I am primarily interested in the modeling and analysis of fractures using various techniques that range from deterministic Discrete Fracture Network models (DFN) to Fractals and Multi-Point Geostatistics (MPS). My research is funded through Stanford Center for Reservoir Forecasting (SCRF), an industrial affiliates program in reservoir modeling and geostatistics. Currently, I am working on three different but related topics that encompass research in reservoir characterization and modeling using multi-scale, multi-dimensional fracture data.

The first one is a BP-sponsored project that deals with evaluating uncertainty in fractured reservoirs. It is a major SCRF endeavor and I am working with supervision from Profs. Tapan Mukerji and Jef Caers while coordinating with other PhD-students and visiting scholars. As a first step, a benchmark synthetic fractured reservoir data set is built with details on geology, rock physics, DFN and seismic attributes. The next phase will use this data to evaluate if DFN models from an explored area can be utilized to predict performance in relatively under-explored adjacent zones.

The second project explores the use of image quilting, a pattern-based MPS technique, for modeling fracture network maps starting with outcrop analogs as training images (TI). A set of nested-fracture maps from the Devonian Sandstone of Hornelen basin created at multiple scales is considered as TIs. This research focuses on choosing the “proper” template-size required for reproducing a desired pattern from a given TI. This project tests the tools that will be implemented in the second phase of project-I.

Finally, I am also working in close association with Prof. Atilla Aydin in Geological and Environmental Sciences on a third project that studies the effects of data resolution on fracture clustering. Lacunarity is used as a tool for quantifying scale-dependent clustering. This technique has been previously employed for identifying scale-dependent heterogeneity in different types of data sets including fracture scanlines, maps and grayscale data, e.g. fracture intensity maps, soil thin-sections etc. Fracture spacing data collected at different resolutions from the same line are compared for their scale-dependent clustering attributes.