GOLDEN, Colo.— The Reservoir Characterization Project (RCP) put its Phase XVI work in the spotlight at its spring meeting held by the Colorado School of Mines. The project, conducted with field sponsor Anadarko Petroleum Corp. (NYSE: APC) in the Wattenberg Field, focused on unconventional shale reservoirs, with students presenting time-lapse 9-C seismic data; regional seismic; microseismic; geological information; and reservoir engineering and completions information to analyze performance within the Niobrara and Codell reservoirs.
“The RCP program seeks to connect student and industry to create a greater understanding of the stimulated reservoir volume and produced reservoir volume to guide future drilling and completions decisions and help diagnose and predict reservoir performance,” said Ali Tura, director of RCP and a professor of geophysics.
The Wattenberg Field study area concentrated on a one-sq-mile section on Anadarko’s property with 11 horizontal wells drilled in two reservoir zones. The section was designated a test area to optimize well spacing, hydraulic fracturing parameters and other engineering-driven questions to study the usefulness of dynamic seismic reservoir characterization.
An initial multicomponent (9-C) 3-D seismic survey was acquired over the focus area after drilling but before fracturing and production. The survey will serve as a baseline for future monitoring with seismic surveys. It also is being used for static parameter analyses including brittleness, natural fracture characterization, geologic structure, stratigraphy and stress state.
A second survey was acquired immediately after the wells were fractured to study differences in the earth’s response to stimulation. Finally, a third survey was conducted during the winter of 2015 to study the response of the reservoir to multiple years of production. The integrated study includes data provided by Anadarko, including core, well logs, microseismic, chemical tracers and the production data.
Colorado School of Mines Assistant Professor Dr. Whitney Trainor-Guitton presented the initial results for the data analytics. In her presentation, Dr. Trainor-Guitton acknowledged that, “The seismic interpretation challenges we are dealing with today are large 3-D cubes, with multiple versions of these cubes and lots of diverse data. It’s hard for humans to look at such large chunks of data to make sense of all of the variables—we’re turning to algorithms to see trends and patterns.”
She described the algorithms for the program families as “supervised learning” and “unsupervised learning.” With supervised learning, an algorithm is given a dataset with classifications of data such as a general (data) regression. A new development algorithm is called “random forest,” where an algorithm might predict “lab quakes” or fault failures.
With unsupervised learning, an algorithm adapts to data to identify natural patterns or clusters, such as neural networks, which have been around for a while but are gaining popularity because of increased computing power. Another example of unsupervised learning is “clustering,” which is useful for creating maps and multidimensional scaling.
Another student, Jake Utley, presented a study to increase Wattenberg Field recovery. His report focused on Niobrara/Codell wells in the study area that are producing about 20 barrels of oil per day in a section, called “The Wishbone Section.”
“The Wishbone Section has 11 horizontal wells—four in Codell and seven in Niobrara—that were completed with zipper fracs where we had the tightest well spacing, we had the best production,” he said.
He said the RCP team is putting together a fully integrated dynamic simulation model to understand heterogeneity, structure and the dynamic effects for stimulation and production.
The model the team used observed geological outcrop studies and initial seismic interpretations. Datasets were incorporated from a 50-sq-mile merged 3-D survey, a 10-sq-mile 3-D survey and a four-sq-mile survey.
A nine-component, time-lapse survey was conducted during stimulation, one after stimulation and one after two years of production in the main study area.
The group also tied the seismic interpretation results from MWD data, picked log tops, as well as the mud logs to track the location of the well itself.
Student Tom Bratton, who is working on developing a rock physics velocity model that varies with stress said his study, “will help establish reasonable values for stress sensitive coefficients to determine where fluid flow and stress and velocity changes have occurred. This will help us better estimate well spacing and fracture spacing to get more hydrocarbons out of the reservoir.”
As part of the rock physics-velocity model, student Ahmed Alfataierge’s investigation looked at integrating hydraulic fracturing and seismic signatures. His study will help determine how fractures will propagate, and to establish geometry and conductivity in non-homogenous rocks with a 3-D geomechanical model that is input into a hydraulic fracture simulation model. The model could be used for enhanced recovery or infill drilling.
Also at the meeting, two proposed studies were presented by companies. Denbury Resources proposed a study on an area on the Cedar Creek anticline in the eastern Montana portion of the Williston Basin. The company plans to conduct a CO2-enhanced recovery using horizontal drilling. The plan is to determine how the reservoir will respond to CO2 from both a cost and production standpoint.
The company also discussed a second proposed project in Bell Creek Field in the northern part of the Powder River Basin in Montana. The company is deciding where to drill wells, how much fiber should be installed, the best way to install the fiber in the well as well as CO2 use.
Steve Cole, of OptaSense, proposed a study on reservoir characterization with digital acoustic seismic, saying it “could be a new and powerful tool for real-time fracture profiling, measuring the effectiveness of a treatment program and calculating the injected proppant a fluid volumes.”
Cole said a fiber-optic cable can be used as a downhole seismic receiver, instead of downhole geophones to record vertical seismic profiling and microseismic. “The entire length of the cable can be used as a sensor and can be interrogated as finely as we like.”
Fiber installation can be permanent, installed behind the casing, or semi-permanent, attached to production tubing. It can also be a temporarily deployed by wireline or with coiled tubing and can monitor hydraulic fracture profiling flow during testing or production.
Larry Prado can be reached at firstname.lastname@example.org.