SBIR-STTR Award

A significant portion of known gas reserves is containedwithin heterogeneous reservoirs for which well logging, usingcurrent analytical techniques, often can provide only qualitativeinformation. This fact is due to the indeterminate nature ofgeologic signal processing, combined with the inherent limitationsof using mechanistic approaches to analyze the relationships ofmultiple signals in complex geologic formations. HydrocarbonSignature Logs (HSLS) identify producing zones with a greaterdegree of accuracy than can be derived using conventional wellboreanalysis, substantially increasing the hydrocarbon discovery ratefor obscure reservoirs. HSLs are created using a proprietaryprocess called PROWLS (Pattern Recognition for Wellbore LogSuites). PROWLS is based on an emergent pattern recognitiontechnology called neural computing, developed primarily throughDepartment of Defense-sponsored research, to address the problemsof identifying military targets in difficult environments. This technology has been successfully adapted to identifyoil-producing zones in a well using a suite of conventionalwellbore logs. The purpose of this project is to adapt PROWLS forsubsequent mapping of development wells. The Phase I research isdeveloping the database for the experiment, producing HSLs thatindicate gas producing zones, and incorporating mapping into PROWLSto demonstrate increased gas recovery.Anticipated Results/Potential Commercial Applications as described by the awardee:PROWLS is an empirical methodology, independent ofheuristics, and does not require expensive re-entry of the borehole to collect additional information. The experience andsystems developed in Phase I will be extended and tested duringPhase II on other heterogeneous reservoirs to confirm conclusivelythat PROWLS is broadly applicable. Phase II results shoulddemonstrate a methodology that can be marketed as a service andpotentially bundled as a software/hardware solution.

A significant portion of known gas reserves is contained withinheterogeneous reservoirs for which well logging, using currentanalytical techniques, often can only provide information of aqualitative type. This is due to the indeterminate nature ofgeologic signal processing, combined with the inherent limitationof utilizing mechanistic approaches to analyze theinterrelationships of multiple signals in complex geologicformations. Hydrocarbon Signature Logs (HSLS) identify producingzones with a greater degree of accuracy than can be derived usingconventional wellbore analysis. This can substantiallyincrease the hydrocarbon discovery rate for obscure reservoirs. HSLs are created using a proprietary process termed PROWLS,(Pattern for Wellbore Log Suites). PROWLS is based upon anemergent pattern recognition technology called neural computing,that has been developed primarily through Department of Defensesponsored research to address the problems of identifyingmilitary targets in difficult environments. Phase Isuccessfully adapted this technology so that it can be used toidentify oilproducing zones in a well, -using a suite ofconventional wellbore logs. Proof of concept was initiallydemonstrated using the Silurian Interlake formation, which is anUpper Interlake Subgroup of the central Williston Basin, thatproduced oil and gas from sequences of thinly interbeddedperitidal dolomites and calcareous dolomites. On the NessonAnticline, the Silurian interval is recognized as an area with ahigh potential for bypassed production because of the extremedifficulty of identifying pay zones using conventional loganalysis. HSLs developed for wells in this formation identifiedproducing intervals to a high degree, that were not achievablewith conventional analysis. Phase I demonstrated that PROWLS canreliably estimate production of tight gas sands based uponpatterns contained within the log suites. Further,PROWLS accurately identified all of the dry wells within thestudy field. HSLs were produced that showed definitivesignatures indicating downhole porosity and permeability of theproducing sand. Maps were produced that identified producingtrends for the study field. Phase II will define processcapabilities and limitations. Procedures for model verificationand validation will be designed. The PROWLS process will beimbedded into an existing commercial well log software system. Extensive demonstration cases will be developed to show systemcapabilities.Anticipated Results/Potential Commercial Applications as described by the awardee:PROWLS will be a software product thatwill allow oil and gas professionals, with no prior experience inneural computing, to take advantage of this state-of-the-artpattern recognition technology. Innovative applications of thistechnology will facilitate additional gas production fromdepleted or nearly depleted fields, enhance development of newproduction, and create new reserves through discovery of bypassedpay of oil and gas.