HORIZONTAL DRILLING FOR ONSHORE GAS FACES MANY CHALLENGES

Iraj A. Salehi
Gas Research Institute
Chicago

Even though cost-effective in oil fields and some prolific offshore gas fields, horizontal well technology has not yet been successfully transferred to onshore gas fields.

The major technical challenges are accurate reservoir characterization, proper well design and placement, and stimulation.

The first two of these challenges can be met with the state-of-the-art technologies but the third requires emerging knowledge, tools, and techniques.

To determine the application of horizontal wells to typical gas reservoirs in the continental U.S., the Gas Research Institute (GRI) initiated an applied research and development project in late 1990. The project's first phase was focused on two principal problems:

1. Development of an analytic screening tool for economic evaluation of horizontal wells vs. vertical completion techniques

2. Assessment of the state-of-the-art technologies for selecting the azimuth of horizontal wells.

The second phase, planned for 1993 and 1994, aims at development of production stimulation methods and techniques.

BACKGROUND

In the late 1980s, horizontal wells emerged as an effective method to develop certain oil producing basins and geological trends. High initial production rates justified the risks and compensated for high drilling costs. The short pay-back periods dramatically improved the return on investment and fostered the 1990-1991 horizontal drilling boom.

Attempts to drill horizontal wells for gas have been sporadic, with mixed results at best. In fact, of the total horizontal wells drilled for gas from 1988 to early 1992, a mere 2.7% have been reported as successful gas completions.'

The lack of success is partly due to the differences between oil and gas reservoirs relative to flow characteristics and reservoir settings. Of equal significance is the market price of natural gas.

Generally speaking, horizontal completion is preferred under these conditions:

• Foremost, better economics in terms of return on investment and/or higher ultimate recovery.

• Reservoir conditions not amenable to conventional completion and production enhancements.

• Geomorphologic, environmental, and cultural conditions inhibiting conventional drilling operations.

If horizontal wells are to compete with conventional methods, these wells must have production rates and ultimate recoveries significantly higher than from techniques such as hydraulic fracturing.

CANDIDATE SCREENING

Considering the high drilling and completion costs of horizontal wells, prediction of production performance is critical in assessing the financial risks and returns.

To develop a simple analytic method for estimating horizontal well production, GRI funded Maurer Engineering Inc. to develop a PC-based software program (GMOD) for determining well performance as a function of well design and reservoir parameters (e.g., porosity, vertical and horizontal permeability, and thickness).

This software does first-cut screening of candidate wells prior to elaborate reservoir studies and sophisticated reservoir simulations.

The analytic model computes production from horizontal, vertical, slant, and vertical hydraulically fractured gas wells, Graphical output (Fig. 1) enables a quick evaluation of a gas reservoir's potential with horizontal wells.

RESERVOIR CHARACTERIZATION

For horizontal wells, integrated reservoir knowledge is the most significant prerequisite for success.

In essence, a horizontal well is a guided probe designed to intersect zones of high permeability that are either naturally fractured or in high matrix permeability trends. As such, the term reservoir characterization translates into the understanding of the reservoir's make-up, environment of deposition, subsequent diagenetic changes, and structural deformations.

In short, the knowledge of reservoir attributes, flow mechanism, and state of fracturing, is critical.

Vertical discontinuity is a common feature of most natural gas reservoirs in the continental U.S.

More often than not, completed zones include a number of sand bodies separated by less permeable to impermeable rocks.

In conventional vertical wells, particularly when the well is hydraulically fractured, all pay zones are open to the well bore. In horizontal completions, the well bore intersects only the targeted zone and the impermeable or low-permeability layers separating multiple reservoir components limit the flow between layers.

Equally detrimental is the presence of thin low-permeability layers that cause low effective vertical permeability within the net pay. Under these reservoir conditions, horizontal completions can be improved by either high-angle slant holes or hydraulic fracturing.

For high-angle slant holes, vertical permeability barriers are removed and the entire reservoir section is opened to the well bore. While this practice is effective in moderate to high-permeability reservoirs, it may not attain economically attractive flow rates from lower-permeability formations.

Because of extensive reservoir rock exposure, both in terms of time and distance, horizontal drilling may severely damage a reservoir. For example, extensive drilling fluid leakoff may damage, beyond remedy, the formation. In tight reservoirs with natural microfractures, fluid imbibition can reduce or eliminate permeability to gas.

Air or foam drilling, and underbalanced drilling will alleviate formation damage to a great extent. However, air or foam can introduce well stability and safety problems that need cautious monitoring and control.

HYDRAULIC FRACTURING

Hydraulic fracturing can improve flow rates and recoveries in multiple pay and low-permeability reservoirs. In essence, the hydraulic fractures act as conduits across all reservoir components. These conduits allow every reservoir segment to feed directly into the fracture.

In moderate to low-permeability reservoirs, the hydraulically induced fractures provide the large surface area necessary for commercially acceptable producing rates. But hydraulic fracturing of horizontal wells opens a whole new category of technological unknowns.

Because hydraulically induced fractures exceeding 1,500 ft are proven to propagate parallel to the in situ intermediate principal stress, 92, the horizontal section should be drilled normal to this stress. The hope is to create one or more hydraulic fractures that propagate normal to the well bore (Fig. 2a). In reality, no evidence exists that such hydraulic fractures can be created.

In fact, a horizontal well may act as a source of asymmetry and problems. The hydraulic fractures are more likely to be initiated along the well bore and twist around to align with the intermediate stress field. Common symptoms are the high breakdown and treatment pressures and very early sandouts because of near well bore tortuosity caused by fracture curling and bending.

On the other hand, if the horizontal well is parallel to the intermediate principal stress, the fracture plane will contain the horizontal section and remain in contact with the well bore. This configuration may work for moderate permeability reservoirs but is not an attractive alternative for low permeability reservoirs where natural fractures contribute most of the flow.

Because open natural fractures are probably parallel to the intermediate stress field, the likelihood is very slight that hydraulic fractures will intersect natural fractures other than those incidentally occurring at the well bore.

Except for multiple and conjugate fracture systems, none of the two mentioned scenarios is a desirable course of action. The common rule of wisdom is that horizontal wells in tight and moderate permeability fractured reservoirs should be aimed at intersecting the existing fracture system (Fig. 2b). in low-permeability reservoirs with few natural fractures, hydraulic fracturing of horizontal wells definitely improves the production rate. In these cases, the well must be oriented parallel to the intermediate principal stress to prevent the hydraulic fracture from twisting. Twists can cause flow restriction and sandouts (Fig. 2c).

Under these conditions, the fracture plane will contain the well bore and slight curling of the fracture is not expected to cause major treatment problems. Naturally, these fractures can be planned as a single fracture or a number of smaller fractures placed along the well bore.

In practice, creating a single long fracture of 2,000 ft or more in length requires a truly homogeneous reservoir, impractical pumping rates, and huge jobs. Therefore, the practice is to pump a number of smaller treatments placed along the well bore (Fig. 2d).

Because the length of these smaller treatments only slightly exceeds the pay thickness, total job cost is not high. However, mechanical devices for zone isolation, or cementing and perforating at selected intervals add to the cost.

Irrespective of the strategy and method in hydraulic fracturing, the number and size of hydraulic fractures depends on reservoir properties. In essence, the optimum number of fractures is determined by reservoir deliverability and fracturing costs.

Simulation of the Mancos B well (Fig. 3) indicates that three hydraulic fractures are optimum for the reservoir.

WELL BORE DIRECTION

Horizontal completions have succeeded best in high permeability and highly fractured reservoirs. As mentioned, horizontal well orientation relative to the existing natural fractures is clearly critical. The maximum number of natural fractures should be intersected by the well bore.

Several methods can determine the azimuth of natural fractures. These include geological, geophysical, well logging, and techniques for measuring the azimuth of hydraulic fractures .

Some of these methods are expensive and each method has its own inherent ambiguity and limitations.

The prudent approach evaluates and calibrates the less costly methods, such as borehole ellipticity analyses, with a more direct method (e.g., fracture overcoming, borehole seismic survey, or tiltmeter survey). Once calibrated in one or two wells in the field, the less costly techniques are used in subsequent Development wells.

REGIONAL STRESS

An area's prevailing tectonic regime can provide information for selecting the azimuth for horizontal laterals. Because natural fractures open most likely in the direction of least pressure, fracture openings are in the direction of minimum in situ stress component. Thus the fracture azimuth should be normal to this direction, i.e., along the intermediate stress component.

The work by Zoback and Zoback2 provides a good starting point for geologic studies. However, one needs to pay attention to a number of pitfalls in using geological interpretation and extrapolation.

First, nearly all publicly available data on stress, including Zoback's work, are regional. Reducing these data to field-size scale involves some degree of extrapolation and interpolation that causes inaccuracies.

Second, local geological variations, particularly in areas on the boundary between two different stress provinces, may prevent accurate extrapolation.

Therefore, regional data are general guidelines that have to be supplemented by site-specific studies such as lineament studies, local structural geology, study of dominant joint systems, and production trends.

SHEAR-WAVES

Shear-wave seismic is a promising geophysical method for delineating the dominant fracture systems. In these surveys, shear-wave velocity anisotropy provides reliable indicators for fracture directions .3 4

However, shear-wave seismic surveys are expensive, cumbersome, and consequently uncommon. Until shear-wave seismic becomes more affordable and more commonly used, fracture azimuth determination will require coupled analysis of geological, geophysical, and reservoir engineering data that converges on a common fracture direction.

WELL LOGGING

Wire line service companies have developed well bore imaging logs and techniques that use sonic or microresistivity tools. Often, natural fractures are clearly identifiable on these logs.

Commonly, borehole image logs are run in the vertical rat hole drilled prior to horizontal kick off, or in a neighboring well. These logs provide information on the state and orientation of natural fractures.

Fracture information from well logs of vertical wells represents a conservative estimate of fracturing. Because well logs are essentially a one-dimensional representation of the formations drilled, even in extensively fractured formations, the probability of intersecting a representative sample in the vertical well is rather low.

In these cases, the orientation of drilling-induced fractures, if present and recognized, can complement the natural fracture data.

For horizontal wells, borehole imaging logs provide direct measures of fracture density along the well bore. However, these logs can be misleading if the well is not drilled normal to the fracture trend.

With a well parallel, or subparallel, to the dominant fracture azimuth, the observed fractures do not represent the reservoir fracture density.

The azimuth of the intermediate principal stress can also be determined from well bore ellipticity analyses.

Under nonuniform horizontal stress conditions, stress concentration around the well bore is highest along the direction of the higher horizontal component. Spalling and breakouts are more likely at these higher stress points.

As a result, a well bore's cross section will become elliptical with the major axis parallel to the intermediate stress field.-5 Because data on cross sectional elongation of wells is readily available from oriented logging tools, this analysis is an inexpensive method for mapping horizontal stresses.

In practice, however, selecting a reliable fracture azimuth is difficult because the direction of well bore elongation is often inconsistent. To be useful, elongation data often have to be combined with information from other methods.

Anelastic relaxation of cores can also estimate the in situ stress orientation.6 In these analyses, an oriented core is placed between pairs of orthogonal strain gauges and the magnitude of strain is measured as the rock relaxes from the in situ stressed condition.

If oriented cores are unavailable, magnetic measurements and paleomagnetic information can arrive at core orientation. This method is inexpensive but may be imprecise.

Similar to all indirect measurements, interpretation of strain recovery faces uncertainty because of rock anisotropy. Nonuniform strain may be due to the rock fabric heterogeneity or grain orientation, as well as nonuniform stress fields.

FRACTURING

Hydraulically induced fractures propagate along the direction of the intermediate principal stress. Therefore, the azimuth of hydraulic fractures indicate the azimuth of open natural fractures. The hydraulic fracture azimuth can be determined in an offset well or by creating a small fracture in the vertical section of the candidate well.

Tiltmeter surveys and borehole seismic surveys are two geophysical techniques for determining the azimuth.

In tiltmeter surveys, the earth's flexure during the fracturing treatment is monitored and measured by a number of instruments. Variations in tilt magnitude and direction provide an estimate of the azimuth of the hydraulic fracture.7

A number of surveys prove that this method is reliable to a depth of 3,000-6,000 ft. This depth limitation is dependent on the instruments' precision, field installation, fracture size, mechanical properties of the fractured rocks and overburden, local geology, and surface conditions.

Seismic signals emanated at the time of rock failure also can indicate hydraulic fracture azimuth. This method analyzes the seismic signal polarization recorded shortly after fracture treatments. Because some low level fracture growth and seismic activity continues, the position of the seismic events can be determined.

The single-well borehole fracture monitoring system (BFMS) developed by GRI, in the late 1980s, has provided valuable information both on fracture azimuth ind fracture height. But because of instrument coupling and other problems of installing the instrument below the tubing end, the method is impractical in through tubing fracturing. Creating a small hydraulic fracture during the initial drilling phase and cutting an oriented core is a third and very direct method for delineation of in situ stress.

In these operations, the induced fracture normally extends a distance below the bottom of the hole and is recovered in the core (Fig. 4)." This method provides the most direct information on magnitude of the minimum principal stress and orientation of the intermediate stress component as well as localized information on fracture density.

FIELD EXAMPLES

Three wells are part of a GRI fieldwide project for developing guidelines for drilling horizontal wells for natural gas.

In this industry-cooperative project with Mitchell Energy Development Co., Dallas Production Co., Chandler & Associates and Belden & Blake, the four companies hosted the field efforts and provided the background reservoir information.

The target formations are the Barnett shale and Davis sands in the Dallas-Fort Worth basin, Mancos B in South Rangely in Colorado, and Clinton sand in Ohio.

The project scope was as follows:

• Develop an analytical tool to estimate production from horizontal wells as a function of key reservoir characteristics.

• Select a number of test sites for field data acquisition and field experiments.

• Gather focused data sets on vertical cooperative wells in selected gas-productive formations.

• Evaluate various methods to determine optimum azimuth angle of horizontal wells.

• Drill and complete horizontal wells in selected formations.

Data were collected and analyzed on vertical cooperative wells from late 1990 to early 1991. Data analysis and evaluation were directed at supporting the subsequent horizontal wells.

As the basis for comparison with horizontal wells, vertical wells in the study areas were fractured to establish the best possible production rates for vertical completions. The techniques for evaluating the stress directions included:

• Overcoring and orientation of an open hole stress test fracture

• Determining the overall height and azimuth of the stress test fracture with borehole image logs

• Obtaining anelastic strain recovery (or core strain relaxation) analyses

• Performing a tiltmeter survey during the hydraulic fracture treatment.

Based on evaluation of the vertical well data, three horizontal wells were drilled in the Barnett shale, Mancos B, and Barnett shale.

Evaluation of the three wells is not yet complete. On each well, additional well testing, stimulation (as required), and reservoir characterization work remains to be performed to fully assess production potential. The performance will then be compared to the vertical wells in the area.

The work, to date, has confirmed that high production rates from horizontal wells drilled in marginal reservoirs depends on successful fracture stimulation.

BARNETT SHALE

The Barnett horizontal well was drilled with fluid to a true vertical depth of 7,868 ft. The total drilled footage was 10,000 ft, of which 2,880 ft, including the curve, were drilled at 85. The 85 angle was selected for intersecting all reservoir segments within the 300-ft thick shale section (Figs. 5a and 6).

The southeasterly direction was normal to the maximum horizontal stress. Borehole-image log data indicated that, similar to the vertical well data, discrete intervals of gas-bearing natural fractures were intersected.

Guided by these fractured intervals, the well was completed with external casing packers to isolate fractured zones and mechanically opened casing ports to produce the isolated zones.

Initial reservoir modeling efforts, based on interconnected natural fractures in the Barnett, indicated a flow potential of up to 1.7 MMcfd.

Although the original plan left the casing uncemented, it became necessary to cement about one half of the lateral section.

Two fracture trends exist at about 60 from one another. This well appears to be an ideal candidate for hydraulic fracturing and a treatment is planned for the fall of 1992.

MANCOS B

The Mancos B well was air-drilled to a measured depth of 4,928 ft. True vertical depth was 3,320 ft and the length of the high-angle lateral was 1,960 ft, including the curve. The main porosity interval is 100-ft thick (Fig. 5b).

The well azimuth of N30W was selected normal to the northeast/ southwest regional trend of faults and approximately normal to the hydraulic fracture azimuth (Fig. 6a).

No noticeable gas show was observed during air drilling, and subsequent testing did not flow substantial gas. However, because direct offsets have good gas production, casing was set and cemented.

Possibly because of near well bore tortuosity, two hydraulic fracture treatments were terminated prematurely due to screenouts during the initial sand stages. Extremely high-stress gradients, greater than psi/ft, also characterized the jobs.

Additional work is under way to meet the technical challenges presented in stimulating gas production from this horizontal well bore.

DAVIS SAND

The horizontal well in the Davis sand (Fig. 5c) was air-drilled to a total measured depth of 7,033 ft. True vertical depth was 4,320 ft, and the length of the westerly lateral extension was 3,000 ft. The well encountered the predicted set of north-south trending natural fractures (Fig. 7b).

The Davis reservoir was found to be very heterogeneous and a significant portion of the well penetrated nonreservoir rock. To take advantage of the natural fractures, the well was completed open-hole.

During an acid clean out, the well severely sloughed and stuck the pipe. At this time the well is producing about 130 Mcfd through the fish. Work on the stuck pipe is planned for the fall of 1992.

ACKNOWLEDGMENTS

The author thanks the Gas Research Institute, Mitchell Energy Development Co., Chandler & Associates, and Dallas Production Co. for permitting publication of this article. Thanks also goes to John K. Aslakson and Jacki L. Kelley for assisting in preparation of this article.

REFERENCES

1. "Horizontal drilling survey," OGJ Energy Database, 1992.

2. Zoback, M.L., and Zoback, D., "Tectonic stress field of the continental United States," Geological Society of America, Memoir 172, 1989.

3. Lynn, H.B., and Thomsen, A., "Reflection shear-wave data collected near the principal axes of azimuthal anisotropy," Geophysics, Vol. 55, No. 2, February 1990.

4. Winterstein, D.F., and Meadows, M.A., "Shear-wave polarization and subsurface stress directions at Lost Hills held," Geophysics, Vol. 56, No. 9, September 1991.

5. Plumb, R.A., and Hickman, S.H., "Stress-induced borehole elongation: A comparison between the four-arm dipmeter and the borehole televiewer in the Auburn geothermal well," Journal of Geophysical Research, Vol. 90, No. B7, 1985.

6. Teufel, L.W., "Strain relaxation method for predicting hydraulic fracture azimuth from oriented core," SPE/DOE Paper No. 9836, 1981.

7. Davis, P.M., "Surface deformation associated with a dipping hydrofracture," Journal of Geophysical Research, Vol. 88. No. 87, 1983.

8. Fix, J.E., Adir, R.G., Fisher, T.W., Mahrer, K.D., Mulcahey, C.C., Myers, B.C., Swanson, J. G., and Worpel, J. C., "Development of microseismic techniques to determine hydraulic fracture dimensions," GRI Report 89/0116, 1989.

9. Daneshy, A.A., "In-situ stress measurements during drilling," Journal of Petroleum Technology, Vol. 38, No. 9, 1986.

Copyright 1992 Oil & Gas Journal. All Rights Reserved.