Indirect Detection Method of Rotor Position Based on DE-SVM

AbstractIn view of the defects and deficiencies of existing detection methods of rotor position for Switched Reluctance Motor SRM, an indirect Detection Method DM based on DE-SVM for Support Vector Machine SVM rotor position is proposed. This method uses the three-phase current and flux linkage within the full angle domain of SRM as input and rotor position angle as output, and utilizes the strong nonlinear mapping capability of SVM to create a predication model for these three parameters offline. The strong global optimization capability of Differential Evolution DE Algorithm is then employed based on the deviation between actual rotor position and model output to optimize the prediction model online, thereby realizing sensorless detection of SRM rotor position. The simulation result shows that this method can accurately predict the position of SRM rotor.