Identificación de la presencia de hidrocarburos en arenas arcillosas usando un algoritmo de machine learning, registros de pozo e imágenes de fluorescencia de corazones

Evaluating the presence of hydrocarbon in sands interbedded with clay laminations below the vertical resolution of resistivity logs is a complex task. Clays, being highly conductive, generate suppression of the resistivity logs, making water saturation calculated using the Archie equation high in otherwise potentially productive zones. Other models such as Poupon, Waxman-Smits or Simandoux require knowledge of the electrical properties of clays from laboratory measurements or calibration in thick end-member shales representative of the laminations. Recent advances in machine learning algorithms and the availability of image processing modules in Python suggest training borehole logs to recognize variations in ultraviolet (UV) fluorescence core images as a promising way to recognize the presence of hydrocarbons in complex reservoirs. In this paper, we propose a new method based on Machine Learning, analogous to the use of facial recognition algorithms, to predict the complex relationship between the presence of hydrocarbon in laminated sands and basic well log measurements without using any explicit analytical expressions.

• Fecha de publicación: 2021
• Institución: Universidad de América
• Repositorio: Lumieres
• Idioma: spa
• OAI Identifier: oai:repository.uamerica.edu.co:20.500.11839/8627
• Acceso en línea: https://hdl.handle.net/20.500.11839/8627
• Palabra clave: Imágenes de corazones; Presencia de hidrocarburos; Saturación de agua; Images of hearts; Presence of hydrocarbons; Water saturation; Tesis y disertaciones académicas
• License: Atribución – No comercial