Neural network mapping of hybrid FOC-DTC induction motor drive

This paper presents the mapping of a hybrid of two most popular control strategies for induction motor drive: field oriented control (FOC) and direct torque control (DTC), with the multilayer feedforward neural network. One neural network is used to map FOC scheme and another network is used to map DTC scheme. The training is made separately with the input-output data obtained from the simulation result of the structural based models. Two trained networks are combined to represent the neural network hybrid control (NNHC). The criterion for switching between the two control schemes is the operating speed of the motor. The simulation results show that this proposed scheme is characterized by fast dynamic performance and good steady state characteristic. Real time implementation of NNHC scheme is discussed.