ray Ristorcelli - Academia.edu (original) (raw)

Papers by ray Ristorcelli

Research paper thumbnail of High Reynolds number Rayleigh-Taylor turbulence

Bulletin of the American Physical Society, Nov 24, 2008

2006. A comprehensive study of the budgets for the kinetic energy, mass flux, and density specifi... more 2006. A comprehensive study of the budgets for the kinetic energy, mass flux, and density specific volume covariance equations is undertaken. It is found that only the large scale quantities, but not the small scale quantities, reach self-similarity. Hypotheses for the variable density turbulent transport necessary to close the second moment equations are studied. The integral length scale does not follow thek 3/2 /ϵ scaling. This is due to the non-equilibrium nature of the flow and the fact that ϵ, a small scale quantity, does not have the self-similar scaling. As a consequence, the popular eddy viscosity expressionk 2 /ϵ does not model the turbulent transport in any of the moment equations. An integral length scale, based on the layer width, does scale the turbulent transport using a gradient transport hypothesis; that integral scale is a global quantity and does not lead to pointwise local closure. Despite the fact that the intermediate scales are nearly isotropic, the small scales have a persistent anisotropy; this is due to a cancellation between the viscous and nonlinear effects, so that the anisotropic buoyancy production remains important at the smallest scales. Various asymmetries in the mixing layer, not seen in the Boussinesq case, are also identified and explained.

Research paper thumbnail of Turbulence in reshocked Richtmyer-Meshkov unstable fluid layers

Bulletin of the American Physical Society, Nov 21, 2010

ABSTRACT Advances in the implementation of high resolution PIV (150um vector-to-vector resolution... more ABSTRACT Advances in the implementation of high resolution PIV (150um vector-to-vector resolution) and PLIF (50um resolution) diagnostics have allowed the experimental measurement and characterization of turbulent mixing in Richtmyer-Meshkov unstable fluid layers after reshock (Balakumar et. al., Phys. Fluids, 2008). Using instantaneous PLIF data obtained at closely spaced intervals of time, we illustrate the rapid disintegration of the primary wavelengths of the initial interface and the beginning of a turbulence cascade generating smaller flow structures after reshock. The enhanced mixing is reflected in the variation of the density probability distribution function between the pre-reshock and post-reshock states. The density self-correlation is observed to exhibit a double-peaked structure and mild non-Boussinesq effects are observed in a layer with varicose initial interfacial perturbations. Density and velocity pdfs are used to examine the streamwise asymmetry of the mixing layer with large fluctuations occurring preferentially upstream of the centerline. Other turbulence statistics including the 2nd and 4th order structure functions, RMS statistics (both velocity and density) and turbulence intensity are also presented.

Research paper thumbnail of Simulations of Richtmyer-Meshkov instabilities in planar and gas-curtain shocktube experiments (u)

In the large eddy simulation ͑LES͒ approach, large-scale energy-containing structures are resolve... more In the large eddy simulation ͑LES͒ approach, large-scale energy-containing structures are resolved, smaller structures are filtered out, and unresolved subgrid effects are modeled. Extensive recent work has demonstrated that predictive under-resolved simulations of the velocity fields in turbulent flows are possible without resorting to explicit subgrid models when using a class of physics-capturing high-resolution finite-volume numerical algorithms. This strategy is denoted as implicit LES ͑ILES͒. Tests in fundamental applications ranging from canonical to complex flows indicate that ILES is competitive with conventional LES in the LES realm proper-flows driven by large-scale features. The performance of ILES in the substantially more difficult problem of under-resolved material mixing driven by under-resolved velocity fields and initial conditions is a focus of the present work. Progress in addressing relevant resolution issues in studies of mixing driven by Richtmyer-Meshkov instabilities in planar shock-tube laboratory experiments is reported. Our particular focus is devoted to the initial material interface characterization and modeling difficulties, and effects of initial condition specifics ͑resolved spectral content͒ on transitional and late-time turbulent mixing-which were not previously addressed.

Research paper thumbnail of Variable density mixing in buoyancy driven turbulence

APS, Nov 1, 2007

The homogenization of a heterogeneous mixture of two pure fluids with different densities by mole... more The homogenization of a heterogeneous mixture of two pure fluids with different densities by molecular diffusion and stirring induced by buoyancy generated motions, as occurs in the Rayleigh-Taylor instability, is studied using Direct Numerical Simulations. The Schmidt number, Sc, is varied by a factor of twenty, 0.1 <=Sc <=2.0, and the Atwood number, A, a factor of ten, 0.05 <=A <=0.5. Important differences between the mixing in a variable density (VD) fluid, as compared to a Boussinesq fluid, are observed. The pure heavy fluid mixes more slowly than the pure light fluid: an initially symmetric double delta density PDF is rapidly skewed and, only at long times and low density fluctuations, it relaxes to a Gaussian-like PDF. Diverse mix metrics and their dependence on A, Re, and Sc are used to examine the homogenization of the two fluids. In particular, it is shown that the specific volume density covariance is a better measure of the mixing state than the density variance for VD flows, as it directly appears in the dynamical equations. The usual mix parameter theta is mathematically related to the variance of the excess reactant of a hypothetically fast chemical reaction. Analytic expressions and bounds relating theta and the normalized product, xi, are derived. In general, no usual mix measure can predict the amount of pure or mixed fluid in the flow, however bounds on the fluid composition, using low order moments of the density PDF, can be derived.

Research paper thumbnail of ICASE/LaRC Workshop on Benchmark Problems in Computational Aeroacoustics (CAA)

This volume contains the proceedings of the ICASE/LaRC Workshop on Benchmark Problems in Computat... more This volume contains the proceedings of the ICASE/LaRC Workshop on Benchmark Problems in Computational Aeroacoustics (CAA). CAA is a relatively new area of research addressing issues relevant to the acoustic propagation of sound generated by fluid flow. Advances in computer technology make addressing these issues in a more detailed manner a possibility. Such advances allow the treatment of the fully nonlinear propagation problem as well as the direct computation of the acoustic sources, the sound generation problem. These possibilities are expected to benefit both the validation of models that have been developed as well as help develop better models for more complex flows. In situations where calculations are not prohibitively expensive a direct computation of the acoustic and source fields becomes possible. When research in this area was first considered several technical challenges were apparent. The intention of these proceedings is to more fully investigate a subset of these numerical issues and make some progress in their resolution. These issues include: 1. The small magnitude of the acoustical quantities of interest and the need to distinguish and extract them from the larger background fields. 2. The sensitive dependence of the acoustical field on phase, dissipation and dispersion when propagated over large spatial distances. 3. The potentially higher frequencies of the quantities of interest in comparison to those of interest in the problems more typically addressed in unsteady aerodynamics or structural vibrations. 4. For the computation of acoustical spectra long time solutions are necessary for computing averages; numerical codes are required to be stable and accurate for long time integrations. 5. Many codes are designed for stationary problems in which the path of approach to the asymptotic solution is not important (except from the viewpoint of cost). These schemes are potentially inadequate for aeroacoustical problems in which time accurate computations are required. The dissipation, dispersion, and anisotropic biases in these schemes are now very relevant to the aeroacoustical problems of interest. 6. Time dependent boundary conditions are also required which will not reflect acoustic waves from imposed computational boundaries yet reflect acoustic waves properly from real physical boundaries. 0 5 x 5 100, 0 < r < 100. The wall and the piston are at x = 0. The cylindrical coordinate system is centered at the center of the piston. With axisymmetry, the linearized Euler equations are The initial conditions are: t = O p = u = v = p = O 1 1 3 Give the time harmonic pressure distribution at the beginning,-andof a period of 4 ' 2 4 piston oscillation. Category 5 Problem to test the suitability of a numerical scheme for direct numerical simulation of very small amplitude acoustic waves superimposed on a non-uniform mean flows in a semi-infinite duct. Use nondimensional variables with the following scales Ax = length scale a, (sound speed far upstream) = velocity scale Ax-= time scale a00 p, (density of gas upstream) = density scale ,o,a2, = pressure scale A small amplitude sound wave is incident on a convergent-divergent nozzle as shown c M=0.5

Research paper thumbnail of Effect of counter-gradient subgrid-scale transport on turbulent mixing

APS Division of Fluid Dynamics Meeting Abstracts, Nov 1, 2019

Research paper thumbnail of Effects of Mass and Volume Fraction Skewness in Variable Density Mixing Processes

Bulletin of the American Physical Society, 2015

by variable density turbulence of fluids involving density variations of a factor of 5 to 10 are ... more by variable density turbulence of fluids involving density variations of a factor of 5 to 10 are the Atwood, Froude, Schmidt, and Reynolds numbers. There is evidence that the amount of each fluid present when the two pure fluids mix, as described by the probability density function of the mass or molar (volume) fraction, also strongly affects the mixing process. To investigate this phenomena, implicit large-eddy simulations (ILES) are performed for binary fluid mixtures in statistically homogenous environments under constant acceleration. These coarse grained simulations are used as data for theory validation and mix model development. ILES has been demonstrated to accurately capture the mixing behavior of a passive scalar field through stirring and advection by a turbulent velocity field. The present work advances that research and studies the extent to which an underresolved active scalar drives the subsequent fluid motion and determines the nature of the mixing process. Effects of initial distributions of the mass and molar (volume) fraction probability density function on the resulting variable density turbulence and mixing are investigated and compared to direct numerical simulations from the Johns Hopkins Turbulence Database.

Research paper thumbnail of Modeling Reaction Rates in Variable-Density Turbulent Mixing

Bulletin of the American Physical Society, 2014

Research paper thumbnail of ILES of Passive Scalar Mixing in Forced Isotropic Turbulence

Bulletin of the American Physical Society, 2012

Submitted for the DFD12 Meeting of The American Physical Society ILES of Passive Scalar Mixing in... more Submitted for the DFD12 Meeting of The American Physical Society ILES of Passive Scalar Mixing in Forced Isotropic Turbulence ADAM WACHTOR, FERNANDO GRINSTEIN, Los Alamos National Laboratory, RICK DEVORE, Naval Research Laboratory, RAY RISTORCELLI, LEN MARGOLIN, Los Alamos National Laboratory — Predictability of scalar mixing by an under-resolved turbulent velocity field is investigated using ILES. Turbulent mixing of a passive scalar by forced, compressible, isotropic turbulence with a prescribed mean scalar gradient is studied. The simulation strategy uses a multi-dimensional FCT algorithm, with low wavenumber momentum forcing imposed separately for the solenoidal and dilatational velocity components. Effects of grid resolution on the flow and scalar mixing are investigated at turbulent Mach numbers 0.13 and 0.27. ILES captures the mixing transition as function of effective Reynolds number determined by grid resolution, including asymptotic behaviors and characteristic turbulent m...

Research paper thumbnail of Investigating the Mixing-Hydrodynamic Cascade Timescale Ratio in Transient Non-Equilibrium Turbulence Using a Fokker-Planck Equation Approach

Bulletin of the American Physical Society, 2018

Research paper thumbnail of Abstract Submitted for the DFD08 Meeting of The American Physical Society

Reversed trajectories of interacting pair of drops in a steady shear at flnite inertia KAUSIK SAR... more Reversed trajectories of interacting pair of drops in a steady shear at flnite inertia KAUSIK SARKAR, PETER OLAPADE, RAJESH SINGH, University of Delaware | Interactions between viscous drops in a steady shear are numerically simulated using front tracking flnite difierence method. The results match well at low Reynolds number with the experimental observations of Guido and Simione (1998). At flnite inertia, the drops behave difierently from that in Stokes ∞ow. Two drops placed initially ∞ow-direction and shear-direction ofisets do not pass each other. Instead, the drops upon interaction reverse their trajectories. Such behaviors were already reported for freely rotating solid particles, where it is ascribed to the reversed streamlines around a particle at flnite inertia. However, drop deformation critically afiects the trajectories, in that for some values of Reynolds numbers, drops pass each other at low and high capillary numbers, but reverse their motion at intermediate capillary...

Research paper thumbnail of The effect of capturing the correct turbulence dissipation rate in BHR

In this manuscript, we discuss the shortcoming of a quasi-equilibrium assumption made in the BHR ... more In this manuscript, we discuss the shortcoming of a quasi-equilibrium assumption made in the BHR closure model. Turbulence closure models generally assume fully developed turbulence, which is not applicable to 1) non-equilibrium turbulence (e.g. change in mean pressure gradient) or 2) laminar-turbulence transition flows. Based on DNS data, we show that the current BHR dissipation equation [modeled based on the fully developed turbulence phenomenology] does not capture important features of nonequilibrium flows. To demonstrate our thesis, we use the BHR equations to predict a non-equilibrium flow both with the BHR dissipation and the dissipation from DNS. We find that the prediction can be substantially improved, both qualitatively and quantitatively, with the correct dissipation rate. We conclude that a new set of nonequilibrium phenomenological assumptions must be used to develop a new model equation for the dissipation to accurately predict the turbulence time scale used by other models.

Research paper thumbnail of ICASE/LaRC Workshop on Benchmark Problems in Computational Aeroacoustics (CAA)

This volume contains the proceedings of the ICASE/LaRC Workshop on Benchmark Problems in Computat... more This volume contains the proceedings of the ICASE/LaRC Workshop on Benchmark Problems in Computational Aeroacoustics (CAA). CAA is a relatively new area of research addressing issues relevant to the acoustic propagation of sound generated by fluid flow. Advances in computer technology make addressing these issues in a more detailed manner a possibility. Such advances allow the treatment of the fully nonlinear propagation problem as well as the direct computation of the acoustic sources, the sound generation problem. These possibilities are expected to benefit both the validation of models that have been developed as well as help develop better models for more complex flows. In situations where calculations are not prohibitively expensive a direct computation of the acoustic and source fields becomes possible. When research in this area was first considered several technical challenges were apparent. The intention of these proceedings is to more fully investigate a subset of these numerical issues and make some progress in their resolution. These issues include: 1. The small magnitude of the acoustical quantities of interest and the need to distinguish and extract them from the larger background fields. 2. The sensitive dependence of the acoustical field on phase, dissipation and dispersion when propagated over large spatial distances. 3. The potentially higher frequencies of the quantities of interest in comparison to those of interest in the problems more typically addressed in unsteady aerodynamics or structural vibrations. 4. For the computation of acoustical spectra long time solutions are necessary for computing averages; numerical codes are required to be stable and accurate for long time integrations. 5. Many codes are designed for stationary problems in which the path of approach to the asymptotic solution is not important (except from the viewpoint of cost). These schemes are potentially inadequate for aeroacoustical problems in which time accurate computations are required. The dissipation, dispersion, and anisotropic biases in these schemes are now very relevant to the aeroacoustical problems of interest. 6. Time dependent boundary conditions are also required which will not reflect acoustic waves from imposed computational boundaries yet reflect acoustic waves properly from real physical boundaries. 0 5 x 5 100, 0 < r < 100. The wall and the piston are at x = 0. The cylindrical coordinate system is centered at the center of the piston. With axisymmetry, the linearized Euler equations are The initial conditions are: t = O p = u = v = p = O 1 1 3 Give the time harmonic pressure distribution at the beginning,-andof a period of 4 ' 2 4 piston oscillation. Category 5 Problem to test the suitability of a numerical scheme for direct numerical simulation of very small amplitude acoustic waves superimposed on a non-uniform mean flows in a semi-infinite duct. Use nondimensional variables with the following scales Ax = length scale a, (sound speed far upstream) = velocity scale Ax-= time scale a00 p, (density of gas upstream) = density scale ,o,a2, = pressure scale A small amplitude sound wave is incident on a convergent-divergent nozzle as shown c M=0.5

Research paper thumbnail of The bipolar behavior of the Richtmyer–Meshkov instability

Physics of Fluids, 2011

A numerical study of the evolution of the multimode planar Richtmyer-Meshkov instability (RMI) in... more A numerical study of the evolution of the multimode planar Richtmyer-Meshkov instability (RMI) in a light-heavy (air-SF6, Atwood number A = 0.67) configuration involving a Mach number Ma = 1.5 shock is carried out. Our results demonstrate that the initial material interface morphology controls the evolution characteristics of RMI (for fixed A, Ma), and provide a significant basis to develop metrics for transition to turbulence. Depending on initial rms slope of the interface, RMI evolves into linear or nonlinear regimes, with distinctly different flow features and growth rates, turbulence statistics, and material mixing rates. We have called this the bipolar behavior of RMI. Some of our findings are not consistent with heuristic notions of mixing in equilibrium turbulence: more turbulent flow—as measured by spectral bandwidth, can be associated with higher material mixing but, paradoxically, to lower integral measures of turbulent kinetic energy and mixing layer width.

Research paper thumbnail of Simulations of Richtmyer–Meshkov instabilities in planar shock-tube experiments

Physics of Fluids, 2011

In the large eddy simulation (LES) approach, large-scale energy-containing structures are resolve... more In the large eddy simulation (LES) approach, large-scale energy-containing structures are resolved, smaller structures are filtered out, and unresolved subgrid effects are modeled. Extensive recent work has demonstrated that predictive under-resolved simulations of the velocity fields in turbulent flows are possible without resorting to explicit subgrid models when using a class of physics-capturing high-resolution finite-volume numerical algorithms. This strategy is denoted as implicit LES (ILES). Tests in fundamental applications ranging from canonical to complex flows indicate that ILES is competitive with conventional LES in the LES realm proper—flows driven by large-scale features. The performance of ILES in the substantially more difficult problem of under-resolved material mixing driven by under-resolved velocity fields and initial conditions is a focus of the present work. Progress in addressing relevant resolution issues in studies of mixing driven by Richtmyer–Meshkov inst...

Research paper thumbnail of High-Reynolds number Rayleigh–Taylor turbulence

Journal of Turbulence, 2009

2006. A comprehensive study of the budgets for the kinetic energy, mass flux, and density specifi... more 2006. A comprehensive study of the budgets for the kinetic energy, mass flux, and density specific volume covariance equations is undertaken. It is found that only the large scale quantities, but not the small scale quantities, reach self-similarity. Hypotheses for the variable density turbulent transport necessary to close the second moment equations are studied. The integral length scale does not follow thek 3/2 /ϵ scaling. This is due to the non-equilibrium nature of the flow and the fact that ϵ, a small scale quantity, does not have the self-similar scaling. As a consequence, the popular eddy viscosity expressionk 2 /ϵ does not model the turbulent transport in any of the moment equations. An integral length scale, based on the layer width, does scale the turbulent transport using a gradient transport hypothesis; that integral scale is a global quantity and does not lead to pointwise local closure. Despite the fact that the intermediate scales are nearly isotropic, the small scales have a persistent anisotropy; this is due to a cancellation between the viscous and nonlinear effects, so that the anisotropic buoyancy production remains important at the smallest scales. Various asymmetries in the mixing layer, not seen in the Boussinesq case, are also identified and explained.

Research paper thumbnail of Planar Richtmyer-Meshkov Instabilities and Transition

Journal of Physics: Conference Series, 2011

A numerical study of the evolution of the multimode planar Richtmyer-Meshkov (RM) instability in ... more A numerical study of the evolution of the multimode planar Richtmyer-Meshkov (RM) instability in a light-heavy (Air-SF6, Atwood number A = 0.67) configuration involving a Mach number Ma = 1.5 shock is carried out. Our results demonstrate that the initial material interface morphology controls the evolution RM characteristics, and provide a significant basis to develop metrics for transition to turbulence. Depending on the initial rms slope of the interface, RM evolves into linear or nonlinear regimes, with distinctly different flow features and growth rates, turbulence statistics and material mixing rates. We have called this the bipolar behavior of the RM instability. We demonstrate an important practical consequence of our results: reshock effects on mixing and transition can be emulated at first shock if the initial rms slope is high enough.

Research paper thumbnail of Probability density function method for variable-density pressure-gradient-driven turbulence and mixing

Probability density function (PDF) methods are extended to variable-density pressure-gradient-dri... more Probability density function (PDF) methods are extended to variable-density pressure-gradient-driven turbulence. We apply the new method to compute the joint PDF of density and velocity in a non-premixed binary mixture of different-density molecularly mixing fluids under gravity. The full time-evolution of the joint PDF is captured in the highly non-equilibrium flow: starting from a quiescent state, transitioning to fully developed

Research paper thumbnail of Turbulence in reshocked Richtmyer-Meshkov unstable fluid layers

ABSTRACT Advances in the implementation of high resolution PIV (150um vector-to-vector resolution... more ABSTRACT Advances in the implementation of high resolution PIV (150um vector-to-vector resolution) and PLIF (50um resolution) diagnostics have allowed the experimental measurement and characterization of turbulent mixing in Richtmyer-Meshkov unstable fluid layers after reshock (Balakumar et. al., Phys. Fluids, 2008). Using instantaneous PLIF data obtained at closely spaced intervals of time, we illustrate the rapid disintegration of the primary wavelengths of the initial interface and the beginning of a turbulence cascade generating smaller flow structures after reshock. The enhanced mixing is reflected in the variation of the density probability distribution function between the pre-reshock and post-reshock states. The density self-correlation is observed to exhibit a double-peaked structure and mild non-Boussinesq effects are observed in a layer with varicose initial interfacial perturbations. Density and velocity pdfs are used to examine the streamwise asymmetry of the mixing layer with large fluctuations occurring preferentially upstream of the centerline. Other turbulence statistics including the 2nd and 4th order structure functions, RMS statistics (both velocity and density) and turbulence intensity are also presented.

Research paper thumbnail of Variable-density mixing in buoyancy-driven turbulence

Journal of Fluid Mechanics, 2008

The homogenization of a heterogeneous mixture of two pure fluids with different densities by mole... more The homogenization of a heterogeneous mixture of two pure fluids with different densities by molecular diffusion and stirring induced by buoyancy-generated motions, as occurs in the Rayleigh–Taylor (RT) instability, is studied using direct numerical simulations. The Schmidt number, Sc, is varied by a factor of 20, 0.1 ≤ Sc ≤ 2.0, and the Atwood number, A, by a factor of 10, 0.05 ≤ A ≤ 0.5. Initial-density intensities are as high as 50% of the mean density. As a consequence of differential accelerations experienced by the two fluids, substantial and important differences between the mixing in a variable-density flow, as compared to the Boussinesq approximation, are observed. In short, the pure heavy fluid mixes more slowly than the pure light fluid: an initially symmetric double delta density probability density function (PDF) is rapidly skewed and, only at long times and low density fluctuations, does it relax to a Gaussian-like PDF. The heavy–light fluid mixing process asymmetry is...

Research paper thumbnail of High Reynolds number Rayleigh-Taylor turbulence

Bulletin of the American Physical Society, Nov 24, 2008

2006. A comprehensive study of the budgets for the kinetic energy, mass flux, and density specifi... more 2006. A comprehensive study of the budgets for the kinetic energy, mass flux, and density specific volume covariance equations is undertaken. It is found that only the large scale quantities, but not the small scale quantities, reach self-similarity. Hypotheses for the variable density turbulent transport necessary to close the second moment equations are studied. The integral length scale does not follow thek 3/2 /ϵ scaling. This is due to the non-equilibrium nature of the flow and the fact that ϵ, a small scale quantity, does not have the self-similar scaling. As a consequence, the popular eddy viscosity expressionk 2 /ϵ does not model the turbulent transport in any of the moment equations. An integral length scale, based on the layer width, does scale the turbulent transport using a gradient transport hypothesis; that integral scale is a global quantity and does not lead to pointwise local closure. Despite the fact that the intermediate scales are nearly isotropic, the small scales have a persistent anisotropy; this is due to a cancellation between the viscous and nonlinear effects, so that the anisotropic buoyancy production remains important at the smallest scales. Various asymmetries in the mixing layer, not seen in the Boussinesq case, are also identified and explained.

Research paper thumbnail of Turbulence in reshocked Richtmyer-Meshkov unstable fluid layers

Bulletin of the American Physical Society, Nov 21, 2010

ABSTRACT Advances in the implementation of high resolution PIV (150um vector-to-vector resolution... more ABSTRACT Advances in the implementation of high resolution PIV (150um vector-to-vector resolution) and PLIF (50um resolution) diagnostics have allowed the experimental measurement and characterization of turbulent mixing in Richtmyer-Meshkov unstable fluid layers after reshock (Balakumar et. al., Phys. Fluids, 2008). Using instantaneous PLIF data obtained at closely spaced intervals of time, we illustrate the rapid disintegration of the primary wavelengths of the initial interface and the beginning of a turbulence cascade generating smaller flow structures after reshock. The enhanced mixing is reflected in the variation of the density probability distribution function between the pre-reshock and post-reshock states. The density self-correlation is observed to exhibit a double-peaked structure and mild non-Boussinesq effects are observed in a layer with varicose initial interfacial perturbations. Density and velocity pdfs are used to examine the streamwise asymmetry of the mixing layer with large fluctuations occurring preferentially upstream of the centerline. Other turbulence statistics including the 2nd and 4th order structure functions, RMS statistics (both velocity and density) and turbulence intensity are also presented.

Research paper thumbnail of Simulations of Richtmyer-Meshkov instabilities in planar and gas-curtain shocktube experiments (u)

In the large eddy simulation ͑LES͒ approach, large-scale energy-containing structures are resolve... more In the large eddy simulation ͑LES͒ approach, large-scale energy-containing structures are resolved, smaller structures are filtered out, and unresolved subgrid effects are modeled. Extensive recent work has demonstrated that predictive under-resolved simulations of the velocity fields in turbulent flows are possible without resorting to explicit subgrid models when using a class of physics-capturing high-resolution finite-volume numerical algorithms. This strategy is denoted as implicit LES ͑ILES͒. Tests in fundamental applications ranging from canonical to complex flows indicate that ILES is competitive with conventional LES in the LES realm proper-flows driven by large-scale features. The performance of ILES in the substantially more difficult problem of under-resolved material mixing driven by under-resolved velocity fields and initial conditions is a focus of the present work. Progress in addressing relevant resolution issues in studies of mixing driven by Richtmyer-Meshkov instabilities in planar shock-tube laboratory experiments is reported. Our particular focus is devoted to the initial material interface characterization and modeling difficulties, and effects of initial condition specifics ͑resolved spectral content͒ on transitional and late-time turbulent mixing-which were not previously addressed.

Research paper thumbnail of Variable density mixing in buoyancy driven turbulence

APS, Nov 1, 2007

The homogenization of a heterogeneous mixture of two pure fluids with different densities by mole... more The homogenization of a heterogeneous mixture of two pure fluids with different densities by molecular diffusion and stirring induced by buoyancy generated motions, as occurs in the Rayleigh-Taylor instability, is studied using Direct Numerical Simulations. The Schmidt number, Sc, is varied by a factor of twenty, 0.1 <=Sc <=2.0, and the Atwood number, A, a factor of ten, 0.05 <=A <=0.5. Important differences between the mixing in a variable density (VD) fluid, as compared to a Boussinesq fluid, are observed. The pure heavy fluid mixes more slowly than the pure light fluid: an initially symmetric double delta density PDF is rapidly skewed and, only at long times and low density fluctuations, it relaxes to a Gaussian-like PDF. Diverse mix metrics and their dependence on A, Re, and Sc are used to examine the homogenization of the two fluids. In particular, it is shown that the specific volume density covariance is a better measure of the mixing state than the density variance for VD flows, as it directly appears in the dynamical equations. The usual mix parameter theta is mathematically related to the variance of the excess reactant of a hypothetically fast chemical reaction. Analytic expressions and bounds relating theta and the normalized product, xi, are derived. In general, no usual mix measure can predict the amount of pure or mixed fluid in the flow, however bounds on the fluid composition, using low order moments of the density PDF, can be derived.

Research paper thumbnail of ICASE/LaRC Workshop on Benchmark Problems in Computational Aeroacoustics (CAA)

This volume contains the proceedings of the ICASE/LaRC Workshop on Benchmark Problems in Computat... more This volume contains the proceedings of the ICASE/LaRC Workshop on Benchmark Problems in Computational Aeroacoustics (CAA). CAA is a relatively new area of research addressing issues relevant to the acoustic propagation of sound generated by fluid flow. Advances in computer technology make addressing these issues in a more detailed manner a possibility. Such advances allow the treatment of the fully nonlinear propagation problem as well as the direct computation of the acoustic sources, the sound generation problem. These possibilities are expected to benefit both the validation of models that have been developed as well as help develop better models for more complex flows. In situations where calculations are not prohibitively expensive a direct computation of the acoustic and source fields becomes possible. When research in this area was first considered several technical challenges were apparent. The intention of these proceedings is to more fully investigate a subset of these numerical issues and make some progress in their resolution. These issues include: 1. The small magnitude of the acoustical quantities of interest and the need to distinguish and extract them from the larger background fields. 2. The sensitive dependence of the acoustical field on phase, dissipation and dispersion when propagated over large spatial distances. 3. The potentially higher frequencies of the quantities of interest in comparison to those of interest in the problems more typically addressed in unsteady aerodynamics or structural vibrations. 4. For the computation of acoustical spectra long time solutions are necessary for computing averages; numerical codes are required to be stable and accurate for long time integrations. 5. Many codes are designed for stationary problems in which the path of approach to the asymptotic solution is not important (except from the viewpoint of cost). These schemes are potentially inadequate for aeroacoustical problems in which time accurate computations are required. The dissipation, dispersion, and anisotropic biases in these schemes are now very relevant to the aeroacoustical problems of interest. 6. Time dependent boundary conditions are also required which will not reflect acoustic waves from imposed computational boundaries yet reflect acoustic waves properly from real physical boundaries. 0 5 x 5 100, 0 < r < 100. The wall and the piston are at x = 0. The cylindrical coordinate system is centered at the center of the piston. With axisymmetry, the linearized Euler equations are The initial conditions are: t = O p = u = v = p = O 1 1 3 Give the time harmonic pressure distribution at the beginning,-andof a period of 4 ' 2 4 piston oscillation. Category 5 Problem to test the suitability of a numerical scheme for direct numerical simulation of very small amplitude acoustic waves superimposed on a non-uniform mean flows in a semi-infinite duct. Use nondimensional variables with the following scales Ax = length scale a, (sound speed far upstream) = velocity scale Ax-= time scale a00 p, (density of gas upstream) = density scale ,o,a2, = pressure scale A small amplitude sound wave is incident on a convergent-divergent nozzle as shown c M=0.5

Research paper thumbnail of Effect of counter-gradient subgrid-scale transport on turbulent mixing

APS Division of Fluid Dynamics Meeting Abstracts, Nov 1, 2019

Research paper thumbnail of Effects of Mass and Volume Fraction Skewness in Variable Density Mixing Processes

Bulletin of the American Physical Society, 2015

by variable density turbulence of fluids involving density variations of a factor of 5 to 10 are ... more by variable density turbulence of fluids involving density variations of a factor of 5 to 10 are the Atwood, Froude, Schmidt, and Reynolds numbers. There is evidence that the amount of each fluid present when the two pure fluids mix, as described by the probability density function of the mass or molar (volume) fraction, also strongly affects the mixing process. To investigate this phenomena, implicit large-eddy simulations (ILES) are performed for binary fluid mixtures in statistically homogenous environments under constant acceleration. These coarse grained simulations are used as data for theory validation and mix model development. ILES has been demonstrated to accurately capture the mixing behavior of a passive scalar field through stirring and advection by a turbulent velocity field. The present work advances that research and studies the extent to which an underresolved active scalar drives the subsequent fluid motion and determines the nature of the mixing process. Effects of initial distributions of the mass and molar (volume) fraction probability density function on the resulting variable density turbulence and mixing are investigated and compared to direct numerical simulations from the Johns Hopkins Turbulence Database.

Research paper thumbnail of Modeling Reaction Rates in Variable-Density Turbulent Mixing

Bulletin of the American Physical Society, 2014

Research paper thumbnail of ILES of Passive Scalar Mixing in Forced Isotropic Turbulence

Bulletin of the American Physical Society, 2012

Submitted for the DFD12 Meeting of The American Physical Society ILES of Passive Scalar Mixing in... more Submitted for the DFD12 Meeting of The American Physical Society ILES of Passive Scalar Mixing in Forced Isotropic Turbulence ADAM WACHTOR, FERNANDO GRINSTEIN, Los Alamos National Laboratory, RICK DEVORE, Naval Research Laboratory, RAY RISTORCELLI, LEN MARGOLIN, Los Alamos National Laboratory — Predictability of scalar mixing by an under-resolved turbulent velocity field is investigated using ILES. Turbulent mixing of a passive scalar by forced, compressible, isotropic turbulence with a prescribed mean scalar gradient is studied. The simulation strategy uses a multi-dimensional FCT algorithm, with low wavenumber momentum forcing imposed separately for the solenoidal and dilatational velocity components. Effects of grid resolution on the flow and scalar mixing are investigated at turbulent Mach numbers 0.13 and 0.27. ILES captures the mixing transition as function of effective Reynolds number determined by grid resolution, including asymptotic behaviors and characteristic turbulent m...

Research paper thumbnail of Investigating the Mixing-Hydrodynamic Cascade Timescale Ratio in Transient Non-Equilibrium Turbulence Using a Fokker-Planck Equation Approach

Bulletin of the American Physical Society, 2018

Research paper thumbnail of Abstract Submitted for the DFD08 Meeting of The American Physical Society

Reversed trajectories of interacting pair of drops in a steady shear at flnite inertia KAUSIK SAR... more Reversed trajectories of interacting pair of drops in a steady shear at flnite inertia KAUSIK SARKAR, PETER OLAPADE, RAJESH SINGH, University of Delaware | Interactions between viscous drops in a steady shear are numerically simulated using front tracking flnite difierence method. The results match well at low Reynolds number with the experimental observations of Guido and Simione (1998). At flnite inertia, the drops behave difierently from that in Stokes ∞ow. Two drops placed initially ∞ow-direction and shear-direction ofisets do not pass each other. Instead, the drops upon interaction reverse their trajectories. Such behaviors were already reported for freely rotating solid particles, where it is ascribed to the reversed streamlines around a particle at flnite inertia. However, drop deformation critically afiects the trajectories, in that for some values of Reynolds numbers, drops pass each other at low and high capillary numbers, but reverse their motion at intermediate capillary...

Research paper thumbnail of The effect of capturing the correct turbulence dissipation rate in BHR

In this manuscript, we discuss the shortcoming of a quasi-equilibrium assumption made in the BHR ... more In this manuscript, we discuss the shortcoming of a quasi-equilibrium assumption made in the BHR closure model. Turbulence closure models generally assume fully developed turbulence, which is not applicable to 1) non-equilibrium turbulence (e.g. change in mean pressure gradient) or 2) laminar-turbulence transition flows. Based on DNS data, we show that the current BHR dissipation equation [modeled based on the fully developed turbulence phenomenology] does not capture important features of nonequilibrium flows. To demonstrate our thesis, we use the BHR equations to predict a non-equilibrium flow both with the BHR dissipation and the dissipation from DNS. We find that the prediction can be substantially improved, both qualitatively and quantitatively, with the correct dissipation rate. We conclude that a new set of nonequilibrium phenomenological assumptions must be used to develop a new model equation for the dissipation to accurately predict the turbulence time scale used by other models.

Research paper thumbnail of ICASE/LaRC Workshop on Benchmark Problems in Computational Aeroacoustics (CAA)

This volume contains the proceedings of the ICASE/LaRC Workshop on Benchmark Problems in Computat... more This volume contains the proceedings of the ICASE/LaRC Workshop on Benchmark Problems in Computational Aeroacoustics (CAA). CAA is a relatively new area of research addressing issues relevant to the acoustic propagation of sound generated by fluid flow. Advances in computer technology make addressing these issues in a more detailed manner a possibility. Such advances allow the treatment of the fully nonlinear propagation problem as well as the direct computation of the acoustic sources, the sound generation problem. These possibilities are expected to benefit both the validation of models that have been developed as well as help develop better models for more complex flows. In situations where calculations are not prohibitively expensive a direct computation of the acoustic and source fields becomes possible. When research in this area was first considered several technical challenges were apparent. The intention of these proceedings is to more fully investigate a subset of these numerical issues and make some progress in their resolution. These issues include: 1. The small magnitude of the acoustical quantities of interest and the need to distinguish and extract them from the larger background fields. 2. The sensitive dependence of the acoustical field on phase, dissipation and dispersion when propagated over large spatial distances. 3. The potentially higher frequencies of the quantities of interest in comparison to those of interest in the problems more typically addressed in unsteady aerodynamics or structural vibrations. 4. For the computation of acoustical spectra long time solutions are necessary for computing averages; numerical codes are required to be stable and accurate for long time integrations. 5. Many codes are designed for stationary problems in which the path of approach to the asymptotic solution is not important (except from the viewpoint of cost). These schemes are potentially inadequate for aeroacoustical problems in which time accurate computations are required. The dissipation, dispersion, and anisotropic biases in these schemes are now very relevant to the aeroacoustical problems of interest. 6. Time dependent boundary conditions are also required which will not reflect acoustic waves from imposed computational boundaries yet reflect acoustic waves properly from real physical boundaries. 0 5 x 5 100, 0 < r < 100. The wall and the piston are at x = 0. The cylindrical coordinate system is centered at the center of the piston. With axisymmetry, the linearized Euler equations are The initial conditions are: t = O p = u = v = p = O 1 1 3 Give the time harmonic pressure distribution at the beginning,-andof a period of 4 ' 2 4 piston oscillation. Category 5 Problem to test the suitability of a numerical scheme for direct numerical simulation of very small amplitude acoustic waves superimposed on a non-uniform mean flows in a semi-infinite duct. Use nondimensional variables with the following scales Ax = length scale a, (sound speed far upstream) = velocity scale Ax-= time scale a00 p, (density of gas upstream) = density scale ,o,a2, = pressure scale A small amplitude sound wave is incident on a convergent-divergent nozzle as shown c M=0.5

Research paper thumbnail of The bipolar behavior of the Richtmyer–Meshkov instability

Physics of Fluids, 2011

A numerical study of the evolution of the multimode planar Richtmyer-Meshkov instability (RMI) in... more A numerical study of the evolution of the multimode planar Richtmyer-Meshkov instability (RMI) in a light-heavy (air-SF6, Atwood number A = 0.67) configuration involving a Mach number Ma = 1.5 shock is carried out. Our results demonstrate that the initial material interface morphology controls the evolution characteristics of RMI (for fixed A, Ma), and provide a significant basis to develop metrics for transition to turbulence. Depending on initial rms slope of the interface, RMI evolves into linear or nonlinear regimes, with distinctly different flow features and growth rates, turbulence statistics, and material mixing rates. We have called this the bipolar behavior of RMI. Some of our findings are not consistent with heuristic notions of mixing in equilibrium turbulence: more turbulent flow—as measured by spectral bandwidth, can be associated with higher material mixing but, paradoxically, to lower integral measures of turbulent kinetic energy and mixing layer width.

Research paper thumbnail of Simulations of Richtmyer–Meshkov instabilities in planar shock-tube experiments

Physics of Fluids, 2011

In the large eddy simulation (LES) approach, large-scale energy-containing structures are resolve... more In the large eddy simulation (LES) approach, large-scale energy-containing structures are resolved, smaller structures are filtered out, and unresolved subgrid effects are modeled. Extensive recent work has demonstrated that predictive under-resolved simulations of the velocity fields in turbulent flows are possible without resorting to explicit subgrid models when using a class of physics-capturing high-resolution finite-volume numerical algorithms. This strategy is denoted as implicit LES (ILES). Tests in fundamental applications ranging from canonical to complex flows indicate that ILES is competitive with conventional LES in the LES realm proper—flows driven by large-scale features. The performance of ILES in the substantially more difficult problem of under-resolved material mixing driven by under-resolved velocity fields and initial conditions is a focus of the present work. Progress in addressing relevant resolution issues in studies of mixing driven by Richtmyer–Meshkov inst...

Research paper thumbnail of High-Reynolds number Rayleigh–Taylor turbulence

Journal of Turbulence, 2009

2006. A comprehensive study of the budgets for the kinetic energy, mass flux, and density specifi... more 2006. A comprehensive study of the budgets for the kinetic energy, mass flux, and density specific volume covariance equations is undertaken. It is found that only the large scale quantities, but not the small scale quantities, reach self-similarity. Hypotheses for the variable density turbulent transport necessary to close the second moment equations are studied. The integral length scale does not follow thek 3/2 /ϵ scaling. This is due to the non-equilibrium nature of the flow and the fact that ϵ, a small scale quantity, does not have the self-similar scaling. As a consequence, the popular eddy viscosity expressionk 2 /ϵ does not model the turbulent transport in any of the moment equations. An integral length scale, based on the layer width, does scale the turbulent transport using a gradient transport hypothesis; that integral scale is a global quantity and does not lead to pointwise local closure. Despite the fact that the intermediate scales are nearly isotropic, the small scales have a persistent anisotropy; this is due to a cancellation between the viscous and nonlinear effects, so that the anisotropic buoyancy production remains important at the smallest scales. Various asymmetries in the mixing layer, not seen in the Boussinesq case, are also identified and explained.

Research paper thumbnail of Planar Richtmyer-Meshkov Instabilities and Transition

Journal of Physics: Conference Series, 2011

A numerical study of the evolution of the multimode planar Richtmyer-Meshkov (RM) instability in ... more A numerical study of the evolution of the multimode planar Richtmyer-Meshkov (RM) instability in a light-heavy (Air-SF6, Atwood number A = 0.67) configuration involving a Mach number Ma = 1.5 shock is carried out. Our results demonstrate that the initial material interface morphology controls the evolution RM characteristics, and provide a significant basis to develop metrics for transition to turbulence. Depending on the initial rms slope of the interface, RM evolves into linear or nonlinear regimes, with distinctly different flow features and growth rates, turbulence statistics and material mixing rates. We have called this the bipolar behavior of the RM instability. We demonstrate an important practical consequence of our results: reshock effects on mixing and transition can be emulated at first shock if the initial rms slope is high enough.

Research paper thumbnail of Probability density function method for variable-density pressure-gradient-driven turbulence and mixing

Probability density function (PDF) methods are extended to variable-density pressure-gradient-dri... more Probability density function (PDF) methods are extended to variable-density pressure-gradient-driven turbulence. We apply the new method to compute the joint PDF of density and velocity in a non-premixed binary mixture of different-density molecularly mixing fluids under gravity. The full time-evolution of the joint PDF is captured in the highly non-equilibrium flow: starting from a quiescent state, transitioning to fully developed

Research paper thumbnail of Turbulence in reshocked Richtmyer-Meshkov unstable fluid layers

ABSTRACT Advances in the implementation of high resolution PIV (150um vector-to-vector resolution... more ABSTRACT Advances in the implementation of high resolution PIV (150um vector-to-vector resolution) and PLIF (50um resolution) diagnostics have allowed the experimental measurement and characterization of turbulent mixing in Richtmyer-Meshkov unstable fluid layers after reshock (Balakumar et. al., Phys. Fluids, 2008). Using instantaneous PLIF data obtained at closely spaced intervals of time, we illustrate the rapid disintegration of the primary wavelengths of the initial interface and the beginning of a turbulence cascade generating smaller flow structures after reshock. The enhanced mixing is reflected in the variation of the density probability distribution function between the pre-reshock and post-reshock states. The density self-correlation is observed to exhibit a double-peaked structure and mild non-Boussinesq effects are observed in a layer with varicose initial interfacial perturbations. Density and velocity pdfs are used to examine the streamwise asymmetry of the mixing layer with large fluctuations occurring preferentially upstream of the centerline. Other turbulence statistics including the 2nd and 4th order structure functions, RMS statistics (both velocity and density) and turbulence intensity are also presented.

Research paper thumbnail of Variable-density mixing in buoyancy-driven turbulence

Journal of Fluid Mechanics, 2008

The homogenization of a heterogeneous mixture of two pure fluids with different densities by mole... more The homogenization of a heterogeneous mixture of two pure fluids with different densities by molecular diffusion and stirring induced by buoyancy-generated motions, as occurs in the Rayleigh–Taylor (RT) instability, is studied using direct numerical simulations. The Schmidt number, Sc, is varied by a factor of 20, 0.1 ≤ Sc ≤ 2.0, and the Atwood number, A, by a factor of 10, 0.05 ≤ A ≤ 0.5. Initial-density intensities are as high as 50% of the mean density. As a consequence of differential accelerations experienced by the two fluids, substantial and important differences between the mixing in a variable-density flow, as compared to the Boussinesq approximation, are observed. In short, the pure heavy fluid mixes more slowly than the pure light fluid: an initially symmetric double delta density probability density function (PDF) is rapidly skewed and, only at long times and low density fluctuations, does it relax to a Gaussian-like PDF. The heavy–light fluid mixing process asymmetry is...