Stony Brook AMS - Downloadable Preprints - 2006
2008 /
2007 /
2006 /
2005 /
2004 /
2003 /
2002 /
2001 /
2000
1999 /
1998 /
1997 /
1996 /
1995 /
1994 /
1993 /
1992 /
1991
Return to main preprint page
Return to AMS home page
Return to main preprint page
Return to AMS home page
Please send suggestions and complaints to
www@ams.sunysb.edu.
| SUNYSB-AMS-06-01 |
A Front Tracking Algorithm for Limited Mass Diffusion
|
Xinfeng Liu, Yuanhua Li, James Glimm, Xiaolin Li
|
The advantage of front tracking over capturing methods for
the numerical simulation of discontinuity surfaces in fluid flow is to
eliminate the numerical diffusion of mass across a fluid
interface. For accurate modeling of an interface between miscible fluids,
however, physical mass diffusion may be significant, and must
be added to the calculation. The main purpose of this paper is to introduce
a new front tracking
algorithm in which the physical diffusion of mass across a tracked interface
is included. The accuracy and
convergence properties of this algorithm are discussed.
The new algorithm is a subgrid algorithm
in the sense that the asymptotic fine grid behavior is
identical to conventional untracked (capturing) methods while
the coarse grid behavior is improved. It is thus most suitable for
modeling small values of physical mass diffusion, for which adequate numerical
resolution is computationally demanding.
The mixing rates for the 3D Rayleigh-Taylor instability of miscible
fluids based on this algorithm agree with experimental values.
Submitted to J. of Comp. Phys.
Available in:
PDF
Over the past decade a growing number of high resolution, three-dimensional,
X-ray computed tomography images of multiphase porous media systems have become
available, as have the number of algorithms to analyze both the pore space and
the fluid partitioning imaged within the voids.
Here we consider the measurement of fluid-fluid and fluid-solid surfaces
at residual fluid conditions in a Berea sandstone sample.
3D reconstructions of the residual fluid blobs show qualitative behaviour
which is in agreeement with conventional wisdom.
We consider the measurement of fluid-solid contact angle via such images.
We elucidate the challenge in identifying triple points (points with three
phases in contact) and describe an algorithm for contact angle measurement.
Results from application of the contact angle algorithm indicate the ability
to distinguish wetting and non-wetting fluid blobs and reveal satisfactory
agreement with experimental wettability measurements.
To appear in the proceedings of International Conference on Computational
Methods in Water Resources XVI, Copenhagen Denmark, June 19-22, 2006.
Available in:
PDF
| SUNYSB-AMS-06-03 |
Classification by Ensembles from Random Partitions
|
Hongshik Ahn, Hojin Moon, Melissa J. Fazzari, Noha Lim,
James J. Chen and Ralph L. Kodell
|
A robust classification procedure is developed based on ensembles of
classifiers, with each classifier constructed from a different set of
predictors determined by a random partition of the entire set of predictors.
The proposed method combines the results of multiple classifiers to achieve a
substantially improved prediction compared to the optimal single classifier.
This approach is designed specifically for high-dimensional data sets for which
a classifier is sought. By combining classifiers built from each subspace of
the predictors, the proposed methods achieve a huge computational advantage in
tackling the growing problem of dimensionality. We have shown empirically that
huge data sets need not be handled as a whole; the subspaces of the feature
space created through partitioning may be treated independently and separately
until after the classifiers are developed. For each subspace of the predictors,
we build a classification tree or logistic regression tree with an optimal size
yielding least cost in terms of misclassification errors. By a random
partition, a logistic regression can be used without losing the ensemble
accuracy for data with a huge number of available predictor variables and a
relatively small number of observations without a variable selection. We
investigate the performance of the proposed ensemble methods compared to widely
used classification methods including Random Forest, Support Vector Machines,
Boosting, Linear Discriminant Analysis, $k$-Nearest Neighbors and single
optimal trees using three real data sets from different areas. Our study shows
that the performance of our methods in terms of overall accuracy is consistently
good compared to the other classification methods considered in this study. For
unbalanced data, our approach maintains the balance between sensitivity and
specificity more adequately than many other classification methods. A primary
area of application is the classification of subjects into cancer-risk or
cancer-type categories based on high dimensional genomics or proteomics data.
It is anticipated that the proposed methods can be used to improve class
prediction in many other areas of application involving high dimensional
prediction sets.
Submitted for publication
Available in:
PDF
| SUNYSB-AMS-06-04 |
A Conservative Front Tracking Method in N-Dimensions
|
Jinjie Liu, Hyun-Kyun Lim, James Glimm and Xiaolin Li
|
We propose a fully conservative Front Tracking
algorithm for systems of nonlinear conservation laws.
The algorithm can be applied uniformly in one, two, three and N
dimensions. Implementation details for this algorithm and
tests of fully conservative simulations are reported.
Submitted for J. of Sci. Comp.
Available in:
PDF
| SUNYSB-AMS-06-05 |
Simulation vs. Theory vs. Experiment for Complex Fluid Mixing Flows
|
James Glimm and D. H. Sharp
|
It is an old saw that science has as its three pillars theory, experiment
and simulation. An ability to believe simulation when the three do not agree
is a testament to a newfound credulence placed on simulations. Our story
concerns widely different experimental and simulation results in a problem of
long standing: the determination of the rates of chaotic mixing resultig from
acceleration across a fluid density discontinuity, namely the classical
Rayleigh-Taylor (RT) instability.
Available in:
PDF
| SUNYSB-AMS-06-06 |
Front Tracking under TSTT
|
James Glimm, Brian Fix, Xiaolin Li, Jinjie Liu Xinfeng Liu,
Tishshi Lu, Roman Samulyak and Zhiliang Xu
|
We report several important developments of the front tracking method
and applications in science and engineering under the TSTT project.
The progress includes the extraction of an independent software library
from the front tracking code, conservative front tracking, applications of
front tracking to the simulation of fusion pellet injection in a magnetically
confined plasma, the study of a fuel injection jet, and the study of fluid
chaotic mixing, among other problems.
Submitted for ASP Conference Series.
Available in:
PDF
| SUNYSB-AMS-06-07 |
Dynamic Phase Boundaries for Compressible Fluids
|
T. Lu, Z. L. Xu, R. Samulyak, J. Glimm, X. M. Ji
|
We present an algorithm for the simulation of a generalized Riemann
problem for phase transitions in compressible fluids.
We model the transition as a tracked jump discontinuity. The emphasis here
is on the coupling of the phase transition
process to acoustic waves, which is required for the study of cavitation
induced by strong rarefaction waves. The robustness
of the proposed algorithm is verified by application to various physical
regimes.
Submitted to Siam J. on Scientific Computing
Available in:
PDF
Postscript
| SUNYSB-AMS-06-08 |
Entropy of Averaging for Compressible Two-Pressure Two-Phase Flow Models
|
H. Jin, J. Glimm, D. H. Sharp
|
We propose here a new closure
for compressible two-pressure two-phase flow models, which
satisfies conservation requirements, boundary conditions at the edges of the
mixing zone, hyperbolic stability (real eigenvalues for the
characteristic version of the equations of motion) and an entropy inequality.
Except for the latter, these properties are direct consequences of the
proposed closures. The entropy, which is the main focus of this paper, inequality (
as opposed to entropy conservation for
microphysically adiabatic processes)
implies positivity for the entropy of averaging.
Submitted to Phys. Lett. A
Available in:
PDF
Postscript
| SUNYSB-AMS-06-09 |
A Front Tracking Method for Multiscale Free Surface Compressible Flows
|
Z. Xu, J. Glimm, Y. Zhang, X. Liu
|
We present an overview of multiscale computations for free surface flows based
on the front tracking method.
Our approach combines theory, numerical algorithm development,
simulation based scientific studies, and the analysis of experimental data.
Submitted to Chemical Engineering Science Journal on complex systems
Available in:
PDF
| SUNYSB-AMS-06-10 |
Numerical Evaluation of Impact of Laser Preheat on Interface Structure and Instability
|
Y. Zhang, R. P. Drake, and J. Glimm
|
This paper presents a computational study of the impact of preheating, in advance of shock
heating, on a structured interface and on the subsequent postshock instability
evolution. The study was performed by applying a method, described previously, of evaluating radiative effects using a multidimensional, front-tracking hydrodynamic code with input from a one-dimensional, radiation-hydrodynamic code. The method is general and could be applied to a range of laser-driven shock experiments. Results of simulations are shown for both high and low levels of preheat, conducted
using a robust front tracking algorithm in the presence of a radiation energy source.
In the low-preheat case, which represents the minimum to be anticipated in
laboratory experiments, some impact of preheat on both preshock conditions and
post-shock evolution are observed. In the high-preheat case, which represents one potential result of preheating by increased radiation and/or energetic electrons, the preheat alters the spectral
content of the interface structure. In this case, before the shock reaches the interface, higher-order harmonic modes are induced, the interface position is shifted, and the perturbation amplitude is reduced. Furthermore, the post-shock evolution of the interface is affected by the amount of preheat and by whether radiative heating after the laser pulse is also included. Such a numerical assessment of preheating can be important to the design and analysis of laboratory experiments. The initial conditions for the interaction of any shock wave with structures in the target may be altered by the presence of preheating. This poses a challenge to the laser experimental study of fluid mixing. Numerical simulations can serve as a useful tool to guide decisions regarding control and/or measurement of this effect.
in press, Physics of Plasma
Available in:
PDF
| SUNYSB-AMS-06-11 |
Microstructure of Natural Hydrate Host Sediments
|
K.W. Jones
K.W. Jones,
P.B. Kerkar, D. Mahajan,
W.B. Lindquist,
and H. Feng
|
There is worldwide interest in the study of natural gas hydrate because of its
potential impact on world energy resources, control on seafloor stability,
significance as a drilling hazard, and probable impact on climate as a
reservoir of a major greenhouse gas.
Gas hydrates can:
(a) be free floating in the sediment matrix,
(b) contact, but not cement, existing sediment grains, or
(c) cement and stiffen the bulk sediment.
Seismic surveys, often used to prospect for hydrates over a large area,
can provide knowledge of the location of large hydrate concentrations because
the hydrate location within the sediment pores can have a profound influence on
its seismic properties.
The ability to image a sample at the grain scale and to determine the porosity,
permeability, and seismic profile is of great interest since these can help
determine the location of hydrates with certainty.
We report here on an investigation of the properties of methane hydrate
sediments at the grain-size scale using the synchrotron radiation-based
computed microtomgraphy (CMT) technique.
Work has started on the measurements of the changes occurring as
tetrahydrofuran hydrate, a surrogate for methane hydrate,
is formed in the sediment.
Proceedings of CAARI 2006: the 19'th International Conference on the Application
of Accelerators in Research and Industry, Aug. 20-25, 2006, Fort Worth, TX.
Available in:
Word
| SUNYSB-AMS-06-12 |
A Comparison Study of Two Methods for Elliptic Boundary Value Problems
|
J. Du, S. Wang, J. Glimm, R. Samulyak
|
In this paper, we compare two methods with regard to speed and accuracy
(the embedded boundary method and several versions of the mixed finite
element method) to solve an elliptic boundary value problem.
Submitted to SIAM Journal on Numerical Analysis
Available in:
PDF
| SUNYSB-AMS-06-13 |
Recent Progress in Turbulent Mixing
|
J. Glimm, X. Li
|
We present recent progress on turbulent mixing, jointly obtained with
collaborators. We have an improved version of front tracking, improved
physics models of Rayleigh-Taylor turbulent mixing, and agreement of
simulations with experiment for immiscible experiments without surfactants
and miscible fluids with mass diffusion. We present improved closures for
averaged equations. We use the experimentally validated simulations to
validate the closures in the averaged equations. Issues of verification are
also addressed.
10th IWPCTM - Paris (France) July 2006
Available in:
PDF
Return to main preprint page
Return to AMS home page