Image logs provide critical data for characterizing fractured reservoirs. However, the nature of such logs, coupled with borehole-modified stresses, creates challenges that may lead to erroneous fracture interpretations. We developed a methodology to integrate image logs, core, and geomechanical models to obtain high-confidence, fracture data sets derived from image logs. The methodology starts with a detailed comparison of overlapping image logs and core, first to properly shift them into alignment (core-to-log shift) and second to determine which features interpreted as fractures on image logs (picks) can be validated as fractures in core. Then we derive guidelines for interpreting image logs without core control to reduce the likelihood of false fracture picks. Geomechanical modeling is used to assist with quality control and interpretation of image logs where drilling-induced tensile fractures occur. We use a fractured carbonate reservoir to demonstrate the application of this methodology. Comparison of image log fracture picks with core led to an average rejection rate of greater than 50% of fractures picked on the image logs. These high rejection rates are mostly caused by overinterpretation of discontinuous sinusoids and misidentification of sedimentary features and induced fractures. We derived rules for interpreting high-quality fractures: (1) only sinusoids with continuity greater than 50% should be interpreted and (2) sinusoids with dip angles less than 55° should be excluded from the interpretation. Application of this methodology has provided much higher quality fracture orientation and frequency data than that resulting from interpretation of borehole images alone. The improved fracture data set provides a more reliable basis for testing hypotheses and deriving models of fracture distribution.

You do not have access to this content, please speak to your institutional administrator if you feel you should have access.