1
Hydraulic Fracture Geometry from Mineback Mapping

R. G. Jeffrey

SCT Operations Pty. Ltd., Wollongong, New South Wales, Australia

KEY POINTS

• Details of full-size hydraulic fracture geometry are only available from mapping of mined fractures.
• Hydraulic fractures mapped in naturally fractured coal, sandstone, and in crystalline rock show many similarities that result from interactions with natural fractures and shear structures.

1.1 Introduction

Hydraulic fractures have been mapped during mining in a range of rock types and in a variety of geologic and in situ stress settings. Mapping has occurred for fractures placed into clay and soil [1], welded tuff [2], coal ([3-5]), andesite [6], and other crystalline and metamorphic rocks in porphyry ore bodies [7-9].

Coal seams are typically fractured for the purpose of stimulating production of seam gas (coalbed methane), either for commercial use of the gas or to improve drainage of gas from the coal before mining. Approximately 50 hydraulic fractures have been mined and mapped in detail in coal seams in the United States and Australia ([3, 4, 10, 11]; Jeffrey et al. 1993). In comparison, fractures mapped in other materials (soil and rock) total less than 20, with 10 of these located in porphyry copper and gold orebodies ([6, 7, 9]; containing intrusive monzonites and metamorphosed volcanic sediments. These fractures were placed as part of investigations into fracture geometry expected to be produced by hydraulic fractures used to precondition the orebody in advance of mining. The fractures placed into welded tuff were part of early research into hydraulic fracture growth [2] and the soil fractures were part of a study of fracturing for waste remediation [1].

By comparing the fracture geometry mapped in these different natural materials, common and disparate features of the fractures are highlighted. To help with the comparison, dimensionless groups that have been shown to be important in hydraulic fracture growth are calculated or estimated for each mapped fracture. There is an extensive body of work using experimental, analytical, and numerical methods to investigate interactions of hydraulic fractures with bedding and natural fractures and faults. This paper limits itself to the mapped geometry exposed by mining of full-size hydraulic fractures and presents a comparison of features found in coal, sandstone, and stronger metamorphic or igneous rocks.

1.2 Summary of Mapped Fracture Geometries

Selected hydraulic fractures that were placed in coal, sandstone, and hard rock are described in this section. For each fracture the treatment parameters, rock properties, and in situ properties are listed.

1.2.1 Fractures in Coal

Two fractures placed into coal seams will be described, one located in Australia as described by Jeffrey et al. (1993) and one located in the United States [10]. The fracture descriptions include details of the treatment and site characterization, including rock mechanical properties and in situ stress data.

1.2.1.1 DHM-7 Fracture

As part of a program to better understand hydraulic fracture stimulation of coal, well DHM-7 was fractured using linear gel with sand proppant [10]. DHM-7 was drilled and completed open hole through the Blue Creek coal seam in Alabama. The well was located over the Oak Grove Mine in the Warrior Basin. Table 1.1 summarizes the parameters of the site and the treatment parameters. It is difficult to measure sH in coal because the cleat and natural fractures make overcoring impractical and the determination of fracture initiation is difficult when using a microfrac stress test. A value for sH that is larger than the vertical stress has been selected because the mine back mapping revealed no development of propped fractures at right angles to the direction of the main vertical fracture branch.

Room and pillar mining exposed the fracture in the coal rib around the sides of two pillars and along the coal rib nearest the well (Figure 1.1). A vertical and horizontal propped fracture (T-shaped) was found with the vertical part consisting of a number of parallel strands. The vertical fracture extended for more than 30?m to the north of the well. A horizontal propped fracture, extending over the vertical fracture, was located at the coal-roof rock interface and was mapped in detail (see Figure 1.2). The fracture was not mined to the south, but based on the area and fracture widths mapped, the propped fracture mapped on the north side of well DHM-7 was estimated to contain approximately 75% of the sand proppant injected [12].

The treatment pumped into DHM-7 used a non-crosslinked guar-based gel fluid that was injected at an average rate of 8.3 barrels/min (0.022?m3/s). A thick resin-coated sand system was pushed into the near-well part of the fracture on the morning following the main treatment. This resin coated sand was used to test its ability to stabilize the wellbore region. It is designed to retain 80% of the sand's permeability after curing. Much of the vertical fracture exposed nearest the wellbore, at location A in Figure 1.1, was filled with this resin-coated sand. Samples of this propped fracture were excavated and taken from the mine for later analysis and display. No resin-coated sand was found at the next exposure at location B. Mapping of the horizontal fracture was done at the detail level represented in Figure 1.2 along all propped exposures.

Table 1.1 DHM-7. Coal, well, and treatment parameters.

The mapped geometry in DHM-7 and other cases presented below rely on the proppant to mark the fracture path. The hydraulic fracture typically extends beyond the proppant, especially when less viscous fluids are used. Hydraulic fractures cannot in general be found or mapped if they do not contain proppant. The distribution of the proppant, especially in a horizontal fracture, depends on both the fracture width and the fluid velocity field. Proppant transport in horizontal fractures is an area of study that has received little attention, primarily because horizontal fracture growth is thought to be a rare occurrence at depths greater than approximately 300?m. The hydraulic fractures described below that were placed into the orebody at Northparkes at a depth of 580?m were horizontal. T-shaped fractures are common in stimulation of coal and better models that can deal with horizontal fracture growth and the associated proppant transport problem would be welcome.

1.2.1.2 DDH 190 Fracture

An uncased HQ-size cored borehole (DDH 190) was drilled through the German Creek coal at Central Colliery in Queensland and was hydraulically fractured using a borate crosslinked hydroxypropyl guar gel fluid. The site was characterized by undertaking well testing, stress measurement, core testing, and fracture testing before the main fracture treatment. Table 1.2 summarizes the site parameters relevant to the treatment as given by Jeffrey et al. [5].

Mapping of the fracture during and after development of the roadways in this area of the mine revealed a vertical fracture in the coal that extended into the roof rock (Figure 1.3). The fracture trace in the roof rock (Figure 1.3a) was primarily a single fracture, but interactions with natural fractures resulted in the formation of some offsets and short parallel branches. The vertical fracture trace in the coal at the north side of 13 cut-through (Figure 1.3b) was typical of other vertical sections mapped at this site, consisting of a single fracture that interacted with bedded and sheared coal. The 150?mm-thick mid-seam shear zone (mssz) runs through much of the German Creek seam and is composed of sheared coal, with particles ranging from clay size to a few centimeters in size. The mssz is softer and weaker than the...