You are on page 1of 5
In J-Rock Mach, Min. Sc & Gromech. Abr, Wo. 3, No.6, pp 627-631, 1983 Pinte in Great Bian, Al ights reserved Technical Note 0148:906293 58.00 + 0.00 Copyright 1993 Pergamon Press Ld Interpretation of Hydraulic Fracturing Pressure: A Comparison of Eight Methods Used to Identify Shut-in Pressure F. GUOt N. R. MORGENSTERN} J.D. SCOTT} INTRODUCTION The minimum principal stress is an important factor in the design of underground openings and massive hydraulic fracturing. Hydraulic fracturing is one of the ‘most promising methods for determining the minimum, principal stress, especially at great depth. In hydraulic, fracturing stress measurements, the minimum principal stress is assumed to be equal to the shut-in pressure. This, concept is attributed to Kehle [1]. One important premise of this concept is that leak-off into the formation, is negligible. As hydraulic fracturing practice illustrates (2,3), an indistinct shut-in pressure appears if leak-off is, not negligible. To deal with the indistinct shut-in pressure, numerous methods have been put forward. A comparison among various methods has also been made by Ageson er al. [4] and Proskin er al. [5]. However, the comparison was performed on field hydraulic fracturing data, The minimum principal stress was unknown in, these tests, and it is difficult to obtain a persuasive conclusion from the comparisons. Considerable shut-in pressure responses were ob- tained in a laboratory single-well hydraulic fracturing, program [6,7]. All the curves yielded indistinct shut-in pressure because high leak-off was incorporated in the tests. Therefore, the data provide an opportunity to evaluate the analysis techniques to deal with indistinct shut-in pressure. This Technical Note compares the pressure evaluated by eight different methods the minimum principal stress applied to a specimen. SHUT-IN PRESSURE IDENTIFICATION TECHNIQUES Various investigators have proposed methods to identify the indistinct shut-in pressure. Proskin [8] summarized the various methods developed by hydraulic fracturing practitioners to determine the shut-in pressure. Eight of these methods will be described in the following paragraphs. tof Civil Engineering, University of Calgary, Calgary, ‘Canada T2N 2N4 {Department of Civil Engineering, University of Alberta, Edmonton, ‘Alberta, Canada T6G 267 Inflection point method The inflection point method suggested by Gronseth and Kry [9,10] is a simple graphical technique. The construction consists of drawing a tangent line to the pressure-time record immediately after shut-in. The pressure at which the pressure-time record departs from the tangent line is defined as the shut-in pressure. This, method was suggested to interpret low-rate hydraulic fracturing data (<50 l/min). Figure | illustrates the application of this method. Po vs log(t + 4i)/ dt Method McLennan and Roegiers [11] suggested that the inflection point (a slope change) in the plot of p, vs log(s + Ar)/Ar represents the shut-in pressure, where p, is the bottomhole pressure, f is the time of injection and At is the time since shut-in. Figure 2 shows how to determine the shut-in pressure from the plot of py vs log(s + Ar)/Av. py vs log dt Method Doe and Hustrulid (12] obtained the shut-in pressure using the plot of p, vs log Ar for the period immediately following the first breakdown, where Ar is the time since shut-in, The shut-in pressure corresponds to a break in the slope of the plot. The method was recommended for interpreting hydraulic fracturing under slow pumping, cycles. Figure 3 demonstrates the application of this method. log(p.— p,) vs At Method (Muskat method) ‘Aamodt and Kuriyagawa [13] thought that the press- ure after shut-in approaches some value asymptotically. ‘A trial value for this asymptotic pressure, p, is chosen and the logarithm of the pressure p. minus p, is plotted against time. p, is varied until the curve, after an initial transient, is best fitted by a straight line. The straight line is extrapolated back to the time of shut-in, giving a pressure. Then, this pressure plus p, is taken as the shut-in pressure. This method is shown in Fig. 4. log p. vs log Method As stated by Zoback and Haimson (14), Haimson recommended selecting the shut-in pressure from the or GUO erat. ‘TECHNICAL NOTE 400) 1200) 600 Bonanbole ese (MP) 400 200 . 0 » wo 0 @ ae) Mydeaulic Fracture Test 3 Shatin in Cyclet Fig. |. Inflection point method. plot of logp, vs log’, where p, is the bottomhole pressure, and ¢ is the time since pumping. The pressure vs time curve in this plot is bilinear. The shut-in pressure is the intersection of the bilinear lines. Figure 5 illustrates this method, plat vs p, Method Tunbridge [15] assumed that the shut-in curve is bilinear in the plot of dp,/dt vs p., where p, is the bottomhole pressure. The intersection of the two lines corresponds to the shut-in pressure. Figure 6 is a plot of dp, dt vs p, and the corresponding shut-in pressure. 0s \/di Method Fracture linear flow will_lead to a linear relation between the pressure and ,/Ar. Therefore, when the plot of p, vs \/At departs from a straight line, the fracture closes. The corresponding bottomhole pressure is the fracture closure pressure of the shut-in pressure (16) This method is illustrated in Fig. 7 000 aw Hydraulic Fracture Test 3 Shatin i Cycle Fig 3. p, vs log Ar method. Maximum curvature method The bottomhole pressure at the point of maximum curvature in the shut-in curve is also recommended as the shut-in pressure [17]. Figure 8 illustrates the shut-in pressure determined through the maximum curvature method, RESULTS AND DISCUSSION ‘The detailed experimental procedure and results have been reported by Guo etal. [6,7] and will not be repeated here. The shut-in pressure obtained by various methods and the actual minimum principal stress applied to the specimens are summarized in Table 1 Inflection point method. This method always produces a high shut-in pressure, especially for a low stress. For example, it gives a shut-in pressure of 8.0-9.5 MPa (1160-1378 psi) in Test 16, but the least principal stress was only 2.92MPa (424 psi). The shut-in pressure obtained by this method is 8.0-9.0 MPa (1160-1305 psi) an 11090 g 1500 z g ' E sa00 j pe70Nm s.00 ao ect i © 700 ‘wo 7” % wane ay Hydraulic Factore Test § Shatin in Cyete-t Fig. 2 p, v6 loge + 1 method, Mydraulle Fracture Test § Shutin in Cyees3 Fig. 4 togta.~p,) v6 A method (Muskat method), Guo eras TECHNICAL NoTE 7 - 7 oe ts 7 J nal i I i estos 7 7 aaasied \ = 2.00} 10! 0.00! Stee pe 7 Ee ee ee ee eee 0, var ois) re ene ee cee Fig. 5. log, vs log method. in Test 5; while the actual least principal stress was 7.0 MPa (1015 psi). This method depends strongly upon, the timescale used in the plot. When the timescale used. expands, the shut-in pressure determined by this method increases. The reason is that the slope of the pressure vs, time curve is not constant after immediate shut-in, and. it increases rapidly with time (Fig. 9). The shut-in pressure determined by this method is subjective. D. 05 log(t + At)|4t Method. This method is promis- ing. It produces a slightly high shut-in pressure. For example, Test 3 had a 0; of 1.9 MPa (276 psi) and the shut-in pressure obtained by this method is 2.6 MPa (377 psi). Test 5 has a 0; of 7.0 MPa (1015 psi), and the shut-in pressure is 7.0 MPa (1015 psi). Test 16 has a a3, of 2.92 MPa (424 psi) and the shut-in pressure is 3.8 MPa. (S51 psi). Under various stress conditions the linear future of p, vs log(t + Ar)/A1 is obvious, and the shut-in, Pressure can easily be identified. Pe ts log dt Method. This method, like p. vs log(t + Ar)/At, produces good results. Test 3 has a 0, of 1.9 MPa (276 psi) and the shut-in pressure obtained by Fig. 7p, vs Ar method, this method is 3.0 MPa (435 psi). Test 5 has a 0, of 7.0 MPa (1015 psi) and the shut-in pressure is 6.5 MPa (943 psi). Test 16 has a 0, of 2.92 MPa (424 psi) and the shut-in pressure is 3.0 MPa (435 psi). However, it should be pointed out that the p, vs log Ar curve is not bilinear. In this analysis, the shut-in pressure is considered as the pressure at which the p, vs log Ar curve departs from a straight line, Log(p,—p,) ts dt Method. This is also a promising method. It is capable of providing a reasonable a. as illustrated in Table 1. It should be stressed that the shut-in curve from the frst injection cycle usually under- estimates the shut-in pressure. For example, while Test 5 has a a, of 7.0MPa (1015 psi), the shut-in pressure from the first injection cycle is only 1.0 MPa (145 psi) The subsequent injection cycles usually give good esti- mations of a. For example, the shut-in pressure from the second and third injections in Test 5 is 7.0 MPa (1015 psi). They are equal to . If the injection rate is, low, the shut-in curve after several injection cycles can produce a shut-in pressure close to a; as shown by Test, jonas) P04) Hydraulic Fratere Test § Shatin iw Cycle? Fig, 6. dp. dt vs py, method, ” é | - Ley Ah otal ML AL VV Bororate Pressre (MPs) cytes Fig. 8 Maximum curvature method. 630 GUO eral Table Method Gee 12 3 Test 3 0 = 190MPa Inflection point 48 prslogt+Aryar 26 ps log. Ar 30 Toate = p,) vs Ar oe logp vs logy 20 dp ar vs p ase pis var 5324 Maximum curvature 4646 Test 5.0; = 200 MPa Inflection point 30 99 99 ps log =Aniar 7.0 vs log At 6S foatp ~ p,) vs Ar Lo 70 70 logp vs log a5 dpiar vs p 10 68 65 pis Vat £0 90 85 Maximum curvature =? 5670 ‘Test 9. 0) 640MPa Inflection point 90 2s vs logit + Arar vs log Ar loatp ~ p.) vs Ar 69 10 logp ws loge apidi vs p 10 38 ps var 1335 Maximum curvature sa 74 Test 6, 0) =292MPa Inflection point 80 80 80 pos log Arar 38 vs log ar 30 Toate — py) vs 8 lo 12 10 logp vs log 20 apidr vs p 33 40 48 pis var gaa Naa Maximum cunature 3930 4d Note 16. The shut-in curve after the fourth injection cycle gives a shut-in pressure close to 0 Log p, vs logt Method. This method produces few significant results. The plot of log p, vs log is usually a smooth curve. Itis difficult to draw a bilinear relation, Therefore, the shut-in pressure is considered the pressure at which the curve departs from the line, As shown in Table 1, the results are not stable. For some tests like 1.0 200 200 ry) ep phat 600 109 00 Ot mse an aan een an saaaten ea Fig. 9, Hydraulic fracture Test 16: shut-in in Cycle | TECHNICAL NOTE pressu suggested not suggested ot suggested not suggested not suggested sugested suggested suggested ot suggested suggested rot suggested ot suggested ot suggested suggested suggested suggested suggested not suggested suggested suggested Suggested suggested not sugested not suggested not suggested 40 5944 39 aoe 427148 teens that no distinct Feature wo that the shutin pressure cannot be determined. Tests 3 and 16, the shut-in pressure obtained is close to oy, However, for the other tests like Test 5, the shut- pressure obtained is meaningless. dp, dt vs p,. Method. For high stress, this method provides a reasonable shut-in pressure. For example, Test 5 has a of 7.0 MPa (1015 psi), while the shut-in pressures obtained by this method are 7.0, 6.5 and 6.5 MPa (1015 and 943 psi). Test 9 has a 0, of 6.4 MPa (928 psi), and the shut-in pressures are 6.9, 7.0 and 6.8 MPa (1001, 1015 and 986 psi). The bilinear feature is, distinct. However, this method produces a high shut-in pressure under ow stress, as happened in Tests 3 and 16. In addition, the bilinear feature is sometimes indistinct under a low stress. DP. 0S s/t Method. This method sometimes produces a reasonable shut-in pressure. As in Test 9 and the second shut-in of Test 3, the shut-in pressure is close to oy. However, some shut-in curves have no linear period as in Test 16, so the shut-in pressure cannot be determined. In addition the bilinear features are not distinct. Therefore, the shut-in pressure is regarded as the pressure at which the curve departs from the linear segment. ‘Maximum curvature method. Although this method hhas an exact definition, it produces poor results. The main error may result from the calculation of the GUO eral: TECHNICAL NOTE curvature. The curvature depends on the second deriva- tive of pressure vs time, so a little noise on the pressure data will lead to a large fluctuation in curvature. This fluctuation often masks the true maximum curvature, Therefore, while the definition is clear, the determination Of shut-in pressure is still subjective. In conclusion, the shut-in pressure obtained from ight methods has been compared with the minimum, principal stress. The results show that the p, vs log(t + Ar)/iAr method, the p, vs log At method and the log(p. —P,) vs At method obtain the shut-in pressure the shut-in curve following the first injection cycle is used in the plot of p, vs log(t + At)/At and the plot of p, vs log Ar, and the shut-in curves in the subsequent injection cycles are used in the plot of log(p,—p,) vs At. The log p, vs log ¢ method, the dp, {dt vs At method, the p, vs \/At method and the maximum curvature method can sometimes obtain a shut-in pressure close to o;. However, the results are unstable. For some tests the results are meaningless. The inflection point method is also subjective because of non-linear shut-in curves immediately after shut-in. Acknowledgements —The authors wish to acknowledge the Natural Sciences and Engineering Research Council of Canada and Shell ‘Canada Lid for the financial support for this projec. ‘Accepied for publication 23 April 193. REFERENCES |. Kelle R_O. The determination of tectonic stresses through analysis of hydraulic well fracturing. J. Geophys. Res, 69, 259-273 (968) 2. Hydraulic Fracturing Stress Measurements Proceedings ofa Work shop on Hydraulic. Fracturing Stress Measurements. National ‘Academy Press, Washington, DC (1983) 3. Haimson BLC., Roegiers J. C. and Zoback M.D. Hydraulic Fracturing Sires Measurements, Proceedings of the Second tnt Workshop on Hydraulic Fracturing Stress Measurements, (1988) 4 6a ‘Aggson J. R.and Kim K. Analysis of hydraulic facturing pressure histories: a comparison of five methods used to identify shut-in pressure, Int. J. Rock Mech, Min. Sci. & Geomech, Abstr. 24, 75-80 (1987). Proskin S. A, Scott J. D. and Chhina HS. Interpretation of the ‘minimum priccipal stress ftom microfrac tests. Rook at Great Depih, Proc. In. Symp. on Rock Mechanics and Rock Physics at Great Depth, Pau, France, pp. 1509-1519 (1990), Guo F.. Morgenstem N. Ry and Scott J. D. An experimental investigation into. hydraulic fracture propagation Part | Experimental facilities. nt. J. Rock Mech. Min. Sci. & Geomech Abstr. 30, 17-188 (1993), Guo F.. Morgenstern N. R. and Scott J. D. An experimental investigation into hydraulic facture propagation—Part 2. Single Well tests. Int. J. Rock Mech, Min. Sci & Geomech. Abstr. 30, 189-202 (1993), Proskin S, Jn stu stress determination in oil sands. M.Se. Thesis, Department of Civil Engineering, University of Alberta, Canada (1989) Gronseth J, M. and Kry P. R. Instantaneous shutin pressure and its relationship tothe minimum i stu sess, Proc. Workshop om Hydraulic Fracture Stress Measurements, Monterey, CA, pp. 55-60 (1981), Gronseth J. M. Determination of the instantaneous. shutin pressure from hydraulic featuring data and its reliably a8 ‘measure of the minimum principal stresk Jesus in Rock Mechanics, Proc. 23d US. Symp. on Rock Mechanics, pp. 183-189, Univesity of California, Berkeley (1982) McLennan J. D. and Roegiers J-C. Do instantaneous shut-in Dressures gocurately represent the minimum principal stress. Proc Workshop on Hydraulle Fracture Siess Measurements, Monterey, CA, pp. 68-78 (1981. Doc W. T. and Husirulid W. A. Determination of the state of ‘ress at the Stripa Mine, Sweden. Proc. Workshop on Hidraule Fracture Siress Measurements, Monterey. CA, pp. 119-129 (1981). ‘Aamodt R. Land Kuriyagawa M. Measurement of instantaneous shutin pressure in erystlline rock. Proc. Workshop on Hydraulic Fracture Siress Measurements, Monterey. CA. pp. 139-143 (198). Zoback M.D. and Haimson B.C. Status of the hydraulic fracturing method for in ste stress measurements. sues in Rook Mechanics, Proc. 2rd U.S. Symp. on Rock Mechanics, University ‘of California, Berkeley, pp. [43-186 (1982). ‘Tunbridge L. W. Interpretation of the shut-in pressure from the rate of pressure decay. Int. J. Rock Mech, Min. Sc. & Geomech. Abstr. 26, 457-489 (1989), Sookprasong P. A. Plot procedure finds closure pressure. Oi! & Gas J. 84, 110-112 (1986), Hayashi K_and Sakurai I. Interpretation of hydraulic fracturing shutin curves for tectonic stress measurements. Pn. J. Rock Mech, Min. Sei. & Geomech. Abstr 26, 477-482 (1989),

You might also like