Reduction of the torque ripple in permanent magnet actuators by a multi-objective minimization technique

In this paper a minimization technique is used to reduce the torque ripple of permanent magnet actuators. In order to do this a global stochastic (1+1) evolution strategy algorithm coupled to a filled function acceleration technique is utilized. The cost function is defined as a combination of the e.m.f. harmonics induced in the stator windings and the harmonic components of the magnet permeance. The relative importance of the different contributions are taken into account by penalty coefficients. The method described is used for a 3-phase, 6-pole Permanent Magnets (PM) synchronous motor.