Turbulent flow of quadrangle mode in interdisk midplane between two shrouded co-rotating disks

Particle image velocimetry (PIV) was employed to study the flow patterns, time-averaged velocity field, and turbulence properties of the flow in the interdisk midplane between two shrouded co-rotating disks at the interdisk spacing to disk radius ratio S = 0.1 and rotating Reynolds number Re = 2.25 × 10⁵. A quadrangle core flow structure rotating at a frequency 75% of the disks’ rotating frequency was observed. The flow in the region outside the quadrangle core flow structure consisted of four cellular flow structures. Five characteristic flow regions—the hub-influenced region, solid-body rotation region, buffer region, vortex region, and shroud-influenced region—were identified in the flow field. Circumferential and radial turbulence intensities, Reynolds stresses, turbulence kinetic energy, correlation coefficients, as well as the Lagrangian integral time and length scales of turbulent fluctuations were analyzed and presented. Features of the turbulence properties were found to be closely related to the rotation motion of the inner and outer characteristic flow structures. The circumferential components of the turbulence properties exhibited local minima in the buffer region and maxima in the solid-body rotation and vortex regions, while the radial components of the turbulence intensity, turbulent normal stress, and Lagrangian integral turbulence time scale exhibited maximum values in the buffer region and relatively low values in the regions near the hub and the shroud.     

Two- and three-dimensional laminar flows between disks co-rotating in a fixed cylindrical enclosure

A numerical investigation is performed for the constant property laminar flow of air in the space between a pair of disks clamped co-axially on a central hub and co-rotating in a stationary cylindrical enclosure. Both two- and three-dimensional flow conditions are examined in relation to the interdisk spacing, H, and the disk angular velocity, Ω. Two interdisk spacings are considered, corresponding to aspect ratios Γ = 0.186 and 0.279 (with Γ = H/(R₂+a−R), where R₂ is the disk radius, a is the disk rim–enclosure wall clearance, and R is the hub radius). A range of rotational speeds encompassing the transition from axisymmetric two-dimensional steady flow to non-axisymmetric three-dimensional unsteady flow are considered for various values of the Reynolds number, Re (with $ Re=\Omega R_2^2/v $, where v is the kinematic viscosity of air). Axisymmetric calculations are first performed for both aspect ratios in the range 3858≤Re≤23 150. Fully three-dimensional calculations are then performed for the configuration with Γ = 0.186 and Re = 23 150, and for the configuration with Γ = 0.279 and Re = 7715, 15 430 and 23 150. The axisymmetric calculations performed with Γ = 0.186 confirm many known features of the flow, including the transition from a steady flow to an oscillatory periodic regime. This occurs at ≈Re = 23 150 for a configuration with a/H = 0, and at ≈Re = 14 670 for one with a/H = 0.28 and a finite disk thickness (b/H = 0.2). Three-dimensional calculations performed for Γ = 0.186 with a/H = 0 and Re = 23 150 reveal a circumferentially periodic flow pattern with eight foci of intensified axial component of vorticity. The axisymmetric calculations performed with Γ = 0.279 and Re > 7715 yield a novel, non-unique steady solution for the velocity field that is asymmetric with respect to the interdisk mid-plane. No experimental verification of this finding exists to date, but similar situations are known to arise in the context of anomalous modes of the Taylor–Couette flow. Relaxing the axisymmetry constraint allows this flow to evolve to an oscillatory three-dimensional regime of increasing irregularity with increasing rotational speed. In this case, the number of foci of intensified axial vorticity varies with time, ranging from six at Re = 7715 to between six and eight at Re = 23 150. © 1998 John Wiley & Sons, Ltd.     

Numerical simulation of the flow field in the vicinity of an axial flow fan

The main purpose of the current investigation is the development and evaluation of a numerical model used to simulate the effect of an axial flow fan on the velocity field in the vicinity of the fan blades. The axial flow fan is modeled as an actuator disc, where the actuator disc forces are calculated using blade element theory. The calculated disc forces are expressed as sources/sinks of momentum in the Navier–Stokes equations solved by a commercially available computational fluid dynamic (CFD) code, Flo⁺⁺. The model is used to determine the fan performance characteristics of an axial flow fan as well as the velocity fields directly up‐ and downstream of the fan blades. The results are compared with experimental data. In general, good agreement is obtained between the numerical results and experimental data, although the fan power consumption, as well as radial velocity downstream of the fan blades, is underpredicted by the fan model. Copyright © 2001 John Wiley & Sons, Ltd.     

Bifurcation analysis of fan casing under rotating air flow excitation

A fan casing model of cantilever circular thin shell is constructed based on the geometric characteristics of the thin-walled structure of aero-engine fan casing. According to Donnelly＇s shell theory and Hamilton＇s principle, the dynamic equations axe established. The dynamic behaviors are investigated by a multiple-scale method. The effects of casing geometric parameters and motion parameters on the natural frequency of the system are studied. The transition sets and bifurcation diagrams of the system are obtained through a singularity analysis of the bifurcation equation, showing that various modes of the system such as the bifurcation and hysteresis will appear in different parameter regions. In accordance with the multiple relationship of the fan speed and stator vibration frequency, the fan speed interval with the casing vibration sudden jump is calculated. The dynamic reasons of casing cracks are investigated. The possibility of casing cracking hysteresis interval is analyzed. The results show that cracking is more likely to appear in the hysteresis interval. The research of this paper provides a theoretical basis for fan casing design and system parameter optimization.     

Internal flow numerical simulation of a cross‐flow fan in refrigerant operating condition using user‐defined functions

In the paper, a cross‐flow fan in refrigerant operating condition is systematically simulated using user‐defined functions. Three‐dimensional simulations are acquired with Navier–Stokes equations coupled with k–ε turbulence model, and internal flow characteristics of an indoor split‐type air conditioner are obtained, which is mainly composed of cross‐flow fan and heat exchanger. It has systematically been simulated in the isothermal flow condition that the performance of cross‐flow fan may be reduced easily with dry or humid air, and in the refrigerant operating condition in which user‐defined functions are applied to the humid air, considered as a mixture of dry air and vapor. A density‐modulated function is adopted to deal with the condensation of the vapor at the heat‐transfer region approximately. The results show flow mechanism of the two gas‐phase flow, including phase‐vary process. The distribution of the parameters is not uniform at the inlet of the machine, the intensity and position of pressure and velocity vary along the axial direction of the fan, the distribution of vapor volume fraction and turbulent intensity in heat‐transfer region is obtained, and the external characteristic data of the indoor machine are obtained and analyzed. Compared with the experimental data, the calculated characteristic curves and designed parameters are on target. © British Crown Copyright 2010/MOD. Reproduced with permission. Published by John Wiley & Sons, Ltd.     

Is the steady viscous incompressible two-dimensional flow over a backward-facing step at Re = 800 stable?

A detailed case study is made of one particular solution of the 2D incompressible Navier–Stokes equations. Careful mesh refinement studies were made using four different methods (and computer codes): (1) a high-order finite-element method solving the unsteady equations by time-marching; (2) a high-order finite-element method solving both the steady equations and the associated linear-stability problem; (3) a second-order finite difference method solving the unsteady equations in streamfunction form by time-marching; and (4) a spectral-element method solving the unsteady equations by time-marching. The unanimous conclusion is that the correct solution for flow over the backward-facing step at Re = 800 is steady—and it is stable, to both small and large perturbations.     

Application of the two-equation k-ϵ turbulence model to a two-dimensional,steady, free surface flow problem with separation

An application of the depth-integrated k-? turbulence model is presented for separated flow in a wide, shallow, rectangular channel with an abrupt expansion in width. The well-known numerical problems associated with the use of upwind and central finite differences for convection are overcome by the adoption of the spatially third-order accurate QUICK finite difference technique. Results show that modification of the depth-integrated k-? turbulence closure model for streamline curvature leads to significant improvement in the agreement between model predictions and experimental measurements.     

Numerical modelling and flow analysis of a centrifugal pump running as a turbine: Unsteady flow structures and its effects on the global performance

The purpose of this paper is to investigate the flow patterns in a centrifugal pump when it works as a centripetal turbine, with special interest in the unsteady behavior in order to explain the shape of the performance curves. Also, we focus on the determination of the radial thrust and other mechanical loads over a pump‐designed machine. The pump studied is commercial, with single axial suction and a vaneless spiral volute casing. A numerical study has been carried out in order to obtain more information about the flow into the volute and the impeller. A numerical three‐dimensional unsteady simulation has been developed using a commercial code that solves the URANS set of equations with a standard k–ε turbulence model. The results show the non‐axisymmetric flow developed in the volute, responsible for a significant radial thrust; the interaction between the tongue and the impeller, generating force fluctuations; the velocity and pressure distributions inside the impeller; and the exit flow, characterized with post‐rotation and low‐pressure. These flow results allow us to understand the behavior of the machine by comparing it with the pump mode. Complementarily, an experimental study was conducted to validate the numerical model and characterize the pump‐turbine performance curves at constant head. Fast‐response pressure taps and a three‐hole pneumatic pressure probe were employed to obtain a complete data set of non‐stationary and stationary measurements throughout the centrifugal machine. As a result, loss of efficiency or susceptibility to cavitation, detected numerically, was confirmed experimentally. The study demonstrates that the numerical methodology presented here has shown its reliability and possibilities to predict the unsteady flow and time‐mean characteristics of centrifugal pumps working as turbines. In particular, it is shown that the commercial design of the pump allows a reasonable use of the impeller as a turbine runner, due to the suitable adaptation of the inflow distributions to the volute casing. Moreover, the efficiency for the inverse mode is shown to be as high as achieved for the pumping operational mode. In addition, it is concluded that both axial and radial thrusts are controlled, though important unsteady fluctuations—up to 25%—clocked with the blade passing frequency appear beyond the nominal conditions. In that case, a moderate use of the pump as a turbine is recommended in order to minimize risks of fatigue failure of the bearings. Copyright © 2009 John Wiley & Sons, Ltd.     

Discrete and continuum modeling of granular flow in silo discharge

Granular material discharge from a flat-bottomed silo has been simulated by using continuum modeling and a three-dimensional discrete-element method (DEM). The predictive abilities of three commonly used frictional viscosity models (Schaeffer, S–S, and μ(I)) were evaluated by comparing them with the DEM data. The funnel-flow pattern (type C) and the semi-mass-flow pattern (type B) that was predicted by DEM simulations can be represented when the Schaeffer or μ(I) model is used, whereas the S–S model gives a consistent type-B flow pattern. All three models over-estimate the discharge rate compared with the DEM. The profiles of the solids volume fraction and the vertical velocity above the outlet show that the larger discharge rates given by the Schaeffer and μ(I) model result from an over-estimation of volume fraction, whereas the deviation in the S–S model stems from the failure to predict a solid vertical velocity and a volume fraction.     

Computational modeling of biomagnetic micropolar blood flow and heat transfer in a two-dimensional non-Darcian porous medium

We study theoretically and computationally the incompressible, non-conducting, micropolar, biomagnetic (blood) flow and heat transfer through a two-dimensional square porous medium in an (x,y) coordinate system, bound by impermeable walls. The magnetic field acting on the fluid is generated by an electrical current flowing normal to the x–y plane, at a distance l beneath the base side of the square. The flow regime is affected by the magnetization B ₀ and a linear relation is used to define the relationship between magnetization and magnetic field intensity. The steady governing equations for x-direction translational (linear) momentum, y-direction translational (linear) momentum, angular momentum (micro-rotation) and energy (heat) conservation are presented. The energy equation incorporates a special term designating the thermal power per unit volume due to the magnetocaloric effect. The governing equations are non-dimensionalized into a dimensionless (ξ,η) coordinate system using a set of similarity transformations. The resulting two point boundary value problem is shown to be represented by five dependent non-dimensional variables, f _ξ  (velocity), f _η (velocity), g (micro-rotation), E (magnetic field intensity) and θ (temperature) with appropriate boundary conditions at the walls. The thermophysical parameters controlling the flow are the micropolar parameter (R), biomagnetic parameter (N _H ), Darcy number (Da), Forchheimer (Fs), magnetic field strength parameter (Mn), Eckert number (Ec) and Prandtl number (Pr). Numerical solutions are obtained using the finite element method and also the finite difference method for Ec=2.476×10⁻⁶ and Prandtl number Pr=20, which represent realistic biomagnetic hemodynamic and heat transfer scenarios. Temperatures are shown to be considerably increased with Mn values but depressed by a rise in biomagnetic parameter (N _H ) and also a rise in micropolarity (R). Translational velocity components are found to decrease substantially with micropolarity (R), a trend consistent with Newtonian blood flows. Micro-rotation values are shown to increase considerably with a rise in R values but are reduced with a rise in biomagnetic parameter (N _H ). Both translational velocities are boosted with a rise in Darcy number as is micro-rotation. Forchheimer number is also shown to decrease translational velocities but increase micro-rotation. Excellent agreement is demonstrated between both numerical solutions. The mathematical model finds applications in blood flow control devices, hemodynamics in porous biomaterials and also biomagnetic flows in highly perfused skeletal tissue. Dedicated to Professor Y.C. Fung (1919-), Emeritus Professor of Biomechanics, Bioengineering Department, University of California at San Diego, USA for his seminal contributions to biomechanics and physiological fluid mechanics over four decades and his excellent encouragement to the authors, in particular OAB, with computational biofluid dynamics research.     

Design of the centrifugal fan of a belt-driven starter generator with reduced flow noise

Large eddy simulations based on the explicit algebraic subgrid-scale stress model were carried out to predict the flow-induced noise generated on the centrifugal fan of a belt-driven starter generator using Lighthill's analogy and the method of Ffowcs Williams and Hawkings. The surrounding air was approximated by an ideal gas at fixed room temperature (T_in = 300 K), and the rotating velocity of the fan was considered to be 6000 rpm. The blade array angles were designed using the modulation method, and a large blade curvature was adopted. We identified several centrifugal fan design parameters that could minimize the flow-induced noise while also minimizing fan efficiency losses. Three design parameters: the top serrated edge (θ_t), the step leading edge (0.52 H_b) and the tail edge (d_b and r_b), played a critical role in preventing vortex generation and collision, significantly weakening the surface pressure fluctuations on the blade. The maximum sound pressure level at 800 Hz at a specific location was reduced by 5.5 dB (at the top serrated edge) and 6.8 dB (at the step leading edge) relative to the baseline case. The sound power, calculated over a hemisphere surface of 950 mm, was reduced by 77.3% (at the top serrated edge) and 61.0% (at the step leading edge) relative to the baseline whereas the mass flow rates were reduced by 5.2% and 10.6%, respectively. Experiments were performed using the optimally designed fan in a semi-anechoic chamber. The predicted sound pressure level and frequency were in good agreement with the experimentally measured values.     

Direct modeling of flow of FENE fluids

Direct simulations of macromolecular fluids are carried out for flows between parallel plates and in expanding and contracting channels. The macromolecules are modeled as FENE dumbbells with soft disks or Lennard-Jones dumbbell-dumbbell interactions. The results are presented in terms of profiles and contour plots of velocity, pressure, temperature, density, and flow fields. In addition the data for potential energy, shear stress, and the normal components of the stress tensor are collected. In general, an excellent agreement is found between the simulated profiles and the well-known flow structures, such as flow separation and formation of viscous eddies, indicating that micro-hydrodynamics is a viable tool in linking macroscopic phenomena with the underlying physical mechanisms. The simulations are performed in the Newtonian regime, for medium-size systems comprising up to 3888 dumbbells. This number is sufficiently large to control boundary and particle number effects. The flow is induced by gravity. The traditional stochastic (thermal) and periodic boundary conditions are employed. Also, diffusive boundary conditions, which could include a stagnant fluid layer and repulsive potential walls, are developed. The scaling problems, which are related to the application of a large external force in a microscopic system (of the size of the order 100 Å), result in extreme pressure and temperature gradients. In addition, the viscosity and thermal conductivity coefficients obtained from velocity and temperature profiles of the channel flow are presented. These results are confirmed independently from modeling of Couette flow by the SLLOD equations of motion and from the Evans algorithm for thermal conductivity.     

Two-phase flow of R-134a refrigerant during flow boiling through a horizontal circular mini-channel

This research focuses on heat transfer to R-134a during flow boiling in a 1.75 mm internal diameter tube. Flow visualisation and heat transfer experiments are conducted to obtain heat transfer coefficients for different flow patterns. The measured data in each flow regime are compared with predictions from a three-zone flow boiling model. The calculations are in fair agreement with the experimental results which correspond in particular to slug flow, throat-annular flow and churn flow regimes under conditions of low heat flux.     

Lattice Boltzmann modeling of transport phenomena in fuel cells and flow batteries

Fuel cells and flow batteries are promising technologies to address climate change and air pollution problems. An understanding of the complex multiscale and multiphysics transport phenomena occurring in these electrochemical systems requires powerful numerical tools. Over the past decades, the lattice Boltzmann(LB) method has attracted broad interest in the computational fluid dynamics and the numerical heat transfer communities, primarily due to its kinetic nature making it appropriate for modeling complex multiphase transport phenomena. More importantly, the LB method fits well with parallel computing due to its locality feature, which is required for large-scale engineering applications. In this article, we review the LB method for gas–liquid two-phase flows, coupled fluid flow and mass transport in porous media, and particulate flows. Examples of applications are provided in fuel cells and flow batteries. Further developments of the LB method are also outlined.     

Measurements and high-speed visualizations of flow boiling of a dielectric fluid in a silicon microchannel heat sink

Experiments were conducted to investigate flow boiling heat transfer to a dielectric fluid in a silicon chip-integrated microchannel heat sink. Twenty-four microchannels, each 389 μm × 389 μm in cross-section, were fabricated into the 12.7 mm × 12.7 mm silicon substrate. High-speed visualizations (at 12,500 frames per second) were performed simultaneously with heat transfer and pressure drop measurements to investigate the physics of flow boiling in parallel microchannel arrays. At low heat fluxes, bubbly flow is dominant, with the bubbles coalescing to form vapor slugs as the heat flux is increased. At high heat fluxes, the flow regimes in the downstream portion of the microchannels are characteristic of alternating wispy-annular flow and churn flow, while flow reversal is observed in the upstream region near the microchannel inlet. Local heat transfer measurements, obtained at three flow rates ranging from 35 to 60 ml/min, show that at lower heat fluxes, the heat transfer coefficient increases with increasing heat flux. The heat transfer coefficient in fully developed boiling is seen to be independent of flow rate in this range. At higher heat fluxes (exceeding 542, 673, 730 kW/m², respectively, for flow rates of 35, 47 and 60 ml/min), this trend is reversed, and the heat transfer coefficient decreases with further increases in heat flux due to partial dryout in some of the microchannels. Heat fluxes at which fully developed boiling is achieved depend on the flow rate. The pressure drop in fully developed boiling increases with increasing heat flux and is independent of flow rate for the test conditions considered in this work.     

Computation of internal turbulent flow with a large separated flow region

An implicit two-equation turbulence solver, KEM. in generalized co-ordinates, is used in conjunction with the three-dimensional incompressible Navier–Stokes solver, INS3D, to calculate the internal flow in a channel and a channel with a sudden 2:3 expansion. A new and consistent boundary procedure for a low Reynolds number form of the κ-ε turbulence model is chosen to integrate the equations up to the wall. The high Reynolds number form of the equations is integrated using wall functions. The latter approach yields a faster convergence to the steady-state solution than the former. For the case of channel flow, both the wall-function and wall-boundary-condition approaches yield results in good agreement with the experimental data. The back-step (sudden expansion) flow is calculated using the wall-function approach. The predictions are in reasonable agreement with the experimental data.     

Numerical modeling of flow past a volumeless and thin rigid body using direct forcing immersed boundary method

The new capability has been added as the numerical method for modeling volumeless and thin rigid bodies to the direct forcing immersed boundary (DFIB) method. The DFIB approach is based on adding a virtual force to the Navier–Stokes equations of incompressible flow to account for the interaction between the fluid and structures. The volume of a solid function (VOS) identifies the stationary or moving solid structures in a given fluid domain. A new VOS-based algorithm was developed to identify thin, rigid structure boundary points in fluid flow and ensure that the fluid cannot cross through the boundary of a thin rigid structure while moving or stationary. The DFIB method was first validated in a three-dimensional (3D) turbulent flow over a circular cylinder. The large-eddy simulation simulated the turbulent flow scales. The proposed algorithm was tested using a 3D turbulent flow past a stationary and rotating Savonius wind turbine that functions as a thin, rigid body. The validation results showed that the selected DFIB approach, combined with the novel algorithm, could simulate a thin, volumeless, rigid structure that is stationary and rotating in incompressible turbulent flows. The current method is also applicable for two-way fluid-structure interaction problems.     

Improvement of a pressure gradient method and its application to an unsteady flow problem

The pressure gradient method using velocity components and components of a pressure gradient as dependent variables has been modified to solve incompressible Newtonian fluid flow problems numerically. Applying this modified method to unsteady-state development of flow in a circular cavity shows that, at least for the case of a low Reynolds number flow, relative errors produced by the proposed method are smaller for most time intervals than those produced by the primitive velocity-pressure variable method and by the standard pressure gradient method. Also it is found that the modified and standard pressure gradient methods can be applied to the unsteady circular cavity flow at a moderate Reynolds number of at least up to 200.     

On flow and stability of a Newtonian fluid past a rotating plane

An exact solution is given for the steady flow of a Newtonian fluid occupying the halfspace past the plane z=0 uniformly rotating about a fixed normal axis (Oz). This solution is obtained in a velocity field of the form considered by Berker [2] and can be deduced as a limiting case, as h+, of the solution to the problem relative to the strip 0zh imposing at z=h either the adherence boundary conditions or the free surface conditions. Furthermore, the stability of this flow, subject to periodic disturbances of finite amplitude, is studied using the energy method and the result is compared with those corresponding to stability of flows in the strip 0zh.

---

Sommario In questa nota si mostra che-oltre alla calssica soluzione di von Karman [1] — esiste, per opportuni valori del gradiente di pressione all'infinito, una soluzione esatta per il moto stazionario di un fluido Newtoniano posto nel semispazio limitato dal piano z=0 uniformemente rotante attorno ad un asse ad esso perpendicolare (Oz). Tale soluzione, ottenuta sulla scia del lavoro di Berker [2], si può dedurre anche come limite, per h+, della soluzione del problema relativo alla striscia 0zh quando sul piano z=h si assegnano o le condizioni di aderenza o le condizioni di frontiera libera. Si studia poi la stabilità di tale moto rispetto a perturbazioni spazialmente periodiche di ampiezza finita col metodo dell'energia e si confronta il risultato ottenuto con quelli relativi alla stabilità dei moti nella striscia 0zh.

     

Boundary conformed co-ordinate systems for selected two-dimensional fluid flow problems. Part I: Generation of BFGs

In this paper the generation of general curvilinear co-ordinate systems for use in selected two-dimensional fluid flow problems is presented. The curvilinear co-ordinate systems are obtained from the numerical solution of a system of Poisson equations. The computational grids obtained by this technique allow for curved grid lines such that the boundary of the solution domain coincides with a grid line. Hence, these meshes are called boundary fitted grids (BFG). The physical solution area is mapped onto a set of connected rectangles in the transformed (computational) plane which form a composite mesh. All numerical calculations are performed in the transformed plane. Since the computational domain is a rectangle and a uniform grid with mesh spacings Δξ = Δη = 1 (in two-dimensions) is used, the computer programming is substantially facilitated. By means of control functions, which form the r.h.s. of the Poisson equations, the clustering of grid lines or grid points is governed. This allows a very fine resolution at certain specified locations and includes adaptive grid generation. The first two sections outline the general features of BFGs, and in section 3 the general transformation rules along with the necessary concepts of differential geometry are given. In section 4 the transformed grid generation equations are derived and control functions are specified. Expressions for grid adaptation arc also presented. Section 5 briefly discusses the numerical solution of the transformed grid generation equations using sucessive overrelaxation and shows a sample calculation where the FAS (full approximation scheme) multigrid technique was employed. In the companion paper (Part II), the application of the BFG method to selected fluid flow problems is addressed.