Date of Award

Spring 4-1-2016

Document Type


Degree Name

Doctor of Philosophy (PhD)


Electrical, Computer & Energy Engineering

First Advisor

Dragan Maksimovic

Second Advisor

Li Shang

Third Advisor

Khurram K. Afridi

Fourth Advisor

Lijun Chen

Fifth Advisor

Chris A. Deline


This dissertation considers a PV system architecture based on series-connected low-voltage ac microinverters (LVACs). In contrast to standard microinverters, LVACs do not require a high step-up conversion stage, resulting in improved efficiency and reduced cost. The dissertation proposes distributed autonomous control of LVACs, which is one of the challenges of this architecture. Each LVAC is operated autonomously by a coupled-loop controller that regulates instantanous output power. Modeling of system dynamics is challenging due to the nonlinear and time-varying nature of the controller. The system is first modeled in a design-oriented manner where a linearized small-signal model is obtained and solved at every operating point under a quasi-static assumption. The modeling approach and the controller design are verified by simulations and experiments on a scaled-down system consisting of three series-connected LVACs, demonstrating appropriate ac voltage sharing across LVACs. In certain cases, such as a low compensator gain, significant total distortion harmonic (THD) is observed in ac line current. Such distortion cannot be predicted by the quasi-static approach. A pre-distortion technique is used to reduce the distortion, and experimental results demonstrate low THD (3~4%) across a wide operating range. A Fourier-expansion based method is then developed to analytically model the system, establish conditions that need to be met in order to avoid distortion, and to find a steady-state solution in terms of current harmonics. This modeling approach is verified by simulations and experiments at multiple operating points, demonstrating accurate predictions of the fundamental and the harmonics of the ac current.