Search

LDV measurements in a 90 degree elbow, of which curvature radius coincides with the diameter have been conducted for cases of developed inflow, undeveloped inflow, swirling inflow, and deflected inflow that became faster at the inside. This paper especially focuses on a result of the deflected inflow, comparing with a result of the undeveloped inflow. The result shows that the deflected inflow reinforces a convex velocity distribution occurring near the curvature inside in the downstream region, concluding that the deflected inflow promotes the secondary flow of Prandtl's first kind in the elbow. Frequency analyses of axial flow fluctuations show that the dominant frequency can be seen in the downstream of the elbow even in the deflected inflow case. However its Strouhal number increases to 0.6 from 0.5, that is for the short pipe case. Results of frequency analyses are also shown for other cases that we have been examined. Dominant Strouhal numbers in most of the cases become 0.5, except for 0.6 in cases of the developed and deflected inflow. This frequency shift might be related with the boundary layer size and the local flow velocity, since the corresponding fluctuation is caused by vortex shedding from the boundary layer at the elbow inside.

The present experimental study deals with a pulsatile turbulent flow simulating the exhaust flow of an automotive engine. In the experiments, a four-cylinder engine is used as an exhaust-flow generator to realize flow conditions close to those in an engine environment. Particle image velocimetry (PIV) measurements visualize the flow field in an S-shaped double-bend duct at a Reynolds number of 48,000 and a Womersley number of 70.9. Stereo PIV, which is classified as a two-dimensional three-component measurement, is conducted in the duct cross sections located downstream of the bends. The stereo PIV system is synchronized with the engine operation to enable phase-locked measurements at particular phases, and the phase-averaged results show the large-scale vortical structures and the duct axial velocity distribution. Downstream of the first bend, the secondary flow consists of vortices that circulate as Dean-type vortices. Downstream of the second bend, by contrast, vortices that circulate in opposite directions to the Dean-type vortices, so-called Lyne-type vortices, form in the core of the cross section. These secondary flows persist without significant changes in their large-scale vortical structures over time. Time-resolved PIV is conducted to track the temporal evolution of the flow in the bend planes. The results show that the flow reverses locally along the inner wall of the bends during flow deceleration. Snapshot proper orthogonal decomposition (POD) is used on the time-resolved PIV data to extract the significant flow structure from the instantaneous field. We propose POD as a good post-processing tool for the instantaneous data of pulsatile cases.

Particle image velocimetry (PIV) and computational fluid dynamics (CFD) modelling of blood flow through a carotid artery bifurcation has been carried out in order to assess the role of haemodynamics in atherosclerosis and validate a novel wall shear stress (WSS) measurement method via detailed quantitative data analysis. Velocity and WSS data, obtained by PIV was compared against CFD-predicted data obtained for the same geometry and boundary conditions. The results demonstrate that both methods are capable of capturing essential flow characteristics, and are in good agreement. The axial velocity data clearly show the important flow features such as skewed velocity profiles and a low-momentum region near the sinus outer wall, although with slight errors between the two techniques in the order of 6.1-13.7%. The secondary velocity plots and three-dimensional particle traces reveal complex flow features within the sinus, such as Dean vortices and helicoidal flow. WSS data obtained by both techniques reveal the presence of a low-momentum and reversed-flow regions, and show agreement with errors in the order of 1% in the common carotid artery and 0.9-9.8% in the sinus. The error sources in both the experimental and numerical methodology are discussed and recommendations for future applications are made.

A high-frequency (20kHz) standing wave was applied to the unburned mixture upstream of a methane-air lifted jet flame using a bolt-clamped Langevin transducer (BLT) to improve stability. The flow field near the flame was visualized using acetone planar-laser-induced fluorescence (PLIF). The standing wave decreased the lifted flame height and increased the blow-off limit. The upstream flow field of the center jet then bent. This phenomenon appeared when there was a density difference between the center jet and the surrounding secondary flow. When the density of the center jet was less than that of the co-flow, the center jet was redirected to the pressure anti-node side. Conversely, when the density of the center jet was greater than that of the co-flow, the center jet was redirected to the pressure node side. This redirection tended to stabilize the laminar lifted flame.

Experimental investigations were conducted to study the effect of micro-bubbles injected into water flowing through straight and helical pipes with different curvature ( 0.025, 0.05 and 0.1). The micro-bubbles were produced using two hydrocyclone micro-bubble generators positioned face to face with opposite direction of swirl flow. The mean diameter of the micro-bubbles generated was about 60μ m, while the maximum diameter observed was 174μ m, with air fraction, which will be defined in Eq.(4), ranging from 0.21% to 0.44%. The effect of the Reynolds number on the drag reduction is also presented. The experimental results show that, the micro-bubble has an effect of drag reduction on both straight and helical pipes, though the drag reduction in a straight pipe is much higher than that in a helical pipe, because of the secondary flow in a helical pipe. It is found that the drag reduction increases with the decrease of curvature and the increase of air fraction. The maximum value of drag reduction ratio was 51% in case of a straight pipe, while that for a helical pipe was 16%.

This study experimentally deals with turbulent flow inside transverse-or angled-rib-roughened square two-pass channel in stationary condition. Details of the two-dimensional flow field were quantitatively measured by using a particle image velocimetry to which was applied Frame-staddling method with high-resolution CCD camera. Angled-ribs generated secondary flow along the rib, flow separation, and reattachment similar to the transverse-rib case. A secondary flow of turning flow caused by a centrifugal force in a sharp turn was stronger than that caused by the angled ribs. Furthermore, the interference between the secondary flow by the sharp turn and the angled ribs was observed and high turbulence intensity between the ribs and after the sharp turn as a result of highly fluctuating unsteady fluid motion was clearly captured.

In the previous study, it was clarified in the right-angle branch with single side branch that the velocity periodically oscillates in the side branch, and the Strouhal number is independent of cross-sectional area ratio, the flow division ratio of side branch to main branch, and the Reynolds number. In the present study, the flow structure in the right-angle branch with double side branch has been experimentally studied in laminar steady flow using laser Doppler anemometry and particle image velocimetry. Consequently, the velocity also oscillates periodically in both side branches, and the Strouhal number is independent of the Reynolds number although the Strouhal number in the downstream side branch is a little larger than that in the upstream side branch. It has been clarified that the flow oscillation in the right-angle branch is a characteristic phenomenon caused from its geometry and the factor inducing oscillation would be high shear rate.

In order to address recent environmental problems such as global warming due to CO2, emissions, continuous efforts are required to improve the performance of the steam turbine, which supports the base of energy conversion. Although the aerodynamic design technologies of steam turbine have been developed remarkably by applying the computational fluid dynamics (CFD), it is not easy to reduce the losses within a low aspect ratio cascade in the control stage and high-pressure stages even if the latest three-dimensional design are used. Then, the authors investigated the loss generation mechanism in a low aspect ratio cascade using the smoke visualization and detailed internal flow measurement with a micro pitot probe to develop the new technologies for reducing the secondary loss. This paper describes the evolution of horse-shoe vortex and passage vortex through a turbine blade in experimental studies, theoretical approach to explain the motion of the vortex and the effects of inlet boundary layer on passage vortex interaction analyzed by three-dimensional viscous flow analysis.

In order to develop the performance improving technology for low aspect ratio cascade of the steam turbine, mechanism of the secondary loss generation has been investigated using the smoke visualization and detailed internal flow measurement in phase I. Next in phase 2, flow control technologies such as the blade fillet, endwall convergence and negative curving are introduced. This paper presents the loss reduction mechanism of these flow control technologies by applying the computational fluid dynamics (CFD) and the test results of air model turbine to verify the effects of endwall convergence and negative curving in rotating flow field. In conclusion, the technologies greatly contribute to improving the performance of the actual steam turbine with low aspect ratio cascade.

A high-frequency (20kHz) standing wave was applied to the unburned mixture upstream of the methane-air lifted jet flame using a bolt-clamped Langevin transducer (BLT) to improve stability. The flow field near the flame was visualized using acetone planar-laser-induced fluorescence (PLIF). The standing wave decreased the lifted flame height and increased the blow off limit. The upstream flow field of the center jet then bent. This phenomenon appeared when there was a density difference between the center jet and the surrounding secondary flow. When the density of the center jet was less than that of the co-flow, the center jet was redirected to the pressure anti-node side. Conversely, when the density of the center jet was greater than that of the co-flow, the center jet was redirected to the pressure node side. This redirection effect tended to stabilize the laminar lifted flame.

Computational fluid dynamics (CFD) simulations were carried out for an oscillatory flow in a 3D realistic model of the human central airways to reveal the effect of airway geometry on the oscillatory flow structure. This 3D realistic airway model is a computational model of multi-branching airway constructed based on X-ray CT images of actual human airways in which the airway diameter ranged from approx. 2 to 14 mm, and the flow in this airway model was simulated using CFD software (Fluent). The resultant inspiratory flow patterns in this 3D realistic airway model were relatively similar to patterns observed in a simplified planar multi-branching airway model, whereas the expiratory flow patterns strongly depended on the realistic airway geometry and showed more complicated secondary flow patterns. Secondary flow velocities were higher in the realistic airway model than in the simplified airway model in both the inspiratory and expiratory flows. Langrangian fluid particle tracking clarified the convective dispersion due to asymmetric inspiratory and expiratory velocity profiles.

Energy loss due to the wall skin friction in a turbomachinery impeller is directly related to the mechanical efficiency. Generally, the flow is modeled as a flow around an enclosed rotating disk, and we can estimate the energy loss of the impeller by applying the analytical or experimental results of the friction moment for a rotating disk. Several analytical and experimental studies have been performed in order to examine the friction moment and reduce the skin friction drag. The flow behavior of the boundary layer on the rotating disk surface is strongly affected by the vortices that are formed the spiral streak for the disk rotation stationary. In order to reduce the drag caused by disk friction, the behavior of the vortices in the boundary layer must be controlled. In the present study, the drag reduction phenomena of an enclosed rotating disk were clarified experimentally using a disk with several fine spiral grooves made by a wet etching process. Experiments were carried to measure the frictional moment of a disk having 120∼160 fine spiral grooves, the depths of which ranged from 100∼200 μm, in Newtonian fluids. The shape of the grooves was obtained by referring to spiral streaks in the boundary layer. The Reynolds number range is 4×104 < Re < 6×105. The maximum drag reduction ratio is approximately 15%, and the phenomena depend on the number, the spiral angle, and the depth of the grooves. The experimental results for the velocity profile and the flow visualization clarified that the fine spiral grooves control the secondary flow of the boundary layer on the disk surface and delays the generation of local turbulence in the transition range.

This study experimentally deals with turbulent flow inside smooth and transverse-rib-roughened square two-pass channels in stationary condition. Details of the two-dimensional flow field were quantitatively measured by using a particle image velocimetry which has high spatial- and time-resolution. In the transverse rib case, reattachment and recirculation behind the rib, and small separation bubble in front of the rib were observed. High turbulence intensity in front of the rib as a result of highly fluctuating unsteady fluid motion was clearly captured. Furthermore vortices caused by the sharp turn were observed in the secondary flow.

Flow field modified by a single large bubble in a horizontal wall turbulent boundary layer is measured by particle tracking velocimetry. We focus on intermediate bubble size being comparable to the thickness of boundary layer to find out the events altering the original turbulent shear stress field. The results are all presented by quantities relative to single-phase flow on Lagrangian coordinate that moves with the bubble. The measurement results reveal bubble-produced secondary flow around itself, which involves twin roll vortices, separation on the bubble surface, and strong sweeping flow. For a small and nearly spherical bubble, the sweeping flow is provided strongly to enhance the turbulent momentum exchange while negative exchange is detected for a large flat bubble. This effective length approximately corresponds to the vertical size of the bubble.

The mean velocity field of turbulent jet issuing from rectangular nozzle (Aspect Ratio=12.5) with a rectangular notch at the midspan, has been investigated experimentally. Four aspect ratios of rectangular notch (NAR) were 2.5, 7.5, 12.5 and 165. The Reynolds number based on the nozzle width d and the exit mean velocity Ue, was kept constant 30000 (NAR=2.5 and 7.5), 15 000 (NAR=12.5) and 13 000 (NAR=165), respectively. In this experiment, the variation of the potential core length and the changing rate of both the velocity and the length scales in the characteristic decay region were examined, and the present study suggested the possibility of the engineering applications. From the relationship between these results and the inward secondary flow producing both on the rectangular jet axis and the rectangular notch axis, effects of the rectangular notch on the development of the rectangular jet were clarified. Furthermore, the difference of the NAR value changes both the magnitude and the direction of the secondary flow. Finally, it is revealed that the potential core length takes the maximum value in the case of NAR=12.5, and the half velocity width bz/d on the rectangular jet axis shows the minimum value in the case of NAR=165.

Active control of the diffusion of a circular jet was attempted by application of a secondary film flow around the jet. Sinusoidal acoustic excitation of the film flow was carried out for VR values of 0.5 and 1.0, where VR is the ratio of the film flow velocity to the main jet velocity. For VR = 0.5, the diffusion of the jet was suppressed compared to that of a single jet but it was somewhat enhanced by the acoustic excitation. The acoustic excitation shortened the potential core of the jet; however, the entrainment of the ambient flow was far less than in the case of a single jet, regardless of degree of excitation. For VR = 1.0, diffusion was enhanced by acoustic excitation. Both turbulence intensity and ambient flow entrainment increased downstream of the jet. We conclude that film flow can control diffusion and that acoustic excitation can enhance the diffusion of jet flows. It is worth investigating this mechanism in detail in future studies.

In a series of present studies, the flows within the stationary curved ducts are analyzed numerically with the aerodynamic or the geometrical parameters affecting the loss generation caused by the passage vortex within a passage of a turbomachinery. In the former reports, the inlet boundary layer thickness and the inlet velocity distortion were taken as the aerodynamic parameter and the aspect ratio of the cross-section was chosen as the geometrical parameter. In this report, the effects of the curved angle of the bend on the loss generation caused by the passage vortex are examined by relating the bend of curved duct to the blade-to-blade surface of an axial flow turbine cascade. The curved angle is changed from 90 to 160 degrees by 10 degrees with fixing the length of arc for the mean radius of curvature of the bend or with fixing the mean radius of curvature.

Jets issuing through small holes into a freestream have proven effective in the control of boundary layer separation. The beneficial effect of separation control is obtained only if the jets are pitched to the lower wall and skewed with respect to the freestream direction. In the previous study, the suppression effect of the orifice shape on flow separation was reported for three types of jet orifice (circular, triangular and square orifices). The triangular orifice generates the strong vortex and makes effective the pressure recovery in the diffuser in comparison with the circular and square orifices. In this study, four types of the jet orifice (model-U, model-D, model-R and model-L) are installed and the effect of the direction of the vertex of the triangular orifice on separation control is investigated in a two dimensional diffuser. Separation control is accomplished by all models, however the suppression of flow separation is most ineffective among the four models in the case of model-L (the vertex points to the direction of the jet pitch angle).

The rate of circulating flow passing through the tongue division and the pressure recovery in the casing are investigated in a centrifugal blower, which has no diffuser and large tongue clearance. At the design point, the ratio of circulating flow rate to total flow rate at the scroll end is about 45 percent, and the pressure recovery in the scroll is about 45 percent. As the remaining 55 percent of the pressure recovery in the casing is recovered in the section between the scroll end and exit duct, this section acts as a high-performance diffuser. It is concluded that the centrifugal force produced by the secondary flow and boundary layer suction are the main causes of the large pressure recovery. Moreover, it is found that the flow conditions from the scroll toward the exit duct are closely related to the low flow rate characteristics of the blower.

Flow mechanisms within a monopivot centrifugal pump were clarified in order to prevent stagnation around the pivot bearing, which may cause thrombogenesis. We focused on the geometric effects of the pump, which included the effects of the washout hole diameter, the pivot friction area and the back gap width of the impeller relative to the washout around the pivot bearing. Flow patterns were carefully examined around the pivot bearing, including the region inside the washout hole and the back gap of the impeller, by computational fluid dynamic analysis. Based on the results from the computational fluid dynamic analyses, we found that a balance relationship between the washout hole diameter and the back gap width of the impeller affected the secondary flow toward the pivot bearing that eliminated the stagnation around the pivot bearing. In addition, while increasing in the pivot friction area eliminated stagnation around the pivot bearing, it also increased hemolysis within the pump.

The mechanism of secondary flow was discussed based on the result of an LES for the turbulence flow parallel to bundled rods. The LES was carried out in the Cartesian grid system using an improved immersed boundary method developed for accurate simulation of complex wall boundary of industrial interest. The method was verified by comparing the LES with a DNS of turbulence flow in a pipe. The LES for bundled rods reasonably reproduced the secondary flow of the second kind. We proposed a theoretical method to analyze the mechanism of the secondary flow. The result of the analysis indicated that the main source of secondary flow was the non-conservative component of Reynolds stress term as was the case for a square duct. The theory proposed that non-conservative force corresponds to the difference between the pressure gradient along the wall and the Reynolds stress term decreasing rapidly near the wall. The present theory will unify the understandings for the secondary flow of the first and second kinds. The present LES technique and theory is useful for investigating the physics in turbulence flow in complex walls.

Secondary mean motions of Prandtl's second kind near an undulatory surface are explained in terms of the turbulent blocking effect and kinematic boundary conditions at the surface, and the order of magnitude is estimated. The isotropic turbulence is distorted by the undulatory surface with a low slope h/λ, where λ is the wavelength and h is the amplitude. The prime mechanism for generating the mean flow is that the far-field isotropic turbulence is distorted by the nonlocal blocking effect of the surface to become the anisotropic axisymmetric turbulence near the surface with a principal axis that is not aligned with the local curvature (slope) of the undulation. Then, the local analysis can be applied and the mechanism is similar to that of the mean flow generation by the impingement of homogeneous axisymmetric turbulence over a planar surface, i.e. the gradient of the Reynolds stress caused by the turbulent blocking effect generates the mean motion. The results from this simple analysis are consistent with the previous exact analysis, in which the effects of the curvature are taken into account. The results also qualitatively agree with the flow visualization over an undulatory surface in a mixing box.

In the previous study, it has been clarified in the right-angle branch that the periodical oscillation flow is induced in the side branch and the Strouhal number based on the variables in the side branch is independent of the Reynolds number, the flow division ratio, and the radius of the side branch. This periodical oscillation appears along the boundary between two pairs of vortices with the opposite rotation at the cross section of the side branch across the separation flow. In the present study, another source of the periodical oscillation is experimentally clarified that the high shear rate flow is induced along the separation shearing layer between the separation flow of low velocity and the main flow with high velocity along the distal wall in the side branch since the high shear rate along this separation shearing layer is several times as large as that in the upstream trunk or in the downstream region of the side branch.

A method for controlling the position of an oscillatory cavity jet flow is demonstrated. The method involves secondary injection of a lower mass flow control jet into the cross-flow region of the primary jet. The primary jet in this case is a turbulent jet (Re = 55 000) which when injected into the rectangular cavity with no secondary control, attains a stable oscillation with a characteristic Strouhal number of S t,w = 0.013. The injection control method is investigated using a combined experimental and numerical approach with a water model test rig and a 2D and 3D computational fluid dynamics (CFD) model. Based on previous work, a baseline cavity, with a depth to width ratio of H/W = 0.16 and entry nozzle submergence of S/W = 0.38, is used to study the effect of secondary jet injection parameters on primary jet deflection angle (δ) as a function of momentum ratio (β) and injection position (Y i ). Results have shown primary jet deflection angles (6) of up to 15° for a momentum ratio (β) of 20% can be achieved for a secondary jet injection position of Y 1 /W = 0.12.

One of the objectives of this study is to present devices of generating ideal oscillating grid turbulence in a rotating tank. Secondary flow which prevents the generation of ideal grid turbulence was suppressed by narrowing the space between the turbulence generation grid and the tank wall, modifying the support of the grid, and installing an inner cylinder. By means of the devices, the secondary flow disappears in the inner cylinder, and the turbulent flow becomes homogeneous in the stationary tank. Then, flow visualization experiments were conducted on the oscillating grid turbulence in the rotating tank. The flow structure is almost homogeneous in the vicinity of the grid, and the longitudinal vortex like structures appear on the rotating flow. The number of the longitudinal vortex structures increases and their shape become thin and straight with increasing the system rotation. The dye injection experiment also shows that the dye diffusion is suppressed by the rotation.

This study is to clarify the flow structure of a three-dimensional wall jet discharged tangentially on a flat surface from a circular nozzle. Distributions of the mean velocities and fluctuating velocities in a cross section of a fully-developed flow region were measured by using a single hot-wire probe and cross-wire probes. It is confirmed in any position of the lateral direction that the distributions of the mean velocities U and W in the direction normal to the wall are made similar to each other when they are made dimensionless by the maximum local velocity while the jet height remains almost invariable in the spanwise direction. The direction of the mean flow vector in the inner layer is remained constant while the direction of the mean flow vectors outside the half-value height is coincided with the direction of the jet axis. The distributions of the turbulent statistics in the direction normal to the wall are almost similar to each other outside the half-value height due to the scaling by the maximum local velocity, and on the other hand, the magnitude of the turbulent energy and Reynolds stress in the wall-side region becomes larger at the position farther from the jet axis.

The vortex generator jets can achieve the adaptive control by adjusting properly the jet speed in response to flow situations. The vortex generator jet method may accomplish separation control only when it is necessary, and therefore it is useful for both design and off-design conditions. In this study, the effect of vortex generator jets on separation control is investigated for three types of rectangular orifices with aspect ratios AR=0.16, 1 (square) and 6.4. The rectangular orifice with AR=0.16 set the long side of the orifice in the streamwise direction and the long side turned to the spanwise direction for AR=6.4. For the rectangular orifices with AR=6.4, the vortices become stronger and exist at lower vertical positions in comparison with those for the other orifice types. Therefore, the rectangular orifices with AR=6.4 promote the boundary layer mixing and make effective the separation control.