A New Exploratory Data Analysis Technique to Aid in Biosteering

Anthony Gary, Glenn Johnson, and Douglas Ekart
University of Utah, Salt lake City, UT

Deviated wellbores are now a common practice within the energy industry as a means of maximizing production and minimizing infrastructure costs. Today's drilling technology allows very precise positioning of the well path in 3D space, but precise geologic information is first required to determine the appropriate attitude of the drill bit. From applications to hundreds of wells worldwide the technique of biosteering has a proven track record in providing this critical information. Biosteering relies on recognizing, during drilling, those particular micropalentological components of an assemblage (i.e., biofacies) that differentiate the target interval from those bounding it. These differences in the assemblages are based on the observations of an experienced biostratigrapher. While a knowledgable biostratigrapher is an absolute necessity for the successful implementation of biosteering, there is an opportunity for more rigorous and exhaustive data analysis methods that can aid the biostratigrapher in knowledge discovery, yielding geologically useful information that may be too subtle or complex to be resolved by observation alone. To fulfill this need we have developed a multivariate procedure, based on fuzzy c-means clustering, that will: 1) assist the biostratigrapher in the recognition of assemblages that may represent biofacies, 2) determine the relative frequencies of the components in each of the assemblages, and 3) calculate the proportions of each of those assemblages in every sample. From these results the biostratigrapher can readily evaluate the significance of each assemblage and recognize mixed or transitional biofacies.

AAPG Search and Discovery Article #90039©2005 AAPG Calgary, Alberta, June 16-19, 2005