance of the neuro-based local contact searching method for contact analyses with 10-node tetrahedral elements is tested through sample analyses.
    Download PDF (310K)
  • Tadashi SHIBUE, Masato YAMAMOTO
    Article type: Article
    Pages 41-42
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A release of connective force or failure of bolt connections under the shock loading sometimes causes severe accidents in the fabricated structures. Behavior of a bolt connection under a shock loading is studied experimentally and numerically in this report. A steel block connected on a table by eight bolts is used for this study. A shock loading is given as an upward impact on the bottom of table by an impact head. Strain time history of connecting bolts is measured by the strain gages set within the bolts. FEM analysis is carried out to reproduce the behavior of a bolt connection. Pre-tension is introduced to show the initial torque given to the bolt. Then, shock loading is introduced by the physical collision between table and impact head. The calculated strain is compared with the measured value to show the validity to current numerical model.
    Download PDF (359K)
  • Kazuhiro ASANO, Kazuaki SASAKI
    Article type: Article
    Pages 43-44
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    This paper suggested one method to analyze a contact problem of two elastic body with sliding. The analysis is carried out by using a geometrical condition defined by an angle change of contact surface, meanwhile, a dynamical condition defined by equilibrium of force and the Coulomb's friction law. From the comparison of present results with ones by Hertz static analyses, the availability of this method is clarified
    Download PDF (260K)
  • Junichi MATSUMOTO, Takeshi SUZUKI, Akira TEZUKA
    Article type: Article
    Pages 45-46
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    For the purpose of the improvement in efficiency and stability and accuracy of the calculation, spatial discretization is applied to the mixed interpolations for the velocity and pressure fields by the stabilized bubble element and linear element, respectively. This paper presents a finite element analysis of a thermal-fluid structure interaction problem, in which the fluid is treaded as incompressible viscous flow with boussinesq approximation, and a structure is idealized by elastic springs.
    Download PDF (325K)
  • Masayuki HARIYA, Ichiro NISHIGAKI, Yoshimitsu HIRO
    Article type: Article
    Pages 47-48
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A technique for automatically generating a hexahedral mesh of a complex solid model was developed. This technique can automate the interactive operations (such as model editing) that are carried out as a pretreatment for hexahedral-mesh generation. Consequently, the time taken by these interactive operations was significantly shortened, and high-speed generation of a hexahedral mesh was thus possible. Furthermore, the developed technique was applied to the generation of a hexahedral mesh of an engine-block model. And it was confirmed that the mesh generation time is shortened to one tenth of that for the conventional method.
    Download PDF (442K)
  • Masakazu INABA, Takashi UCHIKAWA, Toshimitsu FUJISAWA, Genki YAGAWA
    Article type: Article
    Pages 49-50
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A node-based parallel mesh generation technique based on the free mesh method (FMM) is featured by a node-by-node local mesh generation algorithm and it is quite suitable for massively parallel computing. Is this paper, a merit of FMM other than parallelization is present. i.e. the fast re-meshing for large deformation problem. It is shown here that FMM can re-mesh properly local area only that needs re-meshing properly.
    Download PDF (321K)
  • Toshimitsu FUJISAWA, Genki YAGAWA
    Article type: Article
    Pages 51-52
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A parallel node generation technique based on a probabilistic theory is presented. A set of generating points of centroidal Voronoi tessellations can be obtained in parallel using the method. Furthermore, very good-natured finite element mesh can be obtained using these points. By employing the present technique, all the processes from CAD model to finite element mesh can be performed on parallel computers.
    Download PDF (422K)
  • Hajime FUKUDOME, Masakazu INABA, Toshimitsu FUJISAWA, Genki YAGAWA
    Article type: Article
    Pages 53-54
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In this study, we attempted to visualize results of numerical simulations by using commercial CG software, which is originally developed for the purpose of arts. In order to transfer data of numerical simulations to CG software, we developed plug-in programs, because there is an essential gap between output data of numerical simulations and input data required in CG software.
    Download PDF (478K)
  • Ryuzaburo SUGINO, Nobuyoshi TOSAKA
    Article type: Article
    Pages 55-56
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    This paper discusses on applicability of the Gusssian Newton method with damping term for the shape decision problem in Laplacian field. This problem has the inner unknown boundary which is located at center of gradiential field. The obtained results show that the various identification process in two dimensional region. We recognize the effectiveness of damping term of nonlinear least squre optimum method.
    Download PDF (307K)
  • [in Japanese]
    Article type: Article
    Pages 57-58
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A study of the direct method for numerical solution of the inverse boundary value problem using the boundary element method is presented. This method proposes a non-iterative and unified treatment of conventional boundary value problem. In order for a solution to be identified the inverse problem is reformulated in terms of a variational problem, which is then recast into primary and adjoint boundary value problems of the elasto-statics in its conventional form. An inverse boundary value problem for two-dimensional elasto-statics in the domain enclosed by a piecewise smooth curve is considered. While the displacement and the traction data are prescribed on a respective part of the boundary and the second part of the boundary on which no boundary data are given, the unknown values along the whole bou ndary will be find.
    Download PDF (221K)
  • Yoko OHURA, Kazuei ONISHI
    Article type: Article
    Pages 59-60
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    An idea is presented about a possible numerical identification method for the location of an isolated essential singular point of the harmonic function in the plane. The method is based on the boundary element solution to the Cauchy problem of the Laplace equation, in which Dirichlet and Neumann data are measured on arbitrary parts of the boundary to some extent. Discussion is concerned with only one sample problem whose numerical solutions exhibit oscillating odd behavior near the singular point. The solutions are scanned on family of lines in two independent directions.
    Download PDF (293K)
  • Eisuke KITA, Yoichi IKEDA, Norio KAMIYA
    Article type: Article
    Pages 61-62
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A sensitivity analysis scheme for the boundary value problem governed with Poisson equation by using Trefftz method is presented in this paper. An inhomogeneous term of Poisson equation is approximated with the polynomial function to derive the particular solutions. An unknown function is approximated with the superposition of T-complete functions of Laplace equation and the derived particular solutions. So, direct differentiation of the approximate solution leads to the sensitivities. The present scheme is applied to a numerical example.
    Download PDF (256K)
  • Toru TAKAHASHI, Toshikazu EBISUZAKI
    Article type: Article
    Pages 63-64
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    We propose the fast approach to solve boundary integral equation methods with the special-purpose computer MDGRAPE-2 which is originally developed for the acceleration of classical molecular dynamics simulations. The framework of the approach and the numerical results obtained so far are reviewed here.
    Download PDF (331K)
  • Yuki ONISHI, Kenji AMAYA
    Article type: Article
    Pages 65-66
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A new technique of boundary element methods for 2D potential problems to obtain the solution only in a target region quickly was developed. This technique takes advantage of diffusion effects in Laplace fields. In fields governed by diffusion equations, high frequency disturbances of boundary conditions on non-target bounds far from the target region give little infection to the target region. In this technique, boundary conditions on non-target bound are transformed into Fourier series, and their only low frequency terms are utilized by using special weight functions for boundary integral equations. When the solutions only in a target region are needed especially in large size boundary value problems, this technique enable us to obtain them quickly and precisely with solving non-target solutions roughly.
    Download PDF (392K)
  • Masataka URAGO, Takafumi KOYAMA, Kunio TAKAHASHI, Shigeki SAITO, Yoshi ...
    Article type: Article
    Pages 67-68
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Boundary element method is one of the most powerful tools to solve partial differential equations. However, it takes a lot of computational time when applied to a problem with a large number of unknowns. In this paper, we adopted the binary tree structure with tight bounds and the downward pass for that structure for the fast multipole method to reduce the number of M2L translations. The fast multipole boundary element method using the binary tree structure with tight bounds is developed and the analyses of an electrostatic force acting on a matal micro particle above a complex structure to confirm the effectiveness of this method.
    Download PDF (351K)
  • Hirokazu MATSUDA, Seiichi YAMAJI, Tadashi TANIZAWA
    Article type: Article
    Pages 69-70
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Download PDF (330K)
  • Masahiro ARAI, Toru TAKAGI
    Article type: Article
    Pages 71-72
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    In the present study, boundary element method is applied to the evaluation of interlaminar fracture toughness of CFRP composite. The dynamic fracture toughness of CFRP was measured by split Hopkinson bar method, where the dynamic three point bending tests were carried out using short-beam-type specimens with end notch. In order to evaluate the dynamic deformation of CFRP laminates, Laplace-transformed BEM scheme was employed for the numerical computation, and the extrapolation scheme was applied to obtain the stress intensity factor (SIF) of Mode II in the unidirectional CFRP laminates. The basic formulation of BEM analysis for the anisotropic body of CFRP is introduced, and the experimental results of dynamic fracture toughness obtained from SHPB test and BEM analysis are discussed.
    Download PDF (322K)
  • Toshiro MATSUMOTO, Masataka TANAKA, Youhei OGAWA, Hisaya Murakami
    Article type: Article
    Pages 73-74
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    The boundary integral representation for thermoelastic problems involves a domain integral term originated from the thermal strain. In this paper, the dual reciprocity method (DRM) is utilized to convert the domain integral into boundary integral terms. In the DRM formulation, instead of using the particular solutions associated with a suitable radial basis function, particular solutions are first assumed and the corresponding basis functions are then derived. The effectiveness of the proposed method is demonstrated through some numerical examples.
    Download PDF (268K)
  • Masataka TANAKA, Jianming ZHANG, Toshiro MATSUMOTO
    Article type: Article
    Pages 75-76
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    The meshless BEM, called the hybrid boundary method (HBNM), is applied to evaluate the thermal properties of the carbon-nanotube (CNT) based composites. The results obtained are discussed to demonstrate the feasibility and usefulness of the method for further applications.
    Download PDF (245K)
  • Masataka TANAKA, Toshiro MATSUMOTO, Susumu TAKAKUWA
    Article type: Article
    Pages 77-78
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    This paper presents a dual reciprocity boundary element method (DRBEM) applied to the transient heat conduction problem of functionally graded materials. The functionally graded material can be modeled as an inhomogeneous one where the thermal conductivity is a continuous function of coordinates. The integral equation formulation employs the fundamental solution of the Laplace equation for homogeneous materials, and hence from the inhomogeneous part of the governing differential equation a domain integral arises in the boundary integral equation. This domain integral is transformed into boundary integrals by using a new set of radial basis functions. Furthermore, time derivative is approximated by the time-stepping method, and the domain integral also appears from this approximation. The domain integral concerning the "pseudo" initial condition at each time step is also transformed into a boundary integral via the same dual reciprocity method. The details of the proposed DRBEM are presented, and a computer code is developed for three-dimensional problems. Through comparison of the results obtained by the computer code with the result of collocation method, the usefulness of the present DRBEM is demonstrated.
    Download PDF (322K)
  • Masataka TANAKA, Toshiro MATSUMOTO, Takahiro NISHIGUCHI
    Article type: Article
    Pages 79-80
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    The present paper proposes an integral equation formulation for static bending and its implementation of stiffened-anisotropic plates using boundary element method. For an eccentrically stiffened plate, in-plane and out-of-plane deformations are coupled. In this paper, three components of interactive force and two components of interactive moment between the plate and stiffener are treated as line distributed unknown loads. The numerical results obtained by the computer code developed in this study are discussed, whereby the versatility of the proposed method of analysis is demonstrated.
    Download PDF (278K)
  • Yo YAMAGATA, Sigeki SEKIYAMA, Radha Krishnan
    Article type: Article
    Pages 81-82
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    This paper describes the effect to shortening of the design period by morphing, and the example of construction of the shape optimization system. The shape optimization by the CAD-CAE cooperation in the present conceptual design has taken many periods. The morphing technology which makes a direct shape change of the CAE model attracts attention to reduce design period.
    Download PDF (276K)
  • Takashi SASAKI
    Article type: Article
    Pages 83-84
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    CAE begins to be introduced in the section of the design in manufacturing. As a method for the use of CAE in the design section, there is an analysis based on a parametric model. The user simply changes parameters built into a master model and updates it. This paper will demonstrate CAE systems including CAD environment for changing the parameters, presetting geometry modules and mesh generation. A complex operation and special study for CAE can be reduced by these systems.
    Download PDF (356K)
  • Norihiko NONAKA, Takashi AKI, Yoshimitsu HIRO
    Article type: Article
    Pages 85-86
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A system for optimizing the shape of various products by using 3D-CAD has been developed. This shape-optimization system consists of 3D-CAD, a pre- and a post-processor, an analysis solver, an optimization engine, and an automatic geometry-changing tool. The system has three key features : the iterated optimization process is carried out automatically; it can obtain 3D-CAD data as geometry data after optimization; and the parametric function of 3D-CAD is used to change the geometry of a CAD design. As an example application, the system was used to optimize the shape of a pump impeller, and it was found that it saves time and can cut the impeller mass by about 14%.
    Download PDF (342K)
  • Shoji YAGI
    Article type: Article
    Pages 87-88
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Download PDF (364K)
  • Atsuhiko SENBA, Hiroshi FURUYA
    Article type: Article
    Pages 89-90
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    The variable matrices method, that was based on the properties that adaptive structures could identify their own unknown structural characteristics by their variable and controllable structural parameters, was proposed to identify structural matrices. Also, the relation between the number of variable structural parameters and identifiable degrees of freedom of the structural matrix was shown. A simple spring-mass system attached with variable stiffness spring was demonstrated to investigate the performance of the proposed method. Results of the self-identification showed that the exact stiffness matrix was identified by the variable matrices method.
    Download PDF (300K)
  • Masahiro ARAI, Haruki SUZUKI, Mitsuhiro NARITA
    Article type: Article
    Pages 91-92
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    Resonant ultrasound spectroscopy (RUS) is widely used to evaluate the characteristics of small solid specimens of single crystals or anisotropic materials. The elastic moduli of the solid body can be determined by measurement of the natural frequencies exited by a pair of piezo-electronic transducer. In this paper, a numerical technique for the determination of the elastic constants in RUS testing is discussed. An alternative evaluation function for the search algorithm of the elastic moduli is employed, and genetic algorithm is applied to estimate the evaluation function in the search process. Some numerical simulations for anisotropic solid body are conducted to demonstrate the effectiveness of the present method.
    Download PDF (341K)
  • Fumio KOJIMA, Nanaka GOTO
    Article type: Article
    Pages 93-94
    Published: August 05, 2003
    Released on J-STAGE: August 01, 2017
    CONFERENCE PROCEEDINGS RESTRICTED ACCESS
    A computational method is considered for three dimensional shape recovery of a SUS304 sample plate with cracks. A numerical model of the inspection process is given with the hybrid use of finite element and boundary element methods. The impedance change due to a defect is calculated by building the difference of the impedance values with the flaw present and absent. In order to achieve the high reduction of the computational costs, the finite element domain decomposition method is applied to the forward analysis.
    Download PDF (316K)
feedback
Top