Discrete-time Neural Network Control for a Linear Induction Motor

Abstract

This paper presents a discrete-time control for a linear induction motor (LIM). First, an identifier is proposed with a nonlinear block controllable form (NBC) structure. This identifier is based on a discrete-time high order neural network trained on-line with an extended Kalman filter (EKF)-based algorithm. Then, a sliding mode control is used to achieve the purpose of tracking velocity and magnitude flux. The neural control performance is illustrated via simulations.