SPE 108607
Estimation of Fracture Spacing in Naturally Fractured Carbonate Reservoirs
C. Prez-Rosales, SPE, E. Serrano, A. Moctezuma, and E. Luna, Instituto Mexicano del Petrleo
Copyright 2007, Society of Petroleum Engineers
This paper was prepared for presentation at the 2007 International Oil Conference and
Exhibition in Mexico held in Veracruz, Mexico, 2730 June 2007.
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
presented, have not been reviewed by the Society of Petroleum Engineers and are subject to
correction by the author(s). The material, as presented, does not necessarily reflect any
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper
for commercial purposes without the written consent of the Society of Petroleum Engineers is
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, Texas 75083-3836, U.S.A., fax 01-972-952-9435.
Abstract
Fracture spacing is an important concept for characterizing
flow properties of naturally fractured reservoirs, since the
main function of fractures that separate matrix blocks is
transporting fluids through long distances; however, the
estimation of fracture spacing presents some difficulties
mainly due to the fact that fractures occur at different scales,
going from microfractures in thin sections and minifractures in
cores, up to macrofractures in geological outcrops. The scale
of interest in this work is that used in reservoir simulation,
which is of the order of feet or meters.
This article is based on the ideas developed in a previous
paper, where a procedure to locate fractures is presented. That
procedure, which makes use of resistivity data obtained
through well logging, visualizes the fractures as highly
conducting channels within a low conductivity medium (the
rock matrix). By using a special way of data processing, it is
possible to filter out data that are not associated with fractures,
keeping only those data related to fractures. In this way,
fracture spacing can easily be estimated. However, that
procedure exhibits some uncertainties which must be
overcome to make it a more reliable one.
In this work, a study is made to search for an improved
procedure to estimate fracture spacing. For this purpose,
fractures are considered at two scales: local scale which
includes micro- and minifractures present in matrix blocks,
and at reservoir scale which refers to fractures separating
matrix blocks. These latter fractures, called principal fractures,
constitute the main fracture network, and are the subject
matter of this work.
Conductivity studies reveal that local scale fractures have
a frequency distribution quite different from that of principal
fractures. As it will be seen below, this fact facilitates
establishing a procedure for estimating fracture spacing
without uncertainties.
To make the ideas clear, an application to a carbonate
reservoir is presented. The results obtained show that the
improved procedure is a simple, reliable, and practical tool for
establishing the distribution of fractures along a well, from
which fracture spacing can be inferred.
Introduction
Non sealed fractures in naturally fractured reservoirs are high
conductivity channels; hence, fracture spacing is a factor that
controls, to a great extent, the flow properties of such systems.
In spite of its importance in areas such as hydrology, geology,
geophysics, and petroleum engineering, the problem of
estimating fracture spacing has not received the proper
attention from researchers, and the specialized literature
presents relatively few works treating in depth this theme.
Among the currently used techniques for detecting fractures
are well testing, core analysis, direct outcrop observation, and
well logging.1-4 In this work, an improved way to determine
fracture spacing is approached.
In a previous paper,5 a procedure for estimating fracture
spacing was developed. That procedure is based on data
analysis of formation resistivity factor obtained through well
logging. The fundamental consideration of the procedure is
that fractures are high conductivity anomalies in a low
conductivity medium (the matrix) and, consequently, the basic
tool for studying fracture spacing is based on the detection of
contrasts in electrical conductivity. To this end, a special
analyzing process is used to distinguish between data
associated with fractures and non-associated. However, such a
procedure does not allow establishing with certainty a
discriminating threshold between both types of data.
The fractures referred to in this work are those
surrounding matrix blocks. These fractures, called principal
fractures, constitute the main fracture network, which has the
property of transporting reservoir fluids through long
distances, and eventually to the producing wells, in opposition
to micro- and minifractures which act at block scale, and
whose main function is to convey fluids within the matrix
blocks and towards the principal fractures.
Previous work
Essentially, the above referred procedure5 can be summarized
as follows: Given formation resistivity factor data obtained
through well logging, their reciprocal values, called
conductivity factors, are calculated, whose graphical
representation as a function of depth is the original curve. An
example of such a curve appears as the heavy line plot of Fig.
1, corresponding to the 4460-4500 meters interval of the well
which will be identified as RMH-2007. Afterwards, a new
plot, called the trend curve (light line plot of Fig. 1) is
constructed with the moving average of the original data. In
2 SPE 108607
this new curve, the value at each point is given by the average
of the two previous values, the original value, and the two
subsequent values; that is, if FCi is the conductivity factor at
point i, then the corresponding trend curve value, at the same
point, is given by
i+2
1
TCi =
5
F
i =i 2
Ci (1)
In the examples presented in this paper, the data points
are separated by 0.25 m, so that each point of the trend curve
is representative of the values within one meter interval, which
is a reasonable interval when working at reservoir scale. The
next step consists in calculating, for each position i, the
difference DCi between the values of the original curve and
those of the trend curve, as indicted by Eq. 2.
DCi = FCi TCi (2)
This difference, called deviation, can be positive or negative.
Since high conductivity values are looked for, the negative
deviations are discarded by assigning them zero values, and
keeping only the positive values, as shown in Fig. 2. The
graphical representation of the positive values will be referred
to as peaks. Here it is seen that there is a complex distribution
of values going from very small up to relatively large ones. It
is considered that the small peaks are due to local
heterogeneities generated at matrix block scale, including
micro- and minifractures, whereas the large values are
associated with principal fractures separating matrix blocks.
Therefore, from the standpoint of the study of fracture
spacing, the small values represent a kind of background noise
which must be filtered out. By direct observations of plots of
the kind of Fig. 2, it was concluded that values less than 25%
of the highest value represent background noise; so, only the
higher values were considered as being representative of
principal fractures. Fig. 3 is the result of applying the 25% rule
to data appearing in Fig. 2. Although this scheme leads to
more or less reasonable results, it does not have the certainty
required by reservoir engineering, because the discriminating
threshold depends on a single value: the maximum deviation,
which might have an anomalous value.
Improved procedure
Due to the complexity of peak distribution observed in Fig. 2,
the fundamental problem in the development of an improved
procedure to estimate fracture spacing is finding a reliable
quantitative way to ascertain the discriminating peak size
which allows separating both kinds of peaks, on the
understanding that smaller peaks mean background noise
(local effect fractures) and larger ones represent principal
fractures.
To deal with this problem, it should be taken into account
that micro- and minifractures have conductivities much less
than the conductivity of principal fractures. Furthermore, the
former are much more numerous than the latter. Consequently,
in plots of the type of Fig. 2 the fractures at block scale
generate small and abundant peaks, contrary to principal
fractures which give rise to large and limited peaks. These
features can be used to distinguish between local effect
fractures and principal fractures. Consider, for instance, the
plot of Fig. 4, where the peak occurrence frequency of the
same data of Fig. 2 has been plotted as a function of peak size.
Here each point represents the number of peaks within a size
interval of 0.02 au (dimensionless arbitrary units). For the sake
of clarity, the number of peaks per interval was assigned to the
middle point of the interval. At first sight, it is observed that
small peaks have quite a different distribution than large
peaks: the former are numerous and their frequency declines
rapidly as their size increases (negative slope), whereas the
latter are limited and have a uniform distribution (zero slope).
This same distribution is observed, although in a more
pronounced way, in Fig. 5, which corresponds to another
depth interval (4550-4600 meters) of the same well. The point
where negative slope changes to zero slope is interpreted as
corresponding to the passing from block scale fractures to
principal fractures.
By making similar analyses with other wells data, it has
been found that the behavior observed in Figs. 4 and 5 is a
general one; that is, there always exists a pronounced
frequency drop for small peaks, whereas the frequency for
large peaks remains low and uniform. This means that this
kind of plots can be used to establish the discriminating
threshold between both types of peaks. Once the threshold is
found, the smaller peaks can be eliminated and work only with
the peaks representing principal fractures.
Through the study of different cases, it was found that the
discriminating threshold is located at a peak size that is
approximately equal to twice the average of all the peak
values. This value, which will be represented by 2APV, is
consistent in all cases. Thus, for instance, the values obtained
for 2APV in the two cases presented in Figs. 4 and 5 were
0.064 and 0.034 (dashed lines), respectively, which
correspond, approximately, to the peak size where the slope
changes from negative value to zero value. An advantage of
the threshold determined in this way is the simplicity of its
evaluation.
Application to a real case
In a practical application, the starting point is a set of positive
deviations (peaks) as a function of depth, like those used to
generate Fig. 2, and the first step is to calculate the average of
the peak values, APV. Twice the calculated value is used as
separating boundary between background noise and principal
fractures, so that peak values less than 2APV are filtered out,
and the remaining values are plotted against depth. This was
the procedure used to construct the graphs appearing in Figs. 6
to 13. The plots correspond to eight 50-meter intervals of well
RMH-2007. For calculating the APV, each depth interval was
treated independently of the others; also, to facilitate
comparisons between different graphs, the same coordinate
scales were used in all cases.
By analyzing Figs. 6 to 13, it can be seen that fractures do
not have a uniform distribution, but there are varying distances
(i.e. fracture spacing) between consecutive peaks, sometimes
forming well defined clusters.
It is considered that the size of the peaks is related in
some way to fracture aperture: the larger the peak size, the
larger the contrast between fracture and surrounding medium
and, therefore, the larger the fracture aperture. Based on this
SPE 108607
consideration, it can be said that the three plots within the
4500-4650 meters interval (Figs. 6, 7 and 8) are related to
highly fractured zones with good fracture aperture, and
starting at the 4650-4700 meters interval, the peak size begins
to decline, so that at the deepest interval (4850-4900 meters)
the fracture aperture becomes almost imperceptible. By
comparing the eight graphs of Figs. 6 to 13, it can be said that,
in the case here considered, the deeper the interval, the smaller
the fracture aperture.
Conclusions
An improved procedure for estimating fracture spacing along
a well in naturally fractured reservoir has been described. The
procedure makes use of electrical conductivity data obtained
through well logging, and is based on the detection of highly
contrasting conductivities. High conductivity values indicate
the presence of fractures within a low conductivity medium
represented by the matrix.
The central part of this work describes how the original
data are processed to highlight conductivity values related to
main fractures and to filter out data associated with local
effect (matrix block scale) heterogeneities which, for the
purposes of this work, represent background noise.
To illustrate the procedure, an application to a real case
was presented. From the results obtained, it can be concluded
that the procedure is simple and practical, requiring for its
application only the knowledge of conductivity factor (i.e. the
reciprocal of formation resistivity factor) as a function of
depth.
3
Nomenclature
APV = average of peak values
DCi = deviation at point i
FCi = conductivity factor at point i
TCi = trend curve value at point i
References
1. Aguilera, R.: Naturally Fractured Reservoirs, PennWell
Publishing Company, Tulsa, 1995.
2. Saidi, A.M.: Reservoir Engineering of Fractured
Reservoirs, TOTAL Edition Presse, Paris, 1987.
3. Jorden, J.R. and Campbell, F.L.: Well Logging II
Electric and Acoustic Logging, SPE Monograph Series,
Richardson, Texas, 1986.
4. Dietrich, P., Helmig, R., Sauter, M., Htzl, H., Kngeter,
J., and Teutsch, G. (Editors): Flow and Transport in
Fractured Porous Media, Springer-Verlag Heidelberg,
2005.
5. Prez-Rosales, C., Luna, E., and Serrano, E.: Simplified
Method for Characterizing Fractured Carbonate Media,
SPE 102212 presented at the First International Oil
Conference and Exhibition in Mexico held in Cancun,
Mexico, 31 August-2 September 2006.
_______________________________________
SI Metric Conversion Factor
ft x 3.048* E-01 = m
*Conversion factor is exact.
__________________________________________________________
4 SPE 108607

40

0 .4 5
35
0 .4 0
30
0 .3 5

Peak frequency
Conductivity factor

25
0 .3 0

20
0 .2 5

0 .2 0 15

0 .1 5 10

0 .1 0 5

0 .0 5 0
0 .0 0 0 .0 5 0 .1 0 0 .1 5 0 .2 0 0 .2 5
0 .0 0
4460 4470 4480 4490 4500 P e a k s iz e , a u
D e p th , m

Fig. 4 Peak frequency as a function of peak size for peak

Fig. 1 Conductivity factor (heavy line) and moving average (light increments of 0.02 au. Well RMH-2007, 4460-4500 meters interval.
line) as functions of depth. Dashed line at peak size of 0.064 au indicates position of
discriminating threshold.

0 .2 5 70
Positive deviation, au

60
0 .2 0

50
Peak frequency

0 .1 5
40

0 .1 0
30

0 .0 5 20

10
0 .0 0
4460 4470 4480 4490 4500
0
D e p th , m
0 .0 0 0 .0 5 0 .1 0 0 .1 5 0 .2 0 0 .2 5 0 .3 0 0 .3 5
P e a k s iz e , a u

Fig. 2 Positive values of conductivity deviation as a function of

depth Fig. 5 Peak frequency as a function of peak size for peak
increments of 0.02 au. Well RMH-2007, 4550-4600 meters interval.
Dashed line at peak size of 0.034 au indicates position of
0 .2 5 discriminating threshold.

0 .2 0
Peak size, au

0 .1 5

0 .1 0

0 .0 5

0 .0 0
4460 4465 4470 4475 4480 4485 4490 4495 4500
D e p th , m

Fig. 3 Position of principal fractures for well RMH-2007, 4460-

4500 meters interval.
SPE 108607 5

0.24
0.24

0.20
0.20

0.16

Peak size, au
0.16
Peak size, au

0.12
0.12

0.08
0.08

0.04
0.04

0.00
0.00 4700 4710 4720 4730 4740 4750
4500 4510 4520 4530 4540 4550
Depth, m
D ep th, m
Fig. 10 - Well interval: 4700-4750 meters.
Fig. 6 - Well interval: 4500-4550 meters.
0.24
0.24

0.20
0.20

0.16

Peak size, au
0.16
Peak size, au

0.12
0.12

0.08
0.08

0.04
0.04

0.00
0.00 4750 4760 4770 4780 4790 4800
4550 4560 4570 4580 4590 4600
Depth, m
D epth,m
Fig. 11 - Well interval: 4750-4800 meters.
Fig. 7 - Well interval: 4550-4600 meters.
0.24
0.24
0.20
0.20
0.16
Peak size, au

0.16
Peak size, au

0.12
0.12
0.08
0.08
0.04
0.04
0.00
0.00 4800 4810 4820 4830 4840 4850
4600 4610 4620 4630 4640 4650 Depth, m
Depth, m Fig. 12 - Well interval: 4800-4850 meters.
Fig. 8 - Well interval: 4600-4650 meters.
0.24
0.24
0.20
0.20
0.16
Peak size, au

0.16
Peak size, au

0.12
0.12

0.08
0.08

0.04
0.04

0.00
0.00
4850 4860 4870 4880 4890 4900
4650 4660 4670 4680 4690 4700
Depth, m
D epth, m
Fig. 9 - Well interval: 4650-4700 meters. Fig. 13 - Well interval: 4850-4900 meters.