| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 14.16 MB | Adobe PDF |
Authors
Abstract(s)
A rede elétrica atual possui uma enorme dificuldade em gerir a variação nas tensões existentes, resultante de vários produtores de energia renováveis independentes. Neste âmbito foram criados ao longos dos anos vários algoritmos e técnicas de previsões de forma a ser possível obter uma previsão atempada e correta das cargas e da geração e consequentemente poder otimizar a rede elétrica para obter o seu melhor rendimento possível. Este trabalho pretende contribuir para o estudo e análise do comportamento duma rede elétrica através da implementação do algoritmo escolhido para previsão. A previsão é dividida em várias etapas, utilizando repositórios de informação de meteorologia, informação sobre a rede elétrica e cálculo de erros, para que ocorra com os mínimos problemas. A arquitetura geral da solução é de complexidade acrescida com integração de vários sistemas e tecnologias atuais que comunicam entre si, de forma continua a fim de, receber e transmitir os valores mais recentes. A execução da previsão ocorreu numa zona restrita da rede elétrica existente, fazendo ajustes necessários por forma a não existir complicações com esta redução, devido a existir diferenças entre a rede atual e a rede em estudo. Foi elaborada uma página web de forma, a verificar os resultados da previsão e de otimização da rede elétrica que utiliza os valores da previsão. Este estudo identificou diferenças entre as previsões com a realidade, o que levou a uma análise posterior dos resultados da previsão onde, verificou-se que os erros da previsão da potência ativa foram menores que os erros da previsão da potência reativa e a previsão da produção teve erro maior que os erros da previsão das cargas.
The current electrical grid has enormous difficulty in managing the variation in existing voltages, resulting from several independent renewable energy producers. In this context, several algorithms and forecasting techniques have been created over the years in order to obtain a timely and correct forecast of loads and generation and consequently be able to optimize the electrical network to obtain its best possible performance. This work aims to contribute to the study and analysis of the behavior of an electrical network through the implementation of the algorithm chosen for prediction. The forecast is divided into several stages, using repositories of meteorological information, information about the electrical network and error calculation, so that it occurs with minimal problems. The general architecture of the solution is of increased complexity with the integration of several current systems and technologies that communicate with each other continuously in order to receive and transmit the most recent values. The execution of the forecast took place in a restricted area of the existing electrical network, making necessary adjustments so as to avoid complications with this reduction, due to differences between the current network and the network under study. A web page was created in order to verify the results of the forecast and optimization of the electrical network using the forecast values. This study identified differences between the forecasts and reality, which led to a subsequent analysis of the forecast results where it was found that the errors in the active power forecast were smaller than the errors in the reactive power forecast and the production forecast had a greater error than the load forecast errors.
The current electrical grid has enormous difficulty in managing the variation in existing voltages, resulting from several independent renewable energy producers. In this context, several algorithms and forecasting techniques have been created over the years in order to obtain a timely and correct forecast of loads and generation and consequently be able to optimize the electrical network to obtain its best possible performance. This work aims to contribute to the study and analysis of the behavior of an electrical network through the implementation of the algorithm chosen for prediction. The forecast is divided into several stages, using repositories of meteorological information, information about the electrical network and error calculation, so that it occurs with minimal problems. The general architecture of the solution is of increased complexity with the integration of several current systems and technologies that communicate with each other continuously in order to receive and transmit the most recent values. The execution of the forecast took place in a restricted area of the existing electrical network, making necessary adjustments so as to avoid complications with this reduction, due to differences between the current network and the network under study. A web page was created in order to verify the results of the forecast and optimization of the electrical network using the forecast values. This study identified differences between the forecasts and reality, which led to a subsequent analysis of the forecast results where it was found that the errors in the active power forecast were smaller than the errors in the reactive power forecast and the production forecast had a greater error than the load forecast errors.
Description
Keywords
Rede elétrica Algoritmos Previsão de Cargas Previsão de Geração Otimização Electrical network Algorithms Load Forecast Generation Forecast Optimization