Design of a Morlet wavelet control algorithm using super–twisting sliding modes applied to an induction machine

Abstract

In this paper, a Morlet wavelet and super–twisting control algorithm are designed and implemented to a three–phase induction motor. The mathematical model of the squirrel–cage induction motor to be controlled is approximated by the Morlet wavelet artificial neural network, which is trained on–line with the error filtered algorithm in order to reproduce the dynamics of the plant to be controlled. The structure of the artificial neural network is proposed in series–parallel configuration and block control form to design the sliding variety, where the super–twisting control algorithm is applied indirectly. For the non–measurable state variables of the plant, state observers of the super–twisting type are proposed to feed the inputs of the artificial neural network. The simulation of the complete system in closed loop is performed where the variables to be controlled are the angular velocity and the square modulus of flux linkages. The results obtained in Matlab/Simulink validate the efficiency of the proposed neural network for the identification of states and the application of the controller