GA-FEM Method for Optimum Current Search for Torque Ripple Mitigation in Nonsinusoidal PMSM and Analysis of On-Load Parameters

Permanent magnet synchronous motors with spatial harmonics can develop significant torque ripple under sinusoidal stator current feeding strategy. Usual approaches to reduce torque ripple aim the design of stator current harmonics, based on the machine electrical parameters, to induce compensating torque pulsations. However, the machine parameters are highly dependent on the electric load and the current harmonic content. Thus, these dependencies impose a challenge on the search for optimum stator currents. To overcome this drawback, this paper presents a genetic algorithm (GA) optimum current search procedure based on the machine design assisted by the Finite Element Method (FEM). Additionally, the variation of the onload estimated electrical parameters is analyzed considering different current feeding strategies. The results demonstrate that, according to FEM analysis, a null torque ripple with an increased torque per Ampère ratio is achieved and the parameter analysis demonstrates their different variation according to the current strategy.