No description available.

Notre thèse est consacrée à l’étude de la supervision des réseaux de Petri en vue de la conception de systèmes manufacturiers flexibles. L’objectif est la définition de stratégies de pilotage en ligne pour l’évitement de conflits et d’interblocages, dans le cadre de la théorie de la supervision. Le point de départ de notre travail est d’exploiterle graphe de marquage du réseau de Petri, ce qui permet en particulier d’obtenir des stratégies de commande maximalement permissive pour des problèmes d’évitement de conflits et d’interblocages. Nous avons ainsi introduit des techniques originales, manipulations d’inégalités ou réductions d’ensembles de marquages, destinées à diminuerla complexité algorithmique d’une telle méthode. Dans premier temps, nous avons focalisé sur la synthèse de superviseurs dits purs, ce qui correspond au cas particulier où l’ensemble de marquage légaux, est convexe.Cette optimisation est ensuite considérée du point de vue de la facilité de mise en oeuvre. Nous traitons ainsi de la minimisation de la structure du superviseur et de son coût d’implémentation en préservant une structure de supervision qui offre à la fois la permissivité maximale et une complexité de calcul raisonnable en vue d’utilisationsur des installations réelles. Aussi, nous avons cherché à réduire le nombre de places de contrôle nécessaires pour réaliser un superviseur maximalement permissif, pour cela nous avons formule le calcul du nombre minimal de places de contrôle en termes d’un problème de programmation linéaire. Afin d’affaiblir la complexité de ce calcul de superviseur, deux versions de l’algorithme sont proposées. Ce problème de minimisation de la taille dusuperviseur, quoique fondamental, n’est pas abordé aussi directement dans la littérature. Il s’agit là d’une première contribution.Dans u second temps, nous nous sommes intéressés aux réseaux de Petri à boucles (self-loops). Les boucles étant représentées par une variable qui s’ajoute dans la contrainte inégalité définissant l’ensemble de marquages légaux. Après avoir proposé une méthode de réduction du nombre d’inégalités ainsi que du superviseur optimalen se basant sur les approches et résultats précédents, nous avons établi une condition suffisante d’obtention d’un superviseur maximalement permissif permettant de traiter des ensembles de marquages légaux non convexes.Enfin nous proposons une méthode de synthèse de contrôleur pour une nouvelle classe de réseaux de Petri, avec des arcs inhibiteurs correspondant à des contraintes définies par des intervalles. La taille du contrôleur ainsi obtenu et défini en termes d’arcs inhibiteurs à intervalles s’en trouve réduite ainsi que par conséquent sont coût d’implémentation. / Reachability graph analysis is an important technique for deadlockcontrol, which always suffers from a state explosion problem since it requires togenerate all or a part of reachable markings.Based on this technique, an optimal or suboptimal supervisor with high behavioralpermissiveness can always be achieved. This thesis focuses on designing liveness enforcing Petri net supervisors for FMSs by considering their behavioralpermissiveness, supervisory structure, and computationnal complexity.The following research contributions are made in this thesis.1. The design of a maximally permissive liveness-enforcing supervisor for an FMSis proposed by solving integer linear programming problems (ILPPs).2. Structural complexity is also an important issue for a maximally permissivePetri net supervisor. A deadlock prevention policy for FMSs is proposed, which canobtain a maximally permissive liveness-enforcing Petri net supervisor while thenumber of control places is compressed.3. In order to overcome the computational complexity problem in MCPP and ensurethat the controlled system is maximally permissive with a simple structure, wedevelop an iterative deadlock prevention policy and a modified version.4. We consider the hardware and software costs in the stage of controlimplementation of a deadlock prevention policy, aiming to obtain a maximallypermissive Petri net supervisor with the lowest implementation cost. A supervisorconsists of a set of control places and the arcs connecting control places totransitions. We assign an implementation cost for each control place and controland observation costs for each transition. Based on reachability graph analysis,maximal permissiveness can be achieved by designing place invariants that prohibitall FBMs but no legal markings.5. Self-loops are used to design maximally permissive supervisors. A self-loop ina Petri net cannot be mathematically represented by its incidence matrix. Wepresent a mathematical method to design a maximally permissive Petri netsupervisor that is expressed by a set of control places with self-loops. A controlplace with a self-loop can be represented by a constraint and a selfloopassociated with a transition whose firing may lead to an illegal marking.

Indiana University-Purdue University Indianapolis (IUPUI) / With the increasing demand of economy and environmental pollutions, more and more renewable energy systems with clean sources appear and have attracted attention of systems involving solar power, wind power and hybrid new energy powers[1]. However, there are some difficulties associated with combined utilization of solar and wind, such as their intermittent behavior and their peak hours mismatch in generation and consumption[1]. For this purpose, advanced network of a variety of renewable energy systems along with controllable load and storage units have been introduced[1-3]. This thesis proposes some configurations of hybrid energy harvesting systems, including wind-wind-storage DC power system with BOOST converters, solar-solar-storage DC power system with cascade BOOST converters, wind-solar-storage DC power system with BOOST converter and cascade BOOST converter, and wind-solar DC power system with SEPIC converter and BOOST converter. The models of all kinds of systems are built in Matlab/Simulink and the mathematical state-space models of combined renewable energy systems are also established. Several MPPT control strategies are introduced and designed to maximize the simultaneous power capturing from wind and solar, such as Perturb & Observe (P&O) algorithm for solar and wind, Tip Speed Ratio (TSR) control and Power Signal Feedback (PSF) control for wind, and Sliding Mode Extremum Seeking Control (SM-ESC) for wind and solar systems[4]. The control effects of some of these MPPT methods are also compared and analyzed. The supervisory control strategies corresponding to each configurations are also discussed and implemented to maximize the simultaneous energy harvesting from both renewable sources and balance the energy between the sources, battery and the load[2]. Different contingencies are considered and categorized according to the power generation available at each renewable source and the state of charge in the battery[2]. Applying the system architectures and control methods in the proposed hybrid new energy systems is a novel and significant attempt, which can be more general in the practical applications. Simulation results demonstrate accurate operation of the supervisory controller and functionality of the maximum power point tracking algorithm in each operating condition both for solar and for wind power[3]

No description available.

No description available.

No description available.

No description available.

Solar DEC (Desiccant and Evaporative Cooling) air-conditioning is a renewable technological approach to the future air-conditioning of buildings driven with solar-thermal heat. The principal acceptance of solar airconditioning has led to system prototypes mainly across Europe, however the diffusion of this innovative technology is proceeding slowly due to little field testing experience. In climates with coexisting heating demand particularly, a multivalent system approach that utilizes solar-heat not only for air-conditioning but also for hot water preparation and heating has potential as a feasible concept. However, previous research focused on systems using solar heat exclusively for the DEC-process. This research contributes to the advancement of the solar DEC-technology with multivalent use of solar thermal heat. The investigation consists of an initial detailed in-situ monitoring analysis of a system prototype operated in an industrial environment, followed by the development of optimised system concepts and a climate-specific analysis of the solar DEC-technology. The monitoring provided in-depth knowledge about the system operation, revealing the reasons for the insufficient refrigeration capacity achieved in practice. A detailed simulation model for an entire multivalent solar DEC-system including the heat sinks, DEC-system, heating and hot-water preparation was developed and a DEC-control strategy has been formulated. A new optimised control strategy for multivalent systems with simultaneous sink supply concept was devised. A sensitivity analysis was carried out to investigate the key design parameters for the dimensioning of multivalent solar DEC-systems. The research concluded that the auxiliary primary energy consumption of the optimised system was lower by one third compared to the initial system. Finally, a methodological zoning approach was developed, to systematically produce design-specific outline data for the application of the solar DEC-technology at climatically different sites.

A Research Project Abstract submitted in fulfillment of the requirements for Master of Science in Engineering [Electrical]: Telecommunications at the University Of The Witwatersrand, Johannesburg 07 June 2016 / Utilities use Supervisory Control and Data Acquisition systems as their industrial control system. The architecture of these systems in the past was based on them being isolated from other networks. Now with recent ever changing requirements of capabilities from these systems there is a need to converge with information technology systems and with the need to have these industrial networks communicating on packet switched networks there are cyber security concerns that come up. This research project looks at the whether using encryption in an IP/MPLS core network for SCADA in an OT IT environment has an effect on the performance requirements. This was done through an experimental simulation with the results recorded. The research project also looks at the key literature study considerations. The key research question for the research project of this MSc 50/50 mini-thesis is “whether using encryption in SCADA systems, the services performance requirements are still met in OT/ IT environment over an MPLS core network”? The research project seeks to determine if SCADA performance requirements are met over an encrypted MPLS/IP core network in an OT/IT environment. The key focus area of the research project is only encryption in the whole cyber security value chain versus SCADA services performances. This means that the research project only focused on the encryption portion of the whole cyber security value chain and the scope did not focus on other aspects of the value chain. This suffices for an MSc 50/50 mini-thesis research project as a focus on the whole value chain would require a full MSc thesis. Thus the primary objective for the research project is to research and demonstrate that encryption is essential for secure SCADA communication over a MPLS/IP core network. As aforementioned encryption forms an essential part of the Cyber Security value chain which has to achieve the following objectives. Confidentiality: ensuring that the information source is really from that source. Integrity: ensuring that the information has not been altered in any way. Availability: ensuring that system is not comprised but that it is available. These objectives of encryption should be met with SCADA service performance requirements not violated which is the objective of the research project. / M T 2016

"The recent years have seen the innovative system integration of a great many actuator technologies, such as point force actuators for space vehicle applications and the use of single fire actuators; such as pyrocharges to guide a free falling bomb to it’s target. The inherent limitations of these developments, such as nonlinear behavior under extreme environments and/or prolonged/repeated usage leading to a relaxation time component between firing of actuators and inherent system power limitations, have resulted in greater need for sophisticated control algorithms that allow for optimal switching between various actuators in any given embedded configuration so as to achieve the best possible performance of the system. The objective of this investigation is to offer a proof of concept experimental verification of a real time control algorithm, which switches between online piezoelectric actuators, employed for vibration control in an aluminum beam with fixed boundary conditions. In this investigation at a given interval of time, only one actuator is activated and the rest are kept dormant. The reason is to demonstrate the better vibration alleviation characteristics realized in switching between actuators depending on the state of the system, over the use of a single actuator that is always in fire mode. This effect is particularly pronounced in controlling systems affected by spatiotemporal disturbances. The algorithm can be easily adapted for various design configurations or system requirements. The optimality of switching is with respect to the minimal cost of an LQR performance index that corresponds to each actuator. Computer simulations with repeatable disturbance profiles, revealed that this algorithm offered better performance over the non-switched case. Performance measures employed were the time varying total energy norm of the dynamic system and position traces at any particular location on the beam. This algorithm was incorporated on a dSPACE rapid prototyping platform along with suitable hardware. Experimental and simulation results are discussed. "