Torque Ripple Reduction Method of SRM Drives Using Neural Network Technique

신경회로망기법을 이용한 SRM 드라이브의 토오크리플 저감방안

The torque of SRM is developed by phase currents and inductance variation. The inductance of torque generation region is nonlinearly varied according to phase current. By this nonlinear characteristics, torque ripple can be generated on the condition of constant current. Otherwise, phase current should be controlled instantaneously in accordance with inductance to reduce torque ripple. In this paper, the control system with neural network that can reduce torque ripple is suggested. In this control system, instantaneous inductance and optimal current waveform for smallest torque ripple is obtained by neural network. And this required optimal current waveform is regulated by voltage control.