AmericanChemicalSociety.com

Objective: The main objective of the research conducted through this grant was to define the properties of natural fracture networks and their in controlling oil production from the Tensleep Formation reservoir in the Teapot Dome field, Wyoming.

Background: Fracture orientation, intensity, and aperture measurements obtained from the FMI logs available for the field were used to develop layered and composite fracture network model for the Tensleep reservoir. Analysis of post stack 3D seismic data was used to define the properties of the fracture network at field scale.  Flow simulations were initiated with the ultimate goal of making iterative adjustments to field scale fracture network and matrix properties to obtain a reservoir model that accommodated production history match. In year 1, our efforts concentrated on defining the characteristics of reservoir fracture networks inferred from the FMI log analysis. During year 2 our efforts focused on the development of a reservoir scale starting model for flow simulation using ECLIPSE. Seismic data were used to upscale the local fracture networks to field scale. A short extension period was needed to complete flow simulation tests. In this report we provide a summary of results along with an update on the outgrowths of initial flow simulation tests. 

Reservoir Characterization: Production from producing Tensleep wells was examined and a series of maps showing cumulative production were compiled. High producing wells in the field are distributed along a fault bounded structural culmination in the southern part of the dome. The highest producing wells lie along the fault trend in the structurally higher areas of the reservoir. The reservoir model developed for flow simulation incorporated 11 fracture sets distributed in three primary reservoir intervals. Fracture density and variations of density throughout the reservoir model were controlled using Ant Track intensity derived from post-stack processing of the 3D seismic data. Ant tracking is a Schlumberger patented 3D seismic process that locates and maps subtle disruption of seismic amplitude and may be associated with areas of increased fracture density. The fracture aperture distribution used in the model was developed using FMI log derived hydraulic apertures. The fracture length distribution was based in part on an assumed relationship to bed thickness. However, in the field scale model, the length distribution was extrapolated to reservoir scale using a power law distribution. Fracture length, aperture, permeability, and fracture-matrix coupling were upscaled into the 3D gridded reservoir model using the Schlumberger Seismic to Simulation software package Petrel.

Flow Simulations: Four flow simulation cases were considered to assess the influence of reservoir matrix, fracture network, and water drive. Simulations were performed using Schlumberger’s ECLIPSE simulator in “prediction” mode. Case 1 consisted of a single porosity model that incorporated only matrix flow. This simulation case failed. Without a water drive, the reservoir pressure dropped rapidly and the simulation terminated. Case 2 incorporated the matrix and water drive. This case produced an insignificant amount of oil. Case 3 considered the fracture network and water drive without the matrix. This case also produced very little oil. Case 4, incorporated the matrix, fracture network and water drive; however, successful simulation was not obtained. Significant difficulties encountered in the flow simulation step that could not be overcome within the timeframe of the grant. Preliminary flow simulations suggest that upscaled fracture porosity and permeability are too small.

Impact: The PRF ACS grant helped the PI investigate several issues associated with reservoir model development and flow simulation.  The results led to development of another proposal submitted to DOE to continue this effort. Recent advancements in the NETL code NFFLOW permit multiphase dual porosity flow incorporating a discrete fracture network.  Valerie Smith, the student working on this project, presented two papers on her research at the 2008 Annual AAPG meeting in San Antonio, Texas. The project also supported a two week visit to Schlumberger’s Calgary office where she worked closely with Schlumberger’s flow simulation group. During her tenure in the department she also helped as a part time instructor in two Petrel workshops. As a direct outgrowth of her efforts and dedication to the project Valerie was hired by Schlumberger in the Fall of 2008 as a reservoir geophysicist to work in their Carbon Services group.

Conclusions: We developed a detailed reservoir model that could be carried directly in to ECLIPSE for flow simulation. The reservoir model incorporates a fracture network initially derived from FMI log analysis, field observations and empirical relationships published in the literature. Initial flow simulations were undertaken in close collaboration with Schlumberger. Four simulation test runs were completed during the project. These included: 1) a matrix only simulation without water drive; 2) a simulation run including the matrix with water drive; 3) a simulation incorporating the fracture network with  water drive but no matrix; and 4) a dual permeability simulation that incorporated reservoir matrix, fracture network and water drive. Initial runs resulted in underproduction of oil and excessive water production. The dual permeability run encountered run-time difficulties and could not be completed.

This study was undertaken to obtain a better understanding of the properties of the reservoir fracture network and the role it plays in historical oil production from the field. The intent was to update the reservoir model in a manner consistent with the local geology that would allow a good history match to be obtained in flow simulation. In the end we could not reach this goal. Continued work is needed to complete the effort. The potential benefits of continued efforts will yield improved estimates of oil revovery factors in tandem with CO₂ sequestration potential.

Acknowledgements: Appreciation is extended to individuals with the Rocky Mountain Oil Testing Center and National Energy Technology Laboratory for providing data and helpful advice. We thank Alan Brown (Schlumberger) for his comments and perspective and for providing Petrel and Eclipse software used in the project. Special thanks to Isabelle Pelletier Tardy and Samer Mualla of Schlumberger for their help in the modeling process. Discussions with Duane Smith, Mark McKoy and Grant Bromhal ( Morgantown, WV Energy Technology Laboratory) were greatly appreciated.