IPTC 15029

An Integrated Reservoir Simulation-Geomechanical Study on Feasibility of CO2 Storage in M4 Carbonate Reservoir, Malaysia Rahim Masoudi and Mohd Azran Abd Jalil, PETRONAS; David Press, Kwang-Ho Lee, Chee Phuat Tan, and Leo Anis, Schlumberger; Nasir Darman and Mohamad Othman, PETRONAS

Copyright 2011, International Petroleum Technology Conference This paper was prepared for presentation at the International Petroleum Technology Conference held in Bangkok, Thailand, 7–9 February 2012. This paper was selected for presentation by an IPTC Programme Committee 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 International Petroleum Technology Conference and are subject to correction by the author(s). The material, as presented, does not necessarily reflect any position of the International Petroleum Technology Conference, its officers, or members. Papers presented at IPTC are subject to publication review by Sponsor Society Committees of IPTC. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the International Petroleum Technology Conference is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, IPTC, P.O. Box 833836, Richardson, TX 75083-3836, U.S.A., fax +1-972-952-9435

Abstract The M4 Field is located north of Central Luconia Province in the Sarawak Basin, East Malaysia. The reservoir is approximately 2000 m below sea-level where the water depth is approximately 120m. An integrated geomechanical study for CO2 geological storage has been conducted to evaluate the feasibility of injecting and storing CO2 in the M4 depleted carbonate gas reservoir. The storage feasibility of M4 reservoir is impacted by interaction of the reservoir rock with carbonic acid formed by dissolution of injected CO2 in the water which has risen close to the cap-rock. The geomechanical study needs to assess the risk of CO2 leakage from the reservoir due to degradation of the integrity of the cap-rock by the injection operations, and interaction of the injected CO2 and carbonic acid with the cap and reservoir rocks. A scope of work incorporating data review and integration, downhole log and image interpretation, 1-D in-situ stress and pore pressure analyses, rock property determination and 3-D coupled reservoir geomechanical modeling was conducted. In addition, laboratory rock mechanics tests and petrophysical measurements were conducted on core samples before and after injection of CO2 saturated brine solution, and the results were used to develop material strength, elastic and petrophysical property degradation models due to carbonic acid-carbonate interaction. A coupled geomechanical modeling was subsequently performed, which incorporates reservoir pressure and CO2 saturation from dynamic simulation, and subsequent changes in effective stress and the associated changes in porosity and permeability are calculated by a geomechanical modeler which were then passed back to the dynamic reservoir simulation. In addition, modifications were also made to geomechanical and petrophysical rock properties based on the carbonic acid-carbonate interaction degradation models. The paper describes the staged works from 1-D Mechanical Earth Model construction to comprehensive laboratory rock mechanics testing, 3-D geomechanical model construction, pre-production stress modeling and various injection scenario predictions. Examples of key results and utilization of the results and findings from the geomechanical study to develop recommendations for optimizing the CO2 injection and storage in the M4 Field in order to achieve optimal geological storage management and direct cost savings will be presented and discussed. Introduction

Along with capacity and injectivity of CO2, containment is a primary function in geological storage performance. Controlling the trapping of CO2 in the subsurface, i.e. storage containment, is of fundamental importance for safe geological storage of CO2. Rock formations can be impervious enough to act as flow barriers to CO2 over geological periods of time. Delineating such a seal, safeguarding its integrity under operational conditions, and verifying its isolation effectiveness are key objectives in achieving a successful CO2 storage project. During CO2 injection, increasing fluid pressure, temperature variation, and chemical reactions between the gas and rocks inherently affect the state of stress inside the reservoir and its surroundings. Besides, the mechanical properties of the rocks may be altered by their exposure to CO2 and/or pressure and stress changes. Furthermore, a fault may seal if deformation processes have created a membrane seal, or if it juxtaposes sealing rocks against aquifer/reservoir rocks and if the fault has not been re-activated subsequent to the fluid charging the trap. Rock mechanical properties, pore pressure, in-situ stresses and the stress evolvement under injection conditions control re-activation of a fault, and therefore risk of fault seal breach. The

2 IPTC 15029

impact of the resulting stress and pressure change and associated deformation on cap rock and fault seal integrity must therefore be assessed in order to properly manage containment performance and leakage-related risks. In order to address the above issues, a good understanding of the flow dynamic, in-situ stresses, pore pressure and rock mechanical properties in the field is necessary. The fracture initiation, propagation and containment in the injection zones, and cap rock and fault seal integrity are related to the in-situ stresses and coupled pressure and thermal behaviour while injecting. It is worth noting that when a depleted reservoir is targeted for CO2 storage, both the geomechanical effects associated with future CO2 injection as well as those due to past production must be considered in the modelling.

Objectives

The broad objectives of this study were to: • Investigate the feasibility of injecting and storing CO2 in the M4 depleted gas reservoir • Recommend overall best geomechanics practice for the CO2 injection and storage, and operation and management

of the field throughout its life, in such a way as to achieve optimal reservoir management and direct cost savings.

A reservoir geomechanical study based on the workflow shown Fig 1, was conducted in order to achieve the objectives.

Fig 1. 3-D Reservoir simulation-geomechanical study workflow

The reservoir geomechanical study was conducted to asses CO2 storage for the M4 Field, see Fig 2. The M4 field is located in the North of the Central Luconia Province in the Sarawak Basin, East Malaysia approximately 262 km NNW of Bintulu. The Sarawak Basin is the western part of the large Cenozoic depocentre comprising the continental shelf off the north coast of Borneo and most of onshore Sarawak. The field is approximately 2000 m below sea-level where the water depth is approximately 120m (Farah, 2008). M4 is a depleted field whereby the reservoir is carbonate. The water level in the reservoir has risen close to the cap-rock which implies a strong aquifer.

Fig 2. M4 Reservoir topography

IPTC 15029 3

1-D Mechanical Earth Model (MEM) In order to construct a 3-D geomechanical model used to simulate CO2 injection in the depleted carbonate reservoir, 1-D Mechanical Earth Models (MEMs) were constructed for A-1, A-2, AA-1 and AA-2 offset wells. 1-D MEMs were constructed in order to determine stress magnitudes, stress orientations, pore pressures and formation mechanical properties of the field. The MEM consisted of, amongst other data, continuous profiles of the following rock mechanics information and parameters (Ali et al, 2003):

• A description of rock fabric (i.e., the mechanical stratigraphy of the formations considered); • Rock elastic parameters: including Young’s modulus and Poisson’s ratio; • Rock strength parameters: Unconfined compressive strength (UCS), tensile strength and angle of internal friction;

and • Stress model: Vertical stress, minimum and maximum horizontal stress magnitudes and orientations, and pore

pressure.

The MEM for A-1 well is shown in Fig 3. Once the 1-D MEMs were constructed, the 3-D geomechanical model was constructed using data from the 1-D MEMs.

Fig 3. 1-D Mechanical Earth Model for A-1 offset well

4 IPTC 15029

Rock Mechanical and Petrophysical Properties Tests A comprehensive programme of mechanical and petrophysical properties tests was conducted on selected overburden shale and reservoir limestone core materials from well A-2. The aim of this test programme was to conduct a novel experimental evaluation on any potential interaction between injected carbon dioxide (CO2) and reservoir and cap rocks which could result in a change in mechanical properties of the rock materials and impact on CO2 storage and containment. Shale samples from two core depths and limestone samples from three core depths were tested. The study consisted of comprehensive evaluation of the limestone and shale mechanical and petrophysical properties on identical sets of samples from each core depth. Each set of samples was tested untreated with CO2 whilst an identical set of ‘sister samples’ was tested after having undergone CO2 injection treatment to simulate reservoir injection conditions. A standardized CO2 injection procedure was adopted in an effort to quantify the potential impact CO2 injection has on shale and limestone rock properties under controlled test conditions. Using this approach, a direct comparison can be made to determine any changes in the rock properties resulting from the CO2 interaction. Two separate approaches were adopted to simulate the CO2 injection for the shale and limestone samples. The approach adopted for shale samples was to inject liquid CO2 using an upstream injection pressure simulating field injection with a downstream (reservoir) pressure until CO2 breakthrough occurred. At this point, the sample was held with constant pressure for a predetermined treatment period. For the limestone samples, the procedure was to inject a known quantity of CO2 saturated sodium chloride solution (carbonic acid brine solution) at a predetermined flow rate, in-situ hydrostatic confining and reservoir injection pressure conditions. Following the flow period, the CO2 saturated brine solution was held in contact with the rock for a set period to allow any dissolution reaction to occur. All injection simulations and tests for this study were performed at ambient temperature. The comparative rock mechanical test program for both pre-and post-CO2 injection treated samples consisted of:

• Single stage triaxial compression tests with concurrent ultrasonic velocity measurements on shale samples at room temperature for Mohr-Coulomb failure envelope delineation;

• Multi stage triaxial compression tests with concurrent ultrasonic velocity measurements on limestone samples at room temperature for Mohr-Coulomb failure envelope delineation;

• Indirect tensile strength tests (Brazilian method) on limestone samples; • Pore volume compressibility tests using conventional effective stress loading under uniaxial strain conditions to

simulate production and with concurrent axial permeability measurements; • Basic petrophysical properties i.e., porosity and permeability required for analysis of pore volume compressibility,

pulse decay permeability and mercury injection porosimetry tests; and • Petrographic and mineralogical analyses i.e. mineralogy, grain size, texture and fabric including X-ray diffraction

analysis and scanning electron microscopy.

Petrophysical Properties

Pre-CO2 injection porosity was determined for correlation purposes with the pre-CO2 pore volume compressibility tests for the limestone samples. The identical samples were used to determine porosity and permeability following treatment with CO2 saturated brine for pre-injection treatment comparison and for correlation purposes with the post-CO2 treated pore volume compressibility tests.

Porosity of the limestone samples was determined following cleaning by measuring the pore volume at a nominal stress and at effective mean in-situ stress conditions. Permeability to gas was then measured at the same stress conditions. Uncleaned samples were used for the pore volume compressibility tests conducted under uniaxial strain conditions. For each test, porosity was corrected for the volumetric deformation that occurred during reapplication of the net effective mean in-situ stress.

Pulse decay permeability tests were carried out on two samples from each shale core depth. One sample from each depth was subjected to CO2 injection treatment and followed by pulse decay permeability test for comparison with the untreated sample. High pressure, mercury injection porosimetry tests were performed on two samples from each shale core depth. One test was carried out on an untreated sample and the same test repeated on an adjacent ‘sister’ sample following CO2 injection treatment.

Petrological Analysis

Petrographic evaluation of pre-CO2 injection and post-CO2 injection shale samples from two core depths and limestone samples from three core depths were compared for original mineralogy and texture, as well as effects that may be related to CO2 injection. The post-injection plugs were sampled near the surfaces facing upstream during testing and hence, maximizing possible interaction due to exposure to CO2 saturated brine.

IPTC 15029 5

The shales are composed horizontally laminated and tightly crystalline clays, predominantly illite, mixed layer illite-smectite, and a lesser amount of chlorite. No significant difference between pre-and post-CO2 injection samples was observed. The limestones comprise highly porous, slightly recrystalised boundstones composed of coral, algal, and other calcareous skeletal particles cemented by sparry calcite. Mineralogy and original rock texture are essentially the same in the pre-injection and post-injection samples. However, the two deeper post-injection limestone samples exhibit deeply etched and corroded calcite grains in small areas of the samples. Similar textures were not found in the corresponding pre-injection samples.

Rock Mechanical Properties

Triaxial Compression Tests

Single-stage triaxial compression tests were performed on untreated (and unpreserved) shale samples and ‘sister’ samples subjected to CO2 injection treatment from two core depths at varying confining pressures in order to determine Mohr-Coulomb strength parameters and compare the effect of CO2 injection. All the shale samples were tested in their “as-received” saturation condition with a standard axial loading strain rate of 1 x 10-5 in/in/s at room temperature with pore pressure drained to the atmosphere. Multi-stage triaxial compression tests were performed on untreated limestone samples from three core depths at a range of confining pressures in order to calculate the Mohr-Coulomb strength parameters. Similar tests were also conducted on ‘sister’ samples following injection treatment with CO2 saturated brine for comparison with the untreated samples. All the triaxial compression tests were performed at room temperature with pore pressure drained to the atmosphere. The samples, both untreated and treated, were tested with a standard axial loading strain rate of 1 x 10-5 in/in/s. Examples of the pre- and post-CO2 treated test results are shown in Fig 4 and Fig 5. Dynamic elastic properties using ultrasonic wave transmission were determined concurrently with all the triaxial compression tests. The average change in cohesion from pre- to post- injection for the limestones was +5% and the average change in friction angle was -12%. The cohesion and friction angle were used to determine unconfined compressive strength (UCS) using the following equation:

2 45 2

The average change in UCS from pre- to post-injection was -4%. The average change for Young’s modulus of the limestones was 8% reduction from pre- to post-injection whilst the post-injection Poisson’s ratio increased by an average of 10% in comparison with the untreated sample.

Fig 4. Mohr circles and failure envelope for pre- and post-CO2 treated limestone samples.

6 IPTC 15029

Fig 5. Variations of Young’s modulus with effective confining stress for pre- and post-CO2 treated limestone samples.

Indirect Tensile Strength Tests

Indirect tensile strength tests (Brazilian method) were conducted on two untreated limestone samples and two post-CO2 injection samples from each core depth. The untreated samples were tested following saturation with 2.6% sodium chloride brine solution. The treated samples were tested ‘as received’ following injection with CO2 saturated 2.6% sodium chloride brine. All the test results are shown in Fig 6. As the trend of the test results is inconclusive, the tensile strength reduction with CO2 treatment adopted corresponds to the reduction for UCS (see Triaxial Compression Test Section).

Fig 6. Tensile strength for pre- and post-CO2 treated limestone samples.

Pore Volume Compressibility Tests Pore volume compressibility (PVC) tests were conducted on untreated and post-CO2 injection treated limestone samples to determine the bulk and pore volume compressibility, and porosity of the samples, and to evaluate the impact of CO2 interaction on the rock properties. For all the tests, the PVC tests were performed at room temperature with pore pressure reduction to simulate in-situ reservoir depletion.

IPTC 15029 7

Average Rock Mechanical and Petrophysical Properties

The averaged rock mechanical and petrophysical properties determined from the test programme are given in Table 1. Table 1. Average Rock Mechanical and Petrophysical Properties

Property Shale Limestone Depth Range (ft) 5779.7 – 5800.4 5831.5 – 5892.1 Test Condition Untreated CO2 Injected Untreated CO2 Injected

Initial Reservoir Pressure (psi) 3772 3772 3850 3850

In-situ Porosity (Initial conditions) -- -- 33.64% 34.93%

Porosity (Fully depleted conditions) -- -- 29.12% 29.00%

In-situ Permeability to Gas (mD) 0.001161* 0.000963* 272 316 Averaged Pore-Volume Compressibility

(Depletion phase) (Cpp – based on reservoir pressure)

-- -- 61.5 microsips 1 49.4 microsips

Ko (Averaged over the entire range of depletion) -- -- 0.39 0.36

Critical Reservoir Pressure for Accelerated Compaction -- -- 2235 2405

Reduction in Brine Permeability (Ranging from

initial in-situ conditions to abandonment) -- -- 72.4% 76.5%

Average Grain Compressibility (Cg) -- -- Assumed to be 0.2 microsips (10-6/psi)

Coulomb Friction Angle (Entire range) 23.5° 21.2° 14.1° 12.7°

Coulomb Cohesion (Entire range) (psi) 1138 1307 603 550

In-Situ Effective Compressive Strength (psi) 5932 5982 2590 2518

In-Situ Static Young’s Modulus (psi) 256,500 298,850 1,190,265 1,125,765

In-Situ Dynamic Young’s Modulus (psi) 1,537,500 1,498,000 2,351,000 2,335,300

In-Situ Static Poisson’s Ratio 0.15 0.14 0.22 0.23

In-Situ Dynamic Poisson’s Ratio 0.17 0.10 0.33 0.29

Indirect Tensile Strength (psi) -- -- 87 98

Carbonate Classification -- -- Boundstone Boundstone

Carbonate Grain Types -- -- Red algae, coral, foraminifera

Red algae, coral, foraminifera

Authigenic Cements Calcite, chert, pyrite Calcite, chert, pyrite Sparry calcite, minor

dolomite Sparry calcite, minor dolomite

*: Affected by presence of micro-fractures resulting from desiccation of the unpreserved shale samples 1 microsip = 10-6/psi

Based on the results of the novel rock mechanical and petrophysical properties test programme, it was observed that there is a reduction in Young’s modulus, unconfined compressive strength, angle of internal friction and tensile strength whilst there is an increase in Poisson’s ratio and permeability for the post-CO2 treated test samples.

3D Geomechanical Model Construction

Geometry

The 3-D geomechanical model for the M4 reservoir was constructed based on an existing static model, which utilized the same grid as an existing dynamic reservoir model. The existing grid contained approximately 400,000 grid cells with an individual cell size of 60m x 60m in the horizontal dimensions and containing 64 layers with an average thickness of 5m. The static and dynamic models spanned 3.4km in the east-west direction and 7.0km in the north-south direction. Depth ranged from 1723m to 2022m sub-sea level with a sea-water depth of approximately 120m. The geometry of the existing static and dynamic models is shown in Fig 7.

8 IPTC 15029

Fig 7. Static model representation of the M4 reservoir

In order to construct a representative 3-D geomechanical model, the initial static model required the addition of cells to completely surround the reservoir model in a process known as cell embedding. This was an essential process in order to transfer the far-field pre-production stress state to the reservoir model. The overburden was extended to the sea-bed and the underburden to a depth of 15km. Cells were added in the horizontal dimensions to a distance of one reservoir length. The resultant embedding increased the model size to approximately 1 million grid cells as shown in Fig 8.

Fig 8. Embedded 3-D model

Material properties

Elastic and strength material properties with depth in the reservoir and overburden were contained in the 1-D MEMs of the four offset wells. Population of material properties throughout the entire 3-D model was carried out using a deterministic kriging algorithm with a Gaussian variogram. Material models within the analysis were based on the Mohr-Coulomb failure criterion and the property parameters include Young’s modulus, Poisson’s ratio, density, UCS, angle of friction and tensile strength.

IPTC 15029 9

Pore pressure The initial reservoir pressure was taken from the initial step of the dynamic reservoir simulation. For cells inactive to changes in pressure and within the overburden, an average pore pressure profile was used derived from the 1-D MEMs of the four offset wells. Boundary stresses

In order to establish a representative in situ stress condition within the model, prior to production or injection of CO2, horizontal stress gradients were applied to the boundaries of the model. These were oriented such that the maximum horizontal stress was applied at N020oE, this orientation having been established during the 1-D MEM construction and with reference to The World Stress Map (Heidbach et al, 2009). The horizontal stress gradients were also determined from the 1-D MEMs and a normal fault stress regime was apparent within the field. Maximum and minimum applied horizontal stress gradients were 20.78 kPa/m and 19.36 kPa/m respectively. The vertical loading was applied as a gravitational density load determined by the total overburden weight in addition to an applied surcharge at the top of the model representing the weight of sea water.

Pre-Production Stress Analysis Having provided all the required data for analysis, the stress simulator was used to establish a representative initial stress state within the reservoir and surrounding formations prior to production operations. In order to confirm that the computed initial stresses in the 3-D geomechanical model were in agreement with those contained in the 1-D MEMs, a direct comparison was made at the four offset wells. Fig 9 shows the initial stress comparison between the 1-D and 3-D solutions for one of the wells. The agreement between the results from the two approaches suggested that a representative initial stress state in the 3-D model had been established.

Fig 9. 3-D and 1-D initial stress comparison

CO2 Injection and Sequestration Modeling

In the context of CO2 injection and sequestration study on M4 Field, the analysis was focussed on determining the well injectivity potential, CO2 storage capacity and containment. Various sensitivity analyses were conducted to optimize the CO2 storage capacity, well injectivity potential, injector numbers and locations as well as CO2 trapping mechanisms within the M4 structure. In this simulation, the existing gas producer wells are converted into CO2 injector wells whereas proposal for a new CO2 injector well will be based on the results obtained from the sensitivity analysis conducted on well injectivity potential and optimum well location. Results obtained from all the sensitivity analyses were incorporated into the final CO2 injection and sequestration scenario for the ECLIPSE-VISAGE two-way coupled modeling with material property updating. Two CO2 injection and sequestration scenarios were constructed for ECLIPSE-VISAGE coupled modeling. The first CO2 injection scenario simulates field CO2 injection rate of 200 MMscf/day using four CO2 injector wells, i.e. two existing gas producer wells and two new injector wells, with each well injecting 50 MMscf/day. The injection phase is planned for 10 years whereby it starts in 2018 and end in year 2028, beyond which the simulation run continues until year 2060 for monitoring purposes. The injection is constrained by Bottom Hole Pressure (BHP) limit fixed to the initial reservoir pressure of 3850 psi. The objective of the first CO2 injection scenario is to evaluate the cap rock integrity of M4 Field with CO2

10 IPTC 15029

injection up the initial reservoir pressure. The second injection scenario, on the other hand, simulates field CO2 injection rate of 200 MMscf/day using only two CO2 injector wells, i.e. two existing gas producer wells, with each well injecting 100 MMscf/day continuously from year 2018 until year 2065. The objective of the second CO2 injection scenario is to identify the maximum injection pressure the M4 depleted reservoir and cap rock could withstand before hydraulic fractures are initiated. In terms of CO2 trapping mechanisms in the reservoir, only three trapping mechanisms were evaluated in the study. The three trapping mechanisms are hydrodynamic, capillary, and solubility trapping mechanisms. Mineral trapping mechanism was not evaluated since the effect is less pronounced in short time scale. Fig 10 shows the amount of CO2 trapped as solubility (left), capillary (middle) and hydrodynamic (right) trapping mechanisms for both first (red) and second (blue) scenarios. As depicted in Fig 10 the CO2 dissolution in water for Scenario 1 and Scenario 2 is calculated to be approximately 50 Bscf and 128 Bscf respectively. High amount of CO2 dissolved in water, as calculated in Scenario 2, is attributed to continuous CO2 injection which results in larger CO2 spreading in the reservoir and hence, more CO2 come into contact with fresh water, enhancing solubility trapping. In terms of CO2 trapped due to capillary and hydrodynamic effects, a clear separation can be seen between Scenario 1 and Scenario 2. Theoritically, combination of higher CO2 trapped, either in water or due to capillary effect, will results in lower CO2 volumes exists as mobile CO2 phase. However, this effect is not seen in Scenario 2 due to continuous CO2 injection that result in continuous increase of CO2 volume in the reservoir. In addition, sensitivity analysis on the significance of the hysteresis effect in CO2 sequestration was also conducted. Fig 11 shows the comparison between CO2 injection and sequestration case with and without hysteresis effects. As observed, though CO2 dissolution in water is about similar in both cases, with hysteresis effect, more CO2 will be trapped due to capillary effect that will result in less CO2 exists in mobile phase or CO2 plume. The hysteresis effect is favorable since less CO2 will exists in mobile phase or also known as CO2 plume, hence it gives less impact to cap rock integrity.

Fig 10.Amount of CO2 trapped as solubility (left), capillary (middle) and hydrodynamic (right) trapping mechanisms for both first (red) and

second (blue) scenarios.

IPTC 15029 11

Fig 11.Comparison between CO2 injection and sequestration case with hysteresis (red) and without hysteresis (blue) effects

Coupled Modelling with Material Property Updating Once the initial stress state had been established in the 3-D model, the analysis was continued with the evolution of pore pressure changes due to production and CO2 injection. Several “stress-steps” were selected during the schedule at which points the pore pressures calculated from the reservoir simulator were passed to the stress simulator and equilibrium was re-established. In addition to the calculation of stress at each step, volumetric strains were also calculated and related to a change of porosity. These porosity changes gave rise to a modification of permeability in accordance with pore volume compressibility tests carried out on three samples in the laboratory. These test results are shown in Fig 12 along with the volumetric strain-permeability curve used to update permeability during the analysis. The laboratory tests were carried out in compression only, however as the reservoir was likely to expand during CO2 injection, the exponential curve was extended to include tensile volumetric strains. To account for the effects of acid-carbonate interactions resulting from the injection of CO2, the material properties were continually updated as the coupled analysis proceeded. The material parameters were updated in accordance with the results of laboratory tests on specimens saturated with a brine/CO2 solution as discussed in Rock Mechanical and Petrophysical Properties Test Section. Brine/CO2 flooding of the samples was considered to have given rise to 95% CO2 saturation. The magnitude of material parameter degradation was scaled to that determined from laboratory tests based on the prediction of CO2 saturation by the reservoir simulator. The complete coupling process is shown in Fig 13 where ECLIPSE denotes the reservoir flow simulator and VISAGE denotes the stress simulator. Three stress steps are shown only for brevity. Note that once material elastic and strength properties had been deemed to degrade, this change was considered permanent. There was no subsequent increase in elastic/strength properties should the CO2 saturation have reduced during the schedule.

12 IPTC 15029

Fig 12. Permeability – porosity relationship used in the coupled analysis

Fig 13. Stress-flow coupling with permeability updating and acid-carbonate interaction

IPTC 15029 13

The initial average reservoir pressure was 3850 psi in May 1992 which reduces to 2890 psi in 2015 due to production in accordance with the reservoir simulation schedule as shown in Fig 14. Two CO2 injection scenarios were considered as shown in the figure; the first included four injection wells, each injecting at a rate of 50MMSCF/day to 2028 whereby the reservoir pressure was brought back to the initial, pre-production condition. The second scenario considered two injection wells each injecting at 100MMSCF/day continuously to 2065. Results are shown for both scenarios. The vertical lines in Fig 14 indicate the times at which stress step calculations were carried out.

Fig 14. Average field pressure for two scenarios and selected stress steps

The variation in predicted gas saturation for Scenario #1, used to control the degradation of material properties, is shown in Fig. 15 for an upper reservoir layer. CO2 injection was initiated in 2018 and the saturation is seen to increase to a maximum value in 2028 and then reduces during the monitoring period. The resulting variation in derived CO2 saturation gave rise to a reduction in elastic and strength material properties. The reduction in Young’s modulus between 2017 (end of production stage) and 2028 (peak injection) for Scenario #1 on a section through the reservoir is shown in Fig. 16. Further degradation of material parameters was calculated during Scenario #2 where the injection rate and injected CO2 volume was considerably higher.

14 IPTC 15029

Fig. 15. CO2 gas saturation in an upper reservoir layer: Scenario #1

Fig. 16. Young’s modulus degradation between 2017 and 2028 (kPa): Scenario #1

Volumetric strain changes resulting from reservoir expansion within the reservoir during the complete injection schedule for both scenarios are shown in Fig. 17. The two images are plotted on different scales in order to emphasise the magnitude of strain changes. The increased volumetric strain calculations in scenario #2 resulted in larger permeability “modifiers”, the ratio of current to initial permeability. The permeability modifier at the final step in each schedule resulting from both volumetric strain changes and enhancement due to CO2 saturation is shown in Fig. 18.

IPTC 15029 15

Fig. 17. Volumetric strain changes during injection

Fig. 18. Permeability modifiers at final stress step.

A major objective of the study was to determine whether the cap-rock integrity would be breached during CO2 injection. For this reason an examination of the increased shear stresses within the cap-rock during the injection period was made. Large increases in shear stress would give rise to failure of the material in the form of shear bands or fracturing. This was to be avoided at all costs to prevent injected CO2 being vented to the atmosphere. Increased shear stress within the cap-rock due to injection is shown in Fig 19. Shear stress changes in the cap rock were not generally significant, although considerably larger for Scenario #2, due to an isotropic change in effective stress due to injection. Shear stress magnitudes for both scenarios were not large enough to overcome the shear or tensile strengths of the overburden formations. Corresponding vertical displacements at top reservoir level are shown in Fig 20.

16 IPTC 15029

Fig 19. Change of shear stress due to injection (bar)

Fig 20. Total vertical displacement at top reservoir level due to injection (m)

Changes in stress due to both production and injection within the reservoir for Scenario #1 were small enough such that the Mohr-Coulomb failure criterion was not reached at any point, indicating that the injection pressures could be withstood. However the large injection pressures in Scenario #2 gave rise to both shear and tensile failure within the reservoir. Fig 21 shows the development of large plastic shear strains of over 3% in the vicinity of the injecting wells in 2055. Plasticity occurred over almost the entire width of the reservoir indicating a complete breakdown of the intact rock for this volume of injected CO2. Shear strain is plotted against average field pressure over the complete schedule in Fig 22 for a cell exhibiting a high degree of plasticity in Scenario #2. During production the shear strain increased to a peak value of approximately 0.08% as the pressure in the reservoir reduced and effective stresses increased. Injection followed from step 7 and there was a reversal of sign in the predicted shear strain which reduced to a minimum value of 0.02% at step 13. The sign then changed once more and shear strain started to increase. A rapid increase in shear strain followed from step 15 as plastic behaviour was exhibited.

IPTC 15029 17

Fig 21. Plastic shear strain near top reservoir level in 2055: Scenario #2

Fig 22. Shear strain reversal within plasticity zone: Scenario #2

The shear strain reversal at step 13 can be explained by examination of the strain tensor within this particular grid-cell. The strain tensor at step 11 is shown in Fig 23. At this point in the analysis (and all preceding steps), the maximum compressive principal strain was in the vertical direction. However at step 13 the maximum compressive principal strain was in the horizontal direction as shown in the figure. A similar behaviour was observed in other grid-cells in which plasticity takes place.

The fact that the maximum compressive strains are vertical to step 12 would indicate that any fractures which may be induced during this period would be in the horizontal plane. This would be undesirable but not as undesirable as any vertically produced fractures which would take place following step 12 when the maximum compressive strain is horizontal.

18 IPTC 15029

Fig 23. Principal strain directionality: Scenario #2

Conclusions

Results from the two 3-D coupled analysis indicate that the scheduled injection pressure in Scenario #1, in which the initial reservoir pressure prior to production was re-established, could be accommodated. During this analysis there was no prediction of plasticity within the reservoir or cap-rock and the system remained elastic despite a degradation of material elastic and strength properties with increased CO2 saturation and an increase of permeability of up to 70% locally due to injection. Compactions induced during production were fully recovered during the injection phase when reservoir pressures returned to their pre-production values.

The large injection pressures in Scenario #2, on the other hand, could not be accommodated by the reservoir and gave rise to large amounts of plasticity of up to 3% by 2055. Pressures were sufficiently large as to increase the reservoir volume substantially, with a maximum value of volumetric strain in excess of 5% at several grid cell locations. The computed shear strains indicated the risk of vertical fracture initialization at an injected reservoir pressure of 4200 psi.

Although plasticity in Scenario #2 was not computed until the average reservoir pressure had exceeded 5350 psi, it was recommended that the critical field pressure should not exceed a value of 4200 psi based on an injection rate of 50 MMSCF/day at each of the wells. This was mainly because of the uncertainty in the existence of faults/fractures and the mechanical properties used in the model. It was also recommended that the pore pressure, strain and displacement at each injection well was monitored during the CO2 injection phase. The presence of any sub-seismic faults and/or fractures would significantly reduce the capacity of cap-rock sealing and integrity.

References Heidbach, O., Tingay, M., Barth, A., Reinecker, J., Kurfeß, D., Müller, B., The World Stress Map based on the database release 2008, equatorial scale 1:46,000,000, Commission for the Geological Map of the World, Paris, doi:10.1594/GFZ.WSM.Map2009, 2009.

Abdullah, Farah, Sarawak Shell Berhad, M4 Static Modelling Update Report in support of Full Field Review Studies, Report No. EPA-T-DSP/R08/003. May 2008.

Ali A. H. A., Brown T., Delgado R., Lee D., Plumb D., Smirnov N., Marsden J. R., Prado-Velarde E., Ramsay L., Spooner D., Stone T., & Stouffer T., (2003); Watching Rocks Change – Mechanical Earth Modeling; Oilfield Review; Summer 2003, Volume 15, Number 1, pp. 22-39.