Optimization of photovoltaic power plant allocation and smart inverter settings to maximize the distribution network’s hosting capacity

• Ano de defesa: 2025
• Autor(a) principal: Jaramillo Leon, Brian Daniel [UNESP]
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Distribution network; DG allocation; Hosting capacity; Metaheuristic algorithm; Photovoltaic system; Smart inverter; Algoritmo meta-heurístico; Alocação de geração distribuida; Capacidade de hospedagem; Inversor inteligente; Rede de distribuição; Usina solar fotovoltaica
• Link de acesso: https://hdl.handle.net/11449/261096

As the integration of solar photovoltaic (PV) power plants into electrical grids grows, it becomes critical to determine the maximum PV capacity that can be safely connected to distribution networks without compromising the grid operation and service quality. This thesis formulates an optimization problem to maximize the PV hosting capacity (HC) in a medium-voltage distribution feeder by allocating (siting and sizing) ground-mounted PV power plants, considering both the power factor and voltage-reactive power (Volt-VAr) control functions of their corresponding smart inverters (SIs). A simulation-optimization framework that integrates Python and OpenDSS software is proposed to determine the best number, location, and size of PV systems to be installed on a distribution feeder, as well as the best control set-points of the PV inverters. This thesis also quantifies the location-specific PV HC (PVHC) by connecting a single PV plant at a time (i.e., centralized allocation) at each candidate location and considering the maximum values of feeder bus voltages and thermal loadings through the conductors as performance metrics. Differential evolution (DE) and vortex search (VS) algorithms are used to solve the proposed optimization problem. The connection of one, two, and three PV power plants is tested in an Ecuadorian distribution feeder model. The results indicate that VS presents less variability in its solutions and a higher mean objective function value than DE, and installing two PV power plants with their SIs operating with the Volt-VAr control function produces the highest PVHC.