Commutation signal identification for switched reluctance motors based on fuzzy neural networks

This paper presents a new approach of position sensorless control for switched reluctance motors (SRM) based on fuzzy neural networks (FNN). In this method, two different modules are established to identify phase commutation signals during the starting period and steady running period respectively. In each module, FNNs are used to map the relationship of phase current, flux linkage and phase commutation signals. Commutation signals for each phase are identified continuously by two FNNs. One of them estimates the turn-on signal; the other identifies the turn-off signal. Once the motor reaches the given speed, commutation signals will be provided by the steady running module instead of the starting module. The simulation results show that this method can exactly achieve phase commutation identification, and the system can operate steadily with the proposed position sensorless control method.