Optimization and Improvement of Advanced Nonoverlapping Induction Machines for EVs/HEVs

Abstract

This paper presents a comprehensive design optimization and performance improvement guidelines for induction machines (IMs) having advanced nonoverlapping windings (ANWs). The effectiveness of various optimization approaches, such as individual optimization and single- and multi-objective global optimization using the Genetic Algorithm (GA), has been studied. To minimize the potential drawbacks of high bar copper loss, high torque ripple, and low power at high speed due to high magnet-motive force (MMF) harmonics of nonoverlapping windings (NWs), two different performance improvement approaches have been utilized: (a) to redesign the rotor structure to reduce the parasitic effects such as torque ripple and additional bar copper losses due to air-gap flux density harmonics; (b) to increase the stack length to improve the torque at the constant-power region. It has been revealed that the proposed ANW IMs with bridges in their rotor openings, particularly with u-shaped bridges, show better performance in terms of torque ripple, bar copper loss, and bar current density. By using the proposed design method, an advanced IM (AIM) can achieve a 5.3% higher efficiency with ~34% shorter total axial length, compared to its conventional IM (CIM) counterpart with integer-slot distributed windings (ISDWs). A time-stepping 2-D finite element analysis (FEA) based nonlinear magnetic field solution program has been employed to perform all the parametric analyses, optimizations, and evaluate the optimal solutions and improved designs. Moreover, in order to show the reliability of the FEA predictions performed in this study, the FEA predictions of globally optimized CIM are validated by experimental measurements.