The Value of Well Testing and Pressure Monitoring in Injection Site Characterization

A. Shchipanov, L. Kollbotn, R. BerenblyumInternational Research Institute of Stavanger (IRIS)Contact: Anton.Shchipanov@iris.no

CO2GeoNet Open Forum, 24-25 April 2018, San Servolo Island, VenicePre-Open Forum workshop organised by ENOS, 23 April 2018:

Experience-sharing focus groups: Advanced techniques for site characterisation

Outline

› Introduction

› What is well testing and Pressure Transient Analysis (PTA)?

› Testing wells to characterize injection site

› Well test designs

› Radius of investigation

› Interpretation of injection and fall-off responses

› An example of pressure monitoring helping in characterizing site boundaries

› CO2 injection at Tubåen formation of Snøhvit field

› Time-lapse injection and fall-off pressure responses

› Lessons learned from Snøhvit

› Acknowledgements

What is well testing and Pressure Transient Analysis (PTA)?

10

100

1000

10000

0.001 0.01 0.1 1 10 100 1000

Pre

ssu

re a

nd

Der

ivat

ive

[psi

]

Elapsed Time [hr]

Pressure Derivative

Build-up Build-up

2600

3100

3600

4100

4600

Pre

ssu

re [p

si]

Gauge

0

1000

2000

3000

4000

5000

6000

0 100 200 300 400 500

Rat

e [S

TB/D

]

Time [hr]

Production

2600

3100

3600

4100

4600

Pre

ssu

re [p

si]

Gauge

Model

0

1000

2000

3000

4000

5000

6000

0 100 200 300 400 500

Rat

e [S

TB/D

]

Time [hr]

Production

Pressure Transient Analysis (PTA)

Well• Storage• Skin factor• Induced fracture(s) properties• Effective length (horizontal)

Reservoir• Flow capacity / permeability

(directional for horizontal well)• Areal heterogeneity• Dual porosity / permeability

Boundaries• Faults• Aquifer• Well interference

Parameters estimated from PTA

10

100

1000

10000

0.001 0.01 0.1 1 10 100 1000

Pre

ssu

re a

nd

Der

ivat

ive

[psi

]

Elapsed Time [hr]

Pressure Derivative

Build-up Build-up

Model Model

Typical well test (an example from Kappa Eng.):Production and following pressure build-up

PTA is history matching of well test including pressure derivative

Pressure build-up and its derivative in log-log

Effective well length

Reservoir flow capacity

Wellbore storage

Well shut-in = pressure build-up

Testing wells to characterize injection site

A synthetic model of injection site (a fault block)

An aquifer fault-block (U-shape) confined by three faults (no-flow boundaries): West, North and South from the well

Pressure[bar]

N

Sealing fault =

no-flow boundary

Open for flow boundaryWell

280

330

380

430

Pre

ssu

re [

bar

]

1

10

100

-1500

-1000

-500

0

0 400 800 1200 1600 2000 2400 2800 3200 3600 4000 4400 4800

Rat

e [S

TM3

/D]

Time [hr]

1

10

100

Well test designs: from best to more feasible

› ‘100’ (day): Long injection

› 0.4% PV (of the model) injected, may be followed by fall-off, but not necessary

› All boundaries* covered during injection

› ’10’ (day): Shorter injection – Long fall-off

› 0.04% PV injected with long fall-off (100 day)

› All boundaries may be captured during injection, fall-off serves for confirmation

› ‘1’ (day): Short injection – Long fall-off

› 0.004% PV injected with long fall-off (100 day)

› Very small volume injected short-term

› No boundaries detected from injection, all boundaries covered by fall-off

*All boundaries = 3 boundaries in this U-shape example

Shorter injection

290

300

310

320

330

340

350

Pre

ssu

re [

bar

]

-1500

-1000

-500

0

0 400 800 1200 1600 2000 2400

Rat

e [S

TM3

/D]

Time [hr]

Pressure

Derivative

1

10

100

0.1 1 10 100 1000 10000

Pre

ssu

re a

nd

Der

ivat

ive

[bar

]

Elapsed Time [hr]

1, 2, 4, 8…

Design ‘1’: Radius of investigation and flow regimes

1st day:end of

injection

2nd day

4th day

8th day: and similar

picture afterwards

Long pressure fall-off (100 day)

Short-terminjection (1 day)

Fall-off response: Pressure and derivativePressure evolution in U-shape reservoir

during well test. Pressure exceeding 302 bar is colored in red to highlight pressure

evolution deeper into reservoir

Interpretation of injection and fall-off responses

Pressure

Derivative

1

10

100

1000

0.1 1 10 100 1000 10000

Pre

ssu

re a

nd

Der

ivat

ive

[bar

]

Elapsed Time [hr]

1 (injection) 1 (fall-off)10 (injection) 10 (fall-off)100 (injection) 100 (fall-off)Boundary 1, 2, 4, 8 day

Injection

› Boundary dominated flow regime started after ~40 hr

› Reliable boundary indication after additional ~40 hr

Fall-off

› Beginning of boundary dominated regime depends on injection duration prior to fall-off

› 100 day: after ~80 hr

› 10: ~120 hr

› 1: ~170 hr

› Reliable boundary indication after additional ~40 hr

› Flow barriers are most quickly captured from injection pressure interpretation(quick boundary indication = minimum test duration)

› If volume to be injected is limited, fall-off is an alternative (but longer test)

› Permanent Downhole Gauges (measuring P&T) must be installed

› A well test before the main injection phase may consist of a short water production or injection followed by long shut-in (to disclose distant flow barriers)

› Such a test gives information about critical geological features › Results may work as showstopper for the project!

› Using CO2 for injection test leads to larger uncertainty with interpretation

› PTA of pressure dynamics during the main injection phase improves site characterization

Testing wells to characterize injection site

An example of pressure monitoring helping in characterizing site boundaries

CO2 injection at Tubåen formation of Snøhvit field

Difference amplitude map between baseline and monitor at base of the reservoir [from Hansen et al., 2011]

Depth map of top Fuglen formation (left) and geological cross-section N-S through the reservoir sections at Snøhvit (right) [from Hansen et al., 2013]

History of CO2 injection at Tubåen formation

280

300

320

340

360

380

400

Pre

ssu

re [

bar

]

-3000

-2500

-2000

-1500

-1000

-500

0

0 2500 5000 7500 10000 12500 15000 17500 20000 22500 25000

Rat

e [S

TM3

/D]

Time [hr]

Salt precipitation followed by treatment

History periodin focus

280

300

320

340

360

380

400

Pre

ssu

re [

bar

]

Gauge

Model (Fall-offs)

-3000

-2500

-2000

-1500

-1000

-500

0

0 2500 5000 7500 10000 12500 15000 17500 20000 22500 25000

Rat

e [S

TM3

/D]

Time [hr]

Time-lapse fall-off pressure transients

0.1

1

10

100

10 100 1000 10000

Pre

ssu

re a

nd

Der

ivat

ive

[bar

]

Elapsed Time [hr]

Pressure DerivativeFall-off 1 Fall-off 1Fall-off 2 Fall-off 2Fall-off 3 Fall-off 3Fall-off 4 Fall-off 4

Referencefall-off

Fall-off responses / derivatives

› All faults may be captured by the shortest fall-off (10 days fall-off is already enough)

› Follow exactly the same trend indicating no changes in sealing capacity of the faults

FO 1 FO 2 FO 3 FO 4

2-year history of CO2 injection at Snøhvitmay be reproduced with analytical models matched by PTA

Four pressure fall-off responses caused by well shut-ins

All fall-offs indicate U-shape reservoir

280

300

320

340

360

380

400

Pre

ssu

re [

bar

]

Gauge

Model (Injection 1)

Model (Injection 3)

-3000

-2500

-2000

-1500

-1000

-500

0

0 2500 5000 7500 10000 12500 15000 17500 20000 22500 25000

Rat

e [S

TM3

/D]

Time [hr]

0.1

1

10

100

10 100 1000 10000

Pre

ssu

re a

nd

Der

ivat

ive

[bar

]

Elapsed Time [hr]

Pressure DerivativeFall-off 1 Fall-off 1Injection 1 Injection 1Injection 2 Injection 2Injection 3 Injection 3

Time-lapse injection pressure transients

Referenceinjection

INJ 1 INJ 2 INJ 3

Injection responses / derivatives

› Rate fluctuations cause noisy pressure transients

› Spread of transients is observed even after smoothing, but they approach the same trend as fall-offs

The overall pressure build-up may be predicted with ~10% error based on the models matched to noisy injection data

Three injection pressure responses and longest fall-off response

› PTA of real-time pressure measurements (PDG) or continuous ‘well test’ is the most efficient tool to characterize and monitor site boundary conditions

› PTA of PDG data (analytical models)

› Uses gauge data available for the most of modern wells

› Fast and cheap: minimum time, computations and data input

› Efficient in characterizing large geological domains: pressure diffuses mainly in low-compressible saline aquifer, even after CO2 plume appearance

› Site monitoring

› The above advantages makes time-lapse PTA of PDG data a perfect tool for site monitoring, e.g. CO2 reservoir containment

Lessons learned from Snøhvit

Acknowledgements

A part of this presentation was prepared and presented at 2015 AnnualEAGE conference (paper published at EarthDoc*)The authors acknowledge Statoil Petroleum AS and the Snøhvit licensepartners for providing access to the Snøhvit data presented at the EAGEconference

*Shchipanov A., Kollbotn L., Berenblyum R. Pressure Transient Analysis in CO2 StorageProjects - From Reservoir Characterization to Injection Performance Forecast // 77th EAGEConference & Exhibition. IFEMA Madrid, Spain, 1-4 June 2015

Thank you for your attention!

Contact: Anton.Shchipanov@iris.no

Derivative

Pressure

1

10

100

1000

0.1 1 10 100 1000 10000

Pre

ssu

re a

nd

Der

ivat

ive

[bar

]

Elapsed Time [hr]

Reference

Total Compressibility (x2)

Closer Barrier (West 1/2)

Sensitivity to flow barriers and compressibility

Sensitivity to distance to West fault and total compressibility (U-Shape case)

1

10

100

1000

1 10 100 1000 10000

Pre

ssu

re a

nd

Der

ivat

ive

[bar

]

Elapsed Time [hr]

Pressure Derivative1: Single 1: Single2: Parallel 2: Parallel3: U-Shape 3: U-Shape4: Chamber 4: Chamber

Sensitivity of fall-off response to number of sealing faults (no-flow boundaries)

1: Single 2: Parallel 3: U-Shape 4: Chamber