Machine Learning Helps Predict Electrical Properties of Heterogeneous Reservoirs
Chris Carpenter · Journal of Petroleum Technology · 2024
_ This article, written by JPT Technology Editor Chris Carpenter, contains highlights of paper IPTC 23381, “Machine-Learning-Assisted Prediction of Electrical Properties for Heterogeneous Reservoirs: Case Study in Mamey Field, Northern Colombia,” by Carelis Moya, Hocol; Roxiris C. Prado, Halliburton; and Hugo Caycedo, Hocol, et al. The paper has not been peer reviewed. Copyright 2024 International Petroleum Technology Conference. _ Petrophysical characterization in reservoirs with high heterogeneity is a consistent challenge. The case study presented in the complete paper describes a machine-learning (ML) technique to determine electrical properties. The methodology combines logs, rock types, and facies and digital core analyses from the Mamey field in northern Colombia, a reservoir composed of interlaminated mudstones and very-fine to fine sandstones enclosed in a deltaic environment and capped by cross-stratification sandstones associated with incised valley deposits. The results obtained indicate that the technique is feasible for estimating a continuous curve of the Archie parameters m and n associated with the textural changes identified in images and computed tomography. Introduction For many years, the industry has reviewed easily extracted reservoirs and neglected those that were slightly more complex. Rock parameters were obtained from conventional core analysis using Archie’s cementation coefficient (m) and saturation exponent (n) used in petrophysical analysis, parameters that were difficult to pin down, especially in complex reservoirs because of rock typing, facies, stratigraphy, and structural variations. The authors’ study was used to supplement the petrophysical model and was focused on reserves estimation according to a water-saturation model, with m and n obtained from digital and conventional rock analysis. The method involved an ML technique using logs (triple-combination, sonic, petrophysical curves, microresist images, facies, and rock type) to obtain m and n. Geological Background The Saman field is in the northwestern part of the Lower Magdalena Valley (LMV) Basin. The petroleum system in this region is linked closely to the Oligocene and Neogene sequences. The key reservoirs include the Cienaga de Oro (CDO), Chengue, and Lower Porquero formations. Within the LMV, the CDO represents an Oligocene/Miocene sequence associated with a transgressive system originating from shallow marine environments. The shoreline is interpreted as having undergone a west-to-east movement, effectively filling ancient basement topography, particularly in the San Jorge and Plato depocenters. Detailed core interpretations in Saman reveal the presence of deltaic facies within the CDO, which notably prograde toward the northwest.