Identificação do comportamento do escoamento em golfadas para reconhecimento desse padrão em risers de extração de petróleo utilizando RNA NARX

Detalhes bibliográficos
Ano de defesa: 2020
Autor(a) principal: Conrado, Priscilla Perussolo Cunico
Orientador(a): Não Informado pela instituição
Banca de defesa: Não Informado pela instituição
Tipo de documento: Dissertação
Tipo de acesso: Acesso aberto
Idioma: por
Instituição de defesa: Universidade Federal da Paraíba
Brasil
Engenharia Mecânica
Programa de Pós-Graduação em Engenharia Mecânica
UFPB
Programa de Pós-Graduação: Não Informado pela instituição
Departamento: Não Informado pela instituição
País: Não Informado pela instituição
Palavras-chave em Português:
CFD
RNA
Oil
ANN
Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/18487
Resumo: The oil extraction is occurring in environments increasingly inhospitable, which characterizes an engineering challenge in many fronts. One of those challenges is the stabilization of the flow in marine risers, for risks decrease and also increase in productivity. The slugging effect in particular, due to the multiphasic flow in risers, in some situations is a potential issue. This happens due to the fact that it represents a change in the flow rate and pressure in its usual regimen. When the flow regimen assumes the extreme slugging pattern, it qualifies as a risk of hold ups, preventing the passage of liquid. The oil extraction flow can be defined generically as a multiphasic flow liquid/gas and the slugging pattern is distinguished by its intermittent quality of Taylor’s bubbles (gaseous phase) followed by a liquid phase. This research aims to develop, through the making of a database generated from the results obtained in computational fluid dynamic analysis (CFD), a neural network to identify slug flow. The proposal is that this be a tool which can be applied existing installations without the necessity of extra equipment. Thus, we are looking to use the sensors at the BOP (blowout preventer) in order to identify the slug flow pattern, working with the neural network developed, so that the strangle valve can be deployed automatically before the kicks (pressure variation due to the flow pattern) even reach the end of the riser