Proposition of an offline learning current modulation for torque-ripple reduction in switched reluctance motors: design and experimental evaluation

A new offline current modulation using a neuro-fuzzy compensation scheme for torque-ripple reduction in switched reluctance motors is presented. The main advantage of the proposed technique is that the torque signal is unnecessary. The compensating signal is learned prior to normal operation in a self-commissioning run, capturing the necessary current shape to reduce the torque ripple. Simulation results verify first the effects of speed and then load changes on the compensator performance. Implementation of the proposed technique in a laboratory prototype shows the feasibility and accuracy of the respective offline scheme