CLASSIFYING MUDROCKS BY USING A MULTI-COMPONENT CLUSTER ANALYSIS

Mudstone (“shale”) classification is a challenging but essential task for predicting sweet spots in shale resource plays. To date, most existing mudrock classifications are based on grain size and are of limited value to predict the hydrocarbon potential of mudrocks in different stratigraphic intervals and basins. This is because the hydrocarbon storage properties and geomechanical properties that are important to develop these low-permeability plays are often far more dependent on the mineralogic composition and other, mainly geochemical properties than simply the grain size distribution. Additionally, the data available to the petroleum geologist often is restricted to standard bulk mineralogic data and some other brief core description, further making the use of a grain-size based classification impractical.

To overcome some of these deficiencies, we used XRD data from a large data set of (>2000 data points) from 40 different gas-bearing shale formations to look for natural clusters. In the first step we focused our analyses on results from bulk mineralogic data and solved for clusters based on changes in mineralogic composition. This process resulted in seven unique clusters that are defined by major minerals such as quartz, calcite, dolomite, and illite. These primary clusters show significant variations in their respective TOC contents, suggesting predictive capabilities at this level.

In a second step we created sub-clusters within these seven primary mineralogy clusters by including geochemical parameters into the analyses and also putting more emphasis on minor mineral components. This second step resulted in a series of sub-clusters with strong correlation to changing TOC content. Additionally, these sub-clusters correlate well with a series of geochemical environmental indicators and sea-level curves suggesting a dependence on paleo-environmental conditions and a predictable temporal and spatial position within a basin.

In a last step we applied this classification to a series of case studies and were able to demonstrate these aforementioned relationships. Importantly, results from these case studies showed that the primary clusters and sub-clusters are not just good environmental indicators, but are also useful predictive tools in shale-gas studies.