武居 昌宏

In investigating a microchannel, understanding regarding dispersion reaction are need to be acquire first down to the view point of multiphase flow dynamics. Usually, various measurements methods and in-depth analyzation are necessary to understand thoroughly the inside reaction occurs in this kind of system. In this experiment, fabricated multilayer microchannel with commercial connecter system was made for a tomography system with electrodes attached. This fabricated multilayer microchannel measures impedance using 12 electrodes from 5 different positions continuously. In measuring the impedance, it is necessary to consider the electric field caused by the electrodes and the force distribution inside the particles. Because of a non-uniform electric field distribution inside the system, forces inside the particles are not equal thus net force is present. This study, discuss how the frequency of the electric fields affects the particle position as a driving force to the microchannel system.

An adaptive stereoscopic-PIV system has been constructed for flow analyses around a cylinder body (length=60mm, diameter=10mm, polystyrene). The body is floated by the swirling flows in a vertical pipe with a length of 1100mm and an inner diameter of 100mm. Four optical sensors (LEDs) were installed onto the pipe wall to detect the posture of the cylinder body. The constructed stereoscopic-PIV system is activated by the signals of the optical sensors when the cylinder is in the upright posture. The measurement system consists of two cameras, a Nd-Yag laser, and a host computer. Flow characteristics around the floating cylinder body and the motional characteristics of the body itself are investigated when the body is in the upright posture. It has been verified that the motion of the floating cylinder becomes stable when the azimuthal velocity component (W component) of the swirl flow is maintained at stable states.

Currently, wavelet transforms are widely used for the analyses of particle image velocimetry (PIV) velocity vector fields. This is because the wavelet provides not only spatial information of the velocity vectors, but also of the time and frequency domains. In this study, a discrete wavelet transform is applied to real PIV images of bubbly flows. The vector fields obtained by a self-made cross-correlation PIV algorithm were used for the discrete wavelet transform. The performances of the discrete wavelet transforms were investigated by changing the level of power of discretization. The images decomposed by wavelet multi-resolution showed conspicuous characteristics of the bubbly flows for the different levels. A high spatial bubble concentrated area could be evaluated by the constructed discrete wavelet transform algorithm, in which high-leveled wavelets play dominant roles in revealing the flow characteristics. (c) 2008 Elsevier B.V. All rights reserved.

Particles concentration in a fluidized bed with water pulsation for Lithium collection system in seawater has been measured using impedance method. As a result, the void fraction is increased in the lower layer and that is decreased in the upper layer as the frequency increases. Moreover, the void fraction in the lower layer has a reasonable agreement with the Richardson- Zaki equation. The amplitudes of vibration in the lower and upper layers increase as the frequency increases. However, it is not changed among the frequencies in the middle layer.

A new reconstruction method called sampled pattern matching (SPM) was applied to the image reconstruction of an electrical capacitance computed tomography in powder flow in a vertical pipe for petroleum refinery process. This new method is able to achieve stable convergence without the use of an empirical value. Experiments were carried out using fluid catalytic cracking (FCC) catalysts as powder with two air volume flow rates and four powder volume flow rates to measure the capacitance of a pipe cross section. The SPM method is compared with conventional methods in terms of volume fraction, residual capacitance, and correlation capacitance. Overall, the SPM method proved superior to conventional methods without any empirical value because SPM achieves accurate reconstruction by using an objective function that is calculated as the inner product calculation between the experimental capacitance and the reconstructed image capacitance.

The difference between rotary feeder system and the vessel system on transportation efficiency could not be made sure. But the air velocity in the rotary feeder system was slightly faster than that in vessel system under the same air mass flow rate. This means that pressure drop in the rotary feeder system is smaller than that in vessel. The cause is the number of incorporated plugs in the vessel is lager than that in the rotary feeder system, because deviation of plug velocity and plug length in the vessel is far larger than that in the rotary feeder system. Additionally, the rotary feeder system could take the transportation limit of air mass flow rate down. This is also caused from small pressure drop in the rotary feeder system.

The flow and the heat transfer characteristics in quartz microtube with inner diameter of 0.0196mm are investigated experimentally. By measuring the pressure drop between the inlet and outlet of microtube with different working fluid and the average temperature of microtube wall heated by steam, the corresponding friction factors and Nusselt number are obtained. The experimental results show the friction factors in microtubes exceeds the prediction of Hagen-Poiseulle due to the predominance of the electrical double layer effect (EDL effect) when the working fluid is distilled water, and the experimental Nusselt number is less than the classical laminar at Reynolds number<500 due to the effects of the variation of the thermophysical properties with temperature and the conjugate heat transfer. However, the effect of viscous dissipation results in temperature rise of the working fluid, due to which the friction factors are less than the prediction of Hagen-Poisculle with the working fluid of tetrachloromethane. Using nitrogen gas as the working fluid, the rarefaction effect still brings on that the friction factor is less than the classical laminar solution, though gas density and viscosity increases resulted from the stagnant velocity between gas and inner wall and the viscous heating lead to the increase of friction factor.

A new reconstruction method called Generalized Vector Sampled Pattern Matching (GVSPM) has been applied to all ill-posed inverse problem involving an electrical capacitance CT for solid air two-phase flow. The characteristics of GVSPM method were examined using a simulation for pseudo particle concentration distribution images and real experiment data. Overall, the accuracy is strongly dependent upon the image type and the iteration number. However the GVSPM method was proved Superior to the LW and the ITR methods in the case of annular pseudo particle images and particles with relatively low electric charge. Then, a sensor for capacitance CT was designed to visualize the powder concentration in the process of mixing air and FCC catalysts in a vertical pipeline. The concentration distribution images are obtained tinder certain air-catalyst parameter conditions. The relationship between the air-catalyst parameter condition and the powder distribution is analyzed ill detail. The accuracy of the reconstructed image was also discussed systemically in terms of volume fraction, residual capacitance, and capacitance correlation. Moreover, the particle concentration distribution images of a dense two-phase solid/air (plug) flow have been obtained at 10 milli-second intervals in a horizontal pipeline using capacitance computed tomography. The formation Of Plug was clarified by these reconstructed images.

Fluid flow in microchannels has emerged as an important area of research due to its various applications in medical and biomedical industries, computer chips, and chemical separations. To fabricate microchannel devices, a thorough understanding of the diffusion reactions on the inside of the microchannel from the viewpoint of multiphase flow dynamics is needed. Electrical impedance tomography (EIT) is a recently developed imaging technique in which conductivity or permittivity images of a subject can be rapidly collected with channels of external electrodes. Usually EIT is visualized into a cross-sectional image representing the distribution of density. Therefore, the application of EIT to microchannels is considered in this paper. This research is a first step to developing EIT for microchannel nanoparticles concentrations in a microchannel with 60 electrodes and measured using the cross-sectional impedance measurement technique. Using the EIT system in a microchannel, a particle distribution image can be obtained in a magnetic fluid. Based on the reconstructed image, it is easy to estimate the particle diffusion behaviors in a fluid flow.

In a plug transportation that is one of gas-solid two-phase flows, the prediction equations on particle velocity within a plug and pressure drop in a horizontal pipe have been formulated. The agreement between values calculated by these equations and experiments in which solid-air mass flow rate, pipe diameter, kinds of particle were changed was obtained. The difference between them is almost within 10%. In order to confirm the validity of supposition to derive these equations, particle velocity distribution within a plug in the directions of flow and radius, and particle velocity transformation from a stationary bed to a plug have been analyzed by high speed camera and PIV. The results show that there is no particle velocity distribution in a plug, and particles are accelerated uniformly in extra part of a plug. Namely, particles in a plug are fixed relative to each other and so they all move with the same velocity.

A measurement system has been developed to non-invasively detect plug formation in density pneumatic conveying. This plug formation detection system yields quantitative information that is validated by comparison to an electrical capacitance tomography (ECT) system. Both the plug detection system and the ECT system were applied to a horizontal pipeline (50 mm inner diameter) in density pneumatic conveying. The experimental conditions were varied for different heights of the initial particle dense layer and air volume flow rates. For all cases, the results obtained from the plug detector and ECT show similar patterns. This good agreement suggests that, where feasible, this plug detector can be used in place of the ECT system, which is advantageous since plug detection systems are often less expensive and faster than ECT systems. The factors influencing plug formation are discussed in detail. As a result, it can be seen that the height of the initial particle dense layer and the air volume flow velocity contribute to the nature of the plug formation rather than the solid-phase volume fraction in the dense layer. (C) Koninklijke Brill NV, Leiden and Society of Powder Technology, Japan, 2008

Particle concentration distribution images of a dense solid-air two-phase flow have been obtained at 10 milli-second in a horizontal pipeline by means of a capacitance computed tomography. The volume fractions based on the CT measurements were measured as parameters of air flow velocity and particle height of an initial accumulation layer. As a result, a plug is naturally formed at a specific velocity and height condition. Moreover, the factor of the plug natural formation is theoretically discussed with a two-flow model. The parameters, which are the initial height accumulation layer and the air flow velocity, do not make a big difference of the solid-phase volume fraction in the dispersion layer; however, make a difference of the accumulation layer height. Therefore, it seems that accumulation layer height contributes to the plug nature formation rather than a solid-phase volume fraction in a dispersion layer.

Fluid catalytic cracking (FCC) catalysts are used to produce high-quality petroleum from residual oil by the contact cracking process in the petroleum refinery. The key point in producing high-quality petroleum from residual oil is uniform fluidization at the mixing point. The sensor for capacitance computed tomography (CT) is designed to visualize the powder concentration in the process of mixing air and FCC catalysts in a vertical pipeline. The concentration distribution images are obtained under certain air-catalyst parameter conditions. The relationship between the air-catalyst parameter condition and the powder distribution is analyzed in detail. Moreover, the accuracy of the reconstructed image is discussed systemically in terms of volume fraction, residual capacitance, and capacitance correlation. (c) 2007 Elsevier Ltd. All rights reserved.

In a plug transportation that is one of gas-solid two-phase flow, the predicted equations on particle velocity within a plug and pressure drop in a horizontal pipe have been formulated. The agreement between values calculated by these equations and experiments that changed solid-air mass flow rate, pipe diameter, kinds of particle was obtained. The error between them is almost within 10%. In order to confirm the validity of supposition to derive these equations, particle velocity distribution within a plug in the directions of flow and radius, and particle velocity transformation from a stationary bed to a plug have been analyzed by high speed camera and PIV. As a result, there is no particle velocity distribution in a plug, and particles are accelerated uniformly in extra part of a plug. Particles in a plug are fixed relative to each other and so they all move with the same velocity.

Deioned water, used as the working fluid, flowed through the quartz glass microtubes with inner diameters of 45, 92, and 14-1 mu m, respectively, which was heated by a directly electrifying brass wire coiled evenly up around them. As the Reynolds number varied in the range from 100 up to 3000 in the experiments the corresponding Nussult number was obtained. The experimental results showed that the experimental Nussult numbers were less than predictions of the classical laminar correlations at a lower Reynolds number, which resulted from the effects of the variation of the thermophysical properties with temperature and nonlinear temperature variation of the working fluid along the wall. With the increase of the Reynolds number, the experimental data sharply increased and they were higher than the predictions by the classical transitional correlations due to the conjugate effects of the thinner conductive liquid layer, the entrance length, the viscous dissipation, and the wall roughness when the Reynolds number exceeded 1000.

Measurements on pulsed impinging jets with 3D-particle tracking velocimetry (3D-PTV) system and 6-degree-of-freedom (DOF)-motion tracking system were carried out. Pulsed round jets were impinged on an elastic plate and the flow field near the plate was measured with a 3D-PTV system while the motions of the flexible plate was measured with 6D-motion tracking system. The measurement system consists of four cameras, light sources (Nd-Yag laser, Ar-ion laser, Black lamp) and a host computer. The nozzle diameter is D = 15 mm and two major experiments have been carried out for the cases of the distances between the nozzle tip to the elastic wall are z/D =2.3 and 6.0. The pulsed jets were controlled by a solenoid valve and were impinged onto an elastic plate (material: silicon, diameter: 350 mm, thickness: 0.5 mm, hardness: 15). The measurement system was synchronized with the valve opening time. The Reynolds numbers were 20,000 and 24,000 when the jets were impinged with the volume velocities. In the first experiments a macroscopic interprets on the flow-structure interactions (FSI) was made using three-dimensional vector fields of the flow and three-dimensional displacements of the elastic plate. In the second experiments a microscopic interprets on the FSI was made using two-dimensional velocity vectors and three-dimensional displacements of the elastic plate. Experimental results showed that the elastic plate moved slightly to the opposite direction of the jet direction at the time of valve opening. It has been shown that the vortices travelling over the surface of the wall made the elastic wall distorted locally. (C) 2006 Elsevier B.V. All rights reserved.

Distilled water and nitrogen gas used as the working fluids flow through the stainless steel microtube with inner diameter 168 mu m outer diameter 406 mu m. Using the Joule heating, the wall temperature field photos of the microtube is acquired by employing an IR camera and the temperature and the volume flow rate of the inlet and the outlet of microtube are measured. A correlation between the axial wall heat conduction and the convective heat transfer is obtained by theoretical analysis based on the experimental results. The investigative results clearly show that the axial heat conduction can reduce the convective heat transfer in the stainless steel microtube and the decrement may reach 2% compared to the convective heat transfer when the working fluid is nitrogen gas, however, the decrement can be neglected for distilled water as the working fluid.

Experiments were carried out to investigate forced convective heat transfer of deionized water flowing through quartz microtubes with inner diameter of 242, 315 and 520 pin. The microtube was heated by directly electrifying a brass wire that was coiled evenly up around the microtube to keep a constant heat flux, or heated by steam to keep a constant temperature. As Reynolds number varied in the range from 100 up to 5000-7000 in the experiments, the corresponding Nusselt numbers were also measured. The experimental results were compared with those of the classical correlations for convective heat transfer in a laminar, transitional or turbulent regime for conventional tubes, and a difference was found. The comparisons indicated that the Nusselt number tends to be agreement with those of the classical laminar correlations at a lower Reynolds number, and increased sharply with the increasing of Reynolds number. If Reynolds number was larger than 1500, 1600 and 1900 for microtubes with inner diameter of 242, 315 and 520 pm, respectively, the corresponding Nusselt number was higher than the predictions by the classical transitional correlations. When Nusselt number reached the value from the classical turbulent correlations, Reynolds number was approximately equal to 4000, 4500 and 5500, respectively, for these types of microtubes. (C) 2006 Elsevier Masson SAS. All rights reserved.

This paper describes the state of the art in the application of visualization techniche in the field of scientific arts, which cover wide academic interests, such as art, social science, education, archaeology, literature and so on. The special attention is placed on the recent publications of scientific arts in Journal of Visualization, and a few examples are described. These visualization examples provide a unique category in the recent development of visualization research.

Homogeneous distribution of concentration is an important theme in solid air two-phase flow of chemical industry. Particles heterogeneous distribution it influence the quality of a product. However, the analysis is difficult because data are enormous as for the imaging. Therefore the image from capacitance CT analyzed by image vectrization in this study.

Solid air two-phase flow is used for pipeline transportation and powder mixture in a chemical plant. However, a concentration change of the pipe inside, is not elucidated. In this study, the concentration distribution obtained by capacitance CT is performed analyze a change with spectrum analysis.

Recently, as a useful noninvasive measurement, computed tomography is extensively applied to real industry to measure and visualize particle behavior in a cross-section of a pipeline. In this research, particle concentration distribution images of dense multiphase flow are obtained in a horizontal micro channel using impedance computed tomography. The cross section of the channel is rectangle. The impedance between each electrode pair, which is arranged around the channel, is measured for different particle consistency. Then the particle concentration distribution images are reconstructed based on the measured impedance values.

An adaptive stereoscopic-PIV system has been constructed for flow analyses around a cylinder body (length=60mmm, diameter= 10mm, polystyrene). The body is floated by the swirling flows in a vertical pipe With a length of 600mm and an inner diameter of 100mm. Four optical sensors(LEDs) were installed onto the pipe wall to detect the posture of the cylinder body. The constructed stereoscopic-PIV system is activated by the signals of the optical sensors when the cylinder is in upright posture. The measurement system consists of two cameras, a Nd-Yag laser and a host computer. Flow characteristics around the floating cylinder body and the motional characteristics of the body itself are investigated when the body becomes upright posture. It has been verified that the motion of the floating cylinder becomes stable when the azimuthal velocity component of the swirl flow is maintained at stable states.

The prediction equations on particle velocity within a plug and pressure drop in vertical pipe have been formulated by considering fluidized particles. The gas-liquid two-phase theory established by Davidson and Harrison has been assumed as fluidized particles. As a result, the all kinds of calculated values on particle velocity and pressure drop are in good agreement with the experimental values while internal pipe diameter and kinds of particles are changed. The error between calculations and experiments is almost within 10%. Additionally, the plug velocity and particle velocity in the flow and radius directions have been measured by digital video camera, high speed camera and PIV. As a result, gas-liquid two-phase theory can be applied to the relation between plug velocity and particle velocity, and there is no particle velocity distribution in the plug. In short, particles in the plug are fixed relative to each other and so they all move with the same velocity.

Ergun's equation is modified by introducing the ratio of straight length to particle perimeter and the wall effect of particle bed. It can predict the pressure drop of particle bed of any shapes of coarse particles for various air velocities and pipe diameters, with an error of less than 6%. The wall effect is significant when the ratio of pipe diameter to particle diameter is smaller than 10 and the air velocity is lower than 3m/s. Ergun's equation gives poor prediction for the bed of non-spherical particles because these particles are in contact with a large area which is related to the ratio of straight length on perimeter of particles. A large deviation is found for large ratio of straight length to particle perimeter.

Pressure drop of a long packed granular bed is usually calculated by using the air velocity at the mid-length of bed because a decrease in pressure towards the bed outlet brings the expansion of air and therefore an increase in air flow velocity. However, for a long bed consisting of small particles with a low voidage at a high air velocity, the air pressure decreases so significantly that the prediction of pressure drop with the air velocity at the mid-length of the bed is no longer applicable. Consequently, the pressure drop has to be calculated in an integral manner by considering the air velocity increase across the bed. The present work estimated the error between the pressure drop predicted with the mid-length air velocity and that in an integral manner for various lengths of bed. As a result, it is shown that the pressure drops at various air velocities can be readily predicted when the pressure ratio at two locations in the bed is known at a given air velocity.

本研究では，旋回流搬送装置設計の第一段階として，ＴｙｐｅＡ，ＴｙｐｅＢおよびＴｙｐｅＣの３種類の旋回流発生装置を試作した．そして，各装置を比較検討するために，鉛直パイプ内の速度分布を３Ｄ－ＰＩＶで可視化計測し，得られた速度分布の2次元ベクトル画像に対して，モーダルウェーブレットによる多重解像度解析を行った．

For a plug transportation in gas-solid two-phase flow, the prediction equations on particle velocity in a plug and pressure drop in a vertical pipe were formulated by considering fluidized particles. These equations well predicted the experimental data obtained with various solid-air mass flow rates, pipe diameters, and types of particles with the error within around 10%.<br>In order to confirm the validity of the suppositions introduced in deriving these equations, the particle velocity distributions within a plug in the axial and transversial directions of flow were analyzed by a high speed camera and PIV. As a result, there was very little distribution in particle velocity in a plug, indicating that the particles in a plug are fixed relative to each other and therefore they all move with the same velocity.

A new reconstruction method called Generalized Vector Sampled Pattern Matching (GVSPM) was applied to the image reconstruction of an electrical capacitance computed tomography (ECT) in freely falling particles in a vertical pipe. This new method is able to achieve stable convergence without the use of an empirical value. Experiments were carried out using three particle types with various electric charges and four particle flow rates to measure the capacitance of a pipe cross section. The three particle types were polyethylene pellets (PP), silica sand (SS) and polyvinyl chloride (PVC). Four flow rate settings were used resulting in a volume flow rate ranging from 3.08×10^<-5> to 1.04×10^<-3> [m^3/s]. The GVSPM method is compared with conventional methods in terms of volume fraction, residual capacitance, and capacitance correlation. Overall, the GVSPM method proved superior to conventional methods in the case of polyethylene pellets with high electric charge. GVSPM achieves accurate reconstruction by using an objective function that is calculated as the inner product calculation between the experimental capacitance and the reconstructed image capacitance.

A new reconstruction method called generalized vector sampled pattern matching (GVSPM) was applied to the image reconstruction of an electrical capacitance computed tomography in freely falling particles in a vertical pipe. This new method is able to achieve stable convergence without the use of an empirical value. Experiments were carried out using three particle types with various electric charges and four particle flow rates to measure the capacitance of a pipe cross section. The three particle types were polyethylene pellets (PP), silica particles (SP) and polyvinyl chloride (PVC). Four particle flow rate settings were used resulting in a volume flow rate ranging from 3.08 x 10(-5) to 1.04 x 10(-3) m(3) s(-1). The GVSPM method is compared with conventional methods in terms of volume fraction, residual capacitance and correlation capacitance. Overall, the GVSPM method proved superior to conventional methods in the case of polyethylene pellets with relatively low electric charge. GVSPM achieves accurate reconstruction by using an objective function that is calculated as the inner product calculation between the experimental capacitance and the reconstructed image capacitance.

Particle concentration distribution images of a dense solid-air two-phase (plug) flow have been obtained at 10 ms intervals at a bend pipe upstream in a horizontal pipeline by means of a capacitance computed tomography. The three-dimensional images (time and two-dimensional space images) have been decomposed to the wavelet time levels to extract the dominant particle concentration distribution using three-dimensional discrete wavelet multiresolution. As a result, the time dominant particle distribution with specific time frequency level is visualized in a cross-section. In detail, the high concentration of the particle spatial distribution at the dense flow front, which composes high-time frequency levels 6 and 7, is located at the center above the stationary layer.

This paper describes the application of discrete wavelet transforms to the analysis of condensation jets in order to clarify the associated fluid and heat transfer phenomena. An experimentally-obtained, two-dimensional image of the condensation particle density around the jet was decomposed into 7 levels of resolution with their respective wavelengths. Based on the known physical characteristics of turbulent flow around the jet, levels 0 and 1 were shown to represent the large-scale components of the condensation particle density and the higher levels represent the small-scale components. From the wavelet-analyzed images, the width of the condensation zone was obtained and this compared well with the width inferred from temperature measurements. Thus, the method was verified and also provided data not available experimentally.

We have conducted a basic experiment on visualization of a pipe which is used for transport catalyst in industrial use. First two copper-electrodes are attached to inner side and outer side of the container. We have charged 5volt to the interelectrode and have changed frequency from 5kHz to 100kHz to obtain stable capacitance. The experiment has two conditions that container is filled with catalyst or air. The stable conditions have been confirmed at the time more than 50kHz. We have made a CT sensor with 12 electrodes attached to the pipe which has diameter of 270mm based on the experiment result.

The 4th World Congress on Industrial Process Tomography (WCIPT4) was held in Aizu, Japan, from September 5 to 8, 2005 (home page: http://www.vcipt.org.uk/ wcipt4/). The congress venue was OnYado Toho, a traditional Japanese hotel in Higashiyama Onsen, which is approximately 15 minutes from Aizu-wakamatsu station by car and approximately 4 hours from Narita Airport by train and shinkansen. The congress attracted 183 participants from 21 countries. Eight categories of parallel sessions and two poster sessions were held. A total of 174 papers were presented, including 3 invited lectures and 40 posters. In the plenary lectures, Professor Xi-Cheng Zhang presented "Terahertz Wave Imaging Technology", Professor Lynn Gladden explained the "Latest Development in Dynamic MRI of Multi-Phase Systems", and Professor Hideaki Koizumi discussed "Optical Topography for Higher-order Brain-function Imaging and its Practical Applications". The congress also incorporated some local events, including a technical tour of the University of Aizu. The university staff introduced the state-of-the-art computer facility. In addition, a city sightseeing excursion visited Tsurugajo Castle and a sake museum. All attendees enjoyed lavish dinners with traditional Japanese entertainment in the evenings. © 2006 The Visualization Society of Japan and Ohmsha, Ltd.

A modal wavelet transform, which overcomes the intrinsic data number limitation of power of two to conventional wavelet transform, has been applied to analysis of pseudo and real bubbly flow PIV images. The modal wavelet transform is compared with the discrete wavelet transform in order to select the best base function among Neumann, Dirichlet and Green function types base functions. Consequently, it is verified that Neumann type base function is the best because the correlation of Neumann type base function is the highest. From the result of wavelet analysis of the real bubbly flow PIV image, as the relative velocity is higher, the dominant eddy scale becomes smaller. The extraction modal wavelet level depends on the base function. Copyright © 2006 by ASME.

A simultaneous measurement system that can analyze the flow-structure interactions has been developed. This system consists of four CCD cameras, two for capturing instantaneous flow fields and two for tracking a solid body. The three-dimensional vector fields around a cylinder are measured while the motion of the cylinder forced by the flow field is measured simultaneously with the constructed system. The cylinder is suspended in the circulating water channel and the surface of the working fluid is forced sinusoidal to make the cylinder bounced. Reynolds number for the mean main stream is about 3500. The interaction between the flow fields and the cylinder motion is examined quantitatively.

本研究では，モーダルウェーブレット変換と離散ウェーブレット変換を比較し，前者の有効性，ならびにモーダルウェーブレット変換の各基底関数を比較検討し，最適な基底関数を選択することを目的とした。 模擬ＰＩＶ画像, および標準画像に対して , モーダルウェーブレット変換，および離散ウェーブレット変換をそれぞれ施し，加算多重解像度解析結果から，基底関数の比較検討を行った．モーダルウェーブレット変換の基底関数にはノイマン型，ディリクレ型，およびグリーン関数型を用い，離散ウェーブレット変換の基底関数にはＤＡＵＢＥＣＨＩＥＳ４次基底を用いた． 模擬ＰＩＶ画像，標準画像の解析結果より ，本研究においては，総合的に考えると，モーダルウェーブレット変換の最適基底関数は ，ノイマン型であると思われる．

A modal wavelet transform, which overcomes the intrinsic data number limitation of power of two to conventional wavelet transform, has been applied to analysis of axial and eddy pseudo velocity fields, standard PIV velocity field and experimental PIV measurement. The modal wavelet transform is compared with the discrete wavelet transform in order to select the optimum basis function among Neumann, Dirichlet and Green function types basis functions. Consequently, it is verified that Neumann type function is the best basis because the correlation of Neumann type basis function is higher and the root mean square is lower than the other basis functions. Also, the decomposition vector patterns by Neumann type are similar to that by conventional Daubechies basis function of 4th order.