07122-23

Primary Performer
Lawrence Berkeley National Laboratory, Berkeley, CA 94720

Additional Participants
Texas A&M, College Station, TX 77843
University of Houston, Houston, TX 77204
Anadarko Petroleum Corp., Woodlands, TX 77380
BGI Resources, LLC, Oakland, CA 94606

Abstract

Objectives: Using a multi-disciplinary approach, to develop a self-teaching expert system that (a) can incorporate evolving geological, geophysical, fracturing, reservoir and production data obtained from an continuously expanding database of installed wells in unconventional tight gas reservoirs (i.e., tight sands, shale or coalbeds), (b) continuously update the built-in database and refine the underlining decisionmaking metrics and process, (c) can make recommendations about formation fracturing and well stimulation, in addition to well location, orientation, design and operation using the most recently updated metrics and processes, (d) offer predictions of the performance of proposed wells (and quantitative estimates of the corresponding uncertainty) in the stimulated formations, and (e) permit the analysis of data from installed wells for parameter estimation and continuous expansion of the data base of the expert system.

Deliverable: The deliverable of this project is a self-teaching expert system that can be a vital tool in the attempt to increase reserves and successfully produce gas from shale formations, and to increase production from already producing systems. The final product is not just the development of an abstract approach or methodology, but a computer program that is easily installed and executed on a wide variety of computational platforms. To fully realize the benefits of the self-teaching expert system, a promising approach is its storage at a central location and access through a Web-based application. Note that the data that are entered into the database are treated as confidential, with the user not knowing their origin without the explicit consent of the data owners. Although the geographical location associated with the data may be disclosed, the data provenance and ownership will not. Thus, the user benefits from the data availability to design more productive production systems without compromising confidential information belonging to the entities that provided the data.

Potential Impact: Successful development of the proposed self-teaching expert system is expected to result in a significant (possibly quantum) increase in both reserves and production by providing a technology that will significantly reduce the uncertainties associated with such systems, thus bringing previously inaccessible energy resource to production.

Principal Investigator: George Moridis