GOLDEN, Colo.—The fall 2018 sponsors meeting of the Colorado School of Mines’ Reservoir Characterization project (RCP)—an industry funded geophysical research consortium dedicated to the study of integrated reservoir characterization—was held at the college in Golden, Colo., to present up-to-date results from U.S. and international advanced studies of unconventional resources.

The field studies project, now in phase 17, seeks to advance geophysical research to explore new methods and technologies for better reservoir performance, analysis and resource optimization in U.S. and global geological settings.

Eagle Ford

This project focuses on permanent seismic monitoring to understand reservoir behavior from the initial completion through the life well.

At the south Texas project, permanent seismic monitors have been placed at 42 well sites belonging to Penn Virginia and Devon Energy in a 50-sq-mile area in Lavaca County, Texas. Using seismic monitoring will help researchers understand changes in the reservoir during hydraulic fracturing. The long-term goal is to integrate research for reservoir characterization and monitoring including incorporating time-lapse, multicomponent (4-D) surface seismic and vertical seismic profiles (VSP); microseismic monitoring; fiber optic distributed acoustic sensing (DAS) and distributed temperature sensing (DTS) engineering and production data.

Using the data will help measure the effective stimulation from its initial completion through the life of the well and develop a plan for the life of the field’s development. They also hope to evaluate and understand stimulation design and well productivity including optimal well spacing, frac cluster spacing and number of frac stages and ultimately, develop a strategic plan for field development.

Recent progress, described by student Adam Tuppen, includes the set-up of a discrete fracture network (DFN) and they have used well logs, image logging, gamma ray logging and density and dipole sonic logging tools to estimate reservoir rock properties. The next step with the data is to run a pre-stack inversion as well as other inversions.

In addition, 3-C microseismic monitors were put in place and then the well was fractured, monitored with VSP, fractured again and monitored by an array of surface monitors with VSP.

Integrated data analysis after fracturing and monitoring include incorporating reservoir properties, geology and geochemistry with the VSP, DAS and VSP data. One of the unknowns in the data was total organic compound (TOC) measurement. According to Tuppen, they had to make calculations using the density and gamma wireline tools— once the TOC was added to the model, they could begin to evaluate the ‘sweet spot’ for fracture placement. Tuppen also concluded that the tuning affect the data but not the inversion, inversion results need to be analyzed before interpretation begins and some issues with the data still need to be understood.

Student Whitney Schultz discussed 3-D seismic integration and the difference between the geophone and the DAS data. She found that reflectors in P-wave, zero-vertical seismic processing are aligned properly with surface seismic reading. In addition, she noted that geophone rotations must be improved for accurate shear wave analysis.

Via livestream from her home in Poland, student Weronika Kaczmarozyk discussed her Eagle Ford findings on integrated microfracture modeling based on well data. She integrated data from geological, geomechanical and reservoir properties to make a static model. “By integrating a detailed analysis, we could increase well production and reduce completion costs, and all we need is to understand the reservoir.”

Her study on fracture modeling included defining fracture drivers, structural models and discrete fracture modeling. Such modeling could determine fracture networks in three dimensions including geometry, orientation and distribution of fractures. She built a structural model based on seismic data from Austin Chalk, Lower Eagle Ford and Buda. “Integrating the well and seismic data improves the stimulation design, and modeling reduces cost and increases well productivity.”

They have built a compositional reservoir model for one of the study wells using the data sets to get a history match, for the reservoir model. They determined that the history-matched reservoir was able to forecast future production, which was confirmed from public file data.

Wolfcamp, Midland Basin

The Wolfcamp team is studying the use of distributed acoustic sensing (DAS) to characterize the reservoir and stimulation. According to student Diana Tomayo, they are using a unique dataset with vertical seismic profiling (VSP) records taken before and after 78 stages or fracturing. “Our research analyzes velocity changes and scattering effects and uses modeling to assess the possible mechanisms that cause the time-lapse changes. This will allow us to determine fracture height and asses future optimized DAS data.”

They used a DAS data set from a well that was drilled and completed Apache Corp. and they are using a cemented a fiber optic cable in the well from heel-to-toe. Ms. Tomayo explained that the fixed vertical seismic sources were placed about one mile north and south of the well and in line with the horizontal part of the well creating a 30-degree angle of incidence at the toe and 45-degree angle of incidence at the heel and that “this would allow us to see both P- and S-waves in our data set to characterize hydraulic fractures”.

“Our objectives for the study are to look at lossy changes, amplitude changes and scattering effect that are caused by the 78 stage of fracturing.” They looked a P- and S-wave time lapse response to characterize hydraulic fracture geometries and analyze fracture dynamics with the time lapse response. “This will tell us how our fractures are opening and closing.”

Student Gary Binder reported that modeling stress-induce velocity changes from fluid leak-off may quantitatively explain many features of the observed time delays and that more work is need to improve the model and calibrate the logs and cores with pumping data. They may also try to connect the DAS VSP and pumping data into a single dynamic geomechanical model. In addition, amplitude and S-wave shifts may also provide useful or better detail.

Vaca Muerta, Neuquen Basin, Argentina

The Vaca Muerta formation is in western Argentina’s Neuquen Basin, which covers about 120,000 square miles east of the Andes. The basin has a low structural complexity, flat horizons and vertical strike/slip faults. The Vaca Muerta is upper Jurassic/lower Cretaceous with shales, marls and limestones that is up to 600 meters thick. The estimated total organic content is 2-12% and is also estimated to hold 16 billion barrels of oil and 309 trillion cubic feet of gas.

Sponsoring companies Wintershall and Repsol-YPF have provided a study area and legacy seismic data (from 2004), wide-azimuth 3-D seismic and multicomponent 3-D seismic surveying and five wells.

The research focus is on an overpressured gas accumulation in tight sands of the Lajas and Punta Rosada formations. According to student Pablo Benitez, they will analyze depositional controls on rock properties, fracture distributions and the variability of the mechanical properties across the study area.

Fellow student Patrick Corwin said the study will use recently acquired multicomponent seismic data to analyze stress and natural fractures and integrate geomechanical models to improve Differential Horizontal Stress Ratio (DHSR) analysis to tie results to microseismic surveying.

Comments