This action might not be possible to undo. Are you sure you want to continue?
info@serafimltd.com
P. +44 (0)2890 421106
www.serafimltd.com
NPV Formulas
A Fast Method for Calculating
and Optimising NPV in Screening
Studies
Prepared by: Serafim Ltd
NPV Formulas
i
Contents
Contents ......................................................................................... i
Summary ....................................................................................... 1
Uses of the equations....................................................................... 2
An example.................................................................................. 2
Creating an analytic description of a typical economic model................. 3
Other economic assumptions.......................................................... 3
Reserves, well rates and decline rates ............................................. 4
Derivation of the formulas ................................................................ 5
Understanding the optimal number of wells ........................................ 6
Visualising the NPV formula .............................................................. 7
Using the formulas to derive the optimal injector:producer ratio............ 8
Uncertainties in key parameters ........................................................ 9
Extending the method to maximise rate of return...............................10
Dealing with a development spread over several years........................11
Appendix – Mathematical proofs of the formulas.................................11
Assumptions and Definitions .........................................................11
Theorem 1 – Expressing NPV as a function of well numbers ..............12
Proof .......................................................................................12
Theorem 2 – Number of wells giving the highest NPV.......................13
Part I ......................................................................................13
Part II .....................................................................................14
Proof of Part I...........................................................................14
Proof of Theorem 2, Part II ........................................................16
Lemma A .................................................................................17
Lemma B .................................................................................17
Theorem 3 – NPV for a development with the approximately optimal
number of wells...........................................................................19
Proof .......................................................................................19
NPV Formulas
1
Summary
When quick estimates of NPV are needed for oilfield development
options, it is usual to calculate the economics in a spreadsheet model,
using a production profile generated by means of a set of simple
assumptions (e.g. typical initial well rates, STOIIP, recovery factor etc).
Because
 in most countries, oilfield economics can be reduced to linear factors
(for tax effects) and discounting;
 discounting and exponential decline can be combined into a single
exponential function;
it is possible to take a set of assumptions, similar to those used in a
typical spreadsheet model, and derive the following equations (N.B.
commercial use is subject to patent protection)
D
2
q
E . E d . C
q
.E
 L R. wells of number with this NPV
q
E . E d . C
q
E . E d . C
q
.E
 L
.
q
R.d
NPV maximises that wells of Number
−


¹

\

α − +
−
α
=
α − +


¹

\

α − +
−
α
=
where α = (1+d)
Tab
and T
ab
(abandonment time)=

¹

\

E
q . L
ln .
N . q
R
and
q
– initial oil production per well per year, averaged over all wells,
including injectors
N – total number of wells, including injectors
R – technical reserves i.e. the amount of oil that could be recovered if the
field were run for a very long time
L – net revenue per tonne of oil (i.e. after all taxes and royalties,
including profit tax)
d – discount rate
C – net capital cost per well
D – net capital costs not related to numbers of wells, e.g. roads and
pipelines
E – net opex per well
Note: In this context, “net” means expressed in terms of effect on
present value, after all taxes and royalties
NPV Formulas
2
Uses of the equations
These equations offer a number of advantages when used in addition to
existing methods:
 quick and easy to use, especially for screening prospects and initial
concepts, in the early stages of a project
 easily auditable (in contrast to spreadsheets, which are often very
difficult to fathom)
 can provide a totally independent check on project economics
 can help save time in determining the correct well density
 can help to provide an overview of the common task of almost all
petroleum engineers – optimising the economic return.
The equations can also be extended to give the optimal number of wells
when the objective of the oil company is to maximise the overall rate of
return.
In a trial, when applied to a Russian field, the equations quickly yielded
very similar results to one month of conventional work.
An example
Consider a field with the following technical and economical parameters:
Technically recoverable reserves = 100 million barrels
Net oil price = $8 per barrel
Discount rate = 8%
Net fixed capex costs = $20,000,000
Two development options are considered:
a) Vertical wells
Net capex per well = $2,000,000
Net opex per well per year = $300,000
Average initial well rates = 500 barrels per day = 150,000 barrels per
year
b) Horizontal wells
Net capex per well = $5,000,000
Net opex per well per year = $500,000
Average initial well rates = 1,500 barrels per day = 450,000 barrels
per year
The formulas quickly give
a) Vertical wells
Optimal number of wells = 35
NPV with this number of wells = $122,000,000
(Time to abandonment = 26 years)
b) Horizontal wells
Optimal number of wells = 19
NPV with this number of wells = $209,000,000
(Time to abandonment = 23 years)
NPV Formulas
3
One can see that these results match the way that horizontal
developments typically need fewer wells than vertical ones do.
Creating an analytic description of a typical
economic model
In order to derive algebraic expressions for NPV and for the economically
optimal number of wells, it is necessary to have an algebraic model of all
the economic factors, including taxes and royalties. It is suggested that
the following is an appropriate model for most tax regimes:
NPV = Σ (L.Qi + α.OPEXi + β.CAPEXi) / (1+d)
i1
where the summation is over i=1 to i=n, the year of the end of the field
life
OPEXi = OPEX in year i
CAPEXi = CAPEX in year i, not including VAT
Qi = oil production in year i
L = a linear factor, the “net oil price” – it represents the benefit in
present value terms of one extra unit (ton, cubic metre or barrel) of oil
sold
α = a linear factor, representing the effects on NPV of one extra real
terms dollar (or whatever unit of currency) of expenditure on OPEX
β = a linear factor, representing the effects on NPV of one extra real
terms dollar of expenditure on CAPEX
The terms α.OPEXi and β.CAPEXi can be considered to be the “net OPEX”
and the “net CAPEX”.
It is reasonable to model the NPV as a linear function for two reasons
i) although, as can seen from economics spreadsheet models, the
interplay of costs, revenues and taxation is very complicated,
usually all the individual steps in the spreadsheets are linear;
hence combining them will give a linear function
ii) if a taxation system was not linear in its effects, there would be
benefits in artificially splitting or combining projects so as to take
advantage of the nonlinearities; such a possibility would quickly
become well known. Hence, from initial design or from progressive
correction of anomalies, most tax systems have evolved to be
linear.
Other economic assumptions
The other major assumptions about the economics are that
 it is desired to achieve the highest possible NPV (as opposed to the
highest rate of return, for example);
 capex expenditure for the field consists of a fixed element (e.g. cost
of platform, road or pipeline) and a cost per well;
 all capex expenditure occurs in Year 1;
 opex is proportional to the number of wells.
NPV Formulas
4
In analytic models, it is usually necessary to define in advance what
quantity is to be minimised or maximised. (In contrast, in spreadsheet
models, it is possible to run through all the various scenarios and then
choose the one which gives the overall best results. For example “Putting
in 15 wells rather than 10 increases the NPV by only 5%, while it
increases the capex by 40%. On balance, we think that the increase in
NPV is not worth it, given the overall uncertainties, so we will go for the
10 well option.”) In this article, NPV will be maximised, firstly because it
is the simplest, and, secondly, because it is easy to modify the NPV
maximising results to cope with the situation where there is a shortage of
capital, and one is only interested in returns of more than 15%, for
example. The modification required is described later in the article.
The first assumption about capex are that net capex can be expressed as
Net capex = C.N + D
where C is net capex per well and D is a fixed element. The second
assumption is that all the capex expenditure can be considered to take
place instanteously at the beginning of the project. Clearly, this is not
always realistic, but, again,
 it is simpler;
 it is unlikely that the time required to carry out the capital
expenditure will affect questions such as “What is the optimal number
of wells?”
 it is easy to modify the model to deal with capex spread over several
years (again, the modification is described later in the article).
With regards to opex, it should be noted that, whenever the gross liquid
rate is constant, the simple model gives the same results as a more
complicated model, in which opex contains one element proportional to
the number of wells and one element proportional to the gross liquid
production.
Reserves, well rates and decline rates
There are three important petroleum engineering assumptions
 technically recoverable reserves are independent of the number of
wells
 initial well rates are independent of the number of wells
 the field follows exponential decline from the start of production.
Unless there are geological features like isolated fault blocks, the number
of wells does not much affect the technical recovery factor. In theory, for
a field producing under depletion, a single well could, over a very long
time, drain the field as efficiently as twenty wells. One reason for this is
that, whatever the well spacing, the area of the field that is drilled up
(i.e. directly penetrated by a wellbore) is very small in relation to the
total area. With vertical wells, a well spacing of 500m and a wellbore
diameter of 6 inches gives a ratio (total well bore area): (area of field) of
1 x 10
7
. With ten times as many wells, this would still only give a ratio of
1 x 10
6
, so similar recovery factors should apply.
It can be seen from the formula for the steadystate productivity index
(PI) of a vertical well
NPV Formulas
5
PI = (constant x K
o
.h) / [(ln(r
e
/r
w
)+S) x B
o
.µ
o
]
and the slightly more complicated formula for horizontal wells (Refs 1 and
2), that term depending on well separation, r
e
(the effective drainage
radius) only occurs as a logarithmic terms, hence the effect of changes of
it are usually small. So it is reasonable to assume that PI is not much
affected by changes in well spacing.
Even if PIs are constant, it is does not necessarily follow that well rates
are constant, because one well may reduce the average pressure seen by
the other wells. However, if one scales a whole development, including
the number of injectors, then the average pressures should remain the
same, giving constant well rates.
If fraccing is used, it may be incorrect to assume that PIs are
independent of well numbers. Fraccing can be considered to give a large
negative skin, so it can be the case that, even if the proportional change
in ln(r
e
/r
w
) is small, the change in [(ln(r
e
/r
w
)+Skin] may be significant.
The assumption that initial well rates are independent of well numbers
also holds for gas fields being produced under depletion drive. Here,
increasing the number of wells does give pressure interference. However,
providing all the wells start production at the same time, this effect is
captured by the decline rate.
The other major petroleum engineering assumption is that the field
follows exponential decline from the very beginning. There may be two
objections to this asssumption. Firstly, many fields show a plateau
production period. Secondly, once decline starts, it may be exponential,
but changes to hyperbolic decline in the later stages of field life.
In answer to the first objection, it can be argued that many of the fields
showing a production plateaus did so because of staged development. If
this is the case, then the effect can be captured by the modification to
the model described later. However, if the plateau period is a genuine
subsurface phenomenon (maybe a period of stable, fieldwide dry oil
production, before water breakthrough), then the model may require
further development.
The second objection should be less of a concern. A change to hyperbolic
decline usually occurs sufficiently late in field life that, with discounting, it
has little effect on the economics of the field, and on the question of what
is the economically optimal number of wells. All that is required for the
model to work correctly is to use a modification to the reserves – namely
the total production at time=infinity if expontial decline had continued for
the whole life of the field.
Derivation of the formulas
The derivation of the formulas was done in four steps
1) By simple integration, derive a formula for NPV as a function of
(amongst other things) well numbers.
2) Calculate the optimal number of wells by taking partial derivatives
with respect to number of wells and abandonment time.
3) Show that the formula for the optimal number of wells can be
approximated by a simpler expression.
NPV Formulas
6
4) Calculate the NPV for the approximation of the optimal number of
wells.
A full mathematical proof of these results is given in the appendix.
Understanding the optimal number of wells
One of the big advantages of formulas, rather than spreadsheets, is that
they give more of an insight of how all the factors interact. Consider the
formula for expressing NPV as a function of the number of wells. This
gives a simple explanation to the question of what is the optimal number
of wells, a question that is usually answered in vague terms, based on
experience of analogous fields.
If one ignores abandonment (which is reasonable when one is trying to
gain an understanding rather than calculate a specific value), NPV can be
expressed as
D
d
E
C . N
1
d . R
q . N
1
1 . L . R NPV − 
¹

\

+ −




¹

\

+
− =
The first R.L term (reserves x net oil price) corresponds to the value of
the oil in the ground, i.e. the value of the field if the oil could be taken all
at once at no cost.
The
1
d . R
q . N
L . R
+
term represents the value lost because of the discounting
effect of the time taken to get the oil out of the ground. Clearly, if the
number of wells is very large, this term becomes very small – the field is
produced in a short space of time. On the other hand, if the number of
wells is very small, then this term becomes almost as big as R.L, the total
value of the oil – the field is produced very slowly, so the NPV of the
production is very small.
The D term represents the value lost because of net fixed capex (roads,
pipelines etc). The N.(C+E/d) term represents the value lost because of
capex per well and opex. Clearly, this is linear with the number of wells.
Comparing a set of different scenarios, each with a different number of
wells, it can be seen that the value of the oil in the ground and the net
fixed capex do not vary, but the other two terms do. As the number of
wells increases, the bulk of the oil is produced sooner, so less value is
lost because of time effects, but opex and variable capex costs are
increasing. The optimal number of wells is reached when the sum
(value lost in time effects) + (value lost because of opex and variable
capex expenditure)
is at a minimum. This is illustrated in the plot below, for the case of a
field with the following parameters:
Technically recoverable reserves = 100 million barrels
Average initial well rate = 2,800 bbl/day
NPV Formulas
7
Net capex per well = $5,000,000
Net opex per well = $400,000
Net oil price = $4.00
Discount rate = 8%
Effects of changing the number of wells
$0
$50,000,000
$100,000,000
$150,000,000
$200,000,000
$250,000,000
$300,000,000
$350,000,000
$400,000,000
$450,000,000
$500,000,000
0 5 10 15 20 25 30 35 40
Number of wells
P
r
e
s
e
n
t
V
a
l
u
e
Value lost to time effects
Opex and variable capex
Sum of value lost
It can be seen that the optimal number of wells is about 10.
Visualising the NPV formula
To gain a feel for the significence of the NPV formula (for the optimal
number of wells), it may be helpful to consider the formula as follows.
Imagine a square whose area represents the total value (to the oil
company) of the field, if all the oil could be produced immediately at zero
cost i.e. R x L. Let the length of the sides of the square correspond to √L.
If we ignore abandonment effects for the moment, then the scalable
capex and the opex reduce the value to
2
q
E d . C
L . R


¹

\

+
−
This is the area of the square that can be fitted in the R.L square if there
has been already inserted a square of length √[(C.d + E)/q].
Then the value has to be further reduced to pay for the fixed capex, D,
represented by a suitable rectangle. The final step in building up the
picture is to allow abandonment effects to introduce a gap between the
two smaller squares.
NPV Formulas
8
Lost because of time effects
Scalable CAPEX and
OPEX
Unscalable CAPEX
NPV
Value of the hydrocarbons
Such a geometric representation is scarely any more intuitive than the
formula itself, but it is put forward in the hope that it may a starting point
for better visualisations or interpretations of the formula.
Using the formulas to derive the optimal
injector:producer ratio
Another advantage of formulas is that they can be in further calculations.
For example, in a waterflood, given constraints on well bottomhole
pressures, it is straightforward, through a material balance argument, to
express the expected average well rate (averaged over all wells, both
producers and injectors) as a function of the injector:producer ratio,
independently of the number of wells. Feeding this expression into the
NPV formula gives an expression of NPV as a function of injector:producer
ratio. This expression can be diffentiated w.r.t. the injector:producer
ratio, to give a formula for the optimal injector producer ratio as follows
(proof not given in this article)
Optimal injector:producer ratio =
) Ei d . Ci .( Bw . II
) Ep d . Cp .( Bo . PI
+
+
where
PI = productivity index of average production well
NPV Formulas
9
II = injectivity index of average injection well
Bo = oil formation volume factor
Bw = water formation volume factor
Cp = net capex cost per producer
Ci = net capex cost per injector
Ep = net opex per year for each producer
Ei = net opex per year for each injector
d = discount rate
Uncertainties in key parameters
By reducing the complexity of the economics calculations, the formulas
make it possible to examine a wider range of uncertainties in key
parameters. For example, it is easy to generate plots such as the one
below, to compare the overall economic performance of horizontal and
vertical wells, when there is a big uncertainty in the Kv:Kh ratio.
Development Options for Field X  NPV vs Kv_Kh
$35,000,000
$36,000,000
$37,000,000
$38,000,000
$39,000,000
$40,000,000
$41,000,000
$42,000,000
$43,000,000
$44,000,000
$45,000,000
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Kv_Kh
N
P
V
(
U
S
$
)
Slanted injs; horiz prods
Vertical wells
Horizontal wells
Slanted wells
It can be seen that, for such a field, horizontal wells would be preferably
to vertical wells providing the Kv/Kh ratio is greater than 0.02. Such
analysis can help identify the key uncertainties to be addressed during
field appraisal.
NPV Formulas
10
Extending the method to maximise rate of
return
A simple modification to the Capex figures makes it possible to use the
formulas when the company objective is to maximise overall return on
capital employed.
The formulas as they stand show the number of wells required to
maximise NPV. In the real world, oil companies are faced with the
problem of constraints. In particular, the companies usually have only a
fixed amount of capital available, or at least limits on the amount of
capital employed (gearing limits etc). The problem then is to maximise
NPV, subject to the limits on the amount of capital employed. Assuming
that there are sufficient number of projects available and no other
constraints intrude (such as limits on management capacity), the method
to do this is to rank all the projects in order of decreasing NPV:Capex
ratio, and then choose the projects in this order until one has reached the
limit on the amount of capital employed.
For the projects chosen, let H (the “hurdle” rate) be the lower limit to the
NPV:Capex ratio i.e. the NPV:Capex ratio for the last project chosen.To
determine, for a new project, the optimal number of wells, it is useful to
consider the project as consisting of
(Minimalist development option) + Series of increments to the mininalist
development option.
(Note that the split is purely conceptual; all the increments start at time
= 0).
All the increments can be analysed as if they were separate projects.
They pass the screening criteria if their NPV:Capex ratio is greater than H
i.e. δNPV / δCapex ≥ H
So, while the NPV/Capex ratio for additional wells is greater than H, the
wells are worth adding to the development scheme. The optimal total
number of wells is reached when the limit is reached, so the criteria for
determining the total number of wells is
δNPV / δCapex = H
This can be converted into a more workable criteria as follows
δNPV / δCapex = H
⇔ δ(NPVH.Capex) / δCapex = 0
⇔ [δ(NPVH.Capex) / δN] x [δN / δCapex] = 0 where N is the number
of wells
⇔ δ(NPVH.Capex) / δN = 0 since δN / δCapex ≠ 0 (it never costs
an infinite amount of money to drill a new well).
Hence, the new problem (find the number of wells that maximises
corporate NPV subject to a limit on capital employed) can be converted
into the old (find the number of wells that maximises NPV) by using an
artifical NPV, defined to be
NPV′ = NPV – H.Capex = NPV – H.(C.N + D)
NPV Formulas
11
and artificial C′ and D′ defined to be
C′ = (1 + H) x C
D′ = (1 + H) x D
Then
D N C H
q
E E d C
D N C H V NP
q
E E d C
q
E E d C
− +


¹

\

− + ′
− = + + ′ =
− + ′


¹

\

− + ′
−
=
. .
2
. .
q
.E
 L R. ) . .( wells of number with this NPV
. .
. .
q
.E
 L
.
q
R.d
N wells, of number Optimal
α α
α
α α
(the –H.D and +H.D terms cancel each other out).
Dealing with a development spread over
several years
If a field is very large, so that the development is spread over several
years, with different areas coming on stream in different years, then the
situation can be modelled as follows.
Let NPV(N) be the NPV that would have been achieved if the development
had been had carried out in the space of one year.
Let M be the number of years that will be required. (Assume M to be
independent of N, the number of wells).
The new problem is to find N to maximise
NPV′ = ( ) ) N ( NPV .
d . M
) d 1 ( ) d 1 (
d 1
M
) N ( NPV
) 1 M (
M
1 i
) 1 i (
− −
=
− −
+ − +
= +
∑
Clearly, the value of N that maximises NPV(N) also maximises NPV′ and
the above equation gives NPV′. So the problem is solved.
Appendix – Mathematical proofs of the
formulas
Assumptions and Definitions
Let us make the following assumptions to describe an oilfield
development
1) Let N be the number of wells drilled. All the wells start up at time t=0.
2) If the field were run for an infinite time, the total production would be
R (the technically recoverable reserves), independent of the number
of wells.
3) The initial production rate per well is q, independent of the number of
wells i.e. it is not affected by well spacing.
NPV Formulas
12
4) The field oil production rate follows exponential decline i.e.
Field oil production rate = initial rate x e
at
= N.q.e
at
5) The net oil price is a constant, L, after all taxes and deductions.
6) The net capital costs can be expressed as D + C.N; all capital
expenditure happens at time t=0.
7) Net opex can be expressed as E.N per unit time. (The year is probably
the most appropriate unit of time, but any unit can be used, providing
it is the same for both E and q).
Theorem 1 – Expressing NPV as a function of well
numbers
The NPV of a field run until abandonment can be expressed as
( ) ) D N . C ( ) 1 .(
) d 1 ln(
E . N
e . 1 .
N . q
) d 1 ln( . R
1
L . R
NPV
R / Tab . N . q
+ − α −
+
− α −

¹

\
 +
+
=
−
Proof
NPV can be broken down into the component parts of the cashflow
NPV = NPV(revenue stream) + NPV(Opex) + NPV(Capex)
where the NPV(Opex) and NPV(Capex) are, of course, negative.
Start by calculating the NPV of the revenue stream:
By the assumption that there is exponential decline
Oil production rate = N.q.e
at
= N.q.e
(q.N/R)t
[Since, by the definition of technical reserves
a / q . N e .
a
q . N
dt e . q . N R
0
at
0
t . a
=
(
¸
(
\

− = =
∞
−
∞
−
∫
then a = N.q/R ]
Oil revenue per unit time = (production rate) x (net oil price) = N.L.q.e

q.N.t/R
By the definition of NPV
NPV Formulas
13
( )
( ) ( )
( )
R / Tab . N . q
R / Tab . N . q Tab R / Tab . N . q Tab ). d 1 ln(
Tab )). d 1 ln( R / N . q (
Tab
0
t )) d 1 ln( R / N . q (
) d 1 ln( . t ) ) d 1 ln(( t
Tab
0
t )). d 1 ln( R / N . q (
Tab
0
t
e . 1 .
N . q
) d 1 ln( . R
1
L . R
e . ) d 1 ( 1 .
N . q
) d 1 ln( . R
1
L . R
e . e 1 .
N . q
) d 1 ln( . R
1
L . R
e 1 .
) d 1 ln(
R
N . q
q . L . N
)) d 1 ln(
R
N . q
(
e . q . L . N
) e e ) d 1 ( (Since
dt e . q . L . N dt
) d 1 (
unit time per Revenue
stream revenue of NPV
t
−
− − − + −
+ + −
+ + −
+ − +
+ + −
α −


¹

\
 +
+
=
+ −


¹

\
 +
+
= −


¹

\
 +
+
=
−

¹

\

+ +
=
+ + −
=
= = +
=
+
=
−
∫ ∫
Looking at the Opex cashflow
Opex per unit time = N.E
( )
( ) ( ) α −
+
− = − −
+
− =
−
+
− =
+
= − =
+
−
=
−
+ −
+ −
+ −
∫ ∫
1 .
) d 1 ln(
E . N
) d 1 ( 1 .
) d 1 ln(
E . N
e 1 .
) d 1 ln(
E . N
) d 1 ln(
e . E . N
dt e . E . N dt
) d 1 (
E . N
opex of NPV
Tab
Tab ). d 1 ln(
Tab
0
t ). d 1 ln(
Tab
0
t ). d 1 ln(
Tab
0
t
Looking at Capex, since all the capital expenditure is assumed to occur at
time t=0,
NPV of Capex =  (C.N + D)
By adding NPV(Revenue), NPV(Opex) and NPV(Capex), one obtains the
desired formula.
Q.E.D
Theorem 2 – Number of wells giving the highest NPV
Part I
The number of wells giving the highest NPV, N
opt
, can be calculated
(iteratively) from the expressions
NPV Formulas
14
( )
( )( )

¹

\

=
+ = α
α − + + α −

¹

\

α
= ε
(
(
¸
(
¸
− ε − + ε
α − + +
α −
+ =
−
E
q . L
ln .
N . q
R
Tab and
) d 1 ( and
E . E ) d 1 ln( . C . E . q . L 4
E
q . L
ln . E .
where
1 1 x
E . E ) d 1 ln( . C
E . q . L
). d 1 ln( .
q
R
Nopt
Tab
2
Part II
If
a) L.q ≥ 2.[C.ln(1+d) + E] (i.e. the well is reasonably profitable at first)
and
b) 0.2 ≥ α (equates, for d = 8%, to a field life ≥ 20 years)
then the approximation of setting ε = 0 gives approximate value for the
number of wells, N
approx
within the following bounds
1.31 N
opt
≥ N
approx
≥ N
opt
Proof of Part I
Consider NPV as a function of well numbers and abandonment time. For a
normally behaved function
0
Tab
NPV
such that Tab
0
Tab
NPV
N
NPV
maximum a is NPV
=
∂
∂
=
∂
∂
=
∂
∂
⇔
can be obtained by taking the partial derivatives or, more simply, by
seeing that this occurs when revenue has dropped until it equals
operating expenditure
i.e. N.L.q.e
q.N.Tab/R
= N.E
so –q.N.T
ab
/R = ln(E/(L.q))
so Tab = R/(q.N) . ln(L.q/E)
(N.B Also, for Tab optimal, e
q.N.Tab/R
= E/(L.q))
For well numbers, using Theorem 1, and writing out the expression in full
(not using α, for example)
NPV Formulas
15
( )
( )
( )
( )
( )
( )
( )
( ) 
¹

\

− + −
+ +
+
+ −

¹

\

+ +
−
−
+
+ −
− + −
+ +
=




¹

\

−
(
(
(
(
¸
(
¸
−
+
+ −
− + −
+ +
∂
∂
=
∂
∂
− −
− −
−
− −
−
− −
R
Tab . q
. e . ) d 1 ( 1 .
) d 1 ln(
R
qN
Lq . N
R
q
. e . ) d 1 ( 1 .
) d 1 ln(
R
qN
Lq . N
C
) d 1 ln(
) ) d 1 ( 1 ( E
e . ) d 1 ( 1 .
) d 1 ln(
R
qN
Lq
D C
) d 1 ln(
) ) d 1 ( 1 ( E
e . ) d 1 ( 1 .
) d 1 ln(
R
qN
Lq
. N
N N
NPV
R / qNTab Tab
R / qNTab Tab
2
Tab
R / qNTab Tab
Tab
R / qNTab Tab
Setting this partial derivative equal to zero, multiplying both sides of the
resultant equation by (qN/R + ln(1+d))
2
, then using α as shorthand for
(1+d)
Tab
, and the relationships that apply when T
ab
is optimal, e
q.N.Tab/R
=
E/(L.q) etc, we get the following equation. N.B. This equation only applies
when T
ab
is optimal.
( )
0
q . L
E
. . ) d 1 ln(
R
N . q
. q . L .
E
q . L
ln
q . L
E
1 . q . L .
R
qN
C
) d 1 ln(
1 . E
. ) d 1 ln(
R
N . q
q . L
E
. 1 . q . L . ) d 1 ln(
R
N . q
2
= α 
¹

\

+ + 
¹

\

+


¹

\

α − −


¹

\

+
+
α −

¹

\

+ + −


¹

\

α − 
¹

\

+ +
Cancelling terms and multiplying both sides by –1 gives
0 ) E . q . L ).( d 1 ln( ) d 1 ln(
R
N . q
E
Lq
ln . E . ) d 1 ln(
R
N . q
.
) d 1 ln(
) 1 .( E
C
2
= α − + − 
¹

\

+ + 
¹

\

α + 
¹

\

+ +


¹

\

+
α −
+
This can be considered to a quadratic equation, with the “x” term being
(q.N/R + ln(1+d)) and the other terms being as follows:
a = C + E.(1α)/(ln(1+d)
b = α.E.ln(L.q/E)
c = ln(1+d).(L.q  α.E)
We will proceed here in Part I of this proof to solve the quadratic
equation. In Part II, we will show that the “b” term has little effect, and
can be ignored. (It is interesting to consider how the “b” term arose.
Consider the NPV of the field. If the number of wells increases, then
abandonment is brought forward, so the term representing the present
value of the oil lost at abandonment is increased. This decreases the
overall NPV, but as can be imagined, such effects are small. This will be
proved later, in Part II).
NPV Formulas
16
So, ignoring the negative solution, the solution of the quadratic equation
is:
( )( )
( )
( )
( )
(
(
¸
(
¸
− ε − + ε
α − + +
α −
+ =
ε − + ε
α − + +
α −
+ =
ε − + ε
+
α −
+
α − +
= + +
α − α − + +

¹

\

α
=
−
= ε ε
(
(
¸
(
¸
−
−


¹

\

−
+
−
=
− −
=
1 1 .
E . E ) d 1 ln( . C
) E . q . L (
). d 1 ln( .
q
R
N
gives equation this arranging  Re
1 .
E . E ) d 1 ln( . C
) E . q . L (
) d 1 ln(
1 .
) d 1 ln(
) 1 .( E
C
) E . q . L ).( d 1 ln(
) d 1 ln(
R
N . q
gives a and c x, expanding and
E . q . L . E . E ) d 1 ln( . C . 4
E
q . L
ln . E .
c . a 4
b
by Defining
c . a 4
b
c . a 4
b
1 .
a
c
x
gives this arranging  Re
a 2
b c . a 4 b
x
2
2
2
2
2
Proof of Theorem 2, Part II
As a first step, it is useful to note that, for ε ≥ 0 (which is the case,
providing L.q ≥ E – i.e. Year 1 net revenue for a well is greater than the
opex for the well)
1 ≥ (√(ε2+1)  ε) ≥ 1ε hence, Napprox ≥ Nopt
( ) ( )
] 1 since
1 1 . 2 1 1 2 1 1  Proof [
2
2 2 2 2
2
2
ε ≥ + ε
≤ + ε − ε ε + = ε + + ε ε − + ε = ε − + ε
Examining ε
2
and dividing both the denominator and quotient by (L.q)
2
gives
NPV Formulas
17
( ) ( ) ( )
9 . 0 5 . 0 x 2 . 0 1
q . L
E
1
2) assumption (by 0.2 and 1) assumption (by 5 . 0
q . L
E ) d 1 ln( . C
L.q
E
Since
q . L
E
1
q . L
E ). 1 ( ) d 1 ln( . C
4
q . L E E q . L ln
2 2 2
2
= − ≥


¹

\
 α
−
≤ α ≤
+ +
≤


¹

\
 α
−
(
¸
(
¸
α − + +
α
= ε
Also
( ) ( ) ( )
( )
( ) ( )
q . L
E
. E q . L ln 014 . 0
9 . 0 x q . L E x 8 . 0 x 4
q . L E 2 . 0 E q . L ln
, Hence
q . L
E
8 . 0
q . L
E
). 1 (
q . L
E ). 1 ( ) d 1 ln( . C
2
2 2 2
2
= ≤ ε
≥ α − ≥
α − + +
It can be easily shown that for 1 ≥ E/(L.q) ≥ 0, the maximum value of
(ln(L.q/E))
2
.E/(L.q) is achieved when E/(L.q) = 1/(e
2
) = 0.1353, which
gives (ln(L.q/E))
2
.E/(L.q) = 0.541.
[Proof  Differentiate x.(ln(x))
2
and set to zero].
Hence ε ≤ 0.014 x 0.541 = 0.00757
ε ≤ 0.087
(√(1+ε
2
)  ε) ≥ (1 ε) ≥ (10.087) = 0.93
Before moving on to look at a lower limit for N
opt
/N
approx
, it is useful to
establish a couple of small lemmas.
Lemma A
For u, v, w such that u ≥ v > 0 and v > w ≥ 0,
( )
( )
0
w v . v
w . v u
) w v ( v
uw uv vw uv
v
u
w v
w u
oof Pr
v
u
w v
w u
≥
−
−
=
−
+ − −
= −
−
−
≥
−
−
Lemma B
For w constant and greater than zero, the function f (y) = [(1w).y –1] /
(y1) is strictly increasing (i.e. y
1
< y
2
⇒ f (y
1
) < f (y
2
) ).
Proof Expressing f (y) as 1 – wy / (y1) gives
NPV Formulas
18
( ) ( )
0
1 y
w
1 y
y . w
1 y
w
dy
df
>
−
=
−
+
−
−
=
Moving back to the main proof, let us establish a lower bound on
N
opt
/N
approx
( )
1
E . E ) d 1 ln( . C
E . Lq
1 .
E . E ) d 1 ln( . C
E . Lq
93 . 0
1
E . E ) d 1 ln( . C
E . Lq
1 1 .
E . E ) d 1 ln( . C
E . Lq
Napprox
Nopt
2
−
α − + +
α −
−
α − + +
α −
≥
−
α − + +
α −
− ε − ε +
α − + +
α −
=
By Lemma A and assumption (i)
76 . 0
1 2
1 2 93 . 0
1
E . E ) d 1 ln( . C
E . Lq
1 .
E . E ) d 1 ln( . C
E . Lq
93 . 0
B Lemma By
2
E ) d 1 ln( . C
Lq
E . E ) d 1 ln( . C
E . Lq
=
−
−
≥
−
α − + +
α −
−
α − + +
α −
≥
+ +
≥
α − + +
α −
Hence N
opt
/N
approx
≥ 0.76
or equivalently N
approx
≤ 1.31N
opt
Combining this result with the result established at the beginning of the
Part II of the proof gives
1.31 N
opt
≥ N
approx
≥ N
opt
Q.E.D.
Note – In order to express the expression for the approximate optimal
number of wells in the form first quoted
q
E . E d . C
q
E . E d . C
q
.E
 L
.
q
R.d
NPV maximises that wells of Number
α − +


¹

\

α − +
−
α
=
it suffices to rearrange the expression slightly and to note that for normal
values of d (the discount rate), ln(1+d) ≈ d (e.g. ln(1+0.08) = 0.077)
NPV Formulas
19
Theorem 3 – NPV for a development with the
approximately optimal number of wells
The NPV of a development with the approximately optimal number of
wells, as defined in Theorem 2, is
D
q
E ). 1 ( ) d 1 ln( . C
q
E .
L . R NPV
2
−


¹

\

α − + +
−
α
− =
where α = (1+d)Tab
and Tab = R / (N.q) . ln(L.q/E)
Proof
Combining the results from Theorems 1 and 2
( )
D
q
) d 1 ln( . C
q
) 1 .( E
q . L
E .
1 .
b a
L
. 1
b
a
. R
D
C
) d 1 ln(
) 1 .( E
q . L
E .
1 .
) d 1 ln( 1
E ). 1 ( ) d 1 ln( . C
E . q . L
.
q . R
) d 1 ln( . R . q
q . L
x 1
E ). 1 ( ) d 1 ln( . C
E . q . L
q
) d 1 ln( . R
NPV
−
(
(
¸
(
¸
+
−
α −
−


¹

\
 α
−


¹

\

− =
−
(
(
(
(
(
¸
(
¸
−
+
α −
−


¹

\
 α
−
+ +


¹

\

−
α − + +
α − +


¹

\

−
α − + +
α − +
=
(where a = L.q  α.E
and b = C.ln(1+d) +(1α).E)
D
q
E ). 1 ( ) d 1 ln( . C
q
E .
L . R
D
q
b
q
a
.
q
b
q
a
. R D
q
b
a . q
b . a
. 1
b
a
. R
2
−


¹

\

α − + +
−
α
− =
−
(
¸
(
¸
−


¹

\

− = −
(
(
¸
(
¸
−


¹

\

− =
Q.E.D.
NPV Formulas
Contents
Contents ......................................................................................... i Summary ....................................................................................... 1 Uses of the equations....................................................................... 2 An example.................................................................................. 2 Creating an analytic description of a typical economic model................. 3 Other economic assumptions.......................................................... 3 Reserves, well rates and decline rates ............................................. 4 Derivation of the formulas ................................................................ 5 Understanding the optimal number of wells ........................................ 6 Visualising the NPV formula .............................................................. 7 Using the formulas to derive the optimal injector:producer ratio............ 8 Uncertainties in key parameters ........................................................ 9 Extending the method to maximise rate of return ...............................10 Dealing with a development spread over several years........................11 Appendix – Mathematical proofs of the formulas.................................11 Assumptions and Definitions .........................................................11 Theorem 1 – Expressing NPV as a function of well numbers ..............12 Proof .......................................................................................12 Theorem 2 – Number of wells giving the highest NPV .......................13 Part I ......................................................................................13 Part II .....................................................................................14 Proof of Part I...........................................................................14 Proof of Theorem 2, Part II ........................................................16 Lemma A .................................................................................17 Lemma B .................................................................................17 Theorem 3 – NPV for a development with the approximately optimal number of wells...........................................................................19 Proof .......................................................................................19
i
E −D NPV with this number of wells = R. Because in most countries. including injectors R – technical reserves i.g. it is usual to calculate the economics in a spreadsheet model. “net” means expressed in terms of effect on present value. including profit tax) d – discount rate C – net capital cost per well D – net capital costs not related to numbers of wells. typical initial well rates. recovery factor etc). similar to those used in a typical spreadsheet model.E C. oilfield economics can be reduced to linear factors (for tax effects) and discounting.N E where α = (1+d)Tab and Tab (abandonment time)= and q – initial oil production per well per year.E L − q q R. L − q q R L.g. e. after all taxes and royalties 1 . including injectors N – total number of wells. roads and pipelines E – net opex per well Note:. C. commercial use is subject to patent protection) α. using a production profile generated by means of a set of simple assumptions (e. averaged over all wells.E q q 2 α. STOIIP.e. and derive the following equations (N.In this context.E C.q . it is possible to take a set of assumptions. discounting and exponential decline can be combined into a single exponential function.d + E − α. the amount of oil that could be recovered if the field were run for a very long time L – net revenue per tonne of oil (i.d + E − α.NPV Formulas Summary When quick estimates of NPV are needed for oilfield development options. after all taxes and royalties. ln q.d Number of wells that maximises NPV = .d + E − α.B.e.
500 barrels per day = 450.000 barrels per year The formulas quickly give a) Vertical wells Optimal number of wells = 35 NPV with this number of wells = $122.000 Average initial well rates = 1.000 barrels per year b) Horizontal wells Net capex per well = $5.000 Average initial well rates = 500 barrels per day = 150. which are often very difficult to fathom) can provide a totally independent check on project economics can help save time in determining the correct well density can help to provide an overview of the common task of almost all petroleum engineers – optimising the economic return.000 Net opex per well per year = $300.000 (Time to abandonment = 26 years) b) Horizontal wells Optimal number of wells = 19 NPV with this number of wells = $209. in the early stages of a project easily auditable (in contrast to spreadsheets.000.000. the equations quickly yielded very similar results to one month of conventional work. especially for screening prospects and initial concepts. when applied to a Russian field. In a trial. The equations can also be extended to give the optimal number of wells when the objective of the oil company is to maximise the overall rate of return. An example Consider a field with the following technical and economical parameters:Technically recoverable reserves = 100 million barrels Net oil price = $8 per barrel Discount rate = 8% Net fixed capex costs = $20.NPV Formulas Uses of the equations These equations offer a number of advantages when used in addition to existing methods:quick and easy to use.000.000 Net opex per well per year = $500.000.000 (Time to abandonment = 23 years) 2 .000 Two development options are considered:a) Vertical wells Net capex per well = $2.000.
NPV Formulas One can see that these results match the way that horizontal developments typically need fewer wells than vertical ones do. Creating an analytic description of a typical economic model In order to derive algebraic expressions for NPV and for the economically optimal number of wells. cost of platform. the year of the end of the fieldlife OPEXi = OPEX in year i CAPEXi = CAPEX in year i. representing the effects on NPV of one extra realterms dollar of expenditure on CAPEX The terms α. usually all the individual steps in the spreadsheets are linear. opex is proportional to the number of wells.g. the “net oil price” – it represents the benefit in present value terms of one extra unit (ton. It is suggested that the following is an appropriate model for most tax regimes:NPV = Σ (L. including taxes and royalties.OPEXi + β. revenues and taxation is very complicated. it is necessary to have an algebraic model of all the economic factors. capex expenditure for the field consists of a fixed element (e. road or pipeline) and a cost per well. such a possibility would quickly become well known. 3 . most tax systems have evolved to be linear. cubic metre or barrel) of oil sold α = a linear factor. representing the effects on NPV of one extra realterms dollar (or whatever unit of currency) of expenditure on OPEX β = a linear factor. all capex expenditure occurs in Year 1. from initial design or from progressive correction of anomalies. there would be benefits in artificially splitting or combining projects so as to take advantage of the nonlinearities. not including VAT Qi = oil production in year i L = a linear factor. ii) Other economic assumptions The other major assumptions about the economics are that it is desired to achieve the highest possible NPV (as opposed to the highest rate of return.CAPEXi can be considered to be the “net OPEX” and the “net CAPEX”. It is reasonable to model the NPV as a linear function for two reasons i) although.Qi + α. as can seen from economics spreadsheet models.OPEXi and β. the interplay of costs. Hence. for example).CAPEXi) / (1+d)i1 where the summation is over i=1 to i=n. hence combining them will give a linear function if a taxation system was not linear in its effects.
With ten times as many wells. a single well could. firstly because it is the simplest. and one is only interested in returns of more than 15%. in which opex contains one element proportional to the number of wells and one element proportional to the gross liquid production. the modification is described later in the article). this would still only give a ratio of 1 x 106. Clearly. it is unlikely that the time required to carry out the capital expenditure will affect questions such as “What is the optimal number of wells?” it is easy to modify the model to deal with capex spread over several years (again. Reserves. whenever the gross liquid rate is constant. it is possible to run through all the various scenarios and then choose the one which gives the overall best results. for a field producing under depletion. again.”) In this article. One reason for this is that. The first assumption about capex are that net capex can be expressed as Net capex = C. the simple model gives the same results as a more complicated model. it is usually necessary to define in advance what quantity is to be minimised or maximised. directly penetrated by a wellbore) is very small in relation to the total area. it is simpler. in spreadsheet models. because it is easy to modify the NPVmaximising results to cope with the situation where there is a shortage of capital.NPV Formulas In analytic models. the number of wells does not much affect the technical recovery factor. but. With vertical wells. For example “Putting in 15 wells rather than 10 increases the NPV by only 5%. so we will go for the 10 well option. well rates and decline rates There are three important petroleum engineering assumptions technically recoverable reserves are independent of the number of wells initial well rates are independent of the number of wells the field follows exponential decline from the start of production.  With regards to opex. Unless there are geological features like isolated fault blocks. over a very long time. The second assumption is that all the capex expenditure can be considered to take place instanteously at the beginning of the project. a well spacing of 500m and a wellbore diameter of 6 inches gives a ratio (total well bore area): (area of field) of 1 x 107. secondly. so similar recovery factors should apply. this is not always realistic. In theory. the area of the field that is drilled up (i. It can be seen from the formula for the steadystate productivity index (PI) of a vertical well 4 . The modification required is described later in the article. for example. drain the field as efficiently as twenty wells. On balance. given the overall uncertainties. it should be noted that. and. whatever the well spacing. while it increases the capex by 40%. we think that the increase in NPV is not worth it.N + D where C is net capex per well and D is a fixed element. (In contrast. NPV will be maximised.e.
The assumption that initial well rates are independent of well numbers also holds for gas fields being produced under depletion drive.NPV Formulas PI = (constant x Ko. However. then the effect can be captured by the modification to the model described later. once decline starts. If fraccing is used. increasing the number of wells does give pressure interference. many fields show a plateau production period. before water breakthrough). and on the question of what is the economically optimal number of wells. A change to hyperbolic decline usually occurs sufficiently late in field life that. even if the proportional change in ln(re/rw) is small. 3) Show that the formula for the optimal number of wells can be approximated by a simpler expression. including the number of injectors. with discounting. re (the effective drainage radius) only occurs as a logarithmic terms. There may be two objections to this asssumption. providing all the wells start production at the same time. it is does not necessarily follow that well rates are constant. but changes to hyperbolic decline in the later stages of field life. Firstly. then the average pressures should remain the same. Derivation of the formulas The derivation of the formulas was done in four steps 1) By simple integration. derive a formula for NPV as a function of (amongst other things) well numbers. if one scales a whole development. hence the effect of changes of it are usually small. 2) Calculate the optimal number of wells by taking partial derivatives with respect to number of wells and abandonment time. If this is the case. it may be incorrect to assume that PIs are independent of well numbers. this effect is captured by the decline rate. Here. Secondly. giving constant well rates. However. it may be exponential. fieldwide dry oil production. The other major petroleum engineering assumption is that the field follows exponential decline from the very beginning. because one well may reduce the average pressure seen by the other wells. So it is reasonable to assume that PI is not much affected by changes in well spacing. However. it can be argued that many of the fields showing a production plateaus did so because of staged development. Even if PIs are constant.µo] and the slightly more complicated formula for horizontal wells (Refs 1 and 2). then the model may require further development. if the plateau period is a genuine subsurface phenomenon (maybe a period of stable. In answer to the first objection.h) / [(ln(re/rw)+S) x Bo. The second objection should be less of a concern. it has little effect on the economics of the field. All that is required for the model to work correctly is to use a modification to the reserves – namely the total production at time=infinity if expontial decline had continued for the whole life of the field. 5 . Fraccing can be considered to give a large negative skin. that term depending on well separation. so it can be the case that. the change in [(ln(re/rw)+Skin] may be significant.
1 − N. the total value of the oil – the field is produced very slowly.q +1 R.d The first R. Consider the formula for expressing NPV as a function of the number of wells.L. this is linear with the number of wells. it can be seen that the value of the oil in the ground and the net fixed capex do not vary. so less value is lost because of time effects. pipelines etc). i. This is illustrated in the plot below.d effect of the time taken to get the oil out of the ground. then this term becomes almost as big as R. rather than spreadsheets. if the number of wells is very large. each with a different number of wells. The optimal number of wells is reached when the sum (value lost in time effects) + (value lost because of opex and variable capex expenditure) is at a minimum. if the number of wells is very small. so the NPV of the production is very small.L. On the other hand.(C+E/d) term represents the value lost because of capex per well and opex. the value of the field if the oil could be taken all at once at no cost. this term becomes very small – the field is produced in a short space of time. Understanding the optimal number of wells One of the big advantages of formulas.q d + 1 R.e. Comparing a set of different scenarios.L term (reserves x net oil price) corresponds to the value of the oil in the ground. is that they give more of an insight of how all the factors interact. based on experience of analogous fields. If one ignores abandonment (which is reasonable when one is trying to gain an understanding rather than calculate a specific value). a question that is usually answered in vague terms. Clearly. The N. the bulk of the oil is produced sooner. but opex and variable capex costs are increasing.800 bbl/day 6 . NPV can be expressed as 1 − N. Clearly.L term represents the value lost because of the discounting N. The R. This gives a simple explanation to the question of what is the optimal number of wells. As the number of wells increases. The D term represents the value lost because of net fixed capex (roads.NPV Formulas 4) Calculate the NPV for the approximation of the optimal number of wells. A full mathematical proof of these results is given in the appendix. but the other two terms do. for the case of a field with the following parameters:Technically recoverable reserves = 100 million barrels Average initial well rate = 2. C + E − D NPV = R.
If we ignore abandonment effects for the moment.e.NPV Formulas Net capex per well = $5.000 $0 0 5 10 15 20 25 30 35 40 Value lost to time effects Opex and variable capex Sum of value lost Number of wells It can be seen that the optimal number of wells is about 10. Imagine a square whose area represents the total value (to the oil company) of the field. then the scalable capex and the opex reduce the value to C.000 $250.00 Discount rate = 8% Effects of changing the number of wells $500.000.000.000 $100.d + E)/q]. L − q 2 This is the area of the square that can be fitted in the R. D.000 $200.000. Then the value has to be further reduced to pay for the fixed capex.000.000 $150.000.000. it may be helpful to consider the formula as follows.000.000 Net oil price = $4.L square if there has been already inserted a square of length √[(C.000. 7 .000.000 Net opex per well = $400.000.000 Present Value $300.d + E R.000 $450.000 $350.000 $400.000. if all the oil could be produced immediately at zero cost i. represented by a suitable rectangle. R x L. The final step in building up the picture is to allow abandonment effects to introduce a gap between the two smaller squares. Let the length of the sides of the square correspond to √L.000 $50. Visualising the NPV formula To gain a feel for the significence of the NPV formula (for the optimal number of wells).
d + Ei) 8 .r.Bw.(Cp.Bo. to give a formula for the optimal injector producer ratio as follows (proof not given in this article) Optimal injector:producer ratio = where PI = productivity index of average production well PI. the injector:producer ratio.d + Ep) II. Using the formulas to derive the optimal injector:producer ratio Another advantage of formulas is that they can be in further calculations. both producers and injectors) as a function of the injector:producer ratio.NPV Formulas Value of the hydrocarbons Lost because of time effects NPV Scalable CAPEX and OPEX Unscalable CAPEX Such a geometric representation is scarely any more intuitive than the formula itself. to express the expected average well rate (averaged over all wells. This expression can be diffentiated w.(Ci. but it is put forward in the hope that it may a starting point for better visualisations or interpretations of the formula. For example. in a waterflood. independently of the number of wells. given constraints on well bottomhole pressures.t. Feeding this expression into the NPV formula gives an expression of NPV as a function of injector:producer ratio. through a material balance argument. it is straightforward.
000 $43.NPV Formulas II = injectivity index of average injection well Bo = oil formation volume factor Bw = water formation volume factor Cp = net capex cost per producer Ci = net capex cost per injector Ep = net opex per year for each producer Ei = net opex per year for each injector d = discount rate Uncertainties in key parameters By reducing the complexity of the economics calculations.05 0.000.08 0. 9 .000.000. Such analysis can help identify the key uncertainties to be addressed during field appraisal.06 0. when there is a big uncertainty in the Kv:Kh ratio.000 $44.000 $39.000 0 0. Development Options for Field X . horizontal wells would be preferably to vertical wells providing the Kv/Kh ratio is greater than 0.1 Slanted injs. For example.000.000.000. for such a field.04 0.000 $38.02.09 0.000. to compare the overall economic performance of horizontal and vertical wells.000 $40.000 $36.02 0.000 $41.01 0.03 0. it is easy to generate plots such as the one below.000.000.000 $42.NPV vs Kv_Kh $45.000.07 0.000 $37. the formulas make it possible to examine a wider range of uncertainties in key parameters. horiz prods Vertical wells Horizontal wells Slanted wells NPV (US$) Kv_Kh It can be seen that.000.000 $35.
oil companies are faced with the problem of constraints. subject to the limits on the amount of capital employed. it is useful to consider the project as consisting of (Minimalist development option) + Series of increments to the mininalist development option.Capex) / δN = 0 since δN / δCapex ≠ 0 (it never costs an infinite amount of money to drill a new well). all the increments start at time = 0).Capex) / δCapex = 0 ⇔ [δ(NPVH. the companies usually have only a fixed amount of capital available. or at least limits on the amount of capital employed (gearing limits etc). In particular. In the real world.To determine.e.(C. They pass the screening criteria if their NPV:Capex ratio is greater than H i.NPV Formulas Extending the method to maximise rate of return A simple modification to the Capex figures makes it possible to use the formulas when the company objective is to maximise overall return on capital employed. for a new project. the new problem (find the number of wells that maximises corporate NPV subject to a limit on capital employed) can be converted into the old (find the number of wells that maximises NPV) by using an artifical NPV. the optimal number of wells. defined to be NPV′ = NPV – H. the NPV:Capex ratio for the last project chosen. Hence. the method to do this is to rank all the projects in order of decreasing NPV:Capex ratio.Capex = NPV – H. the wells are worth adding to the development scheme. while the NPV/Capex ratio for additional wells is greater than H. and then choose the projects in this order until one has reached the limit on the amount of capital employed. The optimal total number of wells is reached when the limit is reached.Capex) / δN] x [δN / δCapex] = 0 of wells where N is the number ⇔ δ(NPVH. (Note that the split is purely conceptual.N + D) 10 . The problem then is to maximise NPV.e. All the increments can be analysed as if they were separate projects. Assuming that there are sufficient number of projects available and no other constraints intrude (such as limits on management capacity). let H (the “hurdle” rate) be the lower limit to the NPV:Capex ratio i. δNPV / δCapex ≥ H So. For the projects chosen. so the criteria for determining the total number of wells is δNPV / δCapex = H This can be converted into a more workable criteria as follows δNPV / δCapex = H ⇔ δ(NPVH. The formulas as they stand show the number of wells required to maximise NPV.
N + D) = R. then the situation can be modelled as follows. 11 . The new problem is to find N to maximise NPV′ = − ( M −1) NPV( N) (1 + d )−(i −1) = (1 + d) − (1 + d) .NPV Formulas and artificial C′ and D′ defined to be C′ = (1 + H) x C D′ = (1 + H) x D Then α. with different areas coming on stream in different years.d + E −α.E L − q q R. independent of the number of wells.NPV( N) ∑ M M. so that the development is spread over several years.N − D NPVwith this number wells NP ′ + H.d + E −α. Let M be the number of years that will be required. = . Appendix – Mathematical proofs of the formulas Assumptions and Definitions Let us make the following assumptions to describe an oilfield development 1) Let N be the number of wells drilled. independent of the number of wells i. 2) If the field were run for an infinite time. Lof = V − q q (the –H.E q q 2 α.d i =1 M Clearly.E C′.D and +H.E C′. it is not affected by well spacing. Dealing with a development spread over several years If a field is very large.E + H. of N C′.(C. All the wells start up at time t=0. So the problem is solved.C. the value of N that maximises NPV(N) also maximises NPV′ and the above equation gives NPV′. (Assume M to be independent of N.e.d Optimal number wells.d + E −α.D terms cancel each other out). Let NPV(N) be the NPV that would have been achieved if the development had been had carried out in the space of one year. 3) The initial production rate per well is q. the total production would be R (the technically recoverable reserves). the number of wells).
q.L. Start by calculating the NPV of the revenue stream:By the assumption that there is exponential decline Oil production rate = N.t N. 1 − α. L.e − q.t/R By the definition of NPV 12 .eat = N.q. ln(1 + d ) 1+ q.E . 7) Net opex can be expressed as E. after all taxes and deductions.eat 5) The net oil price is a constant.N ( ) Proof NPV can be broken down into the component parts of the cashflow NPV = NPV(revenue stream) + NPV(Opex) + NPV(Capex) where the NPV(Opex) and NPV(Capex) are.NPV Formulas 4) The field oil production rate follows exponential decline i. all capital expenditure happens at time t=0. Field oil production rate = initial rate x eat = N.q / a a 0 ∞ then a = N.e.e 0 ∞ −a . negative. providing it is the same for both E and q). (The year is probably the most appropriate unit of time. Theorem 1 – Expressing NPV as a function of well numbers The NPV of a field run until abandonment can be expressed as NPV = R.L N. by the definition of technical reserves R = ∫ N. 6) The net capital costs can be expressed as D + C.eq.N.N.Tab / R − .e = N. of course.q. N.q.N/R)t [Since.q −at dt = − .N per unit time.q. but any unit can be used.e(q.(1 − α) − (C.q/R ] Oil revenue per unit time = (production rate) x (net oil price) = N.N + D) ln(1 + d ) R.
Tab .L. N.E . 1 − α.(C.q.N q.t dt = 0 ln(1 + d ) (1 + d ) t Tab =− 0 N. 1 − (1 + d) − Tab .N + D) By adding NPV(Revenue).e −ln(1+ d ). Nopt.q.Tab / R R.L = . one obtains the desired formula.N R.L .N = ( ( ) ( ) Looking at the Opex cashflow Opex per unit time = N. NPV of Capex = .L. Q.t dt t 0 (1 + d) ) = e − t . ln(1+ d ) ) = N.E N. 1 − e − ( q.E NPV of opex = ∫ =− Tab 0 Tab − N. 1 − e −ln(1+ d ).e − ln(1+ d ).t dt = − ∫ N.Tab Tab ( ) ) 0 R. N / R + ln(1+ d )) t q. ln(1 + d) 1 + q. 1 − e − ln(1+ d ). ln(1 + d) 1 + 1 + q.e − q.e − q. ln(1 + d) R.(1 − α ) ln(1 + d ) ln(1 + d ) ( ) Looking at Capex.NPV Formulas NPV of revenue stream = ∫ (Since (1 + d) t = eln((1+ d ) = N. N.L. N.e − ( q. since all the capital expenditure is assumed to occur at time t=0.e − ( q.Tab / R = .q q. can be calculated (iteratively) from the expressions 13 .N / R + ln(1+ d )).L R.D Theorem 2 – Number of wells giving the highest NPV Part I The number of wells giving the highest NPV.e − q.N + ln(1 + d) R . 1 − (1 − d ) −Tab = − .E N.Tab / R R.E.E.E . N / R + ln(1+ d )).N −( + ln(1 + d)) R −t Tab 0 Tab Revenue per unit time dt = ∫ N.E.Tab ln(1 + d ) ( ) N. NPV(Opex) and NPV(Capex).
for Tab optimal.q. eq.E R .E.B Also. and writing out the expression in full (not using α.q)) For well numbers.eq. by seeing that this occurs when revenue has dropped until it equals operating expenditure i. Nopt = and α = (1 + d) −Tab and Tab = R L. ln q.q/E) (N.2 ≥ α (equates.N) . using Theorem 1.31 Nopt ≥ Napprox ≥ Nopt Proof of Part I Consider NPV as a function of well numbers and abandonment time. ln E where ε = 4(L. the well is reasonably profitable at first) and b) 0.Tab/R = ln(E/(L. for example) 14 . For a normally behaved function NPV is a maximum ⇔ Tab such that ∂NPV ∂NPV = =0 ∂N ∂Tab ∂NPV =0 ∂Tab can be obtained by taking the partial derivatives or.Tab/R = N. more simply.N. ln(L. Napprox within the following bounds 1.E )(C.E so –q.NPV Formulas L. x ε 2 + 1 − ε − 1 q C.Tab/R = E/(L. for d = 8%.N.E ) .E L.e. N.ln(1+d) + E] (i.q − α.N E ( ) Part II If a) L.L. to a field life ≥ 20 years) then the approximation of setting ε = 0 gives approximate value for the number of wells.q α.q)) so Tab = R/(q.q − α.e.N. ln(1 + d) + E − α. ln(1 + d).[C.q ≥ 2.q . ln(1 + d) + E − α.
E. then using α as shorthand for (1+d)Tab.1 − α.Tab R Setting this partial derivative equal to zero.q − α. then abandonment is brought forward.q.q.q) etc.N. ln E R + ln(1 + d) − ln(1 + d).E) We will proceed here in Part I of this proof to solve the quadratic equation. ( ( ).E.q.α. with the “x” term being (q. eq.NPV Formulas Lq E(1 − (1 + d ) − Tab ) ∂NPV ∂ = − C − D N.e − (qNTab / R ) − ∂N ∂N qN ln(1 + d ) + ln(1 + d ) R ( ) = Lq qN + ln(1 + d ) R N.ln(L.q 2 E.N C + .e −(qNTab / R ) − ( ) E(1 − (1 + d ) −Tab ) −C ln(1 + d ) q )R − qN + ln(1 + d ) R N. 1 − (1 + d ) −Tab .N + ln + ln(1 + d) .q .N q.1 − α L. ln(1 + d) + C − R .q Cancelling terms and multiplying both sides by –1 gives E. so the term representing the present value of the oil lost at abandonment is increased. This equation only applies when Tab is optimal. 1 − (1 + d ) − Tab . and can be ignored. 1 − (1 + d ) −Tab .N Lq q.(1 − α) q.N/R + ln(1+d)) and the other terms being as follows:a = C + E.B. in Part II). E R L. such effects are small. If the number of wells increases.(1 − α ) qN E .Tab/R = E/(L.L. (It is interesting to consider how the “b” term arose. Consider the NPV of the field.α. + ln(1 + d) L. and the relationships that apply when Tab is optimal. E q. 1 − (1 + d ) −Tab .Lq + .e −(qNTab / R ) .q q.N − + ln(1 + d) . multiplying both sides of the resultant equation by (qN/R + ln(1+d))2. This will be proved later.Lq . This decreases the overall NPV. . but as can be imagined. N. In Part II.(L.E) = 0 ln(1 + d) This can be considered to a quadratic equation. =0 .L.q/E) c = ln(1+d). we will show that the “b” term has little effect. 2 15 .(1α)/(ln(1+d) b = α.(L.q R R E L.L.e − (qNTab / R ) qN + ln(1 + d ) R 2 . − q. R + ln(1 + d) + α. we get the following equation.
c − 4a.(1 − α) R C+ ln(1 + d) = ln(1 + d) (L.NPV Formulas So.E) R .E ( Proof of Theorem 2. . Year 1 net revenue for a well is greater than the opex for the well) 1 ≥ (√(ε2+1) . ln(1 + d) + E − α. the solution of the quadratic equation is: b 2 − 4a.c Defining ε by ε = b − 4a. ln(1 + d) + E − α.E ( ) ( ) ) Re .e. ignoring the negative solution. ε 2 +1 − ε + ln(1 + d ) = E.q − α.q − α. ε 2 + 1 − ε − 1 q C. ln(1 + d) + E − α.arranging this equation gives N= (L.arranging this gives x= x= 2 b −c b − . Napprox ≥ Nopt [ Proof  (ε 2 + 1 − ε = ε 2 + 1 − 2ε ε 2 + 1 + ε 2 = 1 + 2ε.E) .(L.E )(L. it is useful to note that.q α. ε − ε 2 + 1 ≤ 1 ) 2 ( ) since ε 2 + 1 ≥ ε ] Examining ε2 and dividing both the denominator and quotient by (L. for ε ≥ 0 (which is the case.q ≥ E – i.E ) .N ln(1 + d ). ln E 4.c = L. providing L. 1+ a − 4a. c and a gives q. ε 2 +1 − ε C.E.(C.E) .q − α. Part II As a first step. and expanding x.c − b 2a Re .q − α. ln(1 + d).q)2 gives 16 .ε) ≥ 1ε hence.
ε) ≥ (10.(v − w ) Lemma B For w constant and greater than zero.8 L.014 x 0.5 = 0.q )2 C.q/E))2.8x (E L.93 Before moving on to look at a lower limit for Nopt/Napprox.014(ln(L.q) ≥ 0.E/(L.y –1] / (y1) is strictly increasing (i.q E )) .q) = 0. which gives (ln(L.q L. ln(1 + d) + (1 − α).E/(L. 2 E L.e. ≥ 0. y1 < y2 ⇒ f (y1) < f (y2) ).087 (√(1+ε2) .q) is achieved when E/(L.541. Proof Expressing f (y) as 1 – wy / (y1) gives 17 .5 (by assumption 1) and α ≤ 0. u−w u ≥ v−w v u − w u uv − vw − uv + uw (u − v ).q )x 0.q C.q ≥ 1 − 0.q E ))2 α 2 (E L.q L. [Proof . the maximum value of (ln(L.q αE 1 − L.1353.2x 0.q E ))2 0.q) = 1/(e2) = 0.q L.ε) ≥ (1. ε 2 (ln(L.00757 ε ≤ 0.E E E ≥ (1 − α). the function f (y) = [(1w).q Hence. v.q L.2 (by assumption 2) ≤ L.q )2 ≤ 4x 0.9 Also ε = 2 (ln(L.2 2 (E L. it is useful to establish a couple of small lemmas.q/E))2. Lemma A For u.541 = 0.E αE 4 1 − L.9 = 0. ln(1 + d) + E E Since ≤ 0. Hence ε ≤ 0.q It can be easily shown that for 1 ≥ E/(L. ln(1 + d) + (1 − α).Differentiate x.w Pr oof − = = ≥0 v−w v v( v − w ) v. w such that u ≥ v > 0 and v > w ≥ 0.087) = 0.NPV Formulas C.(ln(x))2 and set to zero].
E Hence Nopt/Napprox ≥ 0.E.E − C.E Lq − α. ln(1 + d) + E By Lemma B 0. ln(1+0.93 Lq − α.E C. let us establish a lower bound on Nopt/Napprox Nopt = Napprox Lq − α. ln(1 + d) + E − α.E ( ) 0. C. ln(1 + d) + E − α.E By Lemma A and assumption (i) Lq − α.E −1 C.NPV Formulas w. 1 + ε 2 − ε −1 C.08) = 0.E q q R. ln(1 + d) + E − α.D.76 Napprox ≤ 1.α.g. ln(1 + d ) + E − α.E 0.E Lq ≥ ≥ 2 C.d Number of wells that maximises NPV = . ln(1+d) ≈ d (e.d + E − α.E .077) 18 .31 Nopt ≥ Napprox ≥ Nopt Q.d + E − α. −1 C. Note – In order to express the expression for the approximate optimal number of wells in the form first quoted L .76 or equivalently ≥ = 0. −1 C.E −1 C.31Nopt Combining this result with the result established at the beginning of the Part II of the proof gives 1.y df − w w = + = >0 dy y − 1 (y − 1) (y − 1) Moving back to the main proof.E .E q q it suffices to rearrange the expression slightly and to note that for normal values of d (the discount rate). ln(1 + d ) + E − α.E −1 C.93 ≥ Lq − α. ln(1 + d) + E − α. ln(1 + d ) + E − α.E Lq − α.93 2 − 1 2 −1 Lq − α.E .
R. a − q − D = R.q .E α. − 1. − 1 + ln(1 + d) C.ln(1+d) +(1α).q ) a = L. is C.E b = C.E E. ln(1 + d ) − . ln(L.1 − − −D q q b L. ln(1 + d ) + (1 − α).1 − L.(1 − α) L.D.E −D = R.q −D a = R.E α.q/E) 2 Proof Combining the results from Theorems 1 and 2 NPV = R. as defined in Theorem 2.E. ln(1 + d) + (1 − α). b (where and ( L a α.q − ln(1 + d) − C L. b − 1. ln(1 + d) .q) .α.E R. L − − q q where α = (1+d)Tab and Tab = R / (N.E q.E −D NPV = R.E) a a. q − q .NPV Formulas Theorem 3 – NPV for a development with the approximately optimal number of wells The NPV of a development with the approximately optimal number of wells. q − q − D C. q. L − − q q Q. ln(1 + d) + (1 − α). b b a b a b = R. ln(1 + d) + (1 − α).q . 2 19 .E q α.(1 − α) C.q − α.E − 1 x C.q − α.E E. ln(1 + d ) L.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.