Files
Abstract
Residential energy storage (RES) installations have exponentially increased over the last years in the United States. The cluster of residential battery storage plus solar is also known as distributed nanogrids. Although fundamentally similar to conventional large-scale microgrids, residential nanogrids offer the possibility of a widely distributed energy generation and, at the same time, allow the end-user to be ready against externalities that could cause the electric system suspension for weeks or months. Output power and energy storage scalability, battery prognosis, and optimal energy management are among the main challenges these systems face. In most cases, the implementation of control and monitoring strategies at the power converter and energy management level addresses the challenges mentioned above. This research aims to study non-linear control schemes at different hierarchical levels of the energy storage systems (ESS) for the realization of AC residential nanogrids. This dissertation develops a time-varying phasor model for multiple ESS connected in parallel. The model considers the use of non-linear droop control and virtual impedance scheme in the power converters of the ESS. Contraction theory is then employed to analyze the stability of two systems connected in parallel. We also propose a non-intrusive method to estimate lithium-ion battery modules' internal impedance used in residential ESS. Such an approach considers the non-linear intrinsic power transfer characteristic in the DC/AC conversion process for single-phase systems. Such assessment also allows us to develop a non-linear control law for the DC/DC power converter using feedback linearization analysis. Finally, we formulate a non-linear optimization problem for the energy management of islanded residential nanogrids using stochastic dynamic programming. We corroborate and validate the proposed methods and control schemes using numerical analysis and experimental results throughout the dissertation.