Sensorless position control and torque ripple minimization of switched reluctance machine

Because of need for a position sensor and producing acoustic noise the Switched Reluctance Machines (SRM) are not widely used in industrial applications. However, if these two drawbacks are solved in an important degree, SRMs can be used instead of many machines in today?s industry.In the literature, the primary cause of acoustic noise is given as radial forces. If the effect of the radial force is reduced then the acoustic noise will also reduced. One of the aims of this thesis is that reducing radial forces which is a cause of acoustic noise. Hence, some improvements can be made on control and/or magnetic circuits.For this purpose, after determining the size of which 8/6 SRM is designed and based on improving motor magnetic circuit, four types of SRM are designed and are compared. In order to compare with machines used in industry, submersible pump type of 8/6 SRM was produced. Finite Elements Method (FEM) was also used to perform magnetic analysis of the designed models. ANSYS 10.0 package program was used in the analysis. The result of the investigation showed 19.03% reduction in radial force compared with the standard SRM.The closed-loop speed and current controls with and without sensors of SRMs constitute the second stage of this study. It is planned to find a solution for two important problems with this control circuit. One of them is to reduce the acoustic noise and the other is to achieve sensorless operation of SRM. SRM?s magnetic characteristics which are attained by FEM have been used for predicting SRM?s sensorless position. Flux and current are input to artificial neural networks and position is attained as an output.