SPE 84841 Fracture Height Development and Control in the Mississippian Sandstones of Central West Virginia

David J. Mack, BJ Services Company, Joseph Awny, Equitable Production Company and Daniel Kendrick, BJ Services Company. SPE Members
Copyright 2003, Society of Petroleum Engineers Inc. This paper was prepared for presentation at the SPE Eastern Regional/AAPG Eastern Section Joint Meeting held in Pittsburgh, Pennsylvania, U.S.A., 610 September 2003. This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to a proposal of not more than 300 words; illustrations may not be copied. The proposal must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836 U.S.A., fax 01-972-952-9435.

performed to verify the design simulation. Treatment design history and current philosophy will be discussed in this paper. In addition, the results of the data gathering, treatment design and execution and RA tracer survey will be presented. Introduction The Dickenson Field is located in Nicholas County, West Virginia, fig. 1, and consists of 61 wells, which have produced 37 BCFG, since the field discovery in 1944. Initial development targeted the Mississippian Big Lime limestone and Big Injun sandstone. It was not until the early 1980s that formations below the Big Injun were penetrated and the Mississippian Weir sandstone was added as a target zone. The Big Lime is a limestone member of the Greenbrier Group, fig. 2, and unconformably overlies a clastic sequence called the Pocono or Price Formation1. The Big Lime in the Dickenson field is a stratigraphic trap where porous oolites laterally grade into nonporous, nonoolitic limestone. Depths to the top of the formation range from 1,400 to 2,200, with the gross interval being greater than 200 feet with net pay thickness in the range of 10 to 35 feet. The Big Injun is a member of the Pocono or Price Formation; the term was often applied to any productive interval at the base of Big Lime formation. The preGreenbrier unconformity truncates the Lower Mississippian rocks across the field, which causes the variability of the stratigraphic terminology. The Big Injun in the Dickenson field is a medium to fine-grained sandstone representative of a distributary channel. Depths to the top of the formation range from 1,600 to 2,400, with net pay thickness in the range of 15 to 35 feet. The Weir is a member of the Pocono or Price Formation. In the Dickenson field it is a medium to fine-grained sandstone representative of tidal environments. Depths

Abstract Natural gas production in central West Virginia is primarily from the shales of Devonian age upward to the sandstones of Pennsylvanian age. These productive intervals are typically low permeability, under pressured and adjacent pay zones can have similar closure pressures with only a minimally competent barrier separating them. Such a scenario exists in the Weir and Injun sandstones of Nicholas, Fayette and Kanawha Counties, West Virginia. At a depth of 2000 to 2500 ft, both the underlying Weir and overlying Injun exhibit only a few hundred psi reservoir pressure, frac gradients in the 0.4 to 0.6 psi/ft range and are separated by 10 to 20 ft of sandy shale. To make either of these zones commercially productive, both must be hydraulically fractured. Initially it was felt that historical treatment data and the comparison of conventional electric logs to offset wells would provide sufficient data to effectively fracture model stimulation treatments. This technique provided interesting computer simulations, but treatment and production results did not substantiate the modeling. In an effort to better model and stimulate both zones, mechanical properties were obtained from a sonic log. Treatment parameters such as pressure, fluid efficiency/formation permeability and near wellbore connectivity characteristics were obtained from stepdown and mini-frac tests. Real-time job monitoring, post-frac RA tracer survey and production testing were

SPE 84841

to the top of the formation range from 1,650 to 2,500, with net pay thickness in the range of 20 to 40 feet. The initial well completions primarily consisted of drilling the wells to the base of the Big Injun and setting a nitroglycerine charge in open hole as means of stimulation. The Big Lime was acidized with Hydrochloric acid, the formations were commingled and produced. With advances in stimulation methods the Big Injun and the Weir were stimulated with 75Q foam and 20/40-mesh sand. The completion method involved a three-stage completion of the Weir, Big Injun and Big Lime using the Ball and Baffle technique to isolate the zones. The Weir being completed first followed by the Big Injun and then the Big Lime. A detailed field study revealed that wells which were completed in the Weir yielded recoverable reserves approximately 1/3 less than the wells in which only the Big Lime and Big Injun were completed. Pressure depletion accounted for a portion of the decrease in recoverable reserves, however it did not account for all of the decrease. It was theorized that with the presence of the unconformity, the Weir interval is extremely incompetent to the extent that the once fractured, the fracture height grows uncontrollably out of zone. Then when the Big Injun is fractured, the fracture follows the pre-existing Weir fracture thus rendering the stimulation ineffective. Historical Simulation Strategies Typical treatments for the Weir and Big Injun in southern West Virginia had been very conservative in design. Linear gel, cross-link, or foam fluids were the typical carrier fluids. Standard rate would be on the order of 20 BPM down hole. Pads would be on the order of 20% 30% of the total fluid volume. Proppant loads were typically in the 25,000 lb of 20/40 sand range. A few of the wells had as much as 30,000 lb 20/40, but these were the definite exceptions. Treatments were normally of identical size with little attention to the gross size of the interval or potential reserves based on log porosity. Foam seems to have been the only "design feature" that was implemented consistently. The low reservoir pressure of the Weir/Big Injun seems to have convinced operators that in order to recover the maximum amount of the frac fluid, nitrogen foam had to be used. One must bear in mind that with the depletion that has been prevalent in the study area, the Weir/Big Injun will not, for the most part, be a significant contributor to a commingled production scenario. The Weir/Injun has, in some cases, acted as a "thief zone" when attempting to commingle it with higher-pressure reservoirs.

Stimulation Technique Evolution The first step to test our uncontrolled height growth theory was to perform a basic 3-D hydraulic fracture simulation using an industry accepted model2-4. Since at this point in the study measured reservoir properties were not available, field experience correlations based on gamma ray signature, photo electric measurements and 3-D fracture simulator lithology default values were used. A Nicholas County well was selected as the first study candidate. This wellbore exhibited the typical Weir, Injun and Big Lime productive intervals with additional up hole Little Lime and Blue Monday potential. Table 1 shows the stratigraphic configuration. Conventional treatment design for the Weir interval suggested that a nitrogen foam frac pumped at 20 BPM carrying 35,000 lb 20-40 Ottawa sand be performed. Flush of the treatment is typically comprised of hydrochloric acid for breakdown of the next zone to be fraced and treated water. Fig. 3 is the 3-D simulation for this treatment. Perforations are from 2460 to 2470 in the gross Weir interval of 2450 to 2479. There were two results that were of particular interest. The first was the complete removal of proppant from the near wellbore area, as depicted by the darkest coloration. Although the treatment was designed for a 1 bbl under flush, severe over flushing occurred. This overflushing scenario was a result of the hydrostatic head of the flush fluid being greater than the fracturing pressure of the Weir. One solution would be to flush the treatment with a fluid lighter than the frac gradient. Another would be to use a frac baffle and ball to prevent the well from over flushing. The second technique was chosen. Fig 4 illustrates the effect of over flush avoidance. It is also observed that there is appreciable fracture height growth upward. This growth takes the fracture into the Big Injun as theorized. Although proppant is placed in the Big Injun, the concentration is very low, less than 0.5 lb/sq ft, and would thus provide little conductivity for gas production. Since the Big Injun was going to be fraced anyway, we investigated the result of perforating both intervals and fracing them simultaneously. Fig 5 shows this simulation with the typical treatment design using the over flush avoidance techniques discussed above. Proppant placement is better, however, the majority of the proppant is in the Weir interval. To improve the proppant profile we looked at several scenarios using increased proppant volumes at reduced injection rates. Since we had seen improved proppant placement with just the Big Injun interval perforated, we concentrated our study around that perforation scheme. Our desire was to place a higher concentration in both

SPE 84841

the Big Injun and Weir intervals without significantly increasing fracture height growth. We settled on a treatment consisting of 54,000 lb 20-40 mesh white sand, carried in a 75 Quality nitrogen foam and pumped down the 4 1/2" casing at an injection rate of 12 BPM under down hole conditions. Maximum designed sand concentration in the foam carrier fluid was 5 lb/gal. Fig 6 is the 3-D simulation profile for this treatment. Note that both intervals have a concentration in excess of 1.0 lb/sq ft at a distance greater than 50 ft from the wellbore. Even at 100+ ft the concentration is greater than 0.75 lb/sq ft. There were also productive intervals in the up hole carbonate sections. It was decided to go ahead and execute a second stage acid treatment on these intervals. The 3-D simulation for this treatment did not show growth into the predicted Big Injun/Weir fracture. To test out the predictions of fracture growth for both of the treatments, different radioactive isotopes were added to each stage, iridium in the first stage foam frac and scandium in the second stage acid treatment. Both stages were pumped as planned with the exception that the maximum sand concentration in the Big Injun foam frac was limited to 4.5 ppg. This was due to slightly higher than anticipate surface treating pressures. The well was cleaned up from the treatments using common flowback procedures. Five days after the fracs, the spectral tracer-Gamma Ray survey was run. Fig. 7 is the Prism Analysis of the survey. Iridium is the darker shaded area with scandium being the lighter. Perforations are designated as a stippled area for the Big Injun and horizontal lines for the select fire carbonate intervals. Fracture growth within the carbonate intervals was reasonably similar to what was predicted by the 3-D simulator, i.e., staying within the perforated zone. As for the Big Injun, upward growth was not quite as much as anticipated, however, it is felt that there is good agreement between the predicted upward growth and that observed by the tracer survey. What really surprised us was to see the Big Injun foam frac stop in the 20 ft shale section between the Big Injun and Weir. To have such an abrupt truncation of the fracture growth as well as an exceptionally high concentration of radioactive material was definitely not anticipated. Analysis of the open-hole log shows that there is a rather thick, 10 ft, section of shale with a gamma reading greater than 200 API units. It is within this section that the Prism analysis indicates that downward growth stopped and radioactivity increased, meaning that a high percentage of the treatment was injected into this zone. Since this anomalous behavior occurred in a shale section, it is felt that the fracture actually traveled down a bedding plane within the shale section. This turning of the fracture to exhibit a horizontal component could also

contribute to the higher than anticipated surface treating pressure. Since the tracer survey indicated that the Weir had not been fracture stimulated, a tubing and packer frac was attempted. This treatment had to be aborted due to wellbore mechanical problems. The next step in the study was to select another well in the field and obtain mechanical properties with a long spaced sonic tool. Using this measured data a more accurate 3-D simulation could be attained. This second well was about 220 ft up dip from the first well. Both wells had similar gross developments of the Weir and Big Injun, however, the second well did not exhibit any production from the up hole carbonates. A second difference was noted in the shale barrier between the Weir and the Big Injun. In this well there was only 15 ft of gross shale separating the two target pay intervals. In addition, although present, there only existed a 3 ft interval where the gamma trace was above 200 API units. The general consensus was that this 3 ft section would not act as a treatment "thief" zone as was experienced in the first well. With this idea in mind, 3-D simulations with the measured mechanical properties were performed investigating the three perforation combinations, the Big Injun only, the Weir only and both zones together. Only small differences were observed between the different perforation schemes. Also, little variation in gross fracture height was predicted with respect to injection rate. We did observe improved proppant placement into the reservoir with increased injection rate, i.e., improved fluid efficiency. The treatment settled on consisted of 100,000 lb 20-40 mesh white sand, carried in a nitrogen foam and pumped at 40 BPM down the 4 1/2 in. casing. To improve upon the data used for the treatment design, a step down test and a mini-frac were performed the morning prior to the frac job. With the leakoff characteristics and pressures attained from these diagnostics, the pump schedule was fine-tuned. Fig. 8 is the 3-D simulation of this treatment. First it is important to look at the stress curve on the left-hand portion of the presentation. Note that at an approximate depth of 2235 ft, the stress is in the range of 750 psi. This is equivalent to a frac gradient of 0.336 psi/ft, which is appreciably below the hydrostatic gradient of water. This confirms the over flushing potential previously observed. Sand concentrations in excess of 0.5 lb/sq ft were attained over both pay intervals beyond 300 ft. The treatment was executed as designed. A single radioactive tracer, iridium, was added to the proppant stages.

SPE 84841

Five days after the fracturing treatment, the well had cleaned up sufficiently to run a gamma survey to determine fracture height growth. Fig. 9 is the gamma survey. As was observed in the first well, upward growth into the carbonate above was minimal, with the top at about 2200 ft, and was within 10 ft of the upward penetration predicted by the 3-D model. Downward fracture growth could not be determined accurately from this survey as the bottom was reached before the end of the "hot" zone could be read. The deepest reading was at 2280 ft with the radioactivity appearing to be weakening. The model predicted downward growth to 2305 ft. Although a definitive fracture bottom was not observed from the tracer survey, it appears that the model and the physical measurement were in good agreement. Conclusions 1. Creating individual fractures in the Weir and Big Injun intervals is not possible unless a relatively thick, highly organic shale section is present. 2. Upward growth into the carbonates is minimal. This is true regardless of fracture initiation from the Weir or the Big Injun. 3. Improved understanding of rock mechanical properties is key to optimizing the Big Injun and Weir Completions. 4. 3-D simulation should be incorporated in the stimulation design determine frac height and proppant pack concentrations for Big Injun and Weir stimulations. 5. Radioactive tracer logs and pressure testing should be used to better define post fracturing treatment geometry. 6. Due to the recent vintage of the completions, the relationship between the completion method and production performance cannot be fully quantified at this time. Further investigations will focus on defining this relationship.

Acknowledgements The authors would like to thank Equitable Production Company and BJ Services Company for permission to publish this manuscript. We would also like to thank Paul Gerome and Neal Alexandrowicz for their consultations and advice. References

1. 2. 3. 4.

"The Atlas of Major Appalachian Gas Plays," West Virginia Geologic Survey, Publication V25, 1996. Meyer, B.R.: Frac Model in 3D-1: New Simulator Makes Fracture Design Routine, Oil & Gas J. (June 17, 1985) 87. Meyer, B.R.: Frac Model in 3D-2: Proppant Placement Analyzed, Oil & Gas J. (July 1, 1985) 65. Meyer, B.R.: Frac Model in 3D-3: Hydraulic Fracturing Simulators Capabilities Examined, Oil & Gas J. (July 22, 1985) 65.

SPE 84841

Formation Salt Sand Ravencliff Little Lime Blue Monday Big Lime Big Injun Weir

Top 1773 2102 2156 2204 2414 2450

Bottom 1342 1790 2156 2192 2414 2430 2479

Table 1 Formation depths

Kanawaha

Nicholas

West Virginia
Fayette

Fig. 1 Study Area

SPE 84841

WEST VIRGINIA STRATIGRAPHY SYSTEM LITHOLOGY

Fig. 2 Stratographic Column

SPE 84841

Fig. 3 Typical Weir Design with Full Flush

Fig. 4 Typical Weir with No Over Flush

SPE 84841

Fig. 5 Big Injun at 20 BPM No Over Flush

Fig. 6 Treatment Through Big Injun Only

SPE 84841

Fig. 7 Prism Analysis

10

SPE 84841

Fig. 8 Second Well Weir Only

Fig. 9 Second Well Tracer Survey