Experimental investigation of the effect of imbibition on shale permeability during hydraulic fracturing

Journal of Natural Gas Science and Engineering 29 (2016) 413e430

Contents lists available at ScienceDirect

Journal of Natural Gas Science and Engineering journal homepage: www.elsevier.com/locate/jngse

Experimental investigation of the effect of imbibition on shale permeability during hydraulic fracturing Zhou Zhou a, b, *, Hazim Abass b, Xiaopeng Li b, Tadesse Teklu b a b

Petroleum Engineering Department, China University of Petroleum (Beijing), Beijing, 102249, China Petroleum Engineering Department, Colorado School of Mines, Golden, CO, 80401, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 October 2015 Received in revised form 14 January 2016 Accepted 17 January 2016 Available online 25 January 2016

Hydraulic fracturing technology is widely applied in shale reservoirs to significantly increase production. However, when many operators report a large percentage of the fracturing fluid is not recovered, it is unclear how the remaining fracturing fluid affects the shale formation. It is believed that the unrecovered fracturing fluid could be imbibed by shale matrix, micro-fractures, and surfaces of fractures that are already separated. This paper is to investigate the influence of imbibition on the matrix permeability, micro-fracture permeability, and fracture permeability. It is the first time to correlate permeability change with shale imbibition, and provide quantitative results of increase and decrease in permeability due to imbibition process in shale during hydraulic fracturing. This paper uses the pressure build-up method to measure permeability of the shale sample, and applies the under-weighing approach to do the imbibition experiment. The Niobrara, Horn River, and Woodford shale formations are the source of the samples in the experiment. The experimental results show that the imbibed fracturing fluid will damage and seriously reduce the matrix permeability of the shale sample. When the sample imbibes more fluid, the matrix permeability is reduced greater. Imbibition also decreases the fracture permeability of open fractures, but decrease is less than the reduction of matrix permeability. Moreover, there is a lubrication effect that can reopen micro-fractures on shale samples and stimulate and increase the micro-fracture permeability during imbibition. Permeability is a criterion that determines the long-term production from a formation. By studying the permeability change caused by imbibition during hydraulic fracturing stimulation, this paper presents a new observation that imbibition in shale can not only damage, but also potentially stimulate the shale formation by increasing permeability due to open closed or sealed natural fractures. © 2016 Elsevier B.V. All rights reserved.

Keywords: Shale formations Hydraulic fracturing Imbibition impact Permeability change

1. Introduction With the successful application of hydraulic fracturing technology in shale and other unconventional formations, crude oil production in the U.S. is expected to increase from 5 MMbbl/d in 2008 to 10.6 MMbbl/d in 2020; and oil production from shale and other low permeability reservoirs will grow to one half the national total crude oil production during the same period. From 2008, the U.S. shale gas production is expected to increase almost nine times (EIA, 2015).

* Corresponding author. Petroleum Engineering Department, China University of Petroleum (Beijing), Beijing, 102249, China. E-mail address: [email protected] (Z. Zhou). http://dx.doi.org/10.1016/j.jngse.2016.01.023 1875-5100/© 2016 Elsevier B.V. All rights reserved.

The general procedure for hydraulic fracturing stimulation treatments has five main steps, including pad injection, gel slurry injection, flush injection, well shut-in, and water recovery. Water recovery is the last step of the hydraulic fracturing treatment before the well is put on production. This step is important and necessary during hydraulic fracturing because it can control and minimize damage from fracturing fluids. However, many operators reported that less than 50% of the injected fracturing fluid in shale formations could be recovered (Alkouh and Wattenbarger, 2013). This may be because the system energy is low after hydraulic fracturing in shale formations. Generally, the system energy is higher when fractures are more conventional and less complicated. The higher energy can cause a larger volume of flow recovery at a faster rate. But fractures in shale formations are complex so that the percentage of fracture fluid recovery is small and takes several weeks to

414

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

complete flowback, much longer than in conventional formations (King, 2010; Wu et al., 2010). In shale, the impact on production of this large volume of remaining fluid becomes a concern. This is because many studies found that the remaining fracturing fluid could be imbibed by shale rock and fracture surfaces (Roychaudhuri et al., 2011; Makhanov et al., 2012; Yao et al., 2012; Zhou et al., 2014). Imbibition is the process by which one fluid is displaced by another immiscible fluid in porous media. This displacement is the main reason for serious clay damage in clay rich shale formations. In addition to clay damage, the imbibed water also causes water blockage in tight gas reservoirs from massive hydraulic fracturing treatments (Qin, 2007). Imbibition resulting from hydraulic fracturing can cause clay swelling in shale formations (Ghanbari et al., 2014). The swelling can occur in all clay minerals to various degrees; smectite and mixed-layer illite can expand up to 20 times their original volume (Hayatdavoudi, 1999). It is difficult to determine, however, whether clay swelling is harmful or helpful. Dutta et al. (2012) found that more fluid volume was imbibed in clay rich regions and the mobility of gas was expected to be reduced because of clay swelling. On the other hand, Morsy and Sheng (2014) had the opinion that clay swelling, due to imbibition, could create fractures along bedding in shale formations and thus expect to improve permeability and oil production. Water blockages occur when water and other liquids are trapped in the porous media and impede gas production (Charoenwongsa, 2011). The hysteresis and discontinuous capillary pressure cause injected liquid fluids to be extremely difficult to produce. In addition, after production, the invaded zone liquid saturation could decrease to the residual saturation so as to prevent liquid displacement. Hence, the gas permeability and gas production are greatly reduced because of the additional gas flow resistance from the trapped liquid (Hadley and Handy, 1956; Land, 1968; Ehrlich, 1970). Previous studies found that water blockage decreased permeability only temporarily. The permeability was recovered, as long as the drawdown pressure was high enough (Holditch, 1979; Abrams and Vinegar, 1985; Mahadevan and Sharma, 2003; and Bazin et al., 2009). On the contrary, other investigations indicated that water blockage created permanent permeability damage because in very tight formations, it was difficult for drawdown pressure to be high enough (Penny et al., 1983; Soliman and Hunt, 1985). In addition, some numerical models showed that when the rock matrix imbibed fluid from fractures, the relative permeability of gas in the invaded zone decreased. During production, the imbibed fluid was produced first. Then the gas began to flow through the invaded zone to the fractures with the rise of relative permeability of gas in the water blockage zone (Barati et al., 2009; Charoenwongsa, 2011; Putthaworapoom et al., 2012, Zhang et al., 2014). Therefore, the water blockage was temporary. In summary, previous studies show clay swelling is expected to either damage or stimulate formations; and water blockage is either permanent or temporary to damage formations. However, those previous investigations did not have the experimental data to answer the questions regarding damage or stimulation from the imbibition of fracturing fluid in shale formations. These questions are whether imbibition in shale formations results in damage or stimulation on long-term production. Also, if imbibition has a negative impact, is it permanent or temporary? In this paper, permeability is studied as the criterion of impact. Through experiments, this paper investigates permeability changes because of imbibition under various fracturing fluids in shale. It is the first time to quantify permeability changes as a function of shale imbibition. In addition, the results in this paper can explain how slick-water

fracturing increases production in shale formations. 2. Experiment The determination of permeability changes as a function of fluid imbibition is the main objective of the experiment in this paper. Therefore, the experiment consists of two parts: permeability measurement and the fluid imbibition experiment. The details of the permeability measurement and imbibition experiment will be discussed in the following parts. For each sample experimental run, the original sample permeability was first determined through the permeability measurement. Then, the sample was immersed into a test fluid to begin the imbibition experiment. The sample permeability was measured again after one or two days. After that, the sample was put back into the test fluid to continue the imbibition experiment for another day. The permeability measurement was repeated, and the imbibition experiment was again followed. The duration of these repeated experiments was usually one week and sometimes up to one month. Finally, varying permeabilities were recorded as a function of time for the imbibition process. 2.1. Permeability measurement The pressure build-up method was used for the shale sample permeability measurement. This is one of the most efficient methods to measure permeability of tight rocks. 2.1.1. Measurement principle The principle of the pressure build-up method is that the constant inlet pressure of a confined shale sample is higher than the outlet pressure of the sample. The test records and analyzes the rate of outlet pressure increase as fluid is pumped through the sample. Nitrogen is used as the pressure build-up test fluid. Permeability of the shale sample is derived through the following equations. Gas densities at the standard condition and at the test condition are calculated in Eq. (1) and Eq. (2).

rgs ¼

Pgs M RTgs Zgs

(1)

rgt ¼

Pgt M RTgt Zgt

(2)

rgs, rgt are gas densities at the standard condition and at the test condition, respectively; Pgs, Pgt are pressures at the standard condition and at the test condition, respectively; Tgs, Tgt are temperature at the standard condition and at the test condition, respectively; Zgs, Zgt are compressibility factors at the standard condition and at the test condition, respectively; M is gas molar mass; and R is gas constant. Hence, gas density at the test condition is represented in Eq. (3). rgt ¼

Tgs rgs Zgs Pgt Pgs Tgt Zgt

(3)

where: Zgs ¼ 1 A one dimensional gas continuum equation is Eq. (4).



  v rgt vx v 4rgt ¼ vx vt

(4)

vx is velocity in x direction; 4 is porosity; and t is time. When the gas density equation is substituted into the gas continuum equation, Eq. (5) is derived.

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

Pgt vPgt v vx mg Zgt vx

!

2.2. Imbibition experiment

4CðPÞ Pgt vPgt ¼ k Zgt vt

(5)

mg is gas viscosity; k is permeability; C(P) is gas compressibility, C(P) ¼ 1/P1/ZdZ/dP. P0 is introduced and is defined as following. ZPgt2

0

P ¼2 Pgt1

p01  P20 ¼

415

Pgt dPgt mg Zgt

(6)

 1  2 2 Pgt1  Pgt2 mg Zgt

(7)

Imbibition experiments in this paper were through the underweighing approach (Fig. 2). In this method, a sample was suspended under a balance that could automatically record weight change of the sample as a function of time. Weight was changed because there was a displacement of fluid between the outside and inside of rocks. The displacement was caused by imbibition. Two test fluids were used including 7% KCl and 0.07% friction reducer. Both of these fluids are common in field hydraulic fracturing treatments. The recorded measured weight from the balance was transformed to imbibed liquid saturation to compare the results from various rock samples. The transformation is based on fluid density, sample volume, and porosity (see Eq. (10)).

Therefore, the gas continuum equation can be changed to the following expression (Eq. (8)).

Imbibed Liquid Saturation ¼

  v vP 0 1 vP 0 ¼ vx vx K vt

DW is weight change of the sample; rl is liquid density; V is sample volume; 4 is porosity.

(8)

where: K ¼ 4C(P)mg/k This gas continuum equation is similar to the equation in Oort's paper (Oort, 1994). Oort's equation is used to characterize liquid permeability in shale samples. The only difference between the two continuum equations is that pressure in Oort's equation is changed to P’. Therefore, the gas continuum equation has a similar solution through Oort's development. The solution is below.

! k¼

¼

mg CðPÞVd L A mg CðPÞVd L A

Dln

0 0 Pinlet Pinitial 0 Pinlet P 0 ðtÞoutlet

Dt Dln

!

2 2 Pinlet Pinitial

2 Pinlet PðtÞ2outlet

Dt

(9)

Vd is downstream reservoir volume; L is sample length; A is sample cross sectional area; Pinlet is inlet pressure; Pinitial is initial pore pressure; and Poutlet is outlet pressure. 2.1.2. Measurement apparatus Permeability of the shale samples was measured through the Model 6100 Formation Response Tester (FRT) that is designed and applied for permeability measurement of a formation sample (Fig. 1). Maximum pumping pressure and confining pressure are 5500 psi and 6000 psi, respectively. Core holder diameter is 1 inch, and length is up to 6 inch. 2.1.3. Measurement procedure The pressure build-up procedure had four steps. The first step was to open all valves in order to fill the whole system, including the upstream and downstream reservoirs, with nitrogen at a desired pressure. Thus, the inlet pressure of the sample was equal to the outlet pressure. The second step was to close valves except the valve for the gas injector so that the downstream reservoir was isolated. Thus, the inlet pressure that was raised in the next step was the only source that changed the outlet pressure. The third step was to quickly increase the inlet pressure and then make it constant. The last step was to record the increased rate of the outlet pressure. According to the outlet pressure transient, the permeability of the sample could be derived using Eq. (9).

DW rl V4

(10)

3. Shale samples Shale samples are obtained from the Horn River, Woodford, and Niobrara shale formations. Shale samples from the Horn River and Woodford formations are provided through the courtesy of the companies that have operations in those formations. The Niobrara shale samples were outcrop and obtained from a quarry near Lyons, Colorado, United States. The original shale samples from the Horn River formation were cored with potassium silicate mud system. There were seven slabs. Two were in the Muskwa Member; two were in the Otter Park Member; and three were in the Evie Member. Muskwa, Otter Park, and Evie members are the three primary stratums, from top to base, in the Horn River formation. During shipping, the samples were thickly sealed by PVC food wrap film. The sealed condition was kept in the lab until plugging just before the imbibition experiments. When plugging, cooling air-flow was used to cool the bit or saw blade. There was no liquid to contact samples before imbibition experiments in order to prevent interference from other liquid in the test. In addition, the test samples were achieved from very inside the original sample so that the water-based mud used in coring should have the least influence on the imbibition experiment. Hence, the initial saturation of the sample during spontaneous experiments was similar to the original saturation because the sample had limited time to be exposed to air and other liquids. The Woodford shale samples were the plug of one inch diameter and various lengths from half inch to two inch when they were shipped to the lab from the company. The samples were preserved by PVC food wrap film and wax. Hence, the samples were only cut to desired lengths by a power mitre saw. There was no liquid to contact samples before imbibition experiments, neither. The Niobrara shale samples were picked up in the quarry and sealed by PVC food wrap film when transporting. But the initial saturation before the experiment may be different from the original saturation in the reservoir. This is because the sample was not preserved before transporting. The original samples were big blocks. Thus, plugging and cutting were applied to achieve the desired size of the sample. Table 1 summarizes the rock properties of the shale samples that are from the Horn River, Woodford, and Niobrara shale formations. In addition to the measurement, Table 1 also includes total organic carbon (TOC) content data that were provided by the companies that contributed the samples.

416

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

Fig. 1. Laboratory view of FRT.

Table 2 XRD results of Horn River shale samples (% of Weight). No.

Illite

Illite/Smectite

Kaolinite

Chlorite

Total clay

1 2 3 4

22.2 28.3 3.3 9.2

19.2 14.7 0.5 3.0

1.9 0 0 0

8.4 0 0 0.1

52.6 43 3.8 12.3

Table 3 XRD results of Woodford shale samples (% of Weight). Fig. 2. Schematic of the imbibition experimental design.

For mineral analysis, sample cuttings were collected during plugging or cutting the test samples. The cutting was adjacent to the sample used for the experiment. Therefore, it was assumed that the mineral content determined from the X-ray diffraction (XRD) was similar to the mineral content of the samples that were used in experiments. Table 2e4 show selected XRD results for clay content of the samples from three shale formations.

No.

Illite

Illite/Smectite

Chlorite

Total clay

1 2 3

22.8 19.1 16.5

7.4 6.8 0

0 0.7 0

30.2 26.6 16.5

influence of imbibition in shale samples on the matrix permeability, micro-fracture permeability, and fracture permeability, respectively. 4.1. Imbibition impact on matrix permeability

4. Experimental results The section is subdivided into three parts to discuss the

The first sample used in the experiment was from the Niobrara formation with a 0.07% friction reducer test fluid. The experimental

Table 1 Rock and geologic properties of shale samples. Shale formations

Initial water saturation (% of PV)

Porosity (% of BV)

TOC (%)

Horn River Woodford Niobrara

3.6 to 53.3 15.0 to 60.8 12.6 to 17.3

3.48e7.41 0.46e5.33 2.1e7.47

0.55 to 6.83 0.6 to 9.17 N/A

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

417

Table 4 XRD results of Niobrara shale samples (% of Weight). No.

Illite

Illite/Smectite

Chlorite

Total clay

1

4

7

0

11

time was as long as thirty-eight days so that the long-term effect of imbibition could be observed. Fig. 3 and Fig. 4 show the sample No. 1 at the beginning and at the end of the experiment, respectively. No micro-fractures were observed in the photos. Therefore, the permeability measurement in this experiment represented matrix permeability. Later in the paper, it will discuss the influence of micro-fractures on the permeability change. According to the imbibition experimental result in Fig. 5, it can be seen that the imbibed volume of the test fluid was larger on the first day than on any other day. The imbibed liquid saturation increased by more than 45%. After the first day, the rate of imbibed liquid saturation became very slow. After five days, no more fluid was imbibed, and the imbibed liquid saturation was almost constant. The permeability values measured during the 38-day imbibition experiment are summarized in Fig. 6. Permeability on Day 0 was the original permeability of the sample before the imbibition experiments. The original permeability was the standard by which the impact from imbibition was measured. According to the permeability measurements, the sample had the largest reduction in permeability after Day 1 from about five hundred nanodarcies to tens nanodarcies. This more than 90% reduction in permeability coincides with and can be explained by the nearly 45% increase in imbibed liquid saturation on the same day. For the remainder of this sample test, the permeability values varied, but all of them still kept in about tens nanodarcies. On the thirty-eighth day, the permeability was less than 12 nanodarcies higher than after the first day of the imbibition experiment. As was mentioned before, the permeability measurements represent the matrix permeability of the sample. Therefore, in the experiment, the matrix permeability was reduced from hundreds of nanodarcies to tens of nanodarcies. Also, there was no observation to prove any recovery of the matrix permeability. Thus, it can be concluded that the matrix permeability was seriously damaged when the shale sample was immersed in the 0.07% friction reducer fluid during imbibition. The permeability measurement on the twenty-ninth day was conducted under a different drawdown pressure from that in other days. Normally, the pressure difference along the sample was approximately 400 psi. In the measurement on the twenty-ninth day, the drawdown pressure was increased to almost two times

Fig. 4. Top and bottom faces of the Niobrara sample No. 1 at the end of the experiment with 0.07% friction reducer.

to about 800 psi. That change was to investigate whether the damage was temporary and disappeared as long as the drawdown pressure increased. But according to the measurements, values on the twenty-eighth and twenty-ninth day showed no significant difference in matrix permeability when a high drawdown pressure was applied. The second Niobrara sample was tested in 7% KCl for 38 days. For this sample, the experiment also measured matrix permeability. Fig. 7 and Fig. 8 show the sample on Day 0 and Day 38, respectively. The results of the imbibition experiment and permeability measurements are combined in Fig. 9. Based on the results of the imbibition experiment in 7% KCl fluid, the largest volume of imbibition also occurred on the first two days. The imbibed liquid saturation reached 60% and the corresponding matrix permeability was reduced by 75%. After the second day, the measured matrix permeability of the sample continued to decrease. At the end of the experiment, the imbibed liquid saturation was approximately 65%, and the matrix permeability was reduced by 95%. Therefore, the result of this experiment proves the conclusion that imbibition of the fracturing fluid damages and reduces matrix permeability of the shale sample. The third shale sample was from the Horn River formation. This Horn River sample No.1 was not observed micro-fractures before the 7-day experiment either (see Fig. 10). The results of the imbibition experiment with 0.07% friction reducer and permeability measurements are seen in Fig. 11. The original permeability of the Horn River shale sample was only about 72 nanodarcies. As expected, when the test fluid was imbibed by the sample, the matrix permeability was reduced by 90% that was similar to the percentage of the reduction in the previous two cases.

4.2. Imbibition impact on micro-fracture permeability

Fig. 3. Top and bottom faces of the Niobrara sample No.1 at the beginning of the experiment with 0.07% friction reducer.

Permeability of the test Niobrara sample No. 3 was affected by micro-fractures. Micro-fractures can be observed in Fig. 12. The micro-fractures were indicated by acetone, which is an organic compound that evaporates very fast. If there is a fracture, the evaporation in the fracture is slower than on the sample surface. Thus, the fracture was marked. The summary of the imbibition and permeability change are plotted in Fig. 13. The permeability units are in microdarcies. The original permeability was about three microdarcies and was much higher than the matrix permeability of the No. 1 and No. 2 Niobrara samples. All three Niobrara samples were plugged from the original core in close proximity. Therefore, it is believed that the micro-fractures affect the sample, and the permeability measured in Niobrara sample No. 3 is the micro-fracture permeability. During

418

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

Fig. 5. Imbibition results of the Niobrara sample No.1 in 0.07% friction reducer.

Fig. 6. Permeability change of the Niobrara sample No. 1 in 0.07% friction reducer.

Fig. 7. Top and bottom faces of the Niobrara sample No. 2 on Day 1 of the experiment with 7% KCl.

Fig. 8. Top and bottom faces of the Niobrara sample No. 2 on Day 38 of the experiment with 7% KCl.

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

419

Fig. 9. Imbibition experiment and permeability change of the Niobrara sample No. 2 in 7% KCl.

Fig. 10. Top and bottom faces of the Horn River sample No. 1 at the beginning of the experiment with 0.07% friction reducer.

imbibition the micro-fracture permeability increased from three microdarcies to five hundred microdarcies. This increase in permeability was an opposite behavior to the reduction in matrix permeability that was observed in the No. 1 and No. 2 Niobrara samples. Based on the imbibition data between Day 0 and Day 1, the imbibed liquid saturation was increased more than 40%. This first day increase was similar to the previous samples. But over the second day, there was a shapely unusual rising from 40% to 58%. One reasonable explanation was that the micro-fracture was opened and imbibed additional fluid volume. This explanation is supported by Fig. 14 which shows that by the end of the second day, the shale sample was separated along the micro-fractures. Because of the separation, the experiment had to be ceased. The increase in micro-fracture permeability by imbibition was

Fig. 11. Imbibition experiment and permeability change of the Horn River sample No. 1 in 0.07% friction reducer.

420

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

Fig. 12. The Niobrara sample No.3 with micro-fractures at the beginning of the experiment with the mixture of 7% KCl and 0.07% friction reducer.

Fig. 15. The Horn River sample No. 2 with micro-fractures before the experiment with 0.07% friction reducer.

Fig. 13. Imbibition experiment and permeability change of the Niobrara sample No. 3 with micro-fractures in the mixture of 7% KCl and 0.07% friction reducer.

also found in other shale samples. In the Horn River sample No. 2 (Fig. 15), micro-fractures were obvious and influenced permeability measurement. The original permeability of this sample was about sixty microdarcies and was believed to be micro-fracture permeability. After one day of imbibition, the permeability rose to two hundred microdarcies. The results of its permeability measurements are in Fig. 16.

Fig. 14. The Niobrara sample No. 3 with micro-fractures at the end of the second day in the mixture of 7% KCl and 0.07% friction reducer.

The experiment was stopped after the first day when the sample was separated along micro-fractures (see Fig. 17). Additional three experiments found micro-fractures reopened so as to increase permeability during imbibition process in shale samples. In these experiments, the micro-fracture effect was not observed before the imbibition experiment commenced. In the Horn River sample No. 3, there were no micro-fractures observed on the sample before the experiment (Fig. 18). Because there were no micro-fractures to contribute the sample permeability, the original permeability was only 38.6 nanodarcies that were so low to be believed the matrix permeability of the Horn River sample. During the first three days, the permeability measurement indicated that the sample permeability stayed fairly constant. However, on the seventh day, the permeability rose sharply to 305.4 microdarcies. The summary of the permeability measurement is in Table 5. Also, at the seventh day, micro-fractures were observed (Fig. 19). In Fig. 20, the fluid imbibition of the Horn River sample was similar to that of previous samples during the first three days of the experiment. Most of the fluid volume was imbibed on the first day. After Day 1, the imbibed liquid saturation rate became slower and slower. However, the sample permeability did not have the characteristic reduction in matrix permeability that was expected during imbibition. A possible reason for this unusual change is that

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

421

Fig. 16. Permeability measurement results of the Horn River sample No. 2 with micro-fractures in 0.07% friction reducer.

Table 5 Permeability measurement results of the Horn River sample No. 3 in 0.07% friction reducer.

Fig. 17. The Horn River sample No. 2 with micro-fractures after the experiment with 0.07% friction reducer.

Day

Permeability, nD

Day

Permeability, nD

0 2 7

38.6 49.1 305417.3

1 3

37.3 58.9

micro-fractures and reopened them. Thus, the original measured permeability was matrix permeability and was still decreased during imbibition; at the same time, the micro-fracture permeability increased but very slowly. As a result, the total permeability was constant until the seventh day. By Day 7, the micro-fractures

Fig. 18. All faces of the Horn River sample No. 3 before the experiment with 0.07% friction reducer.

during imbibition, the imbibed friction reducer fluid slowly affected

were reopened enough so that the permeability became

422

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

experiment. Fig. 28 is for the Niobrara sample No. 5. In the Horn River sample No. 5 that already had micro-fractures, the permeability was still reduced from about eighty microdarcies to four nanodarcies because of the imbibition of crude oil. The permeability of the Niobrara sample No. 5 was also seriously decreased to only two nanodarcies. Both of these final permeability results suggest that the samples were damaged by crude oil seriously enough that the nitrogen test gas could not flow through the samples. The permeability results of these two samples are in Table 6. The samples after the experiment are shown in Fig. 29 and Fig. 30. 4.3. Imbibition impact on fracture permeability Fig. 19. The Horn River sample No. 3 at the seventh day in 0.07% friction reducer.

dominated by the micro-fracture system and the permeability increased from 38 nanodarcies to 305 microdarcies. In another Horn River shale sample, the original permeability was 23.4 nanodarcies. This permeability was so low to be considered as matrix permeability. During imbibition in 7% KCl, microfractures were reopened and the permeability was slowly increased. Fig. 21 is a graph of the permeability measurements. The sample is shown in Fig. 22 and Fig. 23. Also, the Niobrara shale sample was affected similar. This effect during imbibition in distilled water can be seen from Fig. 24 and in the photos (Fig. 25 and Fig. 26). Besides water-based fluid, crude oil was also used as the test fluid during imbibition in two shale samples. These two shale samples were from the Horn River and Niobrara formations, respectively. For the Horn River sample No. 5, micro-fractures were observed on the sample, and affect the original permeability. The original permeability was about eighty microdarcies that were much higher than tens of nanodarcies for matrix permeability in previous Horn River samples. The original Niobrara sample No. 5 was approximately eighty nanodarcies, and was believed to be matrix permeability as no micro-fractures were visible on the sample surfaces. Fig. 27 is the Horn River sample No. 5 before the

All shale samples used in this paper from the Woodford formation have natural fractures that are separated (see Fig. 31). Since the sample was already separated, steel wires were used to tighten all parts of the sample together and to ensure no movement of faces in fractures during the experiment. The original permeability of samples with separated natural fractures was in the millidarcies and is referred to as fracture permeability. The first Woodford sample was tested in 7% KCl fluid. Its original permeability was larger than one millidarcy. But after five days under imbibition, the fracture permeability was reduced to three hundred microdarcies. The final fracture permeability was reduced by about 70% that was not as large as the decrease in matrix permeability that was decreased by more than 90% in the first two cases of the paper. The imbibed saturation and corresponding permeability change are shown in Fig. 32. During the first two days of imbibition, the imbibed liquid saturation increased the most so that the permeability was reduced by six hundred microdarcies. After two days, the rate of imbibition slowed, and the permeability continued to decrease. Fig. 33 and Fig. 34 shows that the faces of the sample before and after the experiment were not significantly different. However, in another Woodford shale sample, some microfractures were reopened during imbibition in 7% KCl fluid (see Fig. 35 and Fig. 36). In addition to these photos that showed micro-fractures reopened, the permeability measurements also indicated that the

Fig. 20. Imbibition experiment and permeability change of the Horn River sample No. 3 in 0.07% friction reducer.

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

423

Fig. 21. Permeability change of the Horn River sample No. 4 in 7% KCl.

Fig. 22. All faces of the Horn River sample No. 4 before the experiment with 7% KCl.

Fig. 23. The Horn River sample No. 4 after the experiment with 7% KCl.

permeability increase because of the micro-fracture reopening (Fig. 37, note: unit is mD). The original permeability of 3.7

millidarcies was increased to about 4.5 millidarcies. The Third sample from the Woodford shale formation was tested in 0.07% friction reducer. This sample had a main fracture that was separated, but did not have any obvious micro-fractures. The sample is shown in Fig. 38 and Fig. 39. Because there was no micro-fracture observed on the sample before and after the imbibition experiment, the permeability of the sample decreased as expected. Fig. 40 shows the results of permeability measurement and the imbibition experiment. The decrease of the fracture permeability was only one hundred microdarcies, and the percentage of reduction was 15%. There is one question about this result. In the first sample of the Woodford shale samples, the imbibed liquid saturation is less than 30%, but the fracture permeability was reduced by 70%. Whereas, in this sample with 0.07% friction reducer, the imbibed liquid saturation was much larger at 70%, but the fracture permeability only had a reduction of 15%, which was much less than that of the first sample. The reason

424

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

Fig. 24. Permeability measurement results of the Niobrara sample No. 4 in distilled water.

Fig. 25. The Niobrara sample No. 4 at the beginning of the experiment with distilled water.

Fig. 26. The Niobrara sample No. 4 at the end of the experiment with distilled water.

for that difference is unknown. An explanation is that the impacts of imbibition are various because fracture surfaces are different

Fig. 27. The Horn River sample No. 5 before the experiment with crude oil.

from each other. Hence, more tests are needed to investigate that impact.

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

425

Fig. 28. The Niobrara sample No. 5 before the experiment with crude oil.

Table 6 Permeability measurement results of the Horn River sample No. 5 and Niobrara sample No. 5 in crude oil. Horn River sample no. 5

Niobrara sample no. 5

Day

Permeability, nD

Day

Permeability, nD

0 2 5

80408.8 155.9 4.3

0 2 5

78.9 25.5 2.4

Fig. 31. The Woodford sample.

shale causes matrix and fractured permeability reduction, but results in micro-fractured permeability increase. Table 7 summarizes the results. The following discusses the mechanism and reason for the permeability change. 5.1. Permeability reduction

Fig. 29. The Horn River sample No. 5 after the experiment with crude oil.

Imbibition during hydraulic fracturing can reduce permeability in the shale samples. This permeability damage occurs both in shale matrix and natural fractures that are already separated. Based on the experimental results, the more fracturing fluid volume is imbibed, the more serious the damage is. Imbibition from hydraulic fracturing is related to two formation damage mechanisms. The imbibed water-based fracturing fluid in the clay rich shale rock can react with clay mineral to cause clay swelling. In addition, the imbibed fluid can also be trapped in the small pore in shale so as to result in water blockage (Civan, 2007). Therefore, during the experiment in the paper clay swelling and water blockage both damaged and reduced the permeability of the test shale sample. This damage was considered to be permanent. When a high drawdown pressure was applied during the permeability measurement, the permeability of the shale sample was not recovered as it was suggested by previous studies for sandstone samples. This is because the damage that decreased permeability in shale rocks was caused by both clay swelling and water blockage. Even if water blockage could really be disappeared by increasing drawdown pressure, the permeability could still not be recovered because of clay swelling.

Fig. 30. The Niobrara sample No. 5 after the experiment with crude oil.

5.2. Permeability increase However, the conclusion can still be drawn, based on the results in the Woodford shale samples, the imbibition reduces the permeability of open fractures, but the reduction is less than the permeability of the matrix. 5. Result discussion According to the experimental results, imbibition process in

The experiment found micro-fractures could be reopened during imbibition process. From a macroscopic view, imbibition “lubricates” micro-fractures and causes an increase in permeability. There are three possible explanations for the lubrication occurrence. The first possible explanation is that micro-fractures are reopened because imbibition process results in shear and tensile failures. Fjar et al. (2008) indicated that water saturation changes in

426

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

Fig. 32. Imbibition experiment and permeability change of the Woodford sample No. 1 in 7% KCl.

Fig. 33. The Woodford sample No. 1 on Day 0 of the experiment in 7% KCl.

Fig. 34. The Woodford sample No. 1 on Day 5 of the experiment in 7% KCl.

the rock can induce fractures become unstable. During the imbibition experiment, imbibition process changes the saturation in the shale sample. Thus, the imbibed water-based fracturing fluid increases the pore pressure in the shale rock. If total stresses are kept constant, micro-fractures become unstable because the effective stresses are reduced so that shear and tensile failure may happen. In addition, the imbibed fluid may decrease the cohesive strength of micro-fractures to result in a shear failure. The second possible explanation is that the imbibed fluid may

Fig. 35. The Woodford sample No. 2 at the beginning of the experiment in 7% KCl.

Fig. 36. The Woodford sample No. 2 at the end of the experiment in 7% KCl.

have some physical or chemical reactions with micro-fracture surfaces and fill. These reactions may weaken and reopen the micro-fracture so as to increase the permeability. Comparing experimental results of micro-fracture and fracture permeability change, the reaction appears to have affected only micro-fractures and does not influence the separated natural fractures. This may be because the reaction occurs between the imbibed fluid and the fill in micro-fractures, and there is very little fill in separated fractures. The third explanation is that clay swelling resulting from imbibition can crack and reopen micro-fractures. Hence, as

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

427

Fig. 37. Permeability results of the Woodford sample No. 2 in 7% KCl.

5.3. Baseline experiment

Fig. 38. The Woodford sample No. 3 at the beginning of the experiment in 0.07% friction reducer.

Three more Niobrara shale samples were used in the baseline experiments. During the baseline experiment, repeated permeability measurements were conducted without any imbibition experiment. Therefore, the baseline experiment was to prove the validity of these experimental interpretations and eliminate the concern that the permeability increase due to “lubrication” may result from several cycles of pressure increase and release rather than the fluid imbibition. The procedure of the baseline experiment is that the sample was measured permeability under one confining pressure through pressure build-up method. After that, the sample was moved out from FRT machine and waited for 10 min. Then, the permeability measurement was repeated under a new confining pressure. The confining pressure was applied from 1000 psi to 4000 psi during the baseline experiment. The result is in Fig. 41. According to the result, three samples all exhibited declining trend with increasing confining stress. No lubrication effect was observed in the baseline experiment that did not have the imbibition experiment. Hence, imbibition process is believed to be a cause for the micro-permeability opening and result in permeability increase. 6. Conclusions

Fig. 39. The Woodford sample No. 3 at the end of the experiment in 0.07% friction reducer.

opposed to matrix permeability damage, clay swelling causes micro-fracture permeability increase. The evidence for this explanation is that the micro-fracture reopening only occurs in the imbibition experiment with water-based fluid. In the crude oil, there is no reopening observed. Clay swelling also exists only in water-based fluid.

This paper investigated the influence of the imbibed fracturing fluid on shale samples in the Niobrara, Horn River, and Woodford shale formations. Permeability was selected and quantified to be the criterion of the influence. Based on imbibition experiments and permeability measurements, imbibition process can decrease or increase shale permeability. The detailed conclusions are presented: 1. Matrix permeability of the shale sample is damaged and reduced by imbibition. The reduction in some cases may be as high as 95% based on the original matrix permeability. In addition, the increase in the imbibed volume under imbibition causes a corresponding decrease in the matrix permeability. The more volume is imbibed, the worse the reduction of the matrix permeability is.

428

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

Fig. 40. Imbibition experiment and permeability change of the Woodford sample No. 3 in 0.07% friction reducer.

Table 7 Summary of the experiment results for each test sample. Sample

Result

Sample

Result

Sample

Result

Niobrara sample Niobrara sample Niobrara sample Niobrara sample Niobrara sample

Matrix permeability reduction

Horn River sample No.1

Matrix permeability reduction

Woodford sample No.1

Fracture permeability reduction

Matrix permeability reduction

Horn River sample No.2

Woodford sample No.2

Micro-fracture permeability increase Micro-fractures reopened; permeability increase Permeability decrease in crude oil test fluid

Horn River sample No.3

Micro-fracture permeability increase Micro-fractures reopened; permeability increase Micro-fractures reopened; permeability increase Permeability decrease in crude oil test fluid

Micro-fractures reopened; permeability increase Fracture permeability reduction

No.1 No.2 No.3 No.4 No.5

Horn River sample No.4 Horn River sample No.5

Woodford sample No.3

Fig. 41. Baseline experiments with various confining pressures.

2. The matrix permeability damage in shale formations from imbibition is believed to be permanent. Increasing drawdown

pressure, which was suggested in previous studies, cannot eliminate the damage. This is because that both clay swelling

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

3.

4.

5.

6.

and water blockage cause the permeability damage in shale matrix. Even if higher drawdown pressure could reduce the damage from water blockage, clay swelling still declines matrix permeability in shale formations. Fracture permeability of natural fractures that are already separated is also decreased after imbibition experiments. But the decrease is smaller than the decrease in matrix permeability. Micro-fracture permeability in shale samples can increase during imbibition. The imbibed fracturing fluid “lubricates” microfractures and reopens them. The reason for the lubrication occurrence is that the imbibed water-based fracturing fluid weakens the stability of micro-fractures. Thus, the unstable micro-fractures are reopened so as to increase the permeability. It is difficult to distinguish which fracturing fluid, KCl or friction reducer, has the better or worse influence on the permeability change. The combination of permeability increase in micro-fractures and permeability reduction in fractures and matrix determines whether hydraulic fracturing stimulation increase production in shale formations. In addition, shale imbibition is required to be controlled in order to minimize negative impacts and to maximize positive impacts from imbibition.

The key limitations of this paper are stated below, and recommendations for future study are also included: 1. This paper provides a qualitative analysis of the influence of the imbibition process in shale. Further studies are required to quantify the contribution or damage of imbibition on permeability. 2. This investigation finds that the lubrication effect, which is caused by imbibition, can increase micro-fracture permeability. Future studies should investigate the reason for the increase of the micro-fracture permeability. 3. The imbibition experiments were at room temperature and atmospheric pressure. These conditions are much different from the reservoir environment. Further studies should simulate imbibition at reservoir conditions so results can be scaled to field. 4. The sample could be affected under different stresses when applying separated imbibition experiment and permeability measurement. Future studies are required to measure imbibition and permeability under the same condition. Acknowledgments The authors appreciate the Fracturing, Acidizing, Stimulation Technology (FAST) Consortium at Colorado School of Mines for sponsoring this project, and also thank Nexen Energy ULC, INPEX Gas British Columbia Ltd, and Devon Energy Production Company L.P. for their courtesy in providing the shale samples used in the study. In addition, the authors thank Mr. Joe Chen for his help in the experiment, and Ms. Avis Downey for her contribution on this paper. Nomenclature A C(P) L M Pgs/Pgt Pinitial Pinlet Poutlet

sample cross sectional area gas compressibility sample length gas molar mass pressures at the standard and test conditions, respectively initial pore pressure inlet pressure outlet pressure

R Tgs/Tgt t Vd vx Zgs/Zgt

mg rgs/rgt 4

429

gas constant temperature at the standard and test conditions, respectively time downstream reservoir volume velocity in x direction compressibility factors at the standard and test conditions, respectively gas viscosity gas densities at the standard and test conditions, respectively porosity

SI metric conversion factors psi  6.894757, E þ00 inch  2.54, E 02

¼kPa ¼m

References Abrams, A., Vinegar, H.J., 1985. Impairment Mechanisms in Vicksburg Tight Gas Sands. Paper SPE 13883 presented at the SPE/DOE Low Permeability Gas Reservoirs Symposium, Denver, Colorado, 19e22 March. Alkouh, A.B., Wattenbarger, R.A., 2013. New Advances in Shale Reservoir Analysis Using Flowback Data. Paper SPE 165721 presented at the SPE Eastern Regional Meeting, Pittsburgh, Pennsylvania, 20e22 August. Barati, R., Hutchins, R.D., Friedel, T., Ayoub, J.A., Dessinges, M., England, K.W., 2009. Fracture impact of yield stress and fracture-face damage on production with a three-phase 2D model. SPE Prod. Op. 24 (No. 2), 336e345. Bazin, B., Peysson, Y., Lamy, F., Martin, F., Aubry, E., Chapuis, C., 2009. Situ Water Blocking Measurements and Interpretation Related to Fracturing in Tight Gas Reservoirs. Paper SPE 121812 presented at the 8th European Formation Damage Conference, Scheveningen, The Netherlands, 27e29 May. Charoenwongsa, S., 2011. Numerical Simulation of the 3-D Hydraulic Fracturing Process, Cleanup and Relevant Physics. Doctor’s thesis. Colorado School of Mines, Golden, Colorado. Civan, F., 2007. Reservoir Formation Damage, second ed. Gulf Professional Publishing, Burlington, pp. 5e7. Dutta, R., Lee, C., Odumabo, S., Ye, P., Walker, S.C., Karpyn, Z.T., Ayala, L.F., 2012. Quantification of Fracturing Fluid Migration Due to Spontaneous Imbibition in Fractured Tight Formations. Paper SPE 154939 presented at the SPE Americas Unconventional Resources Conference, Pittsburgh, Pennsylvania USA, 5e7 June. Ehrlich, R., 1970. The Effect of Temperature on Water e Oil Imbibition Relative Permeability. Paper SPE 3214 presented at the SPE Eastern Regional Meeting, Pittsburgh, Pennsylvania, 5e6 November. EIA, 2015 (April). The Annual Energy Outlook 2015. U.S. Energy Information Administration (Tech. rept. U.S. Department of Energy). Fjar, E., Holt, R.M., Raaen, A.M., Risnes, R., Horsrud, P., 2008. Petroleum Related Rock Mechanics, second ed. Elsevier Science, , Amsterdam, pp. 75e79. Ghanbari, E., Xu, M., Dehghanpour, H., Bearinger, D., 2014. Advances in Understanding Liquid Flow in Gas Shales. Paper SPE 171653 presented at the SPE/ CSUR Unconventional Resources Conference e Canada, Calgary, Alberta, Canada, 30 Septembere2 October. Hadley, G.F., Handy, L.L., 1956. A Theoretical and Experimental Study of the Steady State Capillary End Effect. Paper SPE 707 presented at the Fall Meeting of the Petroleum Branch of AIME, Los Angeles, California, 14e17 October. Hayatdavoudi, A., 1999. Changing Chemophysical Properties of Formation and Drilling Fluid Enhances Penetration Rate and Bit Life. Paper SPE 50729 presented at the SPE International Symposium on Oilfield Chemistry, Houston, Texas, 16e19 February. Holditch, S.A., 1979. Factors affecting water blocking and gas flow from hydraulically fractured gas wells. J. Pet. Technol. 31 (12), 1515e1524. King, G.E., 2010. Thirty Years of Gas Shale Fracturing: what Have We Learned? Paper SPE 133456 presented at the SPE Annual Technical Conference and Exhibition, Florence, Italy, 19e22 September. Land, C.S., 1968. Calculation of imbibition relative permeability for two- and threephase flow from rock properties. SPE J. 8 (No. 2), 149e156. Mahadevan, J., Sharma, M.M., 2003. Clean-up of Water Blocks in Low Permeability Formations. Paper SPE 84216 presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado, 5e8 October. Makhanov, K., Dehghanpour, H., Kuru, E., 2012. An Experimental Study of Spontaneous Imbibition in Horn River Shales. Paper SPE 162650 presented at the SPE Canadian Unconventional Resources Conference, Calgary, Alberta, 30 October-1 November. Morsy, S., Sheng, J.J., 2014. Imbibition Characteristics of the Barnett Shale Formation. Paper SPE 168984 presented at the SPE Unconventional Resources Conference, The Woodlands, Texas, USA, 1e3 April. Oort, E., 1994. A Novel Technique for the Investigation of Drilling Fluid Induced Borehole Instability in Shales. Paper SPE 28064 presented at the SPE/ISRM Rock

430

Z. Zhou et al. / Journal of Natural Gas Science and Engineering 29 (2016) 413e430

Mechanics in Petroleum Engineering Conference, Delft, The Netherlands, 29e31 August. Penny, G.S., Soliman, M.Y., Briscoe, J.E., 1983. Enhanced Load Water-recovery Technique Improves Stimulation Results. Paper SPE 12149 presented at the Annual Technical Conference and Exhibition, San Francisco, California, 5e8 October. Putthaworapoom, N., Miskimins, J.L., Kazemi, H., 2012. Sensitivity Analysis of Hydraulic Fracturing Damage Factors: Reservoir Properties and Operation Variables paper SPE 151060 presented at SPE/EAGE European Unconventional Resources Conference and Exhibition, Vienna, Austria, 20 e 22 March. Qin, B., 2007. Numerical Study of Recovery Mechanisms in Tight Gas Reservoirs. Master’s thesis. University of Oklahoma, Norman, Oklahoma. Roychaudhuri, B., Tsotsis, T., Jessen, K., 2011. An Experimental Investigation of Spontaneous Imbibition in Gas Shales. Paper SPE 147652 presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado, 30 October-2 November.

Soliman, M.Y., Hunt, J.L., 1985. Effect of Fracturing Fluid and its Cleanup on Well Performance. Paper SPE 14514 presented at the SPE Eastern Regional Meeting, Morgantown, West Virginia, 6e8 November. Wu, Y.S., Fakcharoenphol, P., Zhang, R., 2010. Non-darcy Displacement in Linear Composite and Radial Flow Porous Media. Paper SPE 130343 presented at the SPE EUROPEC/EAGE Annual Conference and Exhibition, Barcelona, Spain, 14e17 June. Yao, Y., Wu, Y.S., Zhang, R., 2012. The transient flow analysis of fluid in a fractal, double-porosity reservoir. Transp. Porous Media 94 (No. 1), 175e187. Zhang, R., Wu, Y.S., Fakcharoenphol, P., 2014. Non-Darcy displacement in linear composite and radial aquifer during CO2 sequestration. Int. J. Oil Gas Coal Technol. 7 (No. 3), 244e262. Zhou, Z., Hoffman, T., Bearinger, D., Li, X., 2014. Experimental and Numerical Study on Spontaneous Imbibition of Fracturing Fluids in Shale Gas Formation. Paper SPE 171600 presented at the SPE Canadian Unconventional Resources Conference, Calgary, Alberta, 29 September- 2 October.