Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “multiphase flows”

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 739 records · Page 41
All sourcesNASA NTRSDOE OSTI
DOE OSTI2014

Thickness-based adaptive mesh refinement methods for multi-phase flow simulations with thin regions

In numerical simulations of multi-scale, multi-phase flows, grid refinement is required to resolve regions with small scales. A notable example is liquid-jet atomization and subsequent droplet dynamics. It is essential to characterize the detailed flow physics with variable length scales with high fidelity, in order to elucidate the underlying mechanisms. In this paper, two thickness-based mesh refinement schemes are developed based on distance- and topology-oriented criteria for thin regions with confining wall/plane of symmetry and in any situation, respectively. Both techniques are implemented in a general framework with a volume-of-fluid formulation and an adaptive-mesh-refinement capability. The distance-oriented technique compares against a critical value, the ratio of an interfacial cell size to the distance between the mass center of the cell and a reference plane. The topology-oriented technique is developed from digital topology theories to handle more general conditions. The requirement for interfacial mesh refinement can be detected swiftly, without the need of thickness information, equation solving, variable averaging or mesh repairing. The mesh refinement level increases smoothly on demand in thin regions. The schemes have been verified and validated against several benchmark cases to demonstrate their effectiveness and robustness. These include the dynamics of colliding droplets, droplet motions in a microchannel, and atomization of liquid impinging jets. Overall, the thickness-based refinement technique provides highly adaptive meshes for problems with thin regions in an efficient and fully automatic manner.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2023

MACHINE LEARNING-ENABLED PREDICTION OF TRANSIENT INJECTION MAP IN AUTOMOTIVE INJECTORS WITH UNCERTAINTY QUANTIFICATION

Accurate prediction of injection profiles is a critical aspect of linking injector operation with engine performance and emissions. However, highly resolved injector simulations can take one to two weeks of wall-clock time, which is incompatible with engine design cycles with desired turnaround times of less than a day. Hence, it is important to reduce the time-to-solution of the internal flow simulations by several orders of magnitude to make it compatible with engine simulations. This work demonstrates a data-driven approach for tackling the computational overhead of injector simulations, whereby the transient injection profiles are emulated for a side-oriented, single-hole diesel injector using a Bayesian machine-learning framework. First, an interpretable Bayesian learning strategy was employed to understand the effect of design parameters on the total void fraction field. Then, autoencoders are utilized for efficient dimensionality reduction of the flowfields. Gaussian process models are finally used to predict the spatiotemporal void fraction field at the injector exit for unknown operating conditions. The Gaussian process models produce principled uncertainty estimates associated with the emulated flowfields, which provide the engine designer with valuable information of where the data-driven predictions can be trusted in the design space. The Bayesian flowfield predictions are compared with the corresponding predictions from a deep neural network, which has been transfer-learned from static needle simulations from a previous work by the authors. The emulation framework can predict the void fraction field at the exit of the orifice within a few seconds, thus achieving a speed-up factor of up to 38 x 10(6) over the traditional simulation-based approach of generating transient injection maps.

machine learning↗
DOE OSTI2014

An efficient algorithm for incompressible N-phase flows

We present an efficient algorithm within the phase field framework for simulating the motion of a mixture of N (N⩾2) immiscible incompressible fluids, with possibly very different physical properties such as densities, viscosities, and pairwise surface tensions. The algorithm employs a physical formulation for the N-phase system that honors the conservations of mass and momentum and the second law of thermodynamics. We present a method for uniquely determining the mixing energy density coefficients involved in the N-phase model based on the pairwise surface tensions among the N fluids. Our numerical algorithm has several attractive properties that make it computationally very efficient: (i) it has completely de-coupled the computations for different flow variables, and has also completely de-coupled the computations for the (N−1) phase field functions; (ii) the algorithm only requires the solution of linear algebraic systems after discretization, and no nonlinear algebraic solve is needed; (iii) for each flow variable the linear algebraic system involves only constant and time-independent coefficient matrices, which can be pre-computed during pre-processing, despite the variable density and variable viscosity of the N-phase mixture; (iv) within a time step the semi-discretized system involves only individual de-coupled Helmholtz-type (including Poisson) equations, despite the strongly-coupled phase–field system of fourth spatial order at the continuum level; (v) the algorithm is suitable for large density contrasts and large viscosity contrasts among the N fluids. Extensive numerical experiments have been presented for several problems involving multiple fluid phases, large density contrasts and large viscosity contrasts. In particular, we compare our simulations with the de Gennes theory, and demonstrate that our method produces physically accurate results for multiple fluid phases. We also demonstrate the significant and sometimes dramatic effects of the gravity, density ratios, pairwise surface tensions, and drop sizes on the N-phase configurations and dynamics. The numerical results show that the method developed herein is capable of dealing with N-phase systems with large density ratios, large viscosity ratios, and pairwise surface tensions, and that it can be a powerful tool for studying the interactions among multiple types of fluid interfaces.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2019

Effect of Ladle Changeover Condition on Transient Three-Phase Flow in a Five-Strand Bloom Casting Tundish

The transient three-phase flow behavior in a five-strand bloom casting tundish during the ladle changeover process considering a constant casting speed was investigated by the VOF model, and the effect of ladle throughput rate, shroud immersion depth, turbulence inhibitor structure and ladle changeover time on the turbulent flow, and level fluctuations during the filling process were analyzed to optimize the flow pattern in the tundish during ladle changeover. The results showed that the exposed area of molten steel in the impact zone increases as the ladle throughput rate increases during the filling process and that the exposure of molten steel is eliminated when the ladle throughput rate is 1.5Q{sub steady}. As the immersion depth of the shroud increases, a slight decrease in slag entrainment happens, while the exposed area of molten steel first decreases and then increases. The shroud immersion depth should be 310 mm. When the optimized turbulence inhibitor is used, the wave crests of the steel level are greatly reduced, the height of the level fluctuation is alleviated with the maximum value decreasing from 33 mm to approximately 10 mm, and the exposure of molten steel is eliminated. When the ladle changeover time is between 2 and 3 minutes, the exposed area of molten steel during the stable period of the filling process is in the range from 150 to 170 cm{sup 2}, which is substantially less than the exposed area of 500 cm{sup 2} when the ladle changeover time is 4 minutes. Therefore, the ladle changeover time should be controlled within 3 minutes. The flaw detection disqualification rate caused by ladle changeover operation is decreased from 43.8 to 25.8 pct of the total disqualified steel rails after optimization.

36 MATERIALS SCIENCE↗
DOE OSTI2016

Pore-scale displacement mechanisms as a source of hysteresis for two-phase flow in porous media

The macroscopic description of the hysteretic behavior of two–phase flow in porous media remains a challenge. It is not obvious how to represent the underlying pore–scale processes at the Darcy–scale in a consistent way. Darcy–scale thermodynamic models do not completely eliminate hysteresis and our findings indicate that the shape of displacement fronts is an additional source of hysteresis that has not been considered before. This is a shortcoming because effective process behavior such as trapping efficiency of CO 2 or oil production during water flooding are directly linked to pore–scale displacement mechanisms with very different front shape such as capillary fingering, flat frontal displacement, or cluster growth. Here we introduce fluid topology, expressed by the Euler characteristic of the nonwetting phase (χn), as a shape measure of displacement fronts. Using two high–quality data sets obtained by fast X–ray tomography, we show that χn is hysteretic between drainage and imbibition and characteristic for the underlying displacement pattern. In a more physical sense, the Euler characteristic can be interpreted as a parameter describing local fluid connectedness. It may provide the closing link between a topological characterization and macroscopic formulations of two–phase immiscible displacement in porous rock. Since fast X–ray tomography is currently becoming a mature technique, we expect a significant growth in high–quality data sets of real time fluid displacement processes in the future. As a result, the novel measures of fluid topology presented here have the potential to become standard metrics needed to fully explore them.

42 ENGINEERING↗
DOE OSTI2023

Pressure fields produced by single-bubble collapse near a corner

Damage produced by repeated bubble collapse to neighboring rigid objects in hydraulic systems is an important consequence of cavitation. Although bubble collapse near a single wall has received significant attention in the past, few studies exist on the dynamics of bubbles collapsing near a corner, i.e., two flat rigid surfaces intersecting at a right angle. In this work we quantify the pressure fields produced by a single bubble collapsing near two perpendicular rigid walls. Using a high-order accurate shock- and interface-capturing method to solve the three-dimensional compressible Navier-Stokes equations for gas and liquid flows, we simulate the dynamics of a single bubble collapsing at different initial stand-off distances from the two walls. In contrast to a bubble collapsing near a single wall, the collapse of bubbles within a critical stand-off distance is not symmetric about the bisecting plane due to the interaction between the bubble and the second wall. The second wall affects the pressure produced during the collapse in the following ways: (i) For bubbles initially located sufficiently close to both walls, the reentrant jet produced during collapse no longer points in the direction normal to the closest wall but at an angle toward the corner, (ii) the part of the emitted shock with the highest amplitude propagates in line with the jet, and (iii) the bubble migrates in that same direction during its collapse with a dependence on the stand-off distance, consistent with predictions made using Kelvin impulse. The location of maximum pressure along the walls is measured for the different initial stand-off distances. Using acoustic arguments, we find a semiempirical relationship to predict the initial stand-off distances for which the maximum pressure occurs in the corner. Here we find that when the bubble is sufficiently close to equidistant from each boundary, the maximum pressure is observed in the corner due to the water-hammer and implosion shocks reflecting off the boundaries and intersecting in the corner. We also show that when bubbles are initially attached to either wall the wall pressure produced can be significantly increased compared to bubbles detached from either wall.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2026

MFC 5.0: An exascale many-physics flow solver

Many problems of interest in engineering, medicine, and the fundamental sciences rely on high-fidelity flow simulation, making performant computational fluid dynamics solvers a mainstay of the open-source software community. Previous work MFC 3.0 was made a published, documented, and open-source solver via Bryngelson et al. Comp. Phys. Comm. (2021) with numerous physical features, numerical methods, and scalable infrastructure. MFC 5.0 is a significant update to MFC 3.0, featuring a broad set of well-established and novel physical models and numerical methods, as well as the introduction of GPU and APU (or superchip) acceleration. Here, we exhibit state-of-the-art performance and ideal scaling on the first two exascale supercomputers, OLCF Frontier and LLNL El Capitan. Combined with MFC’s single-accelerator performance, MFC achieves exascale computation in practice, and achieved the largest-to-date public CFD simulation at 200 trillion grid points as a 2025 ACM Gordon Bell Prize finalist. New physical features include the immersed boundary method, N-fluid phase change, Euler–Euler and Euler–Lagrange sub-grid bubble models, fluid-structure interaction, hypo- and hyper-elastic materials, chemically reacting flow, two-material surface tension, magnetohydrodynamics (MHD), and more. Numerical techniques now represent the current state-of-the-art, including general relaxation characteristic boundary conditions, WENO variants, Strang splitting for stiff sub-grid flow features, and low Mach number treatments. Weak scaling to tens of thousands of GPUs on OLCF Summit and Frontier and LLNL El Capitan achieves efficiencies within 5% of ideal to over 90% of their respective system sizes. Strong scaling results for a 16-times increase in device count show parallel efficiencies over 90% on OLCF Frontier. MFC’s software stack has undergone further improvements, including continuous integration, which ensures code resilience and correctness through over 300 regression tests; metaprogramming, which reduces code length while maintaining performance portability; and code generation for computing chemical reactions

Computational fluid dynamics↗
DOE OSTI2015

Experimental study on nonmonotonicity of capillary desaturation curves in a 2-D pore-network

Immiscible displacement in a porous medium is important in many applications such as soil remediation and enhanced oil recovery. When gravitational forces are negligible, two-phase immiscible displacement at the pore level is controlled by capillary and viscous forces whose relative importance is quantified through the dimensionless capillary number Ca and the viscosity ratio M between liquid phases. Depending on the values of Ca and M, capillary fingering, viscous fingering, or stable displacement may be observed resulting in a variety of patterns affecting the phase entrapment. The Capillary Desaturation Curve (CDC), which represents the relationship between the residual oils saturation and Ca, is an important relation to describe the phase entrapment at a given Ca. In the present study, we investigate the CDC as influenced by the viscosity ratio. A comprehensive series of experiments using a high-resolution microscope and state-of-the-art micromodels were conducted. The CDCs were calculated and the effects of Ca and M on phase entrapments were quantified. The results show that CDCs are not necessarily monotonic for all M.

capillary desaturation curve↗
DOE OSTI2021

Open source implementation of glued sphere discrete element method and nonspherical biomass fast pyrolysis simulation

Here, in this research, a glued-sphere Discrete Element Method (DEM) method was implemented in the open-source, Computational Fluid Dynamics (CFD) software MFiX. The implementation was verified using a cylinder-wall collision and then validated by simulating the packing and fluidization of non-spherical particles. The validated code was applied to simulate fast pyrolysis of non-spherical biomass particles in a bubbling fluidized bed. The glued sphere occupancy ratio was proposed to quantify the quality of shape resolution using glued sphere. Shape resolution showed significant influence on the packing height in the simulation of particle packing and an occupancy ratio of 80 % was recommended. Its influence is minor in fully fluidized bed but can be 8 times higher in fixed fluidized bed. Three tested heat transfer models predicted similar yields of elongated biomass fast pyrolysis. The solver developed in this research can be used to simulate other multiphase reacting flows involving non-spherical particles.

42 ENGINEERING↗
DOE OSTI2022

Very small-scale, segregating-fluidized-bed experiments: A dataset for CFD-DEM validation and uncertainty quantification

IWe report fluidization experiments were conducted on a small scale and with a rapid response (short duration) to enable corresponding simulations at low-computational cost. Rise times are reported for four or fewer polyethylene particles (intruders) in an air-fluidized bed of ~5000 group D glass beads. Experimental inputs were completely characterized—particle properties, system dimensions and operating conditions—which is necessary for validating computational fluid mechanics (CFD)-discrete element method (DEM) including a comprehensive uncertainty quantification (UQ) analysis. Input uncertainties are reported as bounds or cumulative distribution functions of measured values. The staggering number of simulations required to complete a UQ analysis (~O[10 4 ] simulations corresponding to ~5 uncertain inputs) motivates this study. These segregating-bed experiments are designed to permit analogous CFD-DEM simulations to complete in less than a day on a single (~2.5 GHz) computational processor unit (CPU). Segregation times are reported for several operating conditions, intruder sizes, and initial configurations, providing a rich dataset for numerical model testing, validation and UQ.

42 ENGINEERING↗
DOE OSTI2018

Numerical study of multiscale compaction-initiated detonation

A multiscale model of heterogeneous condensed-phase explosives is examined computationally to determine the course of transient events following the application of a piston-driven stimulus. The model is a modified version of that introduced by Gonthier (Combust Sci Technol 175(9):1679–1709, 2003. https://doi.org/10.1080/00102200302373) in which the explosive is treated as a porous, compacting medium at the macro-scale and a collection of closely packed spherical grains capable of undergoing reaction and diffusive heat transfer at the meso-scale. A separate continuum description is ascribed to each scale, and the two scales are coupled together in an energetically consistent manner. Following piston-induced compaction, localized energy deposition at the sites of intergranular contact creates hot spots where reaction begins preferentially. Reaction progress at the macro-scale is determined by the spatial average of that at the grain scale. A parametric study shows that combustion at the macro-scale produces an unsteady detonation with a cyclical character, in which the lead shock loses strength and is overtaken by a stronger secondary shock generated in the partially reacted material behind it. Finally, the secondary shock in turn becomes the new lead shock and the process repeats itself.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2017

Ejecta transport, breakup and conversion

Here, we report experimental results from an initial study of reactive and nonreactive metal fragments—ejecta—transporting in vacuum, and in reactive and nonreactive gases. We postulate that reactive metal fragments ejected into a reactive gas, such as H 2 , will break up into smaller fragments in situations where they are otherwise hydrodynamically stable in a nonreactive gas such as He. To evaluate the hypothesis we machined periodic perturbations onto thin Ce and Zn coupons and then explosively shocked them to eject hot, micron-scale fragments from the perturbations. The ejecta masses were diagnosed with piezoelectric pressure transducers, and their transport in H 2 and He was imaged with visible and infrared (IR) cameras. Because Ce + H 2 → CeH 2 + ΔH, where ΔH is the enthalpy of formation, an observed increase of the relative IR (radiance) temperature T R between the Ce–H 2 and Ce–He gas systems can be used to estimate the amount of Ce that converts to CeH 2 . As a result, the experiments sought to determine whether dynamic chemical effects should be included in ejecta-transport models.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2017

Pore-scale capillary pressure analysis using multi-scale X-ray micromotography

Here, a multi-scale synchrotron-based X-ray microtomographic dataset of residually trapped air after gravity-driven brine imbibition was acquired for three samples with differing pore topologies and morphologies; image volumes were reconstructed with voxel sizes from 4.44 µm down to 0.64 µm. Capillary pressure distributions among the population of trapped ganglia were investigated by calculating interfacial curvature in order to assess the potential for remobilization of residually-trapped non-wetting ganglia due to differences in capillary pressure presented by neighbor ganglia. For each sample, sintered glass beads, Boise sandstone and Fontainebleau sandstone, sub-volumes with different voxel sizes were analyzed to quantify air/brine interfaces and interfacial curvatures and investigate the effect of image resolution on both fluid phase identification and curvature estimates. Results show that the method developed for interfacial curvature estimation leads to reliable capillary pressure estimates for gas ganglia. Higher resolution images increase confidence in curvature calculations, especially for the sandstone samples that display smaller gas-brine interfaces which are then represented by a higher number of voxels when imaged with a micron or sub-micron voxels size. The analysis of sub-volumes from the Boise and Fontainebleau dataset highlights the presence of a residually-trapped gas phase consisting of ganglia located in one or few pores and presenting significantly different capillary pressures, especially in the case of Fontainebleau sandstone. As a result, Ostwald ripening could occur, leading to gas transfer from ganglia with higher capillary pressure to surrounding ganglia with lower capillary pressures. More generally, at the pore-scale, most gas ganglia do present similar capillary pressures and Ostwald ripening would then not represent a major mechanism for residually-trapped gas transfer and remobilization.

54 ENVIRONMENTAL SCIENCES↗
DOE OSTI2023

Model-parallel Fourier neural operators as learned surrogates for large-scale parametric PDEs

Fourier neural operators (FNOs) are a recently introduced neural network architecture for learning solution operators of partial differential equations (PDEs), which have been shown to perform significantly better than comparable deep learning approaches. Once trained, FNOs can achieve speed-ups of multiple orders of magnitude over conventional numerical PDE solvers. However, due to the high dimensionality of their input data and network weights, FNOs have so far only been applied to two-dimensional or small three-dimensional problems. To remove this limited problem-size barrier, we propose a model-parallel version of FNOs based on domain-decomposition of both the input data and network weights. Here, we demonstrate that our model-parallel FNO is able to predict time-varying PDE solutions of over 2.6 billion variables on Perlmutter using up to 512 A100 GPUs and show an example of training a distributed FNO on the Azure cloud for simulating multiphase CO 2 dynamics in the Earth’s subsurface.

58 GEOSCIENCES↗
DOE OSTI2014

Verification of sub-grid filtered drag models for gas-particle fluidized beds with immersed cylinder arrays

The accuracy of coarse-grid multiphase CFD simulations of fluidized beds may be improved via the inclusion of filtered constitutive models. In our previous study (Sarkar et al., Chem. Eng. Sci., 104, 399-412), we developed such a set of filtered drag relationships for beds with immersed arrays of cooling tubes. Verification of these filtered drag models is addressed in this work. Predictions from coarse-grid simulations with the sub-grid filtered corrections are compared against accurate, highly-resolved simulations of full-scale turbulent and bubbling fluidized beds. The filtered drag models offer a computationally efficient yet accurate alternative for obtaining macroscopic predictions, but the spatial resolution of meso-scale clustering heterogeneities is sacrificed.

42 ENGINEERING↗
DOE OSTI2017

Constrained reduced-order models based on proper orthogonal decomposition

A novel approach is presented to constrain reduced-order models (ROM) based on proper orthogonal decomposition (POD). The Karush–Kuhn–Tucker (KKT) conditions were applied to the traditional reduced-order model to constrain the solution to user-defined bounds. The constrained reduced-order model (C-ROM) was applied and validated against the analytical solution to the first-order wave equation. C-ROM was also applied to the analysis of fluidized beds. Lastly, it was shown that the ROM and C-ROM produced accurate results and that C-ROM was less sensitive to error propagation through time than the ROM.

42 ENGINEERING↗