Creep Behavior of Barnett, Haynesville, and Marcellus Shale

LI Yon,Yawei

2012-01-01

Abstract:In order to assess the contribution of creep to closure rate and conductivity loss of hydraulic fractures in gas shale, the visco-elastic characteristics of gas shale have been investigated. A series of creep tests were conducted on gas shale core samples. First, a few uniaxial creep tests were performed on several selected samples, and then multi-stage triaxial creep tests were carried out at room temperature. Samples used in the tests come from three different gas shale reservoirs. It is found that the creep strain can be described by a power-law function of time. The clay and carbonate contents of these shale samples vary noticeably, and it is observed that rocks with more cement and less clay have higher elastic modulus, Pseudo-steady state creep rate increases linearly with deviator stress and higher confining pressures increase the amount of creep strain under the same deviator stress. sets of test have been conducted at room temperature. However, test under reservoir temperature conditions are ongoing and will be reported at a later time. 2. EXPERIMENT DESCRIPTION 2.1. Sample Description Plugs from Barnett, Marcellus and Haynesville have been tested. The Barnett sample 3-20 is from a depth of 5552.35ft, while the Marcellus sample 09-30 is from a depth of 5899.48ft, Haynesville samples 05-21 and 0522 come from depths of 12424.75ft and 12454.55ft. The mineral content of the four samples and their elastic modulus are listed in Table 1. The samples were prepared for testing by TerraTek-Shlumberger core preparation laboratory. All samples were drilled perpendicular to apparent bedding. All samples were kept under room humidity condition prior to the test and experiments were conducted under dry and drained condition, in this way we were able to eliminate poro-elastic effects, so the data represents the behavior of the dry rock. All samples are 1” in diameter and they all have the standard 1 by 2 aspect ratio except Marcellus 09-20 (2.09” length). Table.1.Mineral contents of the four samples, QFP refers to quartz, feldspar, plagioclase, and pyrite component Sample name QFP % Carbonate % Clay % Others % Marcellus 09-30 5899.48ft 28 9 40 24 Barnett 3-20 5552.35ft 60 10 25 5 Haynesville 05-21 12454.75ft 25 23 33 18 Haynesville 05-22 12454.75ft 25 23 33 18 Table.2. Young’s modulus of the four samples (perpendicular to bedding) at different confining pressures Sample name Confining pressure MPa Young’s modulus GPa Marcellus 09-30 19 17.2 16 15.5 5 12.7 Barnett 3-20 16 32.1 0 27.8 Haynesville 05-21 0 14.5 Haynesville 05-22 16 17.6 Fig.1.The four shale samples used for the creep test 2.2. Experimental Procedures All the creep tests were performed on GCTS triaxial rock test system, two additional Teledyne syringe pumps were used to provide constant confining pressure and axial load, axial load is applied to the samples with a hydraulic cylinder (hydraulic jack). The loading paths of the four samples are listed in Table 2. After installing the sample and displacement sensor assembly in the confining cell, a constant confining pressure is applied first, and then the sample is loaded with a desired deviator stress (constant) for a desired period of time. The deviatoric stress is then increased to a higher value if required, and is held constant. During this process, the deformation of the sample is monitored and recorded. A typical multistage loading path & the axial strain response are illustrated in Fig.4. Fig.2. Hydraulic cylinder for axial load, schematic drawing of the creep test setup. Table.3. Loading path of the four samples Sample name Confining pressure MPa Deviator stresses MPa Marcellus 09-30 16 27-36 19 55-64-72-81-89 5 64 Barnett 3-20 16 46.5-53.5-63-73-83 0 45-55-64-72-81 Haynesville 05-21 0 34-45-63.5-72 Haynesville 05-22 16 36.4-55-64 3. TEST RESULT AND INTERPRETATION For the Marcellus 09-30 sample, the first applied confining pressure is 16 MPa and was applied using the GCTS pressure booster. For the first two stages, the confining pressure was around 16 MPa but in the middle of the two stages test it gradually increased to 19 MPa (not planned). To maintain consistency, the confining pressure of all later stages was kept at 19 MPa using a more stable Teledyne ISCO pump. Fig.3.The first two stages of creep tests on Marcellus 09-30. Fig.4. The loading path & the axial strain of Marcellus 09-30, 3 to 7 stage. Fig.5.The axial creep strain and strain rate of Marcellus 09-30, the 7 stage. For the creep strain of a constant stress stage, the strain rate dt d / 1 e decreases with time until a constant value is reached. However, the strain-time curve does not completely level off. A similar result has been reported by Zoback et al [1] for unconsolidated GOM shale that was tested for 6 days. Thus, we only use the last portion of the strain-time curves and consider it as a quasisteady-state strain. The strain rates of the 1 to 7 stage of creep test were obtained by fitting a strain line to the last portion of the strain-time curves, as shown in Fig. 5. The strain rate-stress relation is plotted in Fig. 6. Fig.6.Strain rate-stress relation for Marcellus 09-30 at a confining pressure of 19 MPa, As shown in Fig.6, the quasi-steady strain rate increases almost linearly with deviator stress. At higher deviator stress, the trend deviates from a linear relation indicated by the last point in Fig.6, implying a non-linear strain rate-stress relationship at higher deviatoric stresses; higher deviatoric stresses are not applied for fear of breaking the sample. In the 0-80 MPa range, the creep can be modeled using a linear visco-elastic model such as Burgers and Maxwell. But because the Maxwell model is not able to capture the transient creep portion of the strain-time relation, Burgers model is used to model this multistage creep. Because the confining pressure of the first two stages jumped from 16MPa to 19 MPa, the first two stages are neglected when fitting the data with Burgers model. A least-square fitting method was used to fit Burgers model (Eq.1) to the strain-time curve obtained from 3 to 7 stage creep, and the four parameters are summarized in Table.4. Fig.7. Four element Burgers model.         − + + = − t E