(February 1) Professor Donald Estep:
Fast and Reliable Methods for Determining the Evolution of
Uncertain Parameters in Differential Equations -
A very common problem in science and engineering is the
determination of the effects of uncertainty or variation in
parameters and data on the output of a nonlinear operator. For
example, such variations may describe the effect of experimental
error or may arise as part of a sensitivity analysis of the
model. The Monte-Carlo Method is a widely used tool for
understanding such effects that employs random sampling of the
input space in order to produce a pointwise representation of the
output. It is a robust and easily implemented tool.
Unfortunately, it generally requires sampling the operator very
many times at a significant cost. Moreover, it provides no robust
measure of the error of information computed from a particular
representation. In this paper, we present an alternative approach
for ascertaining the effects of variations and uncertainty in
parameters in a differential equation that is based on techniques
borrowed from a posteriori error analysis for finite element
methods. The generalized Green's function is used to describe how
variation propagates into the solution around localized points in
the parameter space. This information can be used either to
create a higher order method or produce an error estimate for
information computed from a given representation. In the latter
case, this provides the basis for adaptive sampling. Both the
higher order method and the adaptive sampling methods are
generally orders of magnitude faster than Monte-Carlo methods in
a variety of situations.
(February 8) Prof. Alexis S. Mahalov:
Global Regularity of the 3D Navier-Stokes Equations with
Uniformly Large Initial Vorticity
-
We prove existence on infinite time intervals of regular
solutions to the classical incompressible 3D Navier-Stokes
Equations for fully three-dimensional periodic and almost
periodic initial data characterized by uniformly large vorticity
in R^3 and in bounded cylindrical domains; smoothness assumptions
for initial data are the same as in local existence theorems.
There are no conditional assumptions on the properties of
solutions at later times, nor are the global solutions close to
any 2D manifold. The global existence is proven using techniques
of fast singular oscillating limits and the Littlewood-Paley
dyadic decomposition. The approach is based on the computation of
singular limits of rapidly oscillating operators and cancellation
of oscillations in the nonlinear interactions for the vorticity
field. With nonlinear averaging methods in the context of almost
periodic functions, we obtain fully 3D limit resonant
Navier-Stokes equations. We establish the global regularity of
the latter without any restriction on the size of 3D initial
data. With strong convergence theorems, we bootstrap this into
the global regularity of the 3D Navier-Stokes Equations with
uniformly large initial vorticity. We review applications of our
mathematical techniques to numerical analysis of highly
oscillatory PDE's arising in geophysical fluid dynamics. Global
regularity of the 3D Navier-Stokes Equations of Geophysics is
proven for all domain aspect ratios and all small Froude and
Rossby numbers. (February 15) Dr. Kyoung-Sook Moon:
Convergence rate of an adaptive algorithm for differential
equations -
The theory of a posteriori error estimates suitable for adaptive
refinement is well established. But the issue of convergence
rates of adaptive algorithms is less studied. I will present a
simple and general adaptive algorithm applied to ordinary and
stochastic differential equations with proven convergence rates.
The presentation has two parts: The error approximations used to
build error indicators for the adaptive algorithm are based on
error expansions with computable leading order terms. The
adaptive algorithm, performing successive mesh refinements,
either reduces the maximal error indicator by a factor or stops
with the error asymptotically bounded by the prescribed accuracy
requirement. Furthermore, the algorithm stops using the optimal
number of degrees of freedom, up to a problem independent factor.
(February 22) Dr. Soeren Bartels:
Approximation of Harmonic Maps -- Gradient Flow Approaches vs.
Iterative Minimization -
Harmonic maps are stationary points of the Dirichlet energy among
functions with values in the unit sphere. Owing to the nonconvex
constraint, harmonic maps are non-unique and fail to admit higher
regularity properties. Moreover, the constraint prohibits the use
of standard tools for the numerical approximation. In this talk
we discuss stability and weak convergence of three numerical
schemes. The first scheme consists in the minimization of the
Dirichlet energy over suitable tangent spaces and a
renormalization of the update in each iteration. The second
approach penalizes the constraint and leads to a time-dependent
Ginzburg-Landau equation. A projection method for the
discretization of the harmonic map heat flow is the basis for the
third approach. Besides stating sufficient conditions for
stability and weak convergence to an exact solution, we indicate
generalizations to the approximation of p-harmonic maps.
Applications include liquid crystal theory, image processing, and
micromagnetics. Part of this talk is based on joint work with J.W. Barrett, X.
Feng, and A. Prohl.
(March 1) Prof. John A. Pelesko:
Electrostatic-Elastic Interactions -
In 1968, in the context of investigating fundamental questions in
electrohydrodynamics, G.I. Taylor studied the electrostatic
deflection of elastic membranes. Utilizing soap film as the
membrane material and applying a fixed high voltage potential
difference between two supported circular membranes, Taylor
showed experimentally that at a critical voltage the two
membranes snap together and touch. That is, the equilibrium state
where the membranes remained separate that existed at smaller
voltages either became unstable or failed to exist. This
instability is familiar to researchers in the MEMS
(microelectromechanical systems) and NEMS (nanoelectromechanical
systems) fields where it is known as the "pull-in" instability.
In fact, in an interesting historical coincidence H.C. Nathanson
and his coworkers studied this instability in the context of a
primitive MEMS device at roughly the same time as Taylor was
conducting his studies. Nathanson is responsible for the
"pull-in" nomenclature and the analysis of a mass-spring model of
this effect. Taylor, in conjunction with R.C. Ackerberg
developed and numerically analyzed a more accurate membrane based
model of electrostatic deflection. Recently, a rigorous analysis
of this model was completed. Surprisingly, even this simple
model of electrostatic deflection contains a rich solution set
exhibiting a bifurcation diagram with infinitely many folds. In
this talk, we provide an overview of recent results on the
interaction of elastic membranes with electrostatic fields. We
discuss a re-creation of the Taylor experiment, some new
experimental results and discuss the relevance of this research
to MEMS and NEMS systems.
(March 15) Khamron Mekchay:
AFEM for the Laplace-Beltrami operator on graphs: a posteriori
error estimation and convergence -
We develop an adaptive finite element method (AFEM) to
approximately solve the Laplace-Beltrami operator on graphs
$\Gamma$ in $R^3$ which are described by a $C^1$, or piecewise
$C^1$, function on a polygonal domain in $R^2$. We first derive a
posteriori error estimators which account for both the
approximation of the surface by a polyhedral surface in $R^3$,
the geometric error, and the energy error incurred in solving the
PDE. We show that both errors are coupled, and that our AFEM
decreases their collective contribution, thereby becoming a
contraction eventually and yielding convergence. The proof of
convergence deals with the fact that the approximate surfaces are
no longer nested, as in the flat case, which in turn leads to a
novel quasi-orthogonality property in the energy norm. This is
joint work with P. Morin (IMAL, Argentina) and R.H. Nochetto.
(March 29) Simon P. Schurr:
Universal Duality in Conic Convex Optimization -
For a pair of dual convex optimization problems in conic form,
we provide simple necessary and sufficient conditions on the
``constraint matrices'' and cone under which a zero duality gap
occurs for \textit{every} linear objective function and constraint
``right-hand side''. (The optimal objective function values may
be both $+\infty$ or both $-\infty$.) We refer to this property
as ``universal duality''.
We show that universal duality is generic in both a metric sense and
a topological sense, and can be verified by solving a single conic
convex program. We illustrate our theory by studying a class of
semidefinite programs that appear in control theory and signal
processing applications. This talk is based on joint work with Andr\'e Tits and Dianne O'Leary. (April 05) Prof. Nick Trefethen:
Who invented the great numerical algorithms? -
Over coffee one day, a colleague suggested that most of
the famous algorithms of numerical analysis had their roots
with practitioners, not academics. This intrigued me --
could he be right? I decided to see what I could dig up
concerning the origins of thirty of the greatest hits of
our field. In this lecture I'll show you what I found.
(April 12) Prof. Mark Embree:
A Dozen Cautionary Tales from "Spectra and Pseudospectra"
-
This past summer Nick Trefethen and I completed "Spectra and
Pseudospectra," a book describing the theory and applications
of nonnormal matrices and operators. In this talk I will
extract a dozen of favorite examples that show how eigenvalues
can provide misleading information about system behavior.
This tour will begin with Toeplitz matrices and differential
operators, then focus on the performance of numerical algorithms
(for the solution of linear systems, eigenvalue problems,
and differential equations), and conclude with applications
in fluid dynamics and population ecology.
(April 19) Prof. Alfred Schmidt:
Adaptive Finite Element Simulations for Macroscopic and Mesoscopic Models of Steel -
During heat treatment of a steel workpiece and other production processes, fine
structures like boundary layers for temperature, concentration, or phase
fraction fields appear, while in other parts of the work piece the behaviour of
these fields is quite smooth. Nevertheless, it is important to capture these
structures during numerical simulations.
Local mesh refinements in these regions is needed in order to resolve the
behaviour in a sufficient way. On the other hand, these regions of special
interest are changing during the process, making it necessary to move also the
regions of refined meshes.
Adaptive finite element methods present a tool to automatically give criteria
for a local mesh refinement, based on the computed solution (and not only on a
priori knowledge of an expected behaviour).
We will present examples from heat treatment of steel, including phase
transitions with transformation induced plasticity.
On a mesoscopic scale of grains, similar methods can be used to efficiently
and accurately compute phase field models for phase transformations, including
stress dependent phase transformations.
(April 26) Dr. Jian Zhong Zhu:
Computational modeling of hot tearing and cold crack in metal casting -
Hot tearing and cold crack are two of the most detrimental defects occurring in the metal casting process. Therefore, it is practically important to be able to predict such phenomena numerically. Various criteria have been developed in the literature for assessing the susceptibility of hot tearing. Most of them, however, are not suitable for application in numerical simulation of industrial metal casting processes. In this study, we developed a hot tearing indicator based on the criterion of accumulated plastic strain. Combining with an appropriate constitutive model, such indicator is also extended to predict the development of cold crack in the casting parts. Numerical results are presented and compared with experimental results. (May 05) Prof. Yann Brenier:
Derivation of particle, string and membrane motions from Born-Infeld
Electromagnetism -
We derive classical particle, string and membrane motion
equations from a
rigorous asymptotic analysis of the Born-Infeld nonlinear electromagnetic
theory. The Born-Infeld model is a modification of the classical
Maxwell equations, designed in 1934 to cutoff, in a nonlinear
fashion, the infinite electrostatic field generated by point charged
particles, at a scale of order 10 to the -15 meters.
Quickly shadowed by the success of Quantum Electrodynamics in the 40',
this model has known a recent revival through the concept of
Dirichlet-Branes in String Theory.
Our asymptotic analysis starts by adding to the Born-Infeld
equations the corresponding
energy-momentum conservation laws and write the resulting system as a
non-conservative symmetric system of ten first-order evolution PDEs.
(Surprisingly enough, the extended system enjoys Galilean
invariance, just like classical Fluid Mechanics.)
Then, we show that four rescaled versions
of the system have smooth solutions existing in the time interval
where the corresponding limit problems have smooth solutions.
These limits respectively describe the classical linear Maxwell equations
(for weak fields), string motions (for large magnetic fields and moderate
electric fields), particle motions (for large electromagnetic fields)
and membrane motions.
This is partly a joint work with Wen-An Yong, from Heidelberg.
(May 06) Prof. Yann Brenier:
String integration of some MHD equations -
We first review the link between strings and some Magnetohydrodynamics
equations. Typical examples are the Born-Infeld system, the Chaplygin
gas equations and the shallow water MHD model. They arise in Physics
at very different (from subatomic to cosmologic) scales.
These models can be exactly integrated in one space dimension by solving
the 1D wave equation and using the d'Alembert formula.
We show how an elementary "string integrator" can be used
to solve these MHD equations through dimensional splitting.
A good control of the energy conservation is needed due to the repeted
use of Lagrangian to Eulerian grid projections. Numerical simulations
in 1 and 2 dimensions will be shown.
(May 10) Prof. Ramani Duraiswami:
Computations of Singular and Nearly Singular Integrals in Boundary
Integral Methods -
For many problems boundary integral (a.k.a. boundary element) methods
have long been considered as very promising. These methods can easily
handle complex shapes, lead to problems in fewer variables, and further
only require meshing of the boundary rather than the entire domain. Despite these advantages, two issues have impeded the widespread
adoption of boundary integral methods. First, these techniques lead to
linear systems with dense non-symmetric matrices, for which efficient
iterative solvers are not usually available. Second, the computation of
the matrix elements requires the computation of various integrals.
However the diagonal elements of these matrices require computation of
weakly singular, singular or hypersingular integrals. These have to be
dealt with separately, and require a complicated analytical apparatus to
regularize the singularity. Further they make the development of BEM
software difficult as the singular integrals need special separate
treatment in the software. The fast multipole method presents a promising approach to removing the
first difficulty. This talk is concerned with the second issue: the
treatment of the singular integrals. We are developing a new technique,
that appears capable of efficient computation of the integrals that
arise from the singular elements. The method appears to be of general
applicability for the calculation of all the singular and nearly
integrals that arise in different BEM formulations, and further fit
naturally in to the FMM accelerated solution of the linear system. Our
method only employs regular quadrature formulae. Details of the new
method will be presented, along with its tests in the solution of the
Laplace and Helmholtz equations via the BEM. (joint work with Nail Gumerov)
|