MODELING AND SIMULATION WORKFLOW FOR A FRACTURED CARBONATE CO2 HUFF N PUFFNW McGregor Oil Field

E Goetz 1 A CASE STUDY IN THE WILLISTON BASIN NORTH DAKOTA USA(NDIC 3392)

Damion J Knudsen Guoxiang Liu Charles D Gorecki James A Sorensen Steven A Smith Edward N Steadman and John A Harju

For more information contact Damion J Knudsen dknudsenundeercorg (701) 777-5397 Charles D Gorecki cgoreckiundeercorg (701) 777-5355

James A Sorensen jsorensenundeercorg (701) 777-5287 Edward N Steadman esteadmanundeercorg (701) 777-5279 University ofEERC North Dakota

I mile

I kilometer

PCOR AlbertaEERC Innovates

Energy amp Environmental Research Cente r reg sm

Putting Research into PracticePartnership Technology AMERICAN COALITION FOR University of CLEAN COAL ELECTRICITYFuturesNorth Dakota

ENERGYGrand Forks2003 ndash Present

O

Biorecro Carbozyme Inc CANADA LTD

O

EXCELSIOR E N E R G Y Alberta Geological Survey

minnesota power

MISSOURI Natural ResourcesMISSOURI MUREXMUREXRIVERDEPARTMENT OF CanadaPETROLEUM CORPORATIONE N E R G Y S E R V I C E S reg

NATURAL RESOURCES

NORTH DAKOTA

OTTER AIL OFFICE OF RENEWABLE ENERGYAND ENERGY EFFICIENCY POWER COMPANY

TM

POWER SYSTEMSRAMGEN

Shell Canada Limited

tradeXcelADVANCED INTERSTATE PIPELINE COMPANYW E S T M O R E L A N D GEOTECHNOLOGY C O A L C O M P A N Y

1009

Energy amp Environmental Research Center reg

Grand ForksPutting Research into Practice

ABSTRACT A CO2 huff lsquonrsquo puff (HNP) enhanced oil recovery (EOR) project was carried out in the E Goetz 1 well located in the Northwest McGregor Field of Williams County North Dakota USAThe HNP is one of the Plains CO2 Reduction (PCOR) Partnership Phase II pilot projects in which CO2 was injected into the Mississippian Mission Canyon Formation a fractured carbonate reservoir for the dual purpose of CO2 EOR and associated CO2 storage The workflow for building the static geologic model for this study involved data collection and normalization petrophysical and facies modeling and dynamic simulation with history matchingThe small-scale injection model contained only one well so a larger-scale model containing several wells was built using sequential Gaussian simulation (SGS) and indicator simulations to determine trends and anisotropyThen a smaller downscaled injection model was built using discrete and continuous multiple point statistics to model the gradational mudstone to grainstone sequence common with platform carbonates while using a cropped portion of the large-scale model as a covariableThrough the analysis of core and drill stem test (DST) data it was determined that to more accurately model the reservoir a fracture model was needed which was constructed using discrete fracture network (DFN) simulationThe DFN model was then upscaled to the injection grid to produce a heterogeneous dual permeability and porosity modelThis dual property model was then exported into the Computer Modeling Grouprsquos (CMG) generalized equation-of-state model compositional reservoir simulator (GEM) and SGS and indicator simulation were used to adjust the static modelrsquos petrophysical properties assisting in the history match of the reservoirrsquos historical production Finally the modeling and simulation work was integrated with time-lapse reservoir saturation tool (RST) data and vertical seismic profile (VSP) data to accurately account for the injected CO2 and the produced water oil and CO2 By following this type of workflow the complicated nature of the CO2 HNP was modeled and matched to the monitoring verification and accounting (MVA) techniques displaying how this type of a workflow can be applied to other CO2 storage projects

The Plains CO2 Reduction (PCOR) Partnership is a collaborative program assessing regional CO2 storage opportunities Its primary sponsor is the US Department of Energy National Energy Technology Laboratory with additional support from its more than 80 partners

DUAL POROSITY AND PERMEABILITY MODELING WORKFLOW

Effective Porosity vs Kwater NW McGregor PHIE 012

FRA

CTU

RE (F

) M

ATRI

X (M

)

K wat

er(m

D)

10 Micro-Facies

1 002

01 Samples from 8076rsquo

001

0001

00001

0001 001 01 1

Effective Porosity (volvol) Mudstone Packstone Grainstone Anhydrite

y = 00004e317230x y = 00014e352912x y = 00038e343836x y = 00003e118095x

~100

0 ft

Step 2M The MMPA was Step 3M Results were verified with QEMSCAN Step 4M This graph helps illustrate the relationship between micro- Step 5M Seismic inversions were performed on VSP to acquire then completed in all analysisThe sample above was from approxi- facies effective porosity and permeability to water This graph along spatial variability in porosity and absolute acoustic impedence wells in the study area mately 8076rsquo md with the QEMSCAN analysis in Step 3M helps verify the petrophysical (AAI)

analysis

(1)kf = kd 2km

kf = fracture permeability ( mD) kd = average permeability from DST ( mD) km = core permeability (mD)

ke = km + Φfkf

ke = effective permeability ~ DST permeability (mD) Φf = fracture porosity

(2)kf = 844 x 105 x W3Z Φf = WZ x 100

Z = fracture spacing (cm) W = fracture width (cm)

Step 4F Fracture widths spacing permeability Step 5F Using synthetic logs of fracture intensity along with SGS a fractureSteps 2F and 3F The fracture intensity was quantified with core and the otherStep 1 A multimineral petrophysical analysis (MMPA) was completed after well log normalization and synthetic curves fracture properties were populated using MMPA and pressure buildup analysis and porosities were approximated using the intensity volumetric grid was produced which was the input for DFN modeling

were created for missing curves using neural networks above formulas

1

FRA

CTU

RE (F

) M

ATRI

X (M

) Sill 03

SW_MATRIX(06 099) SWCON_MATRIX(031123 074772)

PERMK_FRACTURE(13878 21887)

Sem

ivar

ian

ce

PERMK_MATRIX(000429 006233)1

Par

amet

ers

Par

amet

ers

DKFRAC_CON(0957 3877)

DIJFRAC_CON(01992 10361) ROCKDEN_FRACTURE(16441 16907)

SW_FRACTURE(06 099)

POR_FRACTURE(0004356 0006823) PERMIJ_FRACTURE(53378 8418) 95th Percentile

NETGROSS_MATRIX(022 036)

01

0275 SORW_MATRIX(001463 008345) ROCKDEN_MATRIX(16441 16907)

POR_MATRIX(005089 009429)0025 SOIRW_MATRIX(001463 008345)

SWCRIT_MATRIX(031123 074772) PERMIJ_MATRIX(000865 004732)

Range Step 6M A semivariogram analysis of the inverted VSP Step 7M Reservoir properties were populated using inverted VSP as a 0 5E+07 1E+08 15E+08 2E+08 porosity was used to acquire a better horizontal correlation covariable for SGS The smaller upscaled injection grid is shown in the Cumulative Gas (cubic feet) for short- and long-range spatial variablity center of the larger geostatistical grid

SW_FRACTURE(06 099)

ROCKDEN_MATRIX(16441 16907) PERMIJ_MATRIX(000865 004732)

POR_FRACTURE(0004356 0006823)

SWCON_MATRIX(031123 074772) PERMK_FRACTURE(13878 21887)

SORW_MATRIX(001463 008345)

ROCKDEN_FRACTURE(16441 16907) POR_MATRIX(005089 009429)

SOIRW_MATRIX(001463 008345) 95th Percentile SWCRIT_MATRIX(031123 074772) PERMK_MATRIX(000429 006233)

SW_MATRIX(06 099)

DKFRAC_CON(0957 3877) DIJFRAC_CON(01992 10361)

NETGROSS_MATRIX(022 036)

PERMIJ_FRACTURE(53378 8418)

0 100 200 300 400 500 600 700

Cumulative Oil (reservoir barrels)

Step 6F Based on the volumetric grid of fracture intensity DFN Step 7F The DFN was then upscaled and populated with volumetric properties modeling was performed to produce discrete fractures such as fracture porosity and permeability for input into the simulation grid

HISTORY MATCHING AND MVA CO2 EOR Injection Northwest McGreror CCO2 EOR IO2 EOR Injecnjection Nortion Norththwwest Mest McGrcGrereroror

Gas Volume (baseline) 2009-06-15 J layer 5 GasOil VVolume (baseline) 2009-06-15 J layer 5 5olume (baseline) 2009-06-15 j layer

(T1)(T1) (T2) (T3)

CO2 EOR Injection Northwest McGreror CO2 EOR Injection Northwest McGreror Gas Volume (after 3 days injection) 2009-06-18 J layer 5 Oil Volume (after 3 days injection) 2009-06-18 J layer 5

(T2)

CO2 EOR Injection Northwest McGreror CO2 EOR Injection Northwest McGreror Gas Volume (after 129 days injection) 2009-11-12 J layer 5 Oil Volume (after 129 days injection) 2009-11-12 J layer 5

(T3)

Steps 9 and 10 Time-lapse RST log and production were matched using CMGrsquos GEM and CMOST After hundreds of simulation runs and adjusting the heavy hitter variables a decent match was obtained More simulations will be run to adjust the final few months of history to form a better match for future incremental recovery estimates as well as predicting long-term CO2 fate

Step 8 Prior to simulation an uncertainty analysis was performed using CMGrsquos CMOST to estimate which parameters most affect injection and production history matching

SUMMARY

The primary goals of this PCOR Partnership pilot project included determining the short- and long-term fate of a small volume of injected CO2 (440 tons) and determining if a CO2 HNP in this type of reservoir could be effective In order to determine the fate of the injected CO2 and the potential incremental oil production time-lapse RST and VSP were used along with detailed static and numerical modelingThe VSP was an excellent tool for acquiring modeling parameters such as property semivariograms near wellbore horizon vertical deviations and production of covariables for petrophysical simulation However because of the highly fractured nature of the reservoir and the extremely low pore volume of the fractures which contained most of the CO2

from the repeat surveys it appeared unlikely that the location of the injected CO2 could be detectedThe baseline and repeat RST logs were found to be an excellent tool for determining the vertical extent of the injected CO2 near the wellbore and the modeling and history-matching activities give good support for the overall extent of the injected CO2 and will lead to better predictions of future incremental oil recovery

References (1) Lucia F J 2007 Carbonate reservoir characterization 2d Ed Springer Publishing (2) DjebbarT and Donaldson E 2004 Petrophysics theory and practice of measuring reservoir rock and fluid transport properties 2d Ed Gulf Publishing

Acknowledgments The authors would like to thank Eagle Operating Computer Modeling Group Schlumberger North Dakota Geological Survey and all PCOR Partnership partners for their support