CHAPTER 5

THE BASIC DIFFERENTIAL EQUATION FOR RADIAL FLOW I N A

POROUS MEDIUM
5.1 INTRODUCTION
In this chapter the basic equation for the radial flow of a fluid in a homogeneous
porous medium is derived as
This equation is non-linear since the coefficients on both sides are themselves
functions of the dependent variable, the pressure. In order to obtain analytical
solutions, it is first necessary to linearize the equation by expressing it in a form in
which the coefficients have a negligible dependence upon the pressure and can be
considered as constants. An approximate form of linearization applicable to liquid
flow is presented at the end of the chapter in which equ. (5.1) is reduced to the
form of the radial diffusivity equation. Solutions of this equation and their applica-
tions for the flow of oil are presented in detail in Chapters 6 and 7. For the flow of
a real gas, however, a more complex linearization by integral transformation is
required which will be presented separately in Chapter 8.
5.2 DERIVATION OF THE BASIC RADIAL DIFFERENTIAL EQUATION
The basic differential equation will be derived in radial form thus simulating the
flow of fluids in the vicinity of a well. Analytical solutions of the equation can then
be obtained under various boundary and initial conditions for use in the description
of well testing and well inflow, which have considerable practical application in
reservoir engineering. This is considered of greater importance than deriving the
basic equation in Cartesian coordinates since analytical solutions of the latter are
seldom used in practice by field engineers. In numerical reservoir simulation, how-
ever, Cartesian geometry is more commonly used but even in this case the flow into
or out of a well i s controlled by equations expressed in radial form such as those
presented in the next four chapters. The radial cell geometry is shown in fig. 5.1 and
initially the following simplifying assumptions will be made.
a) The reservoir is considered homogeneous in all rock properties and isotropic
with respect to permeability.
b) The producing well is completed across the entire forn.ation thickness thus
ensuring fully radial flow.
132 RADIAL DIFFERENTIAL EQUATION FOR FLUID FLOW
c) The formation i s completely saturated with a single fluid.
Fig. 5.1 Radial flow of a single phase fluid in the vicinity of a producing well.
Consider the flow through a volume element of thickness dr situated at a distance
r from the centre of the radial cell. Then applying the principle of mass conserva-
tion
Mass flow rate - Rate of change of mass in
OUT - the volume element
-
Mass flow rate
IN
aP
at
= 2nrhq5dr -
where 2nrh$dr is the volume of the small element of thickness dr. The left hand
side of this equation can be expanded as
which simplifies to
(5.2)
By applying Darcys Law for radial, horizontal flow it is possible to substitute for
the flow rate q in equ. (5.2) since
giving
or
RADIAL DIFFERENTIAL EQUATION FOR FLUID FLOW
133
The time derivative of the density appearing on the right hand side of equ. (5.3) can
be expressed in terms of a time derivative of the pressure by using the basic thermo-
dynamic definition of isothermal compressibility
and since
then the compressibility can be alternatively expressed as
m aP P aP
and differentiating with respect to time gives
aP aP
at at
cp- =
Finally, substituting equ. (5.5) in equ. (5.3) reduces the latter to
(5.4)
(5.5)
This is the basic, partial differential equation for the radial flow of any single phase
fluid in a porous medium. The equation is referred to as non-linear because of the
implicit pressure dependence of the density, compressibility and viscosity appearing
in the coefficients kpl p and 'cp. Because of this, it is not possible to find simple
analytical solutions of the equation without first linearizing it so that the coeffi-
cients somehow lose their pressure dependence. A simple form of linearization
applicable to the flow of liquids of small and constant compressibility (under-
saturated oil) will be considered in sec. 5.4, while a more rigorous method, using the
Kirchhoff integral transformation, will be presented in Chapter 8 for the more
complex case of linearization for the flow of a real gas.
5.3 CONDITIONS OF SOLUTION
In principle, an infinite number of solutions of equ. (5.1) can be obtained depending
on the initial and boundary conditions imposed. The most common and useful of
these is called the constant terminal rate solution for which the initial condition is
that at some fixed time, at which the reservoir i s at equilibrium pressure pi, the
well is produced at a constant rate q at the wellbore, r=rw. This type of solution
will be examined in detail in Chapters 7 and 8 but it i s appropriate, at this stage, to
describe the three most common, although not exclusive, conditions for which the
constant terminal rate solution is sought. These conditions are called transient, semi
steady state and steady state and are each applicable at different times after the start
of production and for different, assumed boundary conditions.
134 RADI AL DI FFERENTI AL EQUATI ON FOR FL UI D FLOW
a) Transient condi ti on
This condition is only applicable for a relatively short period after some pressure
disturbance has been created in the reservoir. In terms of the radial flow model this
disturbance would be typically caused by altering the wells production rate at
r=r,. In the time for which the transient condition is applicable it is assumed that
the pressure response in the reservoir is not affected by the presence of the outer
boundary, thus the reservoir appears infinite in extent. The condition is mainly
applied to the analysis of well tests in which the wells production rate is deliberate-
ly changed and the resulting pressure response in the wellbore is measured and
analysed during a brief period of a few hours after the rate change has occurred.
Then, unless the reservoir is extremely small, the boundary effects will not be felt
and the reservoir is, mathematically, infinite.
This gives rise to a complex solution of equ. (5.1) in which both the pressure and
pressure derivative, with respect to time, are themselves functions of both position
and time, thus
P = g(r,t)
and - - - f(r,t)
at
Transient analysis techniques and their application to oil and gas well testing will be
described in Chapters 7 and 8, respectively.
b) Semi Steady State condi t i on
= 0, a t r I r
Fig. 5.2 Radial flow under semi steady state condi ti ons.
This condition is applicable to a reservoir which has been producing for a sufficient
period of time so that the effect of the outer boundary has been felt. In terms of the
radial flow model, the situation is depicted in fig. 5.2. I t is considered that the well
i s surrounded, at i ts outer boundary, by a solid brick wall which prevents the flow
of fluids into the radial cell. Thus at the outer boundary, in accordance with Darcys
law
RADIAL DIFFERENTIAL EQUATION FOR FLUID FLOW
- 0 at r=r,
aP
ar
- -
135
(5.6)
Furthermore, if the well is producing at a constant flow rate then the cell pressure
will decline in such a way that
-
constant, for all r and t.
at
(5.7)
The constant referred to in equ. (5.7) can be obtained from a simple material
balance using the compressibility definition, thus
which for the drainage of a radial cell can be expressed as
(5.1 0)
This is a condition which will be applied in Chapter 6, for oil flow, and in Chapter
8, for gas flow, to derive the well inflow equations under semi steady state con-
ditions, even though in the latter case the gas compressibility is not constant.
One important feature of this stabilized type of solution, when applied to a deple-
tion type reservoir, has been pointed out by Matthews, Brons and Hazebroek' and is
illustrated in fig. 5.3. This is the fact that, once the reservoir is producing under the
semi steady state condition, each well will drain from within its own no-flow bound-
ary quite independently of the other wells.
For this condition dp/dt must be approximately constant throughout the entire
reservoir otherwise flow would occur across the boundaries causing a re-adjustment
in their positions until stability was eventually achieved. In this case a simple tech-
nique can be applied to determine the volume averaged reservoTr pressure
(5.1 1)
in which
Vi
and pi
Equation (5.9) implies that since dp/dt is constant for the reservoir then, if the
variation in the compressibility is small
= the pore volume of the ith drainage volume
= the average pressure within the ith drainage volume.
136 RADI AL DI FFERENTI AL EQUATI ON FOR FL UI D FLOW
Fig. 5.3 Reservoir depl eti on under semi steady state condi ti ons.
qi Vi (5.12)
and hence the volume average in equ. (5.1 1 ) can be replaced by a rate average, as
follows
(5.13)
and, whereas the Vi's are difficult to determine in practice, the qi's are measured
on a routine basis throughout the lifetime of the field thus facilitating the calcula-
tion of Pres, which i s the pressure at which the reservoir material balance i s evaluate-
ed. The method by which the individual pi's can be determined will be detailed in
Chapter 7. sec. 7.
c) Steady State condi t i on
t q = const ant
Pe I const ant
6 fluid Influx
Pwf -
W
r -
e
Fig. 5.4 Radial f l ow under steady state condi ti ons.
RADIAL DIFFERENTIAL EQUATION FOR FLUID FLOW
137
The steady state condition applies, after the transient period, to a well draining a cell
which has a completely open outer boundary. I t i s assumed that, for a constant rate
of production, fluid withdrawal from the cell will be exactly balanced by fluid entry
across the open boundary and therefore,
p = pe = constant, at r =re (5.14)
and
ap = 0 forall r and t
at
(5.1 5)
This condition i s appropriate when pressure is being maintained in the reservoir due
to either natural water i'nflux or the injection of some displacing fluid (refer Chapter
I O).
I t should be noted that the semi steady state and steady state conditions may never
be fully realised in the reservoir. For instance, semi steady state flow equations are
frequently applied when the rate, and consequently the position of the no-flow
boundary surrounding a well, are slowly varying functions of time. Nevertheless, the
defining conditions specified by equs. (5.7) and (5.15) are frequently approximated
in the field since both production and injection facilities are usually designed to
operate at constant rates and it makes little sense to unnecessarily alter these. I f the
production rate of an individual well is changed, for instance, due to closure for
repair or increasing the rate to obtain a more even fluid withdrawal pattern through-
out the reservoir, there will be a brief period when transient flow conditions prevail
followed by stabilized flow for the new distribution of individual well rates.
5.4 THE LINEARIZATION OF EQUATION 5.1 FOR FLUIDS OF SMALL AND
CONSTANT COMPRESSIBI Ll TY
A simple linearization of equ. (5.1) can be obtained by deletion of some of the
terms, dependent upon making various assumptions concerning the nature of fluid
for which solutions are being sought. In this section the fluid considered will be a
liquid which, in a practical sense, will apply to the flow of undersaturated oil. Exp-
anding the left hand side of equ. (5.1), using the chain rule for differentiation gives
I[?(!)
aP raP +k~ r2_P+kPr C P ] = +cpZ aP
r ar p pr % p ar ar p ar p ar*
and differentiating equ. (5.4) with respect to r gives
which when substituted into equ. (5.16) changes the latter to
(5.16)
(5.1 7)
-1-(-) l a k pr- a p k +- cpr ($)
r ar p ar p p ar p ar2
- +c pg (5.18)
138 RADI AL DI FFERENTI AL EQUATI ON FOR FL UI D FLOW
For liquid flow, the following assumptions are conventionally made
- the viscosity /.L is practically independent of pressure and may be regarded as a
constant
- the pressure gradient ap/ar is small and therefore, terms of the order (ap/ar)*
can be neglected.
These two assumptions eliminate the first two terms in the left hand side of equ.
(5.181, reducing the latter to
which can be more conveniently expressed as
(5.19)
(5.20)
Making one final assumption, that the compressibility i s constant, means that the
coefficient &c/k is also constant and therefore, the basic equation has been
effectively linearized.
For the flow of liquids the above assumptions are quite reasonable and have been
frequently applied in the past. Dranchuk and Quon, however, have shown that this
simple linearization by deletion must be treated with caution and can only be
applied when the product
cp <<1 (5.21)
This condition makes it necessary to modify the final assumption so that the com-
pressibility is not just constant but both small and constant. The compressibility
appearing in equ. (5.20) i s the total, or saturation weighted, compressibility of the
entire reservoir-liquid system
C, = coso +c, s, , +Cf
(5.22)
in which the saturations are expressed as fractions of the pore volume. Using typical
figures for the components of equ. (5.22)
c, = 10xIO-6/psi s,, = 0.2
c, = 3~1O- ~/psi p = 3000psi
cf = 6~10- ~/ ps i
then c, in equ. (5.22) has the value 14.6~10-~/psi and the product expressed by
equ. (5.21) has the value 0.04, which satisfies the necessary condition for this
simple linearization to be valid. However, when dealing with reservoir systems which
have a higher total compressibility it will be necessary to linearize equ. (5.1) using
some form of integral transformation as detailed by Dranchuk and Quon. Such an
approach will be required when describing the flow of a real gas since, in this case,
RADIAL DIFFERENTIAL EQUATION FOR FLUID FLOW 139
the compressibility of the gas alone may, to a first approximation, be expressed
as the reciprocal of the pressure and the cp product, equ. (5.21), will itself be
unity. The linearization of equ. (5.1) under these circumstances will be described in
Chapter 8, secs. 2-4.
Before leaving the subject of compressibility, it should be noted that the product of
r$ and c in all the equations, in this and the following chapters, is conventionally
expressed as
(5.23)
since it was assumed in deriving equ. (5.1) that the porous medium was completely
saturated with a single fluid thus implying the use of the absolute porosity. Alterna-
tively, allowing for the presence of a connate water saturation, the @c product can
be interpreted as
(5.24)
in which (l-Swc) is the effective, hydrocarbon porosity, and the com-
pressibility is equivalent to that derived in Chapter 3, equ. (3.19), which is used in
conjunction with the hydrocarbon pore volume. In either event, the products ex-
pressed in equs. (5.23) and (5.24) have the same value, the reader must only be
careful not to mix the individual terms appearing in the separate equations.
Equation (5.20) is the radial diffusivity equation in which the coefficient k/r$pc
is called the diffusivity constant. This is an equation which is frequently applied in
physics, for instance, the temperature distribution due to the conduction of heat in
radial symmetry would be described by the equation
in which T is the absolute temperature and K the thermal diffusivity constant.
Because of the general nature of equ. (5.20) it is not surprising that many reservoir
engineering papers, when dealing with complex solutions of the diffusivity equation,
make reference to a text book entitled "Conduction of Heat in Solids", by Carslaw
and J aeger3, which gives the solutions of the equation for a large variety of bound-
ary and initial conditions and is regarded as a standard text in reservoir engineering.
REFERENCES
1)
Matthews, C.S., Brons, F. and Hazebroek, P., 1954. A Method for Determination of
Average Pressure in a Bounded Reservoir. Trans. AIME. 201: 182-191.
Dranchuk, P.M. and Quon, D., 1967. Analysis of the Da'rcy Continuity Equation.
Producers Monthly, October: 25-28.
Carslaw, H.S. and J aeger, J .C., 1959. Conduction of Heat in Solids. Oxford at the
Clarendon Press, (2nd edition).
2)
3)