Managing Uncertainties of Permanent Magnet Synchronous Machine by Adaptive Kriging Assisted Weight Index Monte Carlo Simulation Method

Abstract

For electromagnetic devices, there exist many uncertainty sources during design stage, performance simulation, components manufacturing and assembling. These uncertainties may deteriorate the practical operation performance of electrical machines from the designed one or result in violation of constraint conditions to a certain extent. To improve robust performance of a permanent magnet synchronous machine (PMSM), uncertainties should be taken into account in its whole life cycle. To manage uncertainties in the design stage of electrical machines, this paper makes some explorations of reliability-based optimal design. Firstly, the basic knowledge of reliability based algorithm is reviewed. Then the efficient adaptive sampling strategy and the surrogate model construction approach are introduced to reliability analysis. Based on these strategies, the new reliability analysis approach is named adaptive Kriging assisted weight index Monte Carlo simulation method. Finally, the new approach is applied to reliability-based optimal design of permanent magnet synchronous motor for cogging torque reduction.