Résumé: This paper deals with a simulation of different power methods to detect and diagnose induction generator faults. Instantaneous partial and total power, active and reactive power, complex apparent power and transformed power from mechanic to electric nature are revisited, simulated and discussed in this paper for induction generator rotor broken bars and stator short cuts faults detection and diagnosis. Fast Fourier transform (FFT) and PQ transform algorithms are used as comparison tools. Simulation results show that, on one hand, active, reactive and complex apparent power can only be used to detect evolution of rotor faults. On the other hand, partial, total and power transferred from mechanical to electrical nature are able to detect induction generator faults evolution with the advantage of eliminating electrical distortions and influence of low quality of supplying voltage. Furthermore, the implementation of the PQ transformation offers the possibility to isolate load influence from rotor faults and stator ones

Résumé: In this paper, we present an Audio-Visual Automatic Speech Recognition System that combines the acoustic and the visual data. The proposed algorithm here, for modelling the multimodal data, is a Hidden Markov Model (HMM) hybridised with the Genetic Algorithm (GA) to determine its optimal structure. This algorithm is combined with the Baum–Welch algorithm, which allows an effective re-estimation of the probabilities of the HMM. Our experiments show the improvement in the performance of the most promising audio-visual system, based on the combination of GA/HMM model compared to the traditional HMM.

Résumé: This paper deals with the flatness-based approach for sensorless control of the induction motor systems. Two main features of the proposed flatness based control are worth to be mentioned. Firstly, the simplicity of implementation of the flatness approach as a nonlinear feedback linearization control technique. Secondly, when the chosen flat outputs involve non available state variable measurements a nonlinear observer is used to estimate them. The main advantage of the used observer is its ability to exploite the properties of the system nonlinearties. The simulation results are presented to illustrate the effectiness of the proposed approach for sensorless control of the considered induction motor.

Résumé: This paper addresses the problem of accurately estimating the mechanical and magnetic state variables as well as the stator and rotor resistances of induction motors using only the stator current measurements and the supplied stator voltages from an appropriate nonlinear parametrization. The involved estimation is carried out by a high gain adaptive observer designed bearing in mind the available fundamental results together with the useful implementation features, namely conception simplicity and computational efficiency. An exponential convergence of the state and parameter estimation errors is established under admissible assumptions, namely the persistent excitation requirement has been particularly reduced thanks to the introduction of unknown parameter characteristic indices. The effectiveness of the adaptive observer is highlighted throughout simulation results involving a typical induction motor.

Résumé: In this paper a sensorless Backstepping control scheme for rotor speed and flux control of induction motor drive is proposed. The most interesting feature of this technique is to deal with non-linearity of high-order system by using a virtual control variable to render the system simple. In this technique, the control outputs can be derived step by step through appropriate Lyapunov functions. A high gain observer is performed to estimate non available rotor speed and flux measurements to design the full control scheme of the considered induction motor drive. Simulation results are presented to validate the effectiveness of the proposed sensorless Backstepping control of the considered induction motor.