Method of determining fracture closure pressures in hydraulicfracturing of subterranean formations

[0002] This invention relates generally to the hydraulic fracture stimulations of subterranean formations. In one aspect, it relates to determining closure pressure of hydraulically induced fractures in formations by pulse testing.

[0003] Hydraulic fracturing is a production stimulation technique involving the pumping of a hydraulic liquid down a wellbore and into a subterranean formation at such a pressure and rate to cause the formation to crack (fracture). In the vast majority of such treatments, the fracture is vertical, extending outwardly into the formation from the wellbore. During the latter stages of a fracturing treatment, a particulate propping agent (proppant) is generally deposited in the fracture. When the injection pressure is released, the formation walls close on the propping agent creating a “propped fracture” which provides a high conductivity channel in the subterranean formation. The conductivity of the propped fracture is the product of fracture width and fracture permeability. Permeability can be estimated by the size of the proppant. However, in order to generate sufficient fracture width, it is necessary to obtain a tip screenout (TSO) in the formation. Obtaining a TSO at the correct time is heavily dependent upon an accurate estimate of the fluid leakoff coefficient (C_L)—the rate at which fluid leaks off from the fracture to the surrounding rock. It is known that an accurate measurement of C_Lis based on an accurate determination of fracture closure time (t_c), which, in turn, is based on fracture closure pressure (P_c).

[0004] For a given volume of fluid pumped into a fracture (V₁), the fracture will require a specific amount of time to close (t_c), depending on how quickly the fluid filling the fracture leaks off to the surrounding formation. The fracture closes only after all the fluid filling the fracture leaks off. As the fluid leaks off, the gradual closure of the fracture is accompanied by a gradual decline of the pressure inside the fracture. The time required for the fracture to completely close (t_c) coincides with the fracture closure pressure (P_c). The fluid volume (V₁) and t_care used in subsequent computer simulations for designing the main propped fracture stimulation treatment. Only by accurately determining P_c, can t_cbe determined, which, in turn, is used to calculate C_L.

[0005] Thus, an accurate determination of P_cis a key to the design of a fracture treatment. The most common techniques for the onsite determination of P_care pressure decline analysis, constant-rate flowback testing, and pulse testing.

[0006] Pressure decline analysis (briefly alluded to above) involves creating a fracture using a known volume of fluid (V₁) pumped at a constant rate. After pumping is complete and the pumps are shut down, the pressure in the fracture will decline as the fluid in the fracture leaks off to the surrounding formation. In many instances, plotting the declining pressure versus the square-root (SQRT) of the elapsed time since pump shut-down (dt) results in a curve with 2 linear sections of different slopes. The intersection of the 2 linear sections is the point of fracture closure and, thus, defines the values of P_cand t_c(see FIG. 3 for an example). In some cases the pressure vs. SQRT (dt) plot (i.e., the pressure decline plot) does not yield a clear slope change for determining P_cand t_c. In these instances, other methods must be used to determine PC, which can then be used with the original pressure decline plot to determine t_c. The most common of these alternate methods are constant-rate flowback testing and pulse testing.

[0007] Constant-rate flowback testing involves creating a fracture followed by flowing fluid back from the fracture at a constant rate. This method results in a relatively slow drop in pressure while the fracture is open followed by a more rapid drop in pressure once the fracture closes. This test works well if the flowback rate is held constant during the test. Maintaining a constant flowback rate, however, is sometimes very difficult when applying this method.

[0008] A recent SPE publication, SPE Production and Facilities(August 1996), by C. A. Wright, et al., describes the use of fluid pulse testing for determining P_c. Wright's concept of pulse testing involves creating a fracture followed by pumping small fluid pulses intermittently as the fracture gradually closes. The pressure response from each fluid pulse is analyzed to determine if the fracture is open or closed at the time the pulse was pumped. The pressure response of an open fracture is different than that of a closed fracture. This method is robust and can be easily utilized in most situations. This method, however, determines only a range of possible P_c's and not a specific P_c.

[0009] There is a need for a robust, on-site technique that is similar to the pulse testing technique described above, but can actually determine a specific, singular value of P_crather than only range of possible P_c's. Being able to determine a specific value of P_cwith such a test would increase the accuracy of determining C_L(a critical variable in the design of fracture stimulation treatments).

[0010] In its broadest aspect, the method of the present invention involves four main steps:

[0011] (a) pumping a large volume of fluid (V₁), relative to step (c), down a wellbore and into a subterranean formation to form a fracture therein for data;

[0012] (b) permitting the fracture to close;

[0013] (c) pumping a small volume of fluid (V₂), relative to step (a), into the formation to reopen the fracture more narrowly than in step (a); and

[0014] (d) shutting down the pumping operation and determining the pressure in the wellbore at shutdown.

[0015] As described below, the pressure determined in step (d) is the initial shutin pressure (ISIP) and is a very close approximation of P_c. The P_cis used to determine the t_cin the pressure decline analysis technique (if t_ccannot be clearly determined from the pressure decline plot itself). The t_cis used, in turn, to determine C_L. The value of C_Lis then used to design the fracture treatment, along with other essential data, using known computer simulation techniques.

[0016] In a preferred embodiment, the process may include additional steps of (a) using an initial water breakdown to insure open perforations and that adequate injection rates can be obtained, and (b) proppant scouring prior to fracture generation with a low density proppant slug to scour tortuous fracture paths and help plug multiple fractures.

[0017] The variables involved in carrying out the four steps may range within wide limits depending on the factors including formation thickness, formation tensile strength, formation toughness, formation pressure, and pumping equipment, etc.

[0018] However, the following parameters are important for the success of the present invention: (1) the data fracturing step must generate a fracture of larger dimensions than the small volume pulse testing and (2) the pulse testing step should be at low volume (e.g. 0.5 to 3 bbls per pulse) and low injection rates (e.g. 2 to 5 bpm).

[0019] Other important factors, discussed in detail below, are minimizing near wellbore tortuosity, use of low viscosity fluids as pulse fluids, and frequent data recording.

[0020] FIG. 1 is a plot of bottom hole pressures and pump rates vs. time recorded in a field test according to the present invention.

[0021] FIG. 2 is a zoomed-in portion of FIG. 1, illustrating pressure at frequent time intervals.

[0022] FIG. 3 is a pressure decline curve useful in correlating closure pressure and (P_c) fracture closure time (t_c).

[0023] Abbreviations shown on the drawings and used herein are defined in the Nomenclature section at the end of Description of the Preferred Embodiments.

[0024] The method of the present invention is a testing method for determining fracture closure pressures P_c, the value of which can be used in the design of a fracturing treatment of subterranean formations. The method can be used in any formation, but is particularly applicable in soft formations (i.e., those having a rock plain-strain modulus, E′, of less than 800,000 psi).

[0025] As indicated above, the method involves four essential steps: generating a fracture, permitting the fracture to close, pulse testing, and determining the ISIP for each pulse test. For optimum results, however, the method preferably includes the following five sequential steps:

[0026] (1) water breakdown,

[0027] (2) proppant scour step,

[0028] (3) generating a fracture-for-data,

[0029] (4) permitting the fracture to close,

[0030] (5) post-closure pulse testing, including determination of ISIP.

[0031] Each of these steps are described below in detail.

[0032] As is customary in the hydraulic fracturing of subterranean formations, the initial step is to pump a small volume of water into the formation to break down the perforations and ensure adequate injection rates can be achieved. The water breakdown step generally, but not always, initiates the fracture.

[0033] After a small volume of water or gel is pumped into the formation, a low-density slurry of particulate material is injected into the formation. The purpose of the scour step is to minimize both the near wellbore fracture tortuosity and multiple fractures. The slugs of slurry scour out tortuous paths and plug multiple fractures. At this point, preliminary pulse testing may be carried out with the fracture open to obtain a rough estimate of P_cbefore generating a fracture-for-data (the next step).

[0034] The injection may be in slugs of several barrels each. The slurry may be water or an aqueous solution of a polymer. Any of the polymers currently used as viscosity fracturing fluids may be used.

[0035] The particulate material may be finely divided fluid loss additive such as silica sand, but preferably is a propping agent. The proppant density in the liquid is low, between about 0.2 to 4 lbs/gal of liquid, preferably between 0.5 to 3, most preferably between 1 to 2 pounds per gallon of liquid. While injection volume rates and pressures will depend on the condition of a particular application, the example described later on exemplifies a typical treatment.

[0036] A volume (V₁) of fracturing fluid is pumped into the formation to generate a fracture therein (if the previous steps were not performed, or did not initiate the fracture) or propagate the fracture if such previous steps were performed. This fracture is created to obtain data (i.e. P_cand t_c) and not to stimulate formation production or injection. As indicated earlier, the data will be used to determine C_Lfor use in designing the fracture treatment.

[0037] The fracturing fluid may be any of the viscosified aqueous and oil-based fluids, but preferably is an aqueous linear polymer gel (e.g. guar). The viscosity, other additives, pumping rates and volume may be in accordance with fracturing techniques well-known in the art; all that is necessary is that the fracture have dimensions larger than that possible with the volume used in pulse testing (V₂). Preferably, at least 50 barrels are used to generate the fracture for data. The upper limit of the volume can be that used to generate the fully propagated fracture (e.g. 1000 bbls). The preferred volume for generating the fracture for data is from 50 to 300 barrels; and the most preferred volume is 100 to 200 bbls.

[0038] This is a key step in the method of the present invention because the pulse testing is carried out after the fracture closes. Once the fracture has been generated, the pumps are shut down. The fracture closes by fluid leakoff into the formation. Closure is indicated when the pressure in the wellbore is below P_cdetermined at the conclusion of the fracture step (but possibly unknown at this point, if previous steps for determining P_cwere inconclusive).

[0039] Normally, fracture closure will occur from 5 to 60 minutes after pump shutdown.

[0040] Once the fracture is closed, the pulse testing may be carried out. The pulse testing fluid is water or a low viscosity aqueous polymeric solution. Linear polymer gels such as guar have been used with success. Other gelling polymers well-known to those skilled in the art may be used. The concentration of the polymer in the water may range within wide limits, but concentrations from 10 to 100 pounds per 1,000 gallons of water are preferred. For example, linear guar gel at 70 pounds per 1,000 gallons of water have given good results. The viscosity of the pulse fluid is preferably less than 70 centipoise.

[0041] The volume (V₂) of each pulse should be small in relation to the volume used to form the fracture (V₁). Each pulse travels through the preexisting fracture and does not extend or propagate the fracture. The relatively low volume pulse travels through the preexisting fracture unimpeded by tip effects such as fracture toughness and rock tensile strength. This permits the pulse to disperse more rapidly down the fracture length. This results in a minimum fracture width (w_f) which, in combination with a low plain-strain modulus (E′) of the rock, results in a negligible fracture net pressure at the wellbore (P_net). As well understood by those skilled in the art, when P_netis small, ISIP becomes a very close approximation of P_c.

[0042] The volume of each pulse should be in the range of 0.5 to 5 bbls, preferably 0.5 to 3 bbls, and most preferably between 1 and 2 bbls. The 1 to 2 bbl range appears to be optimum because it is large enough to ensure fracture reopening and small enough to minimize fracture width.

[0043] The pumping rate for each pulse preferably should be from 1 to 10 bpm, and most preferably from 2 to 5 bpm. Typical pumps used in fracturing operations can achieve a minimum pump rate of about 5 bpm.

[0044] The low rate, low volume pulse dictates the length of pulse pumping time, which typically is from 15 to 60 seconds.

[0045] After the pulse volume is pumped to reopen the fracture, the pump is shut down and the pressure (ISIP) at the wellbore opposite the fracture is determined. The ISIP is equal to, or very close to, P_c.

[0046] The equipment for measuring the pressure at pump shutdown should be accurate (within 3 psi) and should record the pressure data at intervals not longer than 2 second intervals, preferably not longer than 1 second intervals.

[0047] Care must be exercised in identifying the ISIP from the pressure/time chart because of pressure fluctuations at shutdown due to water hammer. The actual ISIP on the pressure chart is the average of the first full cycle of pressure fluctuation at pump shutdown. (This will be demonstrated in the discussion relating to the field experiment.) The ISIP represents a close approximation of the fracture closure pressure (P_c).

[0048] The best mode of the method according to the present invention involves the five steps described above. It, again, is emphasized, however, that the process in some applications may involve only three steps: (1) generating a fracture-for-data, (2) fracture closure, and (3) postclosure pulse testing (which includes determining ISIP). For example, if fracture tortuosity is not a problem, the proppant scour step may be omitted. Also, the water breakdown step may not be needed if injection capabilities for a particular formation are known. However, good field practice suggests that all five steps be used to determine P_c.

[0049] The following presents an example of the best mode for carrying out the process.

[0050] The bottom hole pressure data can be obtained with (a) a downhole transducer or (b) recorded at the surface and converted to downhole pressure.

[0051] During the pulse testing, the pressure is recorded every second and the value of P_cdetermined at pump shut-in for each pulse. The pressure decline curve may be plotted as pressure vs. square root of time (dt) and is used to determine t_c. The field example will demonstrate the complete process and determination of t_ccorresponding to P_c.

[0052] The method of the present invention was performed in an Alaskan North Slope well having a formation with a rock plain-strain modulus (E′) of 300,000 psi. Following water breakdown step, and proppant scour steps, 100 barrels of a linear water gel (guar) was pumped into the formation at a rate of 15 barrels per minute to generate a fracture. The pumps were shut down permitting the pressure in the fracture to bleed off, closing the fracture. Pulse testing was performed using a pulse volume of 1 to 2 barrels pumped at a rate of 5 bpm.

[0053] FIG. 1 is a plot of bottom hole pressure and pump rate vs. time for the field test. FIG. 2 is a zoomed in view of FIG. 1 focusing on the pulse test. The data in both figures was taken at one-second intervals.

[0054] Returning to FIG. 1, the fracture was generated at a pressure of about 2800 psi (as indicated by reference numeral 10), and an injection rate of 15 bpm (as indicated by reference numeral 11).

[0055] The pumps were shut down at time of about 71 minutes (reference numeral 12) and the pressure declined during the time interval indicated by reference numeral 13. When the pressure had declined sufficiently to ensure fracture closure, the postclosure pulse test was performed. One to two barrels of the linear gel were pumped into the formation at 5 bpm, as indicated by reference numeral 14. Returning to FIG. 2, the pumps were shut down at a time of 108.1 minutes (reference numeral 15) and the bhp pressure immediately dropped. During the first few seconds after pump shutdown, the pressure fluctuated between about 2590 and 2640 because of water hammer's effect. The average of first full pulse cycle of the pressure fluctuation was 2623 psi indicated by reference numeral 16. This average reading is the ISIP and represents a close approximation of P_c. The correlation of ISIP and P_cwas confirmed by the plot of FIG. 3. This figure is a plot of the square root of square root of time (SQRT) vs. bottom hole pressure based on portion 13 of FIG. 1.

[0056] As is well-known in the art, the intercept of line 17 and line 18 as at 19 on the pressure decline curve (FIG. 3) is P_c. (Line 17 is the pressure decline while the fracture is still open, and line 18 is the pressure decline after the fracture closes.) The intercept 19 occurs at about 2628 psi, which correlates well with ISIP of 2623 on FIG. 2. The dt value at a P_cof 2623 on FIG. 3 is about 1.8, giving a t_cof about 3.25 min. This value of t_ccan be used in computer simulations to determine C_L, which, in turn, is used to design the fracture treatment for the formation tested.

[0057] In many field operations, the pressure decline curve does not provide a clear indication of fracture closure. For example, the straight-line portions 17 and/or 18 of FIG. 3 may not be discernible, in which case the value of P_ccan not be determined. However, by using the P_cdetermined by the plot of FIG. 2, the value of t_ccan be determined by correlating the value to a dt on the pressure decline curve (FIG. 3).

[0058] In actual field practice, it may be desirable to perform multiple pulses to obtain corroborating data concerning the ISIP. Actual field tests carried out in a North Slope well have demonstrated the repeatability and reliability of multiple pulses in the well following water breakdown and scouring steps.

[0059] In summary, actual field tests have demonstrated that the method of the present invention is a simple and reliable on-site technique for accurately determining P_cand t_c.

[0060] The abbreviations and symbols used herein are based on standards used in the petroleum industry.

[0061] C_L=Fluid leakoff coefficient, ft/min^0.5

[0062] dt=Elapsed time since pump shut-down

[0063] E′=Rock plain-strain modulus, psi

[0064] P_c=Fracture closure pressure, psi

[0065] P_net=Fracture net-pressure at wellbore, psi

[0066] t_c=Fracture closure time after pump shutdown, minutes

[0067] ISIP=Instant shutin pressure (psi)

[0068] SQRT=Square root of time

[0069] V₁=Volume of fluid used to generate a fracture for data

[0070] V₂=Volume of fluid used in each pulse

[0071] ppa=Pounds of proppant per gallon of fluid

[0072] bpm=Barrels per minute

[0073] bbls=Barrels

[0074] psi=Pounds per square inch gage

[0075] w_f=Fracture width