Snhvit CO2 storage project: Understanding the role of

injection induced mechanical deformation

Joshua White, Laura Chiaramonte, Whitney Trainor-Guitton
Project Number: FWP-FEW0174 Task 4

Lawrence Livermore National Laboratory

Carbon Storage R&D Review Meeting, Pittsburgh, August 12-14, 2014

Program Goal No. 4

Develop Best Practice Manuals for monitoring, verification, accounting, and

assessment; site screening, selection and initial characterization; public outreach;
well management activities; and risk analysis and simulation.

Benefit Statement

An understanding of hydromechanical interactions is essential for effective

monitoring and prediction of reservoir performance. This is especially true for
storage systems expected to experience higher overpressure.

This project has developed new analysis methodologies for:

probabilistic assessment of fault reactivation potential and in situ stress

sensitivity
dynamic well-test analyses using multi-rate gauge data

re

Snhvit

Fig. 1. Location of Snohvit gas field (Courtesy of Statoil).

The Snhvit Field is an elongated E-W trending horst

block bounded by normal faults (Figures 2 and 3), and
located in the center of the Hammerfest Basin [3].

Fig. 2. Structural diagram at Middle Jurassic level (~ age of

producing reservoir) of the central Hammerfest Basin. Blue
lines outline the gas fields [4].

The Hammerfest Basin is a fault bounded extensional

basin with an ENE-WSW axis. It presents a dominant EW trending fault system associated with flexural
extension related to a doming episode. The Snhvit
12 Snhvit Field: structure diagram at Middle Jurassic level. The blue lines show
the outlines
of in
thethree
hydrocarbon
pools.
accumulation
occurs
of these fault
blocks
Structural diagram of Hammerfest Basin at Middle Jurassic level (approx. age of
(Figure 2). With the opening of the Atlantic margin in
producing reservoir). Blue lines outline the gas fields [from Spencer et al., 2008]. the Tertiary, the Hammerfest basin experience several
Statoil, Norsk Hydro
Saga.
recent the
years,
the authorities
phasesand
of uplift,
andIntherefore
maximum
burial depth

Snhvit

Snhvit CO2 Tuben Fm. storage capacity and injection

strategy study

Valid from
Doc. No.

Valid from

Rev. no. 1.0

Depth map at top of Fuglen Fm. (below) and stratigraphic section (right)
[from Wennberg et al., 2008]

Figure 1: Depth map of Top Fuglen Formation with well locations (from Wennberg et al.
2008).

at the Hammerfest Basin is shown in Figure 2. The lowermost reservoir unit is the Tuben
The stratigraphy
Producing
natural gas with 5-8% CO2 content,
Formation of Early Jurassic (Hettangian V Sinemurian) age. This is a delta plain environment with fluvial
which
needs
be reduced
before
liquefication.
distributary
channels
and someto
marine-tidal
influence (Helgesen
and Johansen
2005). Above the Tuben Fm is the

Deleted: Figure 2

Nordmela Formation of Sinemurian to Pliensbackian age. This is a lower coastal plain depositional environment
with brackish, shallow-marine deposits (Wennberg et al. 2008). Following this is the St Formation of Early to

MiddleSeparated
CO2 was
originally re-injected into
Jurassic (Pliensbackian-Bajocian)
age. This is a shallow-marine environment with alternating lower to
upperTuben
shoreface deposits
communication,
Lone Christensen). The
St Formation is covered by the
Fm.(personal
at approx.
2400-2600m
depth.
Fuglen Formation of late Middle Jurassic age. The main reservoirs at Snhvit are the Tuben Formation (CO2

injection) and St Formation (producing reservoir). As far as CO2 storage is concerned, the main target storage

formation,
Injection
beganis dominated
in 2008,
but inchannel
2010facies
Statoil
the Tuben Formation,
by distributary
leading to possible restricted flow
compartments
(individual
channels
or
channel
complexes).
The
alternate
storage
the St
Formation, is
announced storage capacity in Tubentarget,
was
lower
likely to have better lateral communication because of its shallow marine depositional setting.
than expected. Operators recompleted well and
have continued injection in the shallower St
formation.

Figure 2: Stratigraphy in the Hammerfest basin (from Wennber

Focus areas where modeling can help monitoring and

project management:
Decision-based Modeling

Using modeling to inform instrument design and deployment schemes

Data Integration and Data Assimilation

Merge diverse monitoring data sets into unified interpretations

Deterministic or stochastic inversion of monitoring data

Uncertainty Quantification and Risk Assessment

Quantify how uncertainty impacts performance and risk

Snhvit Case Study 1

Using modeling to address uncertainty in stress measurements

[Chiaramonte et al. 2014]

A key project concern was whether excess fluid pressure

could reactivate faults and create leakage paths

Figure: Model prediction of excess pressure (MPa) necessary to initiate fault slip with the base case
scenario parameters, which corresponds to a SS/NF environment with N-S SHmax direction.

Sobol sensi7vity index

Sensitivity analyses reveal that critical uncertainty is

SHmax orientation, but overall leakage risk is low

Figure: SHmax azimuth dominates critical pressure sensitivity

Suggests highest-value target for future characterization efforts

Snhvit Case Study 2

Welltest analysis and continuous inversion of gauge data

Figure: 4D dierence amplitude maps, 2003-2009, lower perfora7on.

(Hansen et al. 2012)

Pressure response indicative of a partially compartmentalized

system
400

estimated BHP

pressure, bar

380

360

340

320

300

100

200

300

400

500
600
time, days

700

800

900

1000

1100

BHP estimated from permanent pressure/temperature sensors at 1782

mTVDss, hourly data.

Well tests
P~ commonly used to look for flow barriers and
otherFigure
indications
of reservoir
1.21. One sealing
fault. Pressurestructure
profile at time t4.
I

The fault is reached, and it is seen at the well. Hemi-radial flow.

t~ 9the fault is not reached, radial flow

t2 'the fault is reached
t3 " the fault is seen at the well, transition
t4' hemi-radial flow

Figure 1.22. One sealing fault. Drainage radius.

This pressure regime corresponds to time t2 on Figure

1.22:
boundary
[Figure
from the
Bourdet,
2002] ha

compartment with acceptable reservoir properties but with

reservoir system. However, other geological models and
Falloff
analysis
showed clear indications of flow barriers
eries,
Figure
3.
Results suggested flow barriers at 110, 110, and 3000m
F-2H - Model matched to PLT data (ref)
F-2H - Model matched to FO 2009

G as potential [bar2/cp]

1E+5

10000

1000

Aug 2009 Fall-off

PLT 2011 Fall-off

1E-4

1E-3

0.01

0.1

1
Time [hr]

10

100

1000

Figure: Fallo analyses using permanent gauge (2009) and PLT data (2011).

osses) and modeled pressure. Timing of the acquired seismic 4D surveys are
(Hansen et al. 2012)
acture pressure. b) Log-log plot of (2011 PLT and 2009 FO) gas pseudo

Falloff testing has proven value, but requires shutting in

the well for significant periods
100
Measured
tonnes per hour

80
60
40
20
0
0

100

200

300

400

500

600

700

800

900

1000

days
100
Simplified
tonnes per hour

80

Motivating
question: Can we derive similar information from ongoing
60
injection data, without shutting in for long periods?
40
20
0

1100

Generalized superposition well-test method

single rate injection

multi-rate injection

pressure

rate

Solve for a characteris7c

buildup curve, as a
constrained least-squares
problem.

time

rate

Can use superposi7on

principle to transform a
mul7-rate injec7on into an
equivalent single-rate
test.

pressure

Mul7-rate injec7ons are

dicult to analyze.

time

Single rate:


Mul.-rate:

time

p(t) = q pC (t)
p(t) = (qi+1 qi ) pC (t ti )
i

Automatic calibration to Snhvit data (~5 seconds)

measurement

forecast / precast

calibration period

400

pressure, bar

380

360

340
equivalent buildup test
150
100

320

50
0
300

100

200

300

400

500
600
time, days

700

500
800

900

1000
1000

1100

Superposition tool can potentially be used in two modes:

Reservoir characterization mode

Calibrate to gauge data, extract equivalent falloff test

Apply standard well-test analysis techniques to results

Pressure forecasting mode

Calibrate to gauge data, project forward in time

Quickly explore alternative injection scenarios

Fast-running pressure forecasting

measurement

forecast / precast

calibration period

400

pressure, bar

380

360

340
equivalent buildup test
150
100

320

50
0
300

100

200

300

400

500
600
time, days

700

500
800

900

1000
1000

1100

Fast-running pressure forecasting

measurement

forecast / precast

calibration period

400

pressure, bar

380

360

340
equivalent buildup test
150
100

320

50
0
300

100

200

300

400

500
600
time, days

700

500
800

900

1000
1000

1100

Fast-running pressure forecasting

measurement

forecast / precast

calibration period

400

pressure, bar

380

360

340
equivalent buildup test
150
100

320

50
0
300

100

200

300

400

500
600
time, days

700

500
800

900

1000
1000

1100

Fast-running pressure forecasting

measurement

forecast / precast

calibration period

400

pressure, bar

380

360

340
equivalent buildup test
150
100

320

50
0
300

100

200

300

400

500
600
time, days

700

500
800

900

1000
1000

1100

Fast-running pressure forecasting

measurement

forecast / precast

calibration period

400

pressure, bar

380

360

340
equivalent buildup test
150
100

320

50
0
300

100

200

300

400

500
600
time, days

700

500
800

900

1000
1000

1100

Fast-running pressure forecasting

measurement

forecast / precast

calibration period

400

pressure, bar

380

360

340
equivalent buildup test
150
100

320

50
0
300

100

200

300

400

500
600
time, days

700

500
800

900

1000
1000

1100

Retrospective analysis of a brine pre-production scenario

measurement

w/o production

w/ production

400
380
360

pressure, bar

340
320
300
280

brine
production
phase

equivalent buildup test

260

150
CO2
injection
phase

240
220
200

100
50
0

-300 -200 -100

500

1000

100 200 300 400 500 600 700 800 900 1000 1100
time, days

Summary
This project has explored two useful analysis techniques:

Probabilistic assessment of fault reactivation potential and in situ stress

sensitivity
dynamic well-test analyses using multi-rate gauge data

Directions for future work:

Full poromechanical simulations of fault reactivation and leakage.

Relaxing current assumptions in the welltest analysis methodology
to provide a more general tool.

Acknowledgements

This work was performed under the auspices of the U.S. Department of Energy by
Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344.
Data and co-funding were provided by the DOE Carbon Sequestration Program
and Statoil.

Contact

Laura Chiaramonte

Josh White

Baker Hughes

Lawrence Livermore National Laboratory

laura.chiaramonte@bakerhughes.com

jawhite@llnl.gov

Appendix

Org Chart
Fuel Cycle Innova7ons
(Roger Aines)

Carbon
Management
(Susan Carroll)

LLNL Carbon
Sequestra7on
Program
Task 1. Ac9ve
Reservoir
Management

Task 2. In Salah

Task 4. Snovit

Task 5.
Carbonates

Technical Sta

Boucier
Buscheck Aines

Ezzedine, Hao, White

Chiaramonte, White,
Hao, Wagoner, Walsh

Carroll, Hao,
Mason, Smith

Exper7se

Subsurface
Hydrology

Computa7onal
Geomechanics

Experimental and
Theore7cal
Geochemistry

Seismology

Structural
Geology

Gantt Chart
!"#$
%&'()'
%&'()*
$%&'()*+,
,-./012324
./)&0123%31)&%/43)/560708&595+&:
!
<:5=036+0)%36(>5%)095+&:/68
?&59&1236/13:095+&:/68
55555%.62-"#37895:";135:"71;62
55555<"06.-$5=2:.6/"37.85>5:6"-3;62

!"#
!"!";
!"!

%&'()+

!
!

All technical tasks have been completed.

In the final reporting stage:

Will submit final report and lessons-learned document by Sept 30,

2014.
Several journal manuscripts currently in review or preparation.

Bibliography
Chiaramonte et al. (2014). Probabilistic geomechanical analysis of
compartmentalization at the Snhvit CO2 storage project. Submitted to
JGRSolid Earth.
White et al. (2014). Generalized superposition welltest analysis. In
preparation.
Numerous conference papers and abstracts.