Drilling Risk Reduction with a Fuzzy Expert Exploration Tool

BALCH, ROBERT. S., TONGJUN RUAN, and SUSAN SCHRADER, Petroleum Recovery Research Center, New Mexico Tech, Socorro New Mexico, 87801

    Expert systems are artificial intelligence tools which store and implement expert opinions and methods of analysis. This paper describes an expert system for oil prospecting developed for the Lower Brushy Canyon and Devonian plays of southeast New Mexico. The Fuzzy Expert Exploration Tool (FEE Tool) successfully blind predicted drilling risk and has made low-risk drilling recommendations at 4481 Brushy Canyon prospects with reserves totaling 212 Million Bbls.

    Brushy Canyon regional geologic and geophysical data were mapped on a 40 acre grid, corresponding to 60,478 potential drill-sites. Production data was available at 911 points (existing wells), verified to have Brushy Canyon production. Another 75 wells were verified as Brushy Canyon dry holes. The Expert System was developed using input from Explorationist’s and statistical analyses of the data. Prospect quality was calculated at all 60,478 locations with a mean estimate of 0.476. A total of 5556 prospective well sites had predictions in the range {0.65, 1.0} and are considered favorable drilling sites. To validate these results, 89 wells drilled since the input of new data to the project was frozen, were blind tested with a mean estimate of 0.70, indicating that these wells would have been recommended by the FEE Tool.

    Subtracting existing wells, 4481 low risk prospects remain un-drilled in the lower Brushy Canyon representing reserves of 212 million Bbls. Ongoing work includes migrating the software to the Devonian carbonates of southeast New Mexico, with an ultimate goal of an easily customizable, user-defined tool capable of defining any play.