Keynote Speeches
Keynote 1
Dr. Tariq Samad, Honeywell Automation and Control Solutions, USA
Tuesday, 3 June 2008, 09:00 - 10:00 hrs, Atrium Ballroom
"Energy, Environment and Process Automation"
The first part of this presentation will be on the topic of industrial automation. I will sketch the evolution of distributed control systems, highlighting how the information technology sector has driven advances in DCSs over the last few decades. This trend continues, with wireless as the most recent driver. The presentation will review new developments at Honeywell in industrial wireless technology, including wireless sensors and mesh networks for industrial plants. The remainder of the presentation will focus on clean energy, and in particular on CO2 mitigation in power generation. I will review some data on the role of the power generation sector in anthropogenic CO2 emissions, highlighting the increasing use of coal as a fuel source and the likelihood that growth in coal use will continue for some time. Clean coal technologies will thus be essential for managing atmospheric CO2 levels. These technologies present opportunities for industrial automation and control; energy efficiency, integrated gasification combined cycle (IGCC), oxy-fuel combustion, and carbon capture and sequestration (CCS) will be discussed in this context.
Keynote 2
Professor Romeo Ortega LSS-Supelec, FRANCE
Wednesday, 4 June 2008, 09:00 - 10:00 hrs, Atrium Ballroom
"Power Factor Compensation for Electrical Circuits: A Framework for Analysis and Design in the Nonlinear Nonsinusoidal Case"
Optimizing the energy transfer from an AC source to a load is a classical problem in electrical engineering. The power transmission efficiency for a given load is captured by the power factor, defined as the ratio between real instantaneous power and the apparent power (which is the product of rms values of the voltage and current). To improve the power factor, it is standard in practice to place a compensator between the source and the load. To design the compensator it is typically assumed that the source consists of an ideal generator, that is, without impedance, and with fixed sinusoidal voltage and that the load is linear time-invariant (LTI). Under these assumptions, it can be shown that the power factor is proportional to the phase shift angle between voltage and current at the load terminals. The role of the power factor compensator is then to reduce this phase shift.
Due to economical and environmental considerations, more stringent efficiency requirements have been imposed in the last few years on electrical systems, which has resulted, on one hand, on the widespread use of power semiconductor switching devices that are, essentially, lossless and hence reduce dissipation. On the other hand, to fulfill new performance requirements the electrical devices are pushed to wider operating ranges where nonlinear phenomena are predominant. Switching circuits are used in the loads as well as the sensing and compensation equipment. An unfortunate consequence of the inclusion of switching devices and the presence of nonlinear loads is additional signal distortion, which has two undesirable effects. First, the introduction of harmonics that are not present in the original waveforms may excite unmodeled dynamics, and result in degraded performances. Second, the task of designing power factor compensators which, as indicated above, is well understood for sinusoidal signals is far from clear in the face of distorted signals.
Underlying the phase-shifting action of power factor compensation in the sinusoidal LTI case is a fundamental energy equalization mechanism. Indeed, it can be shown that the power factor is improved if and only if the difference between the average electric and magnetic energies stored in the circuit is reduced. The optimal value is achieved when electric and magnetic energies are equal, that is, when the impedance seen from the source behaves like a resistor for the particular frequency of the source. One of the driving forces of our work is to investigate to which extent this fundamental energy equalization property is valid in the nonlinear nonsinusoidal case. We prove that for general nonlinear, but necessarily time-invariant, circuits the property still holds under quite general assumptions. Unfortunately, the key time invariance requirement rules out the practically important case of switching circuits that massively dominates in applications.
To overcome this obstacle we look for an alternative characterization of power factor compensation that allows us to treat switching devices. The main contribution of our work [1] is the identification of the fundamental role played by the property of cyclo-dissipativity in this problem. (In terms of energy exchange, cyclo-passive systems exhibit a net absorption of energy along closed trajectories, while passive systems absorb energy along any trajectory that starts from a state of minimal energy. Energy here is understood in the broader sense introduced by Willems in his seminal work of 1972 [2].)
Namely, we prove that a necessary and sufficient condition for a shunt lossless compensator to improve the power factor is that the overall system satisfies a given cyclo-dissipativity property. In this way, cyclo-dissipativity provides a rigorous mathematical framework useful to analyze and design power factor compensators for general nonlinear loads operating in nonsinusoidal regimes. We illustrate the usefulness of this framework with some circuit examples, including the ubiquitous half-wave rectifier.
References
[1] E. Garcia-Canseco, R. Gri, R. Ortega, M. Salich and A. Stankovic, \Power factor compensation of electrical circuits: The nonlinear non-sinusoidal case", IEEE Control Systems Magazine, Vol. 27, No. 4, April 2007.
[2] J. C. Willems, \Dissipative dynamical systems. Part I: General theory: Part II: Linear systems with quadratic suply rates." Archive for Rational Mechanics and Analysis, vol. 45, pp. 321{393, 1972.
Keynote 3
Professor Frank L. Lewis The University of Texas at Arlington, USA
Thursday, 5 June 2008, 09:00 - 10:00 hrs, Atrium Ballroom
"Decision and Control in Supervisory Discrete Event Systems: Applications to Condition-Based Maintenance"
Discrete event systems with shared resources and routing choices are ubiquitous in real-world applications and appear in wireless sensor networks, diagnosis and condition-based maintenance, manufacturing workcells, and elsewhere. We present a framework for design, analysis, and implementation of real-time on-line decision controllers that is based on matrices. This matrix formulation allows straightforward design for shared resource dispatching with multi-path coordination, deadlock avoidance, and priority dispatching. A nonstandard or/and matrix algebra is used. On-line control 'supervisors' are designed for certain 'critical subsystems' within which the number of jobs must be limited to avoid blocking phenomena.
A new method is given for deadlock avoidance in Petri nets having 'decision places,' where there is a choice about which of multiple jobs can be performed next. The so-called 'critical siphons' used for non-choice PN must be redefined in such free choice systems. This opens the applicability of DEC to a wide range of task scheduling problems such as decision making in condition-based maintenance systems.
Dempster-Shafer decision is a powerful method based on evidential reasoning that can be used to select which task to perform at decision branches in a free choice Petri net. However, DS computations are tedious to perform since they require summations over set inclusions. A new matrix formulation is given for computations in Dempster Shafer belief propagation that allows straightforward on-line implementation of DS decisions in a supervisory DE controller.
A sample application to machine Condition-based Maintenance is given where DS decision is used to select which jobs to perform in a free choice Petri Net used to control repair of faulty equipments.