REVIEWAND DISCUSSION'

Physical Principles of Oil Production.By MOR- the volume of water, Q, flowing downward
RIS MUSKAT.New York: McGraw-Hill Book through this sand in unit time was given by
Co., Inc., 1949. Pp. xv+922. $15.00. h2- hi
This book, issued under the auspices of Q= -KA- (1)
L
McGraw-Hill's "International Series in Pure
and Applied Physics," is in part a revision and where h1and h2are the heights above a standard
in part an extension of the author's earlier book, datum of the water in input and output manom-
The Flow of HomogeneousFluids throughPorous eters, respectively, and K is a constant of pro-
Media, published in i937. Whereas the earlier portionality. Expressing this differentially, the
book was restricted to homogeneous fluids, the flow per unit area and unit time is given by
present volume has been extended to include the dh
flow of multiphase fluids (oil, gas, and water) v = - K dL" dL (2)
also.
As to its contents, chapters i-iii are devoted Darcy also observed that this relation failed at
to the physical properties of petroleum fluids sufficiently high rates of flow.
and to pertinent properties of reservoir rocks; In I880 Seelheim, by studying the flow
chapters iv-vi present the theory of the flow of through sands of different grain sizes, found that
single-phase fluids through porous solids; chap-
ters vii-xi deal with multiphase flow problems; v oc d2 (3)
and the concluding chapters (xii-xiv) comprise
where d is the mean grain diameter. To this
such miscellaneous topics as secondary recov-
Slichter (1899) added the effect of the fluid
ery, optimum well spacing, and related prob-
viscosity, whereby
lems. In addition, some six to eight hundred
1
citations of published literature have been V oc - (4)
made.
Like its predecessor, this book contains an
Although Darcy's experiments were limited
enormous amount of otherwise widely dispersed
important to flow in the vertical direction, it was soon real-
information on nearly every aspect of
ized that equation (i) was of general validity for
petroleum exploitation. Its avowed purpose,
flow in any direction. This was given analytical
however, is the establishment of "those physical
for general three-dimensional flow by
principles controlling the behavior of oil fields expression
Forchheimer (1914, p. 437), who expressed
and their implications when applied under con-
ditions broadly simulating those which may equation (i), which had long since become
obtain in oil-producing reservoirs occurring in known as "Darcy's law," in the vectorial form
nature." Accordingly, the book is to be ap- v = - KVh . (5)
praised principally on the degree of its success in
achieving this objective. Since the fundamental This signifies that the flow lines are everywhere
relation involved in the flow of fluids through normal to a family of surfaces h = Constant
porous rocks is that known as "Darcy's law," and in the direction of - Vh, the hydraulicgra-
a brief historical account of this relation will be dient. He also pointed out that (for K constant),
useful as a basis for judgment in considering the V2h = o.
treatment that has been employed. Equation (5) may also be expressed in terms
In I856 Henry Darcy studied the flow of of the pressure gradient in the flow field by the
water through filter sands. Using a vertical use of the hydrostatic equation
cylinder containing a sand pack of cross-sec- ~ = rg (h - z), (6)
tional area A and thickness L, he observed that
where p is the fluid pressure at a given point of
'Manuscript received May 17, 1950. elevation z above a standard datum, h the
655

This content downloaded from 154.059.124.102 on October 06, 2016 11:56:23 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
656 REVIEW AND DISCUSSION

height with respect to the same datum of the in three dimensions and are entirely equivalent
fluid in a manometer terminated at the point of physically to Darcy's original formulation.
interest, 7 the fluid density, and g the accelera- Moreover, they are based exclusively upon work
tion of gravity. Solving equation (6) for h gives done before 1914. The hydraulic-gradientequa-
tion is obviously applicable only to liquids, but
h= z+ , the last two forms are valid also for gases.
In the petroleum industry an awareness of
of which the directional derivative is the importance of Darcy's law as a basis for
studies of fluid flow did not become widespread
ah _ Oz 1 p (7 until about 1930, and the usage followed in the
+gas (7)
os ts gyg ds' book under review dates from a paper by
Wyckoff, Botset, Muskat, and Reed (I933) on
or the gradient is the measurement of permeability. In this paper
1 the pressure-gradientequation,
Vh = Vz+- Vp. (8)
yg k ap
v,= ,' (13)
Now if we multiply both sides of equation
(8) by -g and note that IVzl = i, we obtain was erroneously set forth as representing
1 Darcy's law (compare with eq. [12]). This same
- gVh = g -- p , (9) paper also became the basis for Code 27 on
the measurement of permeability, adopted in
which is a vectorial statement of the forces per 1935 by the American Petroleum Institute,
unit of mass acting upon the fluid at a given which, in effect, made equation (13) official for
point. Multiplying by -yg gives the petroleum industry as a supposed statement
of Darcy's law.
--7 gVh = g - Vp , (10) This equation was given further currency
when Muskat's book, Flow of Homogeneous
which are the corresponding forces per unit of Fluids throughPorous Media, appeared in 1937.
volume.It will be noted, in particular, that the There, after correct statement of Darcy's law
fluid-flowdirection, -Vh, is, in general, parallel (p. 56) as given by equation (i) above, the pres-
to neither of the vectors g or -Vp. sure-gradient equation (p. 71) was again put
Now if we combine equation (9) or (io) with forward as representing Darcy's law, and,
equations (3), (4), and (5) and discard the eliminating any possible doubt as to what was
undefined constant K, we obtain meant, the two (p. 72) were equated to each
Nd7 ( i 'v1 other:
v-= -(Kg--Vp)
k dp kygdh - dh
v k (14)
dx J dx dx"

According to these equations,

of which the only physically undefinedparame- dh 1 dp
ter is the residual factor of proportionality, N. (15)
dx yg dx'
A dimensional check will show that this is di-
mensionless and that it depends only upon the which, by comparison with equation (7), is seen
small-scale geometric shape of the flow system. to be significantly in error.
Combining the two geometrical factors Nd2into Although later in the same book (p. 129) a
a single factor k then gives flow equation of the correct form,

v=-KV Kh
= g -Ivp) ^- te+ +F
K=p,- )- (16)
(12)
= g- p) . was derived, the pressure-gradientequation was
repeatedly employed thereafter both in equa-
These are three different expressions of Darcy's tions and in the text.
law, all of which are of general validity for flow An alternative expression, supposedly ap-

This content downloaded from 154.059.124.102 on October 06, 2016 11:56:23 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 REVIEW AND DISCUSSION 657
plicable to any single-phase flow problem, was universallypresent as driving agents affectingthe
given by the velocity-potential equation (p. movementof oil, gas, and water, in the great ma-
129), jority of practical oil-productionsystems the pre-
dominatingdrivingforcesare thosedue to fluid pres-
v = -V , (17) sure gradients. . . . Accordingly the above statement
of Darcy's law implies that pressuregradientsand
where 4 = k/p(p - V) and V is the potential hydraulic heads are equivalent, interchangeable, and
of a field of force (per unit of volume), F. superposable
[italicsmine].
Critics (Richards, 1940; Hubbert, 1940,
In view of the correct relations between pres-
1941) soon pointed out that the pressure-gradi-
ent equation is not an expression of Darcy's law sure and hydraulic gradients set forth in equa-
and that, in view of the fact that fluids can flow tions (7) and (8) above, a more confused state-
from lower to higher pressures quite as readily ment concerning these quantities would be dif-
as in the reverse direction, it obviously is not ficult to make. As to the dominance of forces due
physically valid. They showed, moreover, that to pressure gradient in oil production, calcula-
the true Darcy's law is that expressed by equa- tions on this point indicate that, for an average-
tion (12) or any equivalent expression. It was producing oil well in the United States, the ratio
also shown (Hubbert, 1940, pp. 903-909) that in of the pressure-gradient force to that due to
all but special cases the velocity vector v is rota- gravity becomes of the order of unity at a dis-
tional and not derivable from a velocity poten- tance away from the well of less than o10feet.
tial. From the foregoing and similar statements it
The erroneous statement of Darcy's law in becomes evident that the same lack of under-
Code 27 was also called to the attention of standing which in 1933 led to the representation
the American Petroleum Institute, and in 1942 of the pressure-gradient equation as Darcy's
(A.P.I., 1942) that code was revised to the ex- law, and in 1937 to the writing of equation (14)
tent of abandoning the pressure-gradientequa- above, still persists. A correct statement of
tion and replacing it by the correct equation, Darcy's law (eq. [IS]), which, in substantially
that form, has been in use for half a century, is
v,=--k -a-'g cos 0 , (18) now presented as a new generalization of that
law, rendering it valid for flow in three dimen-
sions, whereas the pressure-gradient equation,
where 0 is the angle between s and the direction for which Muskat and his associates are solely
of gravity. responsible, is credited to Henry Darcy!
The next development in this sequence is the If the confusion resulting from calling unlike
book under review: Physical Principles of Oil things by the same name is to be avoided, it is
Production.Here the subject of fluid flow is in- apparent that a separate name for the pressure-
troduced (p. I24) not by the usual pressure- gradient equation is needed to distinguish it
gradient equation but by equation (i8) above. from Darcy's law. Since this equation owes its
This, however, is stated not to be an expression prominence to its extensive use by Muskat, it
of Darcy's law but a generalization of that law would
therefore appear fitting that the equation
(due to Muskat), rendering it valid for flow in
three dimensions and also applicable to different k
fluids. Darcy's original law, one reads on pages v= -- P ,
184 and 185, is given by
k or any of its physical equivalents, should here-
v= -- vp . (19) after be known as "Muskat's law."
This might be of only academic interest, were
This viewpoint is further elaborated in various it not for the fact that the physical confusion
parts of the text, of which the following passage indicated is recurrent throughout the book. On
(p. 124) is representative: page 193, for instance, it is stated that "the
nature of the fluid enters Darcy's law only
The second generalizationpertainsto the inclu-
through its viscosity," and, further, "the fluid
sion as measuresof the drivingforce both pressure
and hydraulicgradients.Naturally only the latter density does not enter in the 'law of force' equa-
wereconsideredby Darcy, who experimentedexclu- tion." On pages 194 and 195 the problem of the
sively with water.While,strictlyspeaking,in the oil flow through a uniform sand of constant thick-
industrygravitationalor fluid-headforcesarealmost ness into a well completely penetrating the sand

This content downloaded from 154.059.124.102 on October 06, 2016 11:56:23 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
658 REVIEW AND DISCUSSION

is discussed. It is stated that, since gravity ef- It is of general validity for flow of both compres-
fects are not involved, the system will be de- sible and incompressible single-phase fluids in
scribed by Laplace's equation in the fluid pres- three-dimensional space, and, were it consist-
sure; and the solution obtained comprises a ently and correctly employed, together with the
family of equipressuresurfaces, coaxial with the appropriate equations of state, it would lead
well and perpendicularto the lines of flow. This consistently to correct results.
result is physically false, and the actual equi- Equation (21), however, is seriously re-
pressuresurfaces (since gravity does enter in the stricted by the next step. A "velocity poten-
problem) are a family of coaxial trumpet-shaped tial," 1', is defined by
surfaces, divergent and convex upward, which
rapidly approach horizontality with increase of (22)
=k(p
/z + V) ,
distance from the well. This erroris repeated on
pages 242 and 243 in a discussion of the flow
whereby
with radial variations of permeability.
On page 540 the flow equation for a compres- v = -V< . (23)
sible liquid is stated to be:
Here V is the potential of the force F, and on
k k page i87 the value of V for incompressible and
v =--yvp=- vy, (20)
compressible liquids, respectively, is given by

according to which the flow lines would be V = yg z,

orthogonal with the equipressure and equiden- (24)
sity surfaces and the flow rate proportional to V= gfyd z .
the respective density and pressure gradients-
a result in accord with theoretical deductions on As contrasted with the statement in the ear-
pages 187 and 188 but still physically invalid. lier book that equation (23) is of general valid-
The pressure-gradientequation is again im- ity, it is now admitted to being valid only for the
plied in the statement on pages 764-765 con- flow of a fluid of constant viscosity in homoge-
cerning the analogy between electrical and fluid neous and isotropic media.
models: Since gravity is known to have a potential,
For homogeneoustwo-dimensionalsystems, of the potential V of the force F is likely to pass
constant thickness,the basic analogy between the unnoticed. It happens, however, that the gravi-
electricalmodeland the flowsystem simplyrests on tational-field intensity at a point can be defined
the observationthat the electricalpotentialcorre- uniquely only in terms of a force per unit of mass
spondsto the pressure[italicsmine]and the current and not in terms of force per unit of volume.It is
density to the fluidflux.
the former which in all cases has a potential.
When the density is constant, the force per unit
The theoretical development of the flow
of volume, F, also has a potential, as given by
equations and an explanation of the methods to the
first of equations (24). When the density is
be used in solving flow problems are presented
variable and the equidensity surfaces are not
in chapter iv. This begins with a statement of horizontal, which is, in general,
the case for a
the equation of continuity for fluids with dif-
fluid in motion, the force F has no potential,
ferent equations of state. Next, the equations of
since V in the second of equations (24) in this
Navier and Stokes are written but are discarded case
is a function of the path of integration.
on the grounds that they cannot be solved. Hence the inse:tion of V in the equations for
Finally, the "generalized"Darcy's law is given
compressible fluids on page 186 and elsewhere
(p. i85) as
is not justified.
k A closely related error occurs in several
v = --(Vp-F) , (21) places when the flow of multiphase fluids is dis-
cussed. For example, on page 274, the flow equa-
where F is a force ("such as gravity") per unit tion for the gas phase in a water-gas-oil mixture
volume. is given by
This is a correct statement of Darcy's law in
the form of the third of equations (12), above. v, = ---'Lu V (p- 7rg z) . (25)

This content downloaded from 154.059.124.102 on October 06, 2016 11:56:23 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
 REVIEW AND DISCUSSION 659
Since yg is variable, expansion of the gradient of (pressure, velocity potential, density, or their
the terms in parentheses gives normal gradients) are to be assigned. Then with
the appropriateequation of motion the problem
V (p - y7g z) = VP - 7gVz (26) is to be solved as a boundary-value problem.
- g zV7y , This procedureis sound, providedthat proper
which is significantly different from the correct initial and boundary conditions are assigned
force terms in equation (21).
and valid equations of motion are employed.
The importance of this provision is illustrated
On pages I86-190 the "equations of motion"
by the radial-flow problem of pages 194-195, al-
are obtained by combining the Darcy's law
ready mentioned. Here radial flow is assumed
equation with the equations of state and of con-
to start with, and the pressure is assumed cor-
tinuity. For incompressible fluids and homoge-
rectly to satisfy Laplace's equation. However,
neous media this leads (p. 187) to
by assigning an erroneous boundary value
V2p = 0 = V2I . (27) p = Const. in the well, an erroneous solution
for p in space is obtained, and this can then be
For compressible liquids the equation first reconciled with the assumed flow field only by
obtained (p. 187) is means of an erroneous flow equation.
(y± T
V1 f In the light of this recital it is evident that
+ 72 g z) k . (28) the problem of appraising the book under re-
view is complex. As to style and composition, it
Here Kis the compressibility defined by is meticulously written. The chapter on the
physical properties of petroleum fluids is good.
7 = 7o eP, (29) The mathematical treatment of boundary-value
problems, considered as abstract mathematics
without regard to the physical validity of the
and f is the porosity. Then by a circular argu- equations,
is also good. Moreover, in many in-
ment it is reasoned that the second term,
stances correct physical relations have been
y2gVz,of equation (28) is of a negligible magni- employed.
tude as compared with the first, V'y/K,leading
Despite all this, the fact remains that, with
to the desired result (p. i88),
respect to the physical theory of the flow of
fKg &7 fluids through porous rocks, a wrong turn was
V2 - k (30) taken by this author and his colleagues in 1933,
k t '
which was adhered to in his book The Flow of
This is the basis for equation (20), above, HomogeneousFluids throughPorousMedia. Dur-
quoted from page 540. A similar equation is ing the twelve years which have elapsed between
given for gases (p. 189) in terms of V y(i+m)/m, the publication of that book and its present
where m is the ratio of the specific heat at revision and extension, there has been ample
constant volume to that at constant pressure. opportunity to rectify earlier errorsand to bring
The fallacy in this procedure lies in the fact the physical theory to the same level of excel-
that the compressibility of a fluid is in nowise lence as the rest of the treatment. But this has
dependent upon the pressure gradient or the not been done.
state of motion. Hence an equation of motion In consequence of this default the usefulness
for a compressible fluid must retain its validity of the book is severely limited. For those who
when the fluid is at rest or in very slow motion, already have sufficient knowledge of the physics
which is the usual state for undergroundfluids. of fluid flow through porous rocks to supply the
Under these conditions the two terms of equa- necessary corrections, this work is a useful refer-
tion (28) are of comparablemagnitudes, and the ence; but for students and others who have not
fluid may flow in any direction in space with such knowledge and who, on the basis of its
respect to the direction of the density gradient. credentials and sponsorship, assume the book to
The procedure for solving flow problems is be an authoritative treatise, the result can
outlined on pages 189 and 190. Laplace's equa- hardly be other than the perpetuation of a state
tion ([27] above) is taken as the basis for solving of confusion which has already persisted too
problems in the steady-state flow of liquids. The long.
geometry of the space, initial pressureor density M. KING HUBBERT
distributions, and the boundary conditions Shell Oil Company

This content downloaded from 154.059.124.102 on October 06, 2016 11:56:23 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).
66o REVIEW AND DISCUSSION

REFERENCES CITED
AMERICAN PETROLEUM INSTITUTE (1935) Standard
procedure for determining permeability of porous
media (tentative), Am. Petroleum Inst. Code 27.
(1942) Standard procedure for determining
permeability of porous media, ibid.
DARCY,HENRY(1856) Les fontaines publique de la
ville de Dijon, Paris, Victor Dalmont.
FORCHHEIMER, PHILIPP (1914) Hydraulik, Leipzig
and Berlin, Teubner.
HUBBERT,M. KING (1940) The theory of ground-
water motion: Jour. Geology, vol. 48, pp. 785-
944.
--- (94) The theory of ground-water motion,
a reply: ibid., vol. 49, pp. 327-330o.
MUSKAT,M. (1937) The flow of homogeneous fluids
through porous media, New York, McGraw-Hill
Book Co., Inc.
RICHARDS,L. A. (1940) Concerning permeability
units for soils: Soil Science Soc. America Proc.,
vol. 5, PP. 49-53.
SEELHEIM,F. (I880) Methoden zur Bestimmung der
Durchliissigkeit des Bodens: Zeitschr. anal.
Chemie, vol. 19, p. 387.
SLICHTER, C. S. (1899) Theoretical investigation of
the motion of ground waters: U.S. Geol. Survey
i9th Ann. Rept., I897-I898, pt. 2, p. 301.
WYCKOFF,R. D.; BOTSET, H. G.; MUSKAT, M.; and
REED, D. W. (1933) The measurement of per-
meability of porous media for homogeneous
fluids: Sci. Inst. Rev., vol. 4, pp. 394-405.

This content downloaded from 154.059.124.102 on October 06, 2016 11:56:23 AM

All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).