Vous êtes sur la page 1sur 24

Section II/C, Paper 3

AVERAGE RESERVOIR PRESSURE


BY

D. R. HORNER ( * )

SYNOPSIS. I t i s believed that conventional methods for determining m e a n reservoir pressure


are incorrect in that t h q ignore the existence of localized pressure sinks surrounding the produ-
cing wells. I n consequence the quantity of oil initially in place in a reservoir will be over-estimated
if i t i s derived f r o m the material balance equation using a conventionally determined m e a n
reservoir pressure. T h e report therefore treats the material balance equation as a single functional
relation connecting two u n k n o w n s ( t a n k oil initially in place and m e a n reservoir pressure) and
derives a second relation connecting these same two quantities, based o n observed bottomhole pres-
sures and including a correction term f o r the eflect of the pressure sinks. Both functional rela-
tions are plotted o n the same graph and the point of intersection o f the two curves gives values
f o r the tank oil in place and the m e a n reservoir pressure. An example i s given of the applica-
tion of the method, which i s at present restricted to reservoirs w i t h n o initial gas-cap and n o ap-
preciable water-drive.

RI?SUMB. Suivant l’duteur, 1;s methodes ordinaires de determination de la pression moyenne


de gisement sont inexactes parce qu’elles n e tiennent p a s compte de l’existence, autour des p u i t s
de production, de zones caracterisees p a r des chutes locales de pression. De ce fait, la quantiti
d’huile minerale initiale d’un gisement donne sera estimee p a r excds s i l’on emploie l’dquation
de balance de materiel e n tenant compte d’une pression moyenne de g i s e m m t dCterminCe suivant
(( ))

les mkthodes habituelles. L e present rapport envisage l’equation de U material balance N comme u n e
equation fonctionnelle unique reliant deux inconnues (l’huile rdcuperable a u ddbut duns un gise-
ment et la pression nzoyenne de gisement) et obtient u n e deuxikme expression q u i relie d son
tour ces deux memes inconnues; cette dernikre expression, basee sur les pressions mesurees a u
fond d u puits, comprend un terme de correction tenant compte de l’effet des abaissements de pres-
sion autour des puits. Ces deux expressions fonctionnelles sont reportees sur le meme graphique et
le point d’intersection des deux courbes donne les valeurs de l’huile recuperable existant duns
un gisement mesurke a u x condition de surface, ainsi que la pression inoyenne de gisement. L e memoire
donne rcn exemple d’application de cette methode, q u i n e saurait 2tre adoptee d l’heure actuelle que
pour des gisements sans U gas-cap et n e produisnnt p a s p a r poussde de l’eau.
))

RIASSUNTO. I metodi comuni per la deterrninazione della pressione media d i un giacimento


Sono inesatti per il fatto che essi n o n tengono conto dell’esistenza d i zone attorno a i pozzi produttivi,
nelle quali s i hanno degli abbassamenti locali d i pressione. Di conseguenza la quantitd d i olio ini-
zialmente esistente in un giacimmto verra stimata per eccesso qualora s i impieghi l’equazione d i
U material balance D, adottando una pressione media d i giacimento determinata secondo i metodi
comuni. I1 presente rapporto tratta l’equazione d i material balance come una singola reazione
(( ))

funzionale che collega due incognite (l’olio ricuperabile inizialmente in giacimento e la pressione
media d i giacimento) e ricava u n a seconda relazione che coUega fra loro queste stesse d u e inco-
gnite, relazione basata sulle pressioni misurate a fondo pozzo e che comprende un termine d i cor-
rezione per tenere conto dell’effetto degli abbassamenti d i pressione attorno ai pozzi. Entrambe le
relazioni funzionali vengono riportate sul medesimo grafico e i l p u n t o d i intersezione delle due
curve favorisce i valori dell’olio ricziperabile esistente in giacimento e della pressione media d i
giacimento. V i e n e fornito un esempio d i applicazione d i questo metodo, il quale k per ora applicn-
bile solo a giacimenti che n o n abbiano (1 gas-cap iniziale nd spinta d’acqua apprezzabile.
))

( * ) Instituto Colombian0 de Petroleos, Bogota, Colombia.

e
132 PROCEEDINGS F O U R T H WORLD P E T R O L E U M C O N G R E S S - S E C T I O N IT/C

Tntroduetion pressure (M. R. P.) required under e. I t is


with the problem of the accurate determination
The so-called material balance equation (1) of a representative M. R. P. that this report is
for a reservoir producing by dissolved-gas drive concerned. As with nearly all studies of re-
connects three different groups of quantities. servoir flow, it is necessary t o make a number of
They are: sweeping assumptions in order to obtain a set
a. The quantities of oil and of free gas of equations which can be solved. In conse-
initially present in the reservoir, quence the results of this analysis must remain
b . The quantities of oil, gas and water suspect until they have been verified in practice.
withdrawn from the reservoir since the com- In this particular case, the normal assumptions
mencement of exploitation, and are made as to the homogeneity and isotropy
c. Certain volume conversion factors de- of the formation, and the constancy of the com-
pendent upon the P. V. T. relationships of the pressibility of the single flowing phase, although
reservoir fluids. the method is later generalized somewhat arbi-
Furthermore, in order to be able t o evaluate trarily t o include deviations from these highly
the quantities in group c, i t is also necessary t o idealized conditions.
know: In addition, it should be pointed out that
d. The initial reservoir pressure and the the present method was originally developed
solution G. 0 . R., and for application t o one particular field t and is
e. The present mean reservoir pressure. therefore specifically designed for the conditions
The usual method of application of the mate- of that field. For this reason, the method is
rial balance equation is to determine the quanti- restricted to fields producing by dissolved-gas
ties in groups b , d and in the field and the drive in which no significant water-drive is active
P. V. T. relationships of group c in the labora- and in which no large gas-caps were present
tory. The equation then reduces t o a single initially. The presence of either initial gas-caps
relation between the quantities in group a, i. C . or a detectable water-drive would complicate
the quantities of oil and of free gas which were the problem. No study has been made of this
initially present in the reservoir. Thus, in the more complicated system, but i t would seem
case of a reservoir in which it is knoyrn that n o likely that, used in combination with the method
free gas-cap existed initially, the equation may of Hurst and van Everdingen (3) for the eval-
be solved t o give the quantjity of oil initially uation of the volume of encroaching edge-water,
in place. In view of the extreme simplicity of a modified form of the present method could
the concepts involved in the theory of the ma- be devised.
terial balance equation, there can be little doubt
that the results derived from its use must lic Conventional Methods for the Determination
well within the limits of accuracy required for of the M. R. P.
reservoir engineering purposes, always provided
that the basic data enumerated under groups Before proceeding with a discussion of the
b, c, d and e above are known with sufficient proposed new method, it is perhaps of interest
accuracy. to review the conventional methods for deter-
Of these data, it is here assumed that good mining the M. R. P. Any such method must
production records have been kept, that good automatically resolve itself into two stages,
P. V. T. data are available and that both the namely, the determination of a representative
initial reservoir pressure and the solution G. 0. R. pressure corresponding t o each surveyed well
are well-established. All these requirements are and then the averaging process to be applied
so fundamental t o reservoir enginwring studies, to these individual pressures. These two stages
and present-day techniques of measurement are may conveniently be considered separately:
of such high relat,ive accuracy, that it should be a. Determining a representative pressure
possible to eliminate these data as sources of corresponding to each of the wells surveyed.
serious error. With the data of groups b, c and
d excluded, the only possibility of major error t Casabe field, Yondo Concession, Antioquia, Co-
lies in the determination of the mean reservoir lombia, S. A.
D. R. HORNER-AVERAGE R E S E R V O I R PRESSURE 133

It used t o be common practice t o close in pressure ” and which almost wholly ignores the
an arbitrarily selected group of wells and then effects of the set of localized pressure sinks.
to make a spot reading of their bottom hole Thus, in general, the use of the pressures p.
pressures after some fixed closed-in period, will give too high a value for the M. R. P.
which was often set a t 24 hours. I t is now b . Applying some averaging process to
known that, in the majority of wells, such a the individual well pressures to obtain the re-
pressure reading is no more than a single point quired M. R. P.
on a pressure build-up curve and is essentially A straight arithmetic average may be taken
without significance for the determination of of the individual well pressures. This has the
the M. R. P. When it was realized that the disadvantage that undue weight may be given
pressure build-up process was frequently of long to some portions of the reservoir while other
duration, i t became the custom t o assume that portions may not be adequately included. This
if a well were left closed-in for a sufficiently long difficulty is often overcome by drawing equi-
period, the static pressure (ps) which was finally pressure contours based on the pressure values
reached was the correct one on which t o base to be used and then calculating an areally weight-
the calculation of the M. R. P. ed M. R. P. Apart from the great difficulty
This use of pa has a number of disadvantages. of achieving good accuracy when contouring a
The build-up process in a new well is fairly rapid few pressure readings, this method can have no
and of easy interpretation, and so the determi- exact basis in theory, for each pressure observa-
nation of pa in a new well presents no serious tion is normally made a t different producing
problem, but new wells will usually be drilled conditions -f and so the observed pressures are
t o regions of higher-than-average pressure, and values of different functions and thus theoretic-
if the M. R. P. is to be based on such values ally may not be contoured; the effect of the
of pa an excessively high value will result. On individual pressure sinks is still excluded from
the other hand no satisfactory method is known consideration.
t o the writer for the determination of ps in an
old well short of leaving the well closed-in for
a period of perhaps many months until its pres-
General Considerations Applicable
sure finally reaches a state of approximate equi- to any Method for Determining an M. R. P.
librium. The loss of production so entailed is
What, then, are the conditions which have
frequently so costly that very few determina-
to be fulfilled by a set of pressure readings and
tions of pa can be made in old wells, while this
by the averaging process which is applied to them
same cost factor will also severely restrict the
in order to obtain the M. R. P. ?
number of permanently closed-in observation
Theoretically, it is necessary to know the
wells.
pressure a t every point in the reservoir in order
There is a further objection to the use of pa.
to determine its mean value, but this is, of course,
A producing reservoir may be thought of rough-
impossible t o achieve. The alternative is to
ly as a region of gradually varying pressure
use the pressure data from a particular well to
on which is superimposed a set of pressure sinks
calculate in some manner the mean pressure p
centred upon the wells. These sinks, although
for some portion of the reservoir surrounding the
of small areal extent, represent large pressure
well and then t o apply a further averaging process
drops, and in the case of old wells make a con-
siderable contribution t o the average pressure 6
t o the set of values of corresponding t o all the
wells surveyed in order t o obtain the M. R. I?.
drop which has occurred in the reservoir. By The portion of the reservoir surrounding a well
closing-in any particular well and permitting
its pressure t o stabilize, the effect of its own and whose mean pressure p is t o be determined
pressure sink is dissipated over the entire reser- from pressure observations in the well may con-
voir and is thus almost wholly lost to obser- veniently be termed the “ associated reservoir
vation. Thus the use of values of p a obtained volume ”, or A. R. V., of the well.
from old wells will give a value for the M. R. P.
which corresponds almost entirely to what was t The wells will not all be closed-in simulta-
termed above the ‘lregion of gradually varying neously during the pressure survey.
134 P R O C E E D I N G S FOURTH WORLD PETROLEUM CONGRESS-SECTION II/C

I t follows that the entire reservoir must be


divided into A. R. V.’s with one A. R. V. corre-
sponding to each well t o be surveyed. Further-
more, when the values of the quantities p have
been obtained (by some as yet unspecified meth-
od), these values may not be contoured, both
for the reasons given in paragraph 2a above and
also because j5 is, by definition, a mean pressure
over a volume and is not the pressure a t a single
point. Thus, to obtain an M. R. P. for a reservoir,
a value of p has to be determined separately for
every well in the reservoir. t However, for the
method being developed here, it seems that the
individual well surveys do not need to be exces-
sively long (either a 3- or a 6-day build-up is 1
usually sufficient) and so the disadvantage of Fig. 1. Typical division of a reservoir
having t o survey every well is not as serious as it i n t o A. R. V.’s.
at first appears.
Finally, the exact type of averaging process These are the only criteria which need t o
to be used must be decided upon. It has been be laid down for the application of the present
tacitly assumed in the past that an arithmetic theory and it is therefore proposed that the
mean is a sufficiently good approximation; this A. R. V. boundaries be chosen geometrically.
assumption is also used in this report. It prob- In that the present theory has principally been
ably does not involve a serious error and has designed for the case of a reservoir withasym-
the great advantage of simplifying the calcul- metric well pattern, wells will normally have
ations. The NI. R. P. is thus finally t o be deter- square A. R. V.’s (for the case of a square grid)
mined as the arithmetic mean of the values of
- or hexagonal A. R. V.’s (for wells drilled on a
p weighted in proportion t o the A. R. V.’s. triangular spacing). Fig. 1 shows the A. R. V.’s
as they might be chosen for a group of wells
The Choice of the Boundaries of the A.R.V.’* drilled on a square grid in which drilling is not
An A. R. V. must be assigned t o each well yet complete.
in such a manner that the methods of the fol- This method of choosing the boundaries of
lowing paragraph may be applicable. It will the A. R. V.’s has the disadvantage that there
be shown later that, for the application of the is some degree of freedom left in the choice of
proposed method, the A. R. V.’s only have t o the exact position of the A. R. V. boundaries;
satisfy two criteria. -t-f These are: in practice, it is believed that this will not be a
a. The boundary between two A. R. V.’s serious source of error.
should be situated as far as possible from the
wells. The Basic Equation for p
b. The individual wells should be situated The theory being developed is essentially
as closely as possible t o the geometrical centres for use in bounded reservoirs, but it is precisely
of their A. R. V.’s. the presence of boundaries which makes the
interpretation of bottom-hole pressure surveys
t Obviously a few wells will usually have t o be difficult. It is therefore postulated that it is
omitted from the survey because of mechanical dif- possible t o locate a hypothetical set of image
ficulties. So long as the number of omitted wells wells outside the reservoir boundaries so that,
is small, this will not seriously affect the accuracy il the existing producing wells and this set of
of the final M. R. P.
image wells had been simultaneously produced
t t It should be observed that these criteria in no from an infinite reservoir, there would at no
way involve the thought of a drainage area D. In
((

general, the concept of an A. R. V. as here defined time have been any flow in the infinite case
is wholly distinct from that of a drainage area. across the surfaces which correspond to the
D. R. HORNER-AVERAGE RESERVOIR PRESSURE 135

reservoir boundaries in the finite case. The corresponding image wells must also be consid-
flow within these surfaces in the infinite reservoir ered to be closed-in, while wells B, C, ..., K
would therefore have been identical with that and their image wells continue on production.
flow which has in fact occurred in the bounded The pressure build-up in well A is thus identical
reservoir. Such a set of image wells may be in form with that of a well in an infinite reser-
easily constructed for a number o l simple geo- voir until such time as its pressure is noticeably
metrical shapes (e. g. for a reservoir with rectan- altered by the interference effects attributable
gular boundaries) and there would seem to be either to the closing-in of its own image wells or
no basic reason why a set of image wells could to the continued production of wells B, C, ..., K
not be constructed for an irregularly shaped and of their image wells.
reservoir although the wells would probably Now the pressure build-up curve of a single
be infinite in number and might even be infi- well in an infinite field becomes a straight line
nitely dense over portions of the reservoir. -f to + e
Any reservoir may thus be considered as an when the pressure is plotted against log _ _
e
infinite one with the necessary image wells in- (e = closed-in time,
serted. cumulative production
A basic requirement for the ensuing analysis to = last production rate before closing=)
is what may be termed the “ superposition theo- and the extrapolation of this straight line to some
rem ” which states that the pressure drop at pressure p* at infinite closed-in time (corre-
any point in a producing reservoir may be ob-
tained by evaluating the individual pressure drops
at that point due to the various disturbing
to + e
e 1
sponding to Iog -= 0 gives the undisturbed
pressure at the well ( 4 ) .
causes as if these were acting i n d e p e n d e n t l y and Thus the corresponding plot for well A will
adding these individual pressure drops. This also be a straight line until the interference
result follows directly from the linearity of the effects mentioned above become serious, which
basic differential equation of reservoir flow, normally takes a considerable time, even in
which is: a very old well; the extrapolation of this straight
f P C dp
v2 p = ---
k dt line to log - 0 will thus give a pres-
__
to
9
where p is the pressure at a point in the reser- sure p * which is the ‘‘ undisturbed ” pressure
voir a t time t and P, P, c, k are the porosity, visco- at the well A, that is to say, which reflects
sity, compressibility and permeability respec- the pressure drop which has occurred at the
tively, all assumed to be constant. well due to the production of wells R , C,..., K
Consider now the total pressure drop which and all the image wells (i. e. specifically includ-
has occurred at a point in the reservoir close ing the image wells corresponding to well A
to some well A. This pressure drop is the sum but excluding well A itself). The value of p * thus
of the pressure drops due to the following includes all the pressure drops due to the causes
causes, evaluated as if each well were producing (ii), (iii) and (iv) above, and only excludes the
independently from an infinite reservoir: effect of well A itself.
i. Well A itself, In paragraph 2a above, it was stated: “ A
ii. The image wells corresponding to producing reservoir may be thought of roughly
well A, as a region of gradually varying pressure on
...
111. The other existing wells, say B, C,..., E(, which is superimposed a set of pressure sinks
iv. The image wells corresponding to wells centred upon the wells ”. In the region o f
B , C,..., K. well A, the pressure p* may be considered as an
Now suppose well A to be closed-in for a observation made in this region of gradually
bottom-hole pressure build-up test. Then the varying pressure. Thus if the A. R. V. corre-
sponding to the well A is so chosen that its boun-
t This concept of a set of image wells is not daries are not close to any other well, and further-
essential to the development of the theory; it is more so that well A is roughIy in its centre, it
introduced solely because it simplifies the ensuing follows that the value of p* observed at the well
argument. is approximately a mean for the region of gra-
~-

136 PROCEEDINGS FOURTH WORLD P E T R O L E U M CONGRESS-SECTION II/C

dually varying pressure surrounding the well j-. Combining equations [i]and [4] it therefore
If, then, 8 is the mean pressure drop over the follows that the mean pressure over the A. R. V.
A. R. V. of well A due to itself, it follows that of a well is given approximately by
the arithmetic mean p of the reservoir pressure
over the entire A. R. V. is
- -
p=p*-6 [I1 where @ is defined by equation [5].
To determine 8, it may be noted that the
pressure drop 6 a t distance r and time t after An Improvement in the Basic Equation for 5
the commencement of production of a well,
and due solely to the well itself, is There are two theoretical sources of error
in equation [6]. They are caused by:
i. The use of a cylindrical form for the
A. R. V.
ii. The assumption that the value of p*
where q is the well’s rate of production (assumed at the well is a true average for the “ region of
t o have stayed constant since the commencement gradually varying pressure ” surrounding the
of production), h is the producing‘sand thick- well. This may conveniently be termed the
ness and the Ei-function is defined by “ p*-assumption ”.

X
The first error could be eliminated by per-
forming the integration of equation [3] over the

/
n

known A. R. V. However, as has been pre-


Ei(-x)- -du viously stated (para. 4 ) , the theory is developed
:e
m
J mainly for a field drilled on a regular spacing
pattern in which case the majority of the
The required mean value 8 of this variable 6 A. R. V.’s will be regular polygons.
may now be obtained by integrating equation [2] For example, in the case shown in Fig. 1,
over the A. R. V. of the well. Assuming the there are only 5 or 6 A. R. V.’s out of 22 in which
A. R. V. t o be a circular cylinder of radius a and the well is seriously off centre in its A. R. V.
height h, we have and only in these cases would the approximation
t o the circular form of the A. R. V. be seriously
n h (az-r:) 8 = in error. I t is believed therefore that, although
i t is theoretically possible to correct equation [6]
/
,Pa

- ,,“Lh E j (-= r2fpc


)2xrhdr
for the non-circularity of the A. R. V.’s, the
[3] amount of additional work involved is not
J warranted by the small improvement in theo-
TW
retical accuracy which would thereby be
achieved.
where rw is the radius of the well.
The second error mentioned a t the beginning
To a very close approximation, equation [3]
of this paragraph may also be expected t o be
may be reduced to
of minor importance: in this case, however, it
is possible t o estimate the order of the error
involved and t o modify equation [SI so as par-
tially t o eliminate it, without actually increasing
where the amount of work involved.
Consider a single well producing from the
[51 centre of a finite circular reservoir. Then the
pressure drop a t any point in this reservoir may
7 The errors in this approximation may be some- be calculated exactly from the well-known solu-
what reduced by the considerations of t h e next pa- tion for the flow in this case (4). The algebra
ragraph. is somewhat tedious and is therefore omitted,
D. R . H O R N E R - A V E R A G E RESERVOIR PRESSURE 137

but written in a similar form to that of equa- 3. In practice, the diflerences hetween
tion [6], the mean pressure p in the entire reser- equations [6] and [7] are small;
voir may be shown to be: 4. It is thus probably preferable to use
equation [7] throughout.
-
p p* - - CIF [3/2 -y + ln (I)
4nkh
The Quantities A and +
For convenience, define a quantity A b y the
equation:
where

Then, from equation [5], it follows that


rb is the radius of the circular reservoir
tln is the nth positive root of JI (x) = 0
[91
Jo,JI are the Bessel functions of the first
kind of orders zero and unity respectively,
In is the Napierian logarithm
and y is Euler's constant = 0.5772...
and so, eliminating
4nkh
-= by the use of equation
Now equation [6], by its derivation, should [9], equation [6] becomes
also apply to this case of a single well at the
centre of a finite circular reservoir, with the
reservation that equation [6] includes the errors
which are inherent in the " p*-assumption ".
The difference between equations [6] and [7] In a similar manner, equation [7] may be
thus represents the error in equation [6] due written
to the p*-assumption for the special case of
a circular reservoir boundary coinciding with
the A. R.' V.
It is therefore proposed to use equation [7] in
place of equation [6] for all cases, for the reason
that equation [7] completely eliminates the whilebothequations [io] and [11] may be written:
errors of equation (6) in the one case which it -
is possible to investigate completely and will thus p = p*-A + [I21
presumably reduce these errors in other cases.
This appears t o be a very unscientific argu- where J, is a function of @ defined by either:
ment. However, in a practical field application
it will normally be found that equations [6] and
[7] give very similar final results. For example
,+ = 1- e-'/@ - -1 Ej (-
CD
i) 131
in the Casabe field, Colombia, a difference of or
only 2.8 oh in the quantity of oil initially in
place resulted from using equation [7] instead of 141
equation [6] (see para. 12). The argument in
favour of using equation [7] may be briefly
recapitulated thus: dependent upon whether equation [B] or [7] is
1. Equation [6], although derived for very to be used.
generalized conditions, contains an error due The two equations [13], [14], giving J, as a
to the p* -assumption; function of @, have been plotted in Fig. 2. Fur-
2. For the very restrictive case of a single thermore, the second (preferred) form of the
well at the centre of a circular reservoir, equa- function J, has been tabulated, for greater ease
tion [7] is exact; of calculation, in Appendix B.
138 PROCEEDINGS FOURTH WORLD P E T R O L E U M CONGRESS-SECTION II/C

fl
Fig. 2. J , as 8 function Of @

The Determination of jj in order t o read off N, but in fact any second


functional relationship between N and the
The present paper is based on the premise of M. R. P. may also be plotted on the graph of
the inaccuracy of the conventional methods for the material balance equation and the required
determining the M. R. P. and thus on the un- value of N (and also of the M. R. P.) may then
reliability of the value of N (the volume of tank be read off at the point of intersection of the two
oil initially in place in the reservoir) obtained curves. Such a method is the one proposed
directly from the material balance equation for here. In order to derive this second functional
the entire reservoir, but the validity of the equa- relationship, a value is assumed for N and a
tion per se is nowhere questioned. The mate- corresponding value is derived for the M. R. 1’.
rial balance equation may thus be thought of using the method described below; by taking
as a single equation connecting two unknowns, a suitable suite oE values for N this required
namely N and the M. R. P., and so, given only second functional relationship between N and
the cumulative productions of oil, gas, and water the M. R. P. is obtained. In the following
of the wells, and the necessary P. V. T. data, a discussion, N is therefore treated as a known
curve may be drawn using the material balance quantity.
equation for the entire reservoir showing N as From a pressure build-up survey in a par-
a function of the M. R. P. The calculation of ticular well, values may be derived for the quan-
this curve requires no further comment. It
would, a t first sight, appear to be necessary next
tities p * and wh-
q c L If the yuantity A can
t o determine the correct value of the M. R. P. also be determined, equations [9] and [12],
D. R. HORNER-AVERAGE RESERVOIR PRESSURE 139

together with the + - a relation of equation [13] From this form of equation [ 8 ] , it now be-
or [14], are sufficient t o determine p for the comes possible t o give a physical meaning t o the
A. R. V. Again, from the geometry of the quantity A , for the above form implies that A
-4.R. V., the quantity a, may be calculated is the pressure drop which would have occurred
where a is defined as the ratio of the volume of within the A. R. V. due t o the production of the
the particular A. R. V. t o the volume of the well if the A . R. V . had been a sealed reservoir.
entire reservoir. It then follows that the Equation [8] may thus be generalized by
YUP adopting this definition for A , and A may then
M. R. P. is ~ where the summation ex- be determined from a subsidiary material ba-
la
tends over all the A. R. V.'s. The problem lance equation (S. M. B. E.) made for the well
is thus reduced t o the determination of and its A. R. V., assuming no flow t o have occur-
A =
'
nazfhc'
Now theoretically all the quantities
red across the A. R. V. boundaries. In order
t o set up this S . M. B. E. it is necessary t o
g, t, a, f , h, and c may be considered as known, know:
but in practice considerable difficulties may a. The quantity of oil initially in place
arise in the determination of any or all of them. in the A. R. V.
What value, for example, should be taken for b . The initial pressure conditions to be
the compressibility c, or how should the volume used for the S . M. B. E.
rate o l production q be interpreted in a well Point a above presents no problem. As
producing at other than the solution G. 0. R. ? was explained above, the quantity N may be
Such questions are difficult t o answer in practice. considered as known, and so a N , the quantity
Furthermore, it should not be forgotten that of oil initially in place in the A. R. V., is known
the various equations here presented have been also.
derived for a highly idealized reservoir with a Point b , namely the choice of initial pressure
single flowing phase of constant compressibility conditions t o be used for the S . M. B. E., presents
being produced a t rates which are constant for a more difficult problem. Equation is], which
the individual wells. It is thus desirable as was generalized in an attempt t o give a practical
far as possible t o eliminate the use of in- definition t o the quantity A, gives no clue on
dependently determined parameters and only this point, for the theory has been derived assu-
to use those quantities which may be direct- ming c t o be constant and it is precisely the
ly determined from reservoir observations. -f variability of the compressibility which makes
Furthermore, the quantity A as defined by it necessary t o lay down the initial pressure for
equation [8] must be generalized as far as pos- the S . M. B. E.
sible t o permit the practical application of the The initial pressure for the S . M. B. E. is
theory t o actual oil-field conditions. unknown, and so is the final pressure. Suppose,
Now equation [8] may be written: then, that these two pressures are pi and pi

at
A = xazh x f x c
-
Cumulative production of well
(Gross sand volume of A. R. V.) x (porosity) x (compressibility)
Cumulative production of well
or A x (compressibility) =
Total fluid volume initially within A. R. V.
~~

respectively. Then, from the generalized defi-


t This ideal has already been achieved with respect nition of
to the quantities 4shby determining them from
the build-up curves of the individual wells. This
A, it follows that
A = Pi - Pf ~ 5 1

eliminates the diflculties attendant upon the deter- the S. M. B. E* is to be used, in


mination o f the correct values to be used for the to determine a mean compressibility for the
individual quantities q, p, k and h. reservoir fluids in the A. R. V. for the particular
140 P R O C E E D I N G S FOURTH WORLD P E T R O L E U M CONGRESS-SECTION II/C

withdrawal conditions imposed by the well under of N required t o give a known pressure drop
examination. The mean pressure in the entire from the init,ial reservoir pressure po due to the
A. R. V. is, by definition, p; it would seem to be gauged withdrawal of the well, but it may not be
reasonable t o set this pressure equal to the av- used immediately when starting from any other
erage pressure prevailing in the A. R. V. during pressure, although in fact it is required that the
the period of the S. M. R. E., i. e. to set S. M. B. E. be made from pressure pi to pf.
This difficulty may be overcome as follows.
1
P = 2 (pi + pi) [161 The compressibility curve corresponding t o
equation [IS] may be thought of as expressing a
The problem of determining p has already functional relationship between two variables,
been reduced to one of determining A . The 1 1 1 1
introduction of the two pressures pi and pi t o p and -N; then let - - be the two values of ~-
Ni’ Nf N
define A thus requires the setting up of two equa- which correspond to the pressures pi and pf.
tions to define pi and pf. Equation [16] is one Then it follows from the definition of Ni that
such equation; a second equation can be derived if there had been a quantity Ni of tank oil ini-
as follows. tially in place in the reservoir, the known pro-
As stated in paragraph 1, i t is assumed that duction of the well would have been sufficient
no free gas-cap was originally present, and that t o have caused a pressure drop from po to pi in
there is no active water drive. The oil originally the (supposed sealed) A. R. V.. Now if p is
in place in the A. R. V. was o! N and so the any quantity, it follows that the same pressure
S . M. B. E., when made from original reservoir drop from po t o pi would have occurred in the
conditions, may be written (sealed) A. R. V. if there had been a quantity p Ni
UN of tank oil initially in place and if there had
~ (Vt- Vto) = s been a withdrawal from the A. R. V. of a vo-
cb
lume 13 x (actual withdrawal from the well).
where N is measured in barrels of tank oil N
is the shrinkage factor from reservoir Setting p = - i t follows that this same pres-
cb Ni
oil a t bubble-point to tank oil sure drop from po to pi would occur with
Vt is the reservoir volume a t some pres- N
sure p occupied by unit volume of x Ni = N tank oil initially in place conse-
bubble-point oil and any gas liher- N
quent upon a withdrawal of - x (actual with-
ated therefrom Ni
Vto is the volume a t initial reservoir con- drawal from the well).
ditions of unit volume of bubble- Similarly, a pressure drop from po t o pi would
point oil be caused by the production of x (actual
and S is the reservoir volume at pressure p of K
the cumulative production of oil, gas withdrawal from the well), assuming the tank
and water of the well under exami- oil initially in place to be 1\.
nation. The implications of this reasoning are thus
Equation [17] may be written that, always with N tank oil initially in place,
the pressure in the (sealed) A. R. V. drops t o
0,. (Vt-Vto) - [‘*I N
s Cb - N pi after producing a fraction - Ni of the well’s
known Production and then drops to Pf after
and in this form the left-hand side of the equa-
tion is a known function of the pressure p. Thus
from equation [is] there may be constructed a
producing a further fraction f:‘( - Ei)
of the

curve (which may be conveniently termed the


“compressibility curve”) showing p as a function The pressure thus falls from Pi to pf due t o
1 the well’s actual cumulative production provided
of -;
N this compressibility curve may be used
“ g
directly to determine, for example, the value Ni-Ni=
D. R. HORNER-AVERAGE RESERVOIR PRESSURE 141

or g . Using the value of p* derived from the


-1_ - 1 1 build-up curve, equation [12] is now used to
Nf Ni = N [191 calculate a suite of trial values of p.
1 1 h ) At this stage, there are available suites
N being considered as known, and N-i, Ni being of trial values of pi, pi and p. In order t o se-
known functions of pi and pi, this equation [I91 lect the correct value of 6, pi +
pi is plotted
thus represents the additional equation required against p and the required value of p is read off
to determine pi and pr mentioned above. in accordance with equation [16].
It must be strongly emphasized that this It therefore follows that the precise values
discussion is not intended to suggest that the of the intermediate quantities A, +, @, etc., are
boundaries of the A. R. V.’s are sealed, or alter- never actually determined in the course of the
natively that there has been no flow across calculations, although they could, of course, be
them; there is similarly no suggestion that the determined by subsequent interpolation.
pressures pi and pi have had, or will have, any
special physical significance in the pressure his- The Time Correction
tory of the A. R. V., nor has the S. M. B. E.
any physical meaning. The whole concept of By following the method outlined above, a
the S . M.B. E. made within a sealed A. R. V. is set of values of p for the A. R. V.’s may be de-
introduced solely as an artifice t o enable A t o rived for every value of N assumed, and a weight-
be generalized and evaluated in practice, and ed average? of these values of p may then be
must only be considered as such. obtained. That this weighted average is, in
Equations [RI, [12], [15], [16] and [19j to- fact, the required M. R. P. is not immediately
gether with the curves of Fig. 2 (or the tabulation clear for the case when the pressure survey has
of Appendix B) and the particular compressi- extended over a long period, for it would be
bility curve for the A. R. V. corresponding t o expected that some time correction would have
the S. M. B. E. [is] now suffice t o determine a to be applied. This time correction may be
value of p given a value of N, thus: ? eliminated, however, in the following manner.
1 Assume that the pressure at all points in the
a. A suite of trial values is chosen for r\l
reservoir falls at some constant rate w psi/month
b. Using the assumed value of N, equation during the period of the survey. Then if the
[19] is used t o determine a corresponding suite mean pressure in a particular A. R. V. was p
of values of -
1 when i t was measured a t time t , it would have
Ni been p - (t - tl) w a t some datum time t.
c. A pair of values of pi and pi correspond- A t this datum time f. therefore.
1 1
ing t o each pair of values of - and -is read 1 U [p - (i. - t l ) w]
Ni Nf NI. R. P. =
off the compressibility curve for the A. R. V. XU
d ) A corresponding suite of trial values i. e.
of A is derived from equation [is]. M’.R..P. =
X U p
__
C U
e. The value of qEL for the well being if
known from the build-up curve, a suite of values
of D
I is determined from equation [9].
f . Corresponding values of + are derived
from the curves of Fig. 2, corresponding to equa- Thus, by choosing the datum time t as the
tions [13j or [14j. (If equation [14] is to be used, weighted mean of the observation times for the
8 may be determined more easily from the tab- individual wells, there is no need to apply time
ulation of Appendix B). corrections t o the values of p in order to deter-
mine the M. R. P.
t An alternative method for determining p,
which may sometimes be preferable t o this one, is t Weighted, of course, in proportion to the qua.n-
given in Appendix A. tities a.
142 PROCEEDINGS FOURTH WORLD P E T R O L E U M CONGRESS-SECTION Il/C

1500

1250

.-mU
0

1000
c
K
i

750

500
50 75 100 125 150 175
N, MILLIONS OF BARRELS OF TANK OIL

Fig. 3. Calculation of N

The Two N - M.R.P. Curves sities in centipoises, times in seconds, lengths in


centimetres and volumes in cubic centimetres.
Finally, therefore, two curves are derived However, in the final forms in which the equa-
exhibiting functional relationships between N tions are presented, CD and + are dimensionless,
and the M. R. P.
Curve 1 is obtained from the material ba- A, p * , P, Pi, Pf, and m
" all have the di-
lance equation for t,he entire block made a t mensions of a pressure and the quantities N,
time t as defined by equation [21j. Ni and Nf have the dimensions of a volume.
Curve 2 is derived by plotting the M. R. P. I t thus follows that all calculations may be
calculated from equation [20] against the values made measuring the pressure terms in any
assumed for N. convenient units (say Ib./sq. in2 or kg/cmz) and
The point of intersection of these two curves the volume terms in any other convenient units
gives the required value of N (and also, inci- (say in millions of barrels or of cubic metres)
dentally, of the M. R. P. at, time %). provided that the S . M. B. E. [is] be set
up in these units. Theoretically, the build-up
Dimensions, Units, etc. curves should be plotted against In __
to + ', where
e
In the derivation of the equations so far, no In is the Napierian logarithm, in order that the
mention has been made of the units which are slopes of the build-up curves should have the
to be used, although it is implicit in the deriva- correct values, but it is usually more convenient
tion of the equations that such a system of to use the logarithm to base 10 for the abscissae
units must be self-consistent. The obvious of the build-up curves. In this case it must
system is that used in the definition of the not be forgotten that the slope of the build-up
darcy, namely pressures in atmospheres, visco- curve must be factored by log10 e = 0.4343 in
- ~~

D. R. HORNER-AVERAGE RESERVOIR PRESSURE 143

leaks with rising well-head pressure, and so a


fairly extensive repair programme had t o be
undertaken. This introduced considerable de-
lays, and even so some of the older wells could
not be satisfactorily surveyed at all, due t o the
presence of subsurface leaks which could not
be easily located. A number of other operating
difficulties further delayed the completion of the
survey, which was not finally finished until
February 1954.
Of the wells which did not leak, three gave
build-up curves a t variance with normal build-up
theory and two were not surveyed because i t
was considered to be unsafe to do so. Finally,
usable build-up curves were obtained from only
71 y0 of the wells in the block.
I t is to be feared that 29 y0 of unsuccessful
surveys may well prove to be a normal fraction
of rejects in any future application of this method,
and it is strongly recommended that every
effort be made t o eliminate casing leaks before
commencing the survey, even t o the extent of
undertaking major workover jobs in troublesome
wells.
The first step in the calculations consists
of calculating the mean tjme 5 (as determined
4.5 4.0 3.5 3.0 2.5 by equation [21]) a t which the material balance
to + 0
for the entire block should be made. In this
log10 7 particular case, this time 5 corresponds to 13th

Fig. 4 . Build-up curve

order t o obtain the correct values for ~ qII.


k x .
In the following example, volumes are meas-
ured in millions of barrels and pressures in
psia.

A Practical Application
P

A test of this method was commenced in one


fault-block of the Casabe field in the spring of
1952. As the wells were almost wholly pumping
wells, the technique was t o pull the rods and
then t o close in the tubing and run the bomb
as quickly as possible to the bottom of the
tubing. It was found t o be possible t o have
the bomb in position a t about 2,500 feet within
0 ,005 01 015 02 ,025 .03 035
2-3 hours of stopping the well pumping, so that
satisfactory build-up data could be obtained. -
N
However, considerable difficulty was experienced
wit8hthe older wells as most of them developed Fig. 5. Compressibility curve
144 PROCEEDINGS FOURTH W'ORT,D PETROLETJM CONGRESS-SECTION II/C

Pi + P f
Fig. B. Dctermination of I, for N = I00

December 2952 a t which time the material 1


then, assuming N = 100, - - 0.01, the calcula-
balance equation was used to determine curve N-
1 of figure 3 . Lion of p is performed thus (using the @ - 4
This is the first of the two functional rela- relation of equation [14]):
tionships between N and the M.R.P. -1 1
With regard t o the determination of the second Ni
.~
Nf
Pi Pi A @ 4 +A G Pi + P f
of these functional relationships, it would be te- 0 .010 1586 969 617 31.6 .I385 85.5 710.5 2555
.005 .015 1221 777 444 22.8 .I776 78.9 717.1 1998
dious to reproduce the entire calculations made for .010 .020 969 619 350 17.9 .2127 74.4 721.6 1588
all the wells, and so the results of one well only will .015 .025 777 483 294 15.1 .2409 70.8 725.2 1260
.020 .030 619 367 252 12.9 .2698 68.0 728.0 986
be examined in detail here; the calculations for the
other wells are made in an identical manner. and finally, plotting p against pi
A 72-hour closed-in bottom-hole pressure equation [16] gives a value p = 723 psia.
+ pi in Fig. 6,
survey was made in this well, the well being Proceeding in this manner with all the
closed-in on July 25th, 1952; the results of this A. R. V.'s, and finally determining a weighted
survey are plotted in Fig. 4. From this build-up average of the values of p, a value of 959 psia
curve two results may be directly calculated. may be determined for the M. R. P. a t 13th
They are: December, 1952, based on N = 100.
a. Slope of straight-line portion = This calculation of the M. R. P. is now
qP repeated using other assumed values of N,
4txkh -
45 psi/loglo cycle whence ___ -
45 x 0.4343 =
and the corresponding N - MRP relationship
19.5 psi and b. p * = 387 psig at 2190 feet S S is plotted as curve 2 in Fig. 3; the point of in-
which is equivalent to p* = 796 psia a t the tersection of this curve with that corresponding
chosen datum of 3253 feet SS. to the material balance equation for the entire
For this particular well, the 8. M. B. E. [is] block finally gives the required value N = 107
gives the compressibility curve of Fig. 5, and million barrels.
L). R. HOKNER--AVERAGE RESERVOIR PRESSURE i 45

Fig. 3 also shows the curve corresponding the instantaneous compressibility at p. In this case,
to the @ - J, relation of equation [13], resulting instead of the above three equations there must be
in N = 110 million barrels, which is only 2.8 y0 substituted:
/ d D \
in excess of the previous result. [22j
From this same Fig. 3 i t can be seen t h a t
the result of using p = p*, is t o give N = 134
million barrels, some 25 :/,, in excess of the which is the result of substituting the tangent to the
previous result. compressibility curve at for the chord between
pi and pr. This somewhat shortens the tabulation
of paragraph 12, which takes on a somewhat different
Conclusions form. The initial step now is the assumption of a
suite of values for p, thus:
This is the first test which has been made of 1
the method and i t is therefore early t o draw N = 100, - - 0.01
N-
any conclusions, but i t is of interest to note
that this value of N = 107 millions agrees very
well with the figure of 102 millions determined
900 38,500 385 19.7 .1981 76.3 719.7
volumetrically. 800 35,100 351 18.0 .211S 74.3 721.7
700 31,800 318 16.3 .2278 72.4 723.6
600 29,000 290 14.9 .2342 70.5 725.5
Acknowledgement and then, plotting the assumed value o l p against its
The writer wishes, to express his thanks to calculated value, one may derive immediately jJ =
723 psia which agrees exactly with the value
Coricesionaria de Petroleo Shell-Condor, Bogota, previously calculated. It is coincidental that in
Colombia, for permission to publish this paper. this particular case there is no detectable diffe-
rence between the two methods of calculating p,
APPENDIX A but it appears t o be safe to say that normally the
difference between t h e two methods will be so small
An Alternative Method for Using that equation [22] may be substituted at will for
the Compressibility Curves equations [15], [16] and [19]. However, the useof
the equation [22] has the disadvantage that it
The essential equations which are used in conjunc-
is necessary to calculate
dP -
(i)
graphically, which
tion with the compressibility curve t o determine A are
A=pi- Pf [ 151
makes the method both somewhat inaccurate and
time-consuming. It is t>hereforerecommended that
- 1 the method given in paragraph 12 be used normally
P = 2 (Pi + Pi) [if31 and that equation [22] be only used either:
U . when the compressibility curve has a very
small slope (i. e. for wells with relatively small cumu-
lative offtake)
or b. when either pi or pf Pall off the compres-
These equations have been derived with t h ~ sibility curve and are thus not defined. This is an
thought that the S.M. R. E. is being used to determine unusual phenomenon, but is not theoretically impos-
the compressibility of the reservoir fluids over the sible, and then the use of the method o f this Appendix
pressure range pi to pf. It might easily be argued enables a value to be determined for p whrn the
that the S.M. R . E. should be used t o determinr ... method using equations [is],[ 1 R ] and [19] breaks down.

Proceedings 4th W . P . C. - Section rr 10.


P R O C E E D I N G S FOURTH W O R L D PETROLEUM CONGRESS-SECTION II/C

APPENDIX B

Table of the Function

CD .oo .01 .02 .03 .04 .05 .06 .07 .oa .09

.o 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000
.1 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000
.2 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000
.3 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 .9999 .9999 .9999 .9999
.4 .9999 .9998 .999a .9997 .9997 .9996 .9995 .9994 .9994 ,9992

.5 .9991 .9990 ,9988 .99a7 ,9985 .99a3 .9981 .9978 ,9976 .9973
.6 .9970 ,9967 .9962 .9959 .9955 .9951 .9947 .9942 .9937 .9933
.7 .9927 .9921 .9916 .9910 .9903 .gag7 .9890 .9sa3 .9a75 .9a6a
.a .9860 ,9852 .9844 .9a35 .9az6 .gal7 .98oa .979a ,9789 .9779
.9 .9769 .975a .974a .9737 .9726 .9714 .9703 .9692 .9680 .ma
1.0 .9656 .9643 .9631 mia .9605 ,9592 .9579 .9566 .9552 .9539
1.1 .9525 .9512 .9497 .94a3 .wa .9454 .9440 .9425 .9410 ,9395
1.2 ,9380 ,9365 .9350 .9335 .9320 .9304 .wag .9273 .9257 .9242
1.3 .9226 ,9210 .9194 .917a .9162 .9146 .9130 .9114 .9098 .g081
1.4 .go65 .9049 .9032 .9016 .go00 .ma3 .a967 .8950 .a934 .8917

1.5 .ago1 .a884 .a867 .a851 .a834 ,8818 .a801 A784 .ma .a751
1.6 .a735 mia .8701 .a685 .a668 A652 .a635 .a618 .a602 .a585
1.7 .a569 ,8552 .8536 .a519 .a503 .a486 .a470 .a454 .8437 .a421
1.8 .a405 .a388 .a372 .a356 .a340 .a324 ,8307 .ami .a275 .a259
1.9 .a243 .a227 .8211 .8!95 .a180 .a164 .a148 ,8132 .8116 .a101

2.0 .a085 ,8069 .a054 .a038 .a022 .a007 .7992 .I977 .7961 .7946
2.1 .7931 .7916 .7900 .7aa5 .~WO .7855 .~MO .7825 .7aio .'I796
2.2 .77ai .'I766 .7751 .7737 .7722 .7707 .7693 .ma .7664 .7649
2.3 .7635 .7621 .7606 .7592 .757a .7564 .7550 .7536 .7522 .750a
2.4 .7494. ,7480 .7466 .7452 .7439 .7425 .'I411 .739a .73a4 .7370

2.5 .7357 .7344 .7330 .7317 .7304 .7290 .7277 .7264 ,7251 .ma
2.6 .7225 .7212 .7199 .na6 .7173 .7160 .7147 .7135 .7122 .7109
2.7 .7097 .7oa4 .7072 .7059 .7047 .7034 .7022 .7010 .699a .69a5
2.8 .6973 ,6961 .6949 .6937 .6925 .6913 .6901 .mag .6877 .6a65
2.9 .6a54 .6a42 .WO .6ai9 .6ao7 .6795 mac .6772 .6761 .6749

3.0 .6na .6727 .6715 .6704 .6693 .66m .6671 .6660 .6649 .ma
3.1 .6627 .6616 .6605 .6594 .65a3 .6572 .6561 .6551 .6540 .6529
3.2 .6519 .65oa .6497 .64a7 .6476 .6466 .6456 .6445 .6435 .6425
3.3 .6414 .6404 .6394 .63a4 .6374 .6363 .6353 .6343 .6333 .6323
3.4 .6313 .6304 .6294 .e284 .6274 ,6264 .6255 .6245 .6235 .6226

3.5 .6216 .6206 .6197 .6i7a si168 .6159 .6149 .6140 .6131
3.6 .6121 .6112 .6103 .6094 .6oa5 .6075 .6066 .6057 .6048 .6039
3.7 .6030 .6021 .6012 .6003 .5994 3977 .ma A959 .5950
3.8 .5942 .5933 .5924 .5916 .5907 .5a9a .5a90 .5asi .5873 .5a64
3.9 .5856 .5a47 .5a39 .5831 . 5 a ~ .sa14 .5ao6 ,5797 .5m9 mal
4.0 .5713 .5765 .5756 .574a .5740 .5732 .5724 ,5716 .uoa A700
4.1 .5692 .56ac .5676 .ma .5661 .5653 .5645 .5637 .5629 .5621
4.2 .5614 .5606 .5599 .5591 .55a3 .5576 .556a .5561 .5553 .5546
4.3 .553a .5531 .5523 .5516 .5508 .5501 .5494 .54a6 .5479 .5471
4.4 .5464 .5457 A450 3443 A436 .5429 .5422 .5414 .5407 A400

4.5 3393 .53a6 .5379 .5372 A365 .535a .5351 .5344 .5337 .5331
4.6 .5324 5317 .5310 .5303 .5296 .5290 .5za3 3276 .5269 .5263
4.7 .5256 .5249 .5243 .5236 .5230 .5223 ,5216 .5210 .5203 .5197
4.8 .5190 ma4 .5J78 3171 ,5165 .si58 .5152 3146 .5139 3133
4.9 .5127 .5120 3114 A108 .5101 .5095 .5oa9 ,5083 5077 .5071
U. R . H O R N E R - A V E R A G E RESERVOIR PRESSURE 147

(I, .oo .01 .02 .03 .04 .05 .06 .07 .08 .09

5.0 .5064 .5058 ,5052 3046 .5040 3034 3028 .5022 3016 5010
5.1 3004 .4998 .4992 .4986 .4980 .4974 .4969 .4963 .4957 .4951
5.2 .4945 .4939 .4934 .4928 .4922 .4916 .4911 .4905 .4899 .4893
5.3 ,4888 .4882 .4876 .4871 .4865 .4860 .4854 .4848 .4843 .4837
5.4 .4832 .4826 .4821 .4815 .4810 .4804 .4799 .4794 .4788 ,4783

5.5 .4777 .4772 .4761 .4761 .4756 ,4751 .4745 .4740 .4735 .4729
6.6 .4724 .4719 .4714 .4709 .4703 .4698 .4693 .4688 .4683 .4676
5.7 .4672 .4667 ,4662 .4657 .4652 .4647 .4642 .4637 .4532 .4627
5.8 ,4622 .4617 .4612 .4607 ,4602 .4597 .4592 ,4587 .4582 ,4577
5.9 .4572 .4568 .4563 .4558 .4553 .4548 .4543 ,4539 .4534 ,4529

6.0 .4524 .4520 .4515 .4510 .4505 ,4501 .4496 .4491 .4487 .4482
6.1 .4477 .4473 .4468 .4463 .4459 ,4454 .4450 .4445 .4440 .4436
6.2 .4431 .4427 .4422 ,4418 .4413 .4409 .4404 ,4400 .4395 .4391
6.3 .4386 .4382 .4317 .4373 .4369 ,4364 .4360 .4355 .4351 .4347
6.4 ,4342 .4338 .4334 .4329 .4325 ,4321 .4316 ,4312 .4308 .4304

6.5 .4299 ,4295 .4291 ,4287 .4282 .4278 .4274 ,4270 ,4266 ,4262
6.6 .4257 .4253 .4249 ,4245 .4241 .4237 .4233 .4229 .4224 .4220
6.7 .4216 .4212 .4208 .4204 .4200 .4196 .4192 .4188 ,4184 .4180
6.8 .4176 ,4172 .4168 .4164 .4160 .4156 ,4152 ,4148 ,4145 .4141
6.9 .4137 .4133 ,4129 .4125 .4121 .4117 .4113 .4110 ,4106 .4102

7.0 ,4098 .4094 .a091 ,4087 .4083 .4079 ,4075 .4072 .4068 .4064
7.1 ,4060 .4057 .4053 .4049 ,4046 .4042 .4038 .4034 .4031 .c027
7.2 .4023 ,4020 .4016 .4012 .4009 .4005 ,4002 ,3998 .3994 3991
7.3 .3987 .3984 ,3981 .3977 .3973 .3969 .3966 ,3962 .3959 .3955
7.4 ,3958 ,3948 ,3945 .3941 .3938 .3934 .3931 .3927 .3924 .3920

7.5 .3917 .3913 ,3910 .3907 .3903 ,3900 .3896 .3893 .3890 .3886
7.6 ,3883 ,3880 ,3876 ,3873 .3869 .3866 ,3863 .3859 .3856 .3853
7.1 .3849 .3846 ,3843 .3840 .3836 ,3833 .3830 .38'L6 .3823 .3820
7.8 .3817 .3813 .3810 .3807 .3804 ,3800 .3797 ,3794 .3791 .3788
7.9 .3784 ,3781 .3778 ,3775 ,3172 .3769 .3765 .3762 .3759 .3756

8.0 .3753 .3750 .3747 ,3743 .3740 ,3737 .3734 .3731 .3728 .3725
8.1 .3722 .3719 .3716 ,3713 .3710 ,3707 .3703 .3700 .3697 .3694
8.2 .369l ,3688 .3685 .3682 .3679 .3676 .3673 ,3670 .3667 .3664
8.3 .3661 .3659 ,3656 .3653 ,3650 3647 .3644 .3641 .3638 .3635
8.4 .3632 ,3629 ,3626 3624 .3621 ,3618 .3615 .3612 ,3609 .3606

8.5 .3603 ,3600 .3598 .3595 ,3592 .3589 .3586 ,3584 .3581 .3578
8.6 .3575 .3572 .3570 .3567 ,3564 .3561 .3558 .3556 .3553 .3550
8.7 .3547 .3545 .3542 .3539 ,3536 .3534 .3531 ,3528 ,3525 ,3523
8.8 .3520 .3517 .3515 ,3512 .3509 .3506 .3504 .3501 .3498 .3496
8.9 .3493 ,3490 .3488 .3485 ,3482 .3479 ,3471 .3475 .3472 ,3469

9.0 .3467 .3464 .3461 .3459 .3456 .3454 .3451 .3448 .3446 .3443
9.1 .3441 .3438 .3436 .3433 ,3430 .'3428 ,3425 ,3423 ,3420 ,3418
9.2 .3415 ,3413 .3410 ,3408 ,3405 .3403 ,3400 .3398 ,3395 .3393
9.3 .3390 .3388 .3385 ,3383 .3380 .3378 ,3375 .3373 .3370 .3368
9.4 .3365 .3363 ,3361 ,3358 ,3356 .3353 .3351 .3348 .3346 ,3344

9.5 .3341 .3339 .3336 .3334 .3331 .3329 .3327 .3324 .3322 .3320
9.6 .3317 .3315 ,3312 .3310 ,3308 .3305 ,3303 ,3301 .3298 .3296
9.7 .3294 .3291 .3289 .3287 ,3284 .3282 .3280 .3278 .3275 .3273
9.8 .3271 .3268 ,3266 .3264 .3261 .3259 .3257 .3255 .3252 .3250
9.9 .3248 ,3246 .3243 ,3241 ,3239 .3237 .3234 .3232 ,3230 .3228

10.0 ,3225
148 PROCEEDINGS FOURTH WORLD P E T R O L E U M CONGRESS-SECTION II/C

(D .o .1 .2 .3 .4 .5 .6 .7 .8 .9
10 .3225 .3203 .3182 .3160 .3139 .3118 .3098 .3078 .3058 .3038
11 .3019 .3000 .2981 .2962 .2944 .2926 .2908 .2891 .2874 .2857
12 .2840 .2823 .MO7 .2791 .2775 .2759 .2743 .2728 .2713 .2698
13 .2683 .2668 .2654 .2640 .2625 .2611 .2598 .2584 .2571 ,2557
14 .2544 .2531 .2518 ,2506 .2493 .2481 .2468 .2456 .2444 .2432
15 .2421 .2409 .2397 .2386 .2375 ,2364 .2353 .2342 .2331 .2320
16 .2310 .2299 .2289 .2278 .2268 .2258 .2248 .2238 .2229 .2219
17 .2209 .2200 .2191 .2181 .2172 .2163 .2154 .2145 .2136 .2127
18 .2118 .2110 .2101 .2093 .2084 ,2076 .2068 .2060 .2051 .2043
19 .2035 .2027 .2020 .2012 .ZOO4 .1997 .1989 .1981 .1974 .1967
20 .1959 .1952 .1945 .1938 .1931 .1924 .1917 .1910 .1903 .1896
21 .1889 .1883 .1876 .1869 .1863 .1856 .1850 .1843 .1837 .1831
22 .1824 .1818 .1812 .1806 .1800 .1794 .1788 .1782 .1776 .1770
23 .1764 .1759 .1753 .1747 .1742 .1736 .1731 .1725 .1720 .1714
24 .1709 .1703 .1698 .1693 .1687 .1682 .1677 .1672 .1667 .1662
25 .1657 .1652 .1647 .1642 .I637 .1632 .1627 .1622 .1617 .1613
26 .1608 .1603 .1598 .1594 .1590 .1585 .1580 .1576 .1571 .1567
27 ,1562 .1558 .1554 .1549 .1545 .1541 .1536 .1532 .1528 .1524
28 .1520 .1515 .1511 .1507 .I503 .1499 .1495 .1491 .1487 .1483
29 .1479 .1475 .1472 .1468 .1464 .1460 .1456 .1453 .1449 .1445
30 ,1441 .1438 .1434 .1430 .1427 .1423 .1420 ,1416 .1412 .1409
31 .1405 ,1402 .1398 .1395 .1392 .1388 .1385 .1381 .1378 .1375
32 .1371 ,1368 .1365 .1362 .1358 .1355 .1352 .1349 .1345 .1342
33 .1339 .1336 .1333 .1330 .1327 .1324 .1321 .1318 .1315 .1312
34 .1309 .1306 .1303 .1300 .1297 .1294 .1291 .1288 .1285 .1282
35 .1279 .1277 .1274 .1271 .1268 .1265 .1263 .1260 .1257 .1254
36 .1252 .1249 .1246 .1244 .1241 .1238 .1236 .1233 ,1231 .1228
37 .1225 .1223 .1220 .1218 .1215 .1213 .1210 .izon ,1205 .1203
38 .1200 .1198 .1195 .1193 .1190 .1188 .1186 .1183 .1181 .1178
39 .1176 .1173 .1171 .1169 .1167 .1164 .1162 .1160 .1157 ,1155
40 .1153 .1151 .1148 .1146 .1144 .1142 .1140 .1137 .1135 .1133
41 .1131 .1129 .1127 .1124 .1122 .1120 .I118 .1116 .1114 .1112
42 .1110 .1108 .1105 .1103 .1101 .1099 .1097 .1095 .1093 .1091
43 .1089 .1087 .1085 .1083 .1081 .1079 .1077 .1076 .1074 .1072
44 .1070 .1068 1066 .1064 .1062 .1060 .1058 .1057 .1055 .1053
45 .1051 .1049 .1047 .1045 .1044 .1042 .1040 .1038 .1037 .1035
46 .1033 .1031 .1029 .1028 .1026 .1024 .1022 ,1021 .1019 .1017
47 .1016 .1014 .1012 .1010 .loo9 .loo7 .loo5 .loo4 .10n2 .1000
48 .0999 .0997 .0996 ,0994 .0992 .0991, .0989 .0987 .0986 .0984
49 .0983 .0981 .0979 .0978 .0976 ,0975 .0973 .0972 .0970 .0968
50 .0967 .0965 .0964 .0962 .0961 .0959 .0958 .0956 .0955 .0953
51 .0952 .0950 .0949 .0947 ,0946 .0945 .0943 .0942 .0940 .0939
52 .0937 .0936 .0934 .0933 .0932 .0930 .0929 .0927 .0926 .0925
53 .0923 .0922 .0920 .0919 .0918 .0916 .0915 .0914 .0912 .0911
54 .0910 .0908 .0907 .0906 .0904 .0903 .0902 .0900 .os99 .0598
55 .OS96 .OS95 .OS94 .OS92 .OS91 .os90 .OS89 .OS87 .OS86 .0585
56 .OS84 .OS82 .OS81 .OS80 .OS79 .OS77 .OS76 .OS75 .OS74 .0572
57 .OS71 .OS70 .OS69 .OS68 .OS66 .OS65 .OS64 .OS63 .OS62 .0560
58 .0859 .OS58 .OS57 .OS56 .OS54 .OS53 .OS52 .OS51 .OS50 .0549
59 .OS48 .OS46 .OS45 .OS44 .OS43 .OS42 .OS41 .OS40 .OS38 .0537
60 .OS36 ,0835 .OS34 .OS33 .OS32 .OS31 .OS30 .OS28 .OS27 .0526
61 .OS25 .OS24 .OS23 .OS22 .OS21 .OS20 .os19 .OS18 .OS17 .0516
62 .OS15 .OS13 ,0812 .os11 .os10 .os09 .osos .OS07 .OS06 .0505
63 .OS04 .OS03 .os02 .os01 .os00 .0799 .0798 .0797 .0796 .0795
64 .0794 .0793 .0792 .0791 .0790 .0789 .0788 .0787 .0786 .0785
65 ,0784 ,0783 .0782 .0781 .0780 .0779 .0778 .0777 .0777 .0776
66 .0775 .0774 .0773 .0772 .0771 .0770 .0769 .0768 .0767 .0766
67 .0765 .0764 .0763 .0763 .0762 .0761 .0760 .0759 .0758 .0757
68 .0756 .0755 .0754 .0754 .0753 .0752 .0751 .0750 .0749 .0748
69 .0747 .0747 .0746 .0745 .0744 .0743 .0742 .0741 .0740 .0740
D. R. H O R N E R - A V E R A G E RESERVOIR PRESSURE 149

- -
@ .o .1 .2 .3 .4 .5 .6 .7 .8 .9

70 .0739 .0738 .0737 .0736 .0735 .0735 .0734 .0733 .0732 .0731
71 .0730 .0730 .0729 .0728 .0727 .0726 .0725 .0725 .0724 .0723
72 .0722 .0721 .0721 .0720 .0719 .0718 .0717 .0717 .0716 .0715
73 .0714 .0713 .0713 .0712 .0711 .0710 .0709 .0709 .0708 .0707
74 .0706 .0706 .0705 .0704 .0703 .0703 .0702 .0701 .0700 .0699

75 .0699 .0698 .0697 .0696 ,0696 .0695 .0694 .0693 .0693 .0692
76 .0691 .0691 .Of390 .0689 .0688 .0688 .0687 .0686 .0685 .0685
77 .0684 .0683 .0683 .0682 .0681 .0680 .0680 .Of579 .0678 .0678
78 .0677 .0676 .0675 .0675 .0674 .0673 .0673 .0672 .0671 .0671
79 .0670 .0670 .0669 .0668 .0667 .0666 .0666 .0665 ,0664 .0m4

80 .OM3 .0662 .0662 .0661 .0660 .0660 .0659 .0658 .0658 .0657
81 .0656 .0656 .0655 .0654 .0654 .0653 .0653 .0652 .0651 .0651
82 .0650 .0649 .0649 .0648 .0647 .0647 .0646 .0645 .0645 .0644
83 .0644 .0643 .0642 .0642 .0641 .0640 .0640 .0639 .0639 .0638
84 .0637 .0637 .0636 .0635 .0635 .0634 .0634 .0633 .0632 .0632

85 .0631 .0631 ,0630 .0629 .0629 .0628 .Of328 .0627 .0626 .0626
86 .0625 .0625 .0624 .0623 .0623 .0622 .0622 .0621 .0621 .0620
87 .0619 .0619 .0618 .0618 .0617 .0617 .0616 .0615 .0615 .0614
88 .0614 .0613 .0613 .0612 .0611 .0611 ,0610 .0610 .0609 .0609
89 .Of308 .Of307 .0607 .Of306 .0606 .0605 .0605 .0604 .0604 .0603

90 .0603 .0602 .0601 .0601 .0600 .0600 .0599 .0599 .0598 .0598
91 .0597 .0597 .0596 .0596 .0595 .0594 .0594 .0593 .0593 .0592
92 .0592 .0591 .0591 .0590 .0590 .0589 .0589 .0588 ,0588 .0587
93 .0587 .0586 .0586 .0585 .0585 .OB4 .0584 .0583 .OS3 .0582
94 ,0581 .0581 .0580 .OM0 .0579 .0579 .0578 .0578 .0577 .0577

95 .0576 .0576 .0576 .0575 .0575 .0574 .0574 .0573 .0573 .0572
96 .0572 .0571 .0571 .0570 .0570 .0569 .0569 .0568 .0568 .0567
97 .0567 .OM6 .0566 .0565 .0565 .0564 .0564 .0563 ,0563 .0562
89 .0562 .0562 .0561 .OM1 .0560 .0560 .0559 .0559 .055S .0558
99 .0557 .OM7 .0556 .0556 .0556 .0555 .0555 .OM4 .0554 .0553

100 ,0553

For @ > 100, 9 1


= - [.9228 +In @]
CD

APPENDIX C h Sand thickness


Jo Bessel function of the first kind of
List of Symbols order zero
Ji Bessel function of the first kind of
A. 13. V. Associated Reservoir Volume (see the order unity
text) k Permeability
a Mean radius of A. R. V. (assumed cylin- in Logarithm t o base e
drical) 1% Logarithm, base unspecified
Cb Shrinkage factor from reservoir oil log,, Logarithm t o base 10
at bubble point to tank oil. M. R. P Mean reservoir pressure
Ei The exponential integral function, de- N Volume of tank oil initially in place
fined by Nf The value of N corresponding to pressure
E
pr read from the compressibility curve
of the S.M. B. E.
Ni The value of N corresponding to pressure
m
pi read from the compressibility curve
e The base of Napierian logarithms = of the S. M. B. E.
2.7183 ... P Pressure
f Porosity PO Initial reservoir pressure
150 P R O C E E D I N G S F O U R T H W O R L D P E T H O L E U M CONGRESS-SECTION II/C

Pf The final pressure 01 the S. M. B. E.


(D - -4 k t
-
Pi The initial pressure of the S. 111. B. E. a’fpc
PE The static pressure eventually reached 0 A function of @, defined by either equa-
in a closed-in well. tion [13] or [14]
- 3 3
P The mean pressure in the A. R. V. c’ The Laplace operator = *-
d X
+ +
P* The pressure obtained by the extra-
polation t o infinite 0 of the straight
line portion of the pressure build-up Bibliography
curve, when p is plotted against lop
(1) M. Muskat: “Physical Principles of Oil Production ”.
to+
____ e 378 et seq.
e . (2) Ibid, 188, Eq. 6 (Assuming density to be a linear
q Well rate of production, assumed to be function of pressure)
constant (3) A. F. van Everdingen and W. Hurst: The Ap-
r Distance from well centreline plication of the Laplace Transformations to
I’b Radius of finite circular reservoir Flow Problems in Reservoirs, Truns. A I M E
rw Well radius (1949).
S Reservoir volume at pressure p oI tlic ( 4 ) I). R. Horner: Pressure Build-up in WeZZs. Proc.
cumulative production of oil, gas and Third World Petroleum Congrms, The Haguc,.
water of the well under examination. 1951.
S.M. €3. E. Subsidiary material balance equation
t Time This paper was presented on June 10th, 1955
t Datum time for block material balance by D. R. HORNER.
equation
tl Time at which a particular well is closed Discussion
in for BHP survey
- ~-
Well’s cumulative production W. R. VAN WIJK ( N . V . de Bataafsche Pe-
to
Last production rate before closing-in troleum Mij. - Wageningen, N L ) . Mr. Horner
U A real variable has asked for some suggestions f o r improving the
vt Reservoir volume at pressure p occupied calculations. I should like to put a question from
by unit volume of bubble point oil which Mr. Horner could perhaps derive such
and any gas liberated therefrom. a n improvement. He states that a constant
Vto Volume at initial reservoir conditions of
unit volume of bubble-point oil permeability is assumed for the oil in the field.
w Rate of fall of reservoir pressure during But, in the practical case quoted by Mr. Horner,
survey (assumed constant) we have a reservoir in which there are very im-
X A real variable portant pressure differences, and at lower pres-
tc Volume of the A. R. V. expressed as a sure oil releases more gas than i t does at higher
fraction of the entire reservoir volume pressures. Since the evolution of gas reduces
Kn The nth positive root (in order of in- the permeability t o oil, you may expect that
creasing magnitude) of Jl(x) = 0 you will have to take into account a systema-
B An arbitrary constant, later put equal tically varying effective permeability to oil even
N if the original (absolute) Permeability were quite
t o -.
NI’ constant throughout the field.
Y Euler’s constant = 0.5772... I should like t o ask Horner whether he has
A - _ _q
- _t _ any suggestions for taking this phenomenon into
xaahfc account.
6 Pressure drop a t distance r from a well
and time t after the commencement of D. R. HORNER:replies. With regard to the
production, due only t o the well problem that Prof. van Wijk raises, I must admit
~
itself t o not having thought of it directly, although I
6 The mean value of 6 over the A. Ft. V. must agree with him entirely when he says that
e Closed-in time systematically varying oil permeabilities will
tL Viscosity exist. I have considered this only as a part of
x = 3.1416 the more general problem in which neither per-
\x Summation meability, porosity, compressibility nor any of
D . H. HORNER-AVERAGE RESERVOIR PRESSURE 151

the other parameters are constant. All these that you are probably quite right on both points.
things tend t o make the build-up curves di- Perhaps I may say why so many wells were
verge considerably from the theoretical form. omitted. This was a matter of finance, as it
This is the sort of practical difficulty which we was costing a lot of money t o do this survey,
get in the field with this type of study. and I did not feel justified in going t o extreme
At the beginning of the build-up curve, one lengths on a first trial.
gets a “ swans-neck ” period, and I think that In reply to the next remark, the only sug-
it is during this period that van Wijk’s varying gestion which I have for avoiding this error is
permeability is going t o be of major importance. t o repair and survey all wells. I believe this
I believe that once we get into the region of t o be economically justified in many cases even
straight-line build-up the effect of this varying when workovers means are expensive.
permeability will not be so very serious. 1 As for the proposed use of an average pres-
would ask you t o remember that the extrapola- sure sink for the case of a well which has not
tion of these straight lines over a tremendous been surveyed, this seems t o me to be some-
distance is so extremely doubtful anyway, and what dangerous. The basic equation:
there are so .many other factors which can give -
rise t o the most erratic behaviours in the build-up p=p*-YA
curves that I rather doubt whether any system- implies that the mean value of pis the mean value
atic improvement could be made along the lines of p * minus a mean value for YA correction.
suggested by van Wijk, although it is certainly Now p * is the important quantity. It is the
something t o be looked into. To put the prob- one which the well gives you directly, but it is
lem simply: suppose that a well gives us a unfortunately not known with great accuracy
straight line. If it does, we may hope (perhaps although it is the p * which makes the major
fallaciously) that its slope is a fair value, and contribution t o the 6. The YA is only la cor-
that we may use the value which the well gives rection term. Now if a well is omitted from the
as a fair mean for the conditions which existed survey, its p* is lost, and if that quantity is
in the well a t that time. lost I am not certain that there is any logical
J. VAN HEININGEN(Koninklijke Shell La- justification for trying t o estimate a mean PA.
boratorium - Amsterdam, N L ) . In connection If it is argued that the mean value of p* is the
with the practical test of the method, Mr. Hor- same for the surveyed as for the unsurveyed wells,
ner says that 29 yoof the wells were not surveyed. it might just as well be said that the mean value
He also says that some of the older wells could of p is the same for both groups of wells, and
not be satisfactorily surveyed a t all, due t o the this is what the paper does for Casabe. What I
presence of subsurface leaks which could not should like t o emphasize is that p* is the major
be easily located. This would make one think source of error, and not YA.
that the older wells may perhaps have been I should like t o make a point here about
unfairly excluded from the survey, as this may averaging processes in general, because in oil
have unfairly weighted the results, perhaps fields many daily jobs are done by arithmetic
with the result of giving somewhat too high a averaging. This is usually adequate for routine’
mean pressure. Does the author agree and operations, but I do not think it is good enough
has he any suggestion for the avoidance of this for this sort of problem. I think it is very
possible source of error? Finally, has he in doubtful indeed whether taking arithmetic aver-
particular given any thought t o the possibility ages of rather, complicated expressions such as
of evaluating an average correction for an aver- YA or --__ CIP
age pressure sink t o be used with his unsurvey- 4 x k h can have any significance. I feel
ed wells ? that the type of average that would have t o
be used would be quite complicated.
D. R. HORNERreplies. To your first re-
mark, Dr. Van Heiningen, which was that per- J. BIRKS(British Petroleum Co. Ltd. - North
haps some of the older wells have been unfairly Newark, Notts, G B ) . This seems t o be a very
excluded, and that this may have resulted in worrying paper, because it is also a very good
too high a mean pressure: Yes, Sir, I think one and ‘Mr. Horner should be thanked for his
152 PROCEEDINGS FOURTH WORLD PETROLEUM CONGRESS-SECTION II/C

careful thought and detailed mathematical ar- the pressures measured in such fields are
guments. What is really worrying and will pressures in the fissures.
intimately concern most of our reservoir engin- Since the distances between fissures are
eers, is that in the past they have managed usually less than 100 feet there is virtually
quite well and have been very satisfied with complete pressure equilibrium between fis-
their agreements between oil in place calculated sures and rock matrix. Perhaps Mr. Horner
by material balance and oil in place calculated would like t o comment this point.
by the volumetric method. T have the feeling Finally I would say that the method proposed
that although Mr. Horner has made a good for determining oil in place should not replace the
logical case for his method, i. e., that the pressure volumetric method since both methods require
sinks should be included as well as the average sampling of a large number of wells. Would
extrapolated reservoir pressure, it is not a prac- the author agree that, in the fissured type of field,
tical one in general. I feel that Mr. Horner had this method ceases t o have any application ?
a difficult field t o deal with where there was a
high viscosity oil, and this amount of mathemat- D. R. HORNERreplies. I first of all wish
ical argument has probably been necessary t o to thank Dr. Birks for his complimentary re-
get any reasonable agreement between the marks. Birks said that a lot of reservoir engin-
volumetric and the material balance method?. eers have been happy with their material bal-
One of the fundamental requirements of the ances in the past, and suddenly some one comes
method is that a detailed pressure survey should along and says that their method is wrong.
be made in a large number of wells in order t o That is not wholly true, for what I am proposing
calculate the original oil in place. If there are is a refinement of the conventional method for
a large number of wells, there appears no reason the case of an older field. It is only in old
why they should not have been adequately fields (or rather, fields with large withdrawals)
sampled either by coring or by logging tech- that pressure sinks are so big that i t becomes
niques. Since in his conclusions the strength advisable t o worry about correcting for them as I
of the method relies on its agreement with the have done. The case of a new field reduces t o
volumetric one, I would suggest that in itself YA = O a n d s o p = p * .
it is not a practical method for determining the In the particular case of Casabe, we had
original volume of reservoir oil. We must produced something of the order of 50 % of
bear in mind that Mr. Horner has been very the expected recovery of the block and so it
fair about the limitations of the method par- was really a very old field for a material balance,
ticularly in those fields which have variations and even there the average YA correction was
in permeability, variations in oil characteristics, only about 80 psi. in a block with a mean pres-
a partial water drive or an original gas cap. It sure of about 1000 psi. The effect is not too big.
is appalling t o think of the ‘‘ mathematical An engineer who had done a material bal-
tricks ” that must be necessary t o solve these ance without including the YA corrections,
particular cases. but who had extrapolated his build-up curves
However we must remember that pressure t o their p* values would have found about
surveys must be made for a very good practical 134 million barrels in place instead of 107 mil-
reason, that is, to maintain a reasonable con- lion. However, that same engineer perhaps
trol of the gas oil level and the oil water level in may not have extrapolated the build-up curves
oil fields. It is’ usually this data on which re- to obtain p* values. He may perhaps have
servoir engineers have t o base their calculations used spot readings of the bottom hole pressures
and not on the specialised reservoir pressure made after a few hours or days closed in, without
programme. further extrapolation. It is possible that this
I do not wish Mr. Horner t o feel that I dis- engineer could have obtained pressure readings
agree with the subject matter of his paper in this manner which were of the order of 80
because I consider that it is an excellent piece psi less than the p* values. In this case two
of mathematical logic. compensating errors would have cancelled each
My own particular interest is in the fissured other, and he may well have come very near t o
type of field and I would like t o mention that the true mean pressure.
Birks says that the method is not practical even today many operators core wholly ina(lr-
for two reasons. @ne is that the mathematics quately. T do not, of course, appeal to tho
is involved. and one is the pas and/or water volumetric calculations to justify the material
problem. balance - nor d o I use the material balance equw -
As for the amount of computation involved, tion to justify the volumetric calculations. Thrl
I think it is surprising that thP computations point is that the fact that both methotls agrw
can hcl done almost entirely by comparative1)- is grounds for confidence that the)- arr hotli
untrained. labour; the computing time runs a t correct. [ feel that Birks has allowed himselC
between one and two hours per well using a desk l o be influenced surriewhat by his own partic-
calculator. I do not think this is unpractical as ular interest in fissured limestone fields, for which
far as the amount of arithmetic is concerned. 1 believe the method to be wholly inapplicable.
The paper does sa?; specifically that the> Whereas in such fissured limestones pressurc'
method is not designed to cope with Ihe gas and/or survey may be made for gas/oil and oil/water
water problem, but I don't think t h e water drive contact observation purposes, this docs not
case would be difficult to handle. apply in the more common sandstone reser-
I do not know about the gas cap problem, voirs of my own experience, wlierr pressure
and I a m rather unhappy about it. I havc surveys are frequently instigated f'(ir material
said i n the paper that it could perhaps he han- balance purposes.
clled, but I a m now not reallysure that i t could br. I a m not proposing that thci niaterial h 1 -
With regard to the coring and logging of a ance equation should s u p
large number of wells, i t is surely an accepted metric one. nor can I agree that the volumetric
principle that sampling is rarely. if ever. adc- one should supersedp the material balancc eqiin -
quate. Furthermore, many older fields exist tion. Both have their uses, and i f both can
which have never been adequatel>- core(1, and be rmployetl satisfactoril\-, so milch thc he tter.

Vous aimerez peut-être aussi