Robust Design Optimization of Electric Motor Considering Assembly Imperfection of Segmented Stator Core

Abstract

As industrial applications, permanent magnet (PM) motors are well known for their high efficiency and high power characteristics. To improve the productivity of the PM motors, segmented stator core with non-overlapping windings have been widely used. However, assembly imperfections are likely to occur in the segmented stator core while producing various stator tolerances. If manufacturing tolerances are neglected in a motor designing process, a motor output performance can deviate from what the motor designer intended. Considering assembly imperfections of the segmented stator core, an equivalent modeling method for the segmented stator core and a robust design process are proposed in this thesis. Using the equivalent modeling method, modeling the segmented stator core with the assembly imperfections is simplified. Through the proposed robust design process, a motor design can be obtained that has a robust performance to assembly imperfection of segmented stator core. The followings are the overview of this thesis. First, the conventional magnetic circuit design of electric motor without considering the manufacturing tolerances is presented. By reviewing the design results, it is shown that the motor performance of the conventionally designed motor deviates from its expected value. Second, the design parameters related to motor performances are defined based on an equivalent magnetic circuit of electric motor. Then, the effects of manufacturing variations of the parameters on air-gap magnetic field distribution are investigated. Next, effects on back electro-motive force (BEMF) and cogging torque are compared. Accordingly, it is known that the cogging torque is sensitive to the manufacturing tolerances. Third, the manufacturing tolerances in the conventional design model are investigated by measuring the dimension of the manufactured prototype. Through the investigation, it is found that the main cause of the performance deviation is the assembly imperfections of the segmented core. Fourth, an equivalent modeling method for the segmented core is proposed considering assembly imperfections. By using the proposed modeling method, the segmented stator core with the assembly imperfections can be simply modeled with a few uncertainty parameters. Fifth, a sensitivity analysis method is proposed concerning the robustness of motor performance to the assembly imperfections of the segmented stator core. Through the proposed sensitivity analysis, design variables that have statistically significant effects on the robustness of motor performance can be identified. Sixth, a robust design optimization is performed to minimize the effects of the assembly imperfections of the segmented stator core. Accordingly, the robust design optimum model is determined. Finally, the validity of the proposed robust design process is confirmed by Monte Carlo simulation (MCS) and physical experiments.