The design of motor speed controller using neural networks for disturbance reduction

This paper describes feedforward neural networks space vector PWM implementation of a three-level voltage-fed inverter. The general expressions of time segments of inverter voltage vectors for all the regions have been derived and the corresponding time intervals are distributed so as to get symmetrical pulse widths and neural-point voltage balancing. Space vector PWM been implemented by neural networks. Neural networks, in general, are showing very high promise for simplification of control and feedback signal processing in high performance sensorless vector-controlled drives. The proposed neural networks controller has good performance to reject the effects of disturbance. The simulation has been conducted for the verification of a proposed system, and the results show good speed responses.