Doctoral Dissertations

Date of Award

8-2025

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Major

Mechanical Engineering

Major Professor

Hans Desmidt

Committee Members

Seddik Djouadi, Zhenbo Wang

Abstract

This thesis presents the modeling, design, and evaluation of Motor/Generator Integrated Permanent Magnet Gear (MIPMG) systems, with a focus on coaxial magnetic gear (CMG) architectures and their integration into Magnetic Electric Continuously Variable Transmission (MECVT) systems. A theoretical torque model is developed to characterize the magnetic coupling between inner and outer rotors through a ferromagnetic modulator. This model captures the sinusoidal torque–angle relationship based on pole pair configurations and material properties. A key contribution is the derivation of a maximum torque line, allowing early estimation of torque limits without relying on repeated finite element analysis. Parametric studies show that increasing the outer pole pair number and optimizing magnet dimensions significantly improves torque density in space-constrained applications. Dynamic simulations validate the model under both partial and steady-state engagements, demonstrating consistent torque behavior under realistic loads. Loss analysis reveals that eddy current loss dominates at high speeds, contributing to a notable drop in efficiency. Despite this, CMG-based systems show up to 17% total mass savings compared to multi-stage planetary gear transmissions at high gear ratios, due to shared magnetic components and compact design. A system-level comparison between CMG and planetary gear-based MECVTs indicates that CMG systems are more suitable for high-ratio, mass-sensitive scenarios, while planetary systems remain advantageous in low-to-moderate ratios where efficiency is critical. This work provides a comprehensive framework for developing high-torque, lightweight magnetic transmission systems suitable for electric vehicles, aerospace propulsion, and robotics.

Download
Files over 3MB may be slow to open. For best results, right-click and select "save as..."

Share

COinS