FORTIES FIELD

PROJECT

Time Lapse (4D) seismic Interpretation with

application to deep/ultra deep water reservoirs
 Study Focus:

Repeatability dilemma…
- How far can we push the limit when repeatability is poor?

Integrity monitoring of overburden shale

Is the reservoir base immuned from 4D effect?

Identify potentially damaged zones

Is there a seismic approach to predicting sand production problem?

Geomechanical characterization: Nano scale approach

 Field of choice :Forties Field

Field Location:

Forties field is located approximately 140 Km NE of Aberdeen.

It‟s one of the earlier mega-offshore fields in the UK North Sea.

Water column is about 100 – 120 m

 Forties Field : Data Description
Forties Field : Just a few comments -1988 Survey

 First 3D survey

 Acquired after over 1.8b barrels of oil

had been produced

 A relatively larger coverage

2
3  Semi- high density (25 x 12.5m
4 1 5
 Different geometry

-2000 Survey

 Repeat survey

 Acquired 600m barrels post 1988

 Acquired & Processed as base seismic

Legend
 Just within the AOI
-----Licenses
 High density (12.5 x 12.5m)
-----Platforms
 Same geometry as 2005
-----Field Outline
 Forties Field: Well selection

“Few have been chosen”

 23 Explo wells
Charlie Beta

 20 Alpha
Delta
Alpha

 18 Bravo
Explo. wells
Echo
 22 Charlie

 20 Delta

Explo. wells
 15 Echo

118 / 400 wells

Basic Field Geology & Stratigraphy
 Forties Field: Stratigraphy -1
Source: Apache Internal report
N

Northern
Peripheral
Area
RMS amplitude on the Near stack
Delta
Channel
Complex
Charlie
Complex
Southern
Western Interchannel
Peripheral Area
Echo
Area Bravo Channel
Channel Complex
Complex
Southeastern
Interchannel
Area
- Paleocene sands

- Reservoir sands are friable

- striking resemblance to deep/ultra deep water reservoirs

-Interconnectivity: no absolute answer !!!!

-Charlie - In a different pressure compartment SELB – SELB-20ms

 Forties Field: Stratigraphy -2

Section on Litho-cube

Sele

Forties shale
Charlie Alpha
Bravo Delta

Delta

Upper Main

Litho-cube –:

Red: Sand

Blue: Shale Source: Apache Internal report

 Forties Field: Summary of Production & Injectivity
Production & Injection by survey time lines
SEGMENT 1988 ACQ 2000 ACQ 2005 ACQ

Oil Produced 490,683,047.40 127,308,541.20 28,953,855.30 1988 ACQ

ALPHA Water Produced 45,101,382.00 225,670,823.70 119,420,608.20
Total Oil produced = 1,830,327,329.70
Water Injected 383,825,695.20 326,178,318.90 55,607,609.40
Water Produced= 160,029,717.90
Oil Produced 457,676,638.80 93,098,626.20 15,388,710.00

BRAVO Water Produced 35,088,942.90 176,735,791.50 85,533,168.60

Water Injected = 1,553,386,715.70
Water Injected 481,719,895.20 449,294,153.40 78,178,213.80

Oil Produced 404,216,859.30 94,772,396.70 11,681,200.20

CHARLIE Water Produced 37,528,320.00 162,920,734.80 57,287,447.70

Water Injected 382,502,860.50 277,222,670.70 41,186,439.00 2000 ACQ

Total Oil produced = 579,840,241.20
Oil Produced 453,498,722.10 164,962,788.60 28,120,906.20

DELTA Water Produced 41,001,000.60 261,652,708.80 122,913,411.90 Water Produced= 935,307,604.50

Water Injected 305,338,264.80 253,584,267.90 60,394,852.20

Water Injected = 1,306,279,410.90

Oil Produced 24,252,062.10 99,697,888.50 8,787,609.30

ECHO Water Produced 1,310,072.40 108,327,545.70 20,565,245.40

Water Injected - - -

Cumulative Production & Injection

2005 ACQ
Total Oil produced = 2,503,099,851.90 Total Oil produced = 92,932,281.00

Water Produced= 1,501,057,204.20 Water Produced= 405,719,881.80

Water Injected = 3,225,236,104.50 Water Injected = 235,367,114.40

 Reservoir Pressure and Production history
Group: Temporary Items In Group: 218 Format: [p] Simple well viewer

Production History
Psi 560000

Source: Apache Internal report

504000

448000

392000

336000

280000

224000

168000

112000

56000

Years
0
75 80 85 90 95 00

Psi
Oil rate (merged Finder and TOW),FPS bbl VS Time

 Reservoir bubble point : 1150 Psi

 Good pressure maintenance overall

 Is reservoir pressure > bubble point sufficient ?

 Is the slight decrease of any significance?

Years
 1988 versus 2000 /2005

Differences: 1988 cropped

- Geometry

- Acquisition Parameters

- Processing (Binning, algorithm etc)

- Velocity field

Consequences:
Bin size :

Travel time differences

1988; 12.5 X 25
Amplitude /Frequency spectrum ?
2000/2005; 12.5 X 12.5
Phase ?
 Data Balancing: Survey Projection
Before

Grid Projection & Trace Interpolation

1988
Challenges:

Impact of linear projection / interpolation?

Can we implement a dip consistent projection?

After

d
 Data Balancing: Survey Re-gridding

Inverse relation: ??

-Is that due to production?

- Could it be phase reversal ?

or
2000 Amplitude Vrs.
Amplitude
Something else,

Time shift??

 If time shift, what is the value?

 Is it vertically consistent?

1988
 1988 Vs 2000: Background time residue

Estimated time difference : 13.5ms

Is the time shift constant everywhere ?
Is the time shift everywhere constant ?

A A

B C
 1988 vs 2000 seismic surveys
Volume time-shift: 13.5ms Frequency spectrum
relatively wide scattering, large aspect ratio

relatively wide scattering

Aspect ratio; 1:3.88

Frequency distribution: comparable

Can we balance the 2 spectral?

Amplitude transformation:
Y = 0.0275386x
 1988 Vs 2000 seismic surveys: Amplitude spectrum

Amplitude spectrum: 1988 Amplitude spectrum: 2000

Normalized Normalized

Difference could be explained by:

 Acquisition and processing grid

 Fluid substitution

 Improvement in processing algorithm btw 1988 and 2000

Range: Range:
-5000 : 5000 -200 : 200
Absolute
Absolute

Spectral balancing needed

 Amplitude & Frequency spectral
Trace by Trace scaling or Shaping filter?

Trace scaling:

Amplitude transformation:
Frequency Spectral of „88 & „00 seismic surveys Y = 0.0275386x

Can we balance the amplitude and frequency spectral of -Easy to implement

the two surveys without signal distortion -Residual errors are amplified further

-Spectral (freq & amplitude) remain unchanged

- Distortion free
 1988 Vs 2000 seismic spectral Balancing: Shaping filter
Synthetic example
Shaping filter / Matched filter
Before reshaping
- Provides the best means of comparing waveforms

- Compares 2 signals and outputs a function describing the places at

which the 2 signals are most like one another

- Adjusts a waveform to a predefined spectrum

- Improvement in resolution

-Two modes:

- Trace by Trace matching

- Better waveform matching

- Distorts production effect After reshaping

-Global filter design

-Implements a single filter field-wide

-Extensive QC needed
- Over /under balancing
- Filter instability
- Masking of 4D signal
 Amplitude & Frequency Matching: Reshaping filter
Frequency spectrum before balancing
Final shaping filter

Amplitude spectrum after balancing

Frequency spectrum after balancing

 Amplitude & Frequency Matching: Reshaping filter

Before reshaping After reshaping

 Depletion and its field signature
- Concepts/Terms
Reservoir pressure depletion:
 Pore pressure
Pe = Pc - Pp
 Lithostatic pressure
 - Biot constant: 0< <1 Consequences:
 Confining pressure
where =1,
Reservoir : Compaction
 Effective pressure
Pe =Pd = Pc - Pp Overburden : Dilation/collapse

Time strain

Seismic Velocity
Fluid Substitution
 Gassmann‟s law and modified versions

Principal element: positive density change: oil water

Effective Pressure

Modified After: Colin Sayers, TLE December, 2006

- Amplitude Brightening
Rock Physics
- Dimming
 Time shift (t) and Time strain ()

- t : well established concept in 4D seismic Synthetic illustration

- time strain (); introduced in 2007

 = (t / t)
t

Why t &  ?:

: Subsidence
Max. correlation

: Compaction

: Computation of uniaxial strain changes

: Stretching

Cross correlation

Computation of Time shift (t):

Rickett‟s approach
 Time shift (t) and Time strain ()

Rickett‟s approach (Rickett et al. 2007)

- Non linear inversion of time shift

- Use of objective function

E = |d-f(m)|2 + |xm|2 + |ym|2 + |t2m|2 t

Vector d contains the data volume,

m contains the time shift volume

F(m) applies the time shift to the 2nd volume

x, y are first derivative operators in the inline and crossline

direction

Cross correlation of Traces: (continuous functions)

C(t) = (f(t)g(t) =

- f(t) & g(t) = 1988 and 2000 traces respectively

- t() @ max C(t) is the time difference related to 4D effect

 : computation of time strain

-Reservoir Strain computation ;

 Fairly easy to implement

 = = (1+R)*  zz
Hatchell and Bourne, 2005
 Highly subjective

Where  Is the time strain &  is the vertical strain

What about R? :

- It‟s dimensionless

- Represents the ratio between traveltime changes due to

velocity changes and path length changes

- A highly controversial figure (4-30 is a fairly wide range)

 σzz  v2  Objective

+( + )(σlith - p)

Hawkins et al., 2007

 Plenty of room for errors

Where σlith = litho pressure,  = Biot „s constant, and P = pore pressure

“Hard” part
of a cake
 T: at the overburden shale
Random_3

Random3
SELT

SELT

-AOI: High negative T

-Correlate with production zones

-Indication of field geology

-Indication of heterogeneity
 T: at the overburden shale: SELE
SELB +4ms

SELT

SELB +8ms

overburden shale
Does the reservoir BASE actually subside??
 Reservoir Base: Not speared also

t

Model Base

Base
Time strain (): Top of regional shale

Is a time strain of about .20%

sufficient to compromise the seal?
Time strain (): Base of regional shale

Has the top shale been

compromised?
Time strain (): Reservoir Top
Time strain (): Reservoir Base

Base
 Road ahead: 1988 versus 2000

Transform time strain () to uniaxial strain

Extensive QC of balanced cubes on-going – amplitude distortion

Implement wavelet transform & SOM to confirm result 4D effect

Fluid impedance inversion & calibration – CI and 4D inversion

 Road ahead: Geomechanical characterization: Nano scale approach

Deep/ultra deep reservoirs;

characterized by friable formations

Conventional core plug analysis not appropriate

Application of nanoindenter:

purpose built for friable reservoirs

Rapid implementation

Forties field Application:

unconsolidated reservoir

conventional characterization –failed

Drexel Nanotechnology Institute

(www.nano.materials.drexel.edu)
 Road ahead: Reservoir sand Production
Pre-drill Sand production prediction:
- back to the basics

Cost of sand production

- Forties field case

1997: Loss of 8.3 mbopd hole

1995: Loss of 7.4 mbopd hole

1994: Loss of 2.1 mbopd hole

1993: Loss of 10 mbopd hole

1992: $6.64m in remedial activities

Modified from Paul Weimer & Roger Slatt, 2006 Source: Apache Internal report

Incidences / severity of sand production

 Reservoir sand Production: back to the basics

Current prediction effort is post drilling: rock ultimate failure point

Pre-drill prediction not catching up with the magnitude of the problem

Back to the basics:

 Can we unwrap quality factor (Q) & formation transit time (Δt) as an
indicator ?
 Conclusion
Key elements:

Compaction and stretching exist on Forties field –

- Though at a relatively small scale

Compaction and stretching could compromise overburden sediments – weak overburden responds faster

Contrary to conventional assumption, reservoir base does subside as well

Transforming tau () to uniaxial strain is still not straight forward

Other methodologies needed to confirm cross correlation results : wavelet transform (SOM)

4D effect could still be done but will require extensive QC and calibration

Geomechanical characterization will have to involve the use of a nanoindenter

Thanks !!!