| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| Documento principal | 1.38 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A implementação de um mercado eléctrico liberalizado e a crescente integração de energia
eólica na rede eléctrica, particularmente em Portugal, induzem novos desafios associados à
crescente competitividade no sector eléctrico entre empresas produtoras e à elevada
volatilidade e intermitência inerentes ao vento. Assim, torna-se indispensável para os agentes
de mercado a existência de ferramentas computacionais mais eficientes que permitam obter
previsões fiáveis e rigorosas dos preços da energia eléctrica e da potência eólica. Estas
previsões possibilitam desenvolver melhores estratégias de oferta no mercado, maximizando o
lucro, e optimizando a exploração dos recursos energéticos de origem eólica.
Esta dissertação apresenta uma nova metodologia híbrida para a previsão dos preços da
energia eléctrica e da potência eólica em Portugal, considerando o horizonte temporal de
curto prazo, isto é, de um dia a uma semana. Esta nova metodologia baseia-se na combinação
eficaz de sistemas neuro-difusos, programação evolucionária e optimização por enxame de
partículas, sendo aplicada em casos de estudo reais. Os resultados obtidos são posteriormente
comparados com resultados já publicados em revistas internacionais de referência,
permitindo validar a proficiência da nova metodologia proposta.
The implementation of a liberalized electricity market and the growing integration of wind power into the electric grid, particularly in Portugal, pose new challenges associated with the increasing competition in the electricity industry between power producers and the high volatility and intermittence inherent to the wind. Thus, the existence of more effective computational tools becomes essential for market agents, allowing reliable and accurate forecasts of electricity prices and wind power production. These forecasts allow developing enhanced bidding strategies in the market, maximizing the profit, and optimizing the exploitation of the wind energy resources. This dissertation presents a new hybrid methodology to forecast electricity prices and wind power in Portugal, considering a short-term time horizon, i.e., from one day to one week. This new methodology is based on the successful combination of neuro-fuzzy systems, evolutionary programming and particle swarm optimization, being applied to real case studies. Afterwards, the results obtained are compared with previously reported results in reputable international journals, validating the proficiency of the new methodology proposed.
The implementation of a liberalized electricity market and the growing integration of wind power into the electric grid, particularly in Portugal, pose new challenges associated with the increasing competition in the electricity industry between power producers and the high volatility and intermittence inherent to the wind. Thus, the existence of more effective computational tools becomes essential for market agents, allowing reliable and accurate forecasts of electricity prices and wind power production. These forecasts allow developing enhanced bidding strategies in the market, maximizing the profit, and optimizing the exploitation of the wind energy resources. This dissertation presents a new hybrid methodology to forecast electricity prices and wind power in Portugal, considering a short-term time horizon, i.e., from one day to one week. This new methodology is based on the successful combination of neuro-fuzzy systems, evolutionary programming and particle swarm optimization, being applied to real case studies. Afterwards, the results obtained are compared with previously reported results in reputable international journals, validating the proficiency of the new methodology proposed.
Description
Keywords
Energia eléctrica - Cálculo de preços Potência eólica - Cálculo de preços - Sistemas neuro-difusos Mercado eléctrico Recursos energéticos - Energia eólica Energias renováveis
Pedagogical Context
Citation
Publisher
Universidade da Beira Interior