Dynamics and Controls Seminars
Abstracts
2004 -2005
* * * * * * * *
* * * * *
Wednesday, August 24, 2005
2:00 p.m.
584 EBU-II
Dr. Vassilis Theofilis
E.T.S.I. Aeronáuticos
Universidad Politécnica de Madrid
“Recent Progress in BiGlobal
Instability Analysis of Complex Flows”
In a manner analogous to the well-studied plane channel flow, accurate and rapid access to the eigenspectrum of flows in complex domains is key to theoretically-founded flow-control strategies. The talk will focus on the development and first application of methodologies appropriate for the numerical solution of the viscous incompressible BiGlobal eigenvalue problem (EVP) in complex domains. Recent progress will be reported in the instability analysis of two applications of industrial significance, three-dimensional flow over a row of T-106/300 Low Pressure Turbine (LPT) blades and flow over open cavities of arbitrary aspect ratio.
In the first application, our previous analyses have established that the steady two-dimensional low-Reynolds number flow undergoes a two-dimensional bifurcation into a time-periodic state at a chord Reynolds number of O(1000). Presently, the time periodic basic states established past the first bifurcation have been analyzed using Floquet theory. In addition, three-dimensional direct numerical simulations have been performed in order to follow amplified Floquet eigenmodes into transition. The most significant LPT eigenmode has been found to be qualitatively analogous with mode A of cylinder flow. Throughout the range 910 < Re < 2000 investigated, this LPT mode gives rise to a narrow window of instability of the time-periodic 2D flow to 3D disturbances having spanwise periodicity lengths in the neighborhood of four chord lengths.
In the second application, a novel spectral multi-domain approach for the solution of the BiGlobal EVP has been developed, in order to study instability of flows in domains decomposable into rectangular computational subdomains. Iterative methods for the recovery of the eigenspectrum on distributed computing platforms have been exploited. Model one- and two-dimensional EVPs, as well as the Orr-Sommerfeld equation in the plane channel and the BiGlobal EVP in the lid-driven cavity, have been solved in the course of the validations of the new algorithm. Subsequently, the first ever three-dimensional viscous BiGlobal instability analysis of incompressible open cavity flow has been performed. First results point to strong qualitative analogies between the leading eigenmodes of the open- and the lid-driven cavity flow problems.
*
* * * * * * * * * * * *
Thursday, August 18, 2005
11:00 a.m.
479 EBU-II
Dr. Robert W. Stark
Department of Earth and Environmental
“Qualifications of Forces in Atomic Force
Microscopy: A System Theoretic Viewpoint”
Today, atomic force microscopy is a standard technique for surface analysis with nanometer resolution. In an atomic force microscope (AFM) the specimen is raster scanned by a very fine needle in order to acquire a set of 3-dimensional data of the surface topography. Various imaging modes have been developed including quasi static modes, where the tip is scanned in permanent contact with the surface together with huge variety of dynamic measurement methods such as amplitude and frequency modulation mode, or the ultrasonic mode. In addition to surface analysis the AFM has proved to be a versatile tool for surface modification and nanomanipulation.
For both applications of
the AFM – quantitative surface analysis and precise nanomanipulation
– it is essential to understand the system dynamics of the AFM. This means that
beyond the determination of spring constant and quality factor of the cantilever
the mechanical resonator has to be characterized with its full spectral
behavior. Additionally, the feedback perspective [1] of dynamic AFM provides a
powerful tool to investigate the non-linear system dynamics from a system
theoretic point of view. Including the higher order dynamics of the extended
cantilever beam in the model the contact resonances can be reproduced
faithfully [2]. The investigation of the non-linear dynamics provides valuable
insight into the generation of higher harmonics in dynamic AFM [3].
Based on the understanding
of the system dynamics and the transfer characteristics it has become possible
to address the inverse problem: the reconstruction of tip sample forces in
dynamic AFM [4]. Thus, the solution of the inverse problem allows for the
quantification of surfaces forces with nanometer resolution on a
sub-microsecond timescale.
[1] A. Sebastian et al., J.Appl.Phys.,89, 6473 (2001).
[2] R. W. Stark, G. Schitter, M. Stark et al., Phys. Rev. B C. 69 (8),
085412 (2004).
[3] R. W. Stark, Nanotechnology, 15, 347 (2004).
[4] M. Stark, R. W. Stark, W. M. Heckl et al., Proc. Natl. Acad. Sci.
USA 99
(13), 8473 (2002).
* *
* * * * * * * * * * *
Friday, May 20, 2005
1:00 p.m.
479 EBU-II
Dr. Yuri Orlov
Electronics Department,
“Finite Time Stability and Robust Control
Synthesis of Uncertain Switched Systems with Applications to Underactuated Mechanical Manipulators”
Stability analysis is presented for uncertain nonlinear switched systems. While being asymptotically stable and homogeneous of degree q < 0, these systems are shown to approach the equilibrium point in finite time. This feature is additionally demonstrated to persist regardless of inhomogeneous perturbations. Based on this fundamental property, quasihomogeneous switched control algorithms are then developed to asymptotically stabilize underactuated mechanical systems operating under uncertain conditions. In order to locally stabilize an underactuated system around an unstable equilibrium, its output is specified in such a way that the corresponding zero dynamics is locally asymptotically stable. Once such an output has been chosen, the desired stability property of the closed-loop system is provided by applying a quasihomogeneous switched controller, driving the system to the zero dynamics manifold in finite time. The controllers constructed do not rely on the generation of sliding motions while providing robustness features similar to those possessed by their sliding mode counterparts. Performance issues of the proposed synthesis are illustrated in a simulation study of a servo-motor with backlash and a two-link pendulum robot (Pendubot).
* * * * * * * * * * * * *
479 EBU II
Dr. Richard H. Middleton
ARC Centre for Complex Dynamic Systems and Control
“Control over Signal to Noise Ratio
Constrained Channels”
There has recently been substantial interest in the
interplay between feedback control theory and communications, including the
influence of communications channels on feedback control. One of the effects
studied by several authors is the effect of bit rate limitations on the ability
to stabilize an open loop unstable plant. In this talk, I adopt an alternate
approach of considering a communications channels as have a constraint on the
permissible transmitted power, together with noise corruption of the
transmitted signal. It turns out, for example, that there are limitations on
the ability to stabilize an unstable system over an additive white Gaussian
noise channel whilst respecting a power constraint, that is, signal to noise
ratio constrained channels. Some links between these results, bit rate limited
control results and
* *
* * * * * * * * * * *
Dr.
United
“Control of Oscillations in Jet Engines:
from Fixing Problems to the Design of Dynamics”
The talk will summarize the lessons learned during eight years of industrial research on mitigation of flow and structure oscillations in jet engines. We will advocate the need to change the role of control theory from fixing the detrimental dynamics using active control (with narrow focus on control algorithms) to the design for beneficial dynamics early in the design cycle. Particular case studies will include mitigation of thermoacoustic instabilities and turbomachinery flutter. We will show how the decisions on the control system architecture (sensor and actuator location) impact the achievable level of suppression of oscillations (fundamental limitations of performance). We will also show how certain aspects of design of jet engines (symmetry) contribute to the origin of detrimental oscillations and point out how dynamical systems and control theory methods can guide the design of the engines to prevent the oscillations.
* * * * * * * * * * * * *
Dr.
President of the
“Exploring the maximum capability of
feedback in dealing with uncertainties”
To explore the maximum capability and potential limits of
the feedback mechanism, we have to work in a framework that is somewhat beyond
those of the classical ones (including robust and adaptive control), since we
need to study the full capability of the feedback mechanism which includes all
(nonlinear and time-varying) causal mappings, and are not only restricted to a fixed
feedback law or a set of specific feedback laws. We need not only to answer
what the feedback can do, but also to answer, the more difficult question, what
the feedback cannot do. In this seminar, we will present several
concrete results towards understanding the capability (and limits) of the
feedback mechanism in dealing with structural uncertainties. In
particular, for several basic classes of uncertain nonlinear discrete-time (or
sampled-data) dynamical control systems, some “Critical Values” and
“Impossibility Theorems” concerning the capability of feedback will be
presented.
* * * * * * * * * * * * *
Dr.
MAE Department
University of
Techniques and Tools for Complex
Systems
Increasing computational power has sprouted interest in complex systems. On the one hand, "standard" tools can handle larger and more complex systems. On the other hand, simulations reveal complex behaviors from interactions of a number of simple components. In this talk we discuss techniques and tools to approach problems on these two ends. We first explore ideas and concepts that can take advantage of structure on systems and control problems formulated in terms of matrix inequalities. The challenge, with major impact on the analysis and design of complex systems, is to efficiently handle structure from the setup of the problem to its numerical solution. We then present an overview of our efforts to model and simulate the actin-spectrin network of erythrocytes (red blood cells). We show how to efficiently simulate large networks by carefully modeling its repeating structural unit, and how the results of our simulations can lead to new insights on the biological role played by the actin-spectrin network and the nature of its components.
* * * * * * * * * * * * *
479 EBU-II
Dr. Hana
El-Samad
Department of Mechanical and Environmental
“Biological Design Principles For Robustness,
Performance and Selective Interactions with Noise: A Systems Engineering
Perspective”
We advocate the use of a systems approach in the modeling and analysis of gene regulatory networks and summarize our work on the modeling of the heat shock response, a cellular system of utmost physiological importance. The heat shock response refers to an intricate set of feedback and feedforward mechanisms that organisms employ to combat the harmful effects of unmediated temperature increases. We illustrate how mathematical models of these mechanisms provide valuable insight, explaining dynamic phenomena exhibited by wild type and mutant heat shock responses, corroborating experimental data and guiding novel biological experiments. Since gene regulatory networks, including the heat shock system, are permanently affected by various sources of noise, we briefly discuss some principles of noise attenuation and exploitation that seem to be at work in these circuits. We argue that a meaningful study of stochastic behavior in these systems should reach beyond arguments based on the number of molecules. It should instead emphasize their dynamic properties and structural features. We conclude with a perspective on future research directions, including theoretical aspects of multi-scale stochastic methods and moment closure techniques for stochastic modeling and analysis of genetic networks.
* * * * * * * * * * * * *
Monday, March 14, 2005
10:00 a.m.
479 EBU-II
Dr. Sonia Martinez
Department of Mechanical and Environmental
“Motion
Coordination Algorithms for Multi-Agent Systems”
Multi-agent systems, such as groups of mobile robots, biological populations or large-scale information systems, possess unique characteristics that make their analysis and control a challenging problem. Like in schools of fish each member of the group reacts to local information and yet, complex coordinated behaviors appear without the presence of a group leader. In this talk, I will discuss some recent ideas for the control and coordination of multi-agent systems, in particular groups of mobile robots, motivated by recent applications of sensor networks. We present motion coordination algorithms that produce emerging behaviors such as deployment over a given region or gathering at a specified location of the space, pointing out recent developed theoretical tools to analyze them. Finally, I will discuss some exciting directions for future research.
* * * * * * * * * * * * *
584 EBU-II
Dr. Robert L. Kosut
Vice President, Systems & Control Division
SC Solutions, Inc.,
“Control, Identification, and Detection of Quantum Systems for Quantum
Information Processing”
In a 1985 paper in Optics News entitled ``Quantum Mechanical Computers,'' Richard Feynman described how a computer could be built upon the mathematical principals of quantum mechanics. But he also heralded the difficulties in an actual physical implementation:
``This computer seems to be very delicate and these imperfections may produce considerable havoc.''
This talk will describe some of our on-going efforts to alleviate the potential ``havoc'' by appealing to control design, system identification, and convex optimization.
* * * * * * * * * * * * *
479 EBU-II
Brian D.O. Anderson
Chief Scientist, National ICT Australia Limited and
The
“Two Decades of Adaptive Control
Pitfalls”
Adaptive control is a very appealing technology, at least in principle. Yet its use has been conditioned by an attitude of distrustfulness on the part of some practitioners. In this talk, we review (without presupposing knowledge of adaptive control on the side of listeners) some concepts the isolation of which was necessary to engender confidence in the technology. These include the unpredictable failure of the MIT rule; the bursting phenomenon, and how to prevent it; the notion that identification of a plant is only valid conceptually for a restricted range of controllers (with the implication that in adaptive control, certain controller changes may be hazardous); and the concept of multiple model adaptive control, which from some points of view raises as many the problems as it solves.
*
* * * * * * * * * * * *
584 EBU-II
Christopher K. Allen
“Control Applications in Charged
Particle Accelerators and Light Sources”
We first overview the basic
theory and operation of charged particle accelerators then discuss several
high-level control problems that arise in various accelerator application. We
present the basic principles of charge particle acceleration, as illustrated by
the Panofsky equation, then
cover some of the differing accelerating structures. Transverse beam focusing and beam transport
are also outlined from a beam optics perspective. We then concentrate on control problems in
charged particular accelerators. In
particular the challenge of steering the collective beam through the
accelerator makes a good illustrative problem.
We also discuss klystron phase tuning, beam matching and halo mitigation,
and beam stabilization.
* * * * * * * * * * * * *
584 EBU-II
Johan Akesson
Ph.D. Student
Department of Automatic Control
Lund Institute of Technology
“A Framework for Grade Changes: An Optimization and Sequential Control
Approach”
In modern chemical process industry, process state transitions, or grade changes, are common operations. Frequent production changes increase the focus on efficient operator support for such procedures. The aims of a grade change may be complex, and sometimes conflicting. Fast transitions are desirable in order to minimize production loss. There may also be safety aspects that must be considered, in order to achieve acceptable risks of failure. Operator acceptance is a key issue for a decision support systems, and is the topic of this paper. We address the problem by combining dynamic optimization techniques and a tool for sequential control; JGrafchart.
* * * * * * * * * * * * *
479 EBU-II
Jeff Scruggs
Postdoctoral Research Fellow
Division of Engineering & Applied Science
California Institute of Technology
“Structural Control Using Regenerative Force Actuation
Networks”
Much of the research in semi active structural control has made use of forcing devices with variable mechanical parameters; devices such as variable-orifice, rheological fluid, variable friction, variable stiffness, and mechanical switching systems. However, it is also possible for structural energy to be dissipated electrically, using motors to facilitate electromechanical energy conversion, and then using controllable circuitry to regulate dissipation. This approach has advantages over other methods of dissipation regulation, in that if two or more devices are used to control a structure, their associated electrical circuitry may be connected such that they can share power. Additionally, energy removed from the structure may be stored and reused. Such systems of forcing devices, called Regenerative Force Actuation RFA) networks, have only recently received significant attention. Like semi active systems, they require very little external power for operation, due to their inherently dissipative nature. They can be viewed as applying supplemental controllable linear damping forces to structures. However, unlike traditional damping approaches, this supplemental damping can in general be non-local and asymmetric. This added capability can yield a marked improvement in structural vibratory responses, both for transient as well as stationary excitations. In this talk, a detailed model for an arbitrarily-large RFA network is presented, and issues pertaining to the electronic control of this network are discussed. Some simple implementations for civil engineering applications are illustrated through simulations, and it is shown that the more generalized damping capabilities of RFA networks can be used to yield significant reductions in responses to seismic excitation. Furthermore, some nonlinear feedback controller design methods are discussed, which take advantage of the special features of these systems.
* * * * * * * * * * * * *
584 EBU-II
Professor Anders Rantzer
Department of Automatic Control
Lund Institute of
“ON CONVEXITY AND RELAXATION IN NONLINEAR AND HYBRID
CONTROL”
Since the 1950's, the idea of dynamic programming has propagated into a vast variety of applications. The basic Hamilton-Jacobi-Bellman equation is general and very simple, but the "curse of dimensionality" is often prohibitive and restricts most applications to a discrete state space of moderate size. In the last few years, there has been several exciting developments related to convexity of the corresponding Bellman inequality, both theoretical and computational. For example, research on continuous nonlinear control has generated analogs of the primal-dual optimization used in discrete shortest path algorithms. In nonlinear control, this has resulted in new stability criteria and new methods for control synthesis by convex optimization.
Another very exciting development is based on approximations of the cost to go. It turns out that the computational complexity of traditional dynamic programming algorithms often can be drastically reduced by relaxing the demand for optimality in the Bellman inequality. In fact, finding a solution which is guaranteed to be within 10% from the optimum turns out to be drastically less expensive than finding one within 1%. Results of this type will be discussed and illustrated by examples.
*
* * * * * * * * * * * *
For information: Sophia Bligh at (858) 822-1269