On-line instant torque estimation of switched reluctance motor using adaptive B-spline neural network

A novel on-line instant torque estimation scheme for a switched reluctance motor (SRM) is presented. In the proposed method, an adaptive B-spline neural network is used to learn the non-linear flux-linkage and torque characteristics of a SRM. Due to the local nature of its generalisation properties, the training of the B-spline neural network is accomplished on-line and in real-time, and the system does not require a priori knowledge of the machine's electromagnetic characteristics. The potential of the method is demonstrated successfully in simulation and experimentally using a 300 W 12/8 3-phase SRM.