A Novel Approach to Well Barrier Survival Analysis using Machine Learning

Registro completo de metadados
MetadadosDescriçãoIdioma
Autor(es): dc.contributorPETROBRAS Research Center-
Autor(es): dc.contributorFederal Fluminense University-
Autor(es): dc.contributorUniversidade Estadual Paulista (UNESP)-
Autor(es): dc.contributorUniversity Wolverhampton-
Autor(es): dc.creatorColombo, Danilo-
Autor(es): dc.creatorLima, Gilson Brito Alves-
Autor(es): dc.creatorPapa, João Paulo-
Autor(es): dc.creatorPassos, Leandro Aparecido-
Autor(es): dc.creatorSantana, Marcos Cleison Silva-
Data de aceite: dc.date.accessioned2025-08-21T17:57:34Z-
Data de disponibilização: dc.date.available2025-08-21T17:57:34Z-
Data de envio: dc.date.issued2025-04-29-
Data de envio: dc.date.issued2021-12-31-
Fonte completa do material: dc.identifierhttp://dx.doi.org/10.3850/978-981-18-5183-4_S01-04-420-cd-
Fonte completa do material: dc.identifierhttps://hdl.handle.net/11449/308923-
Fonte: dc.identifier.urihttp://educapes.capes.gov.br/handle/11449/308923-
Descrição: dc.descriptionWell safety barriers are critical elements for the integrity and safety of operations in oil and gas wells. This paper introduces a novel approach to characterize the reliability performance of safety barriers in offshore oil wells under different operational and environmental conditions. Traditional statistical methods have been applied to this class of problem; however, these methods assume specific distributions and parameters to model the time to failure. Furthermore, they usually require specialists to evaluate if the data fits into the model's hypothesis. A combination of statistical and machine learning approaches can overcome these limitations and capture complex behavior and non-linearities. A greedy algorithm composed of two machine learning regression models and a nonparametric reliability estimator can handle the problem nicely. A modification in the loss function is proposed to consider the censored data. The proposed approach was tested and validated in simulated data and for reliability estimation of downhole safety valves, one of the most important and must-have well barrier components. The results show that the proposed model is competitive and outperforms classical approaches, besides being more generic and customizable. We can either use the model to forecast the reliability of new valves, identify the most appropriate valve according to the well features, improve the valve performance, and others.-
Descrição: dc.descriptionFederal Fluminense University PETROBRAS Research Center-
Descrição: dc.descriptionSchool of Engineering Production Department Federal Fluminense University-
Descrição: dc.descriptionDepartment of Computing São Paulo State University (Unesp) Campus Bauru-
Descrição: dc.descriptionCMI Lab School of Engineering and Informatics University Wolverhampton-
Descrição: dc.descriptionDepartment of Computing São Paulo State University (Unesp) Campus Bauru-
Formato: dc.format1767-1774-
Idioma: dc.languageen-
Relação: dc.relationProceedings of the 32nd European Safety and Reliability Conference, ESREL 2022 - Understanding and Managing Risk and Reliability for a Sustainable Future-
???dc.source???: dc.sourceScopus-
Palavras-chave: dc.subjectAccelerated Life Test-
Palavras-chave: dc.subjectMachine Learning-
Palavras-chave: dc.subjectReliability Prediction-
Palavras-chave: dc.subjectSurvival Analysis-
Palavras-chave: dc.subjectWell Barriers-
Título: dc.titleA Novel Approach to Well Barrier Survival Analysis using Machine Learning-
Tipo de arquivo: dc.typeaula digital-
Aparece nas coleções:Repositório Institucional - Unesp

Não existem arquivos associados a este item.