Automatic Identification of Calcareous Lithologies Using Support Vector Machines, Borehole Logs and Fractal Dimension of Borehole Electrical Imaging

In this research algorithms of support vector machine (SVM) and a logic function were applied to identify automatically sections of carbonate rocks in wells located in the former Barco Concession, Catatumbo Basin - Colombia. During training stages the SVMs use neutron, photoelectric factor and gamma...

Full description

Autores:
Leal, Jorge Alberto
Ochoa, Luis Hernan
Contreras, Carmen Cecilia
Tipo de recurso:
Article of journal
Fecha de publicación:
2018
Institución:
Universidad Nacional de Colombia
Repositorio:
Universidad Nacional de Colombia
Idioma:
spa
OAI Identifier:
oai:repositorio.unal.edu.co:unal/68581
Acceso en línea:
https://repositorio.unal.edu.co/handle/unal/68581
http://bdigital.unal.edu.co/69614/
Palabra clave:
55 Ciencias de la tierra / Earth sciences and geology
aprendizaje de máquinas
máquinas de soporte vectorial
registros de pozo
registro de imágenes
dimensión fractal
litologías calcáreas
Cuenca de Catatumbo.
machine learning
support vector machines
borehole logs
image logs
fractal dimension
calcareous lithologies
Catatumbo Basin.
Rights
openAccess
License
Atribución-NoComercial 4.0 Internacional
Description
Summary:In this research algorithms of support vector machine (SVM) and a logic function were applied to identify automatically sections of carbonate rocks in wells located in the former Barco Concession, Catatumbo Basin - Colombia. During training stages the SVMs use neutron, photoelectric factor and gamma ray logs as input; also mean and variance of resistivity acquired for image tool and fractal dimension of resistive images. The first SVM employs in the training stage intervals manually interpreted of fossiliferous limestone, performed by a specialized geologist integrating information of core-logs correlation of a pilot well; afterwards, in classification stages, this SVM automatically recognizes intervals with fossiliferous limestone only using logs data of any well of the field. The second SVM was also trained with nuclear logs, resistivity and fractal dimension, but in this case, with information of intervals composed of calcareous shales interbedded with limestone, recognizing automatically these rock associations during classification stage without interpretations of a geologist as input data. Additionally, a logic function was applied to intervals with photoelectric factor ≥ 4 and all sections not classified by the SVMs were grouped as laminated calcareous rocks. The SVMs and logic function show accuracy of 98.76 %, 94.02 % and 94.60 % respectively in six evaluated wells and might be applied to other wells in the field that have the same dataset. This methodology is highly dependent of the data quality and all intervals affected by bad borehole condition have to be removed prior its application in order to avoid wrong interpretations. Finally, the whole model has to be recalibrated to be applied in other fields of the basin.