FIELD-ORIENTED INDUCTION MULTI-MACHINE DRIVE USING SINGLE-INVERTER AND DIAGONAL RECURRENT NEURAL NETWORK

Abstract

This paper describes a strategy to develop a model electric car control of induction multi-machine drive using single-inverter under unbalanced load conditions. It is discusses improving existing multi-machine field-oriented control methods using Diagonal Recurrent Neural Network (DRNN), and analyses and compares them from dynamic performance. By this method the speed and current difference for both motors due to load fluctuation was minimized. Therefore, the motors will remain in sync without any slippage speed on the second drive. Based on simulation results is known that the current steady state, the controller DRNN produce both motors speed difference is smaller (2 rad/sec) when compared using the PI controller (7 rad/sec).