S.E. Brame, J.W. Castle, SPE, O.K. Fawumi*, SPE and R.W. Falta, Clemson University

History Matching of Heavy Oil Production for Comparing New Approaches to

Generating Reservoir Property Distributions,

West Coalinga Field, California

S.E. Brame, J.W. Castle, SPE, O.K. Fawumi*, SPE and R.W. Falta, Clemson University

* Now with Mobil Producing Nigeria Unltd.

SPE 93469

This research was funded by the U.S. Department of Energy, Fossil Energy Oil Technology Program, through the National Petroleum Technology Office under contract number DE-AC26-98BC15119.

Objectives

• Assess the suitability of different permeability distributions generated from geological and fractal modeling.

• Examine the feasibility of creating realistic permeability distributions from fractal theory for reservoir simulation.

• Use steam flood simulations of a portion of the West Coalinga oil field in California for model assessment.

Location of the West Coalinga Field

Heavy Oil Sands of Coalinga• The West Coalinga oil field in California

produces from heavy oil sands of the Miocene Temblor Formation.

• The oil in this field has low API gravity (12o to 15o API) and is highly viscous (900 cp) at the 40oC initial reservoir temperature.

• This makes the field an ideal candidate for enhanced oil recovery through steam injection.

Steam injection well at Coalinga

Methods

• Characterize geology in outcrop and core

• Identify lithofacies

• Group lithofacies into Facies Groups

• Identify Facies Tracts

• Characterize Fractal Facies

• Construct geologic models

• Assign permeability distributions

• Simulate steam flooding of different models and assess results

Geologic Characterization • Coalinga offers a unique opportunity to observe

and characterize the reservoir rocks on the surface, immediately adjacent to the oil production area.

• Thus it was possible to relate cores of the producing formation to nearby outcrops.

Clemson Students examining outcrop of Temblor Fm. in hills north of Coalinga Field

Lithofacies Characterization• A total of fifteen lithofacies were identified from outcrops and cores based on observed lithological differences.• Statistical methods were used to consolidate the 15 different lithofacies groups into five facies groups.

Facies Group Major Lithology Mean Permeability

1 Clean sand 3180 mD

2 Interlaminated sand and clay 500 mD

3 Burrowed clayey sand 255 mD

4 Bioturbated Sand 525 mD

5 Fossiliferous Sand 225 mD

Facies Tract Characterization

Five facies tracts were interpreted based on detailed sedimentological analysis:

Facies Tract Lithology Mean Permeability

Incised Valley Basal conglomerate, cross-bedded sand, silt, and clay

562 mD

EstuarineInterlaminated sand, silt, and clay, burrowed clay intervals, sandy clay intervals

316 mD

Tide-to Wave-dominated shoreline

Crossbedded sand with burrowed sand and clay; fossiliferous sand 316 mD

Diatomite Clay, silt, and fine sand 22 mD

Subtidal Massive burrowed sand with intervals of silt and clay

224 mD

Facies CorrelationsBetween Wells

Correlation of well 239 with Type Well 118A to determine the location of bounding surfaces, facies tracts, and facies groups.

1870

1850

1830

1810

1790

1770

1750

1730

1710

1690

4

1

4

3

4

3

5

4

53

Gamma Radiation(API) Facies Group

Number

2

3

4

5

Facies TractNumber0 100 200 300

4BS-5

Well 118A

BS = bounding surface (depths are in feet)

Model Area

This map shows the 3 adjacent five-spot configurations that were modeled and the wells that were used.

298600

298800

299000

299200

299400

299600

299800

300000

300200

300400

300600

300800

1587500 1587700 1587900 1588100 1588300 1588500

Easting

227

118B

8-2B

228W228

22

8-2

128

8-3

127

8-4 118A

8-1

239W

239

238

238W

238A

128B

237

237W

127B

236W

236

229W229

Nor

thin

g

ProductionWell

Injection Well

Section 36D

Geologic Model Construction

• 3-D geologic models were constructed using GOCAD.

• Inputs included the bounding surface horizons, geophysical logs, and facies group and facies tract data.

• The grid consisted of 9600 cells:

- 32 vertical layers,

- 10 cells in the x direction

- 30 cells in the y direction

Geologic Models in GOCAD

Facies Group 1

Facies Group 2

Facies Group 3

Facies Group 4

Facies Group 5

200 ft

Facies Tract Facies Group

Fractal Group Model Development

• The cores of five West Coalinga wells were analyzed foot by foot.

• Analysis of the individual facies data revealed that a unique Gaussian fractal structure was present in each one.

• These results and others led to the development of a new model for representing natural heterogeneity called the fractal/facies concept.

Reservoir Simulation

• The geologic model grids were used as the framework for the flow simulation mesh.

• Petrophysical properties were assigned to all cells of the mesh.

• Permeability distributions of the facies tract, facies group, and fractal group models were assigned.

• Numerical simulations of steam injection were used to assess the different models.

T2VOC Flow Simulator

• Uses a general finite difference formulation. • Can solve multi-phase, multi-component mass and

energy balance equations. • Has been used to simulate a variety of subsurface

processes such as:

nonaqueous phase liquid (NAPL) migration, soil vapor extraction, air-sparging, steam injection, and direct pumping of water and NAPL.

Permeability Distribution of the Facies Tract Model

-1000

-800

1.5876E+06 1.588E+06 1.5884E+06

Easting

Perm (mD)400300200100500

Permeability Distribution of the Facies Group Model

-1000

-800

1.5876E+06 1.588E+06 1.5884E+06

Easting

Perm (mD)300010005004003002502000

Fractal Permeability Distribution

• Fractal permeabilities were stochastically generated for the 5 different facies groups.

• The values were assigned to the flow grid but using a finer grid (~3,000,000 cells).

• The fractal permeabilities were upscaled to the standard flow grid cell size (9600 cells).

• An arithmetic mean was used for the horizontal permeability.

• A harmonic mean was used for the vertical permeability.

Permeability Distribution of the Fractal Model

-1000

-800

1.5876E+06 1.588E+06 1.5884E+06

Easting

Perm (mD)10000100050040030020010010

Simulation Period• A five year period (Oct. 1995 to Oct. 2000) was used.• The injected steam volume changed monthly.• Some injection wells were not online until 1997.• All changes in production and injection were

honored.

0

2000

4000

6000

0 1 2 3 4 5

Time (years)

vo

lum

e o

f s

tea

m in

jec

ted

(b

bl/d

ay

of

wa

ter)

Vol

um

e of

Ste

am (

bb

ls o

f w

ater

)

Time (years)

Parameters Adjusted for Better Fit

• The oil-water relative permeability curves provided for the field were based on a data fit from a core.

• The initial oil saturation was interpolated from well logs.

• Problem: these values resulted in simulations where the water to oil ratio was off by a factor of 10 or more compared to field values!

Solution: The water relative permeability endpoint was reduced from .56 to .15, and the oil saturations were increased everywhere by 20% (upper limit of 70%).

Kro Krw

■ Kro Field Data

● Krw Field Data

Kro used in Model

Krw used in Model

Normalized Water-oil Relative Permeabilities

Water Saturation (%)

-1000

-800

1.5876E+06

1.588E+06

1.5884E+06Easting

299000

299500

300000

300500

Northing

Temp (C)1751501251007550

-1000

-800

1.5876E+06

1.588E+06

1.5884E+06Easting

299000

299500

300000

300500

Northing

Temp (C)1751501251007550

Reservoir Temperatures at 5 Years

-1000

-800

1.5876E+06

1.588E+06

1.5884E+06Easting

299000

299500

300000

300500

Northing

Temp (C)1751501251007550

-1000

-800

1.5876E+06

1.588E+06

1.5884E+06Easting

299000

299500

300000

300500

Northing

Temp (C)1751501251007550

FaciesTract

FractalGroup

FaciesGroup

Oil Saturations at 5 Years

-1000

-800

1.5876E+06

1.588E+06

1.5884E+06

299000

299500

300000

300500

Oil Saturation0.60.50.40.30.20.1

-1000

-800

1.5876E+06

1.588E+06

1.5884E+06

299000

299500

300000

300500

Oil Saturation0.60.50.40.30.20.1

-1000

-800

1.5876E+06

1.588E+06

1.5884E+06

299000

299500

300000

300500

Oil Saturation0.60.50.40.30.20.1

-1000

-800

1.5876E+06

1.588E+06

1.5884E+06

299000

299500

300000

300500

Oil Saturation0.60.50.40.30.20.1

FractalGroup

FaciesTract

FaciesGroup

Simulated versus Field Production of Oil (Combined Production)

0

200

400

600

800

0 1 2 3 4 5

Time (years)

Oil

Pro

du

ctio

n (

bb

l/day

)

Field

FaciesTract

FaciesGroup

Fractalkz/10

Simulated versus Field Production of Water (Combined Production)

0

1000

2000

3000

4000

5000

6000

0 1 2 3 4 5

Time (years)

Wat

er P

rod

uct

ion

(b

bl/

day

)

Field

Facies Tract

Facies Group

Fractal kz/10

Summary

• Three models were assessed using simulations of steam injection and heavy oil production.

• Lowering the water relative permeability curve endpoint decreased water production.

• Increasing the initial oil saturation resulted in a better prediction of oil production.

• The fractal realization matched oil production up to Year 3.

• Additional realizations could be generated to improve the history match.

Conclusions

• Permeability distributions from the geologic models provided reasonable matches of oil production.

• The fractal/facies approach is a viable method of generating geologically realistic permeability distributions.

• Steamflood simulations demonstrate that the fractal/facies approach is feasible for modeling heavy oil reservoirs.

Acknowledgments

• Chevron contributed core data, production data, and geophysical logs.

• Many thanks to Venton Shoemaker, George Anderson, Louis Klonsky, Paul Henshaw, and Mike Clark of Chevron.

• Fred Molz and Silong Lu of Clemson University performed the fractal analysis

• Ray Christopher of Clemson University assisted with the statistical treatment of the lithofacies.