0


EDITORIAL

J. Fluids Eng. 1991;113(1):1-2. doi:10.1115/1.2926491.
FREE TO VIEW
Abstract
Commentary by Dr. Valentin Fuster

REPORTS

J. Fluids Eng. 1991;113(1):3-4. doi:10.1115/1.2926494.
FREE TO VIEW
Abstract
Topics: Turbulence , Testing
Commentary by Dr. Valentin Fuster

RESEARCH PAPERS

J. Fluids Eng. 1991;113(1):5-30. doi:10.1115/1.2926497.

Under certain conditions, wall-bounded flows separate. To improve the performance of natural or man-made flow systems, it may be beneficial to delay or advance this detachment process. The present article reviews the status and outlook of separation control for both steady and unsteady flows. Both passive and active techniques to prevent or to provoke flow detachment are considered and suggestions are made for further research.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):31-36. doi:10.1115/1.2926492.

Numerical simulations of the large amplitude stage of the Kelvin-Helmholtz instability of a relatively thin vorticity layer are discussed. At high Reynolds number, the effect of viscosity is commonly neglected and the thin layer is modeled as a vortex sheet separating one potential flow region from another. Since such vortex sheets are susceptible to a short wavelength instability, as well as singularity formation, it is necessary to provide an artificial “regularization” for long time calculations. We examine the effect of this regularization by comparing vortex sheet calculations with fully viscous finite difference calculations of the Navier-Stokes equations. In particular, we compare the limiting behavior of the viscous simulations for high Reynolds numbers and small initial layer thickness with the limiting solution for the roll-up of an inviscid vortex sheet. Results show that the inviscid regularization effectively reproduces many of the features associated with the thickness of viscous vorticity layers with increasing Reynolds number, though the simplified dynamics of the inviscid model allows it to accurately simulate only the large scale features of the vorticity field. Our results also show that the limiting solution of zero regularization for the inviscid model and high Reynolds number and zero initial thickness for the viscous simulations appear to be the same.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):37-44. doi:10.1115/1.2926493.

Details of the horseshoe vortex formation around a cylinder were studied to determine the flow parameters that affect the flow separation in front of the cylinder. An experimental setup consisting of a circular cylinder vertically mounted on the floor of the wind tunnel test section was assembled. The approaching turbulent boundary layer was four centimeters thick. Pressures were measured on the cylinder surface and the tunnel floor with surface static pressure taps. Surface flow visualizations were accomplished to locate singlar points and the size of separation region on the endwall surface. Interior mean and fluctuating velocity data and Reynolds stresses in front of the cylinder were nonintrusively measured with a two-component Laser Doppler Anemometer system. Freestream stagnation at the endwall/cylinder surface occurred in all cases, but two types of separation were identified in this investigation. The flow pattern in the separation region depends on the freestream momentum and the boundary layer displacement thickness. A large-scale fully developed vortex was formed in the plane of symmetry for low approaching freestream velocities. A fully developed vortex was not present at higher approach velocities. Maximum production of turbulent kinetic energy was measured around the core of the vortex when fully formed.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):45-50. doi:10.1115/1.2926495.

The turbulent fluid motion established in the wake of two long, smooth circular cylinders arranged perpendicular to each other has been investigated in a steady, low-turbulence, uniform flow at Reynolds numbers of 2×104 and 2×103 (based on cylinder diameter and freestream velocity). A complex three-dimensional regime was found at the center of the configuration, the precise nature of which is dependent upon the spacing of the cylinders. If the distance between the axis of each cylinder is less than three diameters, the fluid motion in the central near wake is dominated by secondary flows associated with trailing vortices and horseshoe vortices, whereas at spacings beyond this critical value there is a considerable reduction in the influence of secondary flow. The paper examines these spacing related regimes in detail and considers the extent of the associated interference effects.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):51-59. doi:10.1115/1.2926496.

Measurements are reported for the separted reattaching flow around a long rectangular plate placed at zero incidence in a low-turbulence stream. This laboratory configuration, chosen for its geometric simplicity, exhibits all of the important features of two-dimensional flow separation with reattachment. Conventional hot-wire anemometry, pulsed-wire anemometry and pulsed-wire surface shear stress probes were used to measure the mean and fluctuating flow field at a Reynolds number, based on plate thickness, of 5 × 104 . The separated shear layer appears to behave like a conventional mixing layer over the first half of the separation bubble, where it exhibits an approximately constant growth rate and a linear variation of characteristic frequencies and integral timescales. The characteristics of the shear layer in the second half of the bubble are radically altered by the unsteady reattachment process. Much higher turbulent intensities and lower growth rates are encountered there, and, in agreement with other reattaching flow studies, a low frequency motion can be detected.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):60-67. doi:10.1115/1.2926498.

This paper presents calculations of the time-averaged separated-reattaching flow around a bluff rectangular plate, using a finite difference procedure and the k-ε turbulence model. Two discretization methods are used: the hybrid differencing scheme, and the bounded skew hybrid differencing scheme. The latter, although superior to the former for all grid distributions, results in a reattachment length about 30 percent shorter than the measured value. When a modification which takes into account streamline curvature is incorporated into the k-ε model, a marked improvement in the predictions is obtained. A reattachment length of 4.3 plate thicknesses (D), compared to an experimental value of 4.7D, is obtained, and the predicted mean flow field, turbulent kinetic energy and pressure distributions within the separation bubble are found to be in good agreement with experiments.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):68-72. doi:10.1115/1.2926499.

The research presented in this paper is an investigation of the interaction between twin side-by-side circular jets issuing into a crossflow, utilizing a real-time video digitization system to obtain quantitative data from the visualized flow fields. The intermittency field results obtained from experiments with different jet spacings and jet velocity/crossflow velocity ratios are reported.

Topics: Jets , Flow (Dynamics)
Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):73-80. doi:10.1115/1.2926500.

The mean thickness of a thin liquid film of deionized water with a free surface on a stationary and rotating horizontal disk has been measured with a nonobtrusive capacitance technique. The measurements were taken when the rotational speed ranged from 0–300 rpm and the flow rate varied from 7.0–15.0 lpm. A flow visualization study of the thin film was also performed to determine the characteristics of the waves on the free surface. When the disk was stationary, a circular hydraulic jump was present on the disk. Upstream from the jump, the film thickness was determined by the inertial and frictional forces on the fluid, and the radial spreading of the film. The surface tension at the edge of the disk affected the film thickness downstream from the jump. For the rotating disk, the film thickness was dependent upon the inertial and frictional forces near the center of the disk and the centrifugal forces near the edge of the disk.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):81-85. doi:10.1115/1.2926501.

The results of an experimental investigation concerning the operating characteristics of a butterfly valve downstream of a mitered elbow are reported. Primary emphasis is given the influences of valve disk angle, valve/elbow spacing, and valve/elbow orientation on the dimensionless pressure drop, mass flow coefficient, and aerodynamic torque coefficient characteristics of the valve. The results show that when the valve is located two pipe diameters downstream of the elbow, the performance characteristics are substantially affected by the relative valve/elbow orientation. However, at a spacing of eight diameters the effect of the elbow on the valve operating characteristics is small.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):86-89. doi:10.1115/1.2926502.

A valve sizing method is developed to permit the correct selection of the size of small flow control valve trim handling minute flowing quantities. Most valves such as these operate primarily in laminar or transitional flow regimes, with either gaseous or liquid fluids, due to the geometries involved. The proposed sizing method can be combined with existing turbulent flow equations by the establishment of a valve style modifier Fd which describes the correct relationship between the hydraulic diameter and the equivalent single orifice diameter of a specific valve trim, thereby permitting calculation of the correct Reynolds number for a given flow condition. Finally, a correction factor FR is used which permits necessary flow corrections when used in connection with the correct valve Reynolds number. This method appears to be reasonably accurate and validated to the extent that the experimental data permit.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):90-97. doi:10.1115/1.2926503.

Experiments are carried out to determine the effects of different inlet geometries on the onset of suction recirculation and its associated power consumption in an axial-flow pump. The critical flow rate is determined by both the “string” visual technique and “pressure” method. The results are correlated with the inlet area and flow velocity distribution upstream of the impeller. Four different conical covers matching the impeller leading edge are employed to cover the impeller inlet completely or partially. Covering the inlet area reduces the critical flowrate corresponding to the onset of suction recirculation and eliminates all recirculation at higher flowrates. The power consumption associated with the suction recirculation flow for the uncovered impeller is determined by comparing the shaft powers with and without inlet covers. At the shut-off condition, the power is estimated from a comparison with the shaft power measured with the impeller inlet completely covered. Experimental studies conclude that the power consumption due to suction recirculation is mainly controlled by the impeller inlet area and is insensitive to the inlet pipe configuration. At shut-off condition, the power coefficient correlates well with the parameter based on the hydraulic radius of inlet area. At a finite through flowrate the analytical model recommended by Tuzson (1983) is adequate, except for a proportionality coefficient determined from the test data.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):98-103. doi:10.1115/1.2926504.

A previously described experimentally based numerical procedure is used to characterize the flow in thin walled polyethylene tubes with uniformly spaced wall orifices (polytubes). The procedure evaluates the required supply pressure for specified geometry, supply flow, and supply swirl angle; and then predicts the pressure and discharge profiles along the polytube. The predicted values are in good agreement with reported experimental results and are used to develop a data base of the most significant ranges of polytube and supply parameters. Then, explicit design correlations and recommendations are obtained for satisfactory and economical performance.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):104-110. doi:10.1115/1.2926479.

A finite-element model of the turbulent flow field in the annular exhaust diffuser of a gas turbine engine is developed. The analysis is based on a modified version of the Petrov-Galerkin weighted residual method, coupled with a highly accurate biquadratic finite element of the Lagrangian type. The elemental weight functions in the finite-element formulation are so defined to ensure upwinding of the convection terms in the flow-governing equations while reverting to the conventional Galerkin’s definition for all other terms. This approach is equivalent to altering the integration algorithm as the convection terms in the element equations are derived, with the exception that the latter technique is tailored for low-order elements of the linear and bilinear types. Numerical results of the current analysis indicate that spurious pressure modes associated with this type of inertia-dominated flow are alleviated while the false numerical diffusion in the finite-element equations is simultaneously minimized. Turbulence of the flow field is modeled using the two-layer algebraic turbulence closure of Baldwin and Lomax, and the eddy viscosity calculations are performed at variably spaced points which are different from those in the finite-element discretization model. This enhances the accuracy in computing the wall shear stress and the inner/outer layer interface location. The computational model is verified using a set of experimental data at design and off-design operation modes of the exhaust diffuser in a commercial gas turbine engine. Assessment of the results in this case is favorable and, as such, provides evidence of the model capability as an accurate predictive tool in the diffuser detailed design phase.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):111-115. doi:10.1115/1.2926480.

Wall pressure distribution and shear stress fields for pulsatile laminar flow in a 90-degree bifurcation with rectangular cross sections are evaluated using the results of the numerical solution of the Navier-Stokes equation. The extent of the adverse pressure gradient on the bottom wall of the main duct and the upstream wall of the branch closely correlate to the behavior of the two dynamic recirculation zones which are formed on these two walls. Multiple zones of high and low shear stresses at various sites in the bifurcation are observed. The extent of the fluctuations of the maximum and minimum shear stress is identified. Next-to-the-wall laser Doppler anemometer velocity measurements are used to estimate the shear stress distribution on the walls. In general, qualitative agreement between the experimental and computed wall shear stress values is observed. The variation of the wall shear stress in the vicinity of the branch is discussed in light of the highly perturbed flow field.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):116-123. doi:10.1115/1.2926481.

This paper concerns an experimental evaluation of the basic aspects of operation of the advanced solar pond (ASP). Experiments were carried out in a laboratory test section in order to assess the feasibility of the density gradient maintenance in stratified flowing layers. The density stratification was caused by a non uniform distribution of temperatures in the flow field. Results of the experiments are reported and analyzed in the paper. Experimental data were used in order to calibrate the numerical model able to simulate heat and momentum transfer in the ASP. The numerical results confirmed the validity of the numerical model adopted, and proved the latter applicability for the simulation of the ASP performance.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):125-129. doi:10.1115/1.2926484.

The single phase and cavitating tip vortex shed by a NACA 66-209 rectangular plan-form, rounded tip hydrofoil has been studied. Single-phase measurements of instantaneous flow velocity were made by taking double-pulsed holograms of microbubbles moving in and around the vortex core. The tailored air bubble technique of Ooi and Acosta (1983) was employed to measure both the mean and fluctuating single phase vortex core static pressure. Cavitation inception was determined visually. The flow in the vortex core is highly unsteady; the r.m.s. axial velocity fluctuation can be as high as 0.2U∞ . Core pressure fluctuations greater than the freestream dynamic pressure have been observed. These fluctuations are commensurate with the axial velocity unsteadiness. Cavitation inception in saturated water occurs at mean core pressures above the vapor pressure. However, the pressure fluctuations are certainly substantial enough to account for these elevated inception indices. The rapid establishment of fully developed trailing vortex cavitation when the cavitation number is reduced below inception is attributable to the small variation in mean core pressure with downstream distance. The inception index increases substantially with dissolved air content. The cause of this dependence is not presently known.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):130-136. doi:10.1115/1.2926485.

A “two-fluid model” has been incorporated with Lumley’s drag reduction model to analyze the mechanism of momentum transfer in the turbulent dilute gas-particle flow in a vertical pipe. The change of the effective viscous sublayer thickness by the presence of particles is modeled by Lumley’s theoretical model. The numerical computations of the friction factor and the pressure drop in a fully developed pipe flow are in good agreement with the corresponding experimental data for an average particle size of 15 μm. It is proved that Lumley’s model is successful in predicting the correct reduction behavior of the drag in the gas-particle flows. It has been confirmed that the effective viscous sublayer thickness for two-phase gas-particle flow is dependent on the particle relaxation time, Kolmogoroff time scale and the solids-gas loading ratio.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):137-142. doi:10.1115/1.2926486.

An analysis has been performed to study conjugate Couette flows in a channel over a packed-spheres bed of semi-infinite extent. The Brinkman model is used as the momentum equation in the porous bed with the variation of porosity approximated by an exponential function. An approximate expression for the permeability of the bed, applicable for a wide range of porosity, is constructed. A singular perturbation solution is obtained for the velocity distribution in the packed-spheres bed. The streamwise velocity and the shearing stress of the conjugate flows in the channel and in the packed-spheres bed are matched at the fluid/porous bed interface. An explicit analytical expression for the slip coefficient is obtained by comparing the results based on the present formulation with those based on Beavers and Joseph’s formulation. It is shown that the value of the slip coefficient decreases as a result of the variable porosity effect.

Commentary by Dr. Valentin Fuster

DISCUSSIONS

TECHNICAL BRIEFS

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):147-149. doi:10.1115/1.2926488.

An analysis of procedures in current use for prediction of transition onset location shows that they are generally in poor agreement with data obtained in test facilities at low freestream turbulence levels. It has been shown elsewhere that under such conditions transition is driven by residual nonturbulent disturbances in the facility. A method is developed for taking such disturbances into account by defining an equivalent free-stream turbulence intensity; values for this parameter are derived for each facility from which onset data are available. A new correlation incorporating this effect is shown to be in good agreement with all available data on two-dimensional flows with pressure gradient. The correlation suggests that the onset Reynolds number (based on boundary-layer thickness) depends inversely on the total disturbance level when the latter is low.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):149-151. doi:10.1115/1.2926489.

A new 90-deg bend shape has been developed for erosion resistance. The design consists of a gradually expanding inlet, a large-diameter curved section, and a gradually contracting outlet. Tests were conducted using alumina particles in air at an average velocity of 27 m/s, passing through a carbon steel bend. The new bend design had an erosion life exceeding 4000 hr, compared to 38 hr for a conventional bend design.

Commentary by Dr. Valentin Fuster
J. Fluids Eng. 1991;113(1):151-154. doi:10.1115/1.2926490.

This paper deals with the incompressible flow through a porous medium bounded by two parallel permeable plates, one of which is stationary and the other slides and oscillates in its own plane. Also uniform crossflow is superimposed through the plates. Analytical solutions for flow parameters are obtained. The effect of crossflow, permeability, and the frequency of the moving plate on the velocity, skin friction has been investigated. It is shown that skin friction can be reduced considerably by injection at the lower plate.

Commentary by Dr. Valentin Fuster

Sorry! You do not have access to this content. For assistance or to subscribe, please contact us:

  • TELEPHONE: 1-800-843-2763 (Toll-free in the USA)
  • EMAIL: asmedigitalcollection@asme.org
Sign In