Submodule Integrated Gallium Nitride Converters for Distributed Maximum Power Point Tracking PV Applications

Abstract

Partial shading in conventional Photovoltaic (PV) systems with series-parallel configuration of PV panels causes two main problems. Firstly, shaded panels are bypassed and do not contribute any power. Secondly, multiple maxima arise in the PV curve which complicates Maximum Power Point Tracking (MPPT). These problems result in significant loss in captured energy. To resolve these problems, Submodule Integrated Converter (SIC) is developed for Distributed Maximum Power Point Tracking (DMPPT) PV applications. Another point of concern in PV systems is the conversion efficiency of the power converter. To achieve high conversion efficiency, Gallium Nitride (GaN) Field Effect Transistors (FETs) are used in synchronous buck converter topology. GaN FETs have low on-resistance and smaller switching times; therefore, they provide higher efficiency. They can also be operated at high frequencies which reduces the size requirement of passive components in a converter. The GaN converter is designed for MPPT and implemented on a Printed Circuit Board (PCB) along with digital control implementation. Microcontroller and measurement filters are housed on the board to implement MPPT and other Pulse Width Modulation (PWM) functions. The board is self-sufficient and all the devices are powered from a single input which will come from a PV panel's submodule to which the board is connected. The prototype GaN converter shows peak power terrain efficiency of 99.07%. Furthermore, a simplified model for enhancement mode Gallium Nitride (eGaN) Field Effect Transistor (FET) is also developed to investigate the switching characteristics of eGaN FETs. A 12V/24V boost converter is prototyped using eGaN FETs to validate the proposed model. The loss model is simple and easy to reproduce; and can be used to determine switching times and switching losses using parameters given in datasheets. The simulation results of the approximate switching model coincide to a great extent with the experimental results.

Date of Award	May 2014
Original language	American English
Supervisor	Michael Weidong Xiao (Supervisor)