Loading...
Publication
Gestão de recursos energéticos em Smart-grid, considerando incertezas
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 9 MB | Adobe PDF |
Authors
Abstract(s)
Nos últimos anos, as redes elétricas tradicionais passaram por uma transformação significativa, evoluindo para sistemas massivamente distribuídos com a integração de diversos tipos de recursos energéticos. A crescente penetração desses recursos distribuídos, especialmente os renováveis, introduz elevados níveis de variabilidade e incertezas nos sistemas elétricos, aumentando os desafios operacionais das redes. Diante desse cenário, as novas tecnologias associadas às Smart grids têm demonstrado a capacidade de criar um ambiente orientado ao mercado, apto a lidar com a alta penetração desses recursos e com a nova dinâmica do consumo de energia. No entanto, para uma operação eficiente e econômica do sistema, são necessárias ferramentas adequadas de gerenciamento de recursos energéticos. Neste contexto, este trabalho propõe duas metodologias para resolver o problema de gerenciamento de recursos energéticos em diferentes horizontes temporais. A primeira metodologia desenvolve um modelo estocástico baseado em uma métrica de risco para o horizonte do dia seguinte. Para isso, um conjunto de cenários é gerado por meio da Simulação de Monte Carlo. Para avaliar o risco, é utilizado a métrica Conditional Value-at-Risk, visando tornar a resposta do modelo mais segura e robusta diante de eventos extremos. A segunda metodologia propõe um modelo determinístico para o horizonte Intra Diário (hora seguinte), utilizando como ponto de partida a resposta do modelo do dia seguinte baseado no risco. Este modelo prevê um mercado local de energia como estratégia para satisfazer a equação de balanço energético, permitindo ao agregador recorrê-lo para satisfazer suas demandas interna. As metodologias serão testadas
em uma rede distribuição real de 180 barramento, com alta penetração de recursos distribuídos. Serão previstos casos para diferentes níveis de exposição ao risco do agregador e erros de previsão dos recursos. Os resultados obtidos demostram que o modelo para o dia seguinte, baseando no risco, é capaz de realizar decisões de agendamento não apenas com base no custo esperado, mas também considerando a mitigações dos riscos de eventos extremos. É
observado reduções de custo de 12,60% na ocorrência de cenários críticos. Em relação ao modelo Intra Diário, os resultados obtidos permitem observar os impactos dos erros de previsão que próximo ao horizonte de operação, bem como os benefícios que são ofertados pelos sistemas de armazenamento de energia e o mercado local, sendo observados reduções de 3% a 19% no custo do agregador.
In recent years, traditional electrical grids have undergone significant transformation, evolving into massively distributed systems with the integration of various types of energy resources. The increasing penetration of these distributed resources, especially renewables, introduces high levels of variability and uncertainties into power systems, thereby escalating the operational challenges of the grids. Against this backdrop, new technologies associated with smart grids have demonstrated the capability to create a market-oriented environment, capable of handling the high penetration of these resources and the new dynamics of energy consumption. However, for efficient and economic system operation, appropriate energy resource management tools are required. In this context, this study proposes two methodologies to solve the energy resource management problem across different time horizons. The first methodology develops a stochastic model based on a risk metric for the day-ahead horizon. For this purpose, a set of scenarios is generated through Monte Carlo Simulation. To evaluate the risk, the Conditional Value-at-Risk metric is used, aiming to make the model's response safer and more robust against extreme events. The second methodology proposes a deterministic model for the intra-day (next-hour) horizon, using as a starting point the response from the day-ahead risk-based model. This model envisions a local energy market as a strategy to satisfy the energy balance equation, allowing the aggregator to utilize it to meet internal demands. The methodologies will be tested on a real 180-bus distribution network with a high penetration of distributed resources. Cases will be projected for different levels of aggregator risk exposure and resource forecast errors. The obtained results demonstrate that the day-ahead risk-based model is capable of making scheduling decisions not only based on expected cost but also considering the mitigation of risks from extreme events, showing cost reductions of 12.60% in critical scenarios. Regarding the intra-day model, the results allow for observing the impacts of forecast errors close to the operational horizon, as well as the benefits offered by energy storage systems and the local market, with observed cost reductions for the aggregator ranging from 3% to 19%.
In recent years, traditional electrical grids have undergone significant transformation, evolving into massively distributed systems with the integration of various types of energy resources. The increasing penetration of these distributed resources, especially renewables, introduces high levels of variability and uncertainties into power systems, thereby escalating the operational challenges of the grids. Against this backdrop, new technologies associated with smart grids have demonstrated the capability to create a market-oriented environment, capable of handling the high penetration of these resources and the new dynamics of energy consumption. However, for efficient and economic system operation, appropriate energy resource management tools are required. In this context, this study proposes two methodologies to solve the energy resource management problem across different time horizons. The first methodology develops a stochastic model based on a risk metric for the day-ahead horizon. For this purpose, a set of scenarios is generated through Monte Carlo Simulation. To evaluate the risk, the Conditional Value-at-Risk metric is used, aiming to make the model's response safer and more robust against extreme events. The second methodology proposes a deterministic model for the intra-day (next-hour) horizon, using as a starting point the response from the day-ahead risk-based model. This model envisions a local energy market as a strategy to satisfy the energy balance equation, allowing the aggregator to utilize it to meet internal demands. The methodologies will be tested on a real 180-bus distribution network with a high penetration of distributed resources. Cases will be projected for different levels of aggregator risk exposure and resource forecast errors. The obtained results demonstrate that the day-ahead risk-based model is capable of making scheduling decisions not only based on expected cost but also considering the mitigation of risks from extreme events, showing cost reductions of 12.60% in critical scenarios. Regarding the intra-day model, the results allow for observing the impacts of forecast errors close to the operational horizon, as well as the benefits offered by energy storage systems and the local market, with observed cost reductions for the aggregator ranging from 3% to 19%.
Description
Keywords
Aversão ao risco Agregadores Conditional Value-at-Risk Gestão de recursos energéticos dia seguinte Gestão de recursos energéticos intra diário Gestão de risco Risk aversion Aggregators Energy resources management day-ahead Energy resources management intraday Risk management
Pedagogical Context
Citation
Publisher
CC License
Without CC licence