An Adapted Constrained Multi-Objective Bayesian Optimization Under Uncertainties: Application on a Permanent Magnet Assisted Synchronous Reluctance Motor

This paper presents a robust Bayesian design optimization of a permanent magnet-assisted synchronous reluctance machine. This robust optimization takes into account the uncertainties related to the geometrical parameters of the machine as well as the uncertainties related to the characteristics of the magnetic materials. To do that, a new version of the Expected Feasible Improvement with Stepwise Uncertainty Reduction sampling (EFISUR) optimization algorithm has been implemented allowing to reduce the number of finite element simulations. The results have shown the effectiveness of the Bayesian optimization approach in enhancing machine performances while requiring minimal computational resources.