Fracture Design Considerations in Naturally Fractured Reservoirs

SPE SPE 17607 Fracture Design Considerations in Naturally Fractured Reservoirs by C.L, Ctpolla, P.T, Branagan, and S,J. Lee, CER Corp. SPE Members Copyright 1908 Society of Petroleum Engineers This paper waa prepared for presentation at the SPE International Meeting on Palroleum Engineering, held In Tianjin, China, November 14, 1988. This paper wee selected for presentation by an SPE ProQram Commlttae 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 Englneere and are subject to correction by the author(s). The material, as presented, doe$ not necessarily refiect any position of the Society of Petroleum Engineers, Ite officere, or members. Papers presented at SPE meetings are eubjsct to publication review by Editorial Commltteee of the Society of Petroleum Englneere. Permission to co~j:2 restricted IOan abstract of not more than 300 words, 1’ustratlone may not be copied. The abstract should contain conspicuous acknowledgment of where and by whom the paper Is presented. Write Publications Manager, SPE, P.O. Box 833838, Richardson, TX 76083.3336, Telex, 730989 SPEDAL. ABSTRACT The ability to effectively enhance production through hydraulic fracturing is dependent on an accurate description of the reservoir production mechanism(s). Fracture designs may differ greatilydepending on the production mechanism(s). The complex nature ofhydraulic- ally fractured reservoirsin whichthepredomi- nant production mechanism is a set of inter- connected, naturally occurring fractures is investigated in this paper. The paper inte- grates general reservoir simulation results with actual field data from a naturally frac- tured reservoir in the Piceance Basin, Color- ado. The study investigates a variety of natural fracture/ntatrixproperties and compares the productivity of these naturally fractured reservoirs to homogeneous reservoirs with the same average flow capacity. The paper also investigates the influence of natural fracture anisotropy on hydraulic fracture design. The effect of damage to the natural fracture system is illustrated and compared to analogous homogeneous reservoirs. The economic considerations associated with many of the reservoir production mechanisms are presented. The results of the reservoir simulations indicate that optimum fracture lengths for isotropic, naturally fractured reservoirs are identical to those estimated for homogen- eous reservoirs having the same average flow capacity. Therefore, accepted fracture design considerations to determine optimal fracture length and conductivity can be used in isotro- pic, naturally fractured reservoirs based on References and illustrations at end of ~aper the average flow capacity of the reservoir. However, fracture design considerations are more complex when the effects of natural fracturedamage andanisotropy areencountered. INTRODUCTION Thebasic fracturedesign criteria forhomogen- eous reservoirs has been discussed in detail by several authors.1-7 This literature also illustrates the interrelationshipof fracture length, fractureconductivityandwell product- ivity, and the economic impactof many fracture design considerations. However, fracture designconsiderations inmorecomplex ,natural- ly fractured reservoirs-are not widely.avail- able in the literature. This paper presents reservoir simulations and field data that illustratemany fracturedesignconsideratims in naturally fractured reservoirs. The initial requirement for designing a hy- draulic fracturing treatment is an accurate description of the reservoir, including the predominantproductionmechanism(s) . Reservoir production mechanisms and characteristics can be obtained from log, core, geological, well test and production data. In many cases, a limited amount of data are available, and reservoir characteristics and production mechanisms are inferred from pre- and post- fracture well performance. There are many uncertainties associatedwith inferringreser- voir properties based on a limited amount of data because reservoirs with vastly different production mechanisms canproducevery similar pressure /production profiles. The reservoir simulations presented will illustrate the similarities inproduotion andpressure buildup behavior for homogeneous and naturally frac- tured reservoirs that have the same average flow capacity. S87 . . .M FRACTURE DESIGN CONSIDERATIONS IN UAIWRAT.T.Y FRAC!ITWRF!D Wi!RIWVnTPC smn 17cn* M many cases, well test results canbe inte- grated with core, log and geological data to identify and quantify reservoir properties and production mechanisms. However, in many cases, the absence of preoise bottomhole pressures and the effects of wellbore stor- age/afterflow reduce the accuracy and detail that can be obtained from well test analysis. Well test analysis methods have been presented to identify many reservoir production mecha- nisms.81g The more recentq ti e curve and derivative analysis methods-o -% have aided in identifying complex production mechanisms and various flow regimes. The derivative plotting techniques can assist in identifying linear, bilinaar, radial and natural fracture flow regimes. The more complex reservoirs that exhibit a high degree of anisotropy and/or more than one predominant production mechanism may preclude the effective use of theabove analysismethods. Complex reservoirs may require additional data and luoresophisti- cated analyeis techniques to successfully interpret well test and production data. To optimize fracture length and conductivity, the post-fracture production resulting from various stimulation designs/alternatives should be predicted, and the economics of each treatr ‘nt should be compared. The tran- sientproduction/pressure behavior forhy&=ul- ically fracturedwells inahomo eneous system ?L has been presentedby Cinco etal 3andAgarwal et a15 and can be ueed to estimate the post- fracture production during transient flow. The pseudo-steady state productivity of hy- draulically fractured welle, presented by McGuire and Sikora,l can predict well perfor- mance during pseudo-steady state flow. Reser- voir simulation techniques can also be used in more complex reservoirs to redict post- fracturewell performance, s,l~,l?which ~cl~e non-Darcy flow and layered resemoirs. The effects ofcomplexnatural fractureproduction mechanisms combined with a hydraulic fracture can be too intricate for analytical well solutions andmay require reservoir simulation techniques to obtain quantitative predictions of post-fracture well performance.16117 There are two major factors that influence post-fracture well productivity associated with the selection of stimulation materials: ● fracture conductivity, and ● reservoir damage. Holditch18 and Pratslg have presented studies on the effect of reservoir damage on well productivity for homogeneous reservoirs at the fracture faces. These studies illustrated that in most cases, reservoir damage does not significantly affect well productivity. However, Branagan et a116~20 have shown that damage to natural fractures intersected by a hydraulic fracture can significantly reduce post-fracture well productivity. In the case of a naturally fractured reservoir, the majority of fluid leakoff is into the natural fracture system. Therefore, the relative effects of damage are magnified due to the large volume of fluid (and polymer) injected into the natural fractures intersected by the hydraulic fracture. The effects of natural fracture damage are illustrated later in the text. The effects and magnitude of in situ fracture conductivity in homogeneous reservoirs have been discussed by many authors.21-23 In general, the required insitu fractureconduct- ivity for a homogeneous reservoir can be ~~Ztda~53 the following equation after Cr = 10 = khfWhf/3.14~ave~f (1) A Cr value of 10 or more is considered suffi- cient for most applications, providing that the fracture conductivity used in Equation 1’ ‘Mwhf’ ie representative of the actual in s u fracture conductivity and that non- Darcy flow effects are minimal. The required hydraulic fracture conductivity for naturally fractured reservoirs is investigated in this paper in terms of the required Cr value. This paper integrates current design criteria for homogeneous reservoirs with a reeervoir simulation study and actual field data to present fracture.design criteria for naturally fractured reservoirs. The results were ob- tained using a finite difference reservoir simulator that was specifically designed to model transient matrix and natural fracture flow in the presence of a hydraulic fracture. The model verification, along with a more *V::: Pypyr:i%tion’ ‘as ‘resented ‘n a PRESSURE BUIIJXJPBEHAVIOR OneWidely-used method forestimating reservoir permeability and the predominant production mechanism is pressure buildup testing. The pressure buildup behavior of naturally frac- turedreservoirs haebeen resented inprevious % works by Branagan et al 4 and otheks.10-12 The pressure buildup behavicr for a set of homogeneous and naturally fracturedreservoirs wae conducted to compare the behavior of the two production mechanisms. The simulations were performed usfnga cartesian and anatural- lyfractured gas reservoirmodel. The natural- ly fractured model is described and verified in SPE Paper 16434,24 while the cartesian model is described in detail in SPE Paper 16219.15 The simulated reservoirs were rela- tively tight gas formations, exhibiting an average permeability of 0,02 md. Table 1 contains the basic reservoir parameters used for the simulations. Figure 1 is a Horner plot of the pressure buildup behavior of two naturally fractured reservoirs and a homogeneous reservoir, all having the same average flow capacity. These simulated buildups do not include the effects of wellbore storage. The figure illustrates how the early time Horner behavior isaffocted by the contrast in natural fracture and matrix conductivity. Natural fractureCaseA exhibits a significantly smaller slope in the early time (Horner time between 50 and 1,000) than 6SU PE 17607 C.L. CIPOLLA. P.T. BRANAGAN AND S.J. LEE —- . . .—.— — Case B which has a much smaller conductivity The simulations are intended to illustrate contrast between the natural fractures and the applicability of Equation 1, C , topredict the matrix. Reviewing this figure shows f the required fracture conduct vity for a that natural fracture Cases A and B and the naturally fractured reservoir. It should be homogeneous reservoir converge to the same noted that the value of kave used in Equation middle time Horner slope. This confirms lshouldrepresent theaverage, bulk permeabil- thatall three reservoirshave the same average ity of the naturally fractured reservoir as flow capacity. obtained from the middle time Horner slope or other appropriate estimations. The simula- Figure 2’ is a log-log pressure/derivative tions are also intended to evaluate the long comparison of the three cases. Natural frac- term productivity of naturally fractured ture Case A exhibits the characteristic pres- reservoirs. The base resemoir data used sure and pressure derivative shapes fornatur- for these simulations are the same as listed ally fractured reeervoirs,24 while Case B in Table 1. The performance for each case does not, dueto the small contrast in natural was simulated for 10 years using a constant fracture and matrix conductivity. In the bottomhole pressure (BHP) of 1,500 psi. It absence of wellbore storage, predominant should be noted that the simulation model natural fracture production is evident from has been verified against accepted analytical well testing. However, when production is solutionswhereapplicable, andtheseverifica- not totally domitiatedby natural fractures, tions have been presented in previous publica- well testing may not identify natural fracture tions.15#24 production. Figure 5 compares the lo-year performance Figure 3 is a Horner comparison identical to for isotropic naturally fracturedandhomogen- Figure 1 except for the inclusion of wellbore eous reservoirs for Cr (Equation 1) values storage. The effect of wellbore storage of 0.1, 1, 10 and 100. The figure shows masks the early time data that aids in identi- that well performance is identical for both fying natural fracture behavior. Figure 4 production mechanisms and is a function of is the log-log comparison of the pressure average, bulk reservoir flow capacity only. buildups. Againr the effects of wellbore The relationship between hydraulic fracture storage mask the characteristic derivative conductivityandaverage reservoirpermeability curve associated with naturally fractured fora naturally fracturedreservoir (asdefined reservoirs. Therefore, in many field appll,ca- by Cr) is the same as that for homogeneous tions, well testing may not provide sufficient reservoirs. Therefore, accepted fracture data to identify natural fracture production design criteria to optimizehydraulic fracture mechanisms. The examples presented are in- length and conductivity for homogeneous reser- tended toillustrate thedifficultly inidenti- voirs is applicable to isotropic naturally fying naturally fractured reservoirs based fractured reservoirs. solely on well test data and the usefulness of a bottomhole shut-in of test wells. It shouldbe noted thatthere are manyhydraul- ic fracture design criteria for naturally Previouswork2 4hasshownthat natural fracture fractured resemoirs relating to fluid loss anisotropyis not easily identifiednorquanti- and natural fracture damage that differ from fied from well test data. There may be no homogeneous reservoirs. The simulations distinguishingcharacteristicsbetweenisotrop- ic and anisotropic naturally fractured reser- presented assume that the process of creating the hydraulic fracture does not impair the voirs. Also, prev+.ouswork24 has discussed flow capacity of the intersected natural the post-fracture pressure buildup behavior fractures. In many cases, the flow capavity of naturally fractured reservoirs. That work of the natural fractures can be significantly emphasized the similarity in pressure buildup impaired by stimulation fluids.16 Also, the characteristics between various ieotropic applicabilityofhomogeneous resenoirfraoture andanisotropic naturally fracturedreservoirs design criteria to naturally fractured reser- containing hydraulic fractures. The conclu- voirs assumes that treatment design and mater- sions indicated that calculated hydraulic ials are of similar nature and cost for both fracture lengths could vary greatly depending reservoirs. The problems associated with on prior knowledge of the degree of reservoir fluid loss and natural fracture damage may anisotropy. dictate different stimulation designs and materials for naturally fractured resenoirs compared to analogous homogeneous reservoirs. POST-FRACTURE WELL PERFORMANCE The prediction of post-fracture well perfor- ANISOTROPXC RESERVOIRS mance of homogeneous reservoirs is well docu- mented,l-5~=3 as are the criteria foroptimiz- Well Performance ing fracture length and conductivity.1g120 The extension of these procedures to naturally In many naturally fractured reservoirs, the fractured reservoirs isevaluatedby comparinq fracture system is anisotropic.17 The degree the simulated post-fracture production for of anisotropy can often be as much as 100 homogeneous and naturally fracturedreservoirs 1 and not be evident from well test data. B having the same average flow capacity. This The direction of the reservoir anisotropy section is limited to isotropic reservoirs. can be directly related to the in situ stress 589 FRACTURE DESIGN CONSIDERATIONS IN NATURALLY FRA Fieldofthe reservoir, with the lowpermeabil- ity natural fractures aligned parallel to the minimum prinaiple stress.16 Therefore, the hydraulic fracturewill probably intersect :he less permeable natural fracture set. ?igure 6 illustrates the minimum and maximum ?rinciplestresses, the orientation of natural Eracture permeability and the most probable orientation of a hydraulic fracture. 4 set of reservoir simulations were conducted to illustrate the effect of natural fracture rnnisotropyon pnst-fracturewell productivity and fracture design criteria. The basic reservoir data were listed in Table 1, while the details of each case are shown in Table 2. A natural fracture anisotropy of 10 to 1 Was used for all anisotropic simulations. rhe simulations predicted well performance for 10 years using a variety of fracture lengths intersecting both the minimum (most probable case) and the maximum permeability natural fracture set. The hydraulic fracture conductivity for all cases was held constant at 250 md-ft, and fractures lengths of 400, 800, 1,200 and 1,600 ft were simulated. The same hydraulic fracture data set was used to simulate post-fracture well performance for a corresponding isotropic naturally fractured reservoir for comparison. Figure 7 compares the predicted 10-year well performance for an 800-ft hydraulic fracture that intersects theminimumand maximumpermea- bility natural fractures (reference Figure 6). The performance of the corresponding isotropic naturally fractured reservoir is shown for comparison. The figure illustrates that significantly higher production rates are obtained if the hydraulic fracture is preferentially oriented to intersect the high permeability set of natural fractures. However, in many reservoirs, the in situ stress field results inanunfavorable fracture orientation (intersectin~the lowpem.aability Sf3t of natural fractures”5) (referenceFigure 6). Although not shown, the long term produc- tion fcrtheunstimulated isotropic and ar.iso- tropic cases is virtually identical. Optimum Fracture Length and Economics The cumulative production after 10 years as a function of propped fracture length is compared in Figure 8 for isotropic and aniso- tropic naturally fractured reservoirs. The figure illustratesagainthatwell performance is significantly affected by reservoir anise- tropy. It should be noted that the base production (no hydraulic fracture) is the same for both the isotropic and anisotropic naturally fractured reservoirs. Therefore, pre-fracture production characteristics, even long term, can not identify natural fracture anisotropy. T1.e figure includes two fracture orientations in the anisotropic case, parallel to the minimum permeability natUral fracture (denoted Miso Min) and parallel to the maximum permeability natural fractures(denotedAniso Max). As discussed, the more probable case is linisoMax, where the hydraulic fracture intersects theminim.us permeability natural fractures that are many PURED RESERVOIRS SPE 1760 times oriented paralleltothe minimum horizon- tal stress (reference Figure 6). Figure 8 illustrates the drastic effect that fracture orientation has on 10-yearcumulative production. If the hydraulic fracture is oriented in a favorable direction, parallel to the minimum permeability natural fractures (Aniso Min), then the cumulative production may be almost twice that expected from the unfavorable orientation. The isotropic case is approximately in the middle of the two extremes. The purpose cf Figure 8 is to emphasize the significanceof natural fracture anisotropyon post-fracturewell productivity. Again, the fracture orientation is prcbably not in the favorable direction.16 Therefore, post-fracture well productivity inanisotropic naturally fractured reservoirs is likely to be less than expected. Without prior knowledge of the anisotropy, post-fracturewell product- ivity may erroneously be interpreted as an ineffective stimulation treatment. To illustrate the effect of reservcir aniso- tropy on optimum fracture length, a simple economic comparison was conducted. Table 3 lists the base economic data used for the comparison. Figure 9 shows the present value prOfit (PVP) for each case. The PVP is defined as the discounted net gas revenue minus base investment and stimulation costs. The figure illustrates that the optimum fractute length is longer for the anisotropic naturally frac- tured reservoir with the hydraulic fra~ture oriented parallel to the low permeability natUral fractures, Aniso Min (this case is not commonly found in actual practice 16/24), compared to the isotropic case. The shorteet optimum fracture length is estimated fo~ the anisotroplc naturally fractured reservoir with the hydraulic fracture oriented parallel to the high permeability natural fractures. There is considerable difference in the PVP dependingon the type of reservoir and fracture orientation, again emphasizing the importance of identifying reservoir anisotropy. NATURAL FRACTURE PERMEABILITY IMPAIRMENT Simulated production The effects of permeability impairment to the natural fractures intersectedby ahydraul- ic fracture can significantly reduce post- fracture well productivity. As discussed, duringa stimulationtreatment,the interjected natural fractureswillbetheprimary mechanism for fluid loss into the reservoir. Thedispro- portionate amount of fluid lost into the natural fractures will magnify the effects of permeability impairment due to fracturing fluid residue and relative permeability/water blccking.16~20 Asetof reservoir simulations was conducted to illustrate the effects of natural fracture permeability impairment. An isotropic naturally fractured reservoir from the previous section was selected, which contained an 800-ft hydraulic fracture. The permeability of the natural fractures inter- sected by the hydraulic,fracture was reduced o . 17tin7 C.L. CIPOLLA. P.T. BRANAGAN AND S.J. LEE , . — .—. , —. —— —. -—.— -—- —— ;O 1 percent of the original value (reference As a further illustration of the effects of r8bleS 1 and 2 for original values). Again, wellbore damage on pressure buildup behavior, ?ost-fractureproduction was simulated for 10 the Horner and log-log plots of the above {ears. Figure 10 compares the production well tests are presented in Figures 13 and #ithandwithout natural fracturepermeability 14 with the well shut-in at the surface. hzpairment. The figure shows that significant- Reviewing the figures shows that the entire ly less production ie realized if the inter- test is influenced by wellbore storage/after- Sected natural fractures are affected by the flow and can provide very little information. Stimulationfluids. Although reservoircharac- The log-log plot, Figure 14, exhibits a unit teristics and stimulation treatments vary slope for most of the buildup period. There- areatly, field data has indicated that natural fore, in many tight reservoirs, a bottomhole l?racture permeability impairment of this shut-in combined with extended test duration nagnitudeis probablewhen water-based stimula- may be required to minimize wellbore stor- tion fluids are employed with no fluid loss age/afterflow and provide reliable data. additives.16~20 The useof foamed stimulation tluidscombined withsolidfluid loss additives has been tested attheMWX, and initial results FIEIJ)DATA me promising.26 There has been extensive geological, log, Pressure Buildup Behavior core, well test and production data gathered at the DOE MWX cite. The reader can refer [n many cases, naturally fractured reservoirs toprevious Ublications foradditional details onthe~. ~7-31 Thewell test and production sre tested using conventional surface shut- lns and relatively short test times. AssUming results for a naturally fractured reservoir negligible permeability impairment of the at the MWX site ie presented in this section. latural fractures near the wellbore, this The results are reproduced from previous ~rocedure may result in adequate test data. publications.16J17 The reservoir was thor- iowever,in the case wherethenatural fracture oughly tested prior to stimulation to obtain system inthevicinity of t!lewellborehas been an accurate reservoir description for subse- lnfluenced by drilling and completion opera- quent hydraulic fracture design and post- Lions, conventional well test procedures and fracture well test analysis. Following the Iurations may not be adequate. stimulation treatment, extended production and well testing data were obtained in an l?hepressure buildup behavior of an unstimu- attempt to quantify the stimulation resulte. Lated,naturallyfracturedreservo ircontaining a 10-ft damage zone around the wellbore was Initial reservoir data were obtained from 6imulated. The base reservoir data are listed log, in Table 1, natural fracture Case A. core, geological, stress testing and The outcrop studies. These studies aided greatly permeability of the damaged zone is 1 percent in the identification of the natural fracture of the original natural fracture permeability production mechanism, reservoir anisotropy, (l percent of l,980md= 19.8 red). Thepres- hydraulic fractureorientationand theorienta- sure buildup behavior of the corresponding tion of minimum and maximum natural fracture homogeneous system was also simulated for permeability for the development of accurate comparison (reference Table 1, homogeneous reservoir models. It should be noted that case). The drawdown period was 72 hours, the pressure buildup data at the MWX was followed by a shut-in lasting 168 hours. obtained using abottomhole shut-intominimize The test duration was selected to reflect wellbore storage/afterflow, thus providing conventional test durations. Each well was excellent reservoir data to identify natural produced at a constant surface rate of 18 fracture flow regimes. MCFD . For reference, the undamaged pressure buildup behavior of both cases is shown in MWX Paludal Zone Figures 1 through 4. The well test and production data gathered Figures 11 and 12 are the Horner and log-log in the Paludal interval at the MWX is summar- plots, respectively, of the simulated pressure ized in this section. buildup behavior of the homogeneous andnatur- The Paludal zone is a channel deposit approximately 700 ft wide. ally fractured reservoirs using a bottomhole Figure 15 is aplotof thepre-fracture produc- shut-in (minimalwellbore storage). Reviewing the log-leg plot in Figure 12, it appears tion data from MWX-1 (production/test well) that the later time portion of the buildup and the bottomhole pressures for MWX-1 and the two observation wells, MWX-2 and MWX-3. test may provide some reliable data. However, the calculated permeability from the Horner Figures 16 and 17 are Horner and log-log plots, respectively, of the final preseure plot is 0.001 md for the naturally fractured buildup shown in Figure 15. Also shown in reservoir and 0.0004 md for the homogeneous Figures 16 and 17 is the simulated pressures reservoir. The actual average permeability using the above mentioned naturally fractured of both reservoirs is 0.02 md, illustrating reservoir model. Table 4 shows the model the magnitude of error in estimating reservoir input data used to match the paludal pre- permeability from well test data of insuffi- fracture well test and production data. The cient duration in wells with wellbore damage. table shows that anatural fracture anisotropy This calculated permeability could result in of 10 to 1 was required to match the pressure a less than optimum stimulation design and buildup behavior of the test well (MWX-1) inaccurate evaluation of post-fracture well and the lack of pressure interference in the test and production data. two.observation wells (MWX-2 and MWX-3). FRACTURE DESIGN CONSXDERATXONS IN NATURALLY ERA( Phe Paludal zone was then hydraulically frac- kured using a water-base stimulation fluid. F@ure18illustrates thepost-fracture produc- tion and bottomhole pressures in MWX-1. The figure shows that thepost-fracture production rate is less than the pre-fracture production rate (reference Figure 15 . & A comprehensive reservoir ZIodelingstudyl ?20 indicated that m conductive hydraulic fracture had been created, but during the fracturing process, the intersectednatural fracturesweredamaged. AS a result, initial post-fracture production was impaired. Following an extended shut-in period, the well was recentered and tested again. Figure 19 shows the re-entry production and well test results. The figure indicates that flow rates are enhanced compared to both the initial post-fracture and pre-fracture rates. The Horner plot of the re-entry pressure buildup data is presented in Figure 20, along with the reservoir simulation history match. Table 5 lists the model input data for the history match of the re-entry well test and production data. The log-log plot of the re-entry data is presented in Figure 21, comparing the pressure and derivative curves of the actual and the simulated data. Review- ing Figures 20 and 21 shows that the rese~oir simulation model accurately predicted the pressure/productionbehavior for this Paludal zone. The conductive fracture length used for the history match was 100 ft, much shorter than the designed length of 400 ft. Again, moredetaileddiscussions ofthePaludal zone well test and stimulation history can be found in previous papers.16~20 The results do illustrate the effects of natural fracture permeability impairment and isotropy on post- fracture well productivity. It should be noted that the hydraulic fracture orientation was estimated to be parallel to the maximum permeability natural fractures, based on geology, well testing and the orientation of situ stresses. The reservoir simulation study and field data presented illustrate that hydraulic fracture design in naturally fractured reser- voirsrequires extensivepre-fracturereservoir data. The reservoir simulation study focused on the feasibility of applying accepted hy- draulic fracture design criteria for homogen- eous reservoirs to naturally fractured reser- voirs. The simulation study selected specific cases for comparison and then simulated pre- and post-fracture well productivity for both homogeneous and naturally fractured reservoirs. The comparisons were based on homogeneous and naturally fractured reservoirs with identical average/bulk rese~oir permea- bilities. The simulation study investigated the pre- fracturepressure buildupbehavior ofnaturally fractured reservoirs compared to analogous homogeneous reservoirs. This portion of the study illustrates the concept of average/bulk MJRED RESERVOIRS SPE 176C reservoir permeability for naturally fractured reservoirs and emphasizes the problems associ- ated with wellbore storage/afterflow. The abilitytodistinguish natural fractureproduc- tion is significantly affected bytheduration of wellbore storage. In cases where wellbore storage is extensive, natural fracture flow regimes may be completely absent, and only a bottomhole shut-in can provide sufficient data to accurately describe the reservoir using well test data. Post-fracture well productivity for naturally fractured wells is compared to that of analog- ous homogeneous reservoirs. The results illustrate the applicability of current frac- ture design criteria inhomogeneous reservoirs for fracture design in naturally fractured reservoirs. The importanceofnatural fra~ture anisotropy is investigated in detail by simu- latingthepost-fracture production forvarious fracture lengths andorientations. The effects of permeability impairment to the natural fractures intersected by a hydraulic fracture is illustrated. The reservoir simulation results are supple- mented by field data from the DOE MultiWell Experiment. The results of extensive well testing and reservoir modeling are provided to illustrate the application of the fracture design criteria presented. The field data shows both natural fracture permeability impairment and anisotropy. CONCIXYSIONS 1. 2. 3. 4. 5* Accepted fracture design criteria for homogeneous resemoirs can be applied directlyto isotropic, naturally fractured reservoirs to predict post-fracture well performance and optimum fracture length and conductivity. Well test data may not distinguish natural fracture production in the presence of wellbore storage. In many field applica- tions, a bottomhole shut-in is required to identify natural fracture flow regimes. Natural fracture anisotropy can alter fracture design and interpretation of post-fracture well test data. Assuming a constant hydraulic fracture conductivity, optimum fracture lengths may be shorter in an anieotropic naturally fractured reservoircomparedtoanisotropic naturally fractured reservoir with the same average flow capacity. That assumes the hydraulic fracture isoriented paralleltothe maximum permeability natural fractures. The post-fracture well productivity and present value profit for an anisotropic naturally fractured reservoir (with the assumed hydraulic fracture orientation stated in Conclusion 3) will be less than the corresponding isotropic naturally fractured reservoir. Natural fracture permeability impairment can significantly reduce post-fracture ~ ( . - SPE 17607 C.L. CIPOLLA, P.T. 1 I well productivity and should be minimized and quantified as much as possible. I ACIWOW’XXDG~S This work was sponsored by the United States Department of Energy in conjunction with the Multiwell Experiment. The technical information presented istheproduct of ajoint ●ffort, and the authors wish to thank the CER/MWX field crew, Sandfa National Labora- tories MWX etaff and the CER engineering and computer etaff. I NOMENCLATURE BSHI = bottomhole shut-in C = compressibility, psi-l Cr = dimensionless fracture conductivity h = thickness of formation, ft = Pay Xso = Isotropic Natural Fracture Reservoir k = permeability, md E= average reservoir permeability, md Lf = hydraulic fracture half-length, ft m = Horner slope P = pressure, psi Pi = initial reservoir pressure, psi PI Group = derivative pressure group, [(tp + Del t)/tpl[(dp2/dt)Del t] Del p2 = (shut-in pressure)2 - (last flowing pressure)2 PVP = Present Value Profit, $ Del P = P-P~f Pwf = pressure at the end of flow in well testing, psi q = flow rate, STB/D for oil, MCCFD fOr gas re = external radius, ft rw = wellbore radius, ft S,G. = epeoific gravity of gas T.D. = total depth, ft Tid = Tubing Inner Diameter, in. tp = production time before shut-in, hours Tr = formation temperature, ‘F Wm = distance between orthogonal sets cf natural fractures, ft width of fracture, in. formation volume factor, RB/MCF viscosity, Cp porosity, fraction t = shut-in time, hrs in situ stress, psi !ANAGAN AND S.J. LEE subscripts g - gas hf - hydraulic fracture HO . Homogeneous Reservoir m = matrix nf = natural fracture NF = Naturally Fractured Reservoir min = minimum direction or value max = maximum direction or value ave = average or bulk value s = skin 1. 2. 3. 4. 5* 6. 7. 8. 9. 10. McGuire, W.J. andV.J. Sikora: I’TheEffect of Vertical Fractures on Well Productiv- ity,”d. Pet. Tech. (October 1960), 72-74. Prats, M. J.S. and Levine: “Effect of Vertical Fractures on Reservoir Behavicr- Results on Oil and Gas FIow,!lSPE 593, presented at the 1963 SPE Rocky Mountain Joint Meeting, Denver, May 23-24, 1963. van Poollen, H.K., J.M. Tinsley and C.D. Saunders: ‘IHydraulicFracturing-Fracture Flow Capacity vs Well Prcductivity,~~SPE 890-G, presented at the 32nd Annual Fall Meetingof SPE, Dallas, October 6-9, 1957. Tinsley, J.M., J.R. Williams, R.L. Tiner andW.T. Malone: lWertical Fracture Height -Ite Effect cn Steady-State Production Increase,!!J. Pet. Tech. (May 1979), 633- 638. Agarwal, R.G., R.D. Carter, and’C.B. Pol- lock: $lEvaluationand perfo~ance predic- tion of Low-Permeability Gas Wells Stimu- lated by Massive Hydraulic Fracturing,” J. Pet. Tech. (March 1979), 362-372. Hclditch, S.A.: Criteria For Selecting Propping Agents, 2ndEdition, Norton-Alcoa Proppants, Dallas, 1984. Norman, M.E. and C.R. Fast.:Proppant Mono- graph, General Abrasive Division of Dresser Industries, 1985. Matthews, C.S. and D.G. Russell: Pressure Buildup and Flow Tests in wells, Society of Petroleum Engineers of AIME, Dallas (1967),Volumel (HenryL. DohertySeries). Earlougher, R.C. Jr.: Advances in Well Test Analysis, Society of Petroleum Engi- neers of AIME, Dallas (1977), Volume. 2 (Henry L. Doherty Series). Pirard, Y.M. and A. Bocock: “Pressure Derivative Enhances Useof Type Curves for the Analysis of Well Tests,llSPE 14101, presented at the International Meeting on Petroleum Engineering, Beijing, China, March 17-20, 1986. — _ _——.—.——- .——-—— —— — L1. Bourdet, D., T.If.Whittle, A.A. Douglas Generation High-Strength Proppant in and YOM. pirard: ‘IANew Set of Type ~rves Tight Gas Reservoirs,ltSPE 11633, pre- Simplifies Well Test Analysis,c! world sented at the 1983 SPE/DOE Symposium on Oil, (May 1983), 95-106. Low Permeability, Denver, Colorado, March 14-16, 1983. 12. Houze, O.P., R. Home and H,J. Ramey, Jr.: ItInfinite Conductivity Vertical 23. Penny, G.S.: ‘An Evaluation of the Effects Fracture in aReSerVOir With DOUble POrOS- of Environmental Conditionsand Fracturing ity Behavior,n SPE 12778, presented at Fluids Upon the Long-Term Conductivity the California Regional Meeting, Long of Proppants,” SPE 16900, preeented at Beach, April 11-13, 1984. the 62nd Annual Technical Conference and Exhibition of the Society of Petroleum 13. cinco-L., H., F. Samaniego-V. and N. Engineers, Dallas, Texas, September 27- Dominquez-A.:‘ITransientPressure Behavior 30, 1987. for a Well with a Finite Conductivity Vertical Fracture,w Sot. Pet. Engr. J. 24. Branagan, P.T., c.L. Cipolla, S.J. Lee (August 1978), 253-264. and J. Chen: tlDe8igningand Evaluating Hydraulic Fracture Treatmentein Naturally 14. Holditch, S.A. andR.A. Morse: “TheEffects Fractured Reservoire,ll SPE/DOE 16434, of Non-Darcy Flow on the Behavior of presented at the SPE/DOELcwPermeability Hydraulically Fractured Gas Wells,t’ J. Reservoirs symposium, Denver, Colorado, Pet. Tech. (October 1976), 1169-1179. May 18-19, 1987. 15. Cipolla, C.L. and S.J. Lee: “The Effect 25. Warpinski, N.R. and P.T. Branagan: “Al- of Excess Propped Fracture Height on tered-Stress. Fracturing,ll SPE 17533, Well Productivity,!! SPE 16219, presented presentedatthe SPERockyMounta in Region- at the SPE Production Operations Sy_mpo- al Meeting, Casper, Wyoming, May 11-13, sium, Oklahoma City, Oklahoma, March 8- 1988. 10, 1987. 26. Sattler, A.R., B.L. Gall, C.J. Raible 16. Branagan, P.T., C.L. Cipolla, S.J. Me and POJ. Gill: llFrac Fluid swtfms for and L. Yan: ItCaseHistory of Hydraulic Naturally Fractured Tight Gas Sandstones: Fracture Performance in the Naturally AGeneralCase Study fromMultiwell Experi- Fractured Paludal Zone: The Transitory ment stimulations,!’SPE 17717, presented Effects of Damage,r?SPE/DOE 16397, pre- at the SPE Gas Technology symposium, sented at the SPE/DOE Low Permeability Dallas, Texas, June 13-15, 1988. Reservoirs Symposium, Denver, Colorado, May 18-19, 1987. 27. Northrcp, D.A., et al: “Muitiwell Experi- ment: A Field Laboratory for Tight Gas 17. Branagan, P.T., S.J. Lee, C.L. CiPOlla Sands,liSPE/DOE 11646, presented at the and R.H. Wilmer: llPre-FracInterference 1983 SPE/DOESymposiumonLcwPermeability Testing of a Naturally Fractured, Tight Gas Reservoirs, Denver, Colorado. Fluvial Reservoir,llSPE 17724, presented at the SPE Gas Technology Symposium, 28. Lorenz, J.C., et al: ‘tFractureCharacter- Dallas, Texas, June 13-15, 1988. istics and Reservoir Behavior of stress- Sensitive Fracture Systems in Flat-Lying 18. Holditch, s.A.: ‘lFactorsAffecting Water Lenticular Formations,” SPE 15244, pre- Blocking and Gas Flow From Hydraulically sented at the SPE Unconventional Gas Fractured Gas Wells, II J, pet. Tech. (Decem- Technology Symposium, Louisville, Ken- ber 1979), 1515-1524. tucky, May 18-21, 1986. 19. prat-, M.: llEffectof vertical Fractures 29. Lorenz, J.c.: ‘Jsedimentologyof the Mesa- on Reservoir Behavior - Incompressible verde Formation at Rifle Gap, Colorado, Fluid Case,f’ Sot. Pet. Engr. J. (June and Implications for Gas-Bearing Intervale 1961), 105-118. in the Subsurface,t’Sandia Report, March 1982. 20. Branagan, P.T., C.L. Cipolla, S.J. Lee and R.H. Wilmer: ItComprehensive Well 30. Lorenzt J.C.: llpredictionSof Size and Testing and Modeling of Pre- and Post- Orientations of Lenticular Reservoir in Fracture Well Performance of the MWX the Mesaverde Group, Northwestern COlOr- Lenticular Tight Gas Sands,’~ SPE/DOE ado,!! SPE/DOE 13851, presented at the 13867, presented at the SPE/DOE 1985 Low SPE/DOE Symposium on Low Permeability Permeability Gas Reservoirs, Denver, Reservoirs, Denver, Colorado, May 19- Colorado, May 19-22, 1985. 22, 1985. 21. Cooke, C.E. Jr.: tlEffeCtof Fracturing 31. Warpinski, N.R., et al: “Fracturing and Fluids on Fracture Conductivity,’8 J. Testing Case Study of Paludal, T:ght, Pet. Tech. (October 1975), 1273-1282. Lenticular Gas Sands,!’ SPE/DOE 13876, 22. Callanan, M.J., C.L. Cipolla and P.E. presented at the SPE/DOE Symposium of Low Permeability Reservoirs, Denver, Lewis: :!TheApplication of a New Second- Colorado, May 19-22, 1985. - . -Tnnrra m ml nnauacau aun Q 7. TX%! 9 — ; 17607 U.k. b&rvum? =*. ● =— -=- —- Table 1 Base Reservoir Simulation Input Data Table 2 Input Data for Hydraulic Fracture Simulatim?s* Common Data Hydraulic Fracture Propartias Pi = 4,000 psi khf = 25,000 md Tr = 24CF F Whf = 0.12 in. Tid = 2.441 in. T,D. = 6,000 ft Fracture Langtfr, ft Cr h . 40 ft 400 10 P . 0.0214 Cp @ pi 0.7862 ResBBUMCF @ Pi 800 # . 1,200 3!3 Cg = 0.000196 psi-l @Pi 1,600 2.5 Homogeneous Reservoir Anisotropic Natural Fracture Propartias kln 0,02 md 0.05 10:1 anisotropy kni .nin = 626 md knf ma% = 6,231 md N@urally Fractured Fksanroir ●All other input data same as in Table 1‘s Natural Fracture A B Case A km 0.0002 md 0.015 md ‘$m 0.05 0.05 . knf 1,980 md 500 md 4mf 0.5 0.5 Table 3 Base Economic Input Data Wnf 0.0006 in. 0.0006 in. Wrn 5 ft 5 ft Stimulation and Initial Wall Test Data Lf, ft Imrestmant Cost Common Data 240 hrs drawdown@ 100 MCFD o $360,000 Gas Price = 1,5 $/MCF 1,000 hrs shut-in – BHSI & Surface Shut-In 400 392,000 Price Escalation = None Soo 450,000 Operating Cost =“800 $/Mo 1,200 590,000 Discount Rate = 10% 1,600 860,000 Net Rev. Int. = so% Working Int. = 100% Tabla 4 Pra-Fraclure Model Input Data for MWX Paludal History Match Basa Reservoir Data Matrix Properties Natural Frectura Properties Channel width = 350 ft km=l,O~d knf mex = 5,000 md Pi= 5,400 psi @m= 0.04 knf min = 500 md T.D. = 7,000 ft rjnf = 1.0 Tr=2100F Wm=5ft h=40ft Wnf = 0.001 in, S.G, = 0.626 skin, 2.5 ft in y-direction p = 0.02 Cp ks= 100md Table 5 Ra-Entry Model hrput Data Hydraulic Fracture Base Raservoir Data Matrix Properties Natural Fractura Properties Properties Channal width= 350 ft km=l.Ogd knf max = 5,000 md Khf = 5 darcy Pi= 5,400 psi @m= 0.04 (4%) knf min = 500 md Whf = 0,25 in. T.D. = 7,000 ft @nf = 1.0 Lf=100ft Tr” 21@ F Wm=5ft h=40ft Wnf = O.OO1 in. S.G. = 0.626 — FRACTURE DESIGN CONSIDERATIONS IN NATURALLY FRACTURED RESERVOIRS SPE 17607 L 0’ r- . dlW~,dWfJdlOfJ [...]... of Naturally Fractured and Homogeneous Resarvoim With Wellbore Dama@ – Minimal Wellbore storage C.L CIPOLLA, P.T BRANAGAN AND S.J SPE 17607 g , 1 I q : I , I q $ q I 1 I I I I I I I 1 1 I I I I I ! ! I I I I I [ I 1 1 1 1 1 1 1 , , I I I I 1 1 I I I I I 1 I 1 I I I I I 1 I I I I I I I I I I I I I I ~wl I I I I , 1 * z a ; I I 1 IJJ I: I I : I I LEE 13 ., FRACTURE DESIGN CONSIDERATIONS IN NATURALLY. .. Profit for isotropic and Anisotropic Reservoirs, L f = O, 4&W, 8# and l,2iM ft Comparison of Post -Fracture Production With and Without Natural Fracmra Permeability Impairment 4,0m 3,000 \, & t; 2,0m \ l.ooa * n a 10 lm Honnr Fi~ra * l,mo Siwt -in Tim., 11 TInm ?1 Homer Plot of Pressure Buildup Behavior of Natutally Fractured and Homogeneous Reservoin With Wellbon? Damage – Minimal Wellbort? Storage Fi@e 12... I [ I 1 1 1 1 1 1 1 , , I I I I 1 1 I I I I I 1 I 1 I I I I I 1 I I I I I I I I I I I I I I ~wl I I I I , 1 * z a ; I I 1 IJJ I: I I : I I LEE 13 ., FRACTURE DESIGN CONSIDERATIONS IN NATURALLY FRACTURED RESERVOIRS 14 SPE 17607 lsd’WHlss3iid \ i! lSd ‘3 HfK3S3Ud adow ‘MOld s ~ 0’ 0 0 0“ 600 q  C.L CIPOLLA, P.T BRANAGAN AND S.J LEE SPE 17607 15 — - o 0 0 e , I , , u I I I I I I1 I1 I1 1 III I1 I1 . fracture length, fractureconductivityandwell product- ivity, and the economic impactof many fracture design considerations. However, fracture designconsiderations inmorecomplex ,natural- ly fractured. fracture flow regimes. Natural fracture anisotropy can alter fracture design and interpretation of post-fracture well test data. Assuming a constant hydraulic fracture conductivity, optimum fracture. simulations and field data that illustratemany fracturedesignconsideratims in naturally fractured reservoirs. The initial requirement for designing a hy- draulic fracturing treatment is an accurate description