The two main objectives of using a hydraulic fracture model in coals are:optimization of job design and placement, andpost-fracture diagnostics.
Though certain limitations still exist, hydraulic fracture modeling in coals has undergone major advancements in the past decade. Pseudo or lumped three-dimensional (3-D) models have usually been employed to try and meet the objectives mentioned above. The effectiveness of using such models in coals has been very limited. Fully functional 3-D models are currently available in the industry and can be used to obtain better estimates of the fracture dimensions. This paper shows that a grid-oriented fully functional 3-D fracture simulator with shear decoupling can be especially useful in coals for post-fracture diagnostics if sufficient input data can be fixed from logs and Diagnostic Fracture Injection Tests (DFIT).
When problems occurred with placement of fracture treatments, especially in the upper Fruitland coals of a San Juan basin project, instead of experimenting with various ideas arbitrarily, technical evaluation using DFIT data and fracture modeling with a grid-oriented, fully functional 3-D fracture simulator, was used to pin-point the issues and address them accordingly. This paper will discuss the deficiencies of using 2-D and pseudo-3-D fracture models in coals and also will discuss the lessons learned from using a grid-oriented, fully functional 3-D fracture simulator in this project along with the successful implementation of the results developed from the modeling work.
Literature Review
Over the last 25 years, numerous papers have been published regarding hydraulic fracturing and modeling in coals.1–19 Trevits, et al.1 investigated how well the fractures are contained in coal during a stimulation treatment. In this study, 12 vertical wells were drilled, completed, and subsequently evaluated via mining excavation. The treatment designs were adjusted using the information gathered from excavation. Layne, et al.2 through their U.S. Department of Energy (DOE) work wanted to advance the hydraulic fracturing techniques in coals by improving the understanding of fracturing treatments using mineback observations. They used five fracture models that were developed under contract to the DOE to analyze the mineback observations. Jones, et al.3 evaluated 130 fracturing treatments in their study. They observed high treating pressures during fracture treatments in Black Warrior basin and attributed the following as possible causes:high pore pressure buildup due to fluid leakoff,multiple fracture propagation, andfracture-tip plugging by coal fines.
They modified a pseudo-3-D model to replicate a fracture tip-screenout caused by coalfines plugging within the pad. Jones, et al.4 presented another paper addressing the unusual treating pressures observed during hydraulic fracture treatments in coal. As part of that work, they presented methodologies to design fracture treatments in coal. Logan, et al.5 performed extensive fracture treatment diagnostics in two wells at the GRI Deep Coal Seam Project test site. Data collected from such diagnostics was presented in their work. Holditch, et al.6 described four basic scenarios to explain fracture growth in coal reservoirs but did not discuss the need for a fully functional 3-D fracture model to address these four scenarios. The fracture model used by Jones, et al.3,4 was used in a study by Saulsberry, et al.7 to determine the fracture geometry and its effectiveness from two separate hydraulic fracturing treatments in the Oak Grove field of Alabama. In the same field, Palmer and Sparks 8 used a downhole camera to study the propped fractures in coalbeds. The information gathered was then used to calibrate a pseudo-3-D fracture model for optimizing stimulation treatment designs. The authors clearly state that more detailed modeling required the use of a fully functional 3-D fracture simulator.
