Application of learning methodologies in control of power electronics drives

This paper initially discusses the learning methodologies as an underlay approach that advances several applications in power electronics and drive systems. Two applications of learning techniques based on fuzzy logic and neural networks are presented as examples of intelligent electrical drives. Initially, an adaptive learning speed control of a PMSM drive is designed, simulated and implemented. A second example deals with neuro-fuzzy control of a SR drive to reduce pulsating torque. Those techniques are corroborated by both simulation and experimental results, achieving good performance results and fast dynamic response.