1. Identifying clay mineral using angular competitive neural network: A machine learning application for porosity estimative.
- Author
-
Pinheiro de Almeida, Thales Luiz, Fonseca Passos, Bruno Andrey, Santos da Costa, Jéssica Lia, and Neves Andrade, André José
- Subjects
- *
CLAY minerals , *ROCK properties , *MACHINE learning , *MINERAL properties , *POROSITY - Abstract
In the daily practice of Formation Evaluation for porosity calculation considering the shale volume is common the adoption of simple hypotheses about the physical properties of the reservoir rock constituents. In this paper, it is argued that an accurate estimate of clay mineral physical properties in the reservoir rock are essential for a realistic porosity calculation, once it may be different from the physical properties of clay minerals present in nearby shale layers. Geologically, the assumption of similarity between the clay mineral physical properties in the reservoir rock and adjacent shales properties means to presume a depositional continuity in sedimentary process and disregarded the innumerous postpositional processes involved in the sedimentary basin. By applying the angular competitive neural network to the Density-Neutron method to identify the clay mineral in the reservoir rock based on all identifiable clay minerals in the oil field, it is shown that a particular clay mineral signals a characteristic angular pattern in the Density-Neutron cross-plot. The methodology is presented with synthetic data and evaluated with actual well logs and core analysis from a borehole drilled in the Namorado oil field, Campos basin, Brazil. • A geologically coherent method to study petrophysical properties of reservoir rock. • The angular competitive neural network is applied to determine the predominant clay mineral in the reservoir rock. • The identified clay mineral in the reservoir is proposed to a realistic porosity estimate. [ABSTRACT FROM AUTHOR]
- Published
- 2021
- Full Text
- View/download PDF