Torque Ripple Minimization in a Sensorless Switched Reluctance Motor Based on Flexible Neural Networks

The switched reluctance motor (SRM) has obtained great potential as an adjustable speed application due to its outstanding merits. However, its application is limited because of the rotor position sensors and torque ripple. This paper proposes an approach to tackle sensorless control and torque ripple minimization of SRM by using flexible neural networks (FNN) which have many great advantages, such as less nerve cells and quick learning speed. Two FNN are built: through measurement of the phase flux linkages and phase currents, the first one is able to estimate the rotor position, thereby facilitating elimination of the rotor position sensor. The second one is for the estimation of the reference currents with a desired torque, then the real currents in the armatures are adjusted according to the reference values, therefore the torque ripple generated by the non-ideal current waveforms is minimized for a sensorless SRM. Simulation and experimental results illustrate the improvements of the proposed method compared with traditional controller.