Date of Award

2-1-2022

Publication Type

Dissertation

Degree Name

Ph.D.

Department

Electrical and Computer Engineering

First Advisor

N. Kar

Second Advisor

B. Balasingam

Third Advisor

M. Azzouz

Keywords

Analytical model, Characterization, Electric vehicle, Gallium nitride, Traction drives, Wide bandgap

Rights

info:eu-repo/semantics/embargoedAccess

Abstract

This thesis explores the techniques of characterization and applications of gallium nitride (GaN) semiconductor switching devices in power conversion areas, especially permanent magnet synchronous motor (PMSM) based electric vehicle (EV) traction drives to achieve improved system performances.

At first, an investigation has been conducted to report the progresses of wide bandgap (WBG), especially GaN devices in power conversion applications. Based on the motivations to bridge the knowledge gap, the switching transient performance of enhancement–mode gallium nitride high–electron–mobility transistor (eGaN HEMT) and its impaction on switching energy loss have been chosen to start the research due to its technical challenges. Based on the investigation and overall project outlook, eGaN HEMT based hard switching topology is proposed to conduct further research to explore the feasibility to improve GaN motor drive efficiency, current ripples, and harmonics performances under higher switching frequency and more critical dead time setup, which is not possible for traditional silicon insulated gate bipolar transistor (IGBT) based motor drive.

In eGaN–based power inverter, the switching energy loss increases naturally along with the switching frequency, which is the dominant loss component and the bottleneck limiting both the device safety range and system performances. Consequently, the switching transient analysis and its impact on switching energy loss are of significance for EV powertrain engineers in research and development (R&D) activities. In this research, a practical analytical model is proposed that can emulate the transistor’s dynamic behavior and predict the switching energy loss without consuming too much computing resources upon public access information, for example, integrated circuits (IC) datasheet, simulation program with integrated circuit emphasis (SPICE) model and application notes. This methodology opens the door to provide guidelines for researchers to conduct circuit design and performance optimization without relying on the business to business (B2B) collaborations. The simulation results and experimental validation demonstrate that the proposed model is a computationally inexpensive and straightforward switching transient model. According to the equivalent circuits, the accuracy of the analytical model mainly depends on the parasitic parameters inside the eGaN HEMT transistor packaging and external circuits under different switching stages. It is challenging to extract the parasitic parameters accurately due to nonlinearity and the complex correlation with induced ringing and electromagnetic interference (EMI). Through a comprehensive study of different methodology, both the internal and external parasitic parameters of half–bridge configurations have been extracted accurately compared with the traditional curve–fitting methods. The internal parasitic parameters including junction capacitance, internal inductance and resistances are extracted via interpreting SPICE library files equations to MATLAB scripts and the external parasitic parameters are extracted by finite element method (FEM) on Ansys Q3D platform. All parasitic parameters have been verified against the datasheet and show good agreement. The success in accurately extraction of all parasitic parameters enhances the accuracy of proposed analytical modeling and paves the way to amplify advantage the high switching capability of eGaN HEMT to achieve optimized performances for any power converter applications where half–bridge configuration is based, especially space vector pulse width modulation (SVPWM) hard–switching three–phase PMSM drive.

Upon previous investigations, due to technical challenges and the immature reality of GaN power switching devices, the insufficiency supply of mature designs of GaN–based motor drive is still a bottleneck that restricts EV powertrain. Only evaluation board (EVB) level prototypes from dominant GaN switching device manufacturers are available on market in tailor–made mode. Efforts to transfer the conventional Si–based engineering skills and experiences have been invested to achieve eGaN HEMT’s full potential as a power semiconductor technology throughout this research. The whole design and development process of eGaN HEMT based three–phase inverter is presented including key components selection, gate driver design to mitigate oscillation, critical power and gate loop layout. The performances of the GaN–based prototypes are analyzed and verified against experiments. In the last part, the conclusion was drawn and future work is forecasted in this area.

Available for download on Friday, June 07, 2024

Share

COinS