Universidad Pontificia Comillas. Madrid (Spain)
November 23rd, 2022
Summary:
Climate change is a confirmed and ongoing phenomenon that implies several changes in temperatures and weather patterns on Earth, including an increase in temperatures, melting of the poles, and an increase in the intensity of hurricanes. The most critical factor behind climate change is the emission of greenhouse gases, with energy being responsible for 72% of emissions. In response to these findings, organizations and countries such as the European Union and the United States have formulated action plans to achieve net-zero greenhouse gas emissions in societies and economies by 2050. The proposed policies are primarily concerned with energy management, with a commitment to phase out the use of fossil fuels and promote energy supplied by renewable sources.
Renewable generation technologies are commonly connected to the electrical distribution network. As a result, the top-down paradigm is being abandoned, with a transition from a highly centralized electricity system to a considerably more decentralized system, in which distribution networks are no longer dominated by demand but instead coexist with distributed generation. This paradigm shift has increased the complexity of the distribution network operation, where the bi-directional power flows and the growing flexibility of network assets are becoming more and more common. Such complexity highlights the need for the design of algorithms that optimize the management and planning of network assets, and the need for representative synthetic distribution networks to test and compare the proposed algorithms.
This thesis addresses these needs from a network planning perspective with a compendium of four papers.
In the first part of this thesis (papers 1 and 2), a review of the published to date distribution test systems that have a U.S. architecture is carried out, pointing out the limitations they present as a test system to evaluate new algorithms developed by the scientific community. Taking this as a starting point, a novel set of algorithms is presented that allows generating large-scale distribution networks with a U.S. architecture, overcoming the limitations identified in previous test systems.
In the second part of this thesis (papers 3 and 4), distribution network planning is approached from the perspective of reinforcing an existing network through microgrids. First, it is presented a methodology that seeks to maximize network reliability while minimizing investment by installing photovoltaic panels, batteries, and diesel generation groups. Secondly, a model is formulated that aims to maximize the system's resilience while minimizing the investment, in this case, through the installation of remote-controlled switches, the undergrounding of lines, and the installation of photovoltaic panels and batteries. In both cases, the profitability of the solutions is analyzed under different cost assumptions.
Spanish layman's summary:
La creciente complejidad de la red eléctrica de distribución resalta la necesidad de redes sobre las que comprobar el funcionamiento de nuevos algoritmos y de nuevas metodologías que permitan optimizar su planificación. En esta tesis se tratan ambas perspectivas a través de cuatro publicaciones.
English layman's summary:
The increasing complexity of the electrical distribution network highlights the need for test systems to test the performance of new algorithms and new methodologies to optimize the planning procedure. Both perspectives are addressed in this thesis through four publications.
Descriptors: Electricity applications, Transmission and distribution, Other, Power distribution
Citation:
F. Postigo Marcos (2022), Algorithms for distribution system planning: applications to U.S. synthetic networks and improving resilience through microgrids. Universidad Pontificia Comillas. Madrid (Spain).
