Carbonate Reservoirs – Challenges in Facies Modeling & Fracture Characterization

By

Subrata K. ChakrabortyMega Ardhiani Puspa

Schlumberger Data Consulting Services, Jakarta

Carbonate Complexity : Characterization, Modeling and SimulationApril 22nd – 25th, 2008, Yogyakarta, Indonesia

Contents

• Introduction• Facies Modeling of Carbonates• Fracture Characterization of Carbonates

Air Serdang

Arun

RawaRamba

Kaji-Semoga Bima

Rama

Krisna

KrisnaKLY Bawean

Ujung Pangkah

PolengKangean

Natuna D-Alpha

Vanda

Serang

Kampung Baru

Oseil

KlalinArar

KlamonoWalio

Salawati

Air Serdang

Arun

RawaRamba

Kaji-Semoga Bima

Rama

Krisna

KrisnaKLY Bawean

Ujung Pangkah

PolengKangean

Natuna D-Alpha

Vanda

Serang

Kampung Baru

Oseil

KlalinArar

KlamonoWalio

Salawati

Carbonate Reservoir in Indonesia Oil Production in Carbonate vsClastic from Indonesia in 1976

60 Carbonate

Fields

Basin Formation Field AgeNorth Sumatra Basin Arun Limestone Arun Early Miocene

Ramba, Rawa, Soka, Kaji-SemogaAir Serdang, MandalaBima, Rama, Krisna, "AA",Selatan, Kandanghaur Timur, Arimbi X, Yvonne, Nora, Kitty, Cinta, GitaArjuna FF, Arjuna FZ, Pondok Tengah, Tambun

East Natuna/Sarawak Basin Natuna D-Alpha block L structure Carbonate complex Mid - Late MioceneKutai Basin Serang 80-6 Limestone, CN-9 zone Late Miocene

Tarakan Basin Vanda Vanda Limestone Late Miocene-Early PlioceneNgimbang Kangean PSC, Poleng Mid-Upper Miocene

Kujung Poleng, Ujung Pangkah, Jaya, Bawean, JS Late Oligocene-Early MioceneRandublatung, Kedunglusi, Kedung Tuban

Seram Manusela Oseil Early-Middle Jurassic

East Sengkang Basin Tacipi Limestone Kampung Baru, Walanga, Sampi Sampi Lower MioceneKais, Textularia II, "U" marker Kasim, Walio

Cendrawasih, Moi, Jaya, Klamono, Linda, Sele, Salawati A,C,D,E,F&NArar, Klalin, Kasim Utara, Kasim Barat

KaisSalawati Basin Miocene

Late Oligocene-Early Miocene

East Java Basin

Sunda Batu Raja

Batu Raja South Sumatra Basin Lower Miocene

8%92%

Ref. : Warren C Leslie

Worldwide 60% HC in Carbonates, in ME 75%

Facies Modeling in Carbonates

Part-I

Carbonate Ramp Model

Rimmed Platform Model

Regional Carbonate Depositional Model

BatuRaja Carbonate ??

Stratigraphy – South Sumatra

Facies Modeling-WorkflowFacies Modeling

Facies InterpretationAt Wells

3D Modeling

PixelBased

ObjectBased

CoreFacies

ImageFacies

ElectroFacies

SeisFacies

Most DefiniteLimited Vertical

Distribution

More DefiniteReasonable

VerticalDistribution

Less DefiniteWide VerticalDistribution

Result dependson Log

Availability

Least DefiniteWide AerialalDistribution

Limited Vertical Resolution

Results dependOn data quality

MostWidelyUsed

LessWidelyUsed

UsedMostly

For Calibration

Works bestWith Knowledge

of Paleo-geography& Paleo-environment

Works well with good quality Seismic

Data & Good Well Density

3D FaciesModel

Facies Interpretation – Neural Net Based Approach9 originally identified Carbonate sub-facies narrowed down to 4 Carbonate Facies

from fluid flow behavior point of view.Deep water Shale FaciesTight Platform FacieCoral Rich Wackstone & PackstoneReworked Skeletal Facies

VCALY, RHOB, PHIE used as key logs for Neural Network based facies Electro Facies Interpretation in “Petrel”.

Neural Network was trained with above core facies where key logs showed significant variation for different facies type and Supervised Neural Network was run. “Petrel-Make Log” utility was used to generate trained electrofacies in other wells having the key logs.

Generated electro facies in each well was examined and corrections made by editing wherever required.

Neural Network Training

Correlation Analysis : Neural Network – BRF Carbonate Facies

Electro Logs in Training FaciesElectro Logs

Neural Network : An algorithm that takes multiple log inputs and returns one or several outputs. Each input is multiplied by a weight, the result is summed and the result passed through a non-linear function to produce the output. It is of two types – “Supervised or Trained” (Sigmoidal Basis Function Regression Network) or “Untrained” (Competitive Selective Learning).

BRF Carbonate 3D Facies Model – General Workflow

Facies Interpretationin Wells as Logs

Upscale Logs to Model Cells

Vertical Proportion Analysis Variogram AnalysisData Analysis

Facies Modeling Object BasedIdentification of

Depositional Elements(Objects)

Hard Data : Facies LogInterpolation Guide : Objects Q/C

Progradational Line Reef Trends Reef Geometry

Well DataGuided by

Object TrendFacies Model Object Based Model “Stochastic Object Modeling”

Algorithm

Reefal Facies Object Modeling

SUMMARYDeveloped 3D Facies Model captures the Field

Carbonate Geology.The 3D Facies Model becomes the basis of modeling

other reservoir properties.

Fracture Characterization of Carbonates

Part-II

Definition : Naturally Fractured Reservoirs are defined as formations in which the fracture permeability substantially enhances productivity.

OUTCROP EXAMPLE : Fractured Carbonates from Oman

EXAMPLE From Cores

A fracture is a commonest type of geological structures and may be seen in any rock exposures.

There are two kinds of fracture:(a) Joint : no displacement, or displacement is too small to be visible. (b) Fault : has measurable displacement across the fracture plane“Carbonates” are more prone to fracturing than “Clastics”.

A – Breccia ClastsB – Breccia MatrixC – Open FracturesD – Healed FractureE – Fractures filled with breccia material

Core from Oseil-1 (Ref IPA 2002)

Dual Porosity – Permeability System

Ref. : Tyler 1988

Fractured Carbonate RocksCarbonate Rocks are More Prone to

Fracturing than Clastic Rocks.In a carbonate reservoir, fracturing

intensity may be controlled by different carbonate facies present.

Quar

tzite

Dolo

mite

Sst.

Limestone

E = ca. 90 GPa(low frac toughness)

E = ca. 20 GPa E-Modulus low -> Storage

E-Modulus high -> Flow Mudloss and rubble in cores – only seen in outcrop

Brittle Ductile

Frac

ture

num

ber

400

100

Lithology

E = tensile stress / tensile strain

Quar

tzite

Dolo

mite

Sst.

Limestone

E = ca. 90 GPa(low frac toughness)

E = ca. 20 GPa E-Modulus low -> Storage

E-Modulus high -> Flow Mudloss and rubble in cores – only seen in outcrop

Brittle Ductile

Frac

ture

num

ber

400

100

Lithology

E = tensile stress / tensile strain

Fracture toughness is regarded as being the opposite of Rock strength:A weak/ soft rock accomadate more strain before fracturing.Youngs Modulus (E) is related to capability of storage of strain energy.

Difference of Fractured Reservoirs from Conventional Reservoirs

High Transmissivity of Fracture Network – Pressure drop around producing well is low. Production is driven by complex mechanisms that governs fracture/matrix-block communication.

GOR of fractured reservoirs remains lower throughout during production, if reservoir is properly managed.

Fracture reservoirs lack transition zones. OWC/GWC’s are knife sharp surfaces. High fracture permeability ensures rapid contact equilibrium even during production.

Water cut is strictly a function of production rate.PVT properties remain constant throughout a fractured reservoir due to

conductive circulation.(REF. : SPE 84590)

Fracture Reservoir Classification (After Nelson, R.A., 1999, 2002)Type 1- Fractures provide the essential porosity and

permeability to the reservoir.Example : Amal (Libya) 1700 MMBL, Ellenburger (Texas) 108

MMBL, Edison (California) 42 MMBL, TaHe (China) (SPE 49221, 106986)

Type 2- Fractures provide the essential permeability to the reservoir.

Example : Agha Jari (Iran) 9500 MMBL, Haft Kel (Iran) 2660 MMBL, Rangely (Colorado) 600 MMBL, La Paz/Mara (Venezuella) 800 MMBL, NWRA (Kuwait). (SPE 35309, 5023, 97834)

Type 3- Fractures provide a permeability assist to the reservoir.

Example : Kirkuk (Iraq) 15000 MMBL, Gachsaran (Iran) 8000 MMBL, Hassi Messaoud (Algeria) 6000 MMBL, Dukhan(Qatar), UmLulu/Asab/Bab (UAE) (SPE 36228, 62608, 65186, 87238, 96955, 102453, 11164)

Type 4- fractures do not provide significant storage capacity or permeability in an already producible reservoir but instead create anisotropy.

If Fracture Porosity-Permeability is Not Modeled – Consequences ?

Unexpected decline in from predicted Field Performance.

In case of thick gas cap & active aquifer, early gas coning & water breakthrough reducing oil recovery.

In case of oil reservoir with active water drive, early water breakthrough reducing oil recovery.

Water injection schemes may get jeopardized by preferential water front movement along fractures.

Severe mud loss during drilling of horizontal wells aligned across unfriendly fractures.

Improperly aligned horizontal wells may have low productivity.

So, it is Important ....

Oil Column

Water

Gas Cap

Fractures

Field Recovery Curve

Recovery Factor

Thin Oil Column in Fractured Reservoir

Fractures in extensional area are parallel to the main fault.Fractures in strike slip area will be concentrated near the master faults and is 60o offset to

the strike slip fault.Fractures in compressive area will be concentrated on the anticlinal bent areas of the folds

and will be perpendicular to the direction of maximum compressive stress.

Left lateral Strike Slip System, e.g., rotation of plates

Extensional System, e.g., Rifting & Drifting of Plates

Compressional System, e.g., Collision of Plates

Tectonic styles and fracture type identified to three plate boundaries :Divergent (tensional – normal faults)Transform (shear – strike slip faults)Convergent (compressive – reverse faults)

Fractures & Tectonic Style

Microsoft owerPoint Presentatio

World stress map of Austral-Asia Plate

Effect of Present Day Stress Field on Faults & FracturesIn Indonesia the dominant present day maximum horizontal stress direction is NE-SW, hence in

Indonesian fractured reservoirs fractures having fracture sets of this orientation are likely to be more “Open” in general (Leading Edge 2005).

http://www.world-stress-map.org/

Input Data

Interpreted Image Logs Ant Tracking Fault from 3D Cube

Display & Analysis

Well Section

Stereo Net Analysis

Tadpole

Rose Diagram

3D Display

CreateIntensity Log

ModelIntensity Log

Create DFN

Fracture Modeling

Create Fracture PropertyUpscale Fracture

Property

Create DFN

Create Fracture Property

Upscale Fracture PropertyAdvanced Options of Using

Collocated Co-KriggingVolumetricEstimation

Fracture Modeling – Workflow(“Petrel”)

UncertaintyAnalysis

Input Data

Interpreted Image Logs Ant Tracking Fault from 3D Cube

Display & Analysis

Well Section

Stereo Net Analysis

Tadpole

Rose Diagram

3D Display

CreateIntensity Log

ModelIntensity Log

Create DFN

Fracture Modeling

Create Fracture PropertyUpscale Fracture

Property

Create DFN

Create Fracture Property

Upscale Fracture PropertyAdvanced Options of Using

Collocated Co-KriggingVolumetricEstimation

Fracture Modeling – Workflow(“Petrel”)

UncertaintyAnalysis

Fracture From Image Log

Aperture

Fracture

Fracture

Tadpole

Fracture Density/ Intensity Log

Conductive Fractures : Assumed openResistive Fractures : Assumed sealed

Basic Fracture Analysis Data

Advance Fracture Analysis Data

Aperture PorosityDensity

DipAzimuth

Interpretation is done for Conductive Fractures

Fracture Interpretation from Array Sonic LogsArray SonicFMI Log

Qualitative Tool

Quantitative Tool

Shallow Penetration

Deeper Penetration

Tadpole & Rose Diagram Display

Fracture Analysis – From Image Log DataStereo-net Display & Filtering

Tadpole Helps to understand fracture Intensity and orientation.

Rose Diagram helps to understand fracture orientation and sets.

Stereo-net helps to understand fracture dip angle and azimuth.

Rose diagram showing dip azimuth

Tadpole

Fracture Intensity Log

Fracture Intensity from Image Log

Rose diagram showing dip azimuth

Tadpole

Intensity log

Fracture log

After Nelson 1999

Aerial Distribution of Rose Diagrams brings out the Imprint of Different Tectonic Events Clearly

Field-X

CreateIntensity Log

UpscaleIntensity Log

ModelIntensity Log

Creation Upscaling Modeling

3D Modeling of Fracture Intensity

Created from the fracture set

Fracture Intensity Cube

Create Discrete Fracture Network (DFN)

Geometry

Orientation

Distribution

Fracture Density

Fracture Length

Fracture Orientation

Fracture Intensity

Discrete Fracture Network (DFN)

Stereonet Display

Analysis of DFN

Real Life Analogue

Consists of no of fracturesof certain orientation, dipand length.

Create Fracture Attributes

Aperture (Required for Fracture Porosity)By 3D ModelingBy Calculator Operation

Fracture_length=Sqrt( Surface_area)Aperture=Fracture_length*Normal( 0.005, 0.0005)

Permeability (Fracture Permeability depends on “Aperture” & “Permeability”)

By 3D ModelingBy Calculator Operations

Permeability=Pow( Aperture, 3)

Oda MethodUpscale Fracture Attributes

OutputFracture PorosityPermeability Tensors (I, j & k directions)Sigma Factor (connectivity between

fractures and matrix)

Main Faults in 3D Model

Ant Tracked Faults/Fractures

3D Fracture Modeling – From AntTracking Interpretation

DFNCreate DFN

Create Fracture Property

Upscale Fracture Property

Create DFN From AntTracked Fractures

Upscaled Fracture Porosity Model

Upscaled Fracture Ki Model

Upscaled Fracture Kj Model

3D Fracture Modeling – Secondary Porosity & Permeability

Matrix Poro/Perm

Fracture Poro Perm

Sigma Factor

Dual Porosity / Permeability Simulation InputCan be done in Petrel RE, using the Define Simulation case process. Matrix properties should be ready, and the fracture properties from upscaling should now be available.

Validation of Fracture ModelDifficult in “Green Field” compared to in “Brown Field”.Transient Well Test matching in “Green Field”.Dynamic history match data is the main data to validate model.Explanation of uneven injection water movement.Explanation of early gas coning, early water breakthrough.Explanation of mud-loss in horizontal wells.Explanation of higher production from some horizontal wells.

Fractured Reservoirs in IndonesiaFractures related to Syn-Rift Tectonics.Fractures related to Collision Tectonics & Inversion.Fractures related to Strike Slip movements. Fractures related to Local Structuring (trap formation).

Examples“Jatibarang” fractured basement, East Java (Operator : Pertamina)“Pase A” Gas Field, Pase PSC, North Sumatra (Operator : Mobil Oil Indonesia)“Darajat Field” (Operator : Amoseas Indonesia Inc)“Oseil Oil Field”, Seram Island, Eastern Indonesia (Operator : KUFPEC).“Kasim”, “Jaya”, “Kasim Utara”, “Cendrawasih”, “Moi”, “Arae” fields,

Salawati Basin, Irian Jaya.

“Ujung Pangkah”, East Java Sea ?? (Operator : Amerada Hess)

25 my 20 my

Talang AkarSyn Rift/Drift (Normal Fault) Batu Raja Carbonate

Syn Rift/Drift (Normal Fault)

Plate Tectonic Reconstruction of Sumatra & Java of Indonesia showing the active Syn Rift/Drift Phase (Likely Fracture Orientation NW-SE)

12 my

Air Benakat TimeStrike Slip Movement + End of Drift

09my

Structural InversionStrike Slip Movement + Collision + Plate rotationReverse & Strike Slip faults

Plate Tectonic Reconstruction of Sumatra & Java of Indonesia showing the active Strike Slip Movement due to Plate Rotation and the Collision Phase (Likely Fracture Orientation N-S to NE-SW and NW-SE)

Data Acquisition Program for Naturally Fractured Carbonate Reservoirs

FMI/Sonic Scanner LogsAdvanced Fracture Interpretation from FMI/Sonic Scanner LogsHigh Quality 3D Seismic Data with acquisition optimized for Target Reservoir

CompletionFracturing might be necessary to Enhance Fractures

Stage Frac Completion Assembly

Image Log

Sonic Scanner Logs

Thanks for Your Attention