2008 (Volume 18)
Structural decomposition of transfer matrix of positive normal hybrid systems
(Faculty of Electrical Engineering, Białystok Technical University, Białystok, Poland)
A new class of positive normal hybrid linear systems is introduced. It is shown that: 1) the inverse matrix of the characteristic system matrix is normal if and only if its greatest common divisor of (n - 1)-th order minors is equal to one, 2) the rational matrix is normal if and only if its McMillan polynomial is equal to the last common denominator, 3) the rational matrix has the structural decomposition of its transfer matrix if and only if the transfer matrix is normal. A procedure for computation of the structural decomposition of a normal transfer matrix is proposed. The considerations are illustrated by numerical examples.
keywords: normal rational matrix, structural decomposition, linear, hybrid system
Optimization of fuzzy PID controllers using Q-learning algorithm
|Hamid Boubertakh, Salim Labiod|
(LAMEL, Faculte des Sciences de l'Ingnieur, University of Jijel, Jijel, Algeria)
(LCP, Departement de Genie Electrique, Ecole Nationale Polytechnique, Alger, Algeria)
(INSA de Rennes, Departement of Informatics, Rennes, France)
In this article, we first chose the design settings of the fuzzy PID controllers (FPIDC) so that the FPIDCs mimic the classical PID controllers. The advantage of these controllers is the combination of the simplicity of the classical PID controllers and the interpretability of fuzzy controllers which makes the task of parameters tuning easier. Secondly, we present a method for optimizing the closed-loop system consisting of a FPIDC and an unknown plant using the Q-learning algorithm (QLA). Specifically, QLA minimizes a cost function which quantifies the performance of FPIDC. Without loss of generality the square error sum cost function is used. The QLA, which is a nonmodel-based method, iteratively search of the best parameters so that the output of the cost function is less then satisfaction threshold. Finally, a simulation example is used to prove the effectiveness of the proposed method.
keywords: fuzzy PID controllers, classical PID controllers, Takagi-Sugeno fuzzy systems, Q-learning
Zeros and the output-zeroing problem in linear fractional-order systems
(Warsaw University of Technology, Warsaw, Poland)
The concept of invariant zeros in fractional order LTI systems with the Caputo derivative is introduced in the paper. The problem of zeroing the system output is discussed. For systems of uniform rank explicit formulas for output-zeroing inputs and the corresponding solutions to the state equation are provided. The zero dynamics and invariant zeros for such systems are also characterized.
keywords: fractional order linear systems, invariant zeros, output-zeroing problem
Contribution to variable structure control based on the approach of fuzzy modeling of Ben-Ghalia for induction motor with Luenberger observer
(Faculty of Electronic and Computing, University of Science and Technology Houari Boumediene (USTHB), Alger, Algerie)
(Department of Computer Science Faculty of Engineering, Philadelphia University, Jordan)
(Department of Engineering of Information and Measurement Systems, Al-Balq'a Applied University, Jordan)
(Groupe de Recherche en Elecronique Industrielle, Universite du Quebec a Trois-Rivieres, Canada)
Sliding mode and fuzzy logic control have been combined together in a variety of ways. In this paper, a novel direction to design fuzzy sliding controllers is proposed based on the approach of Ben-Ghalia et al. Which is able to provide a fuzzy set based on a decoupled and linear system obtained after the application of the feedback- linearization control to the model of an induction motor. On the other hand, a Luenberger observer is conceived in the present work to observe the rotor flux in the motor which is too important to permit the application of fuzzy modeling approach and ensure the optimal and required performances . Furthermore, another objective to use such an observer consists in stabilizing the system and rejecting the disturbances when the system is subjected to uncertainties. Furthermore, to overcome the problem of the chattering phenomenon generated by the application of conventional sliding mode control applied to the cascade structure of the motor, the approach of fuzzy sliding mode control based on the fuzzy modelling proposed by Ben-Ghalia is used. The robust fuzzy control design is made feasible without resorting to model simplification or imposing restrictive conditions on the system uncertainty. Finally, simulation results proves the capability of the proposed process to ensure a good performances and robustness which is the principal aim of this research
keywords: feedback linearization control, sliding mode control, Luenberger observer, induction motor, cascade structure, fuzzy modeling control, fuzzy sliding mode control
Selecting genetic algorithms for assembler encoding
(Naval University, Gdynia, Poland)
Assembler Encoding makes it possible to use genetic algorithms to construct neural networks. Assembler Encoding represents neural network in a form of the so-called Assembler Encoding Program. The task of the program is to create Network Definition Matrix maintaining all the information necessary to construct the network. In Assembler Encoding different components of Assembler Encoding Programs evolve in separate populations. The evolution in each population can be controlled by a different genetic algorithm. In the experiments reported in the paper the following genetic algorithms were used to control the evolution of programs: Canonical Genetic Algorithm, Steady State Genetic Algorithm and Eugenic Algorithm. The programs created by means of the specified algorithms were used to create neural networks. Then, the networks were tested in the so-called predator-prey problem. The results of the experiments are presented at the end of the paper.
keywords: evolutionary neural networks, genetic algorithms, grasping
Enslavement and control in dq synchronous frame for a cascade using multilevel inverter NPC
|F. Bouchafaa, D. Beriber|
(Laboratory of Instrumentation and Engineering System, University of Science and Technology Houari Boumediene, Algiers, Algeria)
|M.S. Boucherit, E.M. Berkouk|
(Laboratory of Process Control, Department Electrical Engineering, Polytechnic National School Algiers, Algiers, Algeria)
The aim of this paper is to give a solution to unbalance problem of the input DC voltages of the nine-level NPC-VSI in dq synchronous frame. For that, we used in high voltage and great power applications, one two-level PWM current rectifier-nine-level NPC VSI-PMSM cascade. In the first part, we develop a knowledge model of the Nine-level NPC VSI and propose a space vector modulation of this converter using eight bipolar carriers. After that, this study shows particularly the problem of the stability of the DC voltages of the inverter and its consequence on the performances of the PMSM speed control. Then, we propose a solution to the problem by employed closed loop regulation using classic PI controller mode in d-q synchronous frame. We add a Clamping bridge to stabilize the eight DC input voltages of the nine-level NPC VSI. This cascade found applications in great power and high voltage fields as electrical traction.
keywords: NPC inverter, clamping bridge, rectifier multilevel, PWM strategy, regulation current, enslavement