Datapages, Inc.Print this page

Integrated Genetic Analysis of Sand Character for Predicting Reservoir Quality in the Deep Frontier

William A. Heins1 and Suzanne Kairo2
1 ExxonMobil Upstream Research Co, Houston, TX
2 ExxonMobil Production Co, Houston, TX

Reservoir quality forward models are crucial tools for evaluating plays and prospects on the deep frontier, but their usefulness is constrained by the availability of physical analogs (wells, outcrop) from which to draw essential modeling parameters like grain size, sorting, mineralogy, and matrix content. It is possible to make reasonable estimates of these parameters with forward models of sand generating and modifying processes. We have defined and quantified seven key genetic factors that control sand character; these factors constitute “provenance” in the broadest sense, including:

  • tectonic setting (determines kinds of rocks exposed at the Earth's surface);
  • hinterland uplift rate (determines mass flux of material to the Earth's surface, magnitude of regional topographic gradient);
  • climate (determines potential for chemical modification, influences transport potential);
  • hinterland transport distance (influences duration of sand evolution);
  • basin transport distance (influences duration of sand dispersal);
  • basin subsidence rate (influences basin-scale deposition/bypass/erosion patterns); and
  • environment of deposition (segregates available sediments by hydrodynamic character).

We capture the interaction of these factors in a Bayesian Belief Network, which quantitatively convolves the relative effects of each genetic factor. The model estimates the probability of particular sand compositions, size distributions, and matrix content based on discrete states of the genetic factors. Predictions from the sand character model are inputs to diagenetic models. This process allows rapid scenario testing and risk estimation for play evaluation in a comprehensive, consistent, and objective manner. The process ensures that all appropriate questions are answered every time an evaluation is made.