Design Optimization of Permanent Magnet Motors Considering Multiple Operating Conditions

When designing motors for complex applications, optimization with multiple operating conditions in mind can greatly improve overall efficiency while ensuring output capability. In this paper, representative points (RPs) for operating condition equivalence are extracted by a hybrid clustering algorithm. Instead of finite element analysis, a parametric equivalent magnetic network (EMN) model with improved air-gap modeling is developed to efficiently calculate motor performance. For the air gap connection permeance, this EMN employs a nonlinear cross-shaped mesh (NCSM), which improves the accuracy of the model for global optimization. With a Sobol sequence-based adaptive particle swarm optimization (Sobol-APSO) algorithm, multi-point optimization of a permanent magnet motor achieves low mass, high torque, and overall high efficiency. Finally, by comparing it with single-point optimization, the effectiveness of the proposed multi-point optimization strategy is verified.