Multi-Scale Analysis of Wellbore Performance

Bill Carey, Hari Viswanathan, Rajesh Pawar, Phil Stauffer, and George Guthrie, Jr.

Earth and Environmental SciencesLos Alamos National Laboratory

Carbon Mitigation Initiative MeetingPrinceton, NJ

November, 2005

OutlineWellbore Performance—PENS-CO2 Sequestration Framework

Components of the frameworkWellbore leakage within framework (e.g., Nordbotten and Celia)

Statistics of Wellbore PerformanceSustained Casing pressure as a performance measure

Observations of Wellbore PerformanceSACROC

Laboratory ExperimentsE.g., Scherer et al.

Numerical IntegrationFlexible Reservoir-Wellbore Simulation

System Scale

ReservoirScale

WellboreScale

BenchtopScale

CO2-PENS: Systems-Level Framework for Analysis of CO2 management/sequestration

CO2 Flux

CO2 Capturefrom Atmosphere

CO2 Storagein Reservoir

CO2 Releasefrom Storage Reservoir

CO2 Flux

EcosystemInteractions

CO 2Flux

CO2 Flux

OceanicInteractions

AtmosphericInteractions

CO2 Flux

CO 2Flux

CO

2 Flux

GroundwaterInteractions

Impact

Impact

CO

2 FluxEnvironmental & Societal

Impacts

CO2 Flux

CO2 Transportand Injection

CO2 Captureat Power Plants

PENS: Predicting Engineered Natural Systems

CO2-PENS: Reservoir Sub-System

StorageReservoir

ReservoirSeal

Potential ReleaseMechanism

CO2 Transportfrom Reservoir

CO2 Releaseat Surface

Frac

ture

Flow

Satu

rate

d &

Uns

atur

ated

Poro

us F

low

Wel

l Bor

eFl

ow

Wel

l Bor

eFa

ilure

Faul

t or

Frac

ture

Seal

Failu

reC

atas

trop

hic

Even

t

Salt

Shal

eM

udst

one

Car

bona

teO

ther

Sand

ston

e Li

mes

tone

Dol

osto

neB

asal

tC

oal B

edO

ther

Hos

t

Terr

estr

ial

Syst

ems

Atm

osph

eric

Proc

esse

s

Suba

queo

usSy

stem

s

Late

ral

Mig

ratio

nO

ther

CO2-PENS output includes tracking of CO2 in different subsystems over time, controlling factors (such as wellbore status),

economics, etc.

Time

Time

Rel

ease

Rat

e

Bor

ehol

es F

ixed

CO2-PENS: Wellbore ModuleNecessary Components

Failure Rates Failure Rates Leak Rates• Time to fail• Frequency of failure

Time

Probability to Fail

Frequency of failure at time t

Failure mechanism distribution

Permeability distribution for fracture flow

Failure Mechanisms• Poor job of cementing• Cement deterioration• Casing failure• Blow-out

Flow Paths• Matrix flow• Fracture flow• Pipe flow

Sustained Casing Pressure (SCP)(a measurable build-up of pressure in any casing string)

A possible analog and measure of wellbore failureOrigins

Casing and tubingThreads, corrosion, thermal and mechanical stress/rupture

Primary cement jobGas invasion, mud cake, formation damage

Cement damageMechanical shocks, T-P stress

Micro-annuliFracture/cracking

ConsequencesCasing failure, hydrocarbon loss, pollution

Sustained Casing Pressure (SCP)Databases

Regulatory agencies; oil and gas companies

Example data set: Gulf of Mexico (Mineral Management Services)

“A review of sustained casing pressure on the OCS” by A. T. Bourgoyne, Jr., S. L. Scott, and W. Manowski (LSU and Dowell-Schlumberger) (2000)

“Diagnosis and remediation of sustained casing pressure in wells” by A. K. Wojtanowicz, S. Nishikawa, and X. Rong (LSU) (2001)

% A

LL C

ASI

NG

S W

ITH

SC

P

11,498 casings with SCP in 8122 wells

0%

10%

20%

30%

40%

50%

60%

Production Intermediate Surface Conductor Structural

800 400600 200 0 Number of wells

0 20 40 60 80 100Percent of wells

Towards a probabilistic distribution of wellbore failure:Outer casingfailure statistics

Bourgoyne et al. (2000)

Calculation of Effective Permeability from SCP Data

Wojtanowicz et al. (2001)Effective Permeability: 1E-18 m2

Total depth 3077 m = 10076 ft Annulus flow area

0.025 m2

T = 54 C

T = 32 C

Wellhead

Bottom pressure estimate: 44.21 MPa

Cement (555 m)

Mud

FEHM Model: 8E-15 m2

Field Studies: SACROC• Pennsylvanian age reef

system• Discovered 1948• 54,000 acres• 3 billion BBLS original oil

in-place• 13th largest in North

America

SACROC (Scurry Area Capital Reef Operations Committee)

81 mi2, 1800 wells, 600 operationalProductive zone at 7000’ and is as thick as 800’Field temperature 50 ºC; Initial pressure 3200 psi (now 2600 psi)CO2 flooding initiated 1972 (only one field in the world is older)CO2 now obtained primarily from McElmo Dome, CO62% of all CO2 injected is not recovered (effectively sequestered)Drilling and production from zones above and below the Cisco/Canyon Reef complex have been free of CO2

Casing Shale

Grout-CasingInterface

Hydrated Cement

Grout-ShaleInterface

Matrix Diffusion

Interface Flow

Cement Degradation: Carbonation

Interface Flow

Decrease in porosity

Decrease in permeability

Increase in strength

Reduction of pH of pore fluidMay allow corrosion to occur at casing interface

Carbonation-induced shrinkageFormation of cracks (potentially filled with carbonate)

• Reduction of casing/cement and/or cement/caprock interface integrity• Loss of structural integrity at ultimate carbonation state

– CaCO3 + amorphous silica, alumina, and ferric hydroxides • Important factors controlling rates of carbonation

– Saturation and relative humidity– Water/Cement ratio– Age of cement

Fracture Flow

Cross-Section Through Well-Bore: 49-6

Gray Cement – Orange Zone – Shale Fragment Zone

Phase Gray Zone

Amorphous Major

Portlandite 15-58%

Calcite 0-28%

Halite 9-32%

Katoite 22-26%Brucite 3-9%Ettringite 3-4%Friedel’s Salt 2-4%

Phase Orange Zone

Calcite 44%

Aragonite 8%

Vaterite 33%

Halite 13%

Reaction Texturesat SACROC

Reaction Textures at SACROC

Zone Air Dried Oven Dried (mD) (mD) Upper Cement 0.09 74.00 Gray Zone C 0.10 Gray Zone A1 0.09 38.54 Gray Zone A2 0.07 48.22 Gray Zone A3 0.11 18.94 Gray Zone B1 5.75 Gray Zone B2 3.33 Gray Zone B3 8.40 Orange Zone A1 0.38 0.43 Orange Zone A2 0.19 0.19 Orange Zone A3 0.11 0.05 Orange Zone B1 0.17 Orange Zone B2 0.14 Orange Zone B3 0.22 Orange Zone B4 1.22 Shale along layers 8.57

Air-Permeability Measurements of Cement and Shale in milliDarcy

Measurements courtesy of Bob Svec, New Mexico Tech

Bulk Composition of Cement and Shale Core (X-ray Fluorescence )

Normalized to non-volatile mass

Depth 6545 6549 6549 6549.2 6550 6551.5

Oxide Gray Cem.

Oran. Cem.

Gray Cem.

Gray Cem.

Delta Orange-

Gray

Shale Frag Zone Shale

CaO 63.61 59.16 63.07 62.45 -3.88 2.92 4.56MgO 3.66 1.38 3.68 3.66 -2.29 2.17 2.97MnO 0.10 0.10 0.11 0.11 -0.01 0.03 0.03P2O5 0.09 0.10 0.10 0.10 0.00 1.11 0.11TiO2 0.24 0.27 0.24 0.25 0.02 0.69 0.73

Fe2O3 2.52 2.69 2.59 2.76 0.06 5.86 6.46K2O 0.04 0.19 0.04 0.05 0.15 4.02 3.82

Al2O3 4.65 5.16 4.83 4.83 0.39 16.81 16.30Na2O 2.36 4.38 2.04 1.93 2.27 1.44 1.15SiO2 22.65 26.51 23.24 23.81 3.27 64.76 63.68LOI 18.67 35.26 19.16 25.23 14.24 9.26 11.24

LOI: Loss-on-ignition (H2O, CO2, organics) values are not normalized.Typical Cement: CaO (61-65%), MgO (0.9-5.0%), Fe2O3 (1.5-4%), Al2O3 (4-6%),

SiO2 (19-22%)

Stable Isotope Studies of Core Samples

Two-phase multicomponent reactive flow and transportMass & energy conservationSingle and dual continuum formulationsDarcy’s law for two-phase liquid-air systemAqueous speciation (Debye-Hückel and Pitzer)Kinetic formulation of solid reactions

Numerical Analysis of Cement Degradation: FLOTRAN

1-D diffusion of CO2-saturated brine into cementIdealized cement: 40% C-S-H (xSiO2=0.36, Ca/Si = 1.78), 25% portlandite, 15% ettringite, 4% hydrogarnet (16% porosity)Idealized limestone country rock: 80% calcite (20% porosity)Mineral kinetics and rock tortuosity (0.02) control extent of reactionSimulation for 30 year life of well

C-S-H Solid Solution ModelEndmembers: Ca(OH)2 and SiO2

Mol-fraction XSiO2 = 1 / (1+R), R = Ca/Si

( )22

2log SiOOHCaaaa +−+Lippman Variable:

Excess Mixing Model:

]))1(())1(([)1( 2210 xxaxxaaRTxxGE −−+−−+−=

Parameter Estimation:µCa(OH)2,µSiO2, a0, a1, a2, (a3)

Ca/Si Ratio

Experimental data of Chen et al. (2004)

Cement Rock

Cement Rock

Cement Rock

Numerical Integration: Flexible Reservoir-Wellbore Simulation

Traditional approach of incorporation of wellbore requires grid refinement

Ideally, grid refinement should be based on hydrogeology and numerical issues

Goal: Add a well at any desired location with any level of complexity without grid refinement

Allow for a variety of wellbore flow processes

FEHM: Finite Element Heat and Mass Transfer CodeControl volume finite element approach: results in an unstructured gridEasy to add or modify connections to all existing/primary grid blocks (e.g., as used in dual permeability calculations)Wellbore grid blocks can be added to the geologic framework without modifying framework connectivities

ProcedureRepresent wellbore with desired spatial resolution vertically and horizontally (capture wellbore annulus, near wellbore physics)For each outer well grid block identify closest primary grid blocksConnect outer well grid block with the primary grid blocksCalculate/modify existing control volumes/areas and add wellbore control volumes taking particular care to maintain correct total control volumeAdd desired wellbore flow physics: incorporated implicitly (can be non-Darcy flow)

ConclusionsGoal: wellbore behavior analyzed within system-level frameworkOil and gas field records offer provide a potential statistical basis for performance measuresField observations provide mechanisms and dataExperimental work provides rates Integrated numerical methods provide benchmarksAnalytical/semi-analytical methods provide methods for rapid calculation sampling parameter and well location distributions

Funding Acknowledgements: DOE’s National Energy and Technology Laboratory, the Los Alamos LDRD program and the Zero Emission Research & Technology project

Colleague at LANL: Peter Lichtner, George Guthrie, Phil Stauffer, George Zyvoloski and Marcus Wigand

Colleagues at Kinder Morgan CO2: Scott Wehner and Mike Raines