USING BOREHOLE-WALL FRACTURE TRACES TO ESTIMATE MEAN FRACTURE SIZE AND ASPECT RATIO
One approach to characterizing subsurface fracture populations is to assume the characteristics of fracture populations exposed at the surface are representative of fractures at depth. Problems with this assumption include the absence of subsurface rock units at the surface, and differences in the fracture-producing deformation history between surface and sub-surface settings for the same rock units. The present investigation attempts to eliminate the need for this assumption through the direct characterization of mean fracture size and aspect ratio of a given fracture set based on traces on the cylindrical walls of boreholes, shafts, or tunnels.
Sets of fractures with non-equant shape and preferred orientation produce characteristic populations of traces on the unrolled wall of a cylinder. As this study focused on joints in sedimentary rocks where the majority of fractures terminate at bedding surfaces, and are much longer parallel to bedding than across bedding, fractures were assumed to be elongated rectangles. The new characterization approach uses counts of the intersection types between these rectangular fractures and the cylinder or borehole. The six types of fracture-cylinder intersection are: complete, long-edge, short-edge, corner, end, and pierced. These types are differentiated on the basis of the completeness of an intersection around a borehole and the position of the trace on the borehole wall as a function of geographic orientation and joint strike.
Typical expected trace populations were derived using a simple probabilistic model, and confirmed by Monte Carlo simulation. Simple estimators for fracture aspect ratio and size were developed, based on these characteristic trace populations.
Joint populations within the Mesaverde Group of the Piceance Basin were analyzed using the cylinder-based estimators from borehole FMI (Formation MicroImager) data logs. The joints accommodated regional extension during the early Laramide orogeny. Considering sandstone beds in the Williams Fork Formation, the subsurface cylinder-based estimators yielded bed-parallel mean tracelengths of 10.0 m to 52.6 m, bed-normal mean fracture heights of 1.8 m to 10.7 m, and mean fracture surface areas of 22.8 m2 to 171.4 m2.