Method of Sensorless Control for Switched Reluctance Electric Drive

The paper presents the development and implementation of an original method of sensorless control for switched reluctance electric drive based on the mathematical apparatus of discrete Fourier transform and artificial neural networks. In order to indirectly determine the angular position of the rotor of the switched reluctance machine, the rise fronts of the phase currents during the successive energizing of phases are considered as responses to the fluctuations of phase voltages. The waveform of these responses depends on the relative position of the rotor poles and the stator poles when the phases are energized. The discrete Fourier transform is used to transform this response into an array of harmonic components, and the artificial neural network identifies the actual angular position of the rotor that it had when the phases were energized. Then, the non-synchronization of the rotor rotation relative to the stator field is detected and synchronization of phase energizing and energy withdrawing with the position of the rotor is provided. The author determines a set of harmonic components that is sufficient for identification. The proposed sensorless control method can be applied to switched reluctance machines of various types by bringing their designs to basic form. Adaptation of sensorless control to change in voltage supply is provided. Compared to other methods based on the analysis of current waveform, the proposed method allows improving the energy characteristics of the electric drive and to reduce the influence of noise in the measuring circuits and switching noise as well.