Вы находитесь на странице: 1из 25

Reservoir Monitoring Consortium (RMC)

Semi- Annual Project Review Meeting

Reservoir Monitoring with 4D Seismic: Three


Case Histories
Yesser HajNasser,

Los Angeles, CA
July 22, 2015

1
Background

Yesser HajNasser, PhD

Work Experience:

- ConocoPhillips GRE, Houston, 2011-2015


- Heriot Watt University ETLP, Edinburgh, 2008-2011

Education:

- PhD Petroleum Engineering\Reservoir geophysics, HWU


- MSc Reservoir Evaluation and Management, HWU
- MEng Geomechanics, Mechanics, and Structures , EPT
2
Outline
 Introduction

 Case studies:

 4D seismic and reservoir geomechanics

 Well2seis correlation for 4D interpretation

 4D seismic fracture detection

 Conclusions

3
Introduction
The potential benefits of time-lapse ( or 4D) seismic monitoring have been well recognized across
the industry worldwide.
4D at the Ekofisk field (North Sea)
Land 4D at the Alpine field (Alaska)

(Tura et al., 2013)

(HajNasser et al., 2014)

4D at a carbonate reservoir (Middle East)

4D at the Mars field (Deep GoM)

(Foster, 2008)

(Soroka et al.., 2005)

(Hatchell et al., 2005)


4
Objectives

 Integrate 4D seismic data and reservoir


geomechanics to assess potential Geohazards for
future field development.

 Correlate 4D seismic data with production data for


a better interpretation of Time Lapse Seismic.

 Extend conventional 4D seismic modeling to


account for different fracture media

5
Case Study 1: 4D seismic and geomechanics

Motivation
Observed 4D seismic response due to pressure depletion
(2001-1989):
4D Amplitude Time Shift

Softening from
4D Time Shift

?
Hardening from
4D Amplitude

6
What might be causing the softening?
2001 - 1989

Gas
The decrease in gas saturation does
NOT lead to positive time shift.

2001 - 1989

Pressure The decrease in Pore Pressure does


NOT lead to positive time shift.

Geomechanics ??
7
Geomechnics and Fluid simulation
Coupled geomechanical modeling
Simulation Model
Fluid Simulation model
Fluid Flow
Simulation
Eclipse
reservoir

Active Shale concept: Pressure & Saturation


permeable and Pp, Sw ε, σ, Φ, K
mechanically active
Geomechanical Model
Geomechanical model
Stress changes The changes in stresses
and strains are computed
& in VISAGE (a 3D finite
Pressure diffusion element based
 teff   teff1   t   Pt  Pt 1 
geomechanics simulation
software)
Geomechanic  sh t 
Simulation
Visage 112 145 111 grid cells
1.8 Million Cells 8
92000 93000 94000 95000 96000 92000 93000 94000 95000 96000
92000 93000 94000 95000 96000 92000 93000 94000 95000 96000

-4400
-4400

-5200

-5200
Heather formation is responding

-4400
-4400

-4400
-4400

to the pressure diffusion

-4400

-4400
-4400

-4400
Increase in
effective stress
93000 94000 95000 96000
92000 93000 94000
Heather 95000 96000  zzeff (bar)
General

-4600
-4600
0 250 500 750 1000 1250m
∆P(bar)

-4600
-4600

-4600
-4600

0 250 500 750 1000 1250m

-4600

-4600
-4600

-4600
General
formation Shale 0 1:32000 350
1:32000
Decrease in 300
Erskine -50 effective stress

-4800
-4800

-4800
-4800

Erskine shales 250

-4800

-4800
-4800

-4800
-100 200

-4800
-4800 -150 150

-5000
-5000

-5000
-5000

Pentland

-5000

-5000
-5000

-5000
100
No pressure -200
50
changes in the -250 Effective stress
0
Erskine shale changes 2001-

-5000
-5000

-5200
-5200

-5200
-5200

-5200

-5200
-5200

-5200
-300 -50
Pressure changes 2001-1989 1989
-350
92000 93000 94000 95000 96000 92000 93000 94000 95000 96000
92000 93000 94000 95000 96000 92000 93000 94000 95000 96000
0 250 500 750 1000 1250m 0 250 500 750 1000 1250m
0 250 500 750 1000 1250m General 0 250 500 750 1000 1250m General
General
92000 93000 94000 95000 96000 General
1:32000 920000 93000 94000 95000 96000 1:32000 350 0
1:32000 0 1:32000

-5200
-5200

-50 300
-50 -50
250
-100

-4400
-4400

-4400
-4400

-100 -100
200
-150
-150 150 -150

Hardening
92000
-200
-200
93000 94000 95000 96000
100 -200
50
-250
Millistrain

-4600
-4600

-250 -250

-4600
0 250 500 750 1000 1250m
-4600

-300
General 0
-300 -50 -300
1:32000
-350
-350 0.5 -350
Softening

-4800
-4800

-4800
-4800

0
Softening
-0.5

-5000
-5000

-5000
-5000

-1

-5200
-1.5
-5200

Strain 2001-
-5200
-5200

-2
1989 Hardening
92000
92000 93000
93000
0
94000
94000
250 500 750 1000 1250m
95000
95000 96000
96000 9
0 250 500 750 1000 1250m General
General
1:32000
1:32000 0
4D Time shifts

Observed Time Shift Predicted Time Shift

2001 - 1989

Comparison of time shifts at Erskine on stacked field data (left) and calculated using
the combined effect of geomechanics and pressure diffusion simulation (right). Both
field data and prediction show up to +5 ms time shifts with similar shape. 10
Case Study 1: conclusions

The integration of geomechanics and pressure diffusion into the 4D


seismic signal analysis suggest:
The trapping mechanism at the Heather Heather shale

shale is ambiguous
Possible Shear
Geohazards: failure
Possible well casing damage and stress changes
due to depletion at the Heather shale

The Erskine shales acts as an effective


pressure barrier

Geohazards:
Erskine Shale is overpressured relative to the
surrounding formations
11
Case Study 2: Well2seis correlation for 4D interpretation
Motivation
Time Shift
S2- S1 Softening

Obscured
zone S3- S2

Obscured
S4 – S3
zone
Hardening
Obscured
zone S5 – S4

Obscured
zone S6 – S5

Obscured
zone

Where is production coming from?


vs.
Where is the signal coming from?

12
4D Well2Seis
Well2seis: Methodology correlation: Methodology

Seismic attributes Well activity

1989 (S1) 1999 (S2) 2003 (S3)


S1 S2 S3 S4 S5 Well 1

V
V: Cumulative
2006 (S4) 2008 (S5) Volume

2006 2008
Time
Seismic surveys Cumulative volume Well 1
S1 S2 S3 S4 S5 V1 V2 V3 V4 V5
A
1 4D 1 Production

Normalized
production
0.8 seismic 0.8 data changes
Normalize

B
Seismic

0.6 0.6
d 4D

0.4 0.4

0.2 Area of 0.2

0 influence 0

89 - 99 89 - 03 89 - 06 89 - 08 99 - 03 99 - 06 99 - 08 03 - 06 03 - 08 06 - 08 89 - 99 89 - 03 89 - 06 89 - 08 99 - 03 99 - 06 99 - 08 03 - 06 03 - 08 06 - 08

4D Time Interval 4D Time


Interval

Production data
1 changes
0.8 4D seismic at A
Normalize 0.6
4D seismic at B
d data 0.4
0.2
0

89 - 99 89 - 03 89 - 06 89 - 08 99 - 03 99 - 06 99 - 08 03 - 06 03 - 08 06 - 08

4D Time 13
Interval
4D attribute analysis:
Time Shift ( Top reservoir -20 ms) Time Strain (Top –Base)
08 - 06 4D signal 08 - 06

Softening
(Slow down)

Hardening
(Speed up)

Volume

(X 100)

Water injection

Gas Oil
Oil
Water
Gas

Water
Production

(Bubble size proportional to


cumulative volume )

Slow down (time Shift) shows good Speed up (time strain) shows good
correlation with the producers correlation with the Injectors
Time shift attribute will be used to Time strain attribute will be used to
investigate the area of influence for investigate the area of influence for
the producers the injectors 14
4D well2seis correlation
Seismic Attributes Area of Influence

Correlation
technique

Injector
Oil
Water
Gas Producer
Time shift Producers

Correlation
technique Area Swept by
water injection

Area of influence of
producers

Potential Unswept Area ?

Well2Seis correlation technique could help to


identify the unswept areas.

Time strain Injectors


15
Case study 2: Conclusions

• Using production and multi vintages time lapse seismic we can


potentially identify the area of influence of active wells

• This can potentially help in :

– Constrain simulation model

– Identify unswept zones

16
Case Study 3: 4D seismic fracture detection
Motivation
RMS (Monitor – Base) (20ms window) over reservoir interval

Field
Data With
conventional
seismic
modeling we are
? unable to match
the large
amplitude
change near
Synthetic horizontal
Data injectors

Injector
Producer
17
4D Modeling with Fractures
fractured fluid saturated
Dry unstressed rock Dry Rock at Fluid saturated rock at
Rock at reservoir stress
reservoir stress reservoir stress

Convolve
with wavelet
1 2 3 4
Add
Apply stress Isotropic Gassmann Insert fluid filled
dependence Fluid substitution fracture noise
Dry initial pore
space Synthetic
Fluid filled pore seismic
space

Fluid filled fracture


pore space

We considered 3 different fractured media:


• Kuster-Toksoz (KT): Isotropic medium from randomly distributed fractures
• Eshelby-Cheng (EC) HTI: Anisotropic medium from single set of vertical fractures
• Eshelby-Cheng (EC) VTI: Anisotropic medium from single set of horizontal fractures

18
Which Fracture Model is Applicable?
Range of Calculated stresses

Sv > Shmin ~ Shmax


• We wont be able to open horizontal fractures (EC VTI not appropriate)
• KT more likely than EC HTI: i) Shmin ~ Shmax, ii) (Sv - Shmin) > 4T
19
Which Fracture Model is Applicable?

Model Selection Fracture Calibration

Half-space Modeling
- KT model shows large change at zero offset and then decreasing
- EC HTI shows small change at zero offset then polarity reversal
- EC VTI not reasonable with Alpine stress field (not plotted)

20
One Example of Pressure Update to Match 4D Spatially

Field
Data

We are able to
match the 4D via
fracturing and
pressure updates

Synthetic Fracture Aspect Ratio =


Data with 0.0055

Updated Fracture Porosity =


0.0055
Pore
Pressures

RMS (Monitor – Base) (20ms window) over reservoir interval


21
Case study 3: Conclusions

• We show a clear example of seismic amplitude changes from


fracturing induced by water and gas injection
• We extend conventional SFM to 3 different types of fractured
media
• We find that the most appropriate model for this case study is KT
• We use the KT model to calibrate fracture parameters to match
field data
• When fracturing effect is more subtle, it can help explain why we
get larger velocity drops at injectors

22
Learning from 4D:

Full-field model
prediction pressure,
saturation, Geomechanics

Reservoir
Simulation

4D
Reservoir
Log Analysis Data Quality
Monitoring
(Vp, Vs, Density) Rock Seismic S/N, Bandwidth
Properties Response

Pressure Response Repeatability


Core analysis Match Processing,
Analogue data Amplitude analysis

(O'Donovan, 2000)
23
Lessons learned and what we can do at USC

 4D seismic geomechanics can provide a more reliable reservoir model to assess


pore pressure and stress changes, allowing more confidence in selecting safe well
paths.

 What we can do at USC: Apply 4D geomechanical analysis on stress-sensitive


reservoirs provide by sponsors.

 4D Well2Seis Correlation technique can be used to distinguish between the


response of injectors and producers and identify potentially unswept area.

 What we can do at USC: Run the well2seis correlation technique on multiple


seismic surveys provided by sponsors.

 Strong 4D amplitude changes can be caused by the creation of fractures or opening


of pre-existing fractures associated with injection.

 What we can do at USC: Develop rock physics models to assess the sensitivity
of the 4D seismic to reservoir changes.
24
Thank you

25

Вам также может понравиться