2019 AAPG Annual Convention and Exhibition:

Datapages, Inc.Print this page

Advanced Quantitative Stratigraphic Data Integration of Conventional and Unconventional Plays


In petroleum system studies, for a better assessment of the presence, types and volumes of hydrocarbons in a prospective structure, numerous types of data from wireline-logging, gas-logging, core analysis, and cutting sample analysis are typically collected, aiming to develop a comprehensive understanding of the basin framework, to establish a robust geophysical model, and eventually to reduce investment risk in oil and gas exploration. These data could include, for example, well logs, mud logs, computed tomography (CT) scan, Hyperspectral Imaging, LECO TOC, Programmed Pyrolysis, X-ray fluorescence (XRF), inductively coupled plasma atomic emission spectrometry (ICP-AES), inductively coupled plasma mass spectrometry (ICP-MS), X-ray Diffraction (XRD), Fourier-transform infrared spectroscopy (FTIR), and biostratigraphic ages using foraminifera, nannofossils, pollen and spores, dinoflagellates, and radiolaria. When such multiple large and complex datasets are available, how could they be integrated and summarized in a way that would make sense and help decision makers? Here we introduce a method of advanced quantitative stratigraphic data integration of multiple complex datasets using multivariate statistical analysis and graphical display. Cluster analysis aims to group objects that are similar and to distinguish them from other dissimilar objects. A widely-used application of classical clustering in oil and gas industry is to do rock typing, which helps describe characteristics of different formations. In many applications cluster analysis (e.g., hierarchical analysis) is not constrained if there are no relationships between objects other than the similarities implied by their attributes. Well data have an additional property in that the data are ordered along depth. This constraint can be used to limit the analysis to the stratigraphically adjacent data, i.e., only stratigraphically adjacent zones may be merged. The statistic results would identify different formation packages and subpackages, particularly useful regarding the depth interval of the pay zone or any other intervals of interest. Stratigraphic well correlations of conventional and unconventional plays can be made with much higher confidence. We present a case study of Permian basin. In this study, the data of well logs, cutting sample color, LECO TOC, programmed pyrolysis, XRF elemental composition, and XRD mineral percentages are integrated using statistical cluster results, categorized, and displayed in log format. The formations of Wolfcamp A, Wolfcamp B, Wolfcamp C and Wolfcamp D are characterized.