vol.16 (2003)

Bulletin of Computational Science Research Center
Vol.16, 2003 ISSN 1347-6726 Published March 20, 2003

Numerical Investigation of Active Heat Exchanger for Fluid Power Systems
Yutaka Tanaka, Toshiyuki Yoshida, Nobuhito Hazama, Ryushi Suzuki, Kazuo Koike
PDF   Abstract
One trend in industrial manufacturing is for fluid power systems, as well as fluid power components, to be designed in a more compact fashion. The benefits of smaller fluid power systems are obvious ― economy of materials, less energy consumption, less square-footage required. But one often overlooked drawback is the increased probability of air entrainment for fluid power systems that incorporate smaller reservoirs. Also a recent trend in fluid power systems, an environmental protection movement has gained momentum under the global issue. Bubbles or foam in working oils greatly influence the performance of fluid power systems and may cause major problems. Recently, one of the authors has developed a new device using swirl flow for bubble elimination capable of eliminating bubbles and of decreasing dissolved gases. We call this eliminating device a bubble eliminator. In our previous study, we have reported that the bubble eliminator is useful for preventing oil temperature rise caused by the bubbles under low, moderate and high system pressure conditions. A new device for mixing gases and liquids is also developed by one of the author’s company. We also call this device a static mixer. In this paper, a new heat-exchanging system for fluid power systems are proposed and developed by using a principle of the static mixer and the bubble eliminator. The heat-exchanging system consists of an injection, mixing and elimination parts. Firstly, surrounding air is directly and occasionally infused in oil line at the injection parts. Secondary, the air is efficiently mixed with the oil by the static mixer at the mixing part. Thermal energy of the oil is transferred to the adjoining air through the surface of the air bubbles in the oil. The oil is cooling down at the mixing part and the temperature rise of the oil is restrained. Finally, the air bubbles with thermal energy are separated from the oil by the bubble eliminator at the elimination part. We call this system an active heat exchanger. Flow pattern in the static mixer and the bubble eliminator has great influence on thermal efficiency of the active heat exchanger. Numerical analysis of the flow in the static mixer and the bubble eliminator is carried out to investigate the influence of mixing and separating the air to the oil.
Numerical Analysis of Flow within Ultra-Highly Loaded Turbine Cascade - Influence of Inlet Flow Angle -
Yoshiki Matsuda, Munetoshi Matsumoto, Shimpei Mizuki, Hoshio Tsujita
PDF   Abstract
Flow in an ultra-highly loaded turbine cascade (UHLTC) for future gas turbines was analyzed by using the CFD(computational Fluid Dynamics) method. The calculations we re performed by changing the inlet flow angle from 75 to 85 degrees in order to reveal the influence of the inlet flow angle on the performance of the UHLTC. The high inlet flow angle not only enlarged the horseshoe and the passage vortices but also induced a separation on the suction surface of the blade. Consequently, the increase of the inlet flow angle enhanced the loss generation.
Computation of Reverse Flow in a Channel with an Obstruction in Front
B.H.L.Gowda, S.Mizuki,Y.Asaga
PDF   Abstract
Reverse flow occurs in a channel when an obstruction is placed at the entry. The phenomenon involves separation, reattachment and shear layer interaction. The various facets of the phenomenon have earlier been investigated experimentally. Presently, an attempt has been made to compute this flow both steady and unsteady equations of motion. Results predict the occurrence of reverse flow, and bring out other features of flow like vortex shedding behind the channel.
Numerical Simulation and Hypervelocity Impact Test of Space Debris Impact on Geostationary Orbit
Yuji Yasuda, Nozomu Masuda, Keita Fukushima, Masahide Katayama, Kazuyoshi Arai, Yutaka Tanaka
PDF   Abstract
In this study, the numerical simulation and the hypervelocity impact test of a space debris collision phenomenon on geostationary orbit were investigated. The numerical simulation results were shown good agreements with the impact test results and the validity of the macroscopic penetration mechanism of collided material was shown. Further, hypervelocity impact tests using three kinds of plastic materials as collision materials were performed, and the effects of properties of plastic materials on the ballistic limit velocity (limit penetration velocity) of collided material were discussed.
Numerical Simulation of Turbulent Diffusion Flame for Methane by Using Model of Overall One Step Reaction Formula
Yasunori Ebisawa, Taku Nishimoto, Takaya Hoshino, Tadashige Kawakami
PDF   Abstract
Numerical Simulations of turbulent diffusion flame for methane in air have been carried out to examine the behavior of temperature, concentration of fuel and distribution of velocity near the flame holder by using model of overall one step reaction formula. Initial conditions for turbulent diffusion flame for methane correspond to 300 K and 0.1 MPa and a mixture of 79% nitrogen and 21% oxygen by volume is used as a substitute for air. The research results show that 1) The highest calculated value of temperature near the flame holder is of the order of 2100 K 2) The influence of combustion of turbulent diffusion flame on distribution of velocity appears negligible.
Vibration and Stress Analysis of Impeller Using the Finite Element Method
Masahiro Ohtake, Takahiro Tsuji, Mitsuo Iwahara, Shimpei Mizuki
PDF   Abstract
It tends to be thought that a finite element method is already established now, and the program will be completely used as a black box. However, the program used now needs to have a question for whether it is actually the best analysis method, and we should examine the program once again. Then, after clarifying the difference between the stress value by Nastran and the theoretical value by manual calculation, we performed analysis of vibration and stress of impeller for micro centrifugal compressors.
Research on the Vibration Characteristic of a Percussion Instrument Cymbal by Modal Analysis - Verification of Experiment Modal Analysis and Theoretical Modal Analysis, Comparison, and Structure Optimization -
Teruyuki Ishiwata ,Mituo Iwahara, Akio Nagamatu
PDF   Abstract
Using nonlinear optimum method, maximum likelihood modal parameters in statistics are identified in the frequency domain with multiple frequency response functions. User interfaces are constructed with MATLAB in personal computer. This system is applied for percussion cymbals and compared with FEM results. Moreover, it developed further and structure change was optimized this time.
Experimental Modal Analysis by Nonlinear Optimum Method
Takashi Ishii, Mitsuo Iwahara, Gaku Minorikawa
PDF   Abstract
Many products are designed by experience and intuition of engineers and are also developed by the repetition of a trial product and an experiment. However, the design conditions which become future stil l severer are fulfilled, and in order to increase efficiency of development. Raising accuracy etc. from the conventional technique is wanted to produce a new technique. In recent years, modal analysis is beginning to be effectively used for anticipation and elucidations of a vibration phenomenon. Experimental modal an alysis with frequency response functions by vibration testing, is a series of methods of extracting the dynamic characteristics of the system in the form of the modal parameters. It is widely used as base technology of vibration analysis. Moreover, use of the simulation analysis of the vibration in a design step is becoming familiar by rapid progress of computer and software. The purpose of this study is identification of modal parameters with the nonlinear optimum method as the experimental modal analysis technique. This technique is correctly extended to a multi-point response in differential iteration method. User interfaces are constructed with MATLAB in personal computer. This system is applied for the cylinder containing the slit and RV rear panel. And the experimental modal parameters are compared with FEM results. Moreover, the impact test equipment is produced for the improvement in accuracy of experimental data, and the validity is verified.
Self Diffusion Coefficient of Hard Sphere System
Yosuke Akino, Yosuke Kataoka
PDF   Abstract
The hard sphere system is idealized model of liquid state of the simple substance. In this study, we carried out molecular dynamics (MD) simulation in the hard sphere potential, and examined thermodynamic properties. And, we obtained self diffusion coefficient both from mean square displacement and velocity auto correlation function, and then, examined those differences and validity.
Monte Carlo Simulation of Linear Polymers for Replica Exchange Method
Yosuke Ueda, Yosuke Kataoka
PDF   Abstract
We simulated fold up linear polymer, which have sixty monomers of thirty positive and thirty negative electron charges in vacuum. We have performed replica exchange Monte Carlo (REMC) simulation and investigated various thermodynamics quantities. Intra-molecular interaction consisted of Coulomb, spring and softcore interaction. A double helical structure is obtained when a straight chain structure set up as an initial conformation.
Electric Circuit Parts Cognition Based on the Equivalent Characteristic Value Method
Yoshiharu Nakajima, Seiji Hayano, Yoshifuru Saito
PDF   Abstract
The principal purpose of our research is to cognize an each of the distinct electric circuit elements, for example resistance, capacitance, inductance, diode, transistor and so on, by measuring an electromagnetic field distribution around the electric/electronic devices based on the equivalent characteristic value method. In the present paper, we have tried to cognize the electrical resistance, capacitance and inductance by measuring a time domain magnetic field measurement as an initial test of our research project. To extract the characteristics of a time domain signal, we employ the equivalent characteristic value approach. After that we employ the 3D Lissajous diagram methodology to visualize the characteristics of time domain signal. Applying least square means to a system of an ill-posed system of equations, which essentially accompany the image cognition methodology, makes it possible to classify each of the basic electric circuit elements.
Current Visualization by Vector Current Viewer
Akihiro Hirota, Seiji Hayano, Yoshifuru Saito
PDF   Abstract
In recent years, PCBs (Printed Circuit Boards) are widely employed to work out the compact as well as lightweight electrical devices. However, inspection of their regularity is not so simple because of their complex electric circuit structures. One of the non-destructive tests to do that is to measure in the vicinity the electromagnetic fields essentially accompanying with their operation. Estimating the field sources from the measured fields becomes useful and important information not only for checking up their regularity but also for the EMC/EMI problems. According to this background, we are motivated to propose the current viewer, which makes it possible to visualize the current distributions in PCBs. Previously, we have proposed the current viewer in order to visualize the current distribution on planar electric circuits. Operation principle of this current viewer is based on those of modified Rogowski coil method. The mechanical structure of this viewer is that the sensor solenoid coils does not cover all the circumferences of the target current carrying conductor, but covers a semicircle. In this paper, we propose the vector current viewer by combining two current viewers in orthogonal way. Our vector current viewer is capable of visualizing the magnitude as well as direction of target current along with the Lissajous diagram methodology.
Magnetization Characteristics Evaluation by Means of Image Cognition
Yusuke Okubo, Hisashi Endo, Seiji Hayano, Yoshifuru Saito
PDF   Abstract
Previously, we have proposed a method of image cognition based on the color component decomposing methodology, which has made it possible to cognize each of the human faces even though a different facial expression. Furthermore, we have succeeded in cognizing the monochrome images utilizing light intensity histogram approach.We are now planning to exploit a magnetization characteristics measurement instrument by applying the image cognition method to the Bitter magnetic domain images. As the first step, in this paper, we apply the monochrome image cognition method to the SEM (Scanning type Electron Microscope) images of silicon steel magnetic domains. As a result, it is revealed that the magnetic domain images take the similar ones at the highly saturated region but take the extremely different ones at the not saturated region.
Transformer Type Pressure Sensor Array and Its Application
Chikara Tabata, Seiji Hayano, Yoshifuru Saito
PDF   Abstract
Principal purpose of this study is to exploit a fully computerized monitoring system while a hospitalized patience is sleeping or waking on a bed without any personal privacy violation by measuring a pressure distribution in each of the patience. In order to develop a system, we have previously proposed a transformer type pressure sensor, which keeps only cushiness for human interface but also toughness to any types of pollution. This transformer type pressure sensor is composed of the spring coils as a primary coil, which keeps cushiness as a bed, and secondary coils to measure a pressure as an induced voltage. In the present paper, we have worked out a pressure sensor array and equipped to a sitting surface of chair. Further, we have measured the static as well as quasi-dynamic pressure distributions when changing the sitting pose. To cognize the situation of sitting pose, we have applied the monochrome image cognition methodology. As a result, it is revealed that major sitting pose can be cognized. Thus, we have succeeded an initial test of our fully computerized patience monitoring system without any violation of human privacy.
Magnetic Nondestructive Inspection by Means of Image Recognition
Takanori Sato, Seiji Hayano, Yoshifuru Saito
PDF   Abstract
A method of image recognitions is proposed and applied to the nondestructive inspection utilizing the database of color-imaged magnetic fields. Our image recognition method has two distinguished features. One is that the image is handled as its eigen pattern. The other is that the recognition is carried out by solving for the ill-posed linear system of equations comprising the eigen patterns as a database. The eigen pattern consists of histograms in terms of intensity, tones and primal color components of the image. This eigen pattern of color components yields a numerical data set which is hard to be changed by the differences in spatial position of target as well as resolution of the image. An application of the proposed image recognition method to the nondestructive inspection based on measuring magnetic fields is carried out to demonstrate the usefulness of our method. Consequently, it is revealed that our methodology is one of the most powerful and effective nondestructive inspections of magnetic devices.
Evaluation of the Frequency Classified Current Distributions on Printed Circuit Board
Kanako Shiraishi, Hisashi Endo, Seiji Hayano, Yoshifuru Saito
PDF   Abstract
For inverse approaches visualizing the current distribution on the printed circuit board operating under different frequencies, we have examined the magnetic flux distribution by finite elements along with the strategic dual image method taking open boundary condition into account. Further, we have carried out the experimental verification to the simulations. The comparisons between simulated and experimented results have been carried out after classifying the estimated current distribution with an inverse calculation in to the distinct frequency components by Fourier transform. As a result, we have concluded that estimation of the current distributions on the multi-layered printed circuit boards is difficult, however, in case of the single layered model could be evaluated with fairly good accuracy.
Magnetic Field Computation with High Accuracy by Wavelets
Hayato Yabunami, Seiji Hayano, Yoshifuru Saito
PDF   Abstract
Previously, in order to carry out the three-dimensional magnetic field analysis in a most efficient manner, we proposed the theory of modern magnetic circuits based on finite elements discretization. This makes it possible to implement the three-dimensional magnetic field computation in a quite efficient manner similar to those of the conventional magnetic circuit theory. In the present paper, we try to apply the wavelet transform to a magnetic flux distribution in order to represent it by small computational effort. As a result, it is shown that the wavelet transform yields a fairly good distribution from small computation.
Visualization of Barukhausen Signal
Masaki Katsumata, Seiji Hayano, Yoshifuru Saito
PDF   Abstract
Ferromagnetic materials, i.e., iron steel and its composites, are widely used as the frame parts of various artificial products and constructions such as automobile, air plane, building, bridge and so on. Because of its mechanical property, iron steel is most popular in use for the frame materials to maintain their mechanical strength. On the other side, nondestructive testing of iron steel is an extremely important technology; and also magnetization process of the ferromagnetic materials always accompanies with the Barkhausen effects. In the present paper, we propose a new nondestructive testing methodology for the ferromagnetic materials fully utilizing the Barkhausen effects to check up their mechanical reliability. Our new methodology is based on the visualization of Barkhausen effects in a two dimensional plane. As an initial test of our methodology, we have tried to visualize the Barkhausen effects in a two dimensional plane and carried out the identification in each of the Barkhausen effects.
A Study of Electrical Impedance Tomography
Shusuke Mabuchi, Seiji Hayano, Yoshifuru Saito
PDF   Abstract
Electrical impedance tomography (EIT) is one of the cheapest ways in order to realize a tomography. Principal purpose of the EIT is to obtain a conductivity distribution of a two-dimensional cross-sectional target area by injecting the current. EIT utilizes a functional relationship between the injection current and surface potential distribution depending on the conductivity difference. This functional relationship of EIT reveals that EIT has versatile possibilities but it is essentially reduced into solving for an ill-posed inverse problem. In order to develop the EIT, we try to apply the generalized vector sampled pattern matching (GVSPM) method to the inverse problem accompanying with EIT. As an initial test experiment of EIT development, we try to evaluate a resistance distributi on in a planer circuit by measuring entire nodal voltages while changing the electrodes for current injection. As a result, it is revealed that a fairly good result could be obtained as an initial test example.
Optimization of Seed Source Position with Genetic Algorithm in Prostate Implantation
Naoki Endo, Koichi Ogawa
PDF   Abstract
In the brachytherapy for the prostate cancer the calculation of the optimum location of seed sources is very difficult. The least square approach is sometimes not effective, because this therapy uses so many seeds and needles. In this paper we applied the genetic algorithm for deciding the location of seed source and we evaluated the parameter which required in the optimization. The results of simulation showed that the dose distribution is very sensitive to the parameter value being selected.
Reconstruction of I-125 Seed Sources for Prostate Implantation
Yuki Noda, Yoshiyuki Nyui, Koichi Ogawa
PDF   Abstract
Recently, permanent implantation of I-125 seeds into a prostate gland is performed for the treatment of a prostate cancer. In order to make an accurate treatment planning, we have to obtain seed positions accurately. The proposed method uses six X-ray films of a seed implanted prostate gland. With the help of a statistical image reconstruction method we could obtain the three dimensional seed positions correctly. The validity was verified by the phantom experiment.
Respiratory Movement Correction with Fiber Grating Device
Kei Tamagawa, Koichi Ogawa
PDF   Abstract
In the reconstruction of a myocardial SPECT image, the respiration causes serious blurring. The aim of this study is to compensate for this blurring by monitoring the movement of thorax. Our system uses a fiber-grating device (FGD) and an ITV camera for detecting the respiratory movement with images sequentially acquired during the data acquisition in SPECT or PET. The FGD makes a dot-matrix pattern on the patient’s surface and when we monitor a group of dot matrix patterns we can obtain the movement of the patient. This paper shows the preliminary results of a system for measuring respiratory movement with a fiber-grating vision sensor.
A New Light Collimator for Detecting Single Photons
Junya Kato, Koichi Ogawa
PDF   Abstract
This paper proposes a new light collimator which is used for detecting a single photon. The weight of the collimator is a half the conventional collimator with septa structure. In the new collimator we gave up to use many plane walls shaping a collimator hole. The performance of the collimator was clarified with the Monte Carlo simulation.
Public-key Cryptosystems and Linear Codes
Shuzo Matsuda, Toyokazu Hiramatsu
PDF   Abstract
In this paper, we will give a link between public-key cryptosystems and linear codes. It is well-known that the decoding problem for linear codes is in generally NP-complete. We propose a public-key cryptosystem by using this property.
On the Collatz Conjecture
Tomoaki Mimuro, Shigeto Nishimura, Toyokazu Hiramatsu
PDF   Abstract
The Collatz conjecture on the rational numbers is that there exists a positive integer n satisfying fn(m)=1 for any rational number m≥3, where f is the function on the rational number field defined by f(m)=m/2 if the numerator of m is even and f(m)=(3m+1)/2 if the numerator of m is odd. We will consider the cycles and generalized cases.
Easy Way of 3-Dimensional Graphics with Self-Organization Call
Takashi Sakamoto, Hajime Hirooka
PDF   Abstract
In many fields, the representation of 3 dimensional graphics gives a very useful tool in practice. Here we propose an easy metod to riconstruct a 3D graphics from two dimensional informations, such as photographs, with neural-net model, combining the consepts of Kohonen’ self-organization map and simple dynamical law. This algorithm of self-organization will give a possibility of easy way to riconstruct of 3D graphics.
Stress Analysis and Fatigue Life Evaluation at Welded Joint between Main and Lateral Girders in I-Shaped Composite Girder Bridges
Shigeyuki Hirayama, Shinsuke Akimoto, Takeshi Mori
PDF   Abstract
A lot of fatigue damage has been reported at welded joint between main and lateral girders in I-shaped composite girder bridges. The cause has been considered as complex stress field induced by stresses on the main girder web and the lateral girder flange. In this study, for the purpose of examining stress properties at the welded joint, three dimensional finite element analysis has been carried out to a composite steel girder of simple span. As a result, high stress concentration due to bi-axial stress is occurred at the main girder web. Furthermore fatigue life evaluation using the analytical results has been carried out.
Application of the Distributed Runoff Model Using GIS to the Ebi River Basin
Kei Mitsuhori, Daichi Nakayama, Yasumiti Oka
PDF   Abstract
In this study, a distributed runoff model, which has been verified by the flood runoff analysis in the Tsurumi River basin, is applied to the Ebi River basin. In addition, as for the division of study basin into small sub-basins, a new method based on the Digital Elevation Model(DEM) is adopted, instead of the former one which uses the generated river channel networks and ridge lines on the map as criteria for the manual work. As a result, the validity of the proposed model and its applicability to the urbanized basin are assessed by comparing the calculated hydrographs with the observed data
Research on the Oscillating Consciousness by the Vision on Condition of the Outside-of-the-window Scene of a Super-high-rise Building
Kenji Hara, Takeshi Goto
PDF   Abstract
A strong wind makes a super-high-rise building generate two kinds of long cycle level vibration called translation vibration and twist vibration. Many researches on a human body reaction or evaluation are done from a viewpoint of evaluation of habitability to translation vibration among them, and the standard and the indicator are enacted. However, this standard and indicator are aimed only at evaluation by physical feeling consciousness. By this research, when vibration has been perceived using the vision information from the outside-of-the-window scene of a super-high-rise building, the difference in the evaluation about the influence of a scene to evaluation of the vibration, translation vibration, twist vibration, and a synthetic vibration is investigated.
A Study of Visualization Simulation
Norimichi Ishida
PDF   Abstract
Computer visualization is essential as a means of consensus making by rapid technologicalchange. Significance of computer visualization is being censthesia of immersive on virtual space in stead of real world. We can use Creator and Vega of simulation software tool by MuliGen-Paradigm Inc. and type of hemisphere Vision Dome for virtual enviroment of display. The culture of visual simulation is most importance key technology of early 21 century.
Data Interpolation for the Discrete Wavelet Transform
Sawa Matsuyama
PDF   Abstract
A number of data for a discrete wavelets transform is required to be a power of 2, and therefore, part of the data obtained by field observations or laboratory experiments are not frequently applied for the analyses. We propose that Fourier transform is useful to interpolate and extrapolate the data for increasing the number of data for the discrete wavelets analysis. The raw data is firstly transformed to the Fourier coefficients by Fourier transform. Then the inverse Fourier transform is carried out to obtain the number of the data, i.e., the number of a power of 2.