Tooth shape optimization of switched reluctance motors for improved torque profiles

This paper investigates the optimal design of switched reluctance motors to obtain higher average torque with lower torque ripple content. This is achieved by shaping both stator and rotor tooth surface regions, using finite element analysis to evaluate the designs. The optimization problem is solved using two methods, namely, a gradient ascent method and a genetic algorithm. In the gradient ascent method, the tooth surfaces are described by a set of flexible nodes. In the genetic algorithm, the tooth surfaces are defined using quadratic Bézier curves. The two methods provide different insights on the optimal tooth shapes, and thus can be thought to be complementary to each other.