PERA

Preparing Oil & Gas PVT Data

for
Reservoir Simulation

Curtis H. Whitson
NTNU / PERA
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA Tasks
• Collecting samples.
• Which PVT lab tests to use.
• Designing special PVT studies.
• Quality controlling PVT data.
• Heptanes-plus data and characterization.
• Initial EOS model.
• Tuning an EOS model.
• Viscosities.
• Fluid initialization.
• Minimizing number of EOS components.
• Black-oil PVT tables.
PERA
Collecting Samples

Why?

1. PVT data to develop a model.

PERA
Collecting Samples

Why?

1. PVT data to develop a model.

2. Compositions for fluid initialization.

PERA
Collecting Samples

Why?

1. PVT data to develop a model.

2. Compositions for fluid initialization.

3. Crude assays for process design.

PERA
Collecting Samples
How?
Oils
• Bottomhole samples.
• Surface separator samples.
• MDT / RCI
Gas Condensates
• Surface separator samples.
• MDT / RCI
PERA
Collecting Samples
How?
Oils
• Bottomhole samples.
• Surface separator samples.
• MDT / RCI
Gas Condensates
• Surface separator samples.
• MDT / RCI
Saturated Gas / Oil Systems
• Gas-cone an oil producer – perfect!.
• ECM (equilibrium contact mixing)
PERA Open-hole Samplers
MDT / RCI
• Potential Problems
• Oil-based muds.
• Oils -- OK usually.
• Gas condensates – Also OK (quite good).

• Surface cooling before removal.

• Bubblepoint suppression.
 MDT Sampling with MPSR bottles
PERA To pump and Fire-open valve
To pump and open
formation
formation
To pump and
Fire-open valve
formation
operated to fill
Fire-close valve
closed

Fire-open valve Fire-close valve

operated post Dead volume
filling (<10cc) – now gas
Fire-close valve
filled and the gas
will be lost

Dead volume
Dead volume (<10cc) – now oil
(<10cc) – initially Manual close
filled valve now
water filled
operated to
extract MPSR
Manual close from MDT tool
Manual close valve
valve

450cc of 2 phase
450cc of single hydrocarbon at
Piston

MPSR
phase oil at res surface temp
MPSR

temp and and some

pressure pressure
MPSR

Water from dead

Water from dead
450cc MPSR volume
volume
bottle
Piston
Piston

Post Sampling Post Sampling

Prior to Sampling but downhole Now at surface
PERA
Which PVT Lab Tests to Use

What are you simulating?

• Depletion.
• Water injection.
• Condensate blockage.
• Gas injection.
• Miscible.
• Immiscible.
PERA
Designing Special PVT Studies

• Condensate Blockage.
• Condensate viscosities.

• Miscible Gas Injection.

• Through-critical swelling test.
• Vro , compositions and K-values!

• Immiscible Gas Injection.

• Vaporization tests.
PERA
Quality Controlling PVT Data

• Compositions !!!
• Recombination.
• Extended GC.
• Mass-to-mole conversion.
• C7+ properties.
• Molecular weight and specific gravity.
• Use trend plots.
• Ps vs wt-% methane and/or C7+.
 J-476XDST4BHS

PERA 100
Reported (GC)
Expontential Model

Molar Composition, mol-%

10

1
80 100 120 140 160 180 200 220 240 260 280
Molecular Weight
 J-482BHS
PERA
100
Reported (GC)
Expontential Model
Expon. (Reported (GC))
Molar Composition, mol-%

Reported GC extended distribution

appears to be in serious error, being
much too "light" with apparent
M7+ = 130
DON'T USE GC DISTRIBUTION !!!

10

1
80 100 120 140 160 180 200 220 240 260 280
Molecular Weight
PERA
C7+ Data and Characterization

• Correlate MW and SG of C7+.

• Define trends & identify ”outliers”.
• Use TBP Data.
• Gamma distribution model fit.
• SCN MW-SG relationship.
• Downstream Assay data always available.
• Extended GC Data.
• Gamma distribution model fit.
• Ignore heaviest amount and MW.
PERA

Soreide Fc Correlation

0.88 Bottomhole Samples

0.87 Separator Samples
0.86 Best-Fit Fc=0.287
0.85 Fc=.28
Specific Gravity

0.84 Fc=.27
0.83

0.82

0.81 Becoming M ore

Paraffinic Due to
0.80 Wax Accumulation
in Samples ???
0.79

0.78
150 170 190 210 230 250 270 290 310
Molecular Weight
PERA
C7+ Properties (Watson Correlation)
930

910
Decreasing
Aromaticity:
C7+ Density, kg/m3

Asphaltene
890 Loss?

870

Reported / Determined
850 Kw=11.2
Kw=11.5
Kw=11.8
830
150 170 190 210 230 250 270 290
C7+ Molecular Weight
PERA
Assay Data
TBP Distillation (Well 15-2-RD-2X)
1.000

0.950

0.900
Specific Gravity

0.850

0.800 Discontinuity?
Maybe associated with
0.750 change in distillation
pressure.
0.700

Riazi Correlation
0.650
Soreide (after fit Riazi)
0.600
0 100 200 300 400 500 600 700 800

Molecular Weight
PERA
Initial EOS Model
• Default Parameters – don’t mess with ’em.
• C6- properties M, Tc, pc, .
• C6- properties volume shift s(=c/b).
• Non-HC / HC BIPs kij.
• C7+ Characterization.
• Minimum 3 fractions (not C7, C8, C9+ !).
• Methane-C7+ BIPs.
• SG-TB-MW relationship; Tc, pc, (Tb).
• Volume shift treatment s().
• Always keep fraction SGs ”fit” by EOS.
PERA
Tuning an EOS Model

• Densities Don’t Need Regressing!

• What’s Left to Fit?

• Nothing but K-values ... but how ???

• Check Consistency!
• Monotonic K-values of hydrocarbons.
• Three-phase existence (from EOS model).
• Serious problem for EOS models!
PERA
Viscosities

• LBC (Lorenz-Bray-Clark / Jossi-Thodos)

• Need accurate densities.
• Modify C7+ Vc values.
• Make sure fraction viscosities are monotonic.
• LBC polynomial coefficients.
• BE CAREFUL!
• Pedersen.
• Better predictions than LBC.
• Regression - ?
PERA
Fluid Initialization

• Plot C6+ versus Depth.

• Initial Oil in Place plot.
• Use error bars.
• Depth and composition.

• Uncertainty Analysis.
• Use isothermal gradient model.
• Defines maximum compositional variation.
• Use constant composition.
• Defines minimum compositional variation.
PERA

Reference Depth
-11000 GOC

-12000
Depth, ft SSL

Field-Data Based Isothermal

Initialization Model
-13000

-14000

-15000
0.2 0.4 0.6 0.8

IOIP / HCPV, (Sm 3 / m3)

PERA

Reference Depth
-11000 GOC

-12000
Depth, ft SSL

Field-Data Based Isothermal

Initialization Model
-13000

-14000

-15000
10 15 20 25 30 35

C7+ Mole Percent

PERA
3800

4000
True Vertical Depth, mSS

4200

Well B
4400

4600 Well D
Well A DST 2
Well E

4800 Well A DST 1

Well C

5000
0 5 10 15 20 25 30 35
C7+ Mole Percent
PERA
3800

4000
True Vertical Depth, mSS

4200

Well B
4400

4600 Well D
Well A DST 2
Well E

4800 Well A DST 1

Well C

5000
0 5 10 15 20 25 30 35
C7+ Mole Percent
PERA
3800

4000
True Vertical Depth, mSS

4200

Well B
4400

4600 Well D
Well A DST 2
Well E

4800 Well A DST 1

Well C

5000
0 5 10 15 20 25 30 35
C7+ Mole Percent
PERA
Fluid Initialization

• Black-Oil vs Compositional.
• Use consistent EOS model.
• Use consistent surface process.
• Use solution GOR (Rs and Rv) for black-oil
model.
• Based on EOS model initialization.
PERA
Minimizing Number of EOS Components

• Basis of Comparison.
• Detailed & Tuned EOS model.
• Stepwise lumping procedure.
• Check entire relevant p-compositioni space.
• Depletion data.
• Gas injection data.
• Miscibility data.

• Delumping ?
• Detailed & Tuned EOS model.
PERA
Black-Oil PVT Tables

• Select Depletion Test.

• Define Surface Separation.
• Consistency.
• Negative compressibilities.
• Saturated gas / oil systems.
• Compositional grading.
• Extrapolation.
• Undersaturated GOC (ECL100).
• Gas injection.
PERA
Black-Oil PVT Tables

• Delumping to Compositional Streams ?

PERA Split Factor
BOz Conversion

z1 2
zi   Sij  q j q1 = qg
z2 q2 = qo
j1
qg
Sij .
qo . (1 rsCog ) rs (Coo  R s )
Si1  yi  xi
. k 1 rsR s  k 1 rsRs 
zn

(Coo  Rs ) Rs (1 rsCog )

Si2  xi  yi
k 1 rsRs  k 1 rsRs 
PERA
North Sea Full-Field
Black-oil to Compositional conversion
• 2 Platforms / 2 Processes.
• ~ 50 wells.
• ~ 1000 well-grid connections.
• Gas injection.
• 2 Black-oil PVT regions.
• Huge (GB) summary files.
• > 100,000 stream conversions.
PERA North Sea Full-Field Model
E100-BO
E300-EOS B
FFM
A
Platform Platform
A B Different Surface
Process A Process B Processes (BO PVT)
~ 30 Wells ~ 15 Wells in Regions A & B

Gas
Injection
PERA
Objective
• Run black-oil full-field reservoir model.

• Convert surface rates to compositional

streams.
• Connection level conversions.

• Summarize results.
• By well, platform, field.
• Annually, quarterly, cummulatives etc.
PERA

Full-Field Rate Forecast

(Following history match from 1987)
2.6E+06 8000

2.4E+06 7000

2.2E+06 6000
Gas Rate, Sm3/Day

Oil Rate, Sm3/Day

2.0E+06 5000
e100-bo Gas Rate
1.8E+06 e100-bo Oil Rate 4000

1.6E+06 3000

1.4E+06 Changing Group Gas Rate due to reduced 2000

contribution from neighbouring fields.
1.2E+06 1000
2000 2001 2002 2003 2004 2005
Time, Year
PERA

Full-Field Molar Rate Predictions

(E100 - BOz conversion)
30000 100000
C1
C6+ 90000
C3C4 & C6+ Molar Rate, kmol/d

25000
80000

C1 Molar rate, kmol/d

C3C4
70000
20000
C6+
C1
60000
C1
15000 50000

40000
C6+
10000
30000
C3C4 C3C4
20000
5000
10000

0 0
2000 2001 2002 2003 2004 2005
Time, Year
PERA

Full-Field Molar Rate Predictions

(E300-BOZ/PSM vs E300 models)
Validation of Conversion Accuracy
30000 100000
C1
C6+
90000
25000
80000
C3C4 & C6+ Molar Rate,

C1 Molar Rate, kmol/d

70000
20000
60000
kmol/d

C1
ECL300: Z
15000 ECL300: BO to Z 50000

C6+ 40000
10000
C3C4 30000
C3C4
20000
5000
10000

0 0
2000 2001 2002 2003 2004 2005
Time, Year