You are on page 1of 32

What’s so “special” about..

Special Core Analysis


Challenges, Pitfalls and Solutions

Colin McPhee
SPE London May 26 2015
The geomodel juggernaut!

• Modelling is ‘finished’, but the forecasts do not match observations,


imagine the reaction to a request to go back & check core data inputs.
• Often happens & each time the team’s protestations are loud.
• Very hard to stop the ‘geomodel juggernaut’, usually built on a tight
budget that is almost spent & to a deadline that is getting closer
2
Cultural resistance to change – “I know my place”

• Cultural issues can prevent the


models from being improved.

• Reluctance to change model


inputs as may have to admit
mistakes were made to peers.

• Misplaced respect for elders.

• Fear of management’s response


when told of model rebuild

3
Core data for static and dynamic models

• Core tests provide fundamental input to static (in place) and


dynamic (recovery factor) reservoir models
N 1
STOIIP  GRV     1  Sw  N, , Sw from RCA & SCAL
G B0

1
fw 
k ro  w
1 . kro and krw from SCAL
k rw  o

• Core data experiments are….

•The ground truth!

4
The elephant in the room

• SCAL data have uncertainties


that few end users want to
discuss or contemplate (or even
want to know about)
• Misinterpretation and poor
practice impact on static and
dynamic modelling

5
The Ground may be shakier than you think

• Based on review of > 50,000


SCAL experiments……
• 70% of SCAL unfit for purpose
• core damage
• variable data quality
• inadequate program planning and
inappropriate design
• poor reporting standards
• method-sensitivity
• vendors reluctant to share
experience and expertise

6
Core damage

• During coring
• Oil-based mud usually alters
wettability
• Difficult to remove sometimes
• Mud invasion and shear failure in
weak rock
• During core recovery
• POOH too fast results in tensile
fracturing if pore pressure cannot
dissipate
• During wellsite/lab handling
• Liners flexing/bending
• Freezing
• Poor stabilisation
• Poor preservation

7
Formation evaluation – examples of SCAL

• Porosity Porosity Permeability


• Permeability
• Capillary Pressure
• Drainage and imbibition
• Relative Permeability

8
Porosity

• Core porosity - Total or Effective?


• Humidity dry for effective porosity?
T > HOD > E

Absolute or Total Porosity Øt

Matrix
Effective Porosity Øe
VClay

Clay Clay surfaces & Small


Grains Large Pores Isolated Pores
Layers Interlayers Pores

Bound Water Capillary Volume


Water available for
Structural Water storage

Irreducible or
Immobile Water

Usually assumed negligible


in Clastics
Often assumed negligible May be significant in
in Carbonates Carbonates
9
Often significant in Clastics
Porosity (RCA)

• Two different methods

Vg & VbHg
Vb  Vg
 
Vb
Vp
 
Vg  Vp

• Two different results!


Vp & Vg

Vp+Vg

10
Vg+VbHg
Porosity compaction at stress

• Sensitive to “insignificant” artefacts


• Two labs – two different results!
• Annulus volume between sleeve & plug
• Check pre- and post-test results

stress/amb
Porosity Change Net confining stress (psi)
1.00

0.80

0.60

0.40
Porosity Change Post-Test (p.u.)

+ 0.25 p.u.
(p.u.)

0.20

0.00
Porosity Change

-0.20

-0.40 - 0.25 p.u.

-0.60

-0.80

-1.00
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Pre-Test Porosity (%)
11
Pre-test porosity (%)
Permeability

• What is the permeability in your static 3D model? Gas vs. Klinkenberg (measured) permeability (20- 30

• Air permeability? 10000


bar NCP)

• Klinkenberg? – measured or from a correlation?


1000

• Brine?
• Ambient or stressed?

Kl (mD)
100

• What stress?
10

• How measured – steady or unsteady-state?


1.020
y = 0.851x
2
R = 1.000

• How were plugs prepared? 1


1 10 100 1000 10000

• Does it matter? Kg (mD)


Kair after harsh drying (mD)

Kg @ Swir @ Stress (mD)

12
Kair after HOD (mD) Kair at 400 psi (mD)
Capillary pressure (drainage)

• Principal application in saturation-height modelling


• Pc (Height) versus Sw by rock type, rock quality and height

Height above FWL (ft)


Water Saturation (-)
J Function

Carbonate J function by R35 bin

Normalised Sw 13
Capillary pressure (drainage)

• Mercury injection capillary pressure


• NOT a capillary pressure test (just looks like one)
• No Swir: Sw goes to zero at high injection pressure
• Lower Sw at high Pc
• Core damage at high injection pressures?

200
67B K=4563.84 mD phi=0.4 RQI=3.572

175
67B K=4563.84 phi=0.36 RQI=3.58

150
251A K=582.62 mD phi=0.4 RQI=1.373
Air -BrineLabCapillaryPressure(psi)

125
251A K=582.61 phi=0.31 RQI=1.38

100
264B K=2441.96 mD phi=0.3 RQI=3.084

75
264B K=2441.96 phi=0.26 RQI=3.09

50
411B K=183.06 mD phi=0.3 RQI=0.791

25
411B K=183.06 phi=0.29 RQI=0.8

0
0.0 0.2 0.4 0.6 0.8 1.0

Sw (frac)

14
Capillary pressure (drainage)

• Centrifuge
• Pc maximum at inlet face of plug

Pci ~ 1.6 x10 7   w   h   re2  ri 2  RPM 2


• Calculation of inlet face saturation
 
Si  d d ( Pc).S .Pci  S  Pci .
dS
d ( Pci ) Sample No. 136S Porosity (%): 21.2
Depth (m): 2825.760 Gas Perm (mD): 52.4

130

120

110 Hassler Brunner


100 Average
Dean-Stark Sw
90

80

Capillary Pressure (psi)


Inlet face Pc (psi)

70

60

50

40

30

20

10

0
0 10 20 30 40 50 60 70 80 90 100
Water Saturation Brine Saturation(%)
15
Capillary pressure (drainage)

• Centrifuge vs MICP vs porous plate (PP)


• MICP
• no wetting phase – no Swir – Sw always lower at higher Pc
• Centrifuge
• No entry pressure (compared to MICP & PP) - Abrupt transition to Swir
Scaled Lab Pc (psi) vs. Sw (frac.)
Color by
120 Set, Comment
Set 2: O-B Pc Cent. » Lab. Rep

110 MICP Set 2: O-B Pc Cent. » Senergy Calc.


Set 3: FF (@stress)/O-B Pc-RI PP » Lab. Rep
Set 6: MICP » Lab. Rep

All values
100

90

80

70

PP Pc
Pc (psi)

Centrifuge
Scaled Lab Pc (psi)

60

50
Scaled Lab

40

30

20

10

0
16
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Sw (frac.)
Water Saturation
Capillary pressure (drainage)

• Porous plate 1.00

• Good but slow


Pc=2.900 psi 1000
Pc=5.075 psi Pc=2.900 psi

Pc=20.01 psi Pc=5.075 psi


0.80 Pc=36.250 psi Pc=20.01 psi

• Potential loss of capillary contact Pc=72.500 psi


Pc=101.500 psi 100
Pc=36.250 psi
Pc=72.500 psi

...
Pc=101.500 psi
0.60

• Potentially slow drainage

Water saturation, Sw

Resistivity Index, RI
0.40
10

0.20

1
0.00 0 50 100 Time, days 150 200
Air-Water Capillary Pressure (psi)

0 50 100days
Time, 150 200

1000
Transient data

Water Saturation
Eqm data-Dr

13.00
100

Resistivity index, RI
10

1
0.01 0.10 Sw (frac.)
Water saturation, 1.00

Time (days)

Water Saturation
17
Imbibition Pc (water-oil)

4
0

• Example results oil-brine 0.0 0.2 0.4 0.6 0.8 1.0

imbibition Pc
-20

• Lab average Sw does not -40 Senergy Average

(psi)
(Forbes-1 Press.)
Senergy Average

agree with Dean-Stark

Pc (psi)
Capillary Pressure
(Forbes-2 Press.)
-60
Senergy Endface
(Forbes-1 Pc)

• If average Sw wrong then end -80


Senergy Endface
(Forbes-2 Pc)
Rep. Lab Average

face Sw and Pc-Sw wrong -100


(Forbes-2 Pc)
Rep. Lab Endface
(Forbes-2)

• Did lab not think Sro = 40%- -120


Rep. Lab DS Sw

50% strange? Water Saturation


Sw (frac.)

• 3 iterations (and about 3 0


10

months) before lab’s 0.0 0.2 0.4 0.6 0.8 1.0

-20
calculated Pc-Sw curves
matched our calculations Capillary Pressure (psi)
-40 Senergy Average
(Forbes-1 Press.)
Senergy Average

• Lab upper-management were


Pc (psi)
(Forbes-2 Press.)
-60 Senergy Endface
(Forbes-1 Pc)

initially unaware of the issues -80


Senergy Endface
(Forbes-2 Pc)
Rep. Lab Average

• errors later corrected -100


(Forbes-2 Pc)
Rep. Lab Endface
(Forbes-2)

• Plugs found to be fractured -120


Rep. Lab DS Sw

18
Water Saturation
Sw (frac.)
Relative permeability

•“Most relative permeability data are rubbish – the


rest are wrong!” Jules Reed, LR Senergy, 2013
1
Clean
0.9 State
>200 samples – 6 usable Fresh
0.8
State
Residual Oil Saturation (v/v)

0.7
Restored
State
0.6
C = 0.6
0.5

0.4 C = 1.5

0.3
C = 2.5
0.2

0.1 C = 10

0
0 0.2 0.4 0.6 0.8 1

Initial Oil Saturation (v/v) 19


Why are they rubbish?

• Plugs unrepresentative or plugged incorrectly


• Swir too high and/or non-uniform
• Wettability contaminated or unrepresentative

30
Dynamic Displacement

25 Porous Plate

WW SWW MW SOW OW
20
1
Swi (%)

15 180 psi 0.9

??? 0.8

10

Relative Permeability
0.7

0.6

5 0.5
200 psi
0.4
0
0.3
Lab A Lab B Lab C Lab D
0.2

0.1

0
0 0.2 0.4 0.6 0.8 201
Saturation
Why are they wrong?

• Coreflood testing invalidates analytical theory


• Flow is linear and uni-directional
• Capillary effects are negligible
Water Saturation

Water Saturation (-)


Ncres x100
Ncres x10
Ncres

Sample Length

Length along core (slice)

21
Capillary end effects
Nc_res x100 Nc_res x10 Nc_res
1

C a p lla ry P re s s u re
Differential Pressure

Ncres x100
Ncres x10 0 Nc_res x100 Nc_res x10 Nc_res
1

Ncres 0

-1 -1

Capllary Pressure
-2

-3

-2 -4

-5

-6

-3 -7

-8
0 0.2 0.4 0.6 0.8 1
Water Saturation

-4

-5

Sample Length -6

-7

-8
Water Saturation

0 0.2 0.4 0.6 0.8 1


Water Saturation

Saturation is controlled by capillary number (Nc)

Ncres x100 Nc = k DP
Ncres x10
Ncres s Dx
Sample Length
What are the solutions?

• Carefully review legacy data


Drilling & Petrophysics & Reservoir
• Identify uncertainties and impact on: Completions Geology Engineering

• In place calculations
• Recovery factor
Focal point

• What is the value of information? Laboratory

• Is it worth doing the experiments at all?


• Or is it because we have a table to fill in in Eclipse
• New core data
• learn from legacy data review
• integrated program design
• focal point
• improved test and reporting documentation

23
What are the solutions?

• Lab audit
• Assess resources, equipment,
experience and expertise of
management and technicians
• Check plugs
• Test data set interpretation
• Design programme with
stakeholders and lab
• Do not “cut and paste” from
previous jobs
• Do not pick from a “menu”
• Draw up flowchart
• Look where value added at little
incremental cost
• Iterate, iterate, iterate

24
What are the solutions?

• Relative permeability
• Ensure wettability is representative
• Test design
• In situ saturation monitoring
• Coreflood simulation

25
Relative permeability - ISSM

X-ray adsorption
• Reveals what is going on in the core plug

0 Sw(NaI) 100
% %

Water Saturation

26
Length along core (slice)
Relative permeability - coreflood simulation

• Recommended practice for ALL relative permeability tests


• Several non-unique solutions are possible so need to sense check

27
Test specifications/data reporting

• Detailed test and reporting specifications


• define test procedures and methods
• Define what, when and how reported
• experimental data essential
• use to verify and check lab calculations
• allows alternative interpretation
• most labs retain experimental data only for short time
• Tedious and time consuming … but
• essential in data audit trail
• invaluable in unitisation
• can save money as you may not have to repeat tests

28
Test specification example – centrifuge Pc

29
Plugbook
Core Plug History Chart

Plug Parameters Digital Images: Side and End Face

• Plug data
Sample No.: 116
Depth (m) : 3906.20 Pre-test photographs & CT images:
Length (cm) : 5.02
Diameter (cm) : 3.88 Paste here Paste here

• Base properties Plug Base Data


Ambient
Air Permeability (md) : 0.340

• porosity and permeability


Porosity (%) : 10.6
Grain Density (g/cc) : 2.648

Overburden 3035 psi

• History
Air Permeability (md) : 0.182
Porosity (%) : 10.1 Paste here

• when/how cut, cleaned & dried


Study Flow Chart

Sample preparation
Drilled with Brine : 23-May-07

• SCAL test history Hot solvent cleaned & oven dried @ 95°C
In 14-Jul; out 21-Jul
Paste here Paste here

• Plug CT scans
CT scan & pre-test plug photography

Permeability, porosity and grain density

• Heterogeneity
Pressure saturate & Archimede's porosity

Formation factor & resistivity index @ NOBP

• Damage?
Dean-Stark
Paste here
Hot solvent cleaned; oven dried @ 95°C; and Kphi

• Plug photographs
Pressure saturate & Archimede's porosity

Centrifuge air-brine capillary pressure

• pre-and post-test
Dean-Stark

Hg injection and CEC on offcuts


Post-test photographs:

• Can be easily customised


Post-test photography
Paste here Paste here
Report

Paste here

30
Summary

• Lab test pitfalls have a huge


impact on core analysis modelling
data input
• But....
• uncertainties are recognisable
and manageable
• best practice, real-time QC, and
robust workflows ensure that core
data are fit for purpose prior to
petrophysical analysis.
• a forensic data quality
assessment can minimise data
redundancy and reduce
uncertainty in reservoir models

Price is what you pay. Value is what you get - Warren Buffet
31
Questions?

32

You might also like