This study aims to reduce torque ripple and cogging torque in interior permanent magnet synchronous motors (IPMSMs) by optimizing a rotor-notch design robust to manufacturing tolerances. It seeks to balance performance and reliability using Six Sigma (6s) principles.
A robust optimization framework is proposed for rotor-notch design. The notch parameters are modeled as normally distributed random variables to represent manufacturing deviations. Finite-element analysis, Kriging surrogate modeling, Monte Carlo sampling and a multi-island genetic algorithm are combined to evaluate and optimize the statistical responses of average torque, torque ripple and cogging torque. Two rotor topologies, namely, Double-V and V + 1 IPMSMs, are investigated and compared with deterministic optimization. Magnetic flux-density distributions are also analyzed to explain the physical influence of the optimized notches.
The results show that uncertainty-based optimization reduces the sensitivity of electromagnetic responses to dimensional deviations. Compared with deterministic optimization, the robust design does not always provide the best nominal value for every objective, but it generally gives a more stable response distribution. For the Double-V rotor, deterministic optimization gives lower nominal torque ripple and cogging torque, whereas uncertainty-based optimization reduces the dispersion of cogging torque and average torque. For the V + 1 rotor, uncertainty-based optimization achieves better mean average torque and cogging torque, as well as smaller standard deviations of all three performance indices, although its nominal torque ripple is higher than that of deterministic optimization. The flux-density distributions show that the optimized notches mainly redistribute the local magnetic field near the rotor surface, magnet ends and air-gap region without causing large-scale magnetic saturation. Additional electromagnetic performance comparisons further confirm that the optimized notch designs do not significantly deteriorate the main operating characteristics.
This work integrates Six Sigma-based uncertainty analysis with surrogate-assisted finite-element optimization for robust rotor-notch design in IPMSMs. The comparison between two rotor topologies shows that uncertainty propagation and robustness improvement are topology-dependent. The proposed method provides a practical approach for improving the consistency of torque performance in mass-produced IPMSMs.
