CSUG/SPE 137352

Tight Oil Production Analysis: Adaptation of Existing Rate-Transient Analysis

Techniques
C.R. Clarkson, P.K. Pedersen, University of Calgary
Copyright 2010, Society of Petroleum Engineers

This paper was prepared for presentation at the Canadian Unconventional Resources & International Petroleum Conference held in Calgary, Alberta, Canada, 1921 October 2010.

This paper was selected for presentation by a CSUG/SPE program committee following review of information contained in an abstract submitted by the author(s). Contents of the paper have not been
reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect any position of the Society of Petroleum Engineers, its officers, or
members. Electronic reproduction, distribution, or storage of any part of this paper without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is
restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.


Abstract
Low-permeability (tight) oil reservoirs have emerged as a significant source of oil supply in North America. Hydraulic fracturing
techniques combined with horizontal well technology have enabled commercial production from oil and gas reservoirs with
absolute permeability < 0.1 md, and recently in gas reservoirs with permeability in the nanodarcy range. Techniques for
quantitative analysis of ultra-low permeability reservoirs completed with hydraulically-fractured horizontal wells are in their
infancy; complications that have slowed development of these techniques include complex reservoir behavior, such as dual
porosity effects, multi-layer behavior, multi-phase flow and non-static absolute permeability, and complex flow behavior
associated simultaneous flow of multiple hydraulic fractures into horizontal wells. Although analytical models have been and
currently are being developed to simulate well performance associated with these complexities, systematic methods for analyzing
production for the purpose of quantifying reservoir and hydraulic fracture properties have not been discussed in detail in the
literature for tight oil reservoirs.
This paper examines the use of classic rate-transient techniques (flow-regime analysis, type-curve methods and simulation) for
analysis of tight oil reservoirs. Single-phase (oil) flow associated with undersaturated black oil reservoirs (above bubble point) is
the focus of this work. Simulated cases of multi-fractured horizontal wells completed in single porosity reservoirs, as well as
naturally-completed horizontal wells in transient dual porosity reservoirs are analyzed. In all cases, the proposed integrated rate-
transient analysis approach provided reasonable estimates of (simulator input) hydraulic fracture and reservoir properties. A field
case of a multi-fractured horizontal well completed in a low-permeability area of the Pembina Cardium Field, Western Canada, is
also given. Finally, a simplified method for forecasting horizontal wells completed in single and dual porosity media, using inputs
derived from flow-regime analysis, is introduced.

Introduction
With the success of the Bakken tight oil play in the U.S. (ex. see Cox et al. 2008), many operators are starting to explore for, or are
in the initial stages of development of, low-permeability oil plays that could not be economically developed using conventional
well/completion technology. The use of multi-fractured horizontal wells is key to development of some tight oil plays, allowing
sufficient productivity/recovery to make the plays economic. However, analyzing performance of these wells has proven to be
challenging not only because of potentially complex reservoir behavior, but also because of complex flow geometries that evolve
over time during production of the wells.
As with other unconventional reservoir types (ex. tight gas, shale gas and coalbed methane), advanced production analysis
methods may be used to extract quantitative information about the reservoir and the stimulation treatment. Complex reservoir
behavior (dual porosity/dual permeability, multi-layer, stress-dependent porosity and permeability, multi-phase flow etc.) for many
tight oil plays is anticipated, which complicates production analysis. For example, Breit et al. (1992) used an anisotropic, stress-
sensitive, dual porosity reservoir model to match welltest and interference data from the Bakken Shale. Further, the use of multi-
fractured horizontal wells is expected to create a complex sequence of flow regimes that will require strict interpretation in order to
glean quantitative information about either the hydraulic fracture stimulation or the reservoir. Based upon our experience with
2 CSUG/SPE 137352
tight/shale gas reservoirs, several conceptual models may be applied to the wellbore/fracture geometry and reservoir type (Fig. 1)
used to analyze tight oil reservoirs developed with horizontal wells:

a) Single Porosity Reservoir b) Dual Porosity Reservoir
Horizontal Well
Scenario 1 Scenario 2

Stimulated Reservoir Volume
d) c)
Scenario 3 Scenario 4

Discrete Hydraulic Fractures
f)
e)
Scenario 5 Scenario 6

g) h)
Fig. 1 Possible combinations of reservoir/hydraulic fracture encountered for tight oil reservoirs.
Scenario 7 Scenario 8

Fig. 1a represents a naturally-completed (openhole) horizontal well in a single porosity reservoir (Scenario 1). This completion
style is likely ineffective in ultra-low permeability reservoirs because of the lack of contacted surface area the pressure-transient
signature associated with this kind of a completion/reservoir has been discussed by Lee et al. (2003) and Medeiros et al. (2008).
Fig. 1b (Scenario 2) can represent a naturally-completed horizontal well in a dual porosity (naturally-fractured) reservoir or a
multi-(transverse) fractured horizontal well where fracture complexity (Stimulated Reservoir Volume or SRV) has been
created. Bello and Wattenbarger (2008) argued that horizontal shale gas wells with multiple hydraulic fracture stages may treated
as a linear dual porosity system, and Bello (2009) stated that the dominant flow regime in early years is transient linear flow from
the matrix to the fractures. In Scenario 3 (Fig. 1c), the background reservoir is single porosity, and the SRV is limited to a region
immediately around the horizontal well. Scenario 4 (Fig. 1d) corresponds to a case where an SRV is created from multi-fracturing
the horizontal well, but the background reservoir is also naturally-fractured (dual porosity) the SRV and background reservoir
would have different fracture spacing/permeability and fracture porosity. Scenarios 5-8 are the same as 1-4, but with discrete
hydraulic fractures which have a different conductivity than the fracture network. The pressure-transient signatures of several of
these cases were generated by Medeiros et al. (2008), assuming that the natural fracture network could be modeled with a pseudo
CSUG/SPE 137352 3
steady-state dual porosity representation, for tight gas reservoirs. They also looked at longitudinally-fractured wells. Scenario 7 is
similar to the conceptual model that Ozkan et al. (2009) used as the basis of their Trilinear Flow (pressure-transient) solution for
analyzing shale gas wells. To our knowledge, scenario 4 or 8 has not been discussed, but remains a possibility.
Because there are several possible combinations of wellbore/fracture geometry/reservoir types that could exist for tight oil
reservoirs, flow-regime identification and integration of surveillance techniques (see Clarkson and Beierle 2010) to assist with
model identification is critical for the proper extraction of reservoir/stimulation properties.
In this work, we apply advanced production analysis methods to tight oil reservoirs completed with multi-fractured horizontal
wells. We begin with a brief discussion of analysis methods and workflows used, followed by illustrations of this workflow using
simulated and field cases. The flexibility of our methods is demonstrated by the selection of different reservoir types/well
completion styles as examples. We also introduce a simplified method for rate-transient forecasting of multi-fractured horizontal
wells completed in single and dual porosity media using inputs derived from flow-regime analysis.

Analysis Methods and Workflow
The analytical techniques used in this work include modern production (rate-transient) analysis methods that can be categorized as:
Flow-regime analysis (straight-line methods) specific flow regimes are identified using rate-transient (semilog and
linear) derivative methods, followed by analysis of the data corresponding to specific-flow regimes using specialty
(usually Cartesian) plots. Transient (linear, bilinear, elliptical and radial) and boundary-dominated flow may be analyzed
using these techniques. Reservoir and effective wellbore/stimulation properties may be extracted from the fit of straight-
lines to the dataset. Derivative analysis techniques may also be used, provided derivatives are clean enough to analyze.
Type-curve analysis production data, cast in dimensionless form, are matched to dimensionless type-curves developed
from analytical or empirical solutions to flow equations corresponding to a specific reservoir model. Reservoir and
stimulation properties may be extracted from the match from the definition of the dimensionless variables.
Reservoir simulation dynamic (production data and/or flowing pressures and reservoir pressure) data are matched
through adjustment of model input parameters in an effort to calibrate the simulation model, which in turn is used to
forecast fluid production.
The techniques are well known and not discussed in detail in this work; example applications of these methods to tight gas
reservoir production are discussed in Clarkson and Beierle (2010). An important step in the use of these techniques is the selection
of the appropriate reservoir model such that the extracted reservoir/stimulation properties are meaningful. Clarkson and Beierle
(2010) discuss the integration of surveillance data (microseismic, production and tracer logs) with production analysis to improve
model selection and calibration. Their proposed rate-transient analysis workflow, adopted here for tight oil reservoirs and specific
to multi-fractured horizontal wells, is briefly summarized below (more detail is provided in Clarkson and Beierle 2010):

1. Starting with commingled-stage production data, identify flow regimes using log-log diagnostics such as the semilog
derivative (as well as linear derivative). This analysis is combined with microseismic and other surveillance data to assist
with the selection of the appropriate model which best matches the flow regime sequence
2. Analyze the flow-regime data using rate-transient straight-line techniques to obtain preliminary estimates of hydraulic
fracture/effective wellbore length or reservoir properties
3. Match the production data on an appropriate type-curve, using the straight-line derived values as a starting point for
matching
4. If surveillance data was collected to assist with assignment of fracture stage contributions, preferably over time, use
analytical flow model to simultaneously match individual stage contributions and total commingled production data
5. Use rate-transient-derived properties (hydraulic fracture properties, kh etc.) to populate a reservoir simulator which in turn
is used to history-match the well dynamic data and generate a production forecast

Additional steps could include:
Develop a conceptual simulation model (prior to analyzing actual well production using rate-transient techniques)
that incorporates all reservoir/stimulation/surveillance data compiled to date. The conceptual simulation model is used to
generate production forecasts which in turn are analyzed with rate-transient techniques to establish the flow-
regime sequence and duration. This exercise will assist with interpretation of the real data.
Analyze vertical wells completed in the same zone as the horizontal wells generally, vertical wells are easier to analyze
and may provide additional constraints for horizontal wells analysis.
.
Example Applications
We now illustrate our analysis approach using simulated and field examples. Table 1 summarizes the cases analyzed.
4 CSUG/SPE 137352
TABLE 1 SUMMARY OF CASES ANALYZED
Case Simulated or Field Well Type Reservoir Type Fracture
Conductivity
Number of
Fracs
Fig. 1 Scenario
1 Simulated Horizontal well,
multi-fractured
Single porosity Infinite

5

5
2 Simulated Horizontal well,
multi-fractured
Single porosity Finite

5

5
3 Simulated Horizontal Transient, Dual
Porosity
N/A N/A 2
4 Field Horizontal well,
multi-fractured
Single porosity ? 10 5

For analytical simulation, we make extensive use of the linear flow (closed boundaries) solution of Wattenbarger et al. (1998),
assuming constant flowing pressure, originally developed for hydraulically-fractured vertical wells (see Fig. 2 for model sketch).
For completeness, the solution is given below:

=
odd
Dxf
e
f
f
e
t
y
x
n
x
y
D
q
2
2 2
4
exp
4
1

..................................................................................................................................... (1)


Where:

( )
qB
p p kh
D
q
wf i
2 . 141
1

=
............................................................................................................................................................ (2)

2
00633 . 0
f t
Dx
x c
kt
t
f

=
................................................................................................................................................................ (3)
xe
ye
xf

Fig. 2 Conceptual model for linear flow solution (Eq. 1-3).

Case 1: simulated multi-fractured (infinite-conductivity) horizontal well completed in single porosity tight oil reservoir. In this
example, a multi-(transverse) fractured horizontal well was numerically-simulated using KAPPAs Topaze software (Table 2).
In this example, the number of hydraulic fractures (5) equals the number of stages; there is also no flow directly to the horizontal
wellbore. The fractures are assumed to be infinite conductivity. Fluid properties are assumed to be constant with pressure, for
simplicity. This example corresponds to Scenario 5 of Fig. 1.
The sequence of flow regimes encountered for the well are those described by Chen and Raghavan (1997), as illustrated from
the derivative plots in Fig. 3. Early formation linear flow to the individual fractures is followed by early radial flow around the
CSUG/SPE 137352 5
fractures, then by fracture interference, late- (compound) linear flow to the well, late radial flow (to the well), and early
development of boundary-dominated flow.
Following the procedure of Clarkson and Beierle (2010) for tight gas wells, the flow-regimes are first analyzed to obtain
preliminary estimates of hydraulic-fracture and reservoir properties (Fig. 4). The late radial flow period (Fig. 4b) yields an
estimate of kh, and knowing h (same as simulator input), k is estimated to be 0.01 md (same as model input) and skin to be 7.65.
If it can be assumed that the wellbore (after frac interference) behaves like an infinite-conductivity fracture, then the estimate of L
w
from skin is 2520 ft, which is 520 ft greater than the actual wellbore length. From the early linear flow period (Fig 4a), x
ftotal
*(k)
1/2

can be estimated, where x
ftotal
is the total fracture half length (sum of the individual hydraulic fractures) assuming k is uniform
along the well (and equal to the value obtained from late radial flow); in this example, x
ftotal
is calculated to be 1112 ft (or 222
ft/frac), which is close to the value input into the simulator. Since boundary-dominated flow is approached at late time, flowing
material balance (Fig 4c) may be used to estimate OOIP and drainage area (given volumetric input); the estimated drainage area is
317 acres, which is close the model input value of 320 acres. The early radial flow-regime may also be analyzed to provide a kh
which is some multiple of the actual kh (Gilbert and Barree 2009) and late-linear (compound linear) may be analyzed to estimate
effective wellbore length (not shown). The early radial period is difficult to select in this example as no true zero slope is evident
from the derivative plot and late-linear flow yields an over estimate of effective wellbore length, as we noted in our previous work
(Clarkson and Beierle 2010).
The parameters obtained from straight-line analysis are then input into the dimensionless rate and time variables defined
above (Eq. 2 and 3), noting that we have used x
ftotal
for x
f
in Eq. 3, as well as model output production, and then superimpose the
transformed data onto previously-created type-curves for multi-fractured horizontal wells (Fig 4d) the match in this case is
perfect because a type-curve set for a 2000 ft lateral with five 225ft fractures was previously created. If for example we incorrectly
calculated one of the dimensionless variable inputs from straight-line analysis, the type-curve match would be incorrect; we
demonstrate this by assuming late-linear flow is early-linear flow and incorrectly calculating x
ftotal
(Fig. 5) in this case the
transformed data do not match the type-curves.

TABLE 2 INPUT PARAMETERS FOR SIMULATED
UNDERSATURATED OIL CASE (MULTI-FRAC
HORIZONTAL WELL) - CASE 1
Input Parameter Parameter
Value
THICKNESS (FT)
45
POROSITY (%)
6
ABSOLUTE PERMEABILITY (MD)
0.01
INITIAL RESERVOIR PRESSURE
(PSIA)
4900
INITIAL OIL SATURATION (%)
100
INITIAL WATER SATURATION (%)
0
RESERVOIR TEMPERATURE (F)
255
INITIAL OIL FVF (RB/STB)
1.369
INITIAL OIL VISCOSITY (CP)
0.419
TOTAL COMPRESSIBILITY (PSI
-1
)
1.76 X 10
-5
DRAINAGE AREA (ACRES) 320
WELL LENGTH (FT) 2000
WELLBORE DIAMETER (IN.) 7.2
TOTAL FRACTURE HALF-LENGTH
(FT)
1125
# HYDRAULIC FRACTURES 5
INDIVIDUAL HYDRAULIC
FRACTURE HALF-LENGTH (FT)
225
HYDRAULIC FRACTURE SPACING 500
FLOWING BOTTOMHOLE
PRESSURE (PSIA)
1000


-3000 -2500 -2000 -1500 -1000 -500 0 500 1000 1500 2000 2500 3000
Length [ft]
-2000
-1000
0
1000
2000
L
e
n
g
th
[ft]
Ref erence well
Length [ft] vs Length [ft]
6 CSUG/SPE 137352

1000
10000
100000
1000000
0.01 0.1 1 10 100 1000 10000 100000
d
(

p
/
q
)
/
d
l
n
t
Material Balance Time, days
Radial Derivative Plot

1000
10000
1 10 100 1000 10000
d
(

p
/
q
)
/
d
t
^
0
.
5
Material Balance Time, days
Linear Derivative Plot

Fig. 3 Use of derivative plots to identify flow regimes: (a) semilog derivative; (b) linear derivative. The dominant flow regimes are
shown on the semilog derivative.
0
10
20
30
40
50
0 10 20 30 40 50
[
p
i
-
p
w
f
]
/
q
o
Superposition Time
Early Linear Flow Plot

0
100
200
300
4 5 6
[
p
i
-
p
w
f
]
/
q
o
Superposition Time
Late Radial Flow Plot

a) b)
B-D flow
Late Linear
flow
Late Radial
flow
Early Linear
flow
Late Linear
flow
Early Radial
flow
Early Linear
flow
a) b)
0
0.01
0.02
0.03
0.04
0.05
0 500000 1000000 1500000 2000000 2500000 3000000
N
o
r
m
a
l
i
z
e
d

R
a
t
e
Normalized Cumulative Production, STB
Flowing Material Balance

0.01
0.1
1
10
100
0.00001 0.0001 0.001 0.01 0.1 1 10
q
D

o
r

d
(
1
/
q
D
)
/
d
l
n
(
t
D
)
tD
Horizontal Well with Transverse Fractures
derivative type-curve qD type-curve derivative data qD data

c) d)
Fig. 4 Straight-line and type-curve match for Case 2: (a) early linear flow straight-line plot; (b) late radial flow straight-line plot;
(c) flowing material balance and (d) type-curve match.

CSUG/SPE 137352 7

0.01
0.1
1
10
100
0.00001 0.0001 0.001 0.01 0.1 1 10
q
D

o
r

d
(
1
/
q
D
)
/
d
l
n
(
t
D
)
tD
Horizontal Well with Transverse Fractures
derivative type-curve qD type-curve derivative data qD data
1
10
0.00001 0.0001 0.001 0.01 0.1 1 10
d
(
1
/
q
D
)
/
d
(
t
D
)
^
0
.
5
tD
Horizontal Well with Transverse Fractures
derivative type-curve derivative data

a) b)
Fig. 5 Demonstration of type-curve mis-match by assuming late linear flow is early-linear flow and using the incorrect x
ftotal
estimate in dimensionless time definition. (a) q
D
and semilog derivative type-curve match; (b) linear derivative type-curve
match.

The last step in the analysis is history-matching the production data (with rate-transient analysis-derived parameters as a
starting point) using an analytical and/or numerical model. In this example, we have used a new analytical approach to history-
match/forecast well production. After the flow-regimes have been identified and analyzed, the derived parameters are input into
flow equations corresponding to the identified flow regimes. In this example we use a two-linear (closed boundary) flow model,
corresponding to early and late linear flow, with a transition from early to late linear flow being represented by a radial flow
period. We note that use of radial flow to represent the transition from one linear flow period to the other is an approximation in
reality, elliptical flow will precede radial flow. The start of the late linear flow period was estimated by using the time to reach
boundary-dominated flow (assuming linear flow) for the flow around the inner fractures, where the drainage area around the
fractures is estimated from geometry. The linear flow solution for both early and late-linear flow/boundary-dominated flow was
calculated with Eq. 1-3. The transient radial flow solution is well known. The start of the radial flow period corresponds to the
deviation of the actual well data from the first linear flow period (see Fig. 3a). We note that this successive flow regime approach
for forecasting is similar to that suggested by Palmer and Moschovidis (2010) with the notable difference that we first identify and
analyze the flow regimes for each case before selecting a flow model Palmer and Moschovidis (2010) assumed that a horizontal
well composite model was applicable for all shale gas cases, which they admit is an approximation.
The derived history-match using this approach is given in Fig. 6. A slight discrepancy is seen (Fig. 6a) between 1 and 10 days
due to the approximation of elliptical flow with radial flow, but otherwise the match is very good. Also note that late radial flow
was not modeled as the match assumed linear flow only (after 110 days) if necessary, this flow regime may also be represented.

1
10
100
1000
10000
0.1 1 10 100 1000 10000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-New
1
10
100
1000
0 200 400 600 800 1000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-New

a) b)
Start of Late
Linear flow
Start of
Early Radial
flow
Fig. 6 History-match of simulated multi-fractured horizontal well using new analytical flow model (a) log-log plot and (b) semilog
plot. Note the transition from early linear flow (ending at about 5 days) to late linear flow (starting after 110 days) is
represented by radial flow. Also note the transition to boundary-dominated flow at late time.


8 CSUG/SPE 137352
For comparison, we also show the match assuming linear flow only (Fig. 7), where only the early linear flow data was fit with
the model. The model fit is only good for the first 10 days, after which the development of transitional flow regimes (early
elliptical and radial flow) and late linear flow cause a model mis-match.
1
10
100
1000
10000
0.1 1 10 100 1000 10000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-Linear

1
10
100
1000
0 200 400 600 800 1000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-Linear

a) b)
Fig. 7 History-match of simulated multi-fractured horizontal well using transient linear flow model (a) log-log plot and (b) semilog
plot. The early linear flow period was matched and then the transient linear flow model was used to forecast. Note the
large deviation of the actual data from the model after 10 days, due the presence of transitional flow regimes and late
linear flow.

This example illustrates the use of our workflow for analyzing multi-fractured horizontal wells completed in single porosity
media; we also introduce a new method for history-matching the wells and forecasting production using parameters derived from
flow regime analysis. The case of finite conductivity fractures will now be discussed.

Case 2: simulated multi-fractured (finite-conductivity) horizontal well completed in single porosity tight oil reservoir. This case
differs from the previous case only by the assignment of finite conductivity (10 md*ft) to the individual fractures; all other
reservoir, well and fracture inputs are the same.
The sequence of flow-regimes encountered for the well is shown in Fig. 8. Early bilinear flow associated with the fractures is
followed by possibly formation linear flow then early radial flow around the fractures, followed by fracture interference, late-
(compound) linear flow to the well, late radial flow (to the well), and early development of boundary-dominated flow. Early linear
flow is not well-developed in this case as evidenced by the absence of an early zero slope on the linear derivative plot (Fig 8b).
1000
10000
100000
1000000
0.01 0.1 1 10 100 1000 10000 100000
d
(

p
/
q
)
/
d
l
n
t
Material Balance Time, days
Radial Derivative Plot

1000
10000
1 10 100 1000 10000
d
(

p
/
q
)
/
d
t
^
0
.
5
Material Balance Time, days
Linear Derivative Plot

b)
B-D flow
Late Linear
flow
Late Radial
flow
Late Linear
flow
Early Radial
flow
Bilinear flow
a)

Fig. 8 Use of derivative plots to identify flow regimes: (a) semilog derivative; (b) linear derivative. The dominant flow regimes are
shown on the semilog derivative. Note the presence of bilinear flow; also note that early linear flow is not well-developed
in this case (finite conductivity).
Although we analyzed all the flow regimes analogously to the previous cases, we only show bilinear flow analysis in Fig. 9
along with the type-curve match. A fracture-conductivity of 9.4 md*ft was obtained from the straight-line analysis, which is close
CSUG/SPE 137352 9
to the input value in the simulator (10 md*ft). As in the previous case, the type-curve match is good because a type-curve set was
previously generated for the case of a 2000 ft lateral with 5 finite-conductivity (10 md*ft) 225ft fractures.
0
50
100
150
200
0 5 10 15 20
[
p
i
-
p
w
f
]
/
q
o
Superposition Time
Bilinear Flow Plot

0.01
0.1
1
10
100
0.00001 0.0001 0.001 0.01 0.1 1 10
q
D

o
r

d
(
1
/
q
D
)
/
d
l
n
(
t
D
)
tD
Horizontal Well with Transverse Fractures
derivative type-curve qD type-curve derivative data qD data

a) b)
Fig. 9 Bilinear flow (a) and type-curve match (b) for Case 3.

To history-match and forecast well production, we again use the approach described for case 2; a 2-linear (closed boundary)
flow model with a transition from early to late-linear flow, approximated with radial flow, was again used, but in this case finite
conductivity of the hydraulic fractures causes a significant flattening of the production profile in log-log coordinates (compare Fig
9b with Fig. 4d). In order to model the effect of finite conductivity, we approximate the effect using the skin approach suggested
by Bello and Wattenbarger (2009); a skin term was thus added to Eq. 1. All parameters derived from straight-line analysis (ex.
total hydraulic fracture half-length) were used to history-match the well (Fig. 10); the skin was used as a history-match parameter,
although theoretically it could be derived from the straight-line analysis.

1
10
100
1000
1 10 100 1000 10000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-New
1
10
100
1000
0 200 400 600 800 1000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-New

a) b)
Fig. 10 History-match of simulated multi-fractured horizontal well using new analytical flow model (a) log-log plot and (b) semilog
plot. Skin was adjusted to match early-time (bilinear) flow using the approach of Bello and Wattenbarger (2009).

Cases 1 and 2 address a hydraulically-fractured horizontal well in a single porosity reservoir. We now analyze a case in which
the well completion/reservoir behaves like a transient, dual porosity system.

Case 3: simulated horizontal well completed in transient dual porosity tight oil reservoir. According to Bello and Wattenbarger
(2009), multi-fractured horizontal wells completed in shale gas reservoir may be modeled using a transient dual porosity system.
Five flow regions were identified by Bello (2009) using a linear dual porosity model, consisting of matrix blocks and fractures.
The dominant transient linear flow regime associated with shale gas reservoirs is interpreted to be associated with linear flow from
the matrix to the fractures (Region 4, using Bellos terminology).
We have assumed that multi-fractured horizontal wells, completed in tight oil (particularly shale oil) reservoirs could behave
in a similar way, depending upon the details of the stimulation treatments (frac spacing, job size etc.) and reservoir properties
(mechanical properties, existence of natural fractures, in-situ stress environment etc.). In order to simulate such a case
10 CSUG/SPE 137352
(corresponding to Scenario 2 in Fig. 1), we used the transient dual porosity model in PROMAT to generate a production forecast
for a horizontal well completed in a circular reservoir. The model inputs are given in Table 3. Only single-phase flow (above
bubble point) conditions were simulated and PVT properties of the oil were assumed constant (equal to initial).


TABLE 3 INPUT PARAMETERS FOR SIMULATED
UNDERSATURATED OIL CASE (OPENHOLE HORIZONTAL
WELL COMPLETED IN TRANSIENT DUAL POROSITY
RESERVOIR) - CASE 3
Input Parameter Parameter Value
THICKNESS (FT)
25
MATRIX POROSITY (%)
4.5
FRACTURE (VERTICAL AND
HORIZONTAL) ABSOLUTE
PERMEABILITY (MD)
0.5
OMEGA (FRACTION)
0.1/0.5
LAMBDA (FRACTION)
1.0 X 10
-9

INITIAL RESERVOIR PRESSURE
(PSIA)
2958
INITIAL OIL SATURATION (%)
75
INITIAL WATER SATURATION (%)
25
RESERVOIR TEMPERATURE (F)
175
INITIAL OIL FVF (RB/STB)
1.326
INITIAL OIL VISCOSITY (CP)
0.506
TOTAL COMPRESSIBILITY (PSI
-1
)
1.76 X 10
-5
DRAINAGE AREA (ACRES) 40
WELL LENGTH (FT) 1000
WELLBORE DIAMETER (IN.) 6.25
FLOWING BOTTOMHOLE
PRESSURE (PSIA)
1000
Lw
re














Two different omega values were used; derivative analysis for the case of = 0.1 is given in Fig. 11.
1000
10000
100000
1000000
10000000
1 10 100 1000 10000
d
(

p
/
q
)
/
d
l
n
t
Material Balance Time, days
Radial Derivative Plot

1000
10000
100000
1 10 100 1000 10000
d
(

p
/
q
)
/
d
t
^
0
.
5
Material Balance Time, days
Linear Derivative Plot

Late Linear
flow
a) b)
Late Linear
flow
B-D flow?

Fig. 11 Use of derivative plots to identify flow regimes: (a) semilog derivative; (b) linear derivative. The dominant flow regimes are
shown on the semilog derivative.

In this example, early transient flow through the fractures is not evident flow through the fracture pore volume reaches
boundaries and starts to deplete before substantial production from the matrix occurs. The late linear flow period represents
transient linear flow from the matrix to the fractures (Bellos Region 4). Analysis of the two dominant flow regimes using
CSUG/SPE 137352 11
straight-line analysis is given in Fig. 12. Application of flowing material balance to the apparent early boundary-dominated flow
period yields an oil-in-place estimate associated with the fracture pore volume using the input fracture porosity, the calculated
drainage area is 43 acres, which is close to the input value of 40 acres. Late linear flow analysis (Fig 12b) can be used to obtain a
contacted matrix surface area A
cm
, given matrix permeability, using procedures described in Bello and Wattenbarger (2008), which
is estimated to be ~ 203,000 ft
2
in this example (assuming matrix permeability = 0.0001 md, because matrix permeability cannot
be determined from the analysis). The fracture spacing can be calculated, using the calculated A
cm
and drainage volume from
flowing material balance analysis, combined with the assumption of a matrix geometry; assuming a slab geometry for the matrix,
the fracture spacing is calculated to be 458 ft.
0.00
0.05
0.10
0.15
0.20
0 10000 20000 30000 40000 50000 60000
N
o
r
m
a
l
i
z
e
d

R
a
t
e
Normalized Cumulative Production, STB
Flowing Material Balance

0
500
1000
1500
2000
0 100 200 300
[
p
i
-
p
w
f
]
/
q
o
Superposition Time
Late Linear Flow Plot

a) b)

Fig. 12 Straight-line analysis of transient dual porosity case (Case 4): (a) FMB analysis of early boundary-dominated flow data; (b)
late linear (matrix-fracture) analysis. = 0.1.
At the present time, we have not generated type-curves for the transient dual porosity case, so no type-curve match was
attempted, but we note that such type-curves do exist in the literature (ex. see Bello and Wattenbarger 2010 for transient linear
dual porosity type-curves). Type-curve matching can also be used to establish estimates of and .
For history-matching/forecast generation, we again use the analytical approach described above early-linear flow (analysis
not shown), apparent early boundary-flow and late linear flow analysis was performed to obtain apparent well length, drainage
area and matrix contacted surface area, respectively, and these values were used as input to the analytical forecasting tool. To
model this case, we used the 2-linear (closed boundary) flow model (Eq. 1-3), where one linear flow segment was used to model
early fracture linear flow/boundaries and the other to model late matrix linear flow/boundaries. The two linear flow solutions were
superposed (added together) because some matrix contribution occurred before the fracture system was depleted. The results of
the matches for = 0.1 and = 0.5 and are given in Fig. 13 and Fig. 14, respectively. The match in both cases is quite good. For
the case of = 0.5, the fracture linear flow period is more evident, due to the larger fracture pore volume.
0.1
1
10
100
1000
1 10 100 1000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-New

0.1
1
10
100
1000
0 200 400 600 800 1000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-New

a) b)
Fig. 13 History-match of simulated transient dual porosity case using new analytical flow model (a) log-log plot and (b) semilog
plot. = 0.1.
12 CSUG/SPE 137352
0.1
1
10
100
1000
1 10 100 1000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-New

0.1
1
10
100
1000
0 200 400 600 800 1000
O
i
l
R
a
t
e
,
S
T
B
/
D

Time, days
Production Rates
Actual Model-New
a) b)

Fig. 14 History-match of simulated transient dual porosity case using new analytical flow model (a) log-log plot and (b) semilog
plot. = 0.5.
As shown in Bello and Wattenbarger (2010), different combinations of and can lead to drastically different production profiles
in our example, because is very small, bilinear flow is not evident. We continue to test our analytical approach for a wide
range of and .

Case 4: field case, multi-fractured horizontal well completed in single porosity tight oil reservoir. This last example is a field case
of a multi-fractured (transverse) horizontal well (primarily cased hole) completed in the Pembina Cardium oil pool of Western
Canada. The Pembina Cardium field is the largest conventional oil pool in Canada covering an area of over 3000km
2
with over
6100 wells (currently approximately 4400 producers and 1700 injection wells, mainly vertical wells). The field has been developed
since the 1950s; waterflooding is the main method of recovery, although recently operators have also tried CO
2
-EOR (Lawton et
al. 2009). OOIP has been estimated at over 7780,000 Mbbl of oil, with less than 1268,200 Mbbl recovered to date. The field
occurs in a stratigraphic trap of northwest-southeast oriented shoreface sands with the eastern updip margin being defined by shale
out of the sands and the western downdip margin by decreasing reservoir quality (Krause et al. 1987). Horizontal wells have been
drilled in both the Cardium sands and conglomerates within the Pembina field with limited success (Adegbesan et al. 1996).
Recently, operators have successfully exploited the lower permeability Cardium sandstones and muddy sandstones outside the
Pembina field using multi-fractured horizontal well technology. An example well from this outside the area of oil production and
water injection is the subject of this case. The analyzed horizontal well occur in an area where the Cardium consists of 5-8meter of
muddy fine-grained sandstones characterized by relative low permeabilities (average 0.3 md and 11% porosity).
The subject well was fracd in 10 stages with an average fracture spacing of 429 ft. At the time of writing, the well had only
been on production for approximately 100 days (not including post-fracture cleanup period) and is currently on pump. Flowing
bottomhole pressure was estimated from periodic fluid shots combined with casing pressure the well took approximately 21 days
to reach a pumped-off state as inferred from a stabilized fluid level in the casing. Well production data (not corrected for back
pressure estimates) along with estimated flowing pressure is given in Fig. 15. After pump-off, GOR appears to be fairly stable so
we have assumed undersaturated oil above bubble point and ignored relative permeability effects. The estimated flowing pressure
(Fig. 15) appears to be below bubble point, so a constant GOR was not expected. Analysis input parameters are given in Table 4.
Filtering out the production data while the well was being pumped off (before 21 days), we attempted to identify the dominant
early flow regime, which appears to be linear flow as evidenced by a 1/2 slope on the log-log diagnostic plot (Fig. 16). The radial
derivative (not shown), although noisier, also indicated a slope. This early linear flow period is interpreted to be formation
linear flow (consistent with Scenario 5 of Fig. 1) due to the relatively high permeability of the matrix (~ .3 md).
Analyzing the linear flow period (Fig 16b), we obtain a total fracture half-length of 1809 ft, or an average of 181 ft per stage.
The fracture lengths appear reasonably consistent with the small job sizes pumped at low-rates.
Although very little data is available for a confident type-curve match, we see that using the total fracture half-length from
straight-line analysis results in a reasonable type-curve match (Fig. 16c) using the Wattenbarger linear flow type-curve (1998).
Although a forecast could be attempted, we have not done so here due to lack of historical data.

CSUG/SPE 137352 13

0
500
1000
1500
2000
2500
3000
0
50
100
150
200
250
300
0 20 40 60 80 100 120
G
O
R

(
s
c
f
/
S
T
B
)

o
r

P
w
f

(
p
s
i
a
)
G
a
s

(
M
s
c
f
/
D
)

o
r

O
i
l

(
S
T
B
/
D
)

Time, days
Oil Rate Gas Rate GOR Pwf


Fig. 15 Production and flowing pressure data for field case (4).


*estimated from air permeability, corrected to in-situ using relationship provided in Adegbesan et al. (1996)
**heal to toe stage length
1
10
100
1 10 100 1000
d
p
/
q
o

(
p
s
i
/
S
T
B
/
D
)
Time , days
0
10
20
30
0 20 40 60 80
[
p
i
-
p
w
f
]
/
q
o
Superposition Time
Early Linear Flow Plot

TABLE 4 INPUT PARAMETERS FOR FIELD CASE
(MULTI-FRAC HORIZONTAL WELL) - CASE 4
Input Parameter Parameter
Value
THICKNESS (FT)
16.4
POROSITY (%)
12
ABSOLUTE PERMEABILITY (MD)*
0.28
INITIAL RESERVOIR PRESSURE
(PSIA)
2017
INITIAL OIL SATURATION (%)
86
INITIAL WATER SATURATION (%)
14
RESERVOIR TEMPERATURE (F)
115
INITIAL OIL FVF (RB/STB)
1.19
INITIAL OIL VISCOSITY (CP)
1.414
TOTAL COMPRESSIBILITY (PSI
-1
)
1.31 X 10
-5
WELL LENGTH (FT)** 3865
WELLBORE DIAMETER (IN.) 7.2
# HYDRAULIC FRACTURES 10
a) b)

0.01
0.1
1
10
100
0.0001 0.001 0.01 0.1 1 10 100
(
y
e
/
x
e
)

q
D
tDye
Linear Flow Type-Curves

c)
Early Linear
flow?

Fig. 16 Log-Log diagnostic plot (a), linear flow analysis (b) and type-curve match (c) for field case (4). Data during pump-off is
filtered out.
14 CSUG/SPE 137352
Discussion and Future Work
A methodology for analyzing low-permeability undersaturated (above bubble point) oil reservoirs has been provided which
incorporates flow-regime (straight-line) analysis, type-curve matching and analytical model history-matching and forward
modeling. We present a simple methodology for analytical modeling that ties the properties obtained from straight-line analysis to
flow equations. We have tested the methodology against simulated cases, including naturally-completed horizontal wells
producing from dual porosity reservoirs and multi-fractured (infinite and finite conductivity transverse fractures) horizontal wells
producing from single porosity reservoirs. Medeiros et al. (2010) recently introduced a method based upon the transient
productivity index to analyze shale oil reservoirs; we believe our methodology could compliment the analytical methods that they
introduced.
The focus of the current study was on undersaturated oils producing above bubble point. Although we have found that this is
a reasonable assumption for several tight oil plays, particularly when flowing pressures are held above bubble point during the
early flow period, we realize that there will also be cases where the well is flowing below bubble point and multi-phase flow will
become important. Future work will include adaptation of the current methodology for multi-phase flow scenarios, starting with
the framework provided by Camacho and Raghavan (1989) using pseudopressure and time transformations.
Other reservoir complexities have been ignored in this work, including stress-dependent permeability, which is likely to be
strong, particularly for naturally-fractured reservoirs. We will continue to refine our techniques to include additional reservoir
complexities in future work. Further, additional scenarios provided in Fig. 1 will be analyzed.
Finally, we will update the field case presented in this work, which has limited production history, to establish if the
methodology can be used to glean additional information from the well. Additional field cases, including other tight oil plays in
Western Canada, will also be examined.

Conclusions
In this work we have successfully adapted rate-transient techniques that have recently been refined to analyze tight gas and shale
gas reservoirs, to tight oil reservoirs. From this work the following conclusions may be drawn:

1) Multi-fractured horizontal wells completed in tight oil reservoirs may exhibit a wide range in rate-transient signatures
depending on the combination of reservoir properties and induced hydraulic-fracture geometries a model should not be
selected until the flow-regimes have been analyzed
2) A combination of rate-transient analysis techniques (straight-line, type-curve, simulation) is required to arrive at a unique
solution
3) A new simple analytical approach for history-matching/forecasting tight oil reservoirs was introduced and appears to
work well for a wide range in reservoir/hydraulic fracture properties

Acknowledgements
Chris Clarkson would like to acknowledge EnCana for support of his Chair position in Unconventional Gas at the University of
Calgary, Department of Geosciences, and the donations of software by both Fekete and KAPPA to his research program. Funding
for this work was provided in part by a Natural Sciences and Engineering Research Council of Canada Discovery Grant to
Clarkson. The authors would like to acknowledge ARC Resources for contribution of the Cardium well production data.

Nomenclature
Field Variables
B = Oil formation volume factor, RB/STB
c
t
= Total compressibility, psi
1

h = Formation thickness, ft
k = Formation permeability, md
L
w
= Horizontal well length, ft
p = Pressure, psia
p
i
= Initial reservoir pressure, psia
p
wf
= Flowing bottomhole pressure, psia
q = Oil flow rate, STB/D
r
e
= Drainage radius, ft
t = Time, days
x
e
= Reservoir half-width (Fig. 2), ft

x
f
= Fracture half length, ft
CSUG/SPE 137352 15
y
e
= Distance from fracture to outer boundary (Fig. 2), ft
Dimensionless Variables
q
D
= Dimensionless rate
t
D
= Dimensionless time
t
Dxf
= Dimensionless time based upon x
f

t
Dye
= Dimensionless time based upon y
e


Greek Variables
= Oil viscosity, cp
= Porosity, fraction
= Transmissivity ratio, fraction
= Storativity ratio, fraction


Subscripts
D = Dimensionless variable
o = Oil
i = Initial
wf = Sandface

References
Adegbesan, K.O., Costello, J.P., Elsborg, C.C., King, D.A., Mills, M. and King, R.W. 1996. Key Performance Drivers for
Horizontal Wells Under Waterflood Operations in the Layered Pembina Cardium Reservoir. The Journal of Canadian
Petroleum Technology 35 (8): 25-35.
Bello, R.O. 2009. Rate Transient Analysis in Shale Gas Reservoirs with Transient Linear Behavior. PhD Dissertation. Texas
A&M U., College Station, Texas.
Bello, R.O. and Wattenbarger, R.A. 2008. Rate Transient Analysis in Naturally Fractured Shale Gas Reservoirs. Paper SPE
114591 presented at the CIPC/SPE Gas Technology Symposium 2008 Joint Conference, Calgary, Alberta Canada, 16-19 June.
Bello, R.O. and Wattenbarger, R.A. 2009. Modeling and Analysis of Shale Gas Production with a Skin Effect. CIPC Paper 2009-
082 presented at the CIPC Conference, Calgary, Alberta Canada, 16-19 June.
Bello, R.O. and Wattenbarger, R.A. 2010. Multi-stage Hydraulically Fractured Shale Gas Rate Transient Analysis. SPE Paper
126754 presented at the SPE North Africa Technical Conference and Exhibition, Cairo, Egypt, 14-17 February.
Breit, V.S., Stright Jr., D.H., and Dozzon, J.A. 1992. Reservoir Characterization of the Bakken Shale from Modeling of Horizontal
Well Interference Data. Paper SPE 24320 presented at the SPE Rocky Mountain Regional Meeting held in Casper, Wyoming,
Colorado, 18-21 May.
Camacho-V, R.G. and Raghavan, R. 1989. Performance of Wells in Solution-Gas-Drive Reservoirs. SPEFE (December 1989):
611-620.
Chen, C. C., and Raghavan, R. 1997. A Multiply-Fractured Horizontal Well in a Rectangular Drainage Region. SPEJ 2
(December 1997): 455-465. SPE-37072-PA.
Clarkson, C.R., and Beierle, J.J. 2010. Integration of Microseismic and Other Post-Fracture Surveillance with Production Analysis:
A Tight Gas Study. Paper SPE 131786 presented at the SPE Unconventional Gas Conference held in Pittsburgh,
Pennsylvania, 23-25 February.
Cox, S.A., Cook, D.A., Dunek, K., Daniels, R., Jump, C., and Barree, B. 2008. Unconventional Resource Play Evaluation: A Look
at the Bakken Shale Play of North Dakota. Paper SPE 114171 presented at the SPE Unconventional Reservoirs Conference
held in Keystone, Colorado, 10-12 February.
Gilbert, J.V., and Barree, R.D. 2009. Production Analysis of Multiply Fractured Horizontal Wells. Paper SPE 123342 presented at
the SPE Rocky Mountain Petroleum Technology Conference, Denver, Colorado, 1416 April.
Krause, F.F., Collins, H.N., Nelson, D.A., Machemer, S.D., and French, P.R. 1987. Multiscale anatomy of a reservoir: geological
characterization of Pembina-Cardium pool, west-central Alberta, Canada: American Association of Petroleum Geologists
Bulletin, 71, 1233-1260
Lawton, D., Alshuhail, A., Coueslan, M., and Chabot, Louis, C. 2009. Pembina Cardium CO2 Monitoring Project, Alberta
Canada: Timelapse Seismic Analysis, - Lessons Learned. Energy Procedia 1: 2235-2242.
Lee, J., Rollins, J.B., and Spivey, J.P. 2003. Pressure Transient Testing, Textbook Series, SPE, Richardson, Texas 9.
16 CSUG/SPE 137352
Medeiros, F., Ozkan, E., and Kazemi, H. 2008. Productivity and Drainage Area of Fractured Horizontal Wells in Tight Gas
Reservoirs. SPEREE (October 2008): 902-911. SPE-108110-PA.
Medeiros, F., Kutoglu, B. Ozkan, E., and Kazemi, H. 2010. Analysis of Production Data from Hydraulically Fractured Horizontal
Wells in Shale Reservoirs. SPEREE (June 2010): 559-568. SPE-110848-PA.
Ozkan, E., Brown, M., Raghavan, R., and Kazemi, H. 2009. Comparison of Fractured Horizontal-Well Performance in
Conventional and Unconventional Reservoirs. Paper SPE 121290 presented at the SPE Western Regional Meeting, San Jose
California, 2426 March.
Palmer, I., and Moschovidis, Z. 2010. New Method to Diagnose and Improve Shale Gas Completions. Paper SPE 134669
presented at the SPE Annual Technical Conference and Exhibition, Florence, Italy, 1922 September.
Pratikno, H., and Blasingame, T.A. 2003. Decline Curve Analysis Using Type-Curves Fractured Wells. Paper SPE 84287
presented at the SPE Annual Technical Conference and Exhibition, Denver, Colorado, 58 October.
Wattenbarger, R.A., El-Banbi, A.H., Villegas, M. E. and Maggard, J.B. 1998. Production Analysis of Linear Flow Into Fractured
Tight Gas Wells. Paper SPE 39931 presented at the SPE Rocky Mountain Regional/Low-Permeability Reservoirs
Symposium, Denver, Colorado, 58 April.