Control Neuro-Fuzzy of a Dual Star Induction Machine (DSIM) supplied by Five-Level Inverter

摘要

This work relates to a hybrid scheme (Neuro-Fuzzy) for speed control of a dual star induction motor (DSIM) with enhanced performance. A 4-layer network is utilized to set the Fuzzy elements in order to minimize error square. To control this machine, two five-level inverters with PWM techniques are introduced and an indirect field oriented method is used. Simulation results are presented for the NF controller; it is observed that the NFC gives better responses and robustness for the speed control of this machine with its load disturbances and parameter variations, such as increased rotor resistance and moment of inertia. The results are compared with results obtained from a conventional inverter. Notably, there is a great drop in the stator currents, and the magnitude of the pulsating electromagnetic torque is reduced for a five-level inverter compared with a conventional inverter.