NSG-D/ Connecting scales

[<eugene.leitl@xxx.xxxxxxxxxxxx.xxx]

From: "Fred Hapgood" <hapgood@zYXHmYBqWFnNA63Tgq_pV0oQgYqxTTYyGk7aWlgtLxA6X_kwrn1VJQDSoYTLmHMeN9qskh1bEP_50dyZ.yahoo.invalid>

And then "Fred Hapgood" <hapgood@zYXHmYBqWFnNA63Tgq_pV0oQgYqxTTYyGk7aWlgtLxA6X_kwrn1VJQDSoYTLmHMeN9qskh1bEP_50dyZ.yahoo.invalid> says:

To: nsg@Ue7OeWCA--vus1N3Mk070hXszkt4ecUgPA2XJY1RHhHrR1DLZkI5oAZcjE4lTkFWsEfTnq6V2w.yahoo.invalid 
Subject: BU CCS Seminars for the Fall Semester 
From: bruceb@IGs2lu29w2z4cQXrJLcxhBh7Pzy8GJRwcONMfaYC_6gs8ZWabc2O3EmEaMHXpllTMvL6H4Q.yahoo.invalid 
Reply-to: bruceb@IGs2lu29w2z4cQXrJLcxhBh7Pzy8GJRwcONMfaYC_6gs8ZWabc2O3EmEaMHXpllTMvL6H4Q.yahoo.invalid


Dear NSG List,

The Boston University Center for Computational has long had a 
weekly seminar during the academic year.  A central theme of the 
talks that are scheduled for the coming year is computational 
modeling that "connects the length scales."  In particular, we would 
like to focus on those algorithms of computational physics, 
chemistry and materials science that allow one (or, at least, 
endeavor to allow one) to "seamlessly connect" the microscopic, 
particulate world with macroscopic behavior.  I enclose below the 
current list of talks for the fall semester.  I hope that it is of interest 
to list members. Further information about the Center for 
Computational Science can be found at http://ccs.bu.edu.  

Sincerely,
Bruce Boghosian
BU Center for Computational Science



----------------------------------------------------------------------
- Friday, October 8, 1999, 1 pm Boston University Joint Center for
Computational Science and Condensed Matter Physics Seminar 
Physics
Research Building, Room 593 "Multiscale Simulations of Complex
Phenomena in Materials Physics" Efthimios Kaxiras Harvard 
University

     We will present recent work which attempts to establish links
     between the microscopic structure and dynamics of solids to 
the
     macroscopic phenomena observed under various conditions of
     external loading or growth. Examples will be drawn mostly from
     covalently bonded solids, where competition between brittle and
     ductile behavior is exhibited. First-principles or empirical
     methods are employed to describe the atomic scale structure 
and
     dynamics, while the connection to the large scale behavior is
     made through stochastic (kinetic Monte Carlo), finite-element
     (quasicontinuum), or phenomenological methods.

----------------------------------------------------------------------
- Wednesday, October 13, 1999, 12 noon Boston University Center 
for
Computational Science Seminar Physics Research Building, Room 
593
"Better Materials and Process Design through Connecting Length Scales"
Mark J. Biggs University of Surrey

     There is ever increasing pressure on industry to deliver more
     innovative functional materials in less time and with smaller
     budgets, and this trend is set to intensify in the coming
     millennium. The traditional approach to the design of materials
     and the processes that make them is time consuming and resource
     intensive due to its heavy reliance upon experiment. Approaches
     based solely upon experiment also limit innovation when dealing
     with complex materials such as colloids, emulsions, powders, and
     functional microporous solids. The pharmaceutical industry has
     gained immense power for innovation whilst maintaining reasonable
     delivery times by deploying information technology solutions such
     as molecular simulation and bioinformatics. A similar approach is
     being advocated for the design of functional complex materials,
     and the processes used in their manufacture.

     The key to innovative design of complex materials and any
     associated processes is to be able to link clearly and explicitly
     the underlying microscopic/mesoscopic details of the system to
     its behavior at the functional level (i.e. the level at which it
     is observed and used).  This explicit and clear connection of
     length scales can be achieved through use of explicit simulation
     and statistical physics. These two approaches have been deployed
     for many years now in the study of simple gases and liquids, and
     crystalline solids. The challenge is to extend these approaches
     to complex materials that are characterized by complex chemistry,
     geometry and topology, and several different length scales.

     I shall present a case for why explicit simulation and
     statistical physics should be used in conjunction with focused
     experiments to achieve more innovative but rapid and cheap design
     of new complex materials and associated processes. This shall be
     done via presentation of a number of case studies from recent
     work undertaken here at Surrey. The first is the determination of
     the active sites for the SCR of NOx in the presence of
     hydrocarbons for the Cu-ZSM-5 catalyst using the molecular
     simulation based Virtual Porous Solid method. The second case
     study is suspension deposition in porous media using the
     lattice-gas automata and smooth-particle hydrodynamic based
     Virtual Colloid method. The final case study is the flow and
     heaping of irregular shaped particulates using the granular
     dynamic based Virtual Granular Solid.

----------------------------------------------------------------------
- Friday, October 15, 1999, 12 noon Boston University Center for
Computational Science Seminar Physics Research Building, Room 593
"Direct Numerical Simulation of Turbulence on PC/Linux Clusters: Fact
or Fiction?" George Em Karniadakis Brown University

     Direct Numerical Simulation (DNS) of turbulence requires many CPU
     days and Gigabytes of memory. These requirements limit most DNS
     to using supercomputers, available at supercomputer centres. With
     the rapid development and low cost of PCs, PC clusters are
     evaluated as a viable low-cost option for scientific computing.
     Both low-end and high-end PC clusters, ranging from 2 to 128
     processors, are compared to a range of existing supercomputers,
     such as the IBM SP nodes, Silicon Graphics Origin 2000, Fujitsu
     AP3000 and Cray T3E. The comparison concentrates on CPU and
     communication performance. At the kernel level, \bf BLAS \rm
     libraries are used for CPU performance evaluation. Regarding
     communication, the free implementations of MPICH and LAM are used
     on fast-ethernet-based systems and compared to myrinet-based and
     supercomputer networks. At the application level, serial and
     parallel simulations are performed on state of the art DNS, such
     as turbulent wake flows in stationary and moving computational
     domains.

----------------------------------------------------------------------
- Friday, October 29, 1999, 12 noon Boston University Center for
Computational Science Seminar Physics Research Building, Room 593
"Theory of the Viscoelasticity of Gases and Simulations with a
Viscoelastic, Two-Component Lattice Boltzmann Method" Alexander Wagner
M.I.T. Department of Materials Science

     Lattice Boltzmann methods have proven useful for the simulation
     of fluid flows, especially in the area of multicomponent flows.
     We have studied the possibility of implementing viscoelastic
     lattice Boltzmann methods and have implemented a two component
     model. In this talk we present some theoretical consideration of
     the viscoelastic limit of the continuous Boltzmann equation and
     show simulation results for a bubble rising in a viscoelastic
     liquid and some new results in viscoelastic phase separation. In
     particular we have observed an anomalously slow growth regime,
     first observed experimentally by Tanaka, and seen
     non-universality in the scaling state.

----------------------------------------------------------------------
- Friday, November 12, 1999, 12 noon Boston University Joint Center
for Computational Science and Condensed Matter Physics Seminar Physics
Research Building, Room 593 "How is a Bose-Condensed System like a
Polymer Melt?" David Ceperley NCSA, Dept. of Physics University of
Illinois Urbana-Champaign

     Feynman(1953) introduced imaginary-time path integrals to
     understand superfluid 4He.  Path integrals are an exact
     "isomorphism" between quantum systems and the classical
     statistical mechanics of ring polymers.  Bose symmetry of the
     wave function implies that the polymers are allowed to
     ``cross-link'' or exchange.  The specific heat singularity is a
     consequence of this cross-linking, momentum condensation is
     related to the end-to-end distribution of a single open-ended
     polymer: if the ends become delocalized, the quantum system is
     bose condensed.  Superfluidity (coupling to the boundaries) is
     proportional to the mean squared flux of polymers through a
     surface.  Thus all three phenomena, specific heat, momentum
     distribution and superfluidity are directly related to
     macroscopic exchange.  We have developed specialized simulation
     methods(Path Integral Monte Carlo) based on the Metropolis Monte
     Carlo method, to simulate boson systems.

     Over the last few decades, there has been a search for new
     bose-condensed systems. One of the likely candidates is molecular
     para-hydrogen, a boson with half the mass of helium which is
     normally a solid at low temperature.  Based on detailed
     simulation, we predict a monolayer of molecular hydrogen will
     undergo a Kosterlitz-Thouless superfluid transition below
     approximately 1.2K if it is placed on a clean silver surface to
     which alkali metal atoms have also been absorbed.

     The methods are generalizable to other quantum system, and with
     difficulty to fermion systems such as electrons.

----------------------------------------------------------------------
- Friday, November 19, 1999, 12 noon Boston University Center for
Computational Science Seminar Physics Research Building, Room 593
"Simulated Spectroscopy of Guests Absorbed in Zeolitic Solids" Amy Bug
Department of Physics and Astronomy Swarthmore College

     The introduction of guest atoms into porous, host solids has
     applications to industrial processes like chemical catalysis, the
     storage of gaseous fuels, and the fabrication of materials with
     novel mechanical and optical properties.  Though some of these
     applications have been known for decades, detailed microscopic
     information on guest behavior has only recently begun to emerge.
     Much of the progress has been accomplished by a close interplay
     between experimental spectroscopies and computer simulation
     studies.  This talk will discuss two particularly versatile
     guests, molecular hydrogen and positronium, within
     aluminosilicate (zeolite) solids.  In our lab, Pseudospectral
     methods and Path Integral simulations allow us to simulate the
     results of optical, neutron, and positron annihilation
     spectroscopies.

----------------------------------------------------------------------
-

------- End of forwarded message -------
www.pobox.com/~fhapgood
- For info on unsubscribing, etc., please ask
- majordomo@7U88CzpcT3tyFmLFiVWrh-QMyvS2ZX8O_fAk-1CMu679u9xQGG3-b7CIF0V5doMCn5dYLWRob-H5tCnCqC7_waMU.yahoo.invalid for 'help'.