You are on page 1of 24

SPE

Society of Petroleum Engineers


SPE 28688
Decline Curve Analysis Using Type Curves--Analysis of Oil Well Production Data Using Material Balance Time:
Application to Field Cases
by L.E. Doublet,- Texas A&M U., P.K. Pande,- Fina Oil and Chemical Company, T.J. McCollum,- UNOCAL-Coastal Califomia,
and T.A. Blasingame,- Texas A&M U.
SPE Members
Copyright 1994. Society of PetIOleum ervn-rs. Inc.
This paper_pI8IllIIIld for pr8S1lation at1he 1994 PelItlIeumConIeI1ll1C8 and Exhibition of Mexico held in Veracruz, MEXICO, 1()'13OclDber 1994.
This JllIIl8r _ selected for plllSllnlalion by an SPE Committee lolowing lllYiewof information oontained in an absllllct submillad by 1he aulhor(s). Contents 011he paper, as preSlted,
hawnot been nMewed by 1he Society of Petroleum Engineers and are subjeclto OOll8Ction by the aulho'1s). The ma18riaJ, as presented, does not necessanly reftecl position 01 1he Society 01
PlIlrdeum Enslir-rs, its OIlicers, or members. Papers pr8ssnted at SPE meetings are subject to publcalion IIIYiew by Edi1DriaJ Commillees of1he Societyof Pelroleum Engineers. Pennission to
copy is restriCted to an abstract of not more than 300 words. lUustrations may not be The absll8cl should contain oonspicuous acknowledgrrient 01 where and by whom 1he paper is
p-m-d. Write Ubnuien, SPE, P.O. Box 833836, Richardson. TX75Oll3--3838, U.S.A. Telex, 163245 SPEUT.
Harmonic: (b=l)
BRIEF SUMMARY
This paper presents rigorous methods to analyze and interpret
production rate and pressure data from oil wells using type curves
to perform decline curve analysis. These methods are shown to
yield excellent results for both the variable rate and variable
bottomhole pressure cases, without regard to the structure of the
reservoir (shape and size), or the reservoir drive mechanisms.
Results of these analyses include the following:
Reservoir properties:
- Skin factor for near well damage or stimulation, S
- Formation permeability, k
In-place fluid volumes:
- Original oil-in-place, N
- Movable oil at current conditions, N
p

mov
- Reservoir drainage area, A
We have thoroughly verified these analyses and interpretation
methods using both synthetic data and numerous field examples.
In addition, we provide illustrative examples to demonstrate the
ease of analysis and interpretation, as well as to orient the reader
as to what are the benefits of rigorous decline curve analysis.
INTRODUCTION
The importance of performing accurate analysis and interpretation
of reservoir behavior using only rate and pressure data as a
function of time simply can not be overemphasized. In most
cases, these will be the only data available in any significant
quantity, especially for older wells and marginally economic wells
where both the quantity and quality of types of data are
limited. The theoretical application of this technique is for newer
wells, at pressures above the bubble point, although we show that
the methods described here can be accurately applied at any time
during the depletion history of a particular well.
The development of modem decline curve analysis began in 1944
when Arpsl published a comprehensive review of previous efforts
for the graphical analysis of production decline behavior. In that
work, Arps developed a family of functional relations based on
the hyperbolic decline model for the analysis of flow rate data.
References and illustrations at end of paper
Arps' efforts provided a variety of results; including the
exponential, hyperbolic, and harmonic rate decline relations that
we use today for empirical decline curve analysis. Due to the
simplicity and consistency of this empirical approach, the Arps
relations remain a benchmark in the industry for the analysis and
interpretation of production data.
The utility of the Arps relations is the applicability of the
hyperbolic family of curves to model a wide variety of production
characteristics. In addition, the simplified analysis of exponential
and hyperbolic data trends (such as the graphical techniques
provided by Nind
2
) maintain the popularity of the Arps relations.
The application of the Arps relations typically includes a semilog
plot of rate versus time where the hyperbolic cases yield gently
declining curves which have the straight-line, exponential decline
case as a lower limit. Nind
2
provides the development and
illustration of plotting functions for the graphical analysis of rate
data for the general hyperbolic decline case as well as the
exponential decline case.
Another attraction of the Arps relations is their use in graphical as
well as functional extrapolation. Many analysts rely uniquely on
the Arps relations for performance predictions, often without
realizing the empirical nature of such extrapolations. In this work
we will use exponential decline case as a basis for estimating
movable oil at current conditions, N
p
mov. We will demonstrate
that this approach can be derived theoretically for the case of a
well produced at a constant bottomhole flowing pressure. We will
also show that this approach works for wells which are not
produced at such restrictive conditions.
The Arps relations for flow rate and cumulative production are
given as follows
Arps Flow Rate Relations
Rate Relation
Exponential: (b=O) q(t) = q; exp(-D;t) (1)
Hyperbolic: (O<b<l) q(t) = qj .oo (2)
[1+bD;t]lib
q(t) = [l+'lj,;t] (3)
2
Decline Curve Analysis Using Type Curves-Analysis of Oil Well Production Data Using Material Balance Time: SPE 28688
Application to Field Cases
Harmonic: (b=I)
Exponential: (b=O)
Arps Cumulative Production Relations
Cumulative Relation
Np(t) = - exp(-D;t)] (5)
or in terms of q(t)
Np(t) = I5t[q; - q(t)] (6)
Hyperbolic: (0<b<1) () q' [ 1 lIbl
Np t =(l-b)v; 1 - (l+bDjt) - J (7)
or in terms of q(t)
Np(t) = (l-b\vJ1 - (8)
Np(t) = 1n(1+Djt) (9)
or in terms of q(t)
Np(t) =!li In{!J.L) =_!li 1jq(t) .. (10)
Dj Dj u, qj
addition to presenting these fundamental relations, Arps3later
1Otroduced methods for the extrapolation of rate-time data to
estimate primary oil reserves using the exponential and hyperbolic
decline curve models.
The use of "type curves" (dimensionless or normalized flow rate
solutions plotted on a scaled graph) for analysis of production data
was introduced to the petroleum industry in the late 1960's and
early 1970'S.4,5 In 1980 (preprint 1973) Fetkovich
6
introduced
the significant development in type curve matching of
productIon data-the creation of a umfied analytical solution
(exponential decline) for a well produced at a constant bottomhole
pressure during boundary-dominated flow conditions.
Further. Fetkovich
6
plotted his unified exponential decline
simultaneously with the Arpsi hyperbolic decline stems.
which are assumed to account for non-ideal reservoir behavior
(changes in mobility. heterogeneous reservoir features and
reservoir layering). The final result is the so-called
type curve, which provides for the simultaneous analysis of
production data during transient and boundary-dominated flow
conditions. While the Fetkovich decline curve is an extraordinary
f<;>r reservoir engineering. this approach is not without
lImItations.
A partic:ular limitati?n .in and interpretation of
production data WhICh exhIbIt sIgmficant variations in wellbore
pressure. as well as the effects of periodic shut-ins and other
constraints imposed by operational considerations. To its credit
the Fetkovich decline curve is the most powerful tool available fo;
the analysis of production data, as demonstrated in refs. 6-10. In
this our efforts serve only to extend the utility and
applicabIlity of this type curve analysis approach.
The in!tial to inco.rporate rate pressure changes into the
analySIS and 1Oterpretation of production data was introduced in
1986 by Blasingame and Lee.
ll
This work provides analysis
for d!ainage area size and shape from
vanable-rate production data 10 closed reservoirs using a Cartesian
plot based on the following relation
Ap -
q= mt + bpss (11)
where Ap =pj PWf' and
m ="'iJc; (12)
bpss = (13)
e CA r
wa

and the deftnition of "material balance time" is given by
t=Nplq (14)
The analysis method derived from Eq. 11 was observed to work
best when rate changes were small, that is, when the transients
induced by rate changes do not obscure the boundary-dominated
flow behavior for long periods of time. Eq. 11 was derived using
the Dietz result
I2
for the constant rate case, and verified by
comparison to the Muskat
I3
solution for a bounded circular
reservoir and by the analysis of simulated well performance data.
Continuing in a chronological fashion, we note that in 1987
Fetkovich, et ar presented a series of field case studies evaluated
by decline curve analysis using type curves. In addition to several
field exat:r'lples,. the auth<;>rs also gave commentary
regarding the analysIs and 1Oterpretation of production data using
decline type curves.
One of the t:r'Iajor conclusions. of the Fe.tkovich, et ar study was
the observation that the analysIS oftransient production data using
the Arps hyperbolic equations is invalid. Transient flow theory
states that the flow rate profile should be concave up, and as a
declining function, the Arps stems are concave down--which
clearly poses an inconsistency in bOth the analysis and
interpretation flow data. A curious development was
the emergence 10 the 10dustry of a "rule-of-thumb" during the
1970's and 1980's where it was suggested that an Arps stem of
b>1 should be used for the analysis of transient flow data.
However, from the previous arguments it is obvious that this
"rule" is without foundation and will ultimately lead to erroneous
results as well as incorrect interpretations.
Put in a practical sense, transient flow data (production data
functions which are concave up) should never be used to estimate
reserVo.ir volume. Specifically, Fetkovich, et al suggest that
reservOIr volumes and volume-related flow characteristics should
not estimated using curve.analysis before boundary-
dommated flow fully eXISts (production data exhibit a concave
downwards behavior).
In 1991 Blasingame, et aI9 expanded on the earlier work of
McCray8 to develop a time function that would transform
production data for systems exhibiting variable rate or pressure
drop performance into an equivalent system produced at a constant
bottomhole pressure. The motivation of this effort was to create
an equivalent constant pressure analysis formulation for the
analysis of variable-rate/variable pressure drop production data.
Unfortunately, the solution provided by Blasingame et al while
theoretically consistent, is somewhat difficult to apply the
approach appears to be very sensitive to erratic changes in rate and
pressure.
However, the Blasingame, et aI9 study provided both insight and
motivation for the development of a more robust and less
approach to analyze and interpret variable-
rate/vanable pressure drop production data which ultimately
resulted in our present efforts. '
McCray8proposed the following relation as a defmition for the
"equivalent constant pressure time,"t
cp
A;t) = (15)
McCray provided a recursive-type trapezoidal rule formulation to
Eq. 5 !cp' In Blasingame, et aI9 provided a
senes of denvative formu}ations for computing tcp' As attractive
as the of an constant pressure model is, the
comp.utatlOnal as'pec.ts of Its application are unsatisfactory,
espeClally for applIcatIon to field data with erratic variations in the
rate and bottomhole pressure profiles.
The utility of the t
cp
concept is significant given the use of the
(liquid and Carter
I4
.
IS
(gas flow) type curves
for analySIS of production data, and given this potential, we
SPE28688
L.E. Doublet, P.K. Pande, T.J. McCollum, and T.A. Blasingame
3
recommend that the equivalent constant pressure concept be
considered for further study.
In 1993, Palacio and Blasingame
lO
developed a solution for the
general case of variable rate/variable pressure drop for the flow of
either single-phase liquid or gas. These authors showed that for
any particular production history using the pressure drop
normalized flow rate function and the material balance time
function will yield a harmonic rate decline (b=l stem on a
Fetkovich decline curve) for liquid flow.
The authors derived this method rigorously from the
pseudosteady-state (or boundary-dominated) flow equation as
follows. Recalling the pseudosteady-state flow equation, Eq. II,
and the definition of the material balance time, i, (Eq. 14) we have
Ap -
q=mt + b
pss
(II)
where Ap = pj - Pwf and
i =Nplq (14)
Taking the reciprocal of Eq. II gives
=[mi}b
pss
] (16)
Rearranging this result gives,
b
pss
=[1 +In.- i]
b
pss
or reducing to shorthand notation we have
qlAp I
(qIAP)jnt [I +Dii] (17)
where the (qIAp)jnt term is defmed as
(qIAn). - -...L--Lkl!. 1 (18)
r mt - b
pss
-70.6 BJlln .
e
Y

and the Dj term is defined as
-.lL-
D - m - 7 9545 10-
2
'Jlc,A (19)
I - bpss -. x In
Making the fmal reduction ofEq. 17 we have
qDd=[1 (20)
where the definitions of ifri and q/Jd for this case are given by
iDd=Dji (21)
and
_ qlAp
qDd - (qlAplint (22)
Recalling the Arps "harmonic" decline relation (b=l) as defined by
Fetkovich
6
(and given as Eq. B-3 in Appendix B) we have
qDd =[1 }tD& (23)
Comparing Eqs. 20 and 23 we immediately recognize that these
relations are identical. And further, if we consider the base
relation for variable-rate/variable pressure drop performance, Eq.
16, we note that during boundary-dominated flow qlAp data
ploued versus i will exactly overlay the Arps b=l stem on the
Fetkovich decline curve. This was the foundation of analysis for
the work by Palacio and Blasingame
lO
as well as the basis for our
efforts in this present work.
In the present work: we focus on the analysis and interpretation of
production data (flow rates and bottomhole pressures) for oil
wells in order to estimate reservoir volumes and flow
characteristics. We focus on using data that operators acquire as
part of normal field operations (e.g., production rates from sales
tickets and pressures from permanent surface and/or bottomhole
gauges). This approach eliminates the loss of production that
occurs when wells are shut in for pressure transient testing, and
provides analysis and interpretation of well and field performance
at little or no cost to the operator. In addition, the methods we
introduce in this paper are not constrained by the requirement of
constant rates or bottomhole pressures, as is the case for the
previously published methods.
As we mentioned earlier, the analysis methods that we present in
this work provide estimates of the following:
Reservoir properties:
- Skin factor for near well damage or stimulation, S
- Formation permeability, k
In-place fluid volumes:
- Original oil-in-place, N
- Movable oil at current conditions, N
p
mov
- Reservoir drainage area, A .
METHODS FOR THE ANALYSIS AND INTERPRET-
ATION OF PRODUCTION DATA
Harmonic Decline Case: General Approach for
Variable-RateNariable Pressure Drop Production Data
As we discussed in the Introduction, the rigorous solution for any
rate and pressure schedule for the case of a well producing under
boundary-dominated flow conditions is given by Eq. 16. Recall-
ing Eq. 16 we have .
=[mi;' b
pss
] (16)
We recognize that Eq. 16 is a "harmonic" type of equation in
which the "material balance time" function, i, is given by Eq. 14
as
i=Nplq (14)
As such, we simply plot the pressure drop normalized rate
function, qlAp, versus material balance time, i, on a scaled log-log
plot and match these data on the FetkovichlMcCray type curve,lO
with the boundary-dominated flow data being force matched (by
definition) on the Arps b=1 depletion stem. The type curve
matching procedures and the associated analysis methodologies
are discussed later in this text, as well as in Appendix C.
Fetkovich-McCray Decline Type Curve
The so called "FetkovichlMcCray type curve" was first presented
as a single entity in ref. 10, although components of this curve
were presented by Fetkovich
6
(1980, preprint 1973) and McCray8
(M.S. thesis 1990). The utility of the resulting "Fetkovichl
McCray" solution is the ability to match flow rate functions as
well as the flow rate integral and integral derivative functions
simultaneously. In addition, the integral functions provide
smoother data trends for clarity and ultimately, improved
matching of data and type curves.
meCurve Development
Although both Fetkovich
6
and McCray8 provide the details of the
development of their respective decline type curves, we believe
that a unifying discussion is in order, particularly for readers
interested in further developments of this type.
It is important to recall that the "analytical" stems (transient stems
and the exponential decline case [b=O stem]) on the Fetkovichl
McCray type curve
lO
(or any "decline" type curve for that matter)
are solutions for a well producing at a constant bottomhole
flowing pressure. However, the methodology indicated by Eqs.
14 and 16 indicate that the FetkovichlMcCray type curve can be
used to analyze any type of production data, including data
4 Decline Curve Analysis Using Type CUNes-Analysis of Oil Well Production Data Using Material Balance Time: SPE 28688
Application to Field Cases
exhibiting arbitrary changes in rate and pressure, so long as the
boundary-dominated flow data are "force matched" on the b=1
(hannonic) stem. The current discussions consider the application
of the Fetkovich/ McCray type curve only for cases of radial flow,
in particular, vertical wells and vertically fractured wells which
exhibit radial flow. The FetkovichlMcCray type curve approach
was recently extended to horizontal wells as described in ref. 16.
In order to be consistent with current literature we use the
Fetkovich
6
definitions of the dimensionless decline variables (tnt
and qnt) which are given below. The tnt function is given in
terms of dimensionless variables as
tDd =t [In r _tJ tD (24)
and in terms of real variables we have
tDd =0.00633 [In r;; _tl (25)
In a similar fashion, the qDd function is given in terms of
dimensionless variables as
qDd =[In rtD - t] qD ........................ (26)
and in terms of real variables we have
qDd= 141.2fr [InrtD-t] .................. (27)
A minor discrepancy in these definitions is that the 1/2 term
should actually be 3/4 as noted by Ehlig-Economides and
RameyP We maintain the convention of using 1/2 rather than
3/4 for the purpose of type curve developments in order to be
compatible with existing literature. But in fact, this "discrepancy"
rarely makes more than a few percent difference in the
interpretation, and is only noted here for completeness.
The rate integral and rate integral derivative functions introduced
by McCray8 are given in dimensionless form below. The.
dimensionless rate integral function, qDdb is given as
I
tDJ
._NpDd --1.-
qDdi - tDd - tDJ 0 qIJri or) dor (28)
and the dimensionless rate integral derivative function, qIJdid, is
given as .
. -- dqDdj -
qDdid - d 1n{tDd) tDd dtDd .......................... (29)
where Eq. 29 can be reduced to the following result as shown in
AppendixB .
qDdid =qDdi -qDd ............................................. (30)
In addition, we introduce the dimensionless rate derivative func-
tion, qDtii, which is defined as
qDdd=- --tDd (31)
Unfortunately, we do not expect Eq. 31 to be of much use in the
analysis of production data due to the volume of random error
found in production data, where these random errors will only be
magnified by the differentiation process.
In order to develop the FetkovichlMcCray type CUNe, we require
values of the solution for a well produced at a constant bottomhole
pressure, qD, as a function of dimensionless time, tD, which are
then converted to tnt and qIJd using Eqs. 24 and 26 respectively.
These qrJ..tD} values can be obtained from tables in van Everdingen
and Hurst
I8
or using numerical inversionI9 of the Laplace
transform solution developed by Matthews and Russel1.
20
The
Laplace transform solution for constant rate production for a well
centered in a bounded circular reservoir is given by Matthews and
Russell
20
as
- _ Ko(-liirD) h{-liirdJ) + KI{-liirdJ)Io(-liirD)
PD (rD'u) - u [-Iii KI(-lii) h(-liirdJ} - -Iii h(-lii} KI(4u"rdJ)] '" (32)
However, we require the solution for a constant flowing
bottomhole pressure rather than a constant flowrate. We can
easily obtain the constant bottomhole pressure solution from the
constant rate solution using the following relation in Laplace space
given by van Everdingen and Hurst.
18
This result is
qD{u)=;':pju) (33)
Once the qnJ.tDJ) values are obtained from qrJ..tD) values, the
associated aeovative and integral functions can be computed using
standard techniques, or these functions can be computed
simultaneously with the qrJ.tnt) values using the numerical
Laplace transform inversion algorithm.l
9
In Fig. 1 we present the original Fetkovich6 type CUNe, along
with the derivative function, q/Jdd' as defined by Eq. 31. We note
in Fig. 1 that the qIJdl stems show a dramatic characterization of
the transfer from transient to boundary-dominated flow, however,
as we suggested before, we would not expect the qIJdl concept to
be particularly applicable due to random noise present in field
data.
Figure 2 presents the FetkovichlMcCray type cUNe
lO
where qIJd,
qDdj, and qDdid are all plotted versus tDd on the type curve grid.
Although this plot appears somewhat busy, we believe that Fig. 2
provides all of the necessary functions for both rigorous and
empirical analysis of production data. Figure 2 is used throughout
our present work for the analysis and interpretation of both
simulated and field data.
ANALYSIS OF OIL PRODUCTION DATA USING
THE FETKOVICHIMCCRAY TYPE CURVES
A step-by-step procedure for the use of the Fetkovich/McCray
type curve is given in Appendix C, and is abbrevialed in this
section for reference and use in applications. Our type curve
analysis technique provides methods to estimate the original-oil-
in-place and other volume-related properties, as well as the flow
characteristics of the reseNoir.
Our methodology is based on the use of the simple material
balance time function, i, that yields a harmonic decline for the case
of liquid production, regardless of the rate and pressure schedule.
We provide the following procedure for the analysis and
interpretation of production data using decline type CUNes.
1. Computation of Material Balance Time from Production Data
i=Nplq (14)
2. ComDutation of the Bow Rate and Rate Functions
Our approach in this study is to work with the pressure drop
normalized rate function, ql6p, in order to be completely
consistent with the theory given by Eq. 16. Our notaHon will
follow this convention throughout the text, including cases where
continuously measured bottomhole pressure data are not available,
and we use the initial reseNoir pressure, Pi, as the normalizing
condition. The pressure drop normalized rate function is given by
(qI6p) =(p q ) =.!L (34)
i - Pwt: &p
where we use &p =Pi - Pwf as a shorthand notation. The rate
integral function is given by
{qI6p)j =t/: d-r (35)
and the rate integral derivative function is given by
(
16 ). __ d[(qI6p);] _ -(qI6p);]
q P id - d In(ij - - t di (36)
SPE 28688 L.E. Doublet, P.K. Pande, T.J. McCollum, and T.A. Blasingame
5
The three plotting functions (Eqs. 34-36) are computed and
plotted versus the material balance time, i, then matched on the
FetkovichlMcCraylo type curve, taking care to "force match" the
boundary-dominated portion of the data onto the Arps b=1
(harmonic decline) stem. The "force matching" of boundary-
dominated flow data is required by theory and provides the best
possible estimate of oil-in-place, N.
3. Estimation of Qjl-in-Place
Estimating the reservoir volume or oil-in-place, N, from type
curve analysis requires that we relate the definitions of tDi and qDi
(given by Eqs. 25 and 27) to yield a "match point" result in terms
of volume. Equating and isolating terms in Eqs. 25 and 27, we
obtain the following relation
{qDiMP {tDiMP =5.
61
t
8B
(qIAp) MP (ilMP (37)
I/IA c,
Solving Eq. 37 for the oil-in-place, N, we obtain
-lJ!.k{qIAplMJ,
N - Ct (tnihdp {qlXlMP (38)
In order to solve for the pseudosteady-state constant, '!Pss, we will
use the generalized definition of qlXl given by Eq. C-5 in
Appendix C. Recalling Eq. C-5 we have
qDd =141.2 !!E..
kh
B
In (39)
e
Y
CA
We note that Eqs. 27 and 39 are equivalent, but Eq. 27 is strictly
valid only for the case of a well centered in a bounded circular
reservoir and Eq. 39 is valid for a general reservoir/well
configuration using the appropriate shape factor, CA.
Recalling the definition ofb
pss
, Eq. 13, we have
b
pss
= (13)
Combining and solving Eqs. 13 and 39 for b
pss
we obtain the
following match point relation
bpss = (40)
4. Estimation of Reservoir Characteristics
The relations given below are used to estimate volumetric and
flow characteristics of the reservoir based on the results of the
type curve match and the available well data.
Reservoir Drainage Area:
A=5.6148 NB (41)
II (I-S
w
irr)
Reservoir Drainage Radius:
re=Vf (42)
Effective Wellbore Radius:
(43)
reD
Formation Penneability:
k= -2IIn[ 4A l[(tAPLMPl (44)
qlXl p]
or combining Eqs. 40 and 44 we have
l:2ln[ 4A l[--L
b
1 (44)
pssJ
Skin Factor:
s = (45)
ANALYSIS AND INTERPRETATION OF LONG-
TERM PRODUCTION DATA
In this section we present the analyses and interpretations of the
simulated and field data cases that we considered in this study.
Our goal is to be able to analyze cases for which data is plentiful,
but also to be able to accurately estimate movable oil volumes and
fluid flow characteristics when high quality production data is
scarce.
We suggest that our proposed methods for the analysis of long-
term production data are easily transferable to any operator, in
particular, operators that lack the ability to perform periodic
pressure transient tests or long-term production tests.
We present as complete an analysis and interpretation as possible
for each data case. We are able to reduce the adverse affects of
production anomalies that occur during the life of a well, and we
obtained unique type curve matches using production rate and
pressure functions, material balance time, and the Fetkovich/
McCraylo type curve. These production rate functions are
pressure drop normalized rate function, (qIAp),
rate integral function, (qll.P)i, and
rate integral derivative function, (qIAp)id.
This process results in excellent estimates of original and movable
oil volumes, as well as good estimates of permeability and skin
factor. The formation flow characteristics can be calculated with
much greater accuracy and confidence if we have accurate early-
time (transient) data.
When the type curve match on either a transient or depletion stem
is indeterminate, anomalies in the production data can be removed
by reinitializing the data past a particular anomaly. Examples of
such "anomalies" are recompletions, mechanical failures, long-
term shut-ins, and fluctuations in flow rate and pressure at early
times in the life of the well.
When data reinitialization is required due to such anomalies in the
production data, the cumulative oil produced remains constant,
regardless of reinitialization. However, the reinitialization process
requires that we account for prior production in the calculation of
material balance time, i. This is accomplished by computing i
based on the total cumulative production and current rates, then
rescaling the i data to yield t=0 at the first data point. This is a
simple procedure and can be easily implemented with a small
computer program or spreadsheet application module.
Data Preparation and Analysis Procedure
We now provide the procedures that we use to interpret and
analyze production data. These procedures are
1. Verification of pertinent rock, fluid, and completion data
using available field records and fluid property correlations.
The critical data required for our analysis include
Total compressibility Porosily
Fluid viscosity Net Pay Interval
Oil formation volume factor Wellbore radius
Irreducible Water Saturation
2. Initial screening of field production data using semilog and
log-log plots
Identify errors or anomalies in the production data
Locate and annotate changes in the completion practices
Time reinitialization of the production data
Perform integral and integral derivative data smoothing
3. Perform type curve analysis using the Fetkovich/McCray
decline type curve to determine the time and rate match
. 6
Decline Curve Analysis Using Type Curves-Analysis of Oil Well Production Data Using Material Balance Time: SPE 28688
Application to Field Cases

Constant Pwf
= 4000psia
= 1.1 RB/STB
= 1.0cp
= 20.0xlO
6
psi-
1
= 744.7 ft
= 10ft
= 0.20
= 0.00
= 40 acres
= 1 md
= 564,210 5TB
Drainage radius, re
Net pay thickness, h
Porosity, , (fraction)
Irreducible water saturation, Swirr
Original nominal well spacing
Formation permeability, k
Original-oil-in-place, N
Fluid Properties:
Oil formation volume factor, B
Oil viscosity, Jl
Total compressibility, Ct
Production Parameters:
Initial reservoir pressure. Pi
Txge Curve Analysis Result'!
The semilog and log-log production plots, together with the rate
function plots are shown for the two simulated cases in Figs. 3-8.
The rate function, (qlf.p), rate integral function. (qlf.P)i, and rate
integral derivative function, (qlf.P)id. are plotted versus material
balance time, i, on the FetkovichIMcCray type curve as shown on
Fig. 9 (constant pressure case) and Fig. 10 (variable-rate! pressure
case). The boundary-dominated portion of the rate functions are
force matched on the b=1 (harmonic) decline stem as dictated by
theory for the use of material balance time, and the appropriate
match points are taken. The dimensionless drainage radius
matching parameter, reD, is estimated from the position of the data
on the transient flow type curve stems. The reP parameter is then
used to estimate formation permeability and skin factor.
We obtained excellent type curve matches on both the transient
stems (for early-time data) as well as the depletion stems (for late
time or boundary-dominated flow data), as shown on Figs. 9 and
10. The drainage area, total and movable oil volumes,
permeability, and skin factor estimated by type curve analysis
exactly matched the input data to the simulator. verifying our
approach for both cases.
Type Curve Match: FetkovichlMcCray Type Curve (Radial
Flow in a Bounded Reservoir).
Constant Bottomhole Pressure (Fig. 9)
Matching Parameter: reD = 3000 (est.)
[tlliJMp =1.0 =1270.6 days
[qn:llMP = 1.0 [qlf.pJMp =0.00888 STBlDlpsi
Type Curve Analysis Results
Original-Oil-in-Place:
-1. (ilMJ, (qlf.plMJ,
N - Ct (t1l:l)Mp (q1l:l)Mp (38)
N =(1270.6 days)(0.00888 STBlDlpsi) = 564 210 STB
-6 1 '
20x 10 psi"
Reservoir Drainage Area:
A =5.6148 NB (41)
h (l-Swirr)
A =(5.6148 ft
3
/RB)(564,210 STB)(1.1 RB/STB)
(0.20)(10 ft)(1 - 0)
A =(1,742,359 ftl(1 acrel43560 ft2) = 40.0 acres
Reservoir Drainage Radius:
re = Vlfi (42)
re ft2)1tr: = 744.7 ft
Analytical approach-constant bottomhole pressure case:
Plot the flow rate, q, versus cumulative oil production,
Np, and extrapolate to q=O. This method is used when
bottomhole pressure data are not available.
For a complete treatment of the procedures used for the
estimation of movable oil please refer to Appendix A
Simulated Data Cases
We used a 2-D, radial, single-phase black oil simulator with 30
geometrically spaced grids blocks to model well performance in a
single-layer reservoir with homogeneous and isotropic properties.
These cases are used for verification of our type curve analysis
and interpretation methods. A constant bottomhole pressure case
was used as a benchmark and a second case with multiple rate and
pressure changes (including shut-ins) was generated to verify the
variable-rate/pressure drop performance of our approach.
The analysis method was verified using simulated data cases with
a wide range of permeability, and numerous changes in rate and
bottomhole pressure. Agreement between simulated performance
and the results of decline curve analysis were checked for
permeabilities of 1, 10, and 100 md. We present the analysis of
simulated performance for the following production histories
Ttme Bottomhole Rate
li1wl Pressure (psial (STBID)
0.0001 100 variable
3000.0 100 variable
points. This type curve matching process was accomplished
using a commercial software graphics package.
21
These
match points are then used to estimate the following
Oil-in-place, N
Pseudosteady-state flow constant, bpss
Transient stem match, rtf)
These results are then used to estimate reservoir drainage
area, formation permeability, and the near-well skin factor.
4. To estimate the movable oil, Np,mov, at current producing
conditions we use the following
Strictly rigorous approach (requires Pwf data):
Plot calculated average pressure, Peal =Pwf+ qb
pss

versus cumulative oil production. N
p
, and extrapolate to
Pear=JJ
Semi-analytical approach::
Plot (qlf.p) versus cumulative oil production, N
p
, and
extrapolate to (qlf.pFO
Variable Pwfwith 0.0001 variable 15.0
multiple sbut-ins 100.0 1000 variable
200.0 variable 0.0
210.0 2500 variable
310.0 1500 variable
410.0 variable .0.0
420.0 2000 variable
520.0 700 variable
620.0 variable 0.0
630.0 1000 variable
730.0 500 variable
1000.0 200 variable
2500.0 100 variable
4000.0 100 variable
The pertinent reservoir, rock, and fluid properties for these
verification runs are summarized in the table below.
Reservoir. Fluid PropertY and Production Data
Reservoir Properties:
Wellbore radius, rw = 0.25 ft
SPE28688 L.E. Doublet, P.K. Pande, TJ. McCollum, and T.A. Blasingame
7
Effective Wellbore Radius:
rlm=..!L- (43)
reD
rmr- 7jdcS<P = 0.2482 ft
Formation Penneability:
_ !lH. l.In [ 4A ] [(ql&p) MP1
k - 141.2 h 2 (qL\:llMp] (44)
k
= 706 (1.0 cp)(1.1 RB/STB) x
. (10ft) .
In [ (4)(1,742,359 ft2) ] [(.00888)] =I md
(1.781)(31.62)(0.2482 ft)2 (1)
Skin Factor:
s = (45)
s =- =0.0
Since most wells are not usually produced at a constant
bottomhole pressure indefinitely, we developed our second
verification case with multiple rate and pressure changes
(including shut-ins). This case more closely models actual field
performance and should be considered representative of the types
of production histories for which our methodologies were
developed.
Variable Bottomhole Pressure with
Multiple Shut-ins (Fig. 10)
Matching Parameter: reD =3000 (est)
[tlll1Mp = 1.0 [ikp = 1270.6 days
[qni]MP = 1.0 [q/&PMP = 0.00888 STBID/psi
In>e Curve Analysis Results
The results for the second case are calculated similarly
N = 564,210 STB
A = 40.0 acres
re =744.7 ft
Twa =0.2482 ft
k = 1.0 md
s =0.0
Material Balance Analysis: (Figs. 11-16)
Plots of calculated average pressure, Peal' normalized daily rate,
(q/&p), and daily rate, q, versus cumulative production, N
p
' were
constructed to estimate the movable oil volume, Np,moV' Extra-
polation of the plotted data to the Ny axis intercept yields movable
volumes of between 46 and 47 MSTB for both verification cases.
The simulated estimate for movable oil was slightly less
(approximately 45 MSTB).
These extrapolated values represent the movable oil volume at the
time when all reservoir energy has been depleted. These volumes
are usually slightly higher than the actual field value of movable
oil due to the practical and economic inability to produce a well to
such a low pressure level.
When bottomhole pressures are available, the Peal or (q/&p) plots
should be used to estimate Np,mov' Even without bottomhole
pressure data, the plot of q versus Np has been shown to yield
accurate estimates of Np,mov'
Yolumetric Analysis
Np,mov = 45.0 MSTB (simulation)
Np,mov =46.0 - 47.0 MSTB (movable oil plots)
(
47,000 STB \,
Recovery Factor = 564,210 STB}'I00) = 8.33%.
Summaty Discussion:
The simulated cases provide an excellent test for the utility of the
type curve analysis method. The results of the type curve analysis
and material balance analysis are essentially the same as the data
input to the simulator. Our method was shown to work well for a
variety of producing scenarios involving both variable rates and
variable bottomhole pressures, which gives us confidence in
applying these methods to field data cases.
Field Data Cases
This work includes field cases from the following areas:
Location Reservoir
West Texas Clearfork Carbonate (Dolomite)
South Central Texas Austin Chalk Carbonate (Chalk)
West Texas Spraberry Oastic (Turbidite)
Offshore California Lower Repetto Clastic (Turbidite)
The quantity and quality of production data varies for each of the
field cases, and the analysis of each case presents unique
challenges. The types of field production data available for
analysis ineIude
Single well daily rate and bottomhole pressure data
Single well daily rate data with surface tubing and casing
pressure data
Average monthly production data allocated on a tract
basis with no bottomhole pressure data
For many of the wells we analyzed, the rock, fluid, and other
pertinent formation properties were unknown and had to be
estimated. The fluid properties were estimated using the available
field data and from correlations provided in the fluid properties
module of a commercial software package.
22
We suggest that fluid properties be evaluated at an average
pressure when the reservoir is between the initial and bubble point
pressures, and at a pressure just above the bubble point when the
reservoir pressure is below the bubble point Our experience has
shown that these practices yield the best results when using this
approach. Due to the difficulty in obtaining representative values
of certain fluid properties, we suggest reporting a value for the
NCt product. This approach allows each individual analyst to
supply their own estimates of fluid properties, and to provide their
own interpretation of the calculated results.
In addition to difficulties in obtaining representative fluid
properties, we also prefer to report a value for the permeability-
thickness product, kh, in place of permeability because we lack
accurate estimates of net pay thickness for each of the reservoirs
analyzed in this work. However, to be consistent, we do present
permeabilities and drainage areas based on estimated values of net
pay thickness for all cases.
The inability to complete all results with a high degree of
confidence is not related to the analysis or interpretation
methodologies we present, but rather, to a lack of reservoir and
fluid data with which to complete these calculations. We use this
opportunity to point out the importance of early and complete data
collection.
North Robertson Unit (Clearfork), Gaines Co., TX
The North Robertson (Clearfork) Field (Fig. 17) was developed
on a nominal 40 acre well spacing beginning in 1956. The
dominant reservoir producing mechanism for the original 141
wells was solution gas drive. The initial reservoir pressure in the
Lower Clearfork (LCF) was estimated to be 2800 psia. As part of
an infill drilling and waterflood project begun in 1987, 116 new
wells were drilled, reducing well spacing to 20 acres, and
resulting in uniform 40 acre 5-spot patterns. Original-oil-in-place
was estimated to be approximately 230 MMSTB, with primary
8
Decline Curve Analysis Using Type Curves-Analysis of Oil Well Production Data Using Material Balance Time: SPE 28688
Application to Field Cases .
= 2800 psia
unknown
= 1.30 RBISTB
= 1.30cp
= 12.OxIO-
6
psi-
1
= 20.OxIO-
6
psi-
1
= 0.31 ft
= 1300ft
= 250ft
= 0.08
= 0.25
< 1.0 md
= 40acres
= 20 acres
production before unitization in 1987 of 20.5 MMSTB.
Individual well primary recovery factors are low, ranging between
5 and 10 percent
The Lower Clearfork is a shallow-shelf carbonate composed
primarily of a massive dolomite section with varying degrees of
anhydrite cement. The geologic setting at the time of deposition
and subsequent diagenesis contributed to the heterogeneous nature
of the Clearfork formation, which is defined by extremely large
reservoir "pay" intervals, poor vertical and lateral continuity, and
low porosity (8% on average) and permeability (often < 1 md).
The wells were initially completed in the Lower, Middle, and
Upper Clearfork, at measured depths of between 6200 and 7200
feet. The majority of the original completion intervals were in the
Lower Clearfork, which is considered the main pay. Additional
completions were added in the Upper Clearfork and Glorieta
during workover programs in the 1970's. At the inception of the
waterflood project in 1987, many of the original wells were
converted to injectors, and the remaining producers were re-
completed up structure.
Although the reservoir may be difficult to characterize
geologically, the Clearfork does behave like a material balance
reservoir, and the decline curve techniques outlined previously are
applicable. Like many older fields, there are limited data available
for analysis. Much of the fluid properties data, as well as the
completion intervals have been estimated. The oil flow rate data
was allocated to individual wells on a tract basis, and may be in
error, although the errors are not likely to be significant because
the wells were tested for allocation on a semi-annual basis. In
addition, there are no bottomhole pressure data available for the
North Robertson Unit and for analysis purposes we assumed
Pwf= 0, which means that the rate function term, (qlfJ.p), was
actually (qIPi)'
NRU CClearforkl Unit Aujd Property. and Production Data:
Reservoir Properties:
Wellbore radius, rw
Estimated gross pay interval
Estimated net pay thickness, h
Average porosity" (fraction)
Average irreducible water saturation, Swirr
Average formation permeability, k
Original nominal well spacing
Current nominal well spacing
Fluid Properties:
Average oil formation volume factor, B
Average oil viscosity, Jl
Initial total compressib.ility, Cti
Average total compressibility, Ct
Production Parameters:
Initial reservoir pressure (LCF), Pi
Flowing bottomhole pressure, Pwf
NRU Well No. 4202
Figure 18 shows the location of NRU Well 4202 with respect to
its well pattern and the unit. This well was drilled in 1962, and
completed in both the Lower and Upper Clearfork. The well was
stimulated with 3,000 gallons of acid, and hydraulically fractured
with 60,000 gallons of fracturing oil and 90,000 pounds of 20/40
sand. The well initially tested at 141 STBO/D. It had produced
approximately 207 MSTB as of July 1994. Semilog and log-log
production plots shown in Figs. 19 and 20 indicate that there were
no significant rate fluctuations during primary production. It is
interesting to note the decrease in decline rate at approximately
5,500 days of producing time. This stabilizing of the production
rate may be a response to an adjacent waterflood project that was
initiated during the same time period. The response to the unit
waterflood can be seen at approximately 9,000 days, when the oil
rate increased sharply.
Type Curve Analysis Results: (Fig. 22)
We now consider the type curve matching of the rate, (qll::.p), rate
integral, (qII::.P)i, and rate integral derivative, (qII::.P)id, functions
plotted versus material balance time, i, on the Fetkovich/McCray
type curve. The three rate functions are force matched on the Arps
b=l (harmonic) decline stem as before, and the appropriate match
points are obtained.
To obtain the best type curve match, the data was reinitialized at a
time of 549 days. After reinitialization, we obtained a good match
on the depletion stems and a unique match on the transient stems
at an reD value of 160. From the log-log production plot (Fig.
20), we note that the transient flow period had not ended at a time
of 549 days, and the transient match should be valid. Using this
dimensionless radius and the time and rate match points, we
calculate values for in-place oil, drainage area, permeability, and
skin.
Type Curve Match: FetkovichlMcCray Type Curve (Radial
Flow in a Bounded Reservoir).
Matching Parameter: reD = 160
[tlliJMp =1.0 [iJMp =3300 days
[qn:ilMP = 1.0 (qll::.pMP = 0.019 STBlDlpsi
J:xpe Curve Analysis Results: (Fig. 22)
Based on our estimated values for total compressibility and net
pay thickness we find
NCt = 62.7 STB/psi
N = 3.13 MMSTB
A = 35.02 acres
re = 696.9 ft
kh = 19.61 md-ft
k =0.08 md
s =-2.6
Material Balance Analysis: (Fig. 23)
Due to the lack of bottomhole pressure data, it is not possible to
use Peal plotted versus N
p
to estimate movable oil. Instead, we
plot the daily oil rate, q, versus Np to find the movable oil
volume. The extrapolation of this line to the N
p
axis intercept
yields a movable volume at the time when all reservoir energy has
been depleted.
Estimates for primary and secondary movable oil were 190 MSTB
and 130 MSTB, respectively. Our results indicate that
approximately 10,000 STB of primary movable oil remained in
the drainage area of the well when the waterflood was initiated in
1987. The analysis of the secondary decline trend is difficult at
present due to a lack of secondary production history. However,
using the present secondary decline rate we estimate that
approximately 113 MSTB of recoverable oil remained as of July
1994. Obviously, the actual movable oil volume will be less than
the volume calculated if the well were produced to zero rate.
Volumetric Analysis
N
p

mov
= 190.0 MSTB (primary)
Np,mov = 130.0 MSTB (secondary)
Recovery Factor = 6.07% (primary)
= 4.15% (secondary)
Summary Discussion
The results of the type curve match and material balance analysis
yield realistic estimates for original-oil-in-place, movable oil,
drainage area, permeability, and skin factor. The primary
recovery factor calculated using the value of original-oil-in-place
from the type curve match is typical for wells in this unit
A pressure build-up test was performed on well NRU 4202 in
1988, and the permeability to oil was estimated to be 0.2 md, and
SPE28688 L.E. Doublet, P.K. Pande, T.J. McCollum, and T.A. Blasingame 9
= 2650 psia
unknown
= 1.33 RBISTB
= 0.9cp
= 12.4xlO-
6
psi-I
= 18.3x10-
6
psj-l
= 0.3 ft
= 190 ft
= 0.09
= 0.30
1.0 md
= 80acres
the calculated skin factor was -3.7. Both of these values are
consistent with the values obtained from our analysis, although it
should be noted that the calculations for drainage area,
permeability, and skin factor are adversely affected by the lack of
an accurate value for the net pay interval.
NRU Well No. 1004
Figure 24 shows the location of NRU Well 1004 with respect to
its well pattern and the North Robertson Unit. The well was
drilled in 1960, and completed in the Lower, Middle, and Upper
Clearfork. It has produced approximately 135.5 MSTB as of July
1994. The semilog and log-log production plots shown in Figs.
25 and 26 indicate that there were several rate variations and an
extended period of an apparently constant production rate during
primary depletion. Due to the fact that the production data is
allocated monthly on a tract basis, we believe that the rate behavior
between 5,500 and 10,000 days may not represent the well's true
depletion behavior. In order to achieve the best estimate of
original oil-in-place, and the correct type curve match, only
production data prior to 5,500 days was used in our analysis.
l)pe Curve Analysis (Fig. 28)
The production rate functions are plotted versus material balance
time, i, on the FetkovichlMcCray type curve and force matched on
the b=l (harmonic) decline stem. Upon further review, we
reinitialized the data at a time of 336 days, at which point the well
achieved a stable decline rate. After reinitialization, we obtained a
good match on the b=l depletion stems as well as a unique match
on the TelF800 transient stem. From the log-log production plot
(Fig. 26), we note that the transient flow period had not yet ended
at 336 days, and therefore the transient type curve match is valid.
Type Curve Match: Fetkovich/McCray Type Curve (Radial
Flow in a Bounded Reservoir).
Matching Parameter: TeD = 800
[tn:iJMp =1.0 =2000 days
[qniJMP =1.0 [ql.c:\pMP =0.013 STBlDlpsi
l)pe Curve Analysjs
From our estimates of total compressibility and net pay thickness
wefmd
NCt = 26.0 STB/psi
N = 1.30 MMSTB
A = 14.52 acres
Te =448.7 ft
kh = 18.41 md-ft
k =0.07 md
s =-0.6
Material Balance Analysis: (Fig. 29)
As with well NRU 4202, we again plot the daily oil production
rate, q, versus N
p
to estimate the movable oil volume. The
extrapolation of the straight line portion of this data to the Np axis
intercept yields the movable oil volume at the time when all of the
reservoir energy has been depleted. Our results indicate that there
were approximately 105 MSTB of primary movable oil, and 75
MSTB of secondary movable oil (using the average secondary
decline for the unit). The analysis of the secondary decline trend
may be inconclusive due to a lack of secondary production
history, however, we estimate that approximately 44 MSTB of
recoverable oil remained as of July 1994.
Yolumetric Analysis
Np.moY = 105.0 MSTB (primary)
Np,moY = 75.0 MSTB (secondary)
Recovery Factor = 8.08% (primary)
= 5.77% (secondary)
SummlUY Discussjon:
The analysis techniques used for this well show that the analyst
must be careful when major events, such as long shut-in periods,
or questionable production data affect a well's producing history.
If a good well history is available, the analysis and interpretation
can be accurately performed. The results of our type curve match
as well as our material balance analysis indicate that the well is
draining a very small area and may require stimulation, although
the primary recovery factor estimated from this analysis is typical
for wells in the unit.
Sprayberry Trend, West Texas
This particular Spraberry reservoir was initially developed on a
nominal 80 acre well spacing and additional out-of-pattern infill
wells were subsequently drilled throughout the field. The original
reservoir producing mechanism was solution gas drive, but is
presently gravity drainage and waterflood in certain areas of the
field.
The Spraberry Trend in this field consists of two distinct zones
(Upper and Lower) with gross sand intervals of 150 to 600 ft and
330 ft, respectively. The average total net sand interval for the
wells in this field is approximately 190 ft. The average porosity is
about 9 percent and permeabilities are extremely low < 1 md).
After approximately 30 years of primary production, a limited
waterflood was initiated in certain areas of the field, but has had
limited success due to the suspected presence of preferential flow
paths within this reservoir. While it is probable that the lack of
waterflood continuity is due to reservoir heterogeneity, it is also
probable that there is a low sweep efficiency due to communica-
tion of hydraulic fractures between individual wells.
The original-oil-in-place for this reservoir is estimated to be 112.8
MMSTB. The estimates for primary and secondary ultimate
recoveries are 1.9 percent and 1.7 percent, respectively, although
individual well primary recovery factors range as high as 7 to 10
percent for Spraberry reservoirs in general. The initial pressure in
this reservoir was estimated to be 2650 psia.
In this case, only monthly oil production data is available for
analysis. In addition, we have no accurate rock, fluid, or bottom
hole pressure data available for analysis. Since bottomhole
pressure data are not available we assumed Pwr= 0, which means
that the rate function term, (ql.c:\p), was actually (qIPi).
Sprabem Reservoir. Fluid Property and Production Data:
Reservoir Properties:
Estimated wellbore radius, rw
Average net pay thickness, h
Average porosity, (fraction)
Estimated irreducible water sat, Swirr
Average formation permeability, k
Original nominal well spacing
Fluid Properties:
Average oil formation volume factor, B
Average oil viscosity, Jl
Initial total compressibility, Cti
Average total compressibility, Ct
Production Parameters:
Initial reservoir pressure, Pi
Flowing bottomhole pressure, Pwf
Spraberry Well A
This well was drilled in 1957 and completed in both the upper and
lower sections of the Spraberry. The well has produced
approximately 123 MSTB as of September 1993. The semilog
and log-log production plots shown in Figs. 30 and 31 indicate
that the oil rate varied significantly during the later stages of
primary depletion. The rate integral and rate integral derivative
functions reduce the affects of the data scatter evident on the rate
function prof1le (Fig. 32). This smoothing allows for a better type
curve match even for rate data with a high degree of scatter.
10 Decline Curve Analysis Using Type Curves-Analysis of Oil Well Production Data Using Material Balance Time: SPE 28688
Application to Field Cases
= 3326 psia
= 80 psia (7/94)
= 1.35 RB/STB
= 0.45cp
= 16.4xlO-
6
psi-
1
= 21.0xlO-
6
psi-
1
= 0.25 ft
= 300ft
= 0.05
= 0.30
= 0.01 - 1.3 md
1)pe Curve Analysis Results: (Fig. 33)
The (q/!:J.p), (q/!:J.p);, and (q/tJ.P)id rate functions are plotted versus
material balance time, i, and then force matched on the b= 1
(harmonic) decline stem as dictated by theory. Upon obtaining a
match of the data and the type curve trends, the appropriate match
point values are taken. We then obtain a value for the
dimensionless drainage radius matching parameter, rd), which is
used to estimate permeability and skin factor. The match of the
rate functions on the rd)=12 transient stem is excellent
Type Curve Match: FetkovichlMcCray Type Curve (Radial
Flow in a Bounded Reservoir).
Matching Parameter: reD = 12
[tOOMP = 1.0 (iJMp = 8500 days
[qw1Mp =1.0 [q/tJ.pJMp =0.0069 STBID/psi
I}!pe Curve Analysis Results
Using our estimates of total compressibility and net pay thickness
wefmd
NCt = 58.65 STB/psi
N =3.20 MMSTB
A = 45.90 acres
re =797.8 ft
kh =2.024 md-ft
k =0.01 md
s =-5.4
Material Balance Analysis: (Fig. 34)
Since we again lack bottomhole pressure data, we plot daily oil
production rate, q, versus Np to estimate the amount of movable
oil. The extrapolation of this line to the Np axis intercept indicates
that the total primary movable oil volume is 160 MSTB, and that
there were approximately 35 MSTB of primary movable oil
remaining in the well's drainage area as of September 1993.
Presently, there is insufficient data available for comment as to the
volume of secondary oil that may be produced.
Yolumetric Analysis
Np,mov = 160.0 MSTB
Recovery Factor = 4.99%
SummaO' Djscussjon
The results of the type curve match and material balance analysis
yield realistic estimates for original-oil-in-place, movable oil,
permeability, and skin.. The recovery factor (4.99 percent)
calculated using the estimate of original-oil-in-place from the type
curve match and material balance analysis is slightly higher than
average for the field, and the reservoir quality in the area of this
well appears to be high.
Although we had to estimate virtually all of the rock and fluid data
required for the calculation of permeability, the resulting value of
0.01 md is representative for this extremely low permeability,
turbidite reservoir. The calculated skin factor of -5.3 is what we
could reasonably expect for a low permeability, hydraulically
fractured well completion. As of September 1993, the well had
produced 123 MSTB, or 77 percent of the recoverable oil volume
calculated for depletion to zero rate.
Giddings (Austin Chalk) Field, Burleson Co., TX
The Austin Chalk is an Upper Cretaceous, naturally fractured
reservoir consisting of a homogeneous micritic limestone chalk
with interbedded black shales. The reservoir has a low matrix
permeability, with a dominant natural fracture system trending
from Northeast to Southwest, but the presence and influence of
this fracture system is not well correlated. The main producing
trend parallels the Texas Gulf Coast between the Pearsall Field to
the Southwest, and the Giddings Field to the Northeast, although
significant exploration and production activities are presently
occurring in East Texas and Louisiana.
The Austin Chalk consists of an immature zone above 6000 ft, a
generation and accumulation zone between 6000 and 7000 ft, and
a more mature oil generation and accumulation zone below 7000
ft, in which the fracture system is most dominant.
23
The Giddings (Austin Chalk) Field was first developed in the late
1970's. Initial field development used vertical wellbores,
however, with the rapid development of horizontal well
technology in the early 1980's, almost all subsequent wells drilled
in the field have been horizontal to take advantage of the Austin
Chalk fracture system.
In the Giddings Field, the Austin Chalk has an average porosity of
approximately 5 percent and an average permeability between 0.01
and 1.3 md, depending on the relative contributions of the matrix
and fracture systems. The reservoir has an average thickness of
between 200 and 800 ft. Total cumulative production as of 1993
was estimated to be 150 MMSTB. The original reservoir pressure
for the Giddings Field was estimated to be 3326 psia.
The quantity and quality of production data was fairly good for the
wells we analyzed. In particular, both daily rates and surface
pressures are available. The problem we face in these analyses is
our inability to accurately convert surface flowing pressure to
bottomhole flowing pressure, as well as the lack of accurate rock
and fluid data. To be consistent, surface tubing pressure,p,ft will
be used instead of Pwf for both of the Austin Chalk cases we
present
Giddin2S Field. Fluid Property and Production Data
Reservoir Properties:
Estimated wellbore radius, rw
Estimated net pay thickness, h
Average porosity" (fraction)
Estimated irreducible water sat, Sw;rr
Average formation permeability, k
Fluid Properties:
Average oil formation volume factor, B
Average oil viscosity, Jl
Initial total compressibility, C'i
Average total compressibility, Cl
Production Parameters:
Initial reservoir pressure, Pi
Flowing surface tubing pressure, Pwf
Barton Lightsey Well No. 64
This well was drilled and completed in 1991, and has produced
approximately 330.5 MSTB of oil as of July 1994. The well
presently has a daily oil rate of 102 STBID, a producing GOR of
5275 scf/STB, and a water cut of 9%. The semilog and log-log
production plots shown in Figs. 35 and 36 indicate that the oil
production rate started out very high and then declined rapidly, as
would be expected from a dual porosity system (fracture/matrix
drainage).
After approximately 500 days of production, the well was placed
on gas lift and the oil production rate increased sharply from 170
STBID to about 400 STBID before resuming the initial decline
rate. The rate integral and rate integral derivative functions shown
in Fig. 37 were slightly affected by periodic rate variations at early
producing times. The availability of daily production and surface
pressure data improves our chances of obtaining a unique type
curve match.
Type Curve Analysis Resultc;: (Fig. 38)
As before, (q/tJ.p), (q/tJ.P)i, and (q/tJ.p)id are plotted versus material
balance time, i, and matched on the FetkovichIMcCray type curve.
From Fig. 35, we see that due to numerous rate changes and shut-
in periods early in the well's life, it is difficult to obtain a unique
match on the transient flow stems. To improve our chances for
obtaining a match of the transient data, we reinitialized the data to
SPE 28688 L.E. Doublet. P.K. Pande, T.J. McCollum, and T.A. Blasingame
11
a time of 132 days to remove the most significant part of the rate
data scatter.
After reinitialization, we obtained a good match on the reD=28
transient flow stem. It is interesting to note the effect that the gas
lift process has on the flow rate profile. The rate profile shows a
spike-like trend in Fig. 36 but is smoothed to a pair of overlapping
trends for the (q/6.p) function in Fig. 37. This behavior does not
affect the overall quality of the type curve match.
Type Curve Match: FetkovichlMcCray Type Curve (Radial
Flow in a Bounded Reservoir).
Matching Parameter: rtD =28
[tDtiMp = 1.0 [iJMP = 330 days
[qn:iJMP = 1.0 [q/6.pMP = 0.31 STBID/psi
1J!pe Curve Analysis Results
For our estimates of total compressibility and net pay thickness we
find
NCt =102.3 STB/psi
N =4.87 MMSTB
A =80.73 acres
re = 1058.0 ft
kh = 68.67 md-ft
k =0.23 md
s =-5.0
Material Balance Analysis: (Figs. 39-41)
Plots of Peal' (q/6.p), and q versus N
p
are used to estimate the
movable oil volume. We assume that because flowing bottomhole
pressure is held constant after gas lift is initiated, the straight line
extrapolation of q to zero yields about the same value for movable
oil as does extrapolation of Peal or (q/6.p) to the N
p
axis intercept
All of the material balance methods yield a movable oil volume of
360 MSTB with gas lift. which means that there are approximately
30 MSTB of movable oil remaining in the reservoir at present
conditions.
We also note that during the period before installation of gas lift,
that the extrapolated movable oil volumes for all material balance
methods are also quite similar (...310 MSTB). The results of the
volumetric analysis are given below.
Volumetric AnalYSis
Np,mov = 360.0 MSTB (with gas lift)
Recovery Factor =7.39%
Summary Discussion
The type curve and material balance analyses yield acceptable
results for original-oil-in-place, movable oil, and the reservoir
flow characteristics. The calculated recovery factor is in the range
of what we would expect for Austin Chalk wells, and we note the
short operating life that is also characteristic of these wells. The
calculated permeability of 0.23 md and skin factor of -5.0 are also
representative values. The calculated permeability may be in error
since we may have underestimated the effective net pay interval.
As this is a horizontal well, it appears that we may be able to
accurately model the behavior of horizontal wells in the Austin
Chalk using the FetkovichlMcCray type curve which was
developed for vertical wells (radial flow). In addition, this
analysis technique may provide a method to estimate the well
drainage area, which is often unknown for Austin Chalk wells.
Scarmardo Carrabba Well No. 225
This well was drilled and completed in 1993, and has produced
approximately 92 MSTB of oil as of July 1994. The well
presently has a daily oil production rate of 58 STBID, a producing
GaR of 5535 scC/STB, and a water cut of 11%. The semilog and
log-log production plots shown in Figs. 42 and 43 exhibit the
characteristic behavior of a dual porosity system. Oil production
rate declines rapidly as the fracture system is drained, and then the
rate of decline is reduced during the period in which the matrix
dominates. As with well Barton Lightsey 64, we again have daily
production rate and surface pressure data for more rigorous
analysis. The rate, rate integral, and rate integral derivative
functions are shown in Fig. 44. These pressure normalized rate
functions are not greatly affected by early-time rate anomalies, and
therefore, data reinitialization is not required.
Ty,pe Curve Analysis Results: (Fig. 45)
The rate functions are once again plotted versus material balance
time, i, and match points are obtained using the FetkovichlMcCray
type curve. We have a good transient match on the reD=800 stem,
and we will use this dimensionless radius along with the time and
rate match points to estimate values for oil-in-place, drainage area,
permeability, and skin factor. This well is presently producing
under boundary-dominated flow conditions, and is probably
nearing the end of its operating life.
From our calculations, this well appears to be draining a much
smaller volume than the Barton Lightsey well, which is not
surprising considering the Scarmardo Carrabba well's per-
formance to date. As we mentioned for the Barton Lightsey well,
the analysis and interpretation from type curve matching may be in
error because we are analyzing a horizontal well with type curves
derived for a vertical well.
Type Curve Match: FetkovichlMcCray Type Curve (Radial
Flow in a Bounded Reservoir).
Matching Parameter: reD = 800
[tn:iJMp =1.0 [iJMP =84 days
[qn:iJMP =1.0 [q/6.pJMp =0.32 STBID/psi
Type Curve Analysis Results
For our estimates of total compressibility and net pay thickness we
find
NCt = 26.88 STB/psi
N = 1.28 MMSTB
A = 21.21 acres
re = 542.3 ft
kh = 162.90 md-ft
k =0.54md
s = -1.0
Material Balance Analysis: (Figs. 46-48)
Plots of Peal' (q/6.p), and q versus N
p
are used to estimate
Np,moV' and again the computed movable volume for all three
methods is exactly the same. Primary movable oil for this well is
estimated to be 100 MSTB, indicating that the remaining movable
oil volume is less than 10,000 STB. The recovery factor is
slightly higher than for the Barton Lightsey 64 even though no gas
lift process was initiated. The comparison of recovery factors is
somewhat misleading when we consider that the Barton Lightsey
well will recover approximately 3.5 times as much oil as the
Scarmardo Carrabba well. We assume that the higher oil recovery
factor is due to better reservoir quality, if not better
communication between the fracture and matrix systems. All
things being equal, one possible recommendation would be to
perform a significant stimulation treatment on this well.
Yolumetric Analysis
Np,mov = 100.0 MSTB
Recovery Factor =7.81%
12
Decline Curve Analysis Using Type Curves-Analysis of Oil Well Production Data Using Material Balance Time: SPE 28688
Application to Field Cases
= 5900psia
= 50- 600
= 1.42 RB/STB
= 2.0cp
= 10.OxIO-6 psi-I
= 1l.OxIO-6 psi-
1
= 0.146 ft
= 120 - 150 ft
= 925
= 0.35
< 3.0md
=40acres
Summary Discussion
The type curve matching and material balance analyses yield
consistent results even though we have used a type curve
developed for vertical wells to analyze horizontal wells. While the
reservoir quality surrounding this well appears to be much higher
than that of the Barton Lightsey well, the movable oil volume is
much lower, which suggests less than optimal communication of
the well and the reservoir.
Santa Clara (Lower Repetto) Field, Offshore, CA
The Santa Clara (Lower Repetto) Field (Fig. 49) was developed
on an approximate 40 acre nominal well spacing beginning in
1984. There are presently 9 producing wells in the field at an
average true vertical depth of 7500 feet The original reservoir
pressure in the Lower Repetto was estimated to be 5900 psia.
The Lower Repetto reservoir is characterized by four distinct
zones consisting of thinly bedded turbidite sandstones, with
interbedded silts and shales which limit both the vertical and lateral
continuity of reservoir properties. These clastic turbidites were
formed as a result of density currents which were deposited on the
mid and outer fan portions of a turbidite lobe.
This depositional process resulted in the formation of poorly
sorted, medium to very fme-grained arkoses and lithic arkoses.
These sands have porosities ranging from 5 to 35 percent, with an
average in-situ oil permeability of less than 3 md, and possibly
much less even though core permeabilities for the Lower Repetto
often average 20 md or higher. The in-situ reservoir permeability
is much lower than the calculated core permeability due to the
unconsolidated nature of the rock, and relatively high oil viscosity
at reservoir conditions.
Due to this lack of consolidation, sand production is a major
problem and gravel-packed completions using slotted liners are
required. The high viscosity of the oil at reservoir conditions
results in the rapid depletion of reservoir energy, therefore,
pressure communication is limited to within a few hundred feet of
any particular well. The Lower Repetto formation dips at 10 to
20 to the west, and due to the placement of the drilling platform
on the Upper Repetto structure, Lower Repetto wells are inter-
sected at angles between 500 and 600 relative to horizontal.
The original-oil-in-place for the reservoir is estimated to be greater
than 300 MMSTB. Total production from the Lower Repetto as
of January 1994 was 3.6 MMSTB oil and 3.9 BCF gas. Ultimate
recovery is expected to be less than 3% due to the heterogeneous,
low permeability nature of the reservoir, in addition to the high
cost of development drilling.
However, given the producibility problems as well as the expense
of operation, the operator has elected to obtain continuous
measurements of flow rate and bottomhole pressure. Sub-
sequently, the quantity and quality of oil production data and
bottomhole pressure data for the wells is very good, and we
expect to perform a rigorous analysis of these data.
Santa Clara Field. Fluid Property and Production Data:
Reservoir Properties:
Wellbore radius, rw
Net pay thickness, h
Average porosity, ~ (fraction)
Average irreducible water saturation, Swirr
Average formation permeability, k
Original nominal well spacing
FluidProperties:
Average oil formation volume factor, B
Average oil viscosity, Jl
Initial total compressibility, Cti
Average total compressibility, Ct
Production Parameters:
Initial reservoir pressure, Pi
Well deviation
S. Gilda Well S-42
Figure 50 shows the location of Well S-42 within the Santa Clara
Field (Lower Repetto Reservoir). Well S-42 was drilled and
completed in 1986 and has produced approximately 620 MSTB of
oil as of January 1994. At present, the present daily oil rate is 113
STBID, with a producing GOR of 737 scf/STB, a flowing
bottomhole pressure of 2126 psia, and a water cut of <5 percent.
This well intersects the Lower Repetto reservoir at 56.6 and has
an estimated net vertical pay thickness of 150 ft
The semilog and log-log production plots are shown in Figs. 51
and 52 and indicate that the oil rate is declining smoothly, but
quite rapidly, which is probably a result of the producibility
problems mentioned above. The rate integral and rate integral
derivative functions, as seen in Fig. 53, show no instances of
erratic rate variations in the production history.
~ Curve Analysis Results: (Fig. 54)
The three rate functions, (q/l1p), (q/l1P)i, and (Q/l1P)id are plotted
versus material balance time, i, and matched on the Fetkovichl
McCray type curve, as shown in Fig. 54. We have obtained a
very good match on the transient flow stems at a value of rtD=4.
While this is a good match, we must speculate as to why the rtD is
so low, which indicates an extremely high level of near-well
stimulation. The obvious explanation is that Well S-42 is highly
deviated.
The rate functions also indicate that the well is just beginning
boundary-dominated flow, and this behavior may adversely
influence the analysis and interpretation of the well performance.
However, we believe that the results of this analysis are
representative and consistent with boundary-dominated flow
theory. The results of this analysis also indicate that the well is
draining a much larger volume than would be indicated by a 40
acre well spacing. This interpretation of a larger drainage area
may be due to the uncertainty of the net vertical pay thickness and
. the significant deviation of the well. Given the difficulties
associated with interpreting the performance of this well, we
recommend the development and application of decline type
curves for the analysis of horizontal wells (see Ref. 16).
Type Curve Match: FetkovichlMcCray Type Curve (Radial
Flow in a Bounded Reservoir).
Matching Parameter: rtD = 4
[tntJMp = 1.0 [ ~ = 5900 days
[qntJMP = 1.0 [q/l1pJMp = 0.068 STBID/psi
Type Curve Analysis Result5
Using the results of our type curve analysis along with our
estimates of total compressibility and net pay thickness we have
developed the following results
NCt = 401.2 STB/psi
N = 36.5 MMSTB
A = 273.82 acres
re = 1948.5 ft
kh = 17.36 md-ft
k =0.12md
s =-8.1
Material Balance Analysis: (Figs. 55-57)
In this case, we estimate consistent values of movable oil from the
plots of (q/l1p) and q versus N
p
' which yield about 1.0 MMSTB
total recovery. However, the Peal versus N
p
plot predicts
approximately 1.7 MMSTB of movable oil volume. In an attempt
to resolve this discrepancy, we consider that the Peal function is
difficult to interpret relative to the actual pressure level in the
reservoir.
SPE 28688 L.E. Doublet, P.K. Pande, T.J. McCollum, and T.A. Blasingame 13
(qltJ.P)int
q
N
Is there really 1.7 MMSTB of movable oil? Probably not at
current operating conditions, especially when we consider that the
bottomhole pressure level has risen and stabilized for the past
three years. This rise and stabilization in the bottomhole pressure
suggests the need for stimulation and probably artificial lift.
Due to the increase and stabilization of Pwf at a high pressure level
and since the well has only just entered the pseudosteady-state
flow regime, the rate and pressure drop normalized rate-
cumulative oil plots probably yield the most accurate value of
movable oil for this case, and will be used for reference in our
analysis.
'Volumetric Analysis
Np,mov =1.0MMSTB
Recovery Factor =2.74%
Summaey Discussjon
The type curve and material balance analyses yield acceptable
results for estimates of original-oil-in-place and movable oil, and
the calculated recovery factor is reasonable for wells producing
from the Lower Repetto. The calculated drainage area is much
larger than what we expected, but we believe this can be attributed
to a lack of knowledge of net vertical pay thickness and the
deviation of the well.
Perhaps the most intriguing result of this entire analysis is the
match of transient data on the rtIF4 stem, which yields an
estimated permeability to oil of 0.12 md and a calculated skin
factor of -8.1. This estimate of skin factor is unrealistic for any
vertical well case, with the possible exception being the case of an
extremely large, high conductivity vertical fracture. As this is
clearly not the case, we can only assume that the skin factor can
be attributed to well deviation.
In contrast, the semilog and log-log analysis performed on a
pressure build-up test taken in 1992 gave an estimated
permeability to oil of 0.8 md and gave a skin factor of +2.0. If
the permeability and skin factor are corrected for the effects of
partial penetration and well deviation, we expect that the computed
values would be comparable to the values obtained from
production data analysis.
In order to use type curves to accurately estimate formation flow
characteristics for the Lower Repetto wells we should probably
use a matching parameter that incorporates deviatedlhorizontal
well length, instead of effective wellbore radius. The
development and application of type curves for the analysis of
production data for horizontal wells will aid in both the analysis
and interpretation of problems like this.
l6
SUMMARY AND CONCLUSIONS
In this work, we have devised a rigorous and consistent procedure
for the analysis and interpretation of long-term oil well production
data using type curve matching techniques. Specifically, we
propose the use of the Fetkovich/McCray type curve to estimate
total and movable reservoir volumes, as well as the flow
characteristics of the reservoir. Further, given a limited quantity
of production data, we show that we can accurately interpret and
predict reservoir behavior.
We also note that the use of rate integral and rate integral
derivative functions allow for the analysis and interpretation of
"noisy" field production data. In addition, the integral functions
provide better type curve matches and increase confidence in our
interpretations.
The analysis techniques that we propose always yield excellent
estimates of original and movable oil volumes, and accurate
estimates of reservoir flow characteristics, provided good early-
time data are available. Our analysis techniques were verified by
evaluation of the simulated data cases, and we again recommend
that quality data be taken early and often to ensure more accurate
analyses and interpretations.
The main conclusions of this work are:
1. For the case of single-phase liquid flow, the analysis of any
production rate and bottomhole pressure schedule is possible
provided that we use the material balance time function, and
the appropriate rate functions for data matching during
boundary-dominated flow onto the b=1 stem of the Fetkovichl
McCray type curve.
2. Using our methodology to analyze and interpret production
data is relatively straightforward and can provide the same
information as conventional pressure transient analysis,
without the associated cost of data acquisition, or loss of
production.
3. The flow rate integral and flow rate integral derivative
functions allow for more accurate decline type curve matches
than would be possible using flow rate data alone. These
integral functions also eliminate problems associated with the
analysis of field production data with erratic production rate
and bottomhole pressure behavior.
4. The use of data reinitialization for the removal of early-time
rate variations can yield improved type curve matches. The
analyst must be aware of major events in the production
history that might have changed the producing conditions of
the well or reservoir.
5. The calculation of movable oil volume using the q versus Np
plot yields acceptable results unless Pwf varies significantly.
The simulated cases verify that the q versus N
p
plot yields
results similar to those predicted by the more rigorous plots of
(qltJ.p) versus N
p
and Peal versus N
p
' This conclusion has
also been confirmed for field data cases for which surface and
bottomhole pressure data are available.
6. The techniques introduced in this work give excellent
estimates of reservoir volumes (total and movable), and
reasonable estimates of formation flow characteristics.
However, all of these estimates could be significantly
improved if high quality transient production data are
available, as well as accurate rock, fluid, and completion data.
7. Additional work should be developed for the analysis of long-
term production data from horizontal wells. In addition,
present decline type curve analysis concepts should be
extended for the analysis of multiphase flow data.
NOMENCLATURE
Field Variables
Formation and Fluid Parameters:
A = drainage area, ft
2
B = formation volume factor, RBISTB
Ct = total system compressibility, psi-
l
Cti = initial total system compressibility, psi-
l
, = porosity, fraction
h = formation thickness, ft
Swirr = irreducible water saturation, fraction
k = formation permeability, md
re = reservoir drainage radius, ft
r
w
= well bore radius, ft
roo = apparent wellbore radius (includes formation damage
or stimulation effects), ft
J.l = fluid viscosity, cpo
PressurelRateffime Parameters:
b = FetkovichlArps decline curve exponent
b
pss
= constant in the pseudosteady-state equation for liquid
flow, as defined by Eq. 13 or Eq. A-4
Di = constant defined by Eq. 19, D-I
m = constant in the pseudosteady-state equation for liquid
flow, as defined by Eq. 12, psilSTB
=constant defined by Eq. 18, STBlDlpsi
= oil flow rate, STBID
= original oil in place, STB
14 Decline Curve Analysis Using Type Curves-Analysis of Oil Well Production Data Using Material Balance Time: SPE 28688
Application to Field Cases .
Pi
Pwf
Ptt
!J.p
r
t
t
tep
= cumulative oil production, STB
= movable oil, STB
= pressure, psia
= average reservoir pressure, psia
= average reservoir pressure at abandonment
conditions, psia
= initial reservoir pressure, psia
= flowing bottomhole pressure, psia
= flowing surface tubing pressure, psia
=P,-Pwf, pressure drop, psi
= radial distance, r
= time, days
= Nplq, material balance time, days
= equivalent constant pressure time as defined by
McCray8, days
'f = dummy variable of integration
Djrnensionless Variables: Real Domain
CA = reservoir shape factor
r = Euler's Constant =0.577216 ...
NpDd = dimensionless decline cumulative production
function
= circumference to diameter ratio =3.1415926 ...
PD = kh !J.p, dimensionless pressure function for
141.2 qBJl
the constant flow rate case
qD = 141.2 ~ Jq, dimensionless flow rate function
k jP
w
for the constant wellbore pressure case
qDd = dimensionless decline rate function as defmed by
Fetkovich
qDdi = dimensionless decline rate integral as defmed by
McCray
qDdid = dimensionless decline rate integral derivative
function as defmed by McCray
rD = rlrw, dimensionless radius
rdJ = dimensionless drainage radius of reservoir
s = skin factor for near well damage or stimulation
tDA = dimensionless time based on drainage area
t
D
= dimensionless time based on wellbore radius
tfri = dimensionless decline time as defined by Fetkovich
Dimensionless Variables: Laplace Transform Domain
PD = Laplace transform of dimensionless pressure for the
constant flow rate case
qD = Laplace transform of dimensionless rate for the
constant wellbore pressure case
u = Laplace space variable, dimensionless
Special functions
lo(x) = modified Bessel function of the 1st kind, zero order
hex) = modified Besselfunction of the 1st kind, 1st order
Ko(x) = modified Bessel function of the 2nd kind, zero order
K1(x) = modified Bessel function of the 2nd kind, 1st order
Special Subscripts
col = calculated
Dd = dimensionless decline variable
MP = match point
pss = pseudosteady-state
i = integral
id = integral derivative
ACKNOWLEDGMENTS
We acknowledge the permission to publish field data provided by:
Fina Oil and Chemical, Co. (Western Division, USA).
Mobil Exploration and Producing, U.S., Inc.,
UNOCAL Corporation (Coastal California Division), and
Union Pacific Resources Co. (upRC).
We also acknowledge the technical assistance of Dr. Anil Kumar
of Mobil Exploration and Producing, U.S., Inc., and Mr. David
Elmer of UPRC regarding the acquisition and interpretation of
their respective field data cases.
And finally, we acknowledge the technical and computing support
services provided by the Department of Petroleum Engineering at
Texas A&M University, as well as the financial support of the
United States Department of Energy (DOE) for funding provided
through the DOE Class IT Oil Program.
REFERENCES
1. Arps, J.J.: "Analysis of Decline Curves," Trans.,
AIME(1945) 160,228-247.
2. Nind, T.W.: Principles ofOil Well Production, 2nd Edition.
McGraw-Hill (1981). .
3. Arps, J.J.: "Estimation of Primary Oil Reserves," Trans.,
AIME(1956) 207, 182-91.
4. Slider, H.C.: "A Simplified Method of Hyperbolic Decline
Curve Analysis," JPT(March 1968) 235-236.
5. Gentry, R.W.: "Decline-Curve Analysis," JPT (Jan. 1972)
38-41.
6. Fetkovich, M.J.: "Decline Curve Analysis Using Type
Curves," JPT(June 1980) 1065-1077.
7. Fetkovich, M.J., et al: "Decline Curve Analysis Using Type
Curves - Case Histories," SPEFE (Dec. 1987) 637-656.
8. McCray, T.L.: Reservoir Analysis Using Production
Decline Data andAdjusted Time, M.S. Thesis, Texas A&M
University, College Station, TX (1990).
9. Blasingame, T.A. McCray, T.C. and Lee, W.J.: "Decline
Curve Analysis for Variable Pressure DropNariable
Flowrate Systems," paper SPE 21513 presented at the 1991
SPE Gas Technology Symposium, Houston, TX, January
23-24.
10. Palacio, J.C. and Blasingame, T.A: "Decline Curves
Analysis Using Type Curves: Analysis of Gas Well
Production Data ," paper SPE 25909 presented at the 1993
SPE Rocky Mountain Regional/Low Permeability
Reservoirs Symposium, Denver, CO. April 12-14.
11. Blasingame, T.A and Lee, W.J.: "Variable-Rate Reservoir
Limits Testing," paper SPE 15028 presented at the 1986
SPE Permian Basin Oil & Gas Recovery Conference,
Midland, TX, March 13-14.
12. Dietz, D.N.: "Determination of Average Reservoir Pressure
from Buildup Surveys." SPEFE (August 1965) 955-959.
13. Muskat, M.: Flow ofHomogeneous Fluids Through Porous
Media, McGraw-Hill Book Co., Inc., New York (1937).
14. Carter, R.D.: "Characteristic Behavior of Finite Radial and
Linear Gas Flow Systems - Constant Terminal Pressure
Case." paper SPE 9887 presented at the 1981 SPEIDOE
Low Permeability Symposium, Denver, Colorado, May 27-
29.
15. Carter, R.D.: "Type Curves for Finite Radial and linear Gas
Flow Systems: Constant Terminal Pressure Case." SPEJ
(Oct. 1985) 719-728.
16. Shih, M.Y.: Decline Curve Analysis for Horizontal Wells.
M.S. Thesis. Texas A&M University, College Station. TX
(1994).
17. Ehlig-Economides, C.A, and Ramey, H.J., Jr.: "Transient
Rate Decline Analysis for Wells Produced at Constant
Pressure," SPEJ (Feb. 1981) 98-104.
18. van Everdingen, AF. and Hurst, W.: "The Application of
the Laplace Transformation to Flow Problems in
Reservoirs," Trans., AIME (1949). 186, 305-324.
SPE28688 L.E. Doublet, P.K. Pande, T.J. McCollum, and T.A. Blasingame 15
19. Stehfest, H.: "Numerical Inversion of Laplace Transforms,"
Communications of the ACM (January 1970), 13, No. I,
47-49. (Algorithm 368 with correction)
20. Matthews, C.S. and Russell, D.G. : Pressure Buildup and
Flow Tests in Wells, Monograph Series, Society of
Petroleum Engineers of AIME, Richardson (1967) 1.
21. Igor-Graphing and Data Analysis Program (Version 2.7),
WaveMetrics, Lake Oswego, OR, USA, 1992.
22. PanSystem-Well Test Analysis Program (Version 1.8),
Edinburgh Petroleum Services, Ltd., Edinburgh, Scotland,
UK, April 1991.
23. Hinds, G.S. and Berg, R.R.: "Estimating Organic Maturity
From Well Logs. Upper Cretaceous Austin Chalk, Texas
Gulf Coast," Trans., GCAGS (1990) 40, 295-300.
24. Dake, L.P.: Fundamentals of Reservoir Engineering,
Elsevier Scientific Publishing Company, Amsterdam
(1978).
25. Johnston, JL.: Variable-Rate Analysis of Transient Well
Test Data Using Semi-Analytical Methods, M.S. Thesis,
Texas A&M University, College Station, TX (1992).
APPENDIX A - DERIVATION OF MATERIAL
BALANCE PLOTTING FUNCTIONS FOR
PRODUCTION DATA
In this appendix, we start with the material balance equation for a
slightly compressible liquid which is given by Dake
24
as
P=Pi - JJc; N
p
.......................................(A-l)
We note that if we plotp versus N
p
then we will obtain a straight
line of slope lINct and intercept Pi' We can also extrapolate the p
versus Np trend to p=O in order to estimate the "movable" liquid
(oil) volume, Np,mov- Of course, pis typically not available in
practice, so we must use an alternate approach to applying this
concept
We now consider the so called "oil flow equation" which relates
rates and pressure drops during boundary-dominated (or pseudo-
steady-state) flow
24
This expression is given as
P=Pwr+ qb
pss
................................................(A-2)
Combining Eqs. A-I and A-2 and solving for the pressure drop,
Ap =PiPwI' we obtain
l1p = PI-PW/= -l...... N
p
+ qb
pss
(A-3)
NCt
where the pseudosteady-state constant, b
pss
' is given by
bpll =141.2 [t In (A-4)
For the interested reader, a complete derivation of Eq. A-3 from
fundamental principles is given in Appendix A of ref. 25.
Normalizing both sides of Eq. A-3 by the flow rate, q, we have
Ap =-l...... i + b
pss
(A-5)
q NCt
where
-!!.
t= q (A-6)
Eqs. A-5 and A-6 were developed and verified by Blasingame and
Lee
ll
for the analysis of oil well production data. Taking the
reciprocal of both sides of Eq. A-5 and rearranging gives
.!l.- =-L 1 (A-7)
l1p bpss 1 + I i
Ncppss
Eq. A-7 shows that a of plot qlAp versus i will yield a "harmonic"
decline on a FetkovichlMcCray type curve as discussed by Palacio
and Blasingame
lO
for the analysis of oil and gas well production
data.
Estimation of Movable Oil: Constant Bottomhole Pressure Case
Solving Eq. A-3 for the flowrate, q, gives
q=....L{PI-Pwfl- I N
p
(A-8)
bpss Ncppss
We immediately note that if Pwl =constant, then a plot of q versus
N
p
will yield a straight line of the following character
slope = Nc,1pss (A-9)
y-intercept = t-{PiPwfl (A-IO)
pss
x-intercept =Np,mov =N
p
at q=O (A-ll)
This result has considerable implications from a practical
standpoint. In particular, we can use a plot of q versus N
p
as a
means to estimate the movable oil for the case of a well produced
at an approximately constant bottomhole pressure. For cases of
variable bottomhole pressures, Eq. A-8 becomes less applicable,
but we can still use the q versus N
p
plot as a "semi-analytical"
method to predict movable oil.
An interesting historical footnote is that Nind
2
developed Eq. A-8
from a completely empirical perspective. His goal was to develop
the exponential decline case using the observation of a linear trend
of q versus N
p
. In this light, we recall that the analytical
development of the exponential decline solution for a well
produced at constant bottomhole pressure is given by Ehlig-
Economides and Ramey.17
Estimation of Movable Oil: Variable Bottomhole Pressure Case
The development of a variable-rate/variable pressure drop form of
Eq. A-8 can be derived by simply dividing through Eq. A-8 by
the pressure drop, Ap = PiPwt This gives
= - Nc
t
1
pss
(A-12)
Eq. A-12 and other variations of this result are developed and
discussed in detail in refs. 8 and 10.
Considering the form of Eq. A-12, we note that a plot of qll1p
versus N
p
ll1p will yield a straight line with the following
parameters
slope = NCt1pss (A-l3)
y-intercept = (A-14)
[
'!L] -'!L - x-intercept = - at qlAp-O (A-15)
Ap mov Ap
Unfortunately, this method does not yield a direct estimate of
Np,mov- However, we can employ a "semi-empirical" approach
that uses a plot of qlAp versus Np from which the movable oil,
Np,mov' is estimated from the linear extrapolation of the qll1p
trend to the x-axis intercept at qIAp=O. This approach, while not
completely rigorous, should provide accurate estimates of Np,mov
while also "fIltering" the influence of variable rates and pressures.
This is simply an intermediate recommendation and further
research on this topic is warranted.
Estimation of Movable Oil: General Approach
To develop a straightforward and rigorous approach to estimate
the movable oil, Np,mov' we can use the material balance
equation, Eq. A-I, as a plotting function where the average
reservoir pressure, p, is computed from Eq. A-2. Recalling Eq.
A-I we have
16
Decline Curve Analysis Using Type Curves-Analysis of Oil Well Production Data Using Material Balance Time: SPE 28688
Application to Field Cases
Rate Relation
qDd = exp(-tDd} (B-l)
qDd = I (B-2)
[1 + btDdJlIb
qDd =[1 }tDdJ (B-3)
Harmonic: (b=I)
drive mechanisms) are also empirical, based more on speculation
than theory.
The purpose of this appendix is to collect the pertinent Arps
relations and to provide an introduction to the auxiliary rate
functions so that interested readers may create their own type
curves. Starting with the Arps dimensionless rate function, qDd '
we have the following cases
Arps Dimensionless Flow Rate Relations
The different cases for the dimensionless flow rate, qDd' function
are given as
~
Exponential: (b=O)
Hyperbolic: (O<b<l)
Arps Dimensionless Cumulative Production Relations
The definition of the dimensionless cumulative production, NpDd,
is given by
NpDd= f:
Dd
qDJT)dT (8-4)
The different cases for the dimensionless cumulative production,
NpDd, function are given as
~ Cumulative Relation
Exponential: (b=O) NpDd =[1- exp(-tDd}] (8-5)
or in terms of qDd
NpDd =[1- qDd] (B-6)
Hyperbolic: (O<b<I) N 1 [ [ b:/1 lib)'
pDd = I-b 1- 1+ t1' - ] .....(B-7)
or in terms of t ~ and qDd
NpDd = l!b[l-qDd(l+btDd)] ...(B-8)
or in terms of qDd
NpDd =n[l-q]J] (B-9)
Harmonic: (b=I) NpDd = 1n(1 +tDd) (B-lO)
or in terms of qDd
NpDd = 1n(lIqDd} =-1n(QDd}'" (B-11)
Arps Dimensionless Rate Integral Relations
T,he definition of the dimensionless rate integral function, qDdio is
given by
. _ NpDd _ -.Ll
tlli
qDd,- - qDJT)d
T
(B-12)
t/Jd tlli 0
The different cases for the dimensionless rate integral function are
given below
~ Rate n t e ~ r l Relation
Exponential: (b=O) qDdi = -L[l- exp(-tlli)] ....... (B-13)
tDd
or in terms of qDd
qDtJj = -L[l- q/Jd] (B-14)
t/Jd
P=Pi - JJc; N
p
(A-l)
If we have an estimate of the pseudosteady-state constant, b
pss
,
from say, type curve analysis or using the qlt:.p versus Nplt:.p plot
as described above, we can calculate the average reservoir
pressure, p, as
Peal =Pwf+ qb
pss
(A-16)
Plotting Pcol versus N
p
gives the following results
slope = -JJc; (A-17)
y-intercept = Pi (A-I8)
x-intercept = Np,mov = N
p
atp=O (A-19)
While Eqs. A-16 to A-19 provide the most rigorous and
comprehensive analysis of movable oil, this analysis requires a
certain degree of interpretation. For example, we will never
produce an oil reservoir to the p=O condition, so we really want to
determine Np,mov at some Pabn, which depends on the producing
conditions.
Obviously this method assumes that the measured flow rates and
bottomhole pressures are reasonably accurate, which is usually
Dn1 the case in practice. So again, we have a call for vigilant data
acquisition--if we want to perform state-of-the-art analysis and
interpretation of production data.
APPENDIX B - THE ARPS EMPIRICAL RATE
DECLINE FUNCTIONS
This appendix summarizes the Arpsl semi-empirical solutions
(depletion stems) used in the FetkovichlMcCraylO type curves.
These solutions are derived from the Arpsl,3 empirical results for
flow rate, presented in the form of the plotting functions given by
Fetkovich
6
and McCray.s A complete development of these
solutions can be found in Appendix B of ref. 16.
In presenting the Arps solutions we provide several auxiliary
functions based on the flow rate, or in this case dimensionless
flow rate function. The rate and auxiliary functions are given as
follows
variable Description
qDd Dimensionless Rate Function
NpDd Dimensionless Cumulative Production
qDdi Dimensionless Rate Integral Function
qDdid Dimensionless Rate Integral Derivative
As a prelude to these developments, we acknowledge that a
special nomenclature has been adopted for the Arps solutions. In
particular, the term "exponential decline" refers to the case where
the flow rate decays in an exponential fashion with respect to
time.
The exponential decline case is the analytical solution for the rate
behavior in a well producing a single phase liquid at a constant
bottomhole pressure as shown by Ehlig-Economides and
Ramey,17 The term "harmonic decline" refers to the case where
the flow rate varies in a reciprocal fashion with time or some time
function for intermediate to large times. This case is also
"analytical" in the sense that flow rate normalized by pressure
drop plotted versus the "material balance" time function yields
exactly a harmonic decline during boundary-dominated flow
conditions, as shown by Eq. A-7.
The "hyperbolic" decline is the general term given to any decline
curve case lying between the exponential and harmonic decline
cases. Hyperbolic cases generally have little if any analytical
basis, the most notable exceptions being certain ideal and real gas
flow cases as described by Fetkovich.
6
The hyperbolic decline
cases are typically used to empirically model data trends and most
attempts to correlate "hyperbolic" behavior with physical
phenomena (e.g., changes in mobility, layer features, and specific
SPE 28688
L.E. Doublet, P.K. Pande, T.J. McCollum, and T.A. Blasingame
17
Ha17TllJnic: (b=I)
Exponential: (b=O)
Hyperbolic: (O<b<I) qDdj = _1_I_[I-[I+bt11-lIb)]
tfrj I-b
............................ (B-IS)
or in terms of qDd
qDdj = _I[...L_qDd (...L+b)ll
I-b tnt tnt
............................ (B-I6)
Ha17TllJnic: (b=I) 1 1n(I) (B 17)
qDdj =- +tDd -
tDd
or in terms of qDd
qDdj = ....LIn(lIqDd) (B-I8)
tDd .
where the dimensionless time based on drainage area, A, is given
by
tDA =0.00633 .....M... (C-2)
flJlc,A
and the dimensionless time based on wellbore.radius. rw. is given
by
tD = 0.00633 -In.- (C-3)

Combining either definition of dimensionless time (Eq. C-2 or C-
3) yields the following expression for the dimensionless decline
time
l.In[ 2,% ] (C-4)
2 eYCAr2
wa
Similarly the definition of the dimensionless "decline" flow rate is
given by
qDd =141.2 In L (C-5)
kh eY CA,..2 I1p
wa
McCrayS defined the dimensionless "decline" integral flow rate
function as
qDdi =t1;f:frj qDd dtDd (C-6)
and McCrayS also defined the dimensionless "decline" flow rate
integral derivative function as
_ dqDdi
qDdid-- d In{fIJd) - tDd dtDd (C-7)
The dimensional forms of these relations are given by
qDdi= 4A 2 ](qII1
P
)i (C-8)
eYCAr
wa
and
qDdid= 4A ] (qII1P)id (C-9)
kh
where
-
(q,I1
P
)i=tf: d (C-IO)
and
(qII1P)id=- (C-II)
dint dt
Type Curve Matchinl: Procedure:
This procedure assumes that we have accurate measured rates and
pressures as a function of time. Unfortunately, pressures are
usually not available, so for the purposes of analysis and
interpretation, we may have to assume a constant pressure drop
term,l1p=p-Pwf' where Pwf is assumed to be constant with time.
The assumption of a constant pressure drop poses little difficulty
in the analysis--although this assumption may cause errors in
interpretation.
1. Compute the material balance time function from the
production rate data. This function is given by
i=Nplq (C-12)
2. Compute the flow rate and flow rate integral functions using
the material balance time function. These functions are
given by
(qll1p) ={Pi (C-l3)
Arps Dimensionless Rate Integral Derivative Relations
The definition of the dimensionless rate integral derivative
function, qDdid, which we assume to be positive, is given by
qDdid=- dqDdj - tDd tDd ....d.-{NpDd) .. (B-I9)
d In(tDd) dtDd dtDd tDd
Or if we use the defmition of the cumulative production function,
NpDd, we have
qDdid=- tDd .....d...-[...Lfctn1qDJ:r) d
r
]
dtDd fIJd 0
Expanding the derivative and reducing gives
fc
tDd
qDdid =t1Ji 0 qDrf.. r) dr- qDd , (B-20)
Combining Eqs. B-I2 with B-20 gives the most useful definition
of the dimensionless rate integral derivative function, qDdid. This
result is given as
qDdid=qIJdj-qDd (B-2I)
Applying Eq. B-21 to our previous results for the qDdi functions
yields

Rate lntel:ra1 Deriyative Relation


qDdid =...L[1- exp(-tDd)] - qDd
tDd
............................ (B-22)
Hyperbolic: (O<b<I) qDdid = ....L[...L_
qDd

I-b tfrj tfrj
............................ (B-23)
qDdid =_I_In (I +tDd) - qDd'" (B-24)
tDd
APPENDIX C PROCEDURE FOR THE ANALYSIS
OF PRODUCTION DATA USING THE FETKOVICW
MCCRAY TYPE CURVES
In this appendix we develop analysis relations for the
Fetkovich/McCraylO type curves. In order to generalize the
analysis concept for application to non-circular reservoir shapes
we have defined modified expressions for dimensionless
"decline" variables. Recall that Fetkovich,6 as well as later efforts
(refs. 7-10), all consider the case of a bounded circular reservoir.
While this solution is usually acceptable for analysis of production
data from vertical wells, we must understand how to interpret
performance responses from non-circular reservoir shapes. The
use of the reservoir shape factor, CA. permits interpretation of
other reservoir geometries.
Starting with the dimensionless "decline" time function, we have
tDd 1. [ )4A ] 2
tr
tDA 1..l1..In [J 4A ] 2trtD
2 In 2 ra
................................................................... (C-l)
18
Decline Curve Analysis Using Type Curves-Analysis of Oil Well Production Data Using Material Balance Time: SPE 28688
Application to Field Cases
-
(qlt1P)i =tJ: d-r ........................... (C-14)
( )
. _ d[(qlt1P)i] _ -d[(qlt1P)i]
qlt1p id - - dIn{ij - - t di (C-15)
A minor computational issue is that the data must be
in terms of the i function for proper calculation of the
integral and integral derivative functions.
3. Plot qlt1p, (qlt1P)i, and (qll1p)id versus i on a scaled log-log
grid. Force match the data trends onto the Arps b= 1
(harmonic) stem on the FetkovichlMcCray type curve.
Record the "time" and "rate" axis match points as well as the
matched transient rtf) stem.
Estimation of Original Oil-in-Place:
If the material balance time function, i, is correctly calculated then
a scaled log-log plot of qll1p versus i will exactly overlay the qOd
versus tOd trend for a harmonic decline on the Fetkovichl
McCraylO type curve. Once a match of the data and the type curve
has been obtained, the time and rate axis match points can be used
to develop the following relations for bpsS' and N
bpss = (C-16)
-1- (ilMJ, (qlt1plMp
N - c, (tllilMp (qllilMp (C-17)
where M.P. refers to the "match point" value.
We then may solve for the drainage area using the estimate of
original oil-in-place
A =5.6148 NB (C-18)
tPh(I-Swirr) .
where the effective drainage radius, re, can be estimated from the
following identity
re= V"f (C-19)
From the rate match point, we can solve for the formation
permeability, k
k-1412!lE..l.ln[ 4A 1[(qII1
P
)MP]
- . h 2 (qllilMp (C-20)
From the match of the data on a particular transient stem (a unique
value of rtf), we can solve for the effective wellbore radius, rwa,
and the skin factor, s. These estimates are obtained using
rKt'l= ..!L. (C-21)
reD
and
s = - (C-12)
SPE 28688 L.E. Doublet, P.K. Pande, T.J. McCollum, and T.A. Blasingame
19
10'
10'
10'
l'
2000
i
.-
1000
3000
3000
SIrooIated with ShIt-n)
(qAP)..,I(qoJ"PII 0.00888 STBIO/pIl
(I..,)"p!(IoJ... 1270.8 Days
1/0.00888 =112.6 psUSTBIO
NCt. 11.2842 STB/pIt _. 47.000 STB
q
10'
10'
1000
Figure 10 Match of Production Data for Simulated Case '2
Variable Pwtwith Shut-ina Radial Flow Type Curve.
10 .....-flT......,,r+--..,..T"'r'l'Tl"lrYt_
10
10' -:i-----+-----ir--....
(I)
fi
10'
t,.-Nlq. Days
Rgure 8 - Rate Functions for Simulated Case #2 (Variable pwfwith Shut-ins).
10" l.-...........a..L""""""................................. ..........--I........"""'"".................... .......... 10'
10"' 10.
'
10' 10' 10' 10' 10'
Figure 9 Match of Production Data for Simulated Case" 1
Constant Pwt Radial Flow Type Curve.
w"' w
4
10'
10' 10'

Type.Cutvel
Cira*I, ReseNoir Solutions .
iii
__.......
10'
10
3
""III---....__--"' ...__..a...-__........__ 4000
10
2
"Days
Rgure 7 - Log-Log Production Plot for Simulated Case #2 (Variable Pwf with Shut-ins).
2000
t. Days
Figure 6 Semilog Production Plot for Simulated Case #2 (Variable Pwt with Shut-ins).
1200
1000
800
600

i'

200
10'
10'
'10"
(I)
10"
t
1
LL 10-3
10"

10' 10"
q
10'
10 -+-____._'II""I""W'.....,.;.-,....,.-rT'I'Yft'r--_r_r-T"Il'Tftri--r-T"TT'T'I'm-
10
10 10' 10' 10
3
10 10
5
t,.-N/q.[)eys
Figure 5 - Rate Functions for Simulated Case'1 (Constant prJ.
10
2
"Days
Rgure 4 - Log-Log Production Plot for Simulated Case #1 (Constant Pwf).
.. ....,.L;;,w....1...... ....... ..
Figure 1 Fetkovich qDd and qDdd Type Curves.
10' 10'
Rgure 2 FetkovichlMcCray q()dl qDd/t and qDdld Type Curves.
10'
(I)
fi

10 -+----.......----.,.-----.------.------.----__+_ 0
500 1000 1500 2000 2500 3000
"Days
Figure 3 - Semilog Production Plot for Simulated Case #1 (Constant pwJ.
10' .,... --"' --1 110- 01.- +
-I-..... .......__,....,.............
10'
10'
J
1
10
'

10"
10"
10'" 10"

(I)' r----- _
fI
,0
' J _
20 Decline Curve Analysis Using Type Curves--Analysis of Oil Well Production Data Using Material Balance Time:
Application to Field Cases
SPE 28688
BLK.
SEC. 8
I
1901
I
___ J
A- .- .-C .-.
- '-wc'- .- :
.-:
":' : lie. II
.. - - ..... --L ;...
A. .-. :
f\oi
4201
____ ,. 1701
170"-
1604 I I
.4202 I
'----
FIgUre 18 NRU Well 4202 - section 8 of North Robertson Unit
Figure 17 North Robertson (C1earfork) Unlt. Gaines County. TX
I
J
I
1
I
L--
50000
50000
50000
20000 30000
N,.STB
20000 30000
N,.ST8
10000
10000
10000 20000 30000
N,.STB
Figure 11 - Movable Oil Estimation from Rate History.
Agora 13 - Movable_Oil from p". .
I
Simulated Cue .2 I
(Variable P."with Shut"'.)
o
100
O-+--__-..,...-..,...-..,...-..,...-.......--r---.-----aw---t-
0.000 -+---T"""--T"""-.....- .....- .....--r--.....
o
Figure 12 - Movable Oil Estimation from Nonnalized Rate History.
10000 +-.......- ......- ......- ......
0.020 -t--....a....-....a....---------------r::===::::::::===;r
.. 8000
a.
J6000 PI_ -4000 psia
l.aoo
2000
.. 0.015
!0.010

lj. 0.005
10'
10 -+-_.....,.'TTt"n+---r-....... __ .......
10'
10'
0-
2000 4000 6000 8000 10000 12000
"Days
Figure 19 Semilog Production Plot for NRU Well 4202 (Clearfork).
10
3
I.Daya
Figure 20 - Log-log Production Plot for NRU Well 4202 (Clearfork).
10 -t--......---t---.---+--.,..--t----r----+--....,.--+-__---f-
102 -:t----'---+--......
10'
0-
50000
50000
50000
40000
40000
20000 30000
N,.STB
20000 30000
N,.STB
10000
10000
10000 20000 30000
N,.STB
Rgure 16 - Movable Oil Estimation from Calculated p". .
Figure 14 - Movable Oil Estimation from Rate History.
O-t--..,...-..,...-..,...-__- __-..,...-__
o
Agure 15 - Movable Oil Estimation from Normalized Rate History.
0.000 -t--.....- .......- .......- .....- .....--r--.......
0.020 -r-..L--I.----I---.--L.------..&;:=;;;;====;:==:::;t-
10000
.. 0.015
!0.010

lj. 0.005
SPE 28688
L.E. Doublet, P.K. Pande, T.J. McCollum, and T.A. Blasingame
21
10'
10'
12000 10000
10
2
I. Days 10
3
Est. PrlrnItty Movable 01
105.000 STB
10 .....--.-...... __--.-.......""""'"-..............."'I""I"I'I....
10'
2000
Figure 27 - Rate Functions for NRU Well 1004 - (ClearfOf1<).
150000 200000 250000
N,.ST8
Figure 29 - Movable Oil Estimation 'rom Rate History.
6000 8000
I. Days
Figure 25 Semilog Production Plot for NRU Well 1004 (Clearfork).
10' .......--_t:_
(I)
c:i

10'
20 Est. Secondary MovabI.
o-+-.......r--'I'....... =to75_.000r-/ ........ ST ....B....,...-.--....-.,....+-
o
60
Figure 26 - Log-log Production Plot for NRU Well 1004 - (Clearfork).
10'
ti
10' ::-::;;.5:-:.:....,=.::1._
(t...JwI(toJ..,.. 2000 08.,..
b".- (qoJwI(qAIfI.... 110.013- 76.92 pIiIST'M)
Nc,. 26.0 STB/p8I 105.000 STB
I
Figure 28 Match of Production Data for NRU Well 1004 Clearfork (RadiaJ Flow Type Curve).
10" """'--""--"..........u.u&_"--I......... .......... .......__"'""""_'I.IoIo"I.Iw.u...a................... 10"
10" 10" 10
a
10' to- 10' 10' 10' 10'
J
I 10'
J ,..,.....
J
10'
10' 10'
Est. Seoondaty Movable 01
= 130,000 STB
\
10'
Est. Prirntlty Movable Oil
190.000 sm
10"
200000 300000 ..00000
N".STB
Figure 23 Movable Oil Estimation from Rate History.
10
3
10
r.,rN/q. Days
Figure 21 - Rate Functions for NRU Well 4202 (Clearfort<).
j;0l
&,301
&:;0'
-'92

.c.
J07 JOB
,(02
.coo



&,3OJ

.cO-l

103- c.

OJ
BOJ 207
20n

29J


&,201
m
2
\ '0J
0-4-
8O",
10,(, 209 210


&,102 !A20{ t;.
200
295

'000
1007 1006
100!)



1001 100?
1005
100,(
I


1003

_______________ J
Figure 2.-. NRU Wen 1004 - section 8 01 North Robertson Unit
-I-__ .........J......,........
10'
100
80
41\.
.
.

. ..
20
0
..
10" .....
fI)
I
u.. 10
4

!

Rgure 22 Match of Production Data for NRU Well 4202 (Clearfork) Radial Flow Type Curve.
10' 10'
J
'I 10'
"i'
J
J
10" 10"
22 Decline Curve Analysis Using Type Curves-Analysis of Oil Well Production Data Using Material Balance Time:
Application to Field Cases
SPE 28688
1500
1200
900

600

.
300

200 400 600 800 1000 1200
t.Days
Figure 35 - Semilog Production Plot for Barton Ughtsey Well 64 (Austin Chalk).
2000 4000 6000 8000 10000 12000 1.cooo
I. Days
Figure 30 - Semilog Production Plot for Well A (Spraberry).
1oJ -:t--.a................"'+---................ ..............It-...a..--............
101 10
10'
10'

102
en
fI
f:i
10'
10
10

f:i
10
2

. .
10 -+-__.--'I"I"I+-.......""'I""Y'ftft.........-P""'I"'I,....,......__"r"I'.....-.....
10' 10'
I. Days
Figure 31 - Log-Log Production Plot for Well A (Spraberry).
10' ...
en
f:i
10'
10'
10'
10'
10' __.............M-
10
10
4

10 10
2
10'
twrcNJq. Days
Figure 37 - Rate Functions for Barton Ughtsey Well 64 (Austin Chalk).
I
10" ..
en
..
c
t
at 10-2
! ..
STBIOIpSl
STBlOlpsi
6 STBIOlpsi
102
I. Days
Figure 36 - Log-Log Production Plot for Barton Ughtsey Well 64 (Austin Chalk).
10'
loJ
t,.,-N/q. Days
Figure 32 - Rate Functions for Well A (Spraberry).
-+-_-r-..........
10'
110
4

en
f
&&. 10-3

Figure 33 - Match of Production Data for Well A (Spraberry) Radial Flow Type Curve.
..
Est. Litr)
500
0.1000
1500
2000
2500 ;--'--'--'--'--'-............................
10" L-........ ........................_"'-"'......... ..................... ...... ........... 10"
10" 10' 10" lew 10' 10' 10' 10'
Figure 38 Match of Production Data for Barton Ughtsey Well 64
(Austin Chalk) Radial Flow Type Curve.
50000 150000 200000
N,. STB
Figure 34 Movable Oil Estimation from Rate History.

0. 20 Est. PrlfNty Movable 01
1: -+- -r- ...,.....,......,.....,..--..--..--.._,._,.._'6..,0_'OO-r\O_S..,T_B-r--r--r
300000 200000
N,.STB
Figure 39 - Movable Oil Estimation from Rate History.
SPE 28688 L.E. Doublet, P.K. Pande, T.J. McCollum, and T.A. Blasingame
23
Est. Movable OJ
- 100,000 STB
0-+-T""'T""T""T""'T""T"",..,.""T""...........
o 25000 50000 75000 125000
N".STB
Figure 46 - Movable Oil Estimation from Rate History.
1500
1000 ...., ....
. "'"',:
500
10
04
10" 10" 1et 1O' 10' 10'
10'
10'
B,oy!ts 1 ChaIt)
10'
plilSTBIO
1O'
He,. 26.88 STB/pli N-._. 100,000 STB
J
1
,0
,
1O'
J
J
10"
10"
10"
10"
10
04
1O" 1O" 10"
lew
10' 10' 10' 10'
Figure 45 Match of Production Data for S Carrabba Well 225
(Austin Chalk) Radial Flow Type Curve.
2000
EIJt. Movable 01 (Gas Uft)
- 360,000 STB
200000 300000
N,.STB
Agure 41 - Movable Oil Estimation from Calculated p"., .
0.0 -+-ooor-ooor-ooor-......-r-......
o 200000 300000
N".STB
Figure 40 - Movable Oil Estimation from Nonnalized Rate History.
8000
.!I 6000
!.
j p,-3326 psia
1

\
j
2000
0
A'
Est. Movable 01
100,000 STB
I
i
I
I
i
i

I I!
1'1
I

/
PI - 3326 psis

ru'
+ ';::'


o 25000 75000 125000 150000
N,.STB
Agure 48 - Movable 011 Estimation from Calculated p,. .
SANTA CLARA FIELD
LOWER REPETTO FORMATION
25000 50000 75000 125000 150000
N".STB
Figure 47 - Movable Oil Estimation from Nonnalized Rate History.
5OOO..,........................&.-L........-'-.........
0.40
..
i.1O
U)
0.20
Ci-
0.10
i 4000
j 3000
Q;
1
i 2000
1000

...


(I)
fi
10'
10
10J
twaN/q.Daya
Rgure 44 - Rate Functions for S Carrabba Well 225 (Austin Chalk).
t.Days
Agure 43 - Log-Log Production Plot for S Carrabba Well 225 (Austin Chalk).
tDays
Agure 42 - Semilog Production Plot for S Carrabba Well 225 (Austin Chalk).
10"
10"
g
(I)
i
&L 10" T";;:;;=====:::;t:;:::;==:;:;::::;:;t:;::;:;::;;---r-----r----,---;-

10 1500
10'
1200

900

101
1.
fi
eoo

10'
300
10 0
100 200 300 400 500
FIgure 49 Santa Clara (lower Repeno) Field - Offshore California
24
Decline Curve Analysis Using Type Curves--Analysis of Oil Well Production Data Using Material Balance Time:
Application to Field Cases
SPE 28688
SANTA CLARA FIELD
LOWER REPETTO At.
VERTICAL NET PAY MAP
LP-l. M. & N ZONES
O'L COLUMN ONLY
- ""... .-.- .....
---------- I
--=::::::::::- .. --
-------_... -.

_. __----
+ I.
....- ....
.....

................. --+------
........., I
..................... J ,I, ----
A
C
..::.._.... _
.: .
Agore 54 Match of Production Data for Gilda Well 5-42 (Lower Repetto) Radial Flow Type Curve.
figure 50 location or 5-42 within Santa Clara (lower Repetto) Ficld
Est. Primaty Movable 01
1.000,000 STB
Est. Primaty Movable 01
1.000.000 STB
500000 100000o 1500000 200000O
N".STB
Figure 55 - Movable Oil Estimation from Rate History.
O-t-..,.....,.....,.....,......,.....,.....,.....,......,;;:--r-......--r-T"-T"-T".....,......,......,......,.-i-
o
200
.. 0.15
( 0.10
t
0.05
. .
10
2

en
0-
2500
2000
l'
1500
1
1000 5"
500
10
1
0
500 1000 1500 2000 2500 3000
t.Days
Figure 51 - 5emilog Production Plot for Gilda Well 5-42 (Lower Repetto).
10
3
+--L...... .....
103 -:tT---...a..----I..----L----I..----L----I- 3000
2000000 1500000 500000 100000o
N".STB
Figure 56 - Movable Oil Estimation from Normalized Rate History.
10
1

10'
102
P, 5900 psis
500000 100000o 1500000 2000000
N" STB
Rgure 57 - Movable Oil Estimation from Calculated PtMr.
....
"

o
.!! 12000
!.
j9000

3000
10
3
'..-NIq.Daya
Rgure 53 .. Rate Functions for Gilda Well 8-42 (Lower Repetto).
Figure 52 - Log-Log Production Plot for Gilda Well 5-42 (Lower Repetto).

You might also like