1st Workshop on Petroleum Geomechanics Testing

The Westin Michigan

Avenue, Chicago

June 23, 2012

Organized by the ISRM Petroleum Geomechanics Commission

Tony Addis, Russ Ewy, Axel Makurat, Maurice Dusseault (PGC Chair)

And by the American Rock Mechanics Association

Antonio Bobet (Vice President), Peter Smeallie (Executive Director)

PRESENTATIONS

TESTING METHODS AND PROTOCOLS

Effects of pre-stressing, sample size and glued instrumentation on Red Wildmoor Sandstone

Magnus Soldal, NGI

Geomechanical test protocol stages that help address sample-to-sample variation assessment

John Dudley, Shell

Testing of oil sands and their associated lithologies

Rick Chalaturnyk, U of Alberta

Effect of wettability and surface tension of pore fluid on the mechanical strength of chalk: theory vs. experiment

Alexander Rozhko, M-I Swaco

SHALES AND MUDSTONES

Shale characterization and test protocols Jørn Stenebråten, Sintef

Geomechanical testing of the mudstone caprock above injection processes in oilsand reservoirs

Patrick Collins, Petroleum Geomechanics, Inc.

Testing the thermohydromechanical behaviour of shales

Lyesse Laloui, Ecole Polytechnique Federale de Lausanne

SPECIALIZED TESTING

On dilatancy, friction, and softening response of rock

Joe Labuz, U of Minnesota

Block testing with polyaxial stresses, and fluid flow-coupling

Rico Ramos, ConocoPhillips

Measurement and application of effective stress coefficient

John Shafer, consultant

Test methods for creeping rocks Paul Hagin, Chevron

DAMAGE & PRESERVATION, and SANDSTONE ANALOGUES

Core and sample handling and damage, and Experience with analogue materials

Mike Myers, Shell

Use of synthetic rocks for quantification of coring induced rock alteration

Rune Holt, Sintef / NTNU

INDEX TESTING

Enhanced rock strength profiling, combining triaxial compressive strength, non-destructive core scratching and index tests

Abbas Khaksar, Baker Hughes

Effects of pre-stressing, sample size and glued instrumentation on Red Wildmoor sandstone

Magnus Soldal (Norwegian Geotechnical Institute), Toralv Berre (Norwegian Geotechnical Institute)

Lately NGI’s rock mechanical laboratory has seen a growing interest in testing smaller rock samples in the triaxial cell, and to measure radial acoustic velocities while doing so. At the same time, internal strain should also be measured. This calls for much instrumentation directly onto the specimen surface, and the effects of this are studied in a series of triaxial tests on Red Wildmoor sandstone.

For decades it has been common procedure at the NGI’s rock mechanical laboratory to measure internal strain during triaxial rock testing. The internal horizontal strain is measured along a diameter at the lower third point and upper third point of the specimen height, and the internal vertical strain is measured as the change in distance between the two horizontal strain sensors. A horizontal strain sensor consists of a closed, very light metal ring which encloses the specimen. A spring loaded core of a LVDT is fixed to the ring and is acting radially with respect to the specimen, against two metal knobs. The knobs may be glued directly onto the surface of the specimen or they may be glued only to the confining membrane. Either way they go through the rubber membrane and act directly towards the surface of the specimen.

For tests where the effective horizontal stress changes during the important part of the test (e.g. stages with K0-consolidation), the knobs are normally glued to the specimen to minimize bedding effects between knob and specimen. For specimens loaded to failure, “loose” knobs which are glued only to the rubber membrane are preferred. Gluing the knobs to the surface of the specimen is considered to give less false deformation between the knob and the surface of the specimen. However, there seem to be a tendency for the material just below the knobs to expand somewhat more horizontally than the rest of the specimen when compressing the sample vertically.

Over the last years, the rock mechanical laboratory has more frequently been asked to measure radial velocity during triaxial testing. Previously this was done by mounting piezocrystals inside the rubber membranes. Due to high degree of wearing on the crystals during shearing, it was decided to put the crystals inside small metallic sensor houses. In this way, the crystals are spared even when the specimen is sheared to the point where membrane damage occurs. To improve the mechanical coupling between specimen and sensor, and hence the signal quality, the sensors houses may be glued directly onto the specimen surface.

Due to low availability of good quality “offshore rock cores”, it has also become more common to perform triaxial testing on smaller rock samples (i.e. 25 mm instead of 38 mm diameter). The instrumentation used on a 25 mm diameter specimen occupies a significantly larger portion of the specimen surface than it does for a 38 mm diameter specimen.

In this presentation some results of a triaxial test series on Red Wildmoor sandstone are shown. The tests contain isotropic pre-stressing, K0-consolidation and shearing, and are performed on both 25 and 38 mm diameter specimens with both glued and “loose” instrumentation. The main focus areas are:

- The potential of pre-stressing as a means to minimize bedding effects.

- The mechanical response of 25 mm diameter specimens compared to 38 mm diameter specimens.

- The mechanical response using instrumentation glued onto the specimen surface compared to instrumentation glued only into the confining membrane.

Geomechanical test protocol stages that help address sample-to-sample variation assessment JWDudley, Shell Int. E&P

Abstract: Geomechanical characterization often requires multiple different tests (e.g., uniaxial-strain, triaxial and isotropic compression) on a suite of ideally similar plugs. Even plugs that seem quite similar before testing (from visual/CT inspection, bulk density, etc.), however, may not all act mechanically similar for a variety of reasons. A suite of triaxial compression tests, for instance, may show one or two tests that are unusually strong/stiff (or weak/soft) relative to the other tests. Often only the triaxial loading stage stress-strain data are reported, and it is difficult to objectively compare the various results to understand the difference. Therefore when testing such a suite of plugs it is useful to perform representative measurements under identical conditions during each test for just such a relative comparison. Adding these specific measurement stages to the test protocols of the suite provides a mechanical measurement on each plug that can be helpful in understanding the aggregate test suite results.

A confining pressure unload-reload cycle for determining the bulk modulus can often serve this purpose well. For example, a triaxial compression test suite at pressures 3, 6, 10, 15 MPa, can have a 2 MPa confining pressure unload-reload cycle at 3 MPa during the isotropic compression stage of all the tests. The bulk modulus determined from these cycles provides a directly comparable measurement made under the same conditions on each sample, which can highlight anomalous behavior/samples. The same test stage can be added to other test types (isotropic compression, constant-mean-stress, uniaxial-strain) that at least begin with an isotropic compression to at least 3 MPa, and so extend the comparison across all tests used for developing or calibrating a constitutive material model.

EFFECT OF WETTABILITY AND SURFACE TENSION OF PORE FLUID ON THE MECHANICAL STRENGTH OF CHALK: THEORY VS. EXPERIMENT

Alexander Y. ROZHKO Schlumberger Norge, Division: M-I SWACO

Email: arozhko01@slb.com Abstract Previous experimental studies have demonstrated that mechanical properties of high porosity chalk are

strongly dependent on the type of fluid in the pores and wettability of the chalk, which could be altered

during the experiment.

Risnes and Flaageng [1] pointed out that wettability is an important parameter in fluid-rock interaction

and demonstrated that water flooding of oil-wet (hydrophobic) oil-saturated chalk samples show

strengthening effect, while water flooding of water-wet (hydrophilic) oil saturated samples show the

weakening effect. It has been observed that water-saturated chalk is mechanically weaker than methanol

saturated chalk, while the methanol saturated chalk is mechanically weaker than oil saturated chalk [2,3].

The correlation between tensile strength of limestone and fluid-air interface tension of various pore fluids

were investigated [4], and found that the tensile strength is decreasing with increase of surface tension.

The effect of water-glycol mixtures on chalk strength were investigated by [3], and demonstrated that

chalk strength is increasing with increase of glycol content in the water-glycol mixture.

Many theories attempted to explain fluid-chalk interaction mechanism; however, the underlying

mechanism is still uncertain. It was argued [3] that capillary force is not relevant for fluid-chalk

interaction, because of the following reasons: 1) Since the water and glycol are fully miscible, thus, no

interface meniscus and hence, the capillary force is zero; 2) it was demonstrated that the capillary force

between two spheres is insufficient comparing to changes in strength observed during chalk-fluid

interaction [3].

In this paper instead of conventional grain-contact approach a novel fracture mechanics approach is

considered. The effect of capillary force between fracture faces on the stress intensity factor is calculated.

It is demonstrated that a fluid molecule cannot penetrate to the crack tip, because the radius of curvature

of the crack tip opening at critical stress intensity factor is smaller than the size of a water molecule. Thus,

crack tips inside chalk remain dry allowing the possibility to prescribe the geometry and location of

interface meniscus between fluid and gas (air or vapor). In calculations of the capillary force it is

considered that small fraction of air can be trapped near crack tips during fluid flooding. In experimental

data, reported by [2,3] all chalk samples were dried before flooding with different fluids, thus small

fraction of air may remain in pores space after finishing flooding. In calculations outlined in this paper,

the volume of air inside a crack is negligibly small comparing to the volume of fluid inside the crack, thus

it can be physically neglected.

To calculate the capillary pressure, it was assumed that the Young-Laplace equation and the concept of

the contact angle are applicable on the small scale near the crack tips, i.e., the width of interface meniscus

is small and the contact angle is the scale-independent physical property of chalk. The chalk is

preferentially water-wet; however, the spontaneous imbibitions of all other fluids (water, methanol, glycol

and mineral oil) into dry chalk were observed as well [2]. Thus, the experimental data suggests that the

wettability of chalk is not a physical property, because it can be altered by chemical processes during

experiment [5]. Also, the wettability could be different in wide and shallow pores [6]. Using the analytical

solution for capillary pressure and for stress intensity factor [7] the following expression for apparent

fracture toughness is derived:

2, Apparent 2

cos( )21IC ICK K E γ θ

ν= −

−

where: ICK is fracture toughness of calcium carbonate (material the chalk is composed of); & E ν are

the Young’s modulus and Poisson’s ratio of calcium carbonate; γ is the surface tension of pore fluid,

while θ is the contact angle. Using the above equation it is possible to estimate theoretical tensile and

uniaxial compressional strength of rock as follows , Apparent0, Apparent

ICKT

lπ= and

, ApparentApparent 8 ICK

UCSlπ

= , where 2l is the fracture length.

According to the above equation the apparent strength of rock, saturated with wetting liquid will decrease,

while the apparent strength of rock, saturated with non-wetting liquid will increase.

In this paper it is discussed how to apply the above equation to explain previously published experimental

data for chalk with different wettability and saturated with various fluids, including water-glycol

mixtures.

References: [1] Risnes R and Flaageng O. Mechanical Properties of Chalk with Emphasis on Chalk-Fluid Interactions and Micromechanical Aspects. Oil & Gas Science and Technology – Rev. IFP, 54 (6): 751-758. [2] Risnes R, Haghighi H, Korsnes RI, Natvik O. Chalk–fluid interactions with glycol and brines. Tectonophysics 370 (2003) 213– 226, doi:10.1016/S0040- 1951(03)00187-2. [3] Risnes R, Madland MV, Hole M and Kwabiah NK. Water weakening of chalk-mechanical effects of water-glycol mixtures. Journal of Petroleum Science and Engineering, 48 (2005) 21– 36. [4] Vutukuri VS. The Effect of the Liquids on the Tensile Strength of Limestone. International Journal of Rock Mechanics and Mining Sciences, Vol. 11 (1974), pp 27-29. [5] Strand S., Hjuler M. L. , Torsvik R., Pedersen J. I., Madland M. V., and Austad T., ”Wettability of chalk: impact of silica, clay content and mechanical properties”, Petroleum Geoscience 2007, v.13; p69-80., doi: 10.1144/1354-079305-696. [6] Hedegaard K, and Graue A. Does Wettability Affect the Strength of Chalk?, 45th US Rock Mechanics / Geomechanics Symposium held in San Francisco, CA, June 26–29, 2011, ARMA 11-599. [7] Rozhko AY (2011), Capillary Phenomena in Partially-Saturated Rocks: Theory of Effective Stress, 45th US Rock Mechanics / Geomechanics Symposium held in San Francisco, CA, June 26–29, 2011, ARMA 11-146.

Shale characterization and test protocols

Jørn Stenebråten

SINTEF Petroleums Research

Department of Formation Physics and Well Integrity

The characterization and test protocols described focuses on in-situ saturated shales.

An initial evaluation of a shale core using CT scan often reveals information that affects sample preparation and testing, such as open fractures, inclusions or general heterogeneities, all which can affect the success of preparing samples for testing, as well as the quality of the sought test results. The condition of the core also determines the types of test (and sample size) that can be utilized; on a heavily fractured shale the focus may have to be on index type of tests, whereas on an intact core larger samples can generally be extracted for general mechanical characterization tests such as triaxial compression, indirect tensile strength, or hollow cylinder tests. During core handling, sample preparation and storage prior to testing it is important to not alter the present saturation of the sample, as this may trigger non-reversible processes altering the static and dynamic properties of the material.

Following the localization of regions of interest, comes the general characterization of the shale assessing the mineralogy, relevant petrophysical parameters and pore water composition.

A complete fluid saturation can not be expected at atmospheric conditions of a shale core retrieved from in-situ temperature and stress, however, it is important to re-establish this as good as possible prior to mechanical or dynamic characterization, both to reduce capillary suction forces and to create contact between the pore water and the fluid used in the pore pressure measurement loop, in case of tests performed in a pressure vessel. Before contacting a shale sample with an aqueous fluid it is vital to have the sample instrumented and stressed to monitor the response of contact with the aqueous fluid.

For tests performed under a confining pressure it is required to allow sufficient time for stabilizing at the consolidation conditions, and to use the consolidation response to estimate proper strain or load rates based on the drainage conditions and the sample properties.

Geomechanical testing of the mudstone caprock

above injection processes in oilsand reservoirs

Mr. Patrick M. Collins, P.Eng

Petroleum Geomechanics Inc.

Calgary, Alberta

Most of the processes used to recovery heavy oil and bitumen from oilsands reservoirs require the high-pressure injection of fluids such as steam, air, or solvents. Containment of these injected fluids and the mobilized reservoir fluids within the oilsand reservoir is essential in order to ensure the recovery of these fluids and to prevent their loss to other formations. Key to ensuring containment is the presence of a “caprock” formation: an overlying tight formation, such as mudstone, that prevents the egress of underlying fluids.

This presentation briefly describes the issues pertinent to caprock integrity and focuses on the laboratory tests commonly conducted on the mudstone caprock. Uncommon features of these tests include high confining stresses, thermal effects, and salinity considerations. The results of these tests are used in caprock integrity studies that predict the performance of the caprock and the containment of reservoir fluids over the life of the recovery processes.

Typical laboratory programs include:

ASTM Fluid Test Description D2216 Moisture content on core D422 Hydro-grain analysis D854 Grain Density (Specific Gravity of solids ) D1418 Atterberg Limits – Blenderized D3080 saline direct shear, on natural joints

direct shear, on intact rock D5084 saline Triaxial Permeability Test D4767 saline Triaxial Test (CUP) - single stage

Triaxial Test (Consolidated Drained) - single stage Cyclic Triaxial Test (Consolidated Drained) for Young's Modulus - single stage

2166 UCS - Unconfined compressive strength D2435 saline Consolidation Test

distilled Swelling Test Thermal Expansion, horizontal Thermal Expansion, vertical X-ray Diffraction Pore Fluid Salinity

ISRM-ARMA Workshop on Petroleum Geomechanics Testing Chicago, 23rd June 2012

Testing the thermo-hydro-mechanical behaviour of shales

L. Laloui, V. Favero, A. Ferrari Swiss Federal Institute of Technology (EPFL), Lausanne

lyesse.laloui@epfl.ch

Abstract:

Thermo-hydro-mechanical behavior of shales is becoming one of the most important issues in modern geomechanics. The mutual influence among the mechanical, thermal and hydraulic responses of the involved materials should be considered for many specific applications. Highly compacted clayey materials are in fact found in numerous geotechnical applications, such as petroleum exploitation or gas shale extraction. Stress history, diagenesis and cementation may cause these materials to have a high yield pressure (usually greater than 10 MPa). Moreover, it is well known that the mechanical response of clayey materials depends on thermal and hydraulic changes. Temperature changes affect the swelling pressure, the thermal dilatation and contraction behavior, the stiffness, the yielding limit and the time dependent behavior (e.g. Hueckel and Baldi, 1990; Cekerevac and Laloui, 2009). On the other hand, the effects of changes in the degree of saturation have also been pointed out, highlighting the influence on deformability, swelling, collapse behavior, irreversible strain accumulation and strength (Laloui and Nuth, 2009). It is evident that the behavioral features of low-porosity clayey materials under large range of mechanical stress, suction and temperature must be analyzed and understood. The experimental analysis of the THM behaviour of the shale requires that the stress state, the temperature and the pore water potential must be simultaneously controlled. Due to the high preconsolidation pressure of the materials, the testing devices must be designed to study material behaviors under high stress levels, usually approximately dozens of MPa. Temperature is applied in the range of 20 to 90°C for most situations and up to 150°C for some specific cases. In addition, the pore water pressure has to be controlled in positive and negative ranges. As stated, high suction values are required in unsaturated conditions to induce a significant reduction in the degree of saturation. With reference to the mentioned issues, this paper presents the experimental devices and techniques which have been developed at the Laboratory for Soil Mechanics of the Swiss Federal Institute of Technology in Lausanne (EPFL, Switzerland). Two apparatuses operating in non-isothermal conditions are described; the effects of suction and temperature on the volumetric response of shales at high vertical stress are evaluated. Figure 1 depicts the layout of the oedometric cell that is devoted to the investigation of the volumetric response of shales under non-isothermal conditions (Salager et al. 2010).

The apparatus allows for independent control of vertical stress, total suction (by the vapour equilibrium technique) and temperature, and the measurement of water content variations and volumetric strains. The cell is designed to hold cylindrical samples (23 mm in height and 80 mm in diameter). The loading ram is positioned in the lower part of the system, while the upper base of the specimen is in contact with a fixed cylinder. The vertical load is applied by a hydraulic jack that is connected to a volume/pressure controller. A maximum 100 MPa vertical load can be applied.

Figure 1: Layout of the high pressure THM oedometric cell.

Figure 2 shows the thermo-hydro-mechanical triaxial cell; this system aims to perform suction controlled tests under wide range of temperature and confining pressure. The cell holds samples with 50 mm in diameter and 100 mm in height. The cell pressure is controlled in the range of 0–30 MPa by means of a pressure/volume controller. The apparatus is equipped with a double cell system for the assessment of sample volume variations. The mechanical deviatoric load on the sample is applied by controlling the pressure in lower hydraulic press which is handled by a pressure controller. The maximum applicable axial force is 450 kN, and the vertical displacement is measured by two LVDTs positioned outside the cell. Suction controlled tests are implemented by the axis translation and vapor equilibrium techniques to control the matrix and total suction. To control the capillarity pressure in the range of 0 –1.5 MPa, air pressure and water pressure are independently regulated inside the sample to control their difference (matric suction). Vapour equilibrium consists of controlling the relative humidity of a closed system in which the soil is immersed. In this way, soil water potential is applied by migration of water molecules through the vapour phase from a reference system of known potential to the soil pores, until equilibrium is achieved. The relative humidity of the reference system is controlled using salt solutions at various concentrations. By this technique pore water potential can be controlled in the range 4 – 400 MPa. Temperature of the system can be controlled in the range of 5–150 °C using a heating/cooling device which regulates the heat of the fluid in the outer cell (Seiphoori et al.,2011). Test results obtained on samples of shales are reported as an example of the use of the high pressure THM oedometric cell. Test results demonstrating the effects of temperature and suction on the deformability of the material are presented. Testing conditions were (22°C, 15 MPa), (22°C, 20.2MPa), (22°C, 4 MPa) and (80°C, 4 MPa). The results highlight the suction and temperature dependent behavior of the material in particular in terms of changes in the preconsolidation pressure and stiffness.

Figure 2: Layout of the THM triaxial cell. References

Cekerevac C. and Laloui L. “Experimental analysis of the cyclic behavior of Kaolin at high temperature”, Géotechnique 60:8, 2010.

Hueckel T. and Baldi G. “Termoplastic Behavior of Saturated Clays: an Experimental Constitutive Study”,

Journal of Geotechnical Engineering, ASCE, vol. 116 no. 12, 1990, p. 1778-1796. Laloui L. and Nuth M. “On the use of the generalised effective stress in the constitutive modelling of

unsaturated soils”, Computer and Geotechnics, vol. 36, 2009, p. 20-23. Salager S., Ferrari A. and Laloui L. New experimental tools for the characterization of highly

overconsolidated clayey materials in unsaturated conditions. In L. Laloui, editor, Mechanics of unsaturated geomaterials, pages 113-126. John Wiley & sons, 2010.

Seiphoori A., Ferrari A., Laloui L. " An advanced calibration process for a thermo-hydro-mechanical

triaxial testing system ". International Symposium on Deformation Characteristics of Geomaterials, Séoul, August 31- Sept. 2, 2011.

ON DILATANCY, FRICTION, AND SOFTENING RESPONSE OF ROCK: PLANE STRAIN TESTING

Joseph F. Labuz

Civil Engineering, University of Minnesota, Minneapolis, MN USA The objective of element testing of rock should be to promote homogeneous deformation but to allow the unrestricted development of the failure plane, also known as a shear fracture or shear band, with accurate measurement of force and displacement. A plane-strain apparatus, of the type first suggested by Vardoulakis and Goldscheider (1981), was designed and built by this premise (Labuz et al. 1996). The University of Minnesota plane strain apparatus (U.S. Patent 5,063,785) allows the unrestricted formation of the shear band so that the softening response can be evaluated, yet it permits the evaluation of constitutive behavior prior to failure so that dilatancy and friction characteristics of the material can be measured directly. The plane-strain condition is achieved by passive restraint, where a stiff “biaxial” frame enforces two-dimensional deformation. By placing the upper platen on a low friction linear bearing, the prismatic specimen, subjected to confining pressure and compressed axially, has the freedom to translate in the lateral direction if the deformation has localized, and the shear band can develop and propagate in an unrestricted manner. A schematic of the apparatus is given in Fig. 1. A prismatic specimen 75-100 mm in height ho x 27-40 mm in width wo x 100 mm in length is placed within a stiff biaxial frame; this thick-walled cylinder limits the displacement of the specimen to very small values to approximate the plane-strain condition. The specimen is subjected to lateral confining pressure by placing the entire assembly within a pressure vessel, and to axial load applied through displacing rigid platens by a servo-hydraulic actuator. In contrast to other plane-strain devices with similar loading conditions, the upper loading platen is attached to a linear bearing. The low friction bearing allows free displacement of the upper part of the specimen upon failure. Due to the presence of the linear bearing and a specimen slenderness ratio of 2-3, the formation of a shear band is spatially unrestricted. The apparatus is internally instrumented with five pressure resistant LVDTs for measuring axial and lateral displacements of the specimen and the linear bearing, two load cells for measuring the axial force, and four strain gages for measuring the axial and width strains on the specimen and the deformation of the biaxial frame. The two surfaces of the specimen exposed to confining pressure are sealed by a polyurethane coating; metal targets glued to the specimen provide firm contact points for the lateral LVDTs. The four surfaces in contact with polished-steel platens are covered with a stearic acid lubricant to reduce the friction between the platens and the rock (Labuz and Bridell 1993).

Figure 1. Sketch of the University of Minnesota Plane Strain Apparatus.

An advantage of plane-strain testing, as opposed to triaxial conditions, is that principal strains can be easily measured and volumetric strain ε and shear strain γ can be conveniently described:

31 εεε += 31 εεγ −= which can be decomposed into increments of elastic and plastic components: ep γγγ ∆+∆=∆ ep εεε ∆+∆=∆ Note that the principal strain directions correspond to the axial and lateral directions defined for the plane strain device and compression (volume decrease) is positive. Plastic strains are determined from the measured total strains and the calculated elastic strains (constant E and ν). Then, the dilatancy angle (Hansen 1958) is

sinψp

pεγ

∆= −

∆

Plane-strain compression experiments were performed on a Berea sandstone and data from three tests (5, 10, 20 MPa confinement) are shown in Figs. 2 and 3 (Riedel and Labuz 2007). The dilatancy angle ψ was calculated from the onset of plastic deformation to near peak stress (Fig. 2). Dilatancy showed strong pressure dependence. The degree of compactive behavior at the onset of plastic deformation, observed as a negative dilatancy angle, was larger for tests with higher confining pressure. The dilatant response of each specimen seemed to approach a limiting value; higher confinement resulted in a smaller value of peak dilatancy. Using the stress parameters s = (σ1+σ3)/2 and t = (σ1−σ3)/2, the Mohr-Coulomb yield condition can be written as t = s⋅sinφ + c⋅cosφ and the friction angle φ can be determined from

sinφ = t/s[1/(1−Q/s)]. The parameter Q is the intercept at t = 0; for the tested Berea sandstone, Q = 9.25 MPa. The linear yield function showed sensitivity to pressure (Fig. 3), while a nonlinear yield function could reasonably represent the response (Riedel and Labuz 2007). Note that the sandstone displayed non-normality.

Figure 2. Dilatancy of a Berea sandstone. Figure 3. Friction of a Berea sandstone.

30

35

40

45

50

55

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5

Plastic shear strain (x10-3)

Fric

tion

an

gle

(deg

rees

)

5 MPa

10 MPa

20 MPa

-30

-20

-10

0

10

20

30

40

50

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5Plastic shear strain (x10-3)

Dila

tanc

y an

gle

(deg

rees

)

5 MPa

20 MPa

10 MPa

References Hansen B. 1958. Line ruptures regarded as narrow rupture zones. Basic equations based on kinematic

considerations. Proc. Conf. Earth Pressure Problems. 1: 39-48. Brussels. Labuz JF, Bridell JM. 1993. Reducing frictional constraint in compression testing through lubrication. Int J Rock

Mech Min Sci Geomech Abstr. 30:451-455. Labuz JF, Dai S-T, Papamichos E. 1996. Plane-strain compression of rock-like materials. Int. J. Rock Mech. Min.

Sci. Geomech. Abstr. 33:573-584. Riedel JJ, Labuz JF. 2007. Propagation of a shear band in sandstone. Int J. Num. Anal. Meth. Geomech. 31: 1281-

1299. Vardoulakis I, Goldscheider M. 1981. Biaxial apparatus for testing shear bands in soils. Proc. 10th Int. Conf. Soil

Mech. Found. Engng: 819-824. Rotterdam: Balkema.

Block Testing with Polyaxial Stresses, and Fluid Flow-Coupling

Rico Ramos, ConocoPhillips

ABSTRACT

There are various techniques use to apply unequal 3D stresses, also called true-triaxial

or poly-axial stresses, on cubical rock specimens, where 3 orthogonal stresses are

unequal, S1 > S2 > S3 > 0, with or without pore fluid flow. There are poly-axial test

systems commercially available from fabricators, and there are testing services from

service laboratories. Contract testing services are available but limited, and only from

few service companies or from industry-academia joint projects. There are no

established testing protocols and there are several hydro-mechanical techniques of

applying the unequal 3D stresses. Pros and Cons between piston-actuators versus flat-

jacks, depend on the type and range of applications. The range of sample dimensions is

as low as a few cubic-centimeters to as large as 1 cubic-meter. A central vertical

borehole is used for the injection fluid, either pre-drilled before the stress-ramps, or

drilled under-stress. Sample selection and availability are usually of prime concern,

specially for cubic-meter tests. Various examples given are from published studies

where S2-effects are paramount, on borehole stability failure criteria, matrix acidizing,

and hydraulic fracturing. Costs and time escalate geometrically, with increase in the

dimensions of rock being tested

Title: Measurement and Application of Effective Stress Coefficients

Speaker: John Shafer

Company: Consultant

I am currently working with some really stiff ultradeep GoM rocks with a uniaxial pore volume compressibility (PVC) of about 1 microsips. The “standard” procedure for measuring Biot’s Alpha (α) or Effective Stress Coefficient (ESC) for bulk volume at the beginning of a pore pressure depletion PVC test leads to unrealistically high values of about 0.9. Biot’s α varies during a test where the high values of bulk compressibility and thus Biot’s Alpha are at stresses significantly below reservoir stress where core plug sleeve conformance and strain calibration can be an issue. Thus at stresses above reservoir stress the average bulk compressibility is about 0.55 microsips and the average value of Biot’s Alpha is about 0.65 assuming that the grain compressibility is 0.20 microsips. According to Zimmerman [2000] a more

direct approach to obtain Biot’s Alpha is given in the equation (εa)/∂(-Pp) = α [(εa)/∂ (σa)]. The ratio of the slopes of axial strain (εa) versus pore pressure as measured during pore pressure depletion PVC test or inferred during an effective stress PVC test was 0.7 that agrees well with that calculated from the average value of the bulk compressibility.

The concept of effective stress for volumetric strain for an isotropic, linear poroelastic solid, is rather

straightforward. Biot’s αb has a clearly defined meaning and applies as the effective stress coefficient for

εv regardless of the nature of the external state of stress. For example, for hydrostatic states of stress εv

= (Pc – αb Pp) / Kbulk and for non-hydrostatic states of stress εv = (σmean – αb Pp) / Kbulk . The coefficient αb

applies in both cases without any need to change the meaning (or value) of αb. It should be noted that this is not the case for effective stress coefficients in general. Even when a rock behaves to first order as

a simple isotropic poroelastic solid, non-hydrostatic changes in stress (e.g. ∆σa ≠ ∆σr ) will generally result in non-isotropic changes in the petrophysical properties. This is certainly the case for properties such as permeability and sonic velocities. In such cases, it is not clear how the concept of effective stress should best be applied.

A general approach to this problem is to assume that the stress dependence of the property of interest

(Z) can be described as a uni-valued function of external stresses [e.g. Z=f(σx) where σx = ƒ(σr,σa)]. A

simple case would be where ƒ(σr,σa)= A σr + B (σa – σr). Now in order to apply the concept of effective

stress, one needs to find α, Α, and Β such that Z = f(σx - αPp). Permeability and sonic velocity data will be presented to illustrate the calculation of effective stress coefficients for uniaxial pore volume

compaction tests. Based on the data analyzed in this study, we find σx≈σmean for axially propagating Vs,

σx≈σr+2σa/3 for axially propagating Vp, and σx≈σr for axially flowing permeability. We would expect that

radially propagating velocities and radial (horizontal) permeability would have their own σx’s and α’s.

Rock mechanics tests to obtain data for well completion are all effective stress tests with zero pore pressure and typically assume that Biot’s effective stress coefficient, alpha, is very near one, but the measured value for the rocks at reservoir stress conditions in this study are potentially significantly lower, 0.7. This could potentially impact well completions rock mechanics data interpretations. It should also impact the calibration of the log derived Poisson’s ratio and Young’s modulus since compression

and shear velocities have different ESC’s and thus effective stress collection of static and dynamic properties will not reflect reservoir conditions of stresses and pore pressures.

Test Methods for Creeping Rocks

Paul Hagin, Chevron Energy Technology Co.

Abstract Creep in rocks can be observed as time or rate dependent deformation of the matrix. All rocks exhibit creep strain, but the contribution of creep to the total deformation varies widely. In most cases, creep strain is irreversible. In an ideal creep strain test, an instantaneous step in stress is applied, after which stress is held constant and creep is observed as continuous strain as a function of time. Creep strain can also be observed during an ideal stress relaxation test, in which a step change in strain is applied and held constant, and the stress required to maintain the applied strain is observed to decrease as a function of time. Because creep strain is time-dependent, it is also possible to indirectly observe the effects of creep by conducting tests at varying stress or strain rates, or at varying frequencies during cyclic loading.

Most descriptions of creep strain are empirical, and take on a variety of functional forms. However, some micromechanical models exist. These can be grouped into two primary categories: frictional and thermodynamic. Frictional creep mechanisms rely on grain rearrangement due to intragranular slip, and are primarily stress dependent. Thermodynamic creep mechanics include dislocation and diffusion creep, and can be both stress and temperature dependent. Frictional creep has been documented in soils, clays, and uncemented and weakly cemented rocks including sandstones and chalks. Thermodynamic creep can be observed in all rocks at high enough temperatures, but has primarily been studied in igneous and metamorphic rocks.

It is difficult to propose a standard test protocol without consideration of the underlying mechanism. For thermodynamic creep, because the underlying physics are understood, several options for standards exist. For frictional creep, the types and number of tests required to describe the creep function depend on the material properties, but it is conceivable that a standard set of fit-for-purpose questions could be used to inform test protocol design.

Use of synthetic rocks for quantification of coring induced rock alteration

Rune M Holt; SINTEF Petroleum research & NTNU, Trondheim, Norway

Abstract:

Synthetic rocks formed under stress have been used to demonstrate how compaction behavior, porosity and wave velocities are altered as a result of stress release, mimicking the situation during core drilling. This technique enables a better understanding of core damage mechanisms, and how they may be corrected for or mitigated. Two sources of such core alteration have been identified, i.e. formation of microcracks in competent sandstone and collapse of pore structure in poorly consolidated sand and sandstone. Discrete particle modeling has been used as an additional tool, permitting simulation of stress paths representative of coring, and extending the studies of stress release effects as seen in the synthetic rock experiments. This includes studies of stress memory (Kaiser effect) in granular rock.

If time permits, the presentation will also contain some results obtained with synthetic analogues to study fractured rock, comparison between compacted clay and shale behaviour, and ideas around manufacturing of artificial shale.

The main message is that use of synthetic rock permits controlled experiments that may contribute to an improved understanding of physical and mechanical processes in rocks. Such understanding can be used to improve models for rock behavior, but one needs to be careful in direct use of results for field applications.

Enhanced Rock Strength Profiling, Combining Triaxial Compressive Strength, Non-Destructive Core Scratching and Index Tests Abbas Khaksar and Feng Gui: Baker Hughes Inc. Australia Christophe Germay and Thomas Richard: EPSLOG SA, Belgium and Australia Knowledge of rock mechanical properties including rock strength is essential for accurate in situ stress analysis and geomechanical evaluations. Reliable quantitative data on rock strength parameters can only be derived at specific depths from laboratory tests on core samples typically through destructive tests on cylindrical samples. However, laboratory data are limited, discontinuous and often biased toward stronger intervals. In practice, many geomechanical problems are often addressed in the absence of core samples for laboratory testing. Consequently, rock strength evaluation is primarily based on log strength indicators, calibrated where possible against limited measurements on cores.

A number of techniques have been developed to replace or supplement plug based destructive tests to measure the strength properties of rocks. Scratch and Schmitt hammer tests test are examples of such techniques that have demonstrated the ability to provide continuous or semi-continuous, fine-scale measurements of rock mechanical properties. In contrast with conventional triaxial tests, both of these index tests are non-destructive and do not cause significant damage to the core and no special core preparation is required. These tests can be conducted either in the lab, core store or, in principle, on the rig, almost immediately after recovery of core material.

The Schmidt hammer, originally designed for concrete testing and is being widely used in civil engineering and more recently in geotechnical works and mining and applications. The hammer contains a spring-loaded mass that is automatically released against an impact plunger when the hammer is pressed against the test surface. Elastic recovery of the rock is dependent upon its surface hardness. Since hardness is related to mechanical strength, the rebound distance travelled by the returning hammer mass is a relative measure of the surface hardness, and therefore the strength. Taylor and Appelby in SPE 101968 and Khaksar et al in SPE 121972 showed the use of Schmitt hammer to characterise rock strength on cores from North Sea wells with application in sanding evaluations. Impact (rebound number) measurements were made and compared between matched depth sections of whole core, half cut core and slab sawn, resinated core, and all were found to correlate well with each other and many other petrophysical properties; e.g. GR, RHOB, NPHI, Dtc, Dts, Helium porosity, air permeability and grain density. Four intervals of high rock strength can be clearly seen and correspond to zones of intense quartz cementation. Direct measurement, semi-continuous impact logs such as this provide an excellent pre-filtering tool for plug selection for rock strength laboratory testing and qualitative assessments of rock strength. Scratch testing- Experimental results with sharp cutters at small depth of cut show linearity between the energy required provided to the cutter and the volume of rock removed by scratching. The intrinsic specific energy or rock strength is well correlated with the uniaxial compressive strength (UCS) of the tested material. The variation of the internal friction angle of the rock can also be assessed by using a blunt cutter. Experimental data on various sedimentary rock types are used to derive generalized correlations between the intrinsic specific energy and UCS and friction angle. Strong correlations are also found between intrinsic specific energy and core porosity and permeability. We present experimental results from laboratory testing using conventional destructive triaxial testing on core plugs and continuous non-destructive scratch testing along on a series of silisiclastics rocks. The rocks types represent a range of strength from very weak and unconsolidated with UCS ~ 2-6 MPa to moderately strong with UCS ~ 20-40 MPa and strong and competent rocks with UCS > 60 MPa and measured friction angle ranging from 17 to 55 deg. The experimental results show that albeit the use of generalized (global) correlations of scratch testing may provide relatively good estimates of rock strength comparable with triaxial tests, a local calibration based on triaxial tests would be required for more accurate and robust rock strength profiling.

It is shown that besides the direct application to geomechanical modeling the scratch test can also be used as a screening tool to optimize the plug location to characterize the range of heterogeneity of cored interval. Once calibrated to the conventional plug tests, the continuous scratch and Schmitt hammer test results can be correlated with core sedimentological and petrophysical properties to generate more reliable rock strength profiling coupled with rock facies analysis and diagenetic classification for improved strength profiling.