Search NASA⌕ Search

SEARCH · Search NASA

Results for “numerical diffusion”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 559 records · Page 31

The mechanism of particles transport induced by electrostatic perturbation in tokamak

Particle transport in tokamak devices due to wave-particle resonance induced diffusion is studied. The diffusion coefficient is derived both analytically using quasilinear theory, and numerically using a test particle code, and the obtained diffusion coefficient agrees with each other in its validity regime. Dependence of the diffusion coefficient on turbulence intensity, turbulence radial mode structures, and particle energy is investigated. It is found that the diffusion coefficient is proportional to the turbulence intensity, and the diffusion is maximized for E{sub t}≃T{sub i}, and k{sub r}Δ{sub 0}≃1. Here, E{sub t} is the test particle energy, T{sub i} is the thermal ion temperature, Δ{sub 0} is the distance between neighboring mode rational surfaces, and 1/k{sub r} is the half width of the fine radial mode structure on each rational surface.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗

The scaling of kinetic and transport behaviors in the solution-phase chemistry of a plasma–liquid interface

The reactions at a plasma–liquid interface often involve species such as the solvated electron or the hydroxyl radical, which initiate the reduction or oxidation of solution-phase reactants (so-called scavengers) or are consumed by their own second-order recombination. Here, the mathematical scaling of the reaction–diffusion equations at the interface is used to obtain a characteristic time that can be used to determine the transition from highly efficient scavenger reduction or oxidation to lower efficiencies due to transport limitations. The characteristic time (tc) is validated using numerical solutions of the reaction–diffusion equations. When the scavenger kinetics are faster than second-order recombination, this characteristic transition time scales proportionally with the scavenger diffusivity (Ds) and the square of the scavenger bulk concentration (SB) and inversely proportional to the electron flux (J) squared; that is, t c = D s S B 2 F 2 /J 2 , where F is Faraday’s constant. However, when the scavenger kinetics are comparable or slower than second-order recombination, this scaling does not hold. Extending this analysis to three-dimensional systems shows that the profile of the electron flux on the surface affects the spatial location where reactions are most effective. Finally, the assessment of the implications of these behaviors for the reactor design highlights how effectively controlling the electron flux and solution transport may be necessary to improve the efficiency of scavenger reactions.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

The effect of gravity modulation on thermosolutal convection in an infinite layer of fluid

The effect of time-periodic vertical gravity modulation on the onset of thermosolutal convection in an infinite horizontal layer with stress-free boundaries is investigated using Floquet theory for the linear stability analysis. Situations for which the fluid layer is stably stratified in either the fingering or diffusive regimes of double-diffusive convection are considered. Results are presented both with and without steady background acceleration. Modulation may stabilize an unstable base solution or destabilize a stable base solution. In addition to synchronous and subharmonic response to the modulation frequency, instability in the double diffusive system can occur via a complex conjugate mode. In the diffusive regime, where oscillatory onset occurs in the unmodulated system, regions of resonant instability occur and exhibit strong coupling with the unmodulated oscillatory frequency. The response to modulation of the fundamental instability of the unmodulated system is described both analytically and numerically; in the double-diffusive system this mode persists under subcritical conditions as a high-frequency lobe.

Saunders, B. V.↗

The Cool Flames Experiment

A space-based experiment is currently under development to study diffusion-controlled, gas-phase, low temperature oxidation reactions, cool flames and auto-ignition in an unstirred, static reactor. At Earth's gravity (1g), natural convection due to self-heating during the course of slow reaction dominates diffusive transport and produces spatio-temporal variations in the thermal and thus species concentration profiles via the Arrhenius temperature dependence of the reaction rates. Natural convection is important in all terrestrial cool flame and auto-ignition studies, except for select low pressure, highly dilute (small temperature excess) studies in small vessels (i.e., small Rayleigh number). On Earth, natural convection occurs when the Rayleigh number (Ra) exceeds a critical value of approximately 600. Typical values of the Ra, associated with cool flames and auto-ignitions, range from 104-105 (or larger), a regime where both natural convection and conduction heat transport are important. When natural convection occurs, it alters the temperature, hydrodynamic, and species concentration fields, thus generating a multi-dimensional field that is extremely difficult, if not impossible, to be modeled analytically. This point has been emphasized recently by Kagan and co-workers who have shown that explosion limits can shift depending on the characteristic length scale associated with the natural convection. Moreover, natural convection in unstirred reactors is never "sufficiently strong to generate a spatially uniform temperature distribution throughout the reacting gas." Thus, an unstirred, nonisothermal reaction on Earth does not reduce to that generated in a mechanically, well-stirred system. Interestingly, however, thermal ignition theories and thermokinetic models neglect natural convection and assume a heat transfer correlation of the form: q=h(S/V)(T(bar) - Tw) where q is the heat loss per unit volume, h is the heat transfer coefficient, S/V is the surface to volume ratio, and (T(bar) - Tw ) is the spatially averaged temperature excess. This Newtonian form has been validated in spatially-uniform, well-stirred reactors, provided the effective heat transfer coefficient associated with the unsteady process is properly evaluated. Unfortunately, it is not a valid assumption for spatially-nonuniform temperature distributions induced by natural convection in unstirred reactors. "This is why the analysis of such a system is so difficult." Historically, the complexities associated with natural convection were perhaps recognized as early as 1938 when thermal ignition theory was first developed. In the 1955 text "Diffusion and Heat Exchange in Chemical Kinetics", Frank-Kamenetskii recognized that "the purely conductive theory can be applied at sufficiently low pressure and small dimensions of the vessel when the influence of natural convection can be disregarded." This was reiterated by Tyler in 1966 and further emphasized by Barnard and Harwood in 1974. Specifically, they state: "It is generally assumed that heat losses are purely conductive. While this may be valid for certain low pressure slow combustion regimes, it is unlikely to be true for the cool flame and ignition regimes." While this statement is true for terrestrial experiments, the purely conductive heat transport assumption is valid at microgravity (mu-g). Specifically, buoyant complexities are suppressed at mu-g and the reaction-diffusion structure associated with low temperature oxidation reactions, cool flames and auto-ignitions can be studied. Without natural convection, the system is simpler, does not require determination of the effective heat transfer coefficient, and is a testbed for analytic and numerical models that assume pure diffusive transport. In addition, mu-g experiments will provide baseline data that will improve our understanding of the effects of natural convection on Earth.

Pearlman, Howard↗

Bernstein dual-Petrov–Galerkin method: application to 2D time fractional diffusion equation

In this paper, we develop a Bernstein dual-Petrov–Galerkin method for the numerical simulation of a two-dimensional fractional diffusion equation. A spectral discretization is applied by introducing suitable combinations of dual Bernstein polynomials as the test functions and the Bernstein polynomials as the trial ones. We derive the exact sparse operational matrix of differentiation for the dual Bernstein basis which provides a matrix-based approach for the spatial discretization. It is shown that the method leads to banded linear systems that can be solved efficiently. The stability and convergence of the proposed method is discussed. Finally, some numerical examples are provided to support the theoretical claims and to show the accuracy and efficiency of the method.

97 MATHEMATICS AND COMPUTING↗

A mass-transfer particle-tracking method for simulating transport with discontinuous diffusion coefficients

The problem of a spatially discontinuous diffusion coefficient (D(x)) is one that may be encountered in hydrogeologic systems due to natural geological features or as a consequence of numerical discretization of flow properties. To date, mass-transfer particle-tracking (MTPT) methods, a family of Lagrangian methods in which diffusion is jointly simulated by random walk and diffusive mass transfers, have been unable to solve this problem. This manuscript presents a new mass-transfer (MT) algorithm that enables MTPT methods to accurately solve the problem of discontinuous D(x). To achieve this, we derive a semi-analytical solution to the discontinuous D(x) problem by employing a predictor-corrector approach, and we use this semi-analytical solution as the weighting function in a reformulated MT algorithm. Furthermore, this semi-analytical solution is generalized for cases with multiple 1D interfaces as well as for 2D cases, including a 2 × 2 tiling of 4 subdomains that corresponds to a numerically-generated diffusion field. The solutions generated by this new mass-transfer algorithm closely agree with an analytical 1D solution or, in more complicated cases, trusted numerical results, demonstrating the success of our proposed approach.

42 ENGINEERING↗

A PARAMETER STUDY FOR BAROCLINIC VORTEX AMPLIFICATION

Recent studies have shown that baroclinic vortex amplification is strongly dependent on certain factors, namely, the global entropy gradient, the efficiency of thermal diffusion and/or relaxation as well as numerical resolution. We conduct a comprehensive study of a broad range and combination of various entropy gradients, thermal diffusion and thermal relaxation timescales via local shearing sheet simulations covering the parameter space relevant for protoplanetary disks. We measure the Reynolds stresses as a function of our control parameters and see that there is angular momentum transport even for entropy gradients as low as {beta} = -dln s/dln r = 1/2. Values we expect in protoplanetary disks are between {beta} = 0.5-2.0 The amplification-rate of the perturbations, {Gamma}, appears to be proportional to {beta}{sup 2} and thus proportional to the square of the Brunt-Vaeisaelae frequency ({Gamma}{proportional_to}{beta}{sup 2}{proportional_to}N {sup 2}). The saturation level of Reynolds stresses, on the other hand, seems to be proportional to {beta}{sup 1/2}. This highlights the importance of baroclinic effects even for the low entropy gradients expected in protoplanetary disks.

79 ASTRONOMY AND ASTROPHYSICS↗

The effect of saturated thermal conduction on clouds in a hot plasma

We numerically investigate the internal evolution of multiphase clouds, which are at rest with respect to an ambient, highly ionized medium (HIM) representing the hot component of the circumgalactic medium. Time-dependent saturated thermal conduction and its implications like condensation rates and mixing efficiency are assessed in multiphase clouds. Our simulations are carried out by using the adaptive mesh refinement code FLASH. The model clouds are initially in both hydrostatic and thermal equilibrium and are in pressure balance with the HIM. Thus, they have steep gradients in both temperature and density at the interface to HIM leading to non-negligible thermal conduction. Several physical processes are considered numerically or semi-analytically: thermal conduction, radiative cooling and external heating of gas, self-gravity, mass diffusion, and dissociation of molecules and ionization of atoms. It turns out that saturated thermal conduction triggers a continuous condensation irrespective of cloud mass. Dynamical interactions with ambient HIM all relate to the radial density gradient in the clouds: (1) mass flux due to condensation is the higher the more homogeneous the clouds are; (2) mixing of condensed gas with cloud gas is easier in low-mass clouds, because of their shallower radial density gradient; and thus (3) accreted gas is distributed more efficiently. A distinct and sub-structured transition zone forms at the interface between cloud and HIM, which starts at smaller radii and is much narrower as deduced from analytical theory.

79 ASTRONOMY AND ASTROPHYSICS↗

Study of the Stability of 3C-SiC Single Crystals Using High-Resolution Diffuse X-Ray Scattering

The stability of (001)-oriented 3C silicon carbide crystals is studied by a method coupling high resolution x-ray diffraction and numerical simulations. The analysis of the diffuse scattering intensity distribution along selected directions in reciprocal space allows us to obtain qualitative and quantitative informations regarding the 3C-6H transition. Our latest results concerning the influence of the initial crystal quality (presence of defects) and of annealing time on the 3C-6H transition are presented in this article.

36 MATERIALS SCIENCE↗

Numerical Modeling of Flow Control in a Boundary-Layer-Ingesting Offset Inlet Diffuser at Transonic Mach Numbers

This paper will investigate the validation of the NASA developed, Reynolds-averaged Navier-Stokes (RANS) flow solver, OVERFLOW, for a boundary-layer-ingesting (BLI) offset (S-shaped) inlet in transonic flow with passive and active flow control devices as well as a baseline case. Numerical simulations are compared to wind tunnel results of a BLI inlet experiment conducted at the NASA Langley 0.3-Meter Transonic Cryogenic Tunnel. Comparisons of inlet flow distortion, pressure recovery, and inlet wall pressures are performed. The numerical simulations are compared to the BLI inlet data at a free-stream Mach number of 0.85 and a Reynolds number of approximately 2 million based on the fanface diameter. The numerical simulations with and without tunnel walls are performed, quantifying tunnel wall effects on the BLI inlet flow. A comparison is made between the numerical simulations and the BLI inlet experiment for the baseline and VG vane cases at various inlet mass flow rates. A comparison is also made to a BLI inlet jet configuration for varying actuator mass flow rates at a fixed inlet mass flow rate. Overall, the numerical simulations were able to predict the baseline circumferential flow distortion, DPCP avg, very well within the designed operating range of the BLI inlet. A comparison of the average total pressure recovery showed that the simulations were able to predict the trends but had a negative 0.01 offset when compared to the experimental levels. Numerical simulations of the baseline inlet flow also showed good agreement with the experimental inlet centerline surface pressures. The vane case showed that the CFD predicted the correct trends in the circumferential distortion levels for varying inlet mass flow but had a distortion level that was nearly twice as large as the experiment. Comparison to circumferential distortion measurements for a 15 deg clocked 40 probe rake indicated that the circumferential distortion levels are very sensitive to the symmetry of the flow and that a misalignment of the vanes in the experiment could have resulted in this difference. The numerical simulations of the BLI inlet with jets showed good agreement with the circumferential inlet distortion levels for a range of jet actuator mass flow ratios at a fixed inlet mass flow rate. The CFD simulations for the jet case also predicted an average total pressure recovery offset that was 0.01 lower than the experiment as was seen in the baseline. Comparisons of the flow features for the jet cases revealed that the CFD predicted a much larger vortex at the engine fan-face when compare to the experiment.

Allan, Brian G.↗

Numerical Modeling of Flow Control in a Boundary-Layer-Ingesting Offset Inlet Diffuser at Transonic Mach Numbers

This paper will investigate the validation of a NASA developed, Reynolds-averaged Navier-Stokes (RANS) flow solver, OVERFLOW, for a boundary-layer-ingesting (BLI) offset (S-shaped) inlet in transonic flow with passive and active flow control devices as well as the baseline case. Numerical simulations are compared to wind tunnel results of a BLI inlet conducted at the NASA Langley 0.3-Meter Transonic Cryogenic Tunnel. Comparisons of inlet flow distortion, pressure recovery, and inlet wall pressures are performed. The numerical simulations are compared to the BLI inlet data at a freestream Mach number of 0.85 and a Reynolds number of approximately 2 million based on the length of the fan-face diameter. The numerical simulations with and without wind tunnel walls are performed, quantifying effects of the tunnel walls on the BLI inlet flow measurements. The wind tunnel test evaluated several different combinations of jet locations and mass flow rates as well as a vortex generator (VG) vane case. The numerical simulations will be performed on a single jet configuration for varying actuator mass flow rates at a fix inlet mass flow condition. Validation of the numerical simulations for the VG vane case will also be performed for varying inlet mass flow rates. Overall, the numerical simulations were able to predict the baseline circumferential flow distortion, DPCPavg, very well for comparisons made within the designed operating range of the BLI inlet. However the CFD simulations did predict a total pressure recovery that was 0.01 lower than the experiment. Numerical simulations of the baseline inlet flow also showed good agreement with the experimental inlet centerline surface pressures. The vane case showed that the CFD predicted the correct trends in the circumferential distortion for varying inlet mass flow but had a distortion level that was nearly twice as large as the experiment. Comparison to circumferential distortion measurements for a 15 deg clocked 40 probe rake indicated that the circumferential distortion levels are very sensitive to the symmetry of the flow and that a miss alignment of the vanes in the experiment could have resulted in this difference. The numerical simulations of the BLI inlet with jets showed good agreement with the circumferential inlet distortion levels for a range of jet actuator mass flow ratios at a fixed inlet mass flow rate. The CFD simulations for the jet case also predicted an average total pressure recovery that was 0.01 lower than the experiment as was seen in the baseline. Comparison of the flow features the jet case revealed that the CFD predicted a much larger vortex at the engine fan-face when compare to the experiment.

Allan Brian G.↗

Hybrid Particle Method for Computational Shock Physics

A long-standing area of research for Eulerian shock wave physics codes has been the treatment of strength and damage for materials. Here we present a method that will aid in the analysis of strength and failure in shock physics applications where excessive diffusion of critical variables can occur and control the solution outcome. Eulerian methods excel for large deformation simulations in general but are inaccurate in capturing structural behavior. Lagrangian methods provide better structural response, but finite element meshes can become tangled. Therefore, a technique for merging Lagrangian and Eulerian treatments of material response, within a single numerical framework, was implemented in the Multiple Component computational shock physics hydrocode. The capability is a Lagrangian/Eulerian Particle Method (LEPM) that uses particles to interface a Lagrangian treatment of material strength with a more traditional Eulerian treatment of the Equation of State (EOS). Lagrangian numerical methods avoid the advection diffusion found in Eulerian methods, which typically strongly affects strength constitutive law internal variables, such as equivalent plastic strain, porosity and/or damage. The Lagrangian capability enhances existing capabilities and permits accurate predictions of high rate, large deformation and/or shock of mechanical structures.

97 MATHEMATICS AND COMPUTING↗

A high-resolution X-ray and optical study of SN 1006: asymmetric expansion and small-scale structure in a type IA supernova remnant

We introduce a deep (670 ks) X-ray survey of the entire SN 1006 remnant from the Chandra X-Ray Observatory, together with a deep Hα image of SN 1006 from the 4 m Blanco telescope at CTIO. Comparison with Chandra images from 2003 gives the first measurement of the X-ray proper motions around the entire periphery, carried out over a 9 yr baseline. We find that the expansion velocity varies significantly with azimuth. The highest velocity of ∼7400 km s{sup –1} (almost 2.5 times that in the northwest (NW)) is found along the southeast (SE) periphery, where both the kinematics and the spectra indicate that most of the X-ray emission stems from ejecta that have been decelerated little, if at all. Asymmetries in the distribution of ejecta are seen on a variety of spatial scales. Si-rich ejecta are especially prominent in the SE quadrant, while O and Mg are more uniformly distributed, indicating large-scale asymmetries arising from the explosion itself. Neon emission is strongest in a sharp filament just behind the primary shock along the NW rim, where the pre-shock density is highest. Here the Ne is likely interstellar, while Ne within the shell may include a contribution from ejecta. Within the interior of the projected shell we find a few isolated 'bullets' of what appear to be supernova ejecta that are immediately preceded by bowshocks seen in Hα, features that we interpret as ejecta knots that have reached relatively dense regions of the surrounding interstellar medium, but that appear in the interior in projection. Recent three-dimensional hydrodynamic models for Type Ia supernovae display small-scale features that strongly resemble the ones seen in X-rays in SN 1006; an origin in the explosion itself or from subsequent hydrodynamic instabilities both remain viable options. We have expanded the search for precursor X-ray emission ahead of a synchrotron-dominated shock front, as expected from diffusive shock acceleration theory, to numerous regions along both the northeast and southwest rims of the shell. Our data require that a precursor be thinner than about 3'', and fainter than about 5% of the post-shock peak. These limits suggest that the magnetic field is amplified by a factor of seven or more in a narrow precursor region, promoting diffusive particle acceleration.

79 ASTRONOMY AND ASTROPHYSICS↗

A High-Resolution X-Ray and Optical Study of SN1006: Asymmetric Expansion and Small-Scale Structure in a Type Ia Supernova Remnant

We introduce a deep (670 ks) X-ray survey of the entire SN 1006 remnant from the Chandra X-Ray Observatory, together with a deep Hα image of SN 1006 from the 4 m Blanco telescope at CTIO. Comparison with Chandra images from 2003 gives the first measurement of the X-ray proper motions around the entire periphery, carried out over a 9 yr baseline. We find that the expansion velocity varies significantly with azimuth. The highest velocity of approx.7400 km/s (almost 2.5 times that in the northwest (NW)) is found along the southeast (SE) periphery, where both the kinematics and the spectra indicate that most of the X-ray emission stems from ejecta that have been decelerated little, if at all. Asymmetries in the distribution of ejecta are seen on a variety of spatial scales. Si-rich ejecta are especially prominent in the SE quadrant, while O and Mg are more uniformly distributed, indicating large-scale asymmetries arising from the explosion itself. Neon emission is strongest in a sharp filament just behind the primary shock along the NWrim, where the pre-shock density is highest. Here the Ne is likely interstellar, while Ne within the shell may include a contribution from ejecta. Within the interior of the projected shell we find a few isolated "bullets" of what appear to be supernova ejecta that are immediately preceded by bowshocks seen in Hα, features that we interpret as ejecta knots that have reached relatively dense regions of the surrounding interstellar medium, but that appear in the interior in projection. Recent three-dimensional hydrodynamic models for Type Ia supernovae display small-scale features that strongly resemble the ones seen in X-rays in SN 1006; an origin in the explosion itself or from subsequent hydrodynamic instabilities both remain viable options. We have expanded the search for precursor X-ray emission ahead of a synchrotron-dominated shock front, as expected from diffusive shock acceleration theory, to numerous regions along both the northeast and southwest rims of the shell. Our data require that a precursor be thinner than about 3, and fainter than about 5% of the post-shock peak. These limits suggest that the magnetic field is amplified by a factor of seven or more in a narrow precursor region, promoting diffusive particle acceleration.

galaxies: active↗

Spectrum of spin waves in cold polarized gases

The spin dynamics of cold polarized gases are investigated using the Boltzmann equation. The dispersion relation for spin waves (transverse component of the magnetic moment) and the spin diffusion coefficient of the longitudinal component of the magnetic moment are calculated without using fitting parameters. The spin wave frequency and the diffusion coefficient for rubidium atoms are estimated numerically.

74 ATOMIC AND MOLECULAR PHYSICS↗

Adaptive Sparse-Grid Stochastic Collocation Uncertainty Quantification Convergence for Multigroup Diffusion

Advanced methods in uncertainty quantification for numerical models in computational physics continue to gain acceptance in nuclear modeling. Previously, the efficiency of sparse-grid stochastic collocation in comparison with Monte Carlo for uncertainty quantification through convergence studies was demonstrated. An adaptive method for anisotropic sparse grid collocation and its convergence compared to previous efforts is considered. The physical system in consideration is a two-dimensional quarter-core reactor, consisting of 5 materials distributed in 121 regions. The two-group neutron diffusion criticality approximation is solved. Vacuum boundaries are applied on the top and right, and reflecting boundaries on the bottom and left. The material cross sections, neutron multiplication factors, and diffusion coefficients are potential sources of uncertainty. Uniformly-distributed uncertainty within 5% of the reference values is introduced for material properties. The distribution of the uncertain parameters make up the uncertainty space Γ included in R{sup N}, where N is the number of uncertain parameters. The input parameter uncertainties are independently distributed. As expected, the convergence of Monte Carlo on the benchmark as a function of computational solves is nearly linear and lethargic compared to the other methods. The static isotropic index set, in which each input dimension is treated with equal importance, converges much more efficiently than Monte Carlo, but lacks the benefit of emphasizing some dimensions over others. The anisotropic case is most efficient, where first-order sensitivity information was used to determine anisotropic weights for polynomials to use in the polynomial chaos expansion. Interestingly, the adaptive algorithm converges slightly less efficiently than the anisotropic case, while still performing better than the isotropic case. Because the adaptive case has to determine the importance of each input dimension, it occasionally predicts inaccurately and wastes computation solves compared to the more ideal anisotropic case. However, in circumstances when an ideal anisotropy is not understood a priori, the adaptive algorithm may be optimally efficient in representing the original model. Given the potential loss of efficiency when using a poorly-chosen anisotropic weighting, the adaptive algorithm may be an effective choice for initial uncertainty quantification in problems with a low-dimensionality input space. (authors)

22 GENERAL STUDIES OF NUCLEAR REACTORS↗

Low Mach number fluctuating hydrodynamics of multispecies liquid mixtures

We develop a low Mach number formulation of the hydrodynamic equations describing transport of mass and momentum in a multispecies mixture of incompressible miscible liquids at specified temperature and pressure, which generalizes our prior work on ideal mixtures of ideal gases [Balakrishnan et al., “Fluctuating hydrodynamics of multispecies nonreactive mixtures,” Phys. Rev. E 89 013017 (2014)] and binary liquid mixtures [Donev et al., “Low mach number fluctuating hydrodynamics of diffusively mixing fluids,” Commun. Appl. Math. Comput. Sci. 9(1), 47-105 (2014)]. In this formulation, we combine and extend a number of existing descriptions of multispecies transport available in the literature. The formulation applies to non-ideal mixtures of arbitrary number of species, without the need to single out a “solvent” species, and includes contributions to the diffusive mass flux due to gradients of composition, temperature, and pressure. Momentum transport and advective mass transport are handled using a low Mach number approach that eliminates fast sound waves (pressure fluctuations) from the full compressible system of equations and leads to a quasi-incompressible formulation. Thermal fluctuations are included in our fluctuating hydrodynamics description following the principles of nonequilibrium thermodynamics. We extend the semi-implicit staggered-grid finite-volume numerical method developed in our prior work on binary liquid mixtures [Nonaka et al., “Low mach number fluctuating hydrodynamics of binary liquid mixtures,” http://arxiv.org/abs/1410.2300 (2015)] and use it to study the development of giant nonequilibrium concentration fluctuations in a ternary mixture subjected to a steady concentration gradient. We also numerically study the development of diffusion-driven gravitational instabilities in a ternary mixture and compare our numerical results to recent experimental measurements [Carballido-Landeira et al., “Mixed-mode instability of a miscible interface due to coupling between Rayleigh–Taylor and double-diffusive convective modes,” Phys. Fluids 25, 024107 (2013)] in a Hele-Shaw cell. We find that giant nonequilibrium fluctuations can trigger the instability but are eventually dominated by the deterministic growth of the unstable mode, in both quasi-two-dimensional (Hele-Shaw) and fully three-dimensional geometries used in typical shadowgraph experiments.

42 ENGINEERING↗

A comparison between implicit and hybrid methods for the calculation of steady and unsteady inlet flows

Numerical simulation of steady and unsteady transonic diffuser flows using two different computer codes are discussed and compared with experimental data. The codes solve the Reynolds-averaged, compressible, Navier-Stokes equations using various turbulence models. One of the codes has been applied extensively to diffuser flows and uses the hybrid method of MacCormack. This code is relatively inefficient numerically. The second code, which was developed more recently, is fully implicit and is relatively efficient numerically. Simulations of steady flows using the implicit code are shown to be in good agreement with simulations using the hybrid code. Both simulations are in good agreement with experimental results. Simulations of unsteady flows using the two codes are in good qualitative agreement with each other, although the quantitative agreement is not as good as in the steady flow cases. The implicit code is shown to be eight times faster than the hybrid code for unsteady flow calculations and up to 32 times faster for steady flow calculations. Results of calculations using alternative turbulence models are also discussed.

Coakley, T. J.↗