Introducing a causal PAR (p) model to evaluate the influence of climate variables in reservoir inflowsa Brazilian case

Resumo

The Brazilian electricity energy matrix is essentially formed by hydraulic sources which currently account for 70% of the installed capacity. One of the most important characteristics of a generation system with hydro predominance is the strong dependence on the inflow regimes. Nowadays, the Brazilian power sector uses the PAR(p) model to generate scenarios for hydrological inflows. This approach does not consider any exogenous information that may affect hydrological regimes. The main objective of this paper is to infer on the influence of climatic events in water inflows as a way to improve the model's performance. The proposed model is called 'causal PAR(p)' and considers exogenous variables, such as El Niño and Sunspots, to generate scenarios for some Brazilian reservoirs. The result shows that the error measures decrease approximately 3%. This improvement indicates that the inclusion of climate variables to model and simulate the inflows time series is a valid exercise and should be taken into consideration.

Ficha do documento

Tipo

Artigo científico

Ano

2017

Instituição

Sociedade Brasileira de Pesquisa Operacional

Fonte

Repositório da FGV

Idioma

Inglês

Acesso

Acesso aberto

DOI

10.1590/0101-7438.2017.037.01.0107

Identificador

oai:repositorio.fgv.br:10438/21645

Temas

EconomiaInfraestrutura

Palavras-chave

Reservoir inflow modellingPeriodic modelsClimate predictorsEconomiaEnergia elétrica - ProduçãoUsinas hidrelétricasClimatologia