Volume 5, No. 1, January 2009


Volume 5, No. 1, January 2009 ISSN: 1743-9310
Title: AN APPROACH TO INITIALIZE FOUR-PARAMETER SINE WAVE FITTING Authors: K. F. Chen.................................................................................................pp. 1-10 Abstract: Sinusoidal model fitting plays an important role in measuring the effective number of bits of an ADC. However, the most attractive fitting, four-parameter sine wave fitting based on the least square error principle, is nonlinear. An iterative approach to such a problem necessitates the initial frequency. In this paper, the complication of inappropriate initial values was reviewed firstly, and then an explicit approach was presented to obtain the initial frequency. This new approach makes use of three spectrum lines around the main lobe of the fast Fourier transform (FFT) and can work with a short record. Finally, it was tested by simulation signals contaminated with different levels of noise, with the lowest signal-to-noise being -3.01dB. The results from noise-free cases indicate that the frequency error of the proposed approach is no greater than 0.0006 times of the FFT resolution. Regarding to the noise cases, insofar as the simulated trials, the average difference between the empirical variance and the Cramer-Rao bound (CRB) is less than four dBs. The CRB can be reached after only one round of iteration starting from the initial frequency obtained from the proposed approach. Keywords: Four-Parameter Sine Wave, Least Square Error, Fast Fourier Transform (FFT), Cramer-Rao Bound, Analog-to-Digital Converter, Iteration.
Title: STABILIZING AND BALANCING OF INPUT DC VOLTAGES OF FIVE LEVELS DIODE CLAMPED INVERTER BY COMBINING SPACE VECTOR MODULATION AND SLIDING MODE CONTROL Authors: N. Lourci, E. M. Berkouk, D. Lalili...........................................................pp. 11-21 Abstract: The stability problem of the input DC voltages in a five levels diode clamped inverter is recalled and illustrated. A knowledge model of this converter using switching functions and a space vector PWM strategy is demonstrated in detail. To investigate the instability problem of the input DC voltages of the inverter, we study a cascade constituted by a two levels rectifier - five levels diode clamped inverter - induction motor. To solve this problem, we stabilize the rectifier output voltage using the sliding mode control and we balance the input DC voltages of the five levels diode clamped inverter by using redundant vectors of the space vector PWM strategy. The obtained results show the effectiveness of our method which can be used to stabilize the input DC voltages of the multilevel converters. Keywords: Five Levels Inverter, Two Levels Rectifier, Induction Motor, Space Vector PWM, Current Hysteresis Strategy, Sliding Mode Control.
Title: NEW I-V CHARACTERIZATION MODEL FOR PHOTOVOLTAIC MODULES AND EXPERIMENTAL DETERMINATION OF INTERNAL RESISTANCES Authors: M. Benghanem, S. N. Alamri, A. Mellit.....................................................pp. 22-30 Abstract: In this paper, an explicit model for accurate simulation of the I-V characteristic of photovoltaic (PV) cells and modules under various climatic conditions is described. The proposed model is found to be reliable and accurate in the situations where the model is a good approximation of cell or module performance. The developed model has been compared with the traditional I-V characteristic and with some experimental results. The obtained results showed the effectiveness of our model. Moreover, an experimental method is proposed to determine the series resistance and shunt resistance of the PV cells and modules. Keywords: I-V Characterization, Explicit Model, Simulation, Photovoltaic, Experimental Measurement, Series Resistance, Shunt Resistance.
Title: PID CONTROLLER DESIGN FOR TIME DELAY SYSTEMS USING GENETIC ALGORITHMS Authors: K. Saadaoui, A. Moussa, M. Benrejeb.....................................................pp. 31-36 Abstract: In this paper, stabilizing regions of a PID controller applied to a class of time delay systems are computed using parametric methods. A necessary condition is used to obtain the admissible ranges of proportional and derivative gains. Then, for a fixed value of one of these parameters within this admissible range, stabilizing regions in the space of the remaining two parameters are determined. Stabilization being the most basic requirement in any controller design problem, once this property is guaranteed we can search among these stabilizing controllers those that satisfy other performance specifications. This step is carried out using the genetic algorithm optimization method. Time domain measures of the closed loop system such as maximum percent overshoot, rise time and settling time are minimized using genetic algorithms and the stabilizing regions of the PID controller values. Examples are given to demonstrate the effectiveness of our proposed approach. Keywords: Stability, Time Delay Systems, PID Controller, Genetic Algorithms, GA.
Title: A NOVEL TECHNIQUE FOR THE DESIGN OF LOCALLY TUNEABLE DECENTRALISED ROBUST CONTROLLERS Authors: L. F. Yeung, G. Lu, A. Nobakhti, D. W. C. Ho, H. Wang........................pp. 37-50 Abstract: This paper proposes a decentralised controller design procedure for weakly coupled systems. A key feature of the design method is that the problem of designing a complete decentralised controller is replaced by several simpler local design problems. This is achieved by introducing a decomposition condition, which imposes local design constraints. This means that the design of each local controller is completely decoupled from other controllers. This has both design, and computational benefits. For the design, as long as each local controller satisfies the local decomposition condition, its design will be independent of the other controllers. In turn, this facilitates tuning and commissioning of the controller and the various design stages such as the choice of weighting functions. Most importantly this means that each of the loop-controllers is locally tunable. To achieve this, the decomposition condition is combined into a set of linear matrix inequalities (LMIs), then a sub-optimal robust solution for this type of control problem is obtained. The approach has been applied to several numerical examples and used for a case study of a gas turbine engine controller design. Keywords: Decentralised Controller, Large Scale Systems, System Decomposition, Assignment Problem.