Desenvolvimento de uma metodologia integrada para estimação e previsão de geração eólica em larga escala para o planejamento e a operação de sistemas elétricos

Loading...
Thumbnail Image

Journal Title

Journal ISSN

Volume Title

Publisher

Universidade Federal de Juiz de Fora (UFJF)

DOI

Abstract

The search for a more sustainable future is increasingly necessary, and the participation of renewable sources in the energy matrix of countries keeps growing. The importance of this energy transformation is undeniable; however, the sharp growth of sources prone to high variability and intermittency makes the operation of electrical systems more complex. This variability generates operational costs and complexity, mainly for forecast models, both supply and demand. Furthermore, unlike conventional sources, controlled by operators, these energy sources are somewhat controlled by nature, generating uncertainty in energy security and impacts on electrical systems. In Brazil, in addition to the high variability and the dependence on meteorological factors, these sources are non-dispatchable, that is, under normal operating conditions their generation is directly associated with the natural resource that arrives at the plant. Superimposed on this scenario is the significant growth of generation curtailment observed in recent years, especially in the Northeast region, which causes the verified generation to no longer reflect the available wind resource and further disturbs the historical time series used by the models. Therefore, it is essential to produce accurate forecasts of their generation to compose the optimized energy production portfolio of the SIN (National Interconnected System). Such forecasts guide operators’ decisions in real time and are important inputs for very short-term planning (Daily Scheduling of Electroenergetic Operations), as well as useful for short- and medium-term planning. Greater accuracy in forecasting these sources enables better use of resources, reducing the need for operating power reserves or thermoelectric generation beyond that defined by economic merit by the optimization models officially in use. Given this scenario, the main objective of this work is the development of wind power generation forecasting models for the Brazilian case, covering horizons from real time up to one month ahead, as required for the operation and planning carried out by the ONS (National Electric System Operator) and incorporating information from the main numerical weather prediction models worldwide. As supporting stages for this objective, a treatment and reconstruction of the historical wind and generation data is proposed, integrating measurements from different institutions of the sector, together with a potential generation estimation model, able to reconstruct the generation that would occur in the absence of curtailment and to quantify the associated losses. The methodologies were applied at large scale, considering more than one thousand utility-scale wind farms connected to the Brazilian transmission grid, one of the largest centralized wind power systems in the world, contributing to increase the reliability of the operation and the electroenergetic planning of the SIN.

Description

Citation

Endorsement

Review

Supplemented By

Referenced By

Creative Commons license

Except where otherwise noted, this item's license is described as Acesso Aberto