(706d) Critical Assessment of Brittleness Indices As Proxies for Rock Mechanical Impacts on Hydrocarbon Recovery: Case Study from the Caney Shale, Oklahoma

In the drive to maximize gas recovery from horizontal wells, one common approach is to assign so-called “Brittleness Index” values to the rocks. The underlying premise is more brittle rocks will be stimulated more effectively by hydraulic fracturing and will be less prone to the creep related closure of fractures which leads to production decline. There are several manifestations to calculate brittleness. One approach is to characterize a brittleness index along a horizontal well, and target simulation in the most brittle sections of the well. Moving back one step in the decision-making process, a second approach makes use of the brittleness index of layers within a stratified reservoir so the landing depth of the horizontal well targets the most brittle layer(s). A third approach views the entire formation, i.e. an entire shale play, as being more or less difficult to develop based on whether it is a “brittle” or “ductile” shale.

While these approaches have some merit and can point to local successes, criticisms stem from the multiplicity of definitions, none of which comprehensively summarize rock behavior relevant to hydrocarbon production and some of which can contradict one another. At a deeper level, the language of “brittle” versus “ductile” describes shale according to two misleading end member labels, when in fact shale is a class of rocks which always fractures in a quasi-brittle manner and never truly brittle (like glass) or ductile (like clay).

This talk will focus on a three-fold critical comparison of brittleness index (BI) as applied to hydrocarbon-bearing shales. The first compares various measures of BI, ranging from evaluation of non-linearity in stress strain curves under triaxial compression testing, to comparisons of rock strengths and/or stiffnesses, to purely mineralogy-based descriptions. Using recent experimental results from a current project in the emerging Caney Shale, Oklahoma, we show the mechanical and mineralogical descriptions of BI are weakly correlated, at best, and in some cases are anti-correlated, pointing to contradiction.

A second comparison evaluates possible connection between brittleness indices and creep properties. Initial laboratory results demonstrate predictions based on brittleness index do not consistently match behavior from triaxial mechanical properties testing. However, creep testing shows nominally ductile zones are considerably more prone to viscoelastic and viscoplastic deformation compared to nominally brittle zones.

Finally, a third comparison examines brittleness indices among different plays, including nominally “brittle” plays (e.g. Barnett) and nominally “ductile” plays (e.g. Haynesville). While difficulty stems from the multiplicity of ways to define brittleness, it is possible to apply multiple brittleness definitions to each formation. Such a comparison gives rise to consensus definition of brittleness and, in this regard, Caney appears to be relatively brittle by all definitions except for the most industry’s most commonly-applied definition based on elastic properties.

Hence, the distinction of “ductile” does not appear to directly apply to the Caney. However, in its vernacular use in regards to how prone a rock is to creep, the Caney does show propensity to time-dependent deformation which could pose a challenge to be addressed during the development of this emerging play.