a new probabilistic approach to proved reserves estimation...

10/10/2014

1

REGIONAL ASSOCIATION OFOIL, GAS & BIOFUELS SECTOR COMPANIES

IN LATIN AMERICA AND THE CARIBBEAN

A New Probabilistic Approach to Proved Reserves Estimation in

Mature Fields

Brayam Valqui OrdoñezPETROPERÚ86th ARPEL Experts Level Meeting (RANE) “Management of Reservoirs” October 6-7, 2014 – Buenos Aires, Argentina

86th ARPEL Experts Level Meeting (RANE) “Management of Reservoirs“ – October 6-7 – Buenos Aires, Argentina

Outline

� Reserves Estimation Framework

� Conventional Deterministic Approach

� New Probabilistic approach to Reserves Evaluation in Mature Fields

� Case Study

o Block UNI: Field Development Plan and Reserves Estimation

� Conclusions

2

10/10/2014

2

86th ARPEL Experts Level Meeting (RANE) “Management of Reservoirs“ – October 6-7 – Buenos Aires, Argentina 3

Reserves Estimation FrameworkThe volume of existing hydrocarbons can be classified and categorized following PRMS guidelines in two

main factors: Degree of technical uncertainty (X axis) and Opportunity of commerciality (Y axis).

Incr

ea

sin

gC

ha

nce

of

Co

me

rcia

bil

ity

Range of Uncertainty

Represent Uncertainty in Our Estimates by Assigning Volumes to Categorize


Reserves Estimation FrameworkResources and Reserves Calculation Procedures

Decline curve analysis and Material Balance Estimations are the preferred

methods to calculate reserves in Brown Fields.

Appraisal and Initial

Development Stages

Volumetric Analysis

(Exploration, Appraisal and

Development Stages)

Material Balance

(Early Production)

Reservoir Simulation

(Early Decline)

Decline Curves

(Late Decline)Election

Oil Project Maturity and Availability of Information

10/10/2014

3


Conventional Deterministic Approach:

Deterministic Decline Curve Analysis

Deterministic Field Production Profile

Based in Experts’ Opinion

Case Type Decline Model

High Case Armónica

Base Case Hiperbólica

Low Case Exponencial

Low Case

Base Case

High Case

Companies usually assess three scenarios of performance

Pro

ve

dR

ese

rve

s

Un

cert

ain

ty


New Probabilistic Approach - Guidelines:• It is recommended to be used for providing consistency and confidence to proved reserves estimation in

mature fields.

• Three deterministic scenarios (Low, Base and High) were conceived based on expert opinion to initialize the

performance model.

• Ranges of uncertainty and statistical distribution functions of the decline rate (Di) and “b” exponent were

estimated.

• Rank correlation factors need to be carefully defined by expert opinion in order to set dependency between

b and D.

• Probabilistic production profiles were obtained by applying Monte Carlo Simulation; an expectation curve of

proved reserves is eventually built and conventional low, base and high scenarios can be extracted to

deterministic cash flow analysis.

• For undeveloped proved reserves, the methodology can be applied in the same way by estimating a type

curve for each field and/or reservoir. In addition, a hyperbolic model is associated to this type curve to be

treated under probabilistic approach by applying Monte Carlo Analysis.

• Total proved reserves are obtained by adding corresponding PDP and PUD production profiles.

10/10/2014

4


New Probabilistic Approach:

Flowchart - Holistic Probabilistic Model

Evaluation of the Production

and Pressure History

Performance Method

Selection: MBE or DCA

Verification of the Productive behavior

with Dynamic Reservoir Model

Built a Type Curve for each Field

and/or Reservoir

Assigning PDFs for Model

Parameters

Production Forecasting

under Uncertainty


New Probabilistic Approach

With the production history and using a decline curve analysis(DCA) software, we plot three scenarios of declination (2hyperbolic and 1 exponential).

195254 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98200002 04 06 08 10 12 14 16 18 20 22 2410

-2

10-1

100

101

102

103

104

105

pe

tday

, bb

l/d

Date

Working Forecast ParametersPhase : OilCase Name : reservesb : 0.867662Di : 0.0186723 M.e.qi : 76.9357 bbl/dt i : 01/31/2014te : 12/31/2025Final Rate : 19.0487 bbl/dCum. Prod. : 0.0769357 bbl/dCum. Date : 01/31/2014Reserves : 157.063 bbl/dReserves Date : 12/31/2025EUR : 157.14 bbl/dForecast Ended By : TimeDB Forecast Date : Not SavedReserve Type : Proven-Developed

195254 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98200002 04 06 08 10 12 14 16 18 20 22 2410

-2

10-1

100

101

102

103

104

105

pet

day,

bb

l/d

Date

Working Forec as t ParametersPhase : OilCase Name : reser vesb : 1.4264Di : 0.0135676 M.e.qi : 76.9357 bbl/dti : 01/31/2014te : 12/31/2025Final Rate : 30.1019 bbl/dCum. Prod. : 0.0769357 bbl/dCum. Date : 01/31/2014Reser ves : 195.881 bbl/dReser ves Date : 12/31/2025EUR : 195.958 bbl/dForecast E nded By : TimeDB For ecast Date : Not SavedReser ve Type : Proven-Developed

195254 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98200002 04 06 08 10 12 14 16 18 2010

-2

10-1

100

101

102

103

104

105

petd

ay,

bbl

/d

Date

Working Forecast ParametersPhase : OilCase Name : reservesb : 0Di : 0.0236751 M.e.qi : 76.9357 bbl /dti : 01/31/2014te : 03/31/2021Final Rate : 9.8129 bbl/dCum. Prod. : 0.0769357 bbl/dCum. Date : 01/31/2014Reserves : 85.2696 bbl /dReserves Date : 03/31/2021EUR : 85.3465 bbl /dForecast Ended By : RateDB Forecas t Date : Not SavedReserve Type : Proven-Developed

Probabilistic density functions are chosen to modeluncertainty in decline curve parameters accordingly tohistorical production performance and expert opinion

b D

The Monte Carlo method is applied, then the simulation is donefor every month until 2027 (A hyperbolic decline model is used)

Año 2014Año 2014

Año 2027Año 2027

biiO tbDQQ

1

)1(−

+=

The results of the monthly production forecasts areranged in percentiles; as a result, probabilisticcumulative production curves are obtained.

0

50

100

150

200

250

300

350

Jan

-14

Jul-1

4

Feb

-15

Sep

-15

Ma

r-16

Oct-1

6

Ap

r-17

No

v-17

Ma

y-18

De

c-1

8

Jul-1

9

Jan

-20

Au

g-2

0

Feb

-21

Sep

-21

Ma

r-22

Oct-2

2

Ma

y-23

No

v-23

Jun

-24

De

c-2

4

Jul-2

5

Jan

-26

Au

g-2

6

Ma

r-27

MS

TB

Pronóstico de Producción Probabilístico

Reservas Probadas: SPE Field

90%

85%

80%

75%

70%

65%

60%

55%

50%

45%

40%

35%

30%

25%

20%

15%

10%

0

50

100

150

200

250

300

350

Jan

-14

Jul-1

4

Feb

-15

Sep

-15

Ma

r-16

Oct-1

6

Ap

r-17

No

v-17

Ma

y-18

De

c-1

8

Jul-1

9

Jan

-20

Au

g-2

0

Feb

-21

Sep

-21

Ma

r-22

Oct-2

2

Ma

y-23

No

v-23

Jun

-24

De

c-2

4

Jul-2

5

Jan

-26

Au

g-2

6

Ma

r-27

MS

TB

Pronóstico de Producción Probabilístico

Reservas Probadas: SPE Field

90%

85%

80%

75%

70%

65%

60%

55%

50%

45%

40%

35%

30%

25%

20%

15%

10%

−

−= −− )1()1(

11)1( bb

ii

bi

P qqDb

qN

10/10/2014

5


� Using the production history and a decline curve analysis software, three declinescenarios are plot (2 hyperbolic and 1 exponential). Each decline scenario yieldscorresponding b and D parameters with a Qi= 76.9 bpd).

195254 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98200002 04 06 08 10 12 14 16 18 20 22 2410

-2

10-1

100

101

102

103

104

105

petd

ay,

bbl/d

Date

Working Forecast ParametersPhase : OilCase Name : reservesb : 0.867662Di : 0.0186723 M.e.qi : 76.9357 bbl/dti : 01/31/2014te : 12/31/2025Final Rate : 19.0487 bbl/dCum. Prod. : 0.0769357 bbl/dCum. Date : 01/31/2014Reserves : 157.063 bbl/dReserves Date : 12/31/2025EUR : 157.14 bbl/dForecast Ended By : TimeDB Forecast Date : Not SavedReserve Type : Proven-Developed

195254 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98200002 04 06 08 10 12 14 16 18 20 22 2410

-2

10-1

100

101

102

103

104

105

petd

ay, b

bl/d

Date

Working Forecast ParametersPhase : OilCase Name : reservesb : 1.4264Di : 0.0135676 M.e.qi : 76.9357 bbl/dti : 01/31/2014te : 12/31/2025Final Rate : 30.1019 bbl/dCum. Prod. : 0.0769357 bbl/dCum. Date : 01/31/2014Reserves : 195.881 bbl/dReserves Date : 12/31/2025EUR : 195.958 bbl/dForecast Ended By : TimeDB Forecast Date : Not SavedReserve Type : Proven-Developed

Hyperbolic Hyperbolic b > 1

195254 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98200002 04 06 08 10 12 14 16 18 2010

-2

10-1

100

101

102

103

104

105

petd

ay, b

bl/d

Date

Working Forecast ParametersPhase : OilCase Name : reservesb : 0Di : 0.0236751 M.e.qi : 76.9357 bbl/dti : 01/31/2014te : 03/31/2021Final Rate : 9.8129 bbl/dCum. Prod. : 0.0769357 bbl/dCum. Date : 01/31/2014Reserves : 85.2696 bbl/dReserves Date : 03/31/2021EUR : 85.3465 bbl/dForecast Ended By : RateDB Forecast Date : Not SavedReserve Type : Proven-Developed

Exponential

b=0

D=2.37%

Model Inputs

Case Study: Block UNI

“b” and “D” have a inverse relationship

b=0.86

D=1.87%

b=1.42

D=1.36%


• Beta-pert distributions were used to model b and D under uncertainty.

• Monte Carlo simulation was applied monthly until 2027; probabilistic production profiles and cumulative curves were

developed.

b

D

Probabilistic Hyperbolic Decline Model for Oil Production

Dependency Di and b

Di

b

Case Study: Block UNI

Year: 2014Year: 2014

Year: 2027Year: 2027

|)1(1

biiO tbDQQ

−

+=

−

−= −− )1()1(

11)1( bb

ii

bi

P qqDb

qN

10/10/2014

6


Case Study: Block UNIProbabilistic Production Profile – Base Curve of the Block UNI

• Nine (9) producing wells from Ostrea Fm. were used to recover proved developed reserves (PD).


Case Study: Block UNIProbabilistic vs. Deterministic

Production Forecasting

Prob. High Case

Det. High Case

Base Case

Prob. Low Case

Det. Low Case

The best way to evaluate the uncertainty

is using the Probabilistic Approach

10/10/2014

7


Case Study: Block UNIProbabilistic Production Profile – Additional Work Program of the Block UNI

• Five (5) infill wells were agreed to be drilled to Ostrea Fm. to produce proved undeveloped reserves (PUDs).


Case Study: Block UNIProbabilistic Production Profile – Total Proved Reserves of the Block UNI

• Fourteen (14) developed wells were used to forecast total proved reserves (PT).

10/10/2014

8


Case Study: Block UNIProbabilistic Production Profile – Cumulative Production of the Block UNI

• From 200 to 300 MSTB of oil would be recovered in an interval confidence of 80%.


Conclusions• It is recommended to be used for providing consistency and confidence to proved reserves

estimation in mature fields.

• The methodology allows creating a parametric analytic model based on production history

and reservoir behavior. This production profile is used to forecast a “band” of proved

recoverable volumes under a confidence interval.

• Dependency between model inputs should be set in order to provide consistency to

forecasting probabilistic production profiles.

• To estimate proved undeveloped reserves is recommended first to build a “type curve” by

reservoir (or at least by well) using analogy principles; then, this type curve should be

converted to a “proxy” analytical model (using both/either DCA and/or MBE) to be run under

probabilistic approach by Monte Carlo simulation.

• A probabilistic cash flow analysis is the next step to evaluate the profitability of the project. A

probabilistic production profile should be the input to make a reliable cash flow analysis for

the project.

10/10/2014

9

17

Thank you

Questions?

a new probabilistic approach to proved reserves estimation...

Documents