Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

A Perspective Look at Fracture Porosity

W.M. Hen.e. Jr., * SPE, Sun E&P co.

Summary. Because overly optimistic estimates of hydrocarbons associated with reservoir fractures may lead to economic disasters,
it is important that fracture porosity values, assigned by whatever means, be examined critically. Some basic fracture models together
with documented examples of fracture width and frequency tend to suggest an upper limit for fracture porosity.

Introduction Fracture Porosity Calculation Methods


Fractures can be an important feature of hydrocarbon reservoirs Aguilera6 documents a means of estimating fracture porosity from
because their presence (either naturally occurring or induced) in- well logs in a dual-porosity system where a partitioning coefficient,
creases the productive capacity. Many low-matrix-permeability V, is calculated. This V factor is applied to the matrix porosity and
reservoirs would not be commercially attractive without a natural results in an estimation of the fracture porosity. Aguilera provides
or induced fracture system. While fractures are unquestionably im- a chart for estimating V, with the basic assumption that the matrix
portant with respect to formation permeability, the importance of and fractures are connected in parallel.
the contribution that fractures make to the reservoir storage capac- Weber and Bakker7 state that the classic method for estimating
ityor, more specifically, the porosity is unclear. The assigmnents fracture porosity is to examine cores, count the number of natural
of porosity in a dual-porosity system (e.g., matrix and fractures) fractures, and measure their width. The qualifications placed on
may be critical in estimating both in-place reserves and ultimate this method include having a knowledge of fracture systems, bed
production of hydrocarbons. An assigmnent of too much porosity thickness, and lithology.
to the fracture system could lead to grossly optimistic production Bergosh and Lord 8 report on two new techniques: computerized
and economic predictions.' tomographic (CT) scanning and fluorescent epoxy impregnation.
Porosity is expressed in terms of void (pore) space per unit bulk Both techniques are enhanced by use of full-diameter cores, and
volume. The terms grain volume, PV, and bulk volume also are the tests are conducted at overburden stress conditions. Bergosh
involved. To calculate porosity, two of the three must be meas- and Lord state that these techniques have advantages over other
ured or known. In a fracture system, the pore space is a function core-analysis methods, such as Boyle's-law helium porosimetry,
of the block size, which relates to the frequency of the fractures, thin-section microscopy, and brine displacement.
and the fracture width. A simplified example of a fracture porosity Well testing and material balance are additional tools that are used
calculation is shown in Fig. 1. The arbitrary assignment of a frac- to estimate fracture porosity. Material balance appears to suffer in
ture width of 254 I-'m results in a porosit} of 1 %. (The permeabil- accuracy if matrix porosity exists, and buildup curves likewise have
ity of this fracture is about 54 darcies. ) limitations, such as insufficient time to obtain truly representative
Jones 3 states that fracture volume and, therefore, fracture porosi- curves. 7
ty are directly proportiOlial to fracture clearance, and that the ef-
fective porosity is proportional to the cube root of permeability. Fracture.Wldth/Block.Helght Considerations
Fracture porosity is a function of block size and fracture widths.
Fracture Models Fractures with large widths occurring at infrequent intervals (i.e.,
Reiss 4 proposes four types of fracture networks for calculating large block sizes) will yield very low porosities. Extensively frac-
fracture porosity (Fig. 2). The rectangular elements (blocks) have tured rock (i.e., small block sizes) with very small fracture widths
sides a, , az, and a3, and the fracture widths are denoted by b (Fig. likewise will yield low porosity values. In other words, to have
3). For the four schemes the fracture porosity is calculated as significant fracture porosity (e.g., 4 to 6%), the formation must
follows. be fractured with a high degree of frequency and the fracture widths
must be of a significant size.
(Type A) Sheets: Table 1 shows the results of a literature survey of docu-
<bJ=b/a. . ........................................ (1) mented',7,9,10 fracture widths and block heights. With one or two
exceptions, the fracture widths fall into the 100- to 200-l-'m range.
(Type B) Matchsticks: Very little is documented regarding fracture frequency or block size.
The one confidential study that reports on extensive fracture den-
<bJ=2b/a. . ....................................... (2)
sity (336 fractures for one foot of core examined) couples this with
(Type C) Cubes with two effective fracture planes: assigned fracture widths of only to I-'m. This is a classic example
of the small-block-size, small-fracture-width scenario that, in this
<bJ=2b/a. . ....................................... (3) case, results in a calculated fracture porosity of 1. 1 %.
Fracture-width studies of cores can be misleading from several
(Type D) Cubes:
standpoints. Fractures may be observed and reported, but commen-
<br3b/a . ........................................ (4) tary may be missing about whether the fractures are open or closed.
Fractures that are open in cores may be closed in situ from
Here, a=matrix size, in. [cm], b=fracture width, I-'m [mm], and mineralization 7 or by tar or kerogen materials. Overburden stress-
<bJ=fracture porosity, %. es are seldom taken into account, and the width of a fracture at
Warren and RootS used an idealized model in their study of the reservoir conditions will be a function of the net confining pres-
behavior of primary and secondary porosity. They assumed that sure; i.e., the effective compressive stresses are the difference be-
all the secondary porosity was contained within an orthogonal sys- tween the total compressive stresses and the pore pressures."
tem of continuous, uniform fractures that are oriented so that each Matrix core porosities normally are adjusted according to reser-
fracture is parallel to one of the principal axes of permeability. This voir stress conditions, and the same approach should be taken to
is the same as the Type D4 fracture model that yields the maxi- evaluate fracture widths.
mum fracture porosity; however, the mathematics suggested by Fracture widths of 200 pm seem reasonable; however, 6,OOO-l-'m
Warren and Root holds true for any type of fracture system. extension fractures existing at overburden-stress conditions appear
unlikely or may have been created during the coring/core-retrieval
processes. The data shown in Table 1 appear to make Aguilera'S'
'Now retired.
"paper-thin," minimum-fracture-width statement somewhat out of
order when one considers that common copy-machine paper is about
Copyright 1989 Society of Petroleum Engineers 400 I-'m thick.
SPE Fonnation Evaluation, December 1989 531
/i-----------------"lI
· ~I-"- - 1 " -----+l~1 / I I

T
/ I I
/ I
/ I

r::'" I
/
/
/
I
I
I

I // a2 I

//1
//
/;=======-=- 1"
/
/

f---- ---...Ll-------
1
-.01'~ // I
T:Z=::::;:=======7 I I
I II
/--------
I
}-.l.------ --I
/
/
/ J---- I
/

I
/ I/ a3 I
I
Fig. 1-Baslc fracture porosity model: .p, =1%. II / I
1 / 1
I
I I I I
b / / I I
-2'
, I ',1
,I
L,_____________________ J ~

Fig. 3-Deflnition of fracture porosity (after Reiss 4 ).

a b Fracture Porosity Discussion


My observations indicate that fracture porosity may be overesti-
mated, particularly in dual-porosity systems. Fractures have such
a dramatic impact on reservoir performance that their hydrocar-
bon storage capacities may tend to be inflated.
Pittman 12 subdivides porosities into f<;>ur groups - intergranu-
lar, dissolution, microporosity, and fractures. He states that frac-
ture porosity is rarely greater than a few percent and is usually
<1%. ,
Table 2 shows the results of a literature survey of docu-
mented 7,8,10,12,13 fracture porosity values. Note that the values
generally support Pittman's conclusion. Bergosh and Lord's8
reported fracture porosity of 9.64 % was determined on one sam-
ple. Bergosh* stated that all porosity measurements were made on
full-diameter samples of an undefined diameter and that the sam-
c d ples were generally 2 to 3 in. [5 to 8 cm] long.
Of the four models suggested by Reiss, 4 maximum porosity re-
Fig. 2-Typical fracture networks. Arrows indicate possible sults from a fracture network represented by cubes. The relation-
directions of flow.
'During discussion portion of Ref. 8 at oral presentation.

TABLE 1-FRACTURE WIDTHS AND SPACINGS

Formation Fracture Widths (mm) Fracture Spacing/


Reference Information Range Frequency Comments
Aguilera' General statement Paper thin. 6 +
Elkins 9 Spraberry sandstone 0.33 (maximum) 0.051 Few inches to a few feet
Snow'o Selected dam sites 0.051 to 0.10 4 to 14 It
Aguilera 1 La paz·Mara field 6.35 (maximum)"
Confidential study Monterey 0.01 * 3 to 336 fractures/It
Weber and Bakker 7 Small joints' 0.01 to 0.10
Weber and Bakker 7 Extension fractures' 0.1 to 1 0.2
Weber and Bakker 7 Major extension 0.2 to 2t
fractures'

'Unidentified reservoirs.
'';Converted to metric units.
Partial infilling.
fAssigned.

532 SPE Formation Evaluation. December 1989


TABLE 2-POROSITIES OF FRACTURED RESERVOIRS

References Formation Information Porosity Units


Pittman 12 General statement 1
Snow 10 Beaver gas field 0.05 to 5
Stearns and Friedman 13 Austin chalk 0.2
Confidential study Monterey 0.01 to 1.1
Weber and Bakker 7 South African karst zone 1 to 2
Bergosh and Lord 8 CT scan examples' 1.53 to 2.57
Bergosh and Lord 8 Epoxy injection examples' 1.81 to 9.64

'Three unidentified samples.

TYPE 0 TYPES B&C TYPE A

TYPE 0

FRACTURE WIDTH-MICRONS FRACTURE WIDTH-MICRONS

Fig. 4-Porosity-fracture-wldth relationships for various frac- Fig. 5-Poroslty-fracture-wldth relationships for various frac-
ture models with a 1-cm block size. ture models with 4-cm block sizes.

ship of fracture porosity with fracture width for a constant block


size may be developed from Fig. A.3.8 of Reiss. 4 Fig. 4 shows
these relationships for a very optimistic block size of 0.4 in. [1 cm].
Note that fracture widths of 200, 300, and 700 ILm are required TYPE 0

to develop 6 % porosity for the four fracture systems. Fig. 5 depicts


the drastic change in porosity as the block size is increased to 1.6
in. [4 cm]. Fig. 6 shows the relationship between porosity and block
size for a constant fracture width. An arbitrary 200-ILm fracture
width was chosen for this example. This graph demonstrates the ....z
reduction in porosity with increasing block size. Note that a w
o
a:
maximum porosity of abOut 2 % is associated with a I-in. [2.5-cm] ~I
block size. >
!::
An exception to this may be found in California's onshore/ <I)

~ 3
offshore Monterey formation, which has reservoirs that may have oa.
abnormally high fracture porosities. Redwine ll focused his work
on the Monterey shale. He qualifies the term "shale" to connote
that the rock is made up of clay-to-silt-sized particles and is lami-
nated, fissile, or both. His shale term does not imply that the rock
has a significant content of clay minerals. He speaks of the Santa
Maria area where fracturing is almost ubiquitous, with the dominant
type being well-ordered and systematic. He states that the rocks
of this area usually are dense and brittle, and extension fracturing
and brecciation are intense. It is conceivable that a breccia-type
rock could contain the porosities (and permeabilities) to explain the BLOCK SIZE-CENTIMETERS

high production rates of heavy, viscous oil recorded from some


wells. Sangree 14 discusses breccias in detail and cautions that a Fig. 6-Poroslty-block-slze relationships for various fracture
models with a 200-"m fracture width.
core box full of drilling-induced rubble should not be called a brec-
cia. Given the previous mathematics for calculating fracture porosi-
ty, it is still unlikely that a breccia-type rock would yield porosities
much greater than 2 to 3 %.

SPE Fonnation Evaluation, December 1989 533


Weber and Bakker7 discuss the chert associated with the Mon- Acknowledgments
terey and state that porosities of some 8 % are possible. On the ba- I thank Sun E&P Co. for permission to publish this paper. I also
sis of their thin-bed chert occurrence, examinations of long intervals thank Sam Bass, Joan Hahne, and Lynda Chandler for their valua-
of Monterey core, 8 and my observations, it is unlikely that such a ble assistance in the preparation of various parts of the paper. A
lithology condition exists for a vertical interval2! 100ft [2! 30.5 m]. special thanks is due Eve Sprunt of Mobil R&D Corp. for her many
Fracture studies often yield extremely low fracture porosity helpful suggestions and overall guidance.
values, as Table 2 shows. Indirect estimates (e.g., wireline logs)
sometimes yield significant storage space assigned to the fracture References
system; however, the fracture-width and block-size considerations I. Aguilera, R.: "Geological Aspects," Naturally Fractured Reservoirs,
previously documented make these estimates appear overly optimis- PennWell Publishing Co., Tulsa, OK (1980).
tic with a few possible exceptions. 2. "The Fundamentals of Core Analysis," Core Laboratories, Dallas, Sec.
Various types of studies may be conducted to provide fracture 2, 10.
porosity values. The same rules should apply for naturally frac- 3. Jones, F.O. Jr.: "A Laboratory Study of the Effects ofConfming Pres-
tured reservoirs as for more conventional reservoirs when the va- sure on Fracture Flow and Storage Capacity in Carbonate Rocks, " JPT
lidity of the data derived is considered. Be it a well test, a series (Jan. 1975) 21-27.
4. Reiss, L.H.: The Reservoir Engineering Aspects of Fractured Forma-
of downhole logs, or core studies, only a small portion of the reser-
tions, Gulf Publishing Co., Houston, TX (1980).
voir is represented. This presents such valid questions as whether 5. Warren, J.E. and Root, P.J.: "The Behavior of Naturally Fractured
the areal representation of the total lithology is adequate; i.e., do the Reservoirs," SPEJ (Sept. 1963) 245-55; Trans., AIME, 228.
data tend to high grade or low grade because of the well's position? 6. Aguilera, R.: "Relative Permeability Concepts for Predicting the Per-
The same holds true regarding the vertical cross section. Ques- formance of Naturally Fractured Reservoirs," J. Cdn. Pet. Tech.
tions to address include (1) was the complete section cored; (2) was (Sept. -Oct. 1982) 41-48.
the core recovery satisfactory both from a quantitative and qualita- 7. Weber, K.J. and Bakker, M.: "Fracture and Vuggy Porosity," paper
tive standpoint; (3) was the wellbore in satisfactory condition to SPE 10332 presented at the 1981 SPE Annual Technical Conference
produce reliable results from logs; and (4) were the well tests ham- and Exhibition, San Antonio, Oct. 5-7.
8. Bergosh, J.L. and Lord, G.D.: "New Developments in the Analysis
pered in any manner. Additional valid questions may deal with the
of Cores From Naturally Fractured Reservoirs," paper SPE 16805
analyses, data handling, and data interpretation. If a core study was presented at the 1987 SPE Annual Technical Conference and Exhibi-
conducted, did the analyst unintentionally high grade or low grade tion, Dallas, Sept. 27-30.
on the basis of sample sites? Was the proper analysis method used? 9. Elkins, L.F.: "Reservoir Performance and Well Spacing, Spraberry
Did the log analyst use proper judgment and analytical procedures? Trend Area Field of West Texas," JPT (July 1953) 177-96; Trans.,
Did the reservoir-simulation engineer use state-of-the-art models, AIME,198.
and were there uncertainties resulting from lack or quality of 10. Snow, D. T.: "Rock Fracture Spacing Openings and Porosities," Ameri-
input data? can Soc. of Civil Engineers, Soil Mechanics and Foundation Div. (1968)
These comments are not meant to demean any of the many ana- 94,73-91.
1I. Redwine, L.: "Hypothesis Combining Dilation, Natural Hydraulic Frac-
lytical tools. Instead, they are intended to present a challenge for
turing, and Dolomitization to Explain Petroleum Reservoirs in Mon-
the prudent engineer/technologist to probe deeply into all facets of terey Shale, Santa Maria Area, California," The Monterey Formation
each evaluation system and not blindly to accept any assigned frac- and Related Siliceous Rocks of California, Soc. of Economic Paleon-
ture porosity value without considering the possible limitations re- tologists and Mineralogists, Pac. Sec. (1981) 221-48.
garding accuracy and representation. 12. Pittman, E.D.: Porosity, Diagenesis and Productive Capability ofSand-
stone Reservoirs, SEPM Special Publication No. 26, Soc. of Econom-
ic Paleontologists and Mineralogists, Tulsa, OK (March 1979) 159-73.
Conclusions 13. Steams, D.W. and Friedman, M.: "Reservoirs in Fractured Rock,"
1. Fractures, either naturally occurring or induced, are the life- AAPG Memoir 16, Tulsa, OK (1972) 82-106.
blood of many reservoirs because of their influence on well deliver- 14. Sangree, J.B.: "What You Should Know to Analyze Core Fractures,"
ability. World Oil (April 1969) 69-72.
2. Overestimating fracture porosity potentially can lead to eco- 51 Metric Conversion Factors
nomic failures in the development of naturally fractured reservoirs. ft x 3.048* E-Ol m
3. Fracture porosity generally may be insignificant with respect in. x 2.54* E+OO cm
to the total reservoir hydrocarbon storage space.
4. Fracture porosity, in many instances, is less than 1 %. 'Conversion factor is exact. SPEFE
5. Technical personnel and financial planners should look criti- Original SPE manuscript received for review Sept. 27, 1987. Paper accepted for publica·
tion June 12, 1989. Revised manuscript received Jan. 17, 1989. Paper (SPE 16806) first
cally at hydrocarbon-storage-space values associated with the frac- presented at the 1987 SPE Annual Technical Conference and Exhibition held in Dallas,
ture network through whatever means of assignment. Sept. 27-30.

534 SPE Formation Evaluation, December 1989

You might also like