Examination of the LBM in simulation of microchannel flow in transitional regime

The gas flows in micro-electro-mechanical systems possess relatively large Knudsen number and usually belong to the slip flow and transitional flow regimes. Recently the lattice Boltzmann method (LBM) was proposed by Nie et al. in Journal of Statistical Physics, vol. 107, pp. 279-289, in 2002 to simulate the microchannel and microcavity flows in the transitional flow regime. The present article intends to test the feasibility of doing so. The results of using the lattice Boltzmann method and the direct simulation Monte Carlo method show good agreement between them for small Kn (Kn = 0.0194), poor agreement for Kn = 0.194, and large deviation for Kn = 0.388 in simulating microchannel flows. This suggests that the present version of the lattice Boltzmann method is not feasible to simulate the transitional channel flow.

Turbulent collision of inertial particles: point-particle based, hybrid simulations and beyond

Point-particle based direct numerical simulation (PPDNS) has been a productive research tool for studying both single-particle and particle-pair statistics of inertial particles suspended in a turbulent carrier flow. Here we focus on its use in addressing particle-pair statistics relevant to the quantification of turbulent collision rate of inertial particles. PPDNS is particularly useful as the interaction of particles with small-scale (dissipative) turbulent motion of the carrier flow is mostly relevant. Furthermore, since the particle size may be much smaller than the Kolmogorov length of the background fluid turbulence, a large number of particles are needed to accumulate meaningful pair statistics. Starting from the relative simple Lagrangian tracking of so-called ghost particles, PPDNS has significantly advanced our theoretical understanding of the kinematic formulation of the turbulent geometric collision kernel by providing essential data on dynamic collision kernel, radial relative velocity, and radial distribution function. A recent extension of PPDNS is a hybrid direct numerical simulation (HDNS) approach in which the effect of local hydrodynamic interactions of particles is considered, allowing quantitative assessment of the enhancement of collision efficiency by fluid turbulence. Limitations and open issues in PPDNS and HDNS are discussed. Finally, on-going studies of turbulent collision of inertial particles using large-eddy simulations and particle- resolved simulations are briefly discussed.

Space-time correlations of fluctuating velocities in turbulent shear flows

Space-time correlations or Eulerian two-point two-time correlations of fluctuating velocities are analytically and numerically investigated in turbulent shear flows. An elliptic model for the space-time correlations in the inertial range is developed from the similarity assumptions on the isocorrelation contours: they share a uniform preference direction and a constant aspect ratio. The similarity assumptions are justified using the Kolmogorov similarity hypotheses and verified using the direct numerical simulation DNS of turbulent channel flows. The model relates the space-time correlations to the space correlations via the convection and sweeping characteristic velocities. The analytical expressions for the convection and sweeping velocities are derived from the Navier-Stokes equations for homogeneous turbulent shear flows, where the convection velocity is represented by the mean velocity and the sweeping velocity is the sum of the random sweeping velocity and the shearinduced velocity. This suggests that unlike Taylor’s model where the convection velocity is dominating and Kraichnan and Tennekes’ model where the random sweeping velocity is dominating, the decorrelation time scales of the space-time correlations in turbulent shear flows are determined by the convection velocity, the random sweeping velocity, and the shear-induced velocity. This model predicts a universal form of the spacetime correlations with the two characteristic velocities. The DNS of turbulent channel flows supports the prediction: the correlation functions exhibit a fair good collapse, when plotted against the normalized space and time separations defined by the elliptic model.

Longitudinal dispersion in laboratory and natural streams

This study concerns the longitudinal dispersion of fluid particles which are initially distributed uninformly over one cross section of a uniform, steady, turbulent open channel flow. The primary focus is on developing a method to predict the rate of dispersion in a natural stream.

Taylor's method of determining a dispersion coefficient, previously applied to flow in pipes and two-dimensional open channels, is extended to a class of three-dimensional flows which have large width-to-depth ratios, and in which the velocity varies continuously with lateral cross-sectional position. Most natural streams are included. The dispersion coefficient for a natural stream may be predicted from measurements of the channel cross-sectional geometry, the cross-sectional distribution of velocity, and the overall channel shear velocity. Tracer experiments are not required.

Large values of the dimensionless dispersion coefficient D/rU* are explained by lateral variations in downstream velocity. In effect, the characteristic length of the cross section is shown to be proportional to the width, rather than the hydraulic radius. The dimensionless dispersion coefficient depends approximately on the square of the width to depth ratio.

A numerical program is given which is capable of generating the entire dispersion pattern downstream from an instantaneous point or plane source of pollutant. The program is verified by the theory for two-dimensional flow, and gives results in good agreement with laboratory and field experiments.

Both laboratory and field experiments are described. Twenty-one laboratory experiments were conducted: thirteen in two-dimensional flows, over both smooth and roughened bottoms; and eight in three-dimensional flows, formed by adding extreme side roughness to produce lateral velocity variations. Four field experiments were conducted in the Green-Duwamish River, Washington.

Both laboratory and flume experiments prove that in three-dimensional flow the dominant mechanism for dispersion is lateral velocity variation. For instance, in one laboratory experiment the dimensionless dispersion coefficient D/rU* (where r is the hydraulic radius and U* the shear velocity) was increased by a factory of ten by roughening the channel banks. In three-dimensional laboratory flow, D/rU* varied from 190 to 640, a typical range for natural streams. For each experiment, the measured dispersion coefficient agreed with that predicted by the extension of Taylor's analysis within a maximum error of 15%. For the Green-Duwamish River, the average experimentally measured dispersion coefficient was within 5% of the prediction.

Development of roll waves in open channels

This study is concerned with some of the properties of roll waves that develop naturally from a turbulent uniform flow in a wide rectangular channel on a constant steep slope . The wave properties considered were depth at the wave crest, depth at the wave trough, wave period, and wave velocity . The primary focus was on the mean values and standard deviations of the crest depths and wave periods at a given station and how these quantities varied with distance along the channel.

The wave properties were measured in a laboratory channel in which roll waves developed naturally from a uniform flow . The Froude number F (F = u_n/√gh_n, u_n = normal velocity , h_n = normal depth, g =acceleration of gravity) ranged from 3. 4 to 6. 0 for channel slopes S_o of . 05 and . 12 respectively . In the initial phase of their development the roll waves appeared as small amplitude waves with a continuous water surface profile . These small amplitude waves subsequently developed into large amplitude shock waves. Shock waves were found to overtake and combine with other shock waves with the result that the crest depth of the combined wave was larger than the crest depths before the overtake. Once roll waves began to develop, the mean value of the crest depths h_nmax increased with distance . Once the shock waves began to overtake, the mean wave period T_av increased approximately linearly with distance.

For a given Froude number and channel slope the observed quantities h^-_max/h_n , T' (T' = S_o T_av √g/h_n), and the standard deviations of h^-_max/h_n and T', could be expressed as unique functions of l/h_n (l = distance from beginning of channel) for the two-fold change in h_n occurring in the observed flows . A given value of h^-_max/h_n occurred at smaller values of l/h_n as the Froude number was increased. For a given value of h /hh^-_max/h_n the growth rate of δh^-_max/h^-_maxδl of the shock waves increased as the Froude number was increased.

A laboratory channel was also used to measure the wave properties of periodic permanent roll waves. For a given Froude number and channel slope the h^-_max/h_n vs. T' relation did not agree with a theory in which the weight of the shock front was neglected. After the theory was modified to include this weight, the observed values of h^-_max/h_n were within an average of 6.5 percent of the predicted values, and the maximum discrepancy was 13.5 percent.

For h^-_max/h_n sufficiently large (h^-_max/h_n > approximately 1.5) it was found that the h^-_max/h_n vs. T' relation for natural roll waves was practically identical to the h^-_max/h_n vs. T' relation for periodic permanent roll waves at the same Froude number and slope. As a result of this correspondence between periodic and natural roll waves, the growth rate δh^-_max/h^-_maxδl of shock waves was predicted to depend on the channel slope, and this slope dependence was observed in the experiments.

Mixed nonlinear LES for DES suitable flows

It is shown that a new mixed nonlinear/eddy viscosity LES model reproduces profiles better than a number of competing nonlinear and mixed models for plane channel flow. The objective is an LES method that produces a fully resolved turbulent boundary layer and could be applied to a variety of aerospace problems that are currently studied with RANS, RANS-LES, or DES methods that lack a true turbulent boundary layer. There are two components to the new model. One an eddy viscosity based upon the advected subgrid scale energy and a relatively small coefficient. Second, filtered nonlinear terms based upon the Leray regularization. Coefficients for the eddy viscosity and nonlinear terms come from LES tests in decaying, isotropic turbulence. Using these coefficients, the velocity profile matches measurements data at Reτ ≈ 1000 exactly. Profiles of the components of kinetic energy have the same shape as in the experiment, but the magnitudes differ by about 25%. None of the competing LES gets the shape correct. This method does not require extra operations at the transition between the boundary layer and the interior flow.

Predictive techniques for turbulent oscillatory flows

The prediction of turbulent oscillatory flow at around transitional Reynolds numbers is considered for an idealized electronics system. To assess the accuracy of turbulence models, comparison is made with measurements. A stochastic procedure is used to recover instantaneous velocity time traces from predictions. This procedure enables more direct comparison with turbulence intensity measurements which have not been filtered to remove the oscillatory flow component. Normal wall distances, required in some turbulence models, are evaluated using a modified Poisson equation based technique. A range of zero, one and two equation turbulence models are tested, including zonal and a non-linear eddy viscosity models. The non-linear and zonal models showed potential for accuracy improvements.

The breakdown of the viscous regime in riblets

We investigate the mechanisms involved in the breakdown of the viscous regime in riblets, with a view to determining the point of optimum performance, where drag reduction ceases to be proportional to the riblet size. This occurs empirically for a groove cross-section $A_g^+ \approx 120^+$. To study the interaction of the riblets with the overlaying turbulent flow, we systematically conduct DNSes in a ribbed turbulent channel with increasing riblet size. The conditionally averaged crossflow above and within the grooves reveals a mean recirculation bubble that exists up to the point of viscous breakdown, isolating the groove floor from the overlying crossflow, and preventing the high momentum fluid from entering the grooves. We do not find evidence of outside vortices lodging within the grooves until $A_g^+ \approx 400$, which is well past the drag minimum, and already into the drag increasing regime. Interestingly, as the bubble breaks down, we observe that quasi-two-dimensional spanwise structures form just above the riblets, similar to those observed above porous surfaces and plant canopies, which appear to be involved in the performance degradation.

The mechanism of energy loss and transition in a flow with submerged vegetation

The mechanism of energy balance in an open-channel flow with submerged vegetation was investigated. The energy borrowed from the local flow, energy spending caused by vegetation drag and flow resistance, and energy transition along the water depth were calculated on the basis of the computational results of velocity and Reynolds stress. Further analysis showed that the energy spending in a cross-section was a maximum around the top of the vegetation, and its value decreased progressively until reaching zero at the flume bed or water surface. The energy borrowed from the local flow in the vegetated region could not provide for spending; therefore, surplus borrowed energy in the non-vegetated region was transmitted to the vegetated region. In addition, the total energy transition in the cross-section was zero; therefore, the total energy borrowed from the flow balanced the energy loss in the whole cross-section. At the same time, we found that there were three effects of vegetation on the flow: turbulence restriction due to vegetation, turbulence source due to vegetation and energy transference due to vegetation, where the second effect was the strongest one. Crown Copyright (C) 2010 Published by Elsevier Ltd. All rights reserved.

FLOW PROPERTIES and TUBE FRICTION FACTOR of MILK CREAM: INFLUENCE of TEMPERATURE and FAT CONTENT

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Fourth-order method for solving the Navier-Stokes equations in a constricting channel

A fourth-order numerical method for solving the Navier-Stokes equations in streamfunction/vorticity formulation on a two-dimensional non-uniform orthogonal grid has been tested on the fluid flow in a constricted symmetric channel. The family of grids is generated algebraically using a conformal transformation followed by a non-uniform stretching of the mesh cells in which the shape of the channel boundary can vary from a smooth constriction to one which one possesses a very sharp but smooth corner. The generality of the grids allows the use of long channels upstream and downstream as well as having a refined grid near the sharp corner. Derivatives in the governing equations are replaced by fourth-order central differences and the vorticity is eliminated, either before or after the discretization, to form a wide difference molecule for the streamfunction. Extra boundary conditions, necessary for wide-molecule methods, are supplied by a procedure proposed by Henshaw et al. The ensuing set of non-linear equations is solved using Newton iteration. Results have been obtained for Reynolds numbers up to 250 for three constrictions, the first being smooth, the second having a moderately sharp corner and the third with a very sharp corner. Estimates of the error incurred show that the results are very accurate and substantially better than those of the corresponding second-order method. The observed order of the method has been shown to be close to four, demonstrating that the method is genuinely fourth-order. © 1977 John Wiley & Sons, Ltd.

A Reynolds stress model for turbulent flows of viscoelastic fluids

A second-order closure is developed for predicting turbulent flows of viscoelastic fluids described by a modified generalised Newtonian fluid model incorporating a nonlinear viscosity that depends on a strain-hardening Trouton ratio as a means to handle some of the effects of viscoelasticity upon turbulent flows. Its performance is assessed by comparing its predictions for fully developed turbulent pipe flow with experimental data for four different dilute polymeric solutions and also with two sets of direct numerical simulation data for fluids theoretically described by the finitely extensible nonlinear elastic - Peterlin model. The model is based on a Newtonian Reynolds stress closure to predict Newtonian fluid flows, which incorporates low Reynolds number damping functions to properly deal with wall effects and to provide the capability to handle fluid viscoelasticity more effectively. This new turbulence model was able to capture well the drag reduction of various viscoelastic fluids over a wide range of Reynolds numbers and performed better than previously developed models for the same type of constitutive equation, even if the streamwise and wall-normal turbulence intensities were underpredicted.

Experimental setup for measuring roll waves on laminar open channel flows

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Comprehensive review and application of particle image velocimetry

For a fluid dynamics experimental flow measurement technique, particle image velocimetry (PIV) provides significant advantages over other measurement techniques in its field. In contrast to temperature and pressure based probe measurements or other laser diagnostic techniques including laser Doppler velocimetry (LDV) and phase Doppler particle analysis (PDPA), PIV is unique due to its whole field measurement capability, non-intrusive nature, and ability to collect a vast amount of experimental data in a short time frame providing both quantitative and qualitative insight. These properties make PIV a desirable measurement technique for studies encompassing a broad range of fluid dynamics applications. However, as an optical measurement technique, PIV also requires a substantial technical understanding and application experience to acquire consistent, reliable results. Both a technical understanding of particle image velocimetry and practical application experience are gained by applying a planar PIV system at Michigan Technological University’s Combustion Science Exploration Laboratory (CSEL) and Alternative Fuels Combustion Laboratory (AFCL). Here a PIV system was applied to non-reacting and reacting gaseous environments to make two component planar PIV as well as three component stereographic PIV flow field velocity measurements in conjunction with chemiluminescence imaging in the case of reacting flows. This thesis outlines near surface flow field characteristics in a tumble strip lined channel, three component velocity profiles of non-reacting and reacting swirled flow in a swirl stabilized lean condition premixed/prevaporized-fuel model gas turbine combustor operating on methane at 5-7 kW, and two component planar PIV measurements characterizing the AFCL’s 1.1 liter closed combustion chamber under dual fan driven turbulent mixing flow.

Comparison of particle dispersion obtained with simple stochastic models and les in pipe flow

A hybrid Eulerian-Lagrangian approach is employed to simulate heavy particle dispersion in turbulent pipe flow. The mean flow is provided by the Eulerian simulations developed by mean of JetCode, whereas the fluid fluctuations seen by particles are prescribed by a stochastic differential equation based on normalized Langevin. The statistics of particle velocity are compared to LES data which contain detailed statistics of velocity for particles with diameter equal to 20.4 µm. The model is in good agreement with the LES data for axial mean velocity whereas rms of axial and radial velocities should be adjusted.