recent work in well performance analysis for tight gas ... · presentation at marathon oil marathon...

Presentation at Marathon OilMarathon Shale Workshop: Reservoir Engineering Session

Recent Work in Well Performance Analysis for Tight Gas Sands and Gas ShalesT.A. Blasingame/Texas A&M U. (06 November 2009) Sl

ide

—1/

63

Recent Work inWell Performance Analysis for

Tight Gas Sands and Gas Shales


06 November 2009 — Houston, TX

Tom BLASINGAMEDepartment of Petroleum Engineering

Texas A&M UniversityCollege Station, TX 77843-3116 (USA)

+1.979.845.2292 — [email protected]



ide

—2/

63

SPE 110187Evaluating the Impact of Waterfrac

Technologies on Gas Recovery Efficiency: Case Studies Using Elliptical

Flow Production Data Analysis

D. Ilk,* Texas A&M U.J.A. Rushing, Anadarko Petroleum Corp.R.B. Sullivan, Anadarko Petroleum Corp.

and T.A. Blasingame, Texas A&M U.*Department of Petroleum Engineering

Texas A&M UniversityCollege Station, TX 77843-3116




ide

—3/

63

Elliptical Flow Type Curves: [Amini et al (2007)]

●Discussion: Elliptical Flow Type Curves■ Equivalent constant rate type curves shown in qD versus tDA format.■ Elliptical fracture conductivity: FE = 1, 10, 100, 1000.■ Elliptical geometry parameter: ξ0 = 0.25, 0.50, 0.75, 1, 1.50, 1.75, 2, 3, 4, 5.

■ Type curve for a fractured well centered in a closed (homogeneous) elliptical reservoir — FE=10, various ξ0-values; qD functions versus tDA format.

■ Type curve for a fractured well centered in a closed (homogeneous) elliptical reservoir — FE=1, various ξ0-values; qD functions versus tDA format.



ide

—4/

63

Elliptical Flow Type Curves: [Amini et al (2007)]

●Discussion: Elliptical Flow Type Curves■Objective: To demonstrate the elliptical boundary model as a diagnostic

to establish the elliptical (transition) flow regime using production data.■We utilize type curve solutions in terms of the equivalent constant rate

case presented in "decline" form (qD and tDA).

■ Properties that can be estimated using production data analysis based on elliptical flow:— Effective gas permeability, k, md — Effective fracture half-length, xf, ft— Effective fracture conductivity, FE— Contacted gas-in-place, G, BSCF— Effective drainage area, A, acres

■ The characteristic elliptical boun-dary parameter (ξ0) correlates all of the aspects of the drainage area and the fracture half-length:

5.0

0 )2sinh(12

⎥⎦

⎤⎢⎣

⎡=

ξπAx f



ide

—5/

63

Procedure:Step 1: Data Overview: "Production History Plot"

General data quality/correlation.Step 2: Data Correlation: "pwf (or ptf) vs. Rate Plot"

Crude comparison — used only for general trends.Step 3: Clean/Edit Data for Clarity: Log-Log Data Plot(s)

Remove spurious data from base data trends (e.g., q/Δp).Step 4: Identify Flow Regimes (Diagnostics): Log-Log Data Plot(s)

Characteristic flow regimes from normalized PI-style plots.Step 5: Compare Data to Reservoir Model:

"Type Curve" matching (data onto a reservoir model).Step 6: Refine Match/Model:

Improve match of model parameters (k, s, xf, FcD, ...) using individual type curves, simulation models, and/or regres-sion methods.

Step 7: Summary History Match:Final "history match" of model and raw data (pwf and q).

Production Analysis: Overview



ide

—6/

63

Case Studies: Small Waterfrac — 20/40 Proppant

●Discussion: Diagnosis of the Well Performance Data■ A consistent character in the rate and pressure functions is observed

— qg and pwf data correlation appears to be good to very good.■ qg/Δpp diagnostic functions appear to be relevant — earliest third of

data appear off-trend (weak qg and pwf correlation at early time?).■ Treatment — 7745 bbls of slick water; 63,000 lbs 20/40 proppant.

■ Diagnostic log-log plot (dimensionless rate decline and rate decline integral functions).

■ Production history plot (daily flowrate and pressure measurements).



ide

—7/

63

Case Studies: Small Waterfrac — 20/40 Proppant

●Discussion: Analysis Results■ Discrepancy at early times is observed — probably an early-time artifact

(shape of integral-derivative), possibly some modest fracture damage.■ Excellent match of the model and the rate history, good match of the

pressure history — "balance" of analysis between the diagnostic functions and the production history.

■ k=0.0097 md, xf=129 ft, FE=10, ξo=2.0 (near-circular), G=1.28 BSCF.

■ Production history plot with model match (excellent flowrate match, good pressure match).

■ Elliptical boundary decline type curve match (moderate conductivity, near-circular drainage geometry).



ide

—8/

63

Case Studies: Large Waterfrac — 20/40 Proppant

●Discussion: Diagnosis of the Well Performance Data■ Erratic pressure and rate data observed from 600-1200 days (we were

told a workover was performed to remedy a "tubing or packer leak").■ Significant well clean-up effects in the pressure data during the first 2

months of production (poor correlation?).■ qg/Δpp diagnostic functions indicate a strong transient flow signature

(≈ 1/2 → high fracture conductivity), BDF is also observed.■ Treatment — 7300 bbls of slick water; 247,500 lbs 20/40 proppant.





ide

—9/

63

Case Studies: Large Waterfrac — 20/40 Proppant

●Discussion: Analysis Results■ Excellent type curve match obtained using an elliptical flow model with a

high fracture conductivity (all flow regimes matched!).■Outstanding match of the model and the rate history (qg), good match of

the pressure history (pwf) — surface pressure data is not representative of the bottomhole condition (up to 600 hr) due to "tubing/packer" leak.

■ k=0.0039 md, xf=184 ft, FE=1000, ξo=1.50 ("fat" ellipse), G=3.31 BSCF.

■ Production history plot with model match (excellent flowrate match, fair pressure match).

■ Elliptical boundary decline type curve match (very high conductivity, "fat" elliptical drainage geometry).



ide

—10

/63

Case Studies: Hybrid Waterfracs

●Discussion: Diagnosis of the Well Performance Data■ Erratic nature of the flowrate profile is seen (liquid loading). Major

features of the liquid loading observed for the flowrate data are reasonably well-correlated with the calculated bottomhole pressures.

■ Very clear diagnostic trends are obtained after judicious editing of the qg/Δpp function (1/2 slope during transient flow, 1/1 slope for BDF).

■ Treatment — 2082 bbls of slick water + 4827 bbls of cross-link gel; 510,140 lbs 20/40 proppant.





ide

—11

/63

●Discussion: Analysis Results■ Extraordinary "type curve match" is observed for this case — the well

is effectively stimulated and a low reservoir permeability is expected.■Model flowrate response corresponds well with the observed flowrate

data. The pressure match is not as poor as it might seem given the very erratic nature of the flowrate function.

■ k=0.0030 md, xf=200 ft, FE=1000, ξo=1.00 ("thin" ellipse), G=1.60 BSCF.

■ Production history plot with model match (very good flowrate match, acceptable pressure match).

■ Elliptical boundary decline type curve match (very high conductivity, "thin" elliptical drainage geometry).

Case Studies: Hybrid Waterfracs



ide

—12

/63

Case Studies: Discussion of Results

●Discussion: Effect of the Contacted Gas-In-Place■ Expectation — higher reservoir permeabilities, higher contacted gas-

in-place estimates.■Most small waterfrac cases are off-trend (a single on-trend point with

high permeability and high contacted gas-in-place estimate is noted).■ Excellent correlation of contacted gas-in-place with fracture half-

length — confirming that "the fracture defines the reserves."

■ Results correlation plot — G versus xf.■ Results correlation plot — G versus k.



ide

—13

/63

Case Studies: Discussion of Results

●Discussion: Effect of the Drainage Aspect Ratio■ Clear correlation of increasing permeability with increasing drainage

aspect ratio (ξo).■ Small waterfracs exhibit the highest values of permeability? Function

of development strategy — early wells targeted to best reservoir?■ Inverse correlation between xf and ξo — intuitive (see definition).

■ Results correlation plot — xf versus ξo. ■ Results correlation plot — k versus ξo.

Definitions:

⎥⎥⎥

⎦

⎤

⎢⎢⎢

⎣

⎡=

⎥⎦

⎤⎢⎣

⎡=

−2

10

5.0

0

2sinh21

Or

)2sinh(12

f

f

x

A

Ax

πξ

ξπ



ide

—14

/63

D. Ilk, Texas A&M UniversityA.D. Perego, Anadarko Petroleum Corp. J.A. Rushing, Anadarko Petroleum Corp.T.A. Blasingame, Texas A&M UniversityDepartment of Petroleum Engineering



SPE 116731Exponential vs. Hyperbolic Decline in Tight Gas

Sands — Understanding the Origin and Implications for Reserve Estimates Using Arps' Decline Curves



ide

—15

/63

Rationale For This WorkASSUMPTION: The Arps decline parameter, b, defines the decline

behavior when tight gas sand reserves are assessed.REALITY: Difficult to identify the correct b-parameter during

the early decline period — selection of the wrong b-parameter greatly impacts reserve estimates.

b. (Log-log plot) Production forecast of a tight gas well.a. (Semilog plot) Production forecast of a tight gas well.



ide

—16

/63

Overview: Loss Ratio (Definition and Behavior)Loss Ratio: (basis for exponential rate decline)

Loss Ratio Derivative: (basis for hyperbolic rate decline)dtdq

qD g

g/

1−≡

⎥⎥⎦

⎤

⎢⎢⎣

⎡−≡⎥⎦

⎤⎢⎣⎡≡

dtdqq

dtd

Ddtdb

g

g/

1

]exp[ tDqq igig −=

)/1()(1 bi

gig

tbD

qq

+=

[From: Johnson, R.H. and Bollens, A.L.: "The Loss Ratio Method of Extrapolating Oil Well Decline Curves," Trans., AIME (1927) 77, 771.]



ide

—17

/63

Overview: Arps' Rate Decline FunctionsCase Rate-Time Relation Cumulative-Time Relation

]exp[ tDqq igig −=

)/1()(1 bi

gig

tbD

qq

+=

)(1 tDq

qi

gig +=

Exponential: (b=0)

Hyperbolic: (0<b<1)

Harmonic: (b=1)

]]exp[[1 tDDq

G ii

gip −−=

])(1[1 )(1

)/1(1 bi

i

gip tbD

Dbq

G −+−−

=

)ln(1 tDDq

G ii

gip +=

[From: Arps, J.J: "Analysis of Decline Curves," Trans., AIME (1945) 160, 228-247.]



ide

—18

/63

New Rate Equation: Ilk et al [2008]

)1(ˆ1 nitDnD

dtdq

qD −−

∞ +≈−≡

●Observed Behavior of Decline Parameter (D):

●Solving for Flowrate:

] ˆ exp[ ˆ nii tDtDqq −−= ∞

●Solving for the b-Parameter:

nn

i

i ttDDn

nDnb −

−∞+

−= 2)1( ] ˆ[

)1(ˆ



ide

—19

/63

Field Examples: Small WF Gas Well (SWF2)

●Discussion: Small Waterfrac Gas Well (SWF2)■ Liquid loading effects are obvious in the latter portion of the flowrate data.■ The onset of the boundary-dominated flow regime is observed.■ D∞ is set to 0 initially, then tuned to the latest data — we obtain a very

good match of the D-parameter data trend with the power-law models.■ We observe a very good match of the flowrate data with the "base" power-

law exponential model (i.e., D∞=0).

"q-D-b" Plot: SWF2D-parameter data trend exhibits a power-law behavior — essentially a straight line.b-parameter data trend is not constant (contrary to hyperbolic formulation).Computation of the b-parameter is severely affected by noise.



ide

—20

/63

●Discussion: Small Waterfrac Gas Well (SWF2)■ The hyperbolic rate relation (b=1) yields the highest reserves estimate.■ Excellent matches of data are achieved using the power-law exponential

model for both the D∞=0 and the D∞≠0 cases.■ The lower bound for the reserves estimate is 2.3 BSCF, which is con-

sistent with our results from the model-based PA study (SPE 114947).

Field Examples: Small WF Gas Well (SWF2)

a. Semi-log plot — empirical matches are shown using power-law exponential and hyperbolic models.

b. Log-log plot — empirical matches are shown using power-law exponential and hyperbolic models.



ide

—21

/63

Field Examples: Large WF Gas Well (LWF2)

●Discussion: Large Waterfrac Gas Well (LWF2)■ Erratic rate behavior caused by liquid loading is seen in the latter portion

of the rate data.■ The behavior of the computed D- and b-parameters is almost identical to

the previous case suggesting (to some degree) consistency of the data.■ Outstanding matches of the computed D- and b-parameters with the

power-law exponential model are observed.

"q-D-b" Plot: LWF2D-parameter data trend exhibits a power-law behavior — essentially a straight line.b-parameter data trend is not constant (contrary to hyperbolic formulation).Computation of the b-parameter is signifi-cantly affected by noise.



ide

—22

/63

●Discussion: Large Waterfrac Gas Well (LWF2)■ Estimation of the reserves using the hyperbolic rate decline relation is

almost five times higher than the contacted gas-in-place predicted previously using a model-based match (SPE 114947).

■ Outstanding matches of the data are obtained using the power-law exponential model, both the D∞=0 and the D∞≠0 cases.

■ The D∞≠0 model provides the most conservative estimate of reserves.

Field Examples: Large WF Gas Well (LWF2)





ide

—23

/63

Field Examples: Hybrid WF Gas Well (HWF1)

●Discussion: Hybrid Waterfrac Gas Well (HWF1)■ Severe liquid loading is observed at late times.■ The computed D- and b-parameters reflect the effects of liquid loading —

however; the D-parameter data trend is essentially power-law.■ Good matches of the computed D- and b-parameters are obtained using

the power-law exponential model.

"q-D-b" Plot: HWF1D-parameter data trend exhibits a power-law behavior — essentially a straight line.b-parameter data trend is not constant (contrary to hyperbolic formulation).Computation of the b-parameter is severely affected by noise.



ide

—24

/63

Field Examples: Hybrid WF Gas Well (HWF1)

●Discussion: Hybrid Waterfrac Gas Well (HWF1)■ The hyperbolic rate relation (b=1) yields the highest reserve estimate.■ Reasonable matches of the rate data are obtained using the power-law

exponential model for both the D∞=0 and the D∞≠0 cases.■ The power-law exponential model applied using D∞≠0 provides the most

conservative estimate of reserves (as is expected) — this result is quite comparable to the model-based results obtained in SPE 114947.





ide

—25

/63

D. Ilk, Texas A&M UniversityJ.A. Rushing, Anadarko Petroleum Corp.T.A. Blasingame, Texas A&M UniversityDepartment of Petroleum Engineering



SPE 125031Decline Curve Analysis for HP/HT Gas Wells:

Theory and Applications



ide

—26

/63

Orientation: Rate-Time Relation

●Discussion: Rate-Time Gas Flow Relation (Knowles et al)■ Basis is the linearization of the nonlinear "μgcg" term (Ansah, et al).■ D-function and b-function are formulated using the definitions for loss-

ratio and the derivative of the loss-ratio.

From: Knowles R.S. 1999. Development and Verification of New Semi-Analytical Methods for the Analysis and Prediction of Gas Well Performance. M.S Thesis, Texas A&M University, College Station, Texas.Ansah, J., Knowles, R.S., and Blasingame, T.A. 2000. A Semi-Analytic (p/z) Rate-Time Relation for the Analysis andPrediction of Gas Well Performance. SPEREE. 3 (6): 525-533.

2

2

])exp[)1()1((]exp[4

DdwDwDwD

DdwDwDDd

tppptppq−−−+

−=

])exp[)1(1(])exp[)1(1(

DdwDwDwD

DdwDwDwDwDD tppp

tppppD++−++−

=

Dd

Dd

DdD dt

dqq

D 1−=

["Loss-Ratio"]

Rate-Time Relation:

Dimensionless D-function (DD):

b-function (b): ["Derivative of Loss-Ratio"]

⎥⎦

⎤⎢⎣

⎡−=

)/( DdDdDd

Dd dtdqq

dtdb

2

2

])exp[)1(1()1(]exp[2

DdwDwDwD

wDDdwD

tpppptpb

++−

−=



ide

—27

/63

Orientation: Rate-Cumulative Production Relation

●Discussion: Rate-Cumulative Gas Flow Relation■ The definition of the loss-ratio can be re-cast in terms of rate and cumula-

tive production.■ A quadratic relationship exists between rate and cumulative production.

From: Knowles R.S. 1999. Development and Verification of New Semi-Analytical Methods for the Analysis and Prediction of Gas Well Performance. M.S Thesis, Texas A&M University, College Station, Texas.Ansah, J., Knowles, R.S., and Blasingame, T.A. 2000. A Semi-Analytic (p/z) Rate-Time Relation for the Analysis andPrediction of Gas Well Performance. SPEREE. 3 (6): 525-533.

pD

DdD dG

dqD −=

["Loss-Ratio"]

Rate-Cumulative Production Relation:

Dimensionless D-function (DD):

b-function (b): ["Derivative of Loss-Ratio"]

22

1 pDpDDd GGq αα +−=

⎥⎥⎦

⎤

⎢⎢⎣

⎡−=

)/(1

pDDdpDDd dGdqdG

dqb

)1( pDD GD −= α

2

2

)1(2

22

−

+−=

pD

pDpD

G

GGb

α

αα

)1(/2 2wDp−=α



ide

—28

/63

Orientation: Analysis Methodology

●Discussion: Methodology■ The main goal is to match the data with the model using the definitions

for the q-D-b functions during the boundary-dominated flow regime.■ b-function → 0.5 for high drawdown cases (almost constant behavior).



ide

—29

/63

Field Example: HP/HT Tight Gas Well

●Field Example: Application of the Methodology■ 3.5 years of daily data are available for a hydraulically fractured well

completed in a HP/HT gas reservoir.■ Well clean-up effects, liquid-loading, and operational changes are

observed in the data trends. ■ The flowrate data are reviewed prior to analysis; and any erroneous/

redundant data points are removed.■ The half-slope trend is evident in the rate-integral derivative function.

pi = 14000 psia and TR = 260oF



ide

—30

/63


a. — qg versus Gp (Cartesian plot). b. — D-function versus t (Cartesian plot). c. — b-function versus t (Cartesian plot).

d. — qg versus t (Semilog plot). e. — D-function versus t (Semilog plot). f. — b-function versus t (Semilog plot).

●Field Example: Application of the Methodology■ For the computation of D- and b-parameter data functions we remove the

outlying data points; then we perform the numerical differentiation.■ Our analysis provides a gas-in-place estimate of approximately 8.0 BSCF.



ide

—31

/63


●Field Example: Application of the Methodology■ Reasonable matches of the D-function with the data using the semi-

analytical model is achieved (post-transient flow only).■ The matches of the b-function data with the semi-analytical model are

problematic — data indicate no unique characteristic behavior.■ Computation of the b-parameter data function is severely affected by

factors such as liquid loading.



ide

—32

/63


●Field Example: Application of the Methodology■ We observe a good match of the flowrate data with the model (except for

the early time data affected by "cleanup").■ The "power-law exponential" model yields Gp,max ≈ 8.0 BSCF.■ Gas-in-place estimates are consistent comparing the methods we used.



ide

—33

/63


●Field Example: Application of the Methodology■ An analytical reservoir model (vertical well with a finite conductivity

vertical fracture) is used to confirm the analysis results.■ Excellent match of the flowrate data and a reasonable match of the

bottomhole pressure history are obtained.■ All of the models yield consistent estimates of reserves/gas-in-place.



ide

—34

/63

Summary and Conclusions:●Summary:■We utilize a semi-analytical rate relation given by Knowles

[1999] and generalized by Ansah [2000] for the direct estimation of gas-in-place.

■The semi-analytical relation is formulated in terms of the Arps' D- and b-functions for diagnosis and analysis.

■ "Power-law exponential" rate decline relation is used to augment and validate our other analyses.

●Conclusions:■For high drawdown cases, the value of the Arps' b-parameter

should converge to approximately 0.5 during boundary-dominated flow regime.

■The straight line linearization scheme is validated — including application to cases of HP/HT gas reservoirs.

■The use of the D- and b-data functions provides a unique insight into flow regime identification and the simultaneous matching of the data. We conclude that this approach is a robust mechanism for estimating gas reserves.

■Model-based analysis (if applicable) should be performed to validate the results of the semi-analytical/empirical methods addressed in this work.



ide

—35

/63

S.M. Currie, Texas A&M UniversityD. Symmons, Consultant

D. Ilk, Texas A&M UniversityT.A. Blasingame, Texas A&M UniversityDepartment of Petroleum Engineering



Reserves Estimation in Tight Gas/Shale Gas Reservoirs using the Continuous EUR

Concept(manuscript in preparation for presentation at the

2010 SPE Unconventional Gas Conference)



ide

—36

/63

Continuous EUR: Simulated Case

●Discussion: Continuous EUR using Arps' Hyperbolic Decline■ Subsets of data are matched using the Arps' hyperbolic decline relation. ■ EUR of each subset is estimated (qgi is fixed).



ide

—37

/63


●Discussion: Continuous EUR using Power-Law Exponential Model■ EUR of each subset of data is estimated progressively.■ We do not fix any model parameter for the power-law exponential model.■ EUR stabilizes when boundary-dominated flow regime is established.



ide

—38

/63


●Discussion: Continuous EUR using Arps' Hyperbolic Decline■ Hyperbolic b-parameter value decreases with time.■ EUR estimate decreases over time and stabilizes at late times.■ EUR estimate using the power-law exponential model is more

conservative at early times.



ide

—39

/63

Continuous EUR: Field Case

●Discussion: Continuous EUR using Arps' Hyperbolic Decline■ Almost 2000 days of daily production data are available.■ b-parameter value ranges between 0.99 (all data) and 2.2 (earliest portion

— 50 days) for the subsets of data.



ide

—40

/63


●Discussion: Continuous EUR using Power-Law Exponential Model■ All of the matches for each part of the data set are almost outstanding.■ Additional data constrain the EUR estimate.■ Boundary-dominated flow regime effects are being established at late

times.



ide

—41

/63


●Discussion: Continuous EUR using Arps' Hyperbolic Decline■ Almost 50 percent decrease is observed in the EUR estimate when Arps'

hyperbolic relation is used.■ We advise caution when Arps' hyperbolic decline is used for estimating

reserves — in particular, at early times.



ide

—42

/63

C.M. Freeman, Texas A&M UniversityG.J. Moridis, Lawrence Berkeley National Lab.

D. Ilk, Texas A&M UniversityT.A. Blasingame, Texas A&M UniversityDepartment of Petroleum Engineering



SPE 124961A Numerical Study of Tight Gas

and Shale Gas Reservoir Systems



ide

—43

/63

●Discussion: van Kruysdijk and Dullaert [1989] Flow Regime Concept:■ Complex pressure profile behavior evolves due to fracture interference.■ Onset of compound linear period marked by rate decline.

a. Formation linear flow. b. Compound linear flow. c. Pseudo-elliptical flow.

Numerical Simulation Model: Concept



ide

—44

/63

Numerical Simulation Model: Concept

●Discussion: van Kruysdijk and Dullaert [1989] Flow Regime Concept:■ Complex pressure profile behavior evolves due to fracture interference.■ Onset of compound linear period marked by rate decline.



ide

—45

/63

Numerical Simulation Model: Modeling

●Discussion: Linear (Planar) Fracture — Repetitive Element■Repetitive element used to represent flow of individual fractures.■ Fracture interference modeled by a no-flow boundary.



ide

—46

/63

●Discussion: Effects of Fracture Spacing■ Rate and auxiliary functions merge during compound linear flow.■ Smaller fracture spacing causes earlier fracture interference effect.

Flowrate Behavior: Effects of Completion



ide

—47

/63

●Nonlinear Depletion:■ Sorption surfaces near

fracture are significantly more depleted.

■ Pressure/Density/SorptiveStorage is not intuitively related.

Flowrate Behavior: Effects of Desorption

x

y

Flowrates with Various Langmuir Volumes.Dimensionless Desorption Map

Dimensionless Density Map



ide

—48

/63

●Discussion: Effect of Matrix Permeability■ Matrix permeability affects the early time rate behavior in particular.■ Convergence of rate profiles are observed at late times.

Flowrate Behavior: Effects of Medium



ide

—49

/63

Conclusions and Remaining Issues:●Conclusions:■For TGSG reservoir systems, many factors affect performance.■Numerical modeling must be robust and tied to flow physics.■Desorption effects originate near fractures.■The well completion controls early time behavior.■Complex fractures (any type) substantially enhance early rates.

x

y

Flowrates with Various Langmuir Volumes.Dimensionless Desorption Map

Dimensionless Density Map



ide

—50

/63

●Issues for Well Performance in TGSG Systems:■Correct assessment of the fracture distribution must be made

to ensure proper flow regime identification and forecasting.■Desorption is highly nonlinear, but it defies direct assessment.■Assessment of fracture conductivity is non-unique.■Well completion issues are critical — particularly well clean-up.

Conclusions and Remaining Issues:



ide

—51

/63




Tight Gas/Shale Gas Well Performance Analysis:

Future View





ide

—52

/63

x

yPressure

MonitoringPoint No. 1

PressureMonitoringPoint No. 2

HydraulicFracture

Wellbore

X

X

x

y

x

yPressure

MonitoringPoint No. 1

PressureMonitoringPoint No. 2

HydraulicFracture

Wellbore

X

X

Numerical Model Considers:●Reservoir Layering.●kv/kh ratio.●Fracture Length, xf.●Fracture Conductivity, FcD.Analysis/Validation Approach:●Fit q(t) with Arps' hyperbolic relation.●Compare reserves to model at 30 years.

Decline Analysis: Tight Gas SystemsSPE 109625 (2007)Estimating Reserves in Tight Gas Sands at HP/HT Reservoir Conditions: Use and Misuse of an Arps Decline Curve MethodologyJ.A. Rushing, A.D. Perego, R.B. Sullivan, Anadarko Petroleum, and T.A. Blasingame, Texas A&M U.



ide

—53

/63

b. Elliptical flow type curve solution — high fracture conductivity case.

a. Elliptical flow type curve solution — low fracture conductivity case.

Vertical TG/SG Wells: Elliptical Flow Domination

c. Elliptical boundary configurations (finite con-ductivity fracture case [Amini, et al (2007)].

SPE 106308 (2007)Evaluation of the Elliptical Flow Period forHydraulically-Fractured Wells in Tight GasSands — Theoretical Aspects and PracticalConsiderationsS. Amini, D. Ilk, and T. A. Blasingame,SPE, Texas A&M U.



ide

—54

/63

a. Pressure profile at 0 year (0 hr).

b. Pressure profile at 1 year (8768 hr).

c. Pressure profile at 5.59 years (49,010 hr).

d. Pressure profile at 9.26 years (81,200 hr).

e. Pressure profile at 18.44 years (161,700 hr).

f. Pressure profile at 44.10 years (386,600 hr).

Results Generated Using:Ecrin Product Suite, Kappa Engineering, Sophia-Antipolis, France (2008).

Vertical TG/SG Wells: Elliptical Flow Domination



ide

—55

/63

a. Rate performance behavior for a horizontal well with 4 trans-verse fractures — infinite-acting reservoir (analog to van Kruysdijk and Dullaert work). Fine-scale numerical model .

Horizontal TG/SG Wells: Compound Linear Flow

c. Schematic diagram for the "compound linear flow" concept [van Kruysdijk and Dullaert (1989)].

Presented at the 2nd European Conference on the Mathematics of Oil Recovery, Cambridge, England (1989).A Boundary Element Solution of the Transient Pressure Response of Multiply Fractured Horizontal WellsC.P.J.W. van Kruysdijk and G.M. Dullaert, Shell

b. Specialized derivative plot (ref: van Kruysdijk and Dullaert) for a horizontal well with 4 transverse fractures — infinite- and finite-acting reservoir cases. Fine-scale numerical model .



ide

—56

/63

Base Simulation Modelfor Horizontal Well with Multiple Hydraulic Fractures

Top View HorizontalWell

MultipleVertical

Fractures

SPE 119897: Schematic Model for Simulation



ide

—57

/63

●Discussion: Horizontal Wells with Transverse Fractures■ Very high resolution simulation case.■ Very good D-parameter and good b-parameter computed from results.■ Excellent rate match using the new model (all regimes).

SPE 119897: q-D-b Plot — Numerical SimulationFinite-Acting

Reservoir Case



ide

—58

/63

●Discussion: Horizontal Wells with Transverse Fractures■ Very high resolution simulation case.■ Transient D- and b-parameters at late times (quasi-radial flow).■ Impossible to predict reserves.

SPE 119897: q-D-b Plot — Numerical SimulationInfinite-Acting

Reservoir Case



ide

—59

/63

Horizontal TG/SG Wells: Diagnostics?Q1.Compound Linear Flow Domination? (transient flow)A1. Possibilities for estimating reservoir properties:

a. Just give up — impossible to resolve anything. (default)b. "Lump" k, xf, and Lwell into a "parameter." ("mechanistic model")c. Develop testing practices to estimate properties. (maybe…)d. Other model concepts (e.g., propagating ellipse). (very tedious)

Q2.Estimating Reserves?A2. Issues:

a. Extremely long transition to boundary-dominated flow. (reality)b. Hyperbolic rate relation will overestimate reserves. (as always)c. Power-law/exponential rate relation? (more validation)

Q3.Role of simulation/modeling?A3. In the short-term, simulation/modeling is the primary

tool at our disposal — the analogy of using a hammer in place of a screwdriver comes to mind… (sometimes effective, but always a sub-optimal solution — particu-larly in the hands of children).



ide

—60

/63




BlasingameCurrent Projects/Interests





ide

—61

/63

Current Work: Blasingame●Projects: Status■ Simple Rate-Time Models for Shale Gas Systems (active)■ Diagnostics and Simplified Production Analysis (active)■ Analytical/Numerical Models for Shale Gas (active)■ IPR for Solution Gas-Drive Systems (nearing closure)

●Concepts: Status■ Production Mechanisms for Shale Gas Systems (start-up)■ Petrophysical Properties of Shales (2010)■ Continuous EUR (start-up)

●Focus Priority■ Simple Rate-Time Models for Shale Gas Systems very high■ Analytical/Numerical Models for Shale Gas very high■ Continuous EUR very high■ Production Analysis for Shale Gas Systems high■ Correlation/Quality Control for Data Analysis high■ Petrophysical Properties of Shales/Tight Gas Sands medium



ide

—62

/63

Students: Blasingame●Boulis: Extended Hyperbolic Models for DCA (M.S./complete)●Carballo: TBA (Gas Reservoir Eng.?) (Ph.D. coursework — DL)●Currie: Simple Rate-Time Models for Shale Gas Systems (M.S./active)●Freeman: Numerical Models for Shale Gas Well Performance (M.S./active)● Ilk: Production Analysis — Tight Gas Systems (Ph.D./active)● Jam: Analytical Models for Shale Gas Well Performance (M.S./active)● Johnson: Simplified Production Analysis for Gas Wells (B.S./complete)●Mendel: PTA/PA for Heavy Oil Production Systems (B.S./complete)●Nass: IPR For Solution Gas-Drive Reservoirs (M.S./active — DL)●Olsen: Reserves Practices (M.S./start-up — DL)



ide

—63

/63

Recent Work inWell Performance Analysis for

Tight Gas Sands and Gas ShalesEnd of Presentation




Presentation at Texas A&M U.17 November 2009 — College Station, TX

recent work in well performance analysis for tight gas ... · presentation at marathon oil marathon...

Documents