Advanced search
Start date
Betweenand

Geostatistical methods for definition of geological model of a pre-salt reservoir

Grant number: 20/01306-0
Support type:Scholarships in Brazil - Doctorate
Effective date (Start): March 01, 2020
Effective date (End): February 28, 2023
Field of knowledge:Physical Sciences and Mathematics - Geosciences - Geology
Cooperation agreement: Equinor (former Statoil)
Principal Investigator:Alexandre Campane Vidal
Grantee:Jean Carlos Rangel Gavidia
Home Institution: Centro de Estudos do Petróleo (CEPETRO). Universidade Estadual de Campinas (UNICAMP). Campinas , SP, Brazil
Associated research grant:17/15736-3 - Engineering Research Centre in Reservoir and Production Management, AP.PCPE

Abstract

This project proposes the use of geostatistical methods, with an emphasis on Multiple-Point Geostatistics (MPG), for the generation of the geological model, honoring the main heterogeneities. One of the main challenges of pre-salt exploration is related to geological complexity. In this case, it is possible to define as main heterogeneities: 1) distinct geological environment without analogous in recent; 2) influence of karst structures; 3) structural features related to the basin evolution; 4) heterogeneity of the porous medium common to carbonatic rocks. The multiple-points are calculated through image filters that perform weighted transformations by the inverse of the distance between the pixels and the center of the neighborhood, previously determined, scoring the patterns of the image. As a unique filter is not able to predict all the patterns of the image, a set of filters are used: mean, gradient, and curvature. Each filter creates a map of scores of the training image which similar values for analogous patterns, so the patterns can be regrouped. In this project, the k-means method is proposed, for small time-consuming reason. Once the training image is transformed into a map of centroids of clusters, represented by the k-means groups, it is possible to calculate different scenarios of conditional probability for assigning a value to be predict to a particular value of the centroid training image. In this study we propose predict permeability, with high-resolution scale, into large rock models with MPG method. The objective of this project is to generate a geological model that honors the main geological heterogeneities of the porous system in a pre-salt reservoir. (AU)