Defective Multi Inverter on speed sliding neural control of Squirrel Cage Motor

Abstract

the sliding mode is a robust command use for nonlinear control, but can present the high frequency chattering phenomenon that occurs from the discontinuous control action and can become a serious problem when the state of the system is close to the sliding surface. To reduce the chattering phenomenon, we have replaced the discontinuous control by neural network to have a new controller named by sliding neural used in speed control appear to be a solution to overcome these difficulties and ensure good control performance. Squirrel cage motor has a very important role in the industry, but the existence of a certain defect reduces its performance. The faults presented in this work can be classified into switching fault of three levels Neutral Point Clamped inverter (NPC inverter). Switching faults can occur in diodes, the capacitor and IGBT switches of the inverter. This article makes a study of the speed sliding Neural (SM-ANN) controller with the existence of open switch fault of three level Neutral Point Clamped inverter to see the influence of these defects on the physical parameters of the squirrel cage motor (speed, electromagnetic torque and current). The result of the simulation shows the influence of defects on the motor parameters given by the controller. The simulation results are done by the use of Matlab/Simulink that show clearly and robustness of controller S.A.N.N.