| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.72 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Cada vez mais existe a necessidade de uma transição energética para um modelo mais
sustentável, resiliente e descarbonizado devido ao aumento dos preços dos combustíveis
fósseis, às mudanças climáticas e à crescente preocupação ambiental. Essa transição
implica a passagem de um sistema elétrico sustentado por combustíveis fósseis para um
sistema elétrico baseado em energias renováveis. Por sua vez, para ser integrada na rede
de distribuição, a energia elétrica descentralizada tem de cumprir determinadas normas
e padrões. Estas normas ditam os princípios operacionais básicos, operação da rede de
distribuição e resposta do sistema a condições anormais de funcionamento da rede
elétrica. Recomendam ainda que a interligação dos recursos com a rede de distribuição
deve ocorrer com um fator de potência unitário. Assim, para garantir o funcionamento
destes recursos é necessário o uso de algoritmos de sincronização com a rede elétrica.
Concretamente, é realizado um estudo comparativo que considera os modelos
comummente usados na literatura (Notch-PLL e SOGI-PLL). Nesta dissertação, para
mitigar algumas desvantagens associadas a estes algoritmos de sincronismo com a rede
são usados controladores fracionários. Estes controladores consistem numa
generalização dos controladores clássicos, sendo caraterizados por uma função
transferência de ordem fracionária. De modo a determinar os valores dos ganhos de cada
controlador, bem como as suas respetivas ordens fracionárias, foi considerado um
problema de otimização. Portanto, para resolver este problema diferenciadamente foram
selecionados quatro métodos meta-heurísticos: Differential Evolution, Grey Wolf
Optimizer, Particle Swarm Optimization e Whale Optimization Algorithm. Para cada
destes métodos os resultados obtidos para os ganhos dos controladores foram analisados
em função da integral do erro absoluto (IAE). Já para realizar as aproximações das
ordens fracionárias foi utilizada a técnica de aproximação Oustaloup. Para avaliar e
comparar o desempenho dos controladores clássicos e fracionários nos algoritmos de
sincronização com a rede elétrica, os algoritmos usados nesta dissertação foram
submetidos a três casos de estudo em ambiente de simulação e ambiente real de
operação.
There is an increasing need for an energy transition to a more sustainable, resilient, and decarbonized model due to fossil prices, climate change and growing environmental concerns. This transition implies the transition from an electrical system sustained by fossil fuels to an electrical system based on renewable energies. In turn, to be integrated into the distribution grid, decentralized electrical energy must comply with certain norms and standards. These standards dictate the basic operating principles, operation of the distribution grid and the system's response to abnormal operating conditions of the electrical grid. They also recommend that the interconnection of resources with the distribution grid should occur with a unity power factor. Thus, to guarantee the functioning of these resources, it is necessary to use synchronization algorithms with the electrical grid. Specifically, a comparative study is carried out that considers the models commonly used in the literature (Notch-PLL and SOGI-PLL). In this dissertation, to mitigate some disadvantages associated with these network synchronization algorithms, fractional controllers are used. These controllers consist of a generalization of the classical controllers, being characterized by a fractional order transfer function. To determine the gains values of each controller, as well as their respective fractional orders, an optimization problem was considered. Therefore, to solve this problem differently, four metaheuristic methods were selected: Differential Evolution, Gray Wolf Optimizer, Particle Swarm Optimization and Whale Optimization Algorithm. For each of these methods, the results obtained for the controller gains were analysed as a function of the absolute error integral (IAE). To perform the approximations of the fractional orders, the Oustaloup approximation technique was used. To evaluate and compare the performance of classical and fractional controllers in the synchronization algorithms with the electrical grid, the algorithms used in this dissertation were submitted to three case studies in a simulation environment and a real operating environment.
There is an increasing need for an energy transition to a more sustainable, resilient, and decarbonized model due to fossil prices, climate change and growing environmental concerns. This transition implies the transition from an electrical system sustained by fossil fuels to an electrical system based on renewable energies. In turn, to be integrated into the distribution grid, decentralized electrical energy must comply with certain norms and standards. These standards dictate the basic operating principles, operation of the distribution grid and the system's response to abnormal operating conditions of the electrical grid. They also recommend that the interconnection of resources with the distribution grid should occur with a unity power factor. Thus, to guarantee the functioning of these resources, it is necessary to use synchronization algorithms with the electrical grid. Specifically, a comparative study is carried out that considers the models commonly used in the literature (Notch-PLL and SOGI-PLL). In this dissertation, to mitigate some disadvantages associated with these network synchronization algorithms, fractional controllers are used. These controllers consist of a generalization of the classical controllers, being characterized by a fractional order transfer function. To determine the gains values of each controller, as well as their respective fractional orders, an optimization problem was considered. Therefore, to solve this problem differently, four metaheuristic methods were selected: Differential Evolution, Gray Wolf Optimizer, Particle Swarm Optimization and Whale Optimization Algorithm. For each of these methods, the results obtained for the controller gains were analysed as a function of the absolute error integral (IAE). To perform the approximations of the fractional orders, the Oustaloup approximation technique was used. To evaluate and compare the performance of classical and fractional controllers in the synchronization algorithms with the electrical grid, the algorithms used in this dissertation were submitted to three case studies in a simulation environment and a real operating environment.
Description
Keywords
Algoritmos de Sincronismo com a Rede Controladores Fracionários Métodos Meta-Heurísticos Pll Técnica de Aproximação