Search

Conventional formation evaluation provides fast and accurate estimations of petrophysical properties in conventional formations through conventional well logs and routine core analysis (RCA) data. However, as the complexity of the evaluated formations increases conventional formation evaluation fails to provide accurate estimates of petrophysical properties. This inaccuracy is mainly caused by rapid variation in rock fabric (i.e., spatial distribution of rock components) not properly captured by conventional well logging tools and interpretation methods. Acquisition of high-resolution whole-core computed tomography (CT) scanning images can help to identify rock-fabric-related parameters that can enhance formation evaluation. In a recent publication, we introduced a permeability-based cost function for rock classification, optimization of the number of rock classes, and estimation of permeability. Incorporation of additional petrophysical properties into the proposed cost function can improved the reliability of the detected rock classes and ultimately improve the estimation of class-based petrophysical properties. The objectives of this paper are (a) to introduce a robust optimization method for rock classification and estimation of petrophysical properties, (b), to automatically employ whole-core two-dimensional (2D) CT-scan images and slabbed whole-core photos for enhanced estimates of petrophysical properties, (c) to integrate whole-core CT-scan images and slabbed whole-core photos with well logs and RCA data for automatic rock classification, (d) to derive class-based rock physics models for improved estimates of petrophysical properties. First, we conducted formation evaluation using well logs and RCA data for estimation of petrophysical properties. Then, we derived quantitative features from 2D CT-scan images and slabbed whole-core photos. We employed image-based features, RCA data and CT-scan-based bulk density for optimization of the number rock classes. Optimization of rock classes was accomplished using a physics-based cost function (i.e., a function of petrophysical properties of the rock) that compares class-based estimates of petrophysical properties (e.g., permeability and porosity) with core-measured properties for increasing number of image-based rock classes. The cost function is computed until convergence is achieved. Finally, we used class-based rock physics models for improved estimates of porosity and permeability. We demonstrated the reliability of the proposed method using whole-core CT-scan images and core photos from two siliciclastic depth intervals with measurable variation in rock fabric. We used well logs, RCA data, and CT-scan-based bulk-density. The advantages of using whole-core CT-scan data are two-fold. First, it provides high-resolution quantitative features that capture rapid spatial variation in rock fabric allowing accurate rock classification. Second, the use of CT-scan-based bulk density improved the accuracy of class-based porosity-bulk density models. The optimum number of rock classes was consistent for all the evaluated cost functions. Class-based rock physics models improved the estimates of porosity and permeability values. A unique contribution of the introduced workflow when compared to previously documented image-based rock classification workflows is that it simultaneously improves estimates of both porosity and permeability, and it can capture rock class that might not be identifiable using conventional rock classification techniques.

CT scan images provide valuable three-dimensional information on the mineralogical composition and overall internal structure of cores. X-ray computerized tomography (CT) imaging of whole cores has therefore become a routine step in core analysis workflows. This new data type gives new possibilities in reservoir characterization. Lithological classification of reservoir rocks, in its turn, is an essential step to better understand the depositional environment and for subsequent effective reservoir characterization: the chemical composition of the minerals, combined with their grain size, sorting and pore size distribution is known to highly affect the transport properties of reservoir rocks. Lithological classification on the extracted whole core material is thus consequential; however, it also requires significant investments, being traditionally conducted through visual inspection performed by expert geologists. This manual process is time consuming, and prone to subjective interpretations and human errors. Therefore, a current research and development trend is to find automated methods for computer-assisting the assessment of this type of data, eventually reducing time and costs of core analysis, and improving the overall business decisions. In this study we explore the application of Convolutional Neural Networks (CNN) to automatically classify lithofacies. We propose a workflow for high resolution lithofacies classification using whole core three-dimensional CT images, and we assess the validity of our approach on a field-example from the Norwegian continental shelf. The novelty of our approach is thus learning, through a CNN, the relationship between convolution-derived three-dimensional features and expert-derived lithofacies classes. We thus extend approaches working on two-dimensional images into a workflow that uses high-resolution three-dimensional CT images as direct input. In our work the training data set includes information obtained from manual core description. Prior to training, the three-dimensional CT images are pre-processed so that undesired artefacts are automatically flagged and removed before being fed into the network. The approach is validated using the trained CNN classifier to predict lithofacies in a set of unseen three-dimensional CT data. The trained model can predict lithofacies classes with high accuracy, with a misclassification rate of about 3%. We found that these misclassifications are mainly associated with the presence of high-density material such as pyrite nodules and drilling mud invasions. Dipping fractures and missing values, not completely removed by image pre-processing, are additional reasons for model deficiency in some of the incorrectly classified images. Overall, the trained classifier exhibits higher pixel-wise precision and captures the high-resolution heterogeneities more accurately compared to the manual core descriptions.