Date of Award

December 2021

Degree Type

Dissertation

Degree Name

Doctor of Philosophy

Department

Engineering

First Advisor

Adel Nasiri

Committee Members

Necmi Altin, Robert Cuzner, Kaan Kuzu, Brian Armstrong

Keywords

Electric vehicles, Extreme fast charger, Modified Lyapunov function, RMS current minimization, Total power loss minimization, Triple phase shift control

Abstract

There is a shift in the technology of vehicles from gas and diesel engines to electric vehicles(EVs). Approximately ten million EVs were available globally in 2020 and it is projected that number will reach 145 million by 2030. To power the increasing number of EVs, the number of EV charging stations is growing at a significant rate. In order to provide flexibility and longer driving ranges to customers, the trend is to install DC fast charging stations. These chargers demand high power at low voltage, which our existing electrical distribution system cannot accommodate without major upgrades. Currently, bulky transformers are used to step down to a lower voltage level in order to directly connect to the medium voltage (MV) utility grid. This results in lowered efficiency and increased cost and size of the charging system. This dissertation formulates a solution to this significant problem. The proposed work addresses the development of a control scheme and multilevel converter for MV AC to low voltage DC intended for fleet EV charging stations. This architecture removes the shortcomings of the existing systems and offers modular structure, scalability, galvanic isolation, and high efficiency. This topology is investigated for a 1 MW system connected to the 13.8 kV AC grid to create 1 kV DC for EV charging. A robust control structure is proposed for voltage balancing and current sharing among various stages of the converter. The converter and high frequency transformer are also investigated for the DC/DC conversion. In order to mitigate power losses, root mean squared current minimization and power loss minimization controls are evaluated and the power loss minimization method is found to be superior. A three module single-phase prototype using hardware in the loop is developed and tested to verify the viability of the system.

Share

COinS