1

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

© 2014 BAKER HUGHES INCORPORATED. ALL RIGHTS RESERVED. TERMS AND CONDITIONS OF USE: BY ACCEPTING THIS DOCUMENT, THE RECIPIENT AGREES THAT THE DOCUMENT TOGETHER WITH ALL INFORMATION INCLUDED THEREIN IS THE CONFIDENTIAL AND PROPRIETARY

PROPERTY OF BAKER HUGHES INCORPORATED AND INCLUDES VALUABLE TRADE SECRETS AND/OR PROPRIETARY INFORMATION OF BAKER HUGHES (COL LECTIVELY " INFORMATION"). BAKER HUGHES RETAINS ALL RIGHTS UNDER COPYRIGHT LAWS AND TRADE SECRET LAWS OF

THE UNITED STATES OF AMERICA AND OTHER COUNTRIES. THE RECIPIENT FURTHER AGREES THAT THE DOCUMENT MAY NOT BE DISTRIBUTED, TRANSM ITTED, COPIED OR REPRODUCED IN WHOLE OR IN PART BY ANY MEANS, ELECTRONIC, MECHANICAL, OR OTHERWISE, WITHOUT

THE EXPRESS PRIOR WRITTEN CONSENT OF BAKER HUGHES, AND MAY NOT BE USED DIRECTLY OR INDIRECTLY IN ANY WAY DETRIMENTAL TO BAKER HUGHES’ INTEREST.

Waterflood Evaluation using Communication Analysis Tool

Atul Kshirsagar

Reservoir Engineer

2

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Outline

Typical waterflood workflows

The dilemma

Communication analysis tool

Field case study

Conclusions

3

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Recap: Review Conventional WF Profile

Waterflooding provides approximately

double the primary reserves

Majority of oil production in the World

comes from WF

Key Mechanisms are:

– Pressure Maintenance

– Displacement / Sweep

Waterflood decline period accounts for

majority of waterflood life (volumetric

sweep efficiency becomes key factor

governing WF recovery)

Fillup Incline

P

E

A

K

Decline

4

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Logical Work Flow for Waterflood Analysis

Key Questions:

What is the OOIP?

How much oil is left?

Where is remaining oil?

How do we optimise recovery?

Is it economic?

5

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Routine Waterflood Surveillance

Field level analysis

Well pair analysis

Composite plots

Decline analysis

RF vs. HCPVI

Conformance plot

(Qo, Qfld, Wp) vs. Iw

Watercut / GOR / WOR analysis

Time lapse

Bubble maps (Qo, Qfld, Wp, Iw)

Hall Plots

Static reservoir pressure data

Communication Analysis

6

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

What Controls Waterflood Recovery?

Ed is the microscopic displacement efficiency

~ 60 to 65%

Ev is the volumetric efficiency

– Volumetric sweep ranges between 10 to 80% for

most floods

– Low Volumetric Sweep Efficiency (Ev) is the main

reason Waterfloods fail

– Low sweep efficiency usually means high water

cycling and high unswept oil volumes

Communication Analysis will give indication of

pattern Ev by examining cause and effectHigh Ev

Target for Waterflood

optimization

Low Ev

7

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

The Dilemma

It’s a plumbing problem

Combination of RF vs. HCPVI, analogy,

decline analysis indicates field or area

upside but…

Where do we:

– place wells?

– perform workovers?

– convert producers to injectors?

8

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Communication Analysis (SweepSCANTM)

We want to identify key wells that communicate

12

3

Map view

Use all available DYNAMIC data (prod/inj/pressure) toidentify high permeability, channelling, baffles, barriers

9

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

12

3

Communication Analysis (SweepSCANTM )

Objective

– fluid flow between well pairs

– magnitude of communication

– water cycling

Different variables used for correlation

– Water Injected

– Liquid produced

– GOR

– Gas Injected

– Gas Production

– User defined variable

Communication Strength

10

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Manyberries Sunburst Q• OOIP = 25 MMSTB• RFcurrent = 41.5% OOIP• 81 Wells

Manyberries Sunburst JJ•OOIP = 15 MMSTB•RFcurrent = 33% OOIP•37 Wells

Manyberries Sunburst O•OOIP = 14 MMSTB•RFcurrent = 17% OOIP•33 Wells

Field Example – Manyberries Sunburst Field (Canada)

11

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Water Injected (HCPV)

good performance

average performance(unit performance)

poor performance

(low)

favorable geology

low Sorw,

excellent sweep Ev

high Sorw,water channeling,

poor areal/ vert. sweep

unfavorable geology

Cum

ula

tive R

e c

ove

ry (

% O

OIP

)

early break over point

(< 20% recovery)

Recovery Factor vs. HCPVI Plot

Slope of the curve

indicates waterflood

efficiency

Compare RF vs.

HCPVI plots

between fields with

consistent geology

and compare

performance to avg.

trend of field

Group individual

patterns into high,

medium, poor

performers (compare

to avg. trend of field)

12

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Field Example – RF vs. HCPVI

Sunburst Q and JJ fields track each

other and are expected to recover

+40% OOIP

Sunburst O field breaks over earlier

and appears will only recover ~25%

OOIP

– Has poor waterflood efficiency in

comparison to analogues

– Channelling or cycling of water

to one or more producers?

Communication Analysis will

pinpoint which wells are problem

wells.

13

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

SweepSCAN Analysis Manyberries Sunburst Q

Shows relatively good communication in most injection patterns

14

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

SweepSCAN Analysis of Sunburst O

Less communication between injectors and producers

Liquid production not correlate with injection changes at most wells

Shows good communication

15

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Analyse Pressure Trends and Couple with Prod/Inj Response

Sunburst O Pressure Data

16

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Field Example – Sunburst O

We see areas with high pressure and areas with low pressure

Looking at water injection and water production bubble map, it is evident that a lot of the water

injected is being produced by 1 producer - cycling

No communication between injector and producer (low pressure and no water production)

Strong communication between injectors and producers (high pressure and high water production))

17

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Case Study – Weyburn Midale – WF Conformance Optimization

Weyburn Midale beds discovered in 1955

Delineation drilling defined field limits 1956 - 1957

Unitized for waterflood 1962 - 1963

IOR techniques implemented 1985 -1992. Including infill drilling, horizontal drilling and injection modifications (line drive)

Miscible Flood starts in 2000

Naturally Fractured Reservoir

18

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Good communication with all producers

Possible Water shut off candidate (injector has strong communication with one producer )

Using SweepSCAN Analysis to determine poor areas of WF

Use communication analysis software to select water shut off candidates

Use injection and production data to determine areas of low waterflood volumetric

sweep efficiency

Quickly pinpoint areas that have upside potential!

19

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

A Closer Look at Selected Pattern

Well produces at very high watercut

Water injection rate tracks production wells water

production rate

High permeability pathway between injector and

producer resulting in poor sweep to other

producers

Pattern would benefit from water shut off treatment

- directing waterflood support to nearby producers

GOR

20

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Weyburn Water Shut Off Treatment

Gel Treatment Dates/Volumes:

Injector A

November 26-30, 2008: 7250kg

polymer, 497m3 gel solution

Injector B

January 12-22, 2011: 6225kg

polymer, 661m3 gel solution

High Pressure

LowPressure

Inj A

Inj B

Waterflood Diagnostics indicates a volumetric sweep efficiency of around 53% in 2008. Evident that SE region is unswept

D

C E

D

21

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

2008: Injector A- 3 hours to breakthrough to

producer C- 16 hours to E

2011: Injector B< 9 hours to breakthroughTo producer D

Flow of Polymer/GelProducers that show increase in Pressures and Oil Rates after Treatment

A

B

D

CE

Location of Wells with Polymer Breakthrough

22

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

0

2

4

6

8

10

12

14

16

18

0

50

100

150

200

250

300

350

400

0 200000 400000 600000 800000 1000000 1200000 1400000 1600000

Oil

Rat

e (

m3

/day

)

Cumulative Oil (m3)

OIL RATE (P) 2008 Injection 2011 Injection Producing Well Count

Prior to Gel injection After Gel Injection Incremental

Est. Oil Recovery (thousand Sm3)

840 1,180 340

Est. Oil Recovery (MMSTB)

5.3 7.4 2.1

Oil Rate vs. Cumulative Oil Produced

23

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

North Sea Experience

3 completed projects in UKCS for majors

Ideal technique for mature UKCS fields

Complex reservoirs, e.g. multi-layered, faulted

Supplements static geological model

Has indicated poorly swept areas for infill well opportunities

Communication analysis results supported by tracer test results

24

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Conclusions

SweepSCANTM is a powerful tool and workflow that provides valuable

insight into the unknown plumbing of a reservoir

Reduces risk in reservoir characterisation at start of evaluation

Simple map view facilitates easy communication between

disciplines/teams to improve decision cycle time

Can be used successfully in multi-layered reservoirs

Fast - Identifies optimisation target locations in days not months

Thank You

25

© 2

014

Bak

er H

ughe

s In

corp

orat

ed. A

ll R

ight

s R

eser

ved.

Questions