Propane-based fracturing improves well performance in Canadian tight reservoirs

DON LeBLANC, Eastex Petroleum Consultants; LARRY HUSKINS, Corridor Resources; and ROBERT LESTZ, GasFrac Energy Services

Applied onshore New Brunswick, a frac fluid based on liquid petroleum gas resulted in longer fracture half-lengths and higher production rates compared with water-based fluid.

Equipment for a propane-based frac job at McCully field includes a well-test package with a sand trap, two line heaters, separation equipment and flaring for the initial propane flowback.

The McCully tight gas field in New Brunswick, Canada, is being developed by Corridor Resources and began producing gas from two wells in 2003. Full development occurred in 2007 with the startup of the McCully gas plant and tie-in to the Maritimes and Northeast pipeline. The unconventional nature of the reservoir requires the wells to be fracture stimulated to enable economic gas production. Prior to 2009, the preferred technique was to use water as the fracturing fluid. That year, fracture stimulations using gelled propane were performed in an attempt to improve fracture characteristics and flowback.

Because propane is 100% compatible with the formation, when using gelled propane as the frac fluid, multiple frac treatments can be completed without the need for flowback and cleanup between frac stages. Extended shut-ins can simplify and possibly accelerate cleanup without detriment to the formation because of propane’s ability to either mix completely with natural gas in the formation and its 100% solubility (miscibility) into crude oil within the formation, which also results in a lower oil viscosity. At McCully field, not only has the use of gelled propane resulted in savings due to the elimination of flowback, cleanup and water handling processes, but it has also significantly enhanced initial well performance compared with wells fractured using gelled water.

FIELD GEOLOGY

McCully field is comprised of a number of sand packages in the Hiram Brook (HB) formation containing about 500 Bcf of gas in place. The sands overlie the Frederick Brook (FB) shale, which contains about 67 Tcf of free gas in place.

The tight (low-permeability) HB sands are located in a highly stressed environment. Porosities range from 4% to 8% with water saturations ranging from 10% to 30%. The HB sands occur at about 1,800 m (5,900 ft) and have gross thickness up to 900 m (3,000 ft). Net pay per well can be as high as 95 m (310 ft). Permeability has been calculated to be as low as 0.001 mD and as high as 1.8 mD. Reservoir pressures range from 2,900 psi to 5,100 psi. Reservoir temperature is anomalously low: 40°C at 2,250 m (7,380 ft).

Rotary sidewall cores have been taken from most of the McCully wells. Comprehensive logging suites, good formation water samples, extensive core analyses and drilling and completion experiences have led to the conclusion that the HB sands are undersaturated to below irreducible water saturations (S_w). In many cases, S_w has been calculated to be less than 10%. The sub-irreducible water saturations found in the HB sandstones have led to concerns that phase trapping of the fracture fluids may be occurring. Laboratory regain-permeability tests were performed on McCully core, and permeability reduction due to water imbibition was measured at about 50%.

The FB shale formation was found to be productive in late 2007 in the McCully F-58 well, which was placed on production in early 2008. A number of FB shale evaluation wells have been drilled and completed to evaluate the formation’s development potential.
 
STIMULATION PROGRAMS

A typical McCully well has one to seven fracture stimulation treatments along a 400-m (1,300 ft) gross vertical column. Due to the large overall gross thickness of the HB sands, vertical “S”-profile wells have been used to develop most of the field. High-pressure and flow-through composite bridge plugs have been used to isolate each fracturing stage in the wellbores.

From the beginning of 2005 to the end of 2008, 74 water-based hydraulic fractures were placed in 26 wells, using from 9 tonnes to 100 tonnes of proppant. Job sizing was based on extensive 3D fracture modeling using a conservative fracture half-length of 100 m (330 ft) as an initial design basis. Economic optimization of the job size was a priority, and modeling suggested optimum job sizes ranging from 30 tonnes to 60 tonnes of total proppant per zone. In the actual water-based frac jobs, the optimum job size was found to be 40–50 tonnes; fracture sizes greater than 50 tonnes did not appear to increase the gas rate, Fig. 1a.

Frac water cleanup is an ongoing issue at McCully. Ultimate frac water recovery is low, which impacts fracture cleanup/flow effectiveness. As a result, in 2009, nine gelled propane (i.e., liquid petroleum gas, or LPG) -based hydraulic fractures were placed in four wells in an attempt to improve flowback performance and fracture characteristics. Propane fractures appear to have improved fracture cleanup and well productivity. In addition, the gas rate seems to trend upward with increasing job size, but this result must be taken with caution because of the small dataset, Fig. 1b.

Fig. 1. Flowback gas rate as a function of fracture stimulation size for a) water-based frac jobs during 2005–2008 and b) propane-based frac jobs during 2009.

One significant difference between the water and propane fracture treatments was in the time it took to recover the frac fluids. In all cases, 100% of the propane-based frac fluid was recovered within 20 days after start of the flowback. Water-based fluid flowback continued to be produced even after 1,000 days of production.

FRACTURE PERFORMANCE ANALYSIS

All wells had a number of extensive post-fracture flow and buildup tests performed, including pressure transient analysis. Radioactive tracers were included in most fracture treatments, and the wells were logged post-frac for placement analysis. These analyses helped with the understanding of the reservoir flow capacity, permeability thickness, hydraulic fracture effectiveness, fluid damage issues, completion efficiency and other parameters. On wells with multizone fractures, production logs were run to identify which zones were contributing gas.

To evaluate and compare frac flowback performance for the water and propane fracture stimulations, dimensionless type curves were developed. These type curves were based on dimensionless parameters of productivity index (J_D), time (t_D) and fracture conductivity (FCD).

A commercial simulator that includes templates for fractured well models was used to generate a number of forecasts for different reservoir properties, fracture conductivities (k_fW) and fracture half lengths (x_f). The simulation results were then used to generate type curves for J_D and t_D as a function of FCD. The dimensionless productivity index was used because of the length of the transient period for unconventional reservoirs and because J_D is independent of whether the well is exhibiting radial flow, is hydraulically fractured, or is vertical or horizontal. By using the dimensionless productivity index, it was possible to apply data from all of the wells without the  need for additional normalization.

The dimensionless equations used for the frac flowback analysis of McCully wells are the conventional equations for gas in terms of pseudo-pressure:

where PI is the productivity index, q is the gas flowrate (in MMcfd), ψR is the reservoir pseudo-pressure (in psi2/cP), ψ_wf is the flowing pseudo-pressure, k is reservoir permeability (in mD), h is net thickness (in ft), T is the temperature (in °R), Φ is porosity, µ_gi is gas viscosity at initial reservoir pressure (in cP), c_ti is total compressibility at initial reservoir pressure (in psi⁻¹) and r_w is the wellbore radius (in ft).

Using Eqs. 2 and 3 with the well test data and plotting the results on the type curves, it was found that the fracture conductivity (FCD) appeared to improve over time. The initial tests of the water-fractured wells, for both sand and shale intervals, appear to indicate that the fracture was ineffective (FCD ~ 0). After additional testing, FCD would reach a maximum, Fig. 2. This normally occurred after three or four test cycles and appeared to be independent of the amount of frac water recovered. In other words, for water-based fractures in McCully field, fracture cleanup, in terms of maximizing FCD, required three to four pressure cycles performed as flow/shut-in tests. The length of these flow/shut-in periods was not as important as the actual occurrence of the pressure cycles. This result was also observed for wells with multiple water fractures.

Fig. 2. Type-curve analyses of three wells fractured with water: a) McCully B-58 in the HB sand, b) McCully D-57 in the HB sand and c) McCully F-58 in the FB shale.

Type-curve analysis of the propane-fractured wells showed a dramatic difference from the water-based fractures, Fig. 3. The most significant observation was that all flow tests followed the maximum-FCD type curve. There were no fracture cleanup issues apparent with the propane fracs. The only deviation from the maximum-FCD curve occurred during the very early stages of shale flowback in the Green Road G-41 well, where data fell below the maximum-FCD curve but very rapidly increased toward it. In all cases, the very early stages of the initial test had measured gas specific gravities greater than 1.0, indicating high propane content in the flow stream. As the specific gravity decreased, the dimensionless data moved toward the maximum-FCD type curve.

Fig. 3. Type-curve analyses of three wells fractured fractured with propane: a) McCully L-38 in the HB sand, b) McCully P-47 in the HB sand and c) Green Road G-41 in the FB shale.

PROPANE FLOWBACK

With propane, the only required “overhead” for the cleanup period was during the initial 24 hours of flow, when the production from the well was 100% propane and it had to be flared using the limited flowback equipment available during this initial frac campaign. After that time, gas specific gravity declined from about 1.5 to about 1.0, and it was possible to direct the well flow to the McCully gas plant for processing and sale. In almost all cases, propane content was less than 10% within five days after the start of the flowback.

The initial flow period required a full well-test package including a sand trap, two line heaters, separation equipment and flaring. Line heaters were employed to provide a complete flash of the liquid propane, which enabled drop-out of all impurities, including frac sand that may have been flowed back. This also ensured that no liquid propane was sent to the flare or to onsite liquid storage tanks. For future applications, a propane recovery system could be employed to convert the propane back to liquid state, which would enable its reuse in future frac operations.

PROPANE FRAC ADVANTAGES

For propane-based fracs at McCully field, reservoir gas was observed at surface as soon as the wellbore had been cleaned out, whereas for water-based fracs gas was not observed until 5%–20% of the frac fluid had been recovered. Since propane is non-damaging and vaporizes as it mixes with methane, there is no requirement to flow back the well immediately following the fracture, as is preferred for any water-based fracture. In some cases, zones hydraulically fractured with propane remained plugged for a number of weeks before being flowed back. For these wells, startup gas gravity of the flow stream was always 1.0 or less, and no apparent fracture conductivity impairment resulted from the extended post-frac shut-in.

In terms of well productivity, fracturing improves the well’s initial rate and production index by maximizing xf and FCD. Fracture half-lengths for wells at McCully field—as determined from pressure transient analysis, rate-time analysis and analytical modeling—were significantly greater in the propane-fraced wells than in the water-fraced wells, Fig. 4. In wells with single fractures, xf for the propane fracture was twice the average xf for water fractures, even though the job sizes (amounts of proppant pumped) were equivalent.

Fig. 4. Fracture half-length for selected propane-fraced (red) vs. water-fraced wells at McCully field.

An analytical model was built incorporating average McCully sand and fracture parameters to generate a zonal incremental production curve for a propane-based fracture vs. a water-based fracture in the field, Fig. 5. Comparing the curves for propane vs. water fracturing, the production rates are observed to equalize after 10 years. This suggests that the incremental production of a propane-based fracture over a water-based fracture includes a recovery acceleration component.

Fig. 5. Extrapolated zonal incremental production rate curves for a propane-based frac, a water-based frac and a non-fractured zone in McCully field, along with cumulative production curves.

It should be noted that these production curves are based only on the dataset from McCully field. Additional verification would need to be performed in other fields before assuming that comparable results for propane-based vs. water-based fracturing would be obtained generally. 

ACKNOWLEDGEMENTS
This article was prepared from SPE 144093 presented at the SPE North American Unconventional Gas Conference and Exhibition held in The Woodlands, Texas, June 14–16, 2011.