The goal of the project was to investigate applicability of soft computing techniques such as fuzzy logic, pattern recognition, and genetic algorithm to identification of the most influential parameter impacting production rate and ultimate recovery and development of an optimized approach using conventional (crisp as opposed to fuzzy) data analyses. The ultimate goal was increased production by achieving maintainable production rate and increased ultimate recovery through optimization of the upstream oilfield practices.
The project was selected under the Preferred Upstream Management Practices (PUMP) solicitation DE-PS26-01BC15304 issued in the fall of 2000. PUMP is aimed at pairing "best practices' and solutions coming from new technologies to an active campaign of dissemination information to domestic producers. PUMP goals are to slow the decline of domestic oil fields and to maintain the infrastructure to continue to produce oil as a vital part of National Security.
Gas Technology Institute
Des Plaines, IL
West Virginia University
The U.S. Department of Energy has been supporting research and development projects aimed at increased domestic production in terms of ultimate recovery as well as production rate, through optimization of the upstream practices. Optimization of any process involves characterization of the processes and improvements upon them and characterization entails analysis of the existing data. However, in the case of oil field data, specifically regarding fields with long production history, the data is inaccurate, imprecise, and incomplete; and as such, accurate operations characterization is next to impossible.
Soft computing techniques, collectively referred to as Virtual Intelligence techniques, mimic the human brain in that it can decipher incomplete data, recognize patterns, and identify the net result of the cause-and-effect phenomena without resolving the intrinsic makeup of the phenomena. As such, application of the virtual intelligence techniques would identify the areas where elaborate conventional engineering and geoscientific analyses would resolve the cause-and-effect process. At that point, the operations parameter would be re-designed (optimized) to produce the most desirable result. This approach was used on data from three producing fields in Oklahoma.
The project was implemented in three fields in the Golden Trend of Oklahoma. Application of soft computing techniques pointed to hydraulic fracturing as the most influential operation relative to production rate and ultimate recovery. A set of guidelines and recommendations for optimization of hydraulic fracturing was developed. A virtual intelligence software, specially developed for these fields, was developed and delivered to the participating producers. In the meantime, production data analyses pointed to wells that although in close vicinity of one another, had different production history. Advanced seismic data processing using wavelet based spectral decomposition techniques revealed the presence of severe faulting between these wells. Results were presented to these companies for their consideration in their in-fill drilling projects.
The fundamental benefits from this research and development project was the proof that soft computing has the capability of analyzing the heterogeneous and often incomplete data from producing fields time- and cost-effectively. Results from these virtual intelligence techniques would then point to specific area/operation for detailed geoscientific and engineering studies. The net benefits would thus be increased production efficiency and increased production.
The project was implemented in three fields operated by three independent operators in the Golden Trend in Oklahoma. Detail stimulation and production data from more than 230 wells were collected and analyzed using the virtual intelligence methodology. These analyses revealed that the most influential controllable parameter effecting the production rate and ultimate recovery was fracture stimulation, and as such, the bulk of soft computing efforts were focused on determination of optimized hydraulic fracturing procedure. Results of these studies were presented to the participating producers as lists of re-completion candidates and the recompletion characteristics.
Production data analysis showed that in parts of the area wells that were in close vicinity of one another had vastly different production. Elaborate seismic data processing using wavelet based spectral decomposition techniques revealed the presence of severe faulting between these wells. Results were presented to the participating producing companies for their consideration in their in-fill drilling projects. Figures 1 is the wideband seismic section and figures 2, and 3 show results of spectral decomposition on data from the same line.
Technology transfer was achieved through several in-house workshops carried out in the offices of the participating producers. In these workshops, the computer model developed specifically for the golden trend was explained in a series of hands-on tutorial sessions.
The three Oklahoma producers that participated in the project are using the results and one company has indicated their interest in participating in any follow-up work that would expand the technique beyond the Golden Trend. Since the publication of the final report we have received several enquiries from the US and foreign producers.
Final report, DE-FC26-01BC15271: "Development of a Virtual Intelligence Technique for the Upstream Oil Industry"; Salehi, I. Gas Technology Institute, Mohaghegh, S.D., Intelligent Solutions Inc., Ameri, S, West Virginia University, September 19, 2004, 177 pages.
Mohaghegh, S.D. et. al.; "Analysis of Best Hydraulic Fracturing Practices in the Golden Trend Fields of Oklahoma," SPE 95942, SPE Annual Conference, Dallas, October 2005.
Salehi, I., Walker, G.; "PUMP Project: Quantifying Best Practice Analysis to Cut Costs and Boost Output", GasTIPS, April 2004, p 25-30.
$614,863 (52 % of total)