Formation evaluation of a turbidite system in the Albacora field: a statistical approach using R

Abstract

Well logging is basically a set of geophysical measurements taken in open or cased boreholes that provides essential information about the properties of geological formation in subsurface required by exploration and exploitation of important mineral and energetic resources. As a proper geophysical method itself, well logging was developed in the petroleum industry and is the very base for formation evaluation, the most relevant way to characterize the rocks and their exploitation viability, as it embodies the best causal relations between the geophysical measurements and the geological and petrophysical properties of the formation rocks. In the ordinary workflow of the industry, common practice is using commercial suites to perform this task. In this work, we propose an alternative approach, using free platforms (R and RStudio), in which the training evaluation workflow is fully transparent, reproducible, and easily documented by the log analyst. This article presents a consistent and robust analytical flow for an example well of an important Brazilian siliciclastic reservoir, using exclusively free domain tools and language adaptable to other geological scenarios and different teamwork, elegant and fully documentable. As a main result, it is demonstrated that, based on descriptive statistics, this flow is a solid auxiliary tool for formation evaluation, in which parameterization experiments applying common tools such as histograms and 1D and 2D probability density, and diagrams of dispersion (crossplots) compete for the recognition of consistent patterns, which result in a realistic model of electrofacies for the reservoir, with a variety of classes suitable for the scale of inference and geology.