Name: | Description: | Size: | Format: | |
---|---|---|---|---|
5.99 MB | Adobe PDF |
Authors
Abstract(s)
Tipicamente as redes elétricas de distribuição apresentam uma topologia parcialmente
malhada e são exploradas radialmente. A topologia radial é obtida através da abertura das
malhas nos locais que otimizam o ponto de operação da rede, através da instalação de
aparelhos de corte que operam normalmente abertos. Para além de manterem a topologia
radial, estes equipamentos possibilitam também a transferência de cargas entre saídas,
aquando da ocorrência de defeitos. As saídas radiais são ainda dotadas de aparelhos de corte
que operam normalmente fechados, estes têm como objetivo maximizar a fiabilidade e isolar
defeitos, minimizando a área afetada pelos mesmos. Assim, na presente dissertação são
desenvolvidos dois algoritmos determinísticos para a localização ótima de aparelhos de corte
normalmente abertos e fechados, minimizando a potência ativa de perdas e o custo da energia
não distribuída.
O algoritmo de localização de aparelhos de corte normalmente abertos visa encontrar a
topologia radial ótima que minimiza a potência ativa de perdas. O método é desenvolvido
em ambiente Matlab – Tomlab, e é formulado como um problema de programação
quadrática inteira mista. A topologia radial ótima é garantida através do cálculo de um
trânsito de potências ótimo baseado no modelo DC. A função objetivo é dada pelas perdas
por efeito de Joule. Por outro lado o problema é restringido pela primeira lei de Kirchhoff,
limites de geração das subestações, limites térmicos dos condutores, trânsito de potência
unidirecional e pela condição de radialidade.
Os aparelhos de corte normalmente fechados são localizados ao longo das saídas radiais
obtidas pelo anterior algoritmo, e permite minimizar o custo da energia não distribuída. No
limite é possível localizar um aparelho de corte normalmente fechado em todas as linhas de
uma rede de distribuição, sendo esta a solução que minimiza a energia não distribuída. No
entanto, tendo em conta que a cada aparelho de corte está associado um investimento, é
fundamental encontrar um equilíbrio entre a melhoria de fiabilidade e o investimento. Desta
forma, o algoritmo desenvolvido avalia os benefícios obtidos com a instalação de aparelhos
de corte normalmente fechados, e retorna o número e a localização dos mesmo que minimiza
o custo da energia não distribuída.
Os métodos apresentados são testados em duas redes de distribuição reais, exploradas com
um nível de tensão de 15 kV e 30 kV, respetivamente. A primeira rede é localizada no distrito
do Porto e é caraterizada por uma topologia mista e urbana. A segunda rede é localizada no
distrito de Bragança e é caracterizada por uma topologia maioritariamente aérea e rural.
Usually, distribution networks have a partial physically meshed topology, but are run in a radial way. The radial topology is achieved through the mesh opening in the network local to optimize the operating point, by switches installation which operate normally open. In addition to maintaining the radial topology, these devices also allow the transfer of loads between feeders, on the occurrence of defects. The radial feeders are also provided with switches which operating normally closed, these are designed to maximize the reliability and to isolate defects, minimizing the area affected by the same. Therefore, in this thesis are developed two deterministic algorithms for optimal location of normally open and close switches, minimizing the active power losses and energy not supplied cost. The optimal location of normally open switch algorithm aims to find the optimal radial topology that minimizes the active power losses. The method is developed in Matlab – Tomlab, and is formulated as a mixed integer quadratic programming problem. The optimal radial topology is guaranteed by calculating an optimal power flow based on the DC model. The objective function is given for the losses by Joule effect. On the other hand the problem is constrained by the first law of Kirchhoff, generation substation limits, lines thermal limits, unidirectional power flow and the radial condition. The normally closed switches are located along the radial feeders obtained by the above algorithm, and allows minimizing the energy not supplied cost. In the limit, it is possible to install one normally close switch on all lines of a given feeder, which is the solution that minimizes the energy not supplied. However, taking into account that each switch is associated with an investment, it is essential to find a balance between improvement of reliability and investment. As a result, the developed algorithm evaluates the benefits obtained with the normally close switch installation, and returns the number and location of switch that minimize the energy not supplied cost The methods presented are tested on two real distribution networks, operated with a voltage level of 15 kV and 30 kV, respectively. The first network is located in Porto district and is characterized by a mixed urban topology. The second network is located in Bragança district and is characterized by a mostly overhead and rural topology.
Usually, distribution networks have a partial physically meshed topology, but are run in a radial way. The radial topology is achieved through the mesh opening in the network local to optimize the operating point, by switches installation which operate normally open. In addition to maintaining the radial topology, these devices also allow the transfer of loads between feeders, on the occurrence of defects. The radial feeders are also provided with switches which operating normally closed, these are designed to maximize the reliability and to isolate defects, minimizing the area affected by the same. Therefore, in this thesis are developed two deterministic algorithms for optimal location of normally open and close switches, minimizing the active power losses and energy not supplied cost. The optimal location of normally open switch algorithm aims to find the optimal radial topology that minimizes the active power losses. The method is developed in Matlab – Tomlab, and is formulated as a mixed integer quadratic programming problem. The optimal radial topology is guaranteed by calculating an optimal power flow based on the DC model. The objective function is given for the losses by Joule effect. On the other hand the problem is constrained by the first law of Kirchhoff, generation substation limits, lines thermal limits, unidirectional power flow and the radial condition. The normally closed switches are located along the radial feeders obtained by the above algorithm, and allows minimizing the energy not supplied cost. In the limit, it is possible to install one normally close switch on all lines of a given feeder, which is the solution that minimizes the energy not supplied. However, taking into account that each switch is associated with an investment, it is essential to find a balance between improvement of reliability and investment. As a result, the developed algorithm evaluates the benefits obtained with the normally close switch installation, and returns the number and location of switch that minimize the energy not supplied cost The methods presented are tested on two real distribution networks, operated with a voltage level of 15 kV and 30 kV, respectively. The first network is located in Porto district and is characterized by a mixed urban topology. The second network is located in Bragança district and is characterized by a mostly overhead and rural topology.
Description
Keywords
Aparelho de Corte Normalmente Aberto e Fechado Energia Não Distribuída Potência Ativa de Perdas Redes de Distribuição Active Power Losses Distribution Networks Energy Not Supplied Normally Open and Closed Switch