Effects of Perforation-Entry Friction

on Bottomhole Treating Analysis·

J.B. Crump, Halliburton Services
M.W. Conwa." * SPE, Halliburton Services

-.:5PE 154/~
Sum......,. Analysis of field and laboratory data shows that variations in pressure drop caused by changing perforation-entry
friction tends to influence the prediction of fracturing treatment performance. This paper presents experimental data on perforation-
entry friction as it affects fracturing treatment design.
Prefracturing treatment planning practices include examination of numerous treating-pressure charts to determine formation type
curves, which are used to anticipate fracturing treatment performance and screenout modes. Perforation-entry friction may vary
greatly because of erosion of the perforation and near-wellbore fracture, and this changing friction pressure is often not properly
accounted for in planning.
This paper presents discussion and data (laboratory and field) that show the degree of perforation erosion encountered in
fracturing operations and proposed guidelines to determine when to alter pumping schedules to account for proppant erosion to
perforations, cement sheath, and formation.

Introduction
Auid rheology measurements, densimeters, and flowmeters, com- Components of fracture-entry friction include perforation fric-
bined with recent advances in computing power, allow determina- tion, cement-sheath friction, formation damage (resulting from per-
tion ofbottomhole treating pressure (BHTP), Pbht, on a real-time forating), and fracture friction. Fracture friction can be calculated
basis. This pressure is actually the BHTP inside the casing. The and depends on the fracture treatment design and fracture width.
true BHTP is the pressure inside the fracture. BHTP and forma- Formation damage can be minimized by shooting under-
tion bottomhole pressure (BHP) are used interchangeably. balanced,6,7 by proper design of perforation schedule. and by
The missing link and principal unknown in hydraulic fracturing ~medial acid-cleanup jobs. Cement-sheath friction owing to per-
is fracture-entry friction, Pfq. It is usually assumed to be equal to forating damage and its erosive properties is shown to have only
zero, to be a constant, or to be a negligible influence on fracture a minor effect on total perforation-entry friction when slurries are
treating pressure. Fracture-entry friction is the total pressure drop pumped.
experienced by the fluid from the casing through the perforation The major component of Pfq is Pf' which is detailed below. The
and perforation tunnel to the fracture tip. Perforation friction, Pf' other components of Pfq are minor constituents. When sand-laden
is the pressure drop of the fluid passing through the restriction of slurries are pumped at high differential pressure across perfora-
the perforation in the casing. Current technology can determine the tions, Pf changes constantly. Field and laboratory data have been
fracturing pressure in the casing but may not properly account for combined to derive coefficients theoretically and empirically and
the changing BHTP caused by changing Pfq' This paper addresses to check this equation. The equation commonly used to predictPfis
the changing Pfq that occurs during pumping of sand-laden
slurries. 0.2369 q2p
Pioneering work by Nolte and Smith 1 created increased indus- Pf= , .................................. (2)
try awareness of the necessity for accurately determining forma- nl d4Kd 2
tion BHP during a fracturing treatment. In an extension of Nolte
and Smith's work, Conway et al. 2 proposed the analysis and use where
of treating-pressure type curves to predict well type and screenout q = total flow rate,
mode during the treatment. Many hydrocar:bon zones are bounded p = fluid density,
by a delicate boundary layer that may be fractured by pressure a np= number of perforations,
few hundred psi over design pressure. Analysis of the Nolte and d = perforation diameter,
and
Smith plot is used to determine whether the fracture has broken Kd = discharge coefficient.
out of zone. Changing (decreasing) perforation friction pressure
during a treatment can be interpreted on a Nolte and Smith plot Erosion of perforations in tubular goods and the subsequent drop
as evidence of breaking out of zone. in Pf are the main points of emphasis in this paper. Laboratory data
Eq. 1 is used to calculate BHTP: show that perforation-friction changes alone can cause errors in in-
Pbht=Pw+Ph-P,-Pfq' ............................. (1) terpretation of Nolte and Smith plots.

where Experimental
Pbht = BHTP, Apparatus. Three series of tests were run that pumped sand-laden
Pw = wellhead pressure, slurries through perforated casing. Sand slurries are referred to by
Ph = hydrostatic pressure, their concentration-i.e., pounds of sand per gallon of gelled fluid.
p, =
pressure caused by fluid friction in tubulars, and Sand concentrations varied from 2 to 20 Ibm!gal [240 to 2400
Pfq =
pressure caused by fracture-entry friction. kg/m3], and differential pressure across the perforations ranged
up to 1.500 psi [10.4 MPa].
Two calculated values are present in this equation, p, and Pfef' A centrifugal pump was used to circulate the slurries and to boost
Several recent papers3-S described means by which the calculation a positive-displacement pump with protective fluid end covers. This
of p, can be improved, particularly in the case of sand-laden pump was used to circulate slurry through the test fixture. Slurry
slurries. was pumped into a short, perforated section of casing suspended
'Now with Slim-lab Inc. in the slurry mixing tank (Fig. 1). The flow rate was adjusted to
Copyright 1988 Society of Petroleum Engineers achieve the desired perforation pressure drop; a radioactive den-

Journal of Petroleum Technology, August 1988 1041

 tr--------------~----
I
.. ----E-
I
='-[Link]J
: :r:;:;,;;;l
I I I
I I I
... I I
I I
I
I
I
I
I
'If
I
FLOW
..
I
I

Fig. 1-Pumplng setup.

ORIFICE-SQUARE EDGE
simeter was used to monitor sand concentration. A magnetic flow- (PERFORATION)
rate pickup and a pressure transducer were used to monitor and
to control the test. A 100bbl [1.6-m 3] mixing tub was used on Test
Series 1 and 2 and a 50-bbl [8-m 3 ] tank on Test Series 3. Safety
relief valves were used on the positive-displacement pump.

Materials. Selected materials included a common fracturing prop-

pant, 20/40-mesh sand, on all tests. Casing selected for Test Ser-
ies 1 was 4~-in., l1.6-lbm [114.3-mm, 5.3-kg] J-55 casing with
round, drilled perforations. For Test Series 2, 4Ih-in., 11.6-1bm
FLOW
..
. . . . . 7-'--.. "
[114.3-mm, 5.3-kg] N-8O casing with round, drilled perforations
was chosen. Test Series 3 used 4"h-in., l1.6-lbm [1l4.3-mm,
5.3-kg] N-8O casing. The 4~-in. [114.3-mm] N-8O casing was
cemented (on Test Series 3) inside a piece of 8.O-in. [203.2-mm]
-OD by O.I09-in. [2.77-mm] thin-wall tubing. Cement-sheath thick-
ness was 1.64 in. [41.66 mm]. The 8.O-in. [203.2-mm] thin-wall VENA
tubing is not a component of normal wen completions; it served CONTRACT A
only to hold the cement in place. Test fixtures were then jet
perforated with shaped charges. Perforation diameter in the cement
was approximately twice the perforation diameter in the casing.
Perforations penetrated the exterior 8-in. [203.2-mm] tubing, leaving
the cement sheath intact. Perforation exits in the exterior 8-in.
ORIFICE-ROUNDED
[203.2-mmJ tubing were then drilled out to prevent any backpres- (ERODED PERF)
sure or effects on the data. Fluid selected was gened fracturing fluid
containing 40- to 70-1bmll,OOO-gal [4800- to 8400-g/m3] hydroxy-
propyl guar (HPG).

Procedure. Test Series 1 involved pumping sand-laden slurries

ranging up to 20 Ibm/gal [2400 kg/m3] through one each of four
r.,
sizes (0/)2, Y16, and ~ in. [7.14,9.53, 11.11, and 12.7 mm])
of drilled perforations. A water-calibration test and a gel test were FLOW
..
- f--
made on each set of perforations before sand was pumped; only
a water test was made after sand was pumped. Slurry was recircu-
lated through the tub and test fixture. Sand breakage was not a con- ..... ..... "
cern because of results reported by Roll et al. 8 .
Test Series 2 involved pumping a sand-laden slurry in a "typi-
cal" job procedure through a set of three r.-in. [9.53-mm] drilled VENA
perforations. After the pumping rate was established with gelled
fluid at 400 psi [2760 kPa], sand was added to obtain 2 Ibm/gal
CONTRACTA
[240 kg/m3] for the first stage. About 10 seconds was needed to
add sand for each increment. Each stage was pumped for 2.5 minutes
at 3.6 bbllmin [0.6 m 3/min]. The sand stages were incremented
in 2-lbm/gal [240-kg/m 3] stages up to 14 Ibm sand/gal [1680 FLOW NOZZLE
kg/m3] with flow rate held constant.
Fig. 2-Perforatlon va. flow nozzle.
Test Series 3 involved pumping a representative low sand con- '
centration of 2.5 Ibm/gal [300 kg/m3] and a medium-range sand
concentration of6lbm/gal [720 kglm 3J. The slurry was circulated round, drilled perforations. Drilled perforations were chosen to
for about 30 minutes in each test through sets of four 0.28-in. achieve a perfectly round perforation of known diameter. A perfo-
[7 .ll-mm] jet perforations. These tests at higher pressure drops ration discharge coefficient cannot be determined unless the exact
through the perforations simulated a limited-entry job design. cross-sectional area is known.
Test Series 3 featured jet perforations and a cement sheath to es-
Application to Downhole Conditions tablish whether the cement sheath was a major contributor to
Concerns always arise as to the direct application of data generated fracture-entry friction. These tests indicated that cement-sheath fric-
on the surface to downhole conditions. Test Series 1 and 2 used tion is minor and can be neglected. The perforation cross-sectional
 TEST SERIES 1 - CASING : J-55 TEST SERIES 1 - CASING : J-55
PERfS : DRlU£D PERFS : DRIUED

.. T------r------~----~----~----__, 1100

IlOO

~
~+-----fl+--~~~~--~~~~----~
5'
~
z
~ mN+----H+-~~~~--++~--~--~

i
~ IOOO+_---[Link].+---,P---'i+----I-..f-----_+----___f
j
! _r--*~~~~~----+---~
~ J
, Jr!r!J
7
~

1- ~+_~r_~~.--~~----~----_+----___f
r .
j I
200+----0 200 / 1//-7?
d
0

A
I

Q
Legend
D WliTtJlIlf1Oll£ &!IIIIIY
"[Link]. S!.I.!!!IIT
() WAl[II AF'TtlI SWIIIIY

o+-----i.+------~----~----_+----~ o
o 50 100 150 o 50 100 150 200 no
nDWRATE (GPW) nDWRATE (GPM)

Fig. 3-Water and gel flow ratea. FIg. 4-Pressure drop V8. flow rate of a o/I-In. perforation.

TABLE 1-WATER K d • TEST SERIES 1- lIST SER1ES 2 - CASING : N-80

DRIUED PERFORATIONS PERFS : DRIUlD
Kd of Eroded
Perforations - ........,
Initial
Perforation Kd of
New
After
10-Ibm/gal
After
2O-Ibm/gal - 1\.
~2\ ~
-
Diameter
(In.) Perforations Sand Sand
0.49
0.54
0.86
0.85
0.87
0.88 !
z ... \ r..
0.58 0.88 0.88 •,., ... \1 \
0.64 0.93 0.88
I- I
l\.~'" ,.
it
~-
II I'~ ... \ 1Of,..
flow area through the cement was approximately four times the area
of the perforation. so the j~ of fluid exiting the perforation would
...9
!- IV -
\
\j '\ ~'."
experience only minor effects from the cement. /(.. ~ lAND
I '1\:- ...
i-
COlIC ,..".1
Because Test Series 3 featured jet perforations. a "shock-

-
hardened" erosion-resistant perforation in the casing would be ex-
pected. 9 Even with this hardened area, the pressure changed rapid- \
ly, as discussed below. \
The effects of backpressure on erosion rates may be a concern. J.
In theory. backpressure suppresses the effects of the vena contrac-
fa and changes the erosion profile. (Vena contracta is defined as
any of the contracted parts of the minimum cross section of a jet
of fluid discharging from an orifice.) Our laboratory research found
that no measurable effects on the erosion rate could be obtained
- ,

Fig. 5-Preasure drop va. time through three %-In.

• II
PIJIoIPNO 1K (111M)
12 .. • • »

by holding various backpressures with a constant /lp across the per- perforatlons-3.8 bbUmln.
foration.

DI8ouulon of Full-Scal. T••t. nozzle (Fig. 2). Pull-scale tests show that Kd approximately equals
One standard fluid handbook 10 recommends a discharge coeffi- 0.56 for water and gelled fluids (50 Ibm HPG/l.000 gal [6000
cient. Kd • from 0.61 for a sharp-edged orifice up to 0.98 for a flow glml]) in round, sharp-edged. drilled perforations. The gelled fluid
nozzle. The coefficient is a measure of the perforation' s efficiency and water Kls varied about 3% and, allowing for accuracy oftest-
at letting fluid pass through. ing equipment, can be assumed to be equal (Fig. 3). After sand-
An orifice discharge coefficient of 0.95 indicates an efficient laden fluids are pumped at concentrations from 10 to 20 Ibm/gal
orifice with a smooth, tapered inlet approaching the shape of a flow [1200 to 2400 kg/ml] through %2-, Yw. and Ih-in. [7.14-, r.-.
Journal of Petroleum Technology, August 1988 1043
 Fig. 6-Test Series 1-drllled perforations, after sand Is pumped.

9.53-, 11.11-, and 12.7-mm] perforations for 4 minutes, the aver- than is the flow rale for water (Curve 2 vs. Curve I): this results
age perforation discharge coefficient for both water and gel was because of the increased diameter and an improved (or higher) K d,
found to be Kd=0.89 (Table I). which occurred immediately.
Water and gelled fluids were shown to be nonerosive at pressure In Curve 2, the top ponion of the curve has an almost vertical
drops up to 1,500 psi [10.34 MPaJ. Under these conditions. the slope because of rapid erosion: the bottom ponion shows a more
Kd can be determined ellplicitly. When sand-laden slurries were parabolic slope. The erosion rate is much lower below 5OO-psi
used, the actual Kd of the slurry itself could be calculated only after [3450-kPa] pressure drop; therefore, the curve of flow rate vs. pres-
the test. Fig. 4 shows data for one such test. Because the sand slur- sure drop is more realistic. Curve 3 shows another water-calibration
ry is more dense than the base fluid, the pressure drop should have run made after sand erosion on the perforations. Because of a
been greater for the slurry flow rate. smoothed entrance and increased area and Kd, Curve 3 is displaced
This instantaneous increase in pressure resulting from increased to the right and the slope is changed.
density is shown in Fig. 5: however, the erosion rate is so high When sand slurries are pumped. the K d is lower than for water,
that the only place where the Kd can be measured is after it has showing reduced efficiency of slurry flow: i.e .. the solid and liq-
increased to ils equilibrated value and the diameter of the hole has uid phases try to ellit a perforation. Velocity can also be plotted
enlarged. In Fig. 4, Curve I is a calibration run in which water on thex axis. Velocity for Curve I (water) would be plotted from
is pumped through drilled holes. Curve 2 is a plot of pressure vs. o to 725 ftlsec [0 to 221 mls] with major divisions of 145 ftlsee
flow rate for 16-lbm/gal I1920-kg/m 31sand slurry. Note that the [44.2 m/s]. Velocity for Curve 3 (water) would be plaited from
flow rate of sand-laden gel is actually higher for a given pressure Oto 450 ftIsec [0 to 137 mls] with major divisions of90 ftJsec [27.4

1044 loumal of Petroleum Technology, August 1988

 ~O.:- ________ ~ ______ ~ ________ ~ ________ ~

TEST SERIES 3 - CASING: N-80

IX " 25 AND 6 LB/GAL SWRRY
1.1 .~~ ~APE CHARGE PERFS WITH CEMENT SHEATH

I
~ o . o
- ft----+---+---4----------.--r--~

to--o..
t '", .n: ~J---+---+-+----t
wiw.r
U • •7 to---------1'-------+111---------1111----------420 ~- ~ --------~--+-------
SAND CONceNTRAnON (LB/GAU \ - - - 0 . - ....

fig. 7-Perforatlon K" va. und concentration while und Is

'-r--JJ;J ~ -- --.--~---
~
---
pumped.
I--()..~

TEST SERIES 1 - 50 LB/M HPG GEL

16 LB/GAL SAND SWRRY ~---~---~----~--~-----+---~
PERFS: DRUID J-55
[Link]'"---......,~-----r-----~------,....----
!
...
"III'
DlA (IN.I
..Allu+----l---~
INCIIIAII!
I'SI (.211 "

! .... -I-----1~7,;;s::;;:.~::::~ f--:;,-+---+-----t

112 (.101 M

~t_--~----;-----l~--~--~

1OO+---~~--_+----~----4_----~--~
OS. e 20
PUMPING nME (MIN)

FIg. 9-Perforatlon friction pressure va. time.

WELL. IXPnt • aooo FT • MINI ...... STEPS lZE MAXII4.I4

CASING SIZE. 5 . 5 IN .
TWING SI ZE • 2 . a~ IN • • 6.5 L8 OIQTY RATE 0 0 .0 l .O 30 . 0
N..I49ERr7 PERFS • 21 SA,.., CCNXNTRATIDII 0 0 .0 1 .0 10 . 0
PERf: DIAMETER· 0. 28 IN . PERFOAATION FRICTIC»IIi:1 0.0 '00 . 0 '000 . 0
JO to AOOO

27 • 3600

2' e 3 200

21 7 2&00

, 18 to 2 AOO

0'+-----~----~------+_----~-------4
o 100 150
flDWRAlt (GPM)
200 .
~
>-
15 ~ 2 000

; 12 A 1600
C
9 1 1200
FIg. 8-Preaaure drop va. flow rate.
6 2 800

3 1 '00
m/s]. Erosion affects the data on Curve 2 (Fig. 4) because the ini-
tial diameter is 0.375 in. [9.53 mm) and the final diameter is 0.48 ~+,OO~~-13~,.=~=-==
.. r,JO~--,-.~"~----1-6,~OO----,~.,-••-=~l7jl~ 0
in. [12.19 mm); therefore, velocity axes are not plotted. sc .(

An attempt was made to determine the Kd for the sand slurries

themselves. The Kd for slurries can be calculated only after the FIg. 10-Fleld case showing perloratlon friction VB. time. Note
test, when the area of the eroded perforation can be determined. that as sand Is added (14:50), perforation friction (4) de-
Fig. 6 shows some typical drilled perforations after sand is pumped. creased because of perforation erosion.
Perforations 2 (upper right) and 4 (lower right) have almost per-
fectly circular cross sections with smooth, tapered inlets. Perfora-
tion 1 (upper left) has a slightly elliptic configuration. The major stream of slurry eroded preferentially toward the top of the cement
axis of the ellipse is vertical because of the vertical momentum of tunnel. Note also the elliptic cross section of Perforation 1 in Fig. 6.
the sand slurry. Perforation 3 (lower left) shows an extreme ir- Fig. 7 is a plot of the perforation discharge coefficient for vari-
regularity caused when a hard spot in the casing left a protrusion, ous sand concentrations up to 20 Ibm/gal [2400 kglm 3 ] while sand
or nipple, extending into the otherwise circular cross section. slurry is pumped. The water or gel (no sand) discharge coefficient
An interesting observation was made. Slurries traveling verti- (0.89) would be applicable to a smooth, tapered inlet (Fig. 2) after
cally down a weUbore do not "tum the comer" and exit perfectly sand had been pumped.
horizontally through the perforations. Some of the slurry impacts
on the lower side of the perforation. This imparts an upward com- Factors Affecting Erosion Rate
ponent to the velocity, as evidenced in the wear pattern in the per- [Link] Pressure Drop. Fig. 8 illustrates the effects of ve-
foration tunnels observed in the cement (not pictured). The deflected locity and pressure drop on erosion rates. In this test, the first flow-

10urnal of Petroleum Technology, August 1988 1045

 [Link]!t _ - ..... PSI
P\.OTST_IlllE-_

,." 13 . ~ 9600

12 . 0 9200

--
10.5

9 .0 "00

..
6.0,-
,.~

~
aooo

1200

J.O 6100

1.5 ...00

".00 .... 30
0
15.00
SC
.-
6000

FIg. 12-Nolte and SmHh ·plot postJob analysis of field case

In fig. 11. Although _lysis shows that well was In ecreenout
FIg. 11-Fleld case Mowing effect of perforation erosion on mode In Ita late atagea, this Indication wae masked by treat-
BHTP anelysla. BHTP (4) decrea8es . . . .net 18 added. Ing preaure dropbecauee of perforation eroelon.

rate data point was the highest pressure·. Pressure was decreased 6-lbmlgal (720-kg/m 3 ) run. The stable, pseudo-ftrst-order be-
in increments of 100 or 200 psi [690 or 1380 kPa) and another flow havior of the later-time data is taken to indicate perforation enlarge-
rate data point taken. The portion afthe curves above Section A-A ment at a relatively constant Kd • The faster rate for the 6-1bm
bas an almost vertical slope. Perforations eroded and enlarged so sandlgal (720-kg/m 3) test clearly indicates the relationship of sand
quickly that the lower pressures bad almost the same flow rate as concentration to rate of erosion.
the higberpressures. Below SeCtion A-A, the curves follow the ex- Fig. 5 confirms the higher-sand-concentration effect above 12
pected parabolic form because of the diminishing effects of ero- Ibmlgal [1440 kg/m 3). Smoothing of the entrance and application
sion. The fmal diameters were measured and the corresponding oflower pressure should reduce the erosion rate. Hoies that were -
velocities calc:u1ated. A critical velocity of ± 120 ftlsec [±36.6 mls] smoothed previously by lower sand concentrations erode much more
is shown, above which erosion occurred so rapidly that the effects rapidly than they did at lower concentrations, even though they are
of velocity changes could DOt be assessed accurately. Note that these experiencing a lower differential pressure.
boles were drilled and therefore were not subject to shock hardening
from jet perforation. 9
Manufacturers of discharge iron and manifolding recommend Field R••ulta
limiting sand-laden fluid velocities to less than 35 ftlsec [10.7 mls]. The most common perforation practice encountered in wells to be
Ata velocity of35 ftlsec [10.7 mls) ina r,-in. (9.53-mm] perfora- stimulated hydraulically is that of shooting a rather limited number
tion, flow rate would be 0.28 bbUmin-perforation [0.045 of perforatio~ compared with the density of perforations that would
m3 /min·perforation), and at Kd=O.60, expected Il.p would be be used for natural production. Typically, the minimum number
22.7 psi [156.5 kPa). Therefore, we concluded that most fractur- of holes to be shot would be the number required to give less than
ing treatments are conducted in such a manner that the perforation l00-psi [690-kPa) perforation friction at the designed treating rate.
velocity is high enough to cause erosion during the treatment. From field studies on measured perforation friction, we fmd that
the effective cross-sectional area and flow regime through the per-
Perfontioo Size. Smaller perforations erode faster for the same forations is much lower than calculated. After breakdown and ball-
Il.p than do larger perforations (Fig. 8). It is very difficult to separate out, it is common to encounter extremely high perforation friction
experimentally the variables of rate, 4p, and perforation diameter pressures during hydraulic fracturing treatments. A rather dramatic
because none remain constant while sand-laden slurry is pumped, example of what can be encountered in field operations is shown
but the erosion rate is initially highest for the smaller perforations. in Fig. 10.
At constant flow rate in shot perforations (Fig. 9), erosion rate The well was completed with 5~-in. [139.7-mm] casing, and
decreased more slowly at later times (smaller Ap). Erosion and cor- 21 holes 0.28 in. [7.11 mm) in diameter were shot in two zones
responding pressure drop occur by two mechanisms. The fust re- over a 200-ft [60-m) interval at about 8,000 ft [2440 m] deep. De-
quires very little metal loss to change the orifice coefficient sign included stimulation with a 4O-lbm delayed crosslinked
effectively as it becomes more like a nozzle. Second, a slow in- geUl,OOO gal fluid [4800 g/m 3 ) . Treatment was to be performed
crease in total diameter also contributes to a reduction in the pres- down 2¥.-in., 6.5-lbm (73.03-mm, 2.9-kg] tubing. Once pumping
sure drop. Therefore, in these tests, both factors that influence began, maximum wellhead pressure was reached at 12.5 bbllmin
erosion rate (4p and diameter) are changing in a way that reduces [1.99 m 3 /min) . .
the erosion rate. In hindsight, the effect of Il.p on the later-time Field experience bas demonstrated that the friction pressure for
erosion rate could have been quantified by changing the flow rate this gel system can be calculated reliably. Once an instantaneous
to increase the Il.p and to observe the change in erosion rate. But shut-in pressure (!SIP) was obtained in the pad, it was easy to dem-
because the diameter cannot be determined accurately without dis- onstrate that the perforation friction was about 3,600 psi at 12.5
mantling the test fixture, diameter determination was not one of bbUmin [24.8 MPa at 1.99 m 3 /min).
the original test objectives and was not done at that time. During the pad, stages of O.I-lbmlgal [12-kg/m 3 ] 20/40-mesb
sand were added to determine whether perforation erosion could
Sand Concentration. Fig. 9 shows that 6 Ibm sandlgal [720 be induced with a low concentration of sand. There was no change
kg/m 3 ) erodes much more quickly and to a lower total Ap than in the surface pressure from these additions. Note, however, that
does 2.5 Ibm sandlgal [300 kg/m 3 ) under the same conditions. At as significant sand concentrations were pumped, perforation fric-
the end of the test, the perforations were enlarged by an average tion continually decreased throughout the treatment and at the end
of 3.4% in diameter after 29 minutes for the 2.5-1bmlgal of pumping was less than 100 psi at 20 bbUmin "(690 kPa at 3.18
r300-kg/m 3 ] run and 15% in diameter after 21 minutes for the m3 /min).
 TABLE 2-EFFECT OF DESIGN PARAMETERS ON
EXPECTED PERFORATION PRESSURE DROP
WITH WATER AT 1 bbl/mln-perforatlon
TABLE 3-RECOMMENDED
Assumed Expected PERFORATION-FRICTION VALUES
Diameter Pressure Drop
(In.) (psi) Maximum
0.26 0.9 395 p,
0.28 0.6 890 (psi)
0.25 0.6 1,400 Typical fracturing treatment 200
0.22 0.9 1,040 Limited-entry design 1,000
0.22 0.6 2,340 High differential pressure, Iimited-entry design 1,500

During a treatment of this type, there is no way to determine in that these sizes may still be optimistic. Saucier and Lands 12 found
real time what is happening to the formation fracturing treating pres- that some published perforation-penetration-depth data, collected
sure, and it is only after the fmal ISIP is obtained that the perfora- when shaped charges were tested under simulated field conditions,
tion erosion history can be calculated. were optimistic. The example in Table 2 demonstrates the design
A more common example of what is encountered in field opera- situation one might face if designing for a 400-psi [2760-kPa1 Ap
tions is shown in the next examples (Figs. 11 and 12). This well in a treatment. Minor errors in the assumed diameter or Kd cause
was completed with 4 Ih-in. [l14.3-mm] casing. The treatment was major errors in the prediction of perforation friction. Work in the
to be performed down the 41h x2o/.-in. [114.3 x60.33-mm] annu- area of pumping gravel-pack slurries through crossover tools is well
lus. A bottomhole gauge was run down the tubing, and the static documented 13,14 and provides additional insight into the flow of
surface pressure on the tubing was measured during the treatment. slurries through restricted openings.
In this configuration, the total fracture-entry friction can be calcu- Table 3 shows some recommended values for use in designing
lated by the pressure drop observed at the time of shut-in. Early fracturing treatments. These values are not absolute but are rec-
in the treatment, an ISIP taken in the pad demonstrated a perfora- ommended as a starting point on the basis of field and laboratory
tion friction of 1,400 psi at 11 bbVmin [9650 kPa at 1.75 m 3 /min). experience. The limited-entry design value of 1,000 psi [6900 kPa)
Surface pressure declined slowly throughout the treatment. The final is recommended, noting that the K d , and hence perforation fric-
ISIP showed a fracture-entry friction of only 400 psi at 15.2 bbVmin tion pressure, will be modified by sand slurries. It is desirable to
[2760 kPa at 2.42 m 3 /min) . . maintain approximately 300-psi [2070-kPa1 perforation pressure
A Nolte and Smith plot (Fig. 12) constructed from these data throughout the proppant stages to help ensure limited-entry treat-
would have indicated that the well fractured out ofzone. However, ment. The high-differential-pressure limited-entry design in Table
from examination of the early and late ISIP, one can conclude that 3 would involve two formations with a difference in treating pres-
the BHTP actually increased by 700 psi [4830 kPa]; in fact, the sure of up to 700 psi [4830 kPa1. Because perforation friction is
well was on a screenout mode for the last 20 minutes of the modified by sand-slurry erosion of perforations, fracture of for-
treatment. mations having fracture gradients greater than 700 psi [4830 kPa1
Both laboratory and field data have shown that perforation fric- should not be attempted as a single treatment.
tion, whether it exists by design, such as that used in limited-entry Cramer IS studied perforation erosion during massive hydraulic
treatments, or by accident, will be significantly modified in the treat- fracturing in the CodelllNiobrara formations. He listed several sig-
ment by the pumping of sand-laden slurries. This has two major nificant conclusions and recommendations to be used with limited-
implications in hydraulic fracturing treatments. entry techniques.
1. In limited-entry designs, the operator attempts to place fluid Fig. 8 shows that the pressure drop in large perforations decreases
in at least two zones of differing fracturing pressures by control- less rapidly from sand-laden fluid erosion than does the pressure
ling the number of holes in each zone such that the treating pres- drop in small perforations. This has a major implication for limited-
sure is in excess of the fracture gradient for both zones. Assuming entry designs in suggesting that a smaller number of large perfora-
that the proper perforation density is obtained for the pad fluid, tions be shot. The data also suggest that the overall treatment rate
erosion of the perforations by sand slurries will result in most of should be increased during the sand-laden stages to help maintain
the proppant being placed in the lower-pressure zone. Such is the limited-entry control.
case for the treatment described in Fig. 10.
2. Much time and effort were expended in the second field ex- Conclusions
ample in an attempt to obtain accurate BHTP during the treatment.
1. Perforation erosion may mask the true BHP behavior of a wen.
Perforation friction changes during the treatment, however, preclude
Loss of perforation friction may be interpreted as breaking out of
accurate analysis of the data (Fig. 12).
zone in a Nolte and Smith plot and the screenout mode may be in-
determinate.
Implications for Field Operations 2. An increase in sand concentration may not be detected as an
It is common in field operations to break down and to ball out per- increase in pressure as a result of rapid perforation erosion, which
forations before a fracturing treatment to ensure that all perfora- reduces perforation friction.
tions are open. Fluid is then pumped to determine the number of 3. Perforation erosion and changing perforation friction pressure
open perforations. Three calculated or assumed values are needed occur by two mechanisms. The rounding or smoothing of the en-
to calculate the number of perforations taking fluid : (I) tubular fric- trance to a perforation causes an immediate and significant pres-
tion pressure, (2) perforation diameter, and (3) Kd • sure reduction. Also, the perforation diameter increases very slowly,
Lord and McGowen 3 provided field-validated correlations that causing a further drop in pressure vs. time.
can calculate tubular friction pressure accurately. Friction pressure 4 . The erosion rate increases with higher sand concentrations,
for thin fluids depends strongly on pipe roughness, 11 the effects higher differential pressure, and smaller perforations.
of which diminish as the viscosity of the fluid increases. A 5. Because of shock hardening, jet perforations eroded less than
4O-lbmll,OOO-gal [4790-g/m 3 1 HPG gel exhibits the same friction drilled holes.
pressure in rough and smooth pipe and therefore should be used
if knowledge of the number of open perforations is critical. Nomenclature
This study indicates that for design purposes, a discharge coeffi-
cient of approximately 0.60 should be used for new perforations. d = perforation diameter, in. [mm)
The most critical parameter is still perforation diameter, and service- Kd = perforation coefficient
company literature is available to estimate the expected effective np =: number of perforations
perforation size. A general implication from field observations is Pbhl = bottomhole treating pressure, psi [kPa1

Journal of Petroleum Technology, August 1988 1047

 PI = differential pressure across perforations, psi [kPa) 7. Regalbuto, 1.A. and Riggs, R.S.: "Underbalanced Perforation Char-
PIe! = pressure caused by fracture-entry friction, psi [kPa) acteristics as Affected by Differential Pressure," SPEPE (Feb. 1988)
83-88.
Ph = hydrostatic pressure, psi [kPa) 8. Roll, D. et al.: "Effects of Pumping Equipment on Sand-Laden Slur-
P, = pressure caused by fluid f~ction -in tubulars, psi ries," SPEPE (Nov. 1987) 291-96; Trans., AIME, 289.
[kPa) 9. Early, W.L.: "Metals and Their Resistance To Penetration By Shaped
Pw = wellhead pressure, psi [kPa) Charges," MS dissertation, U . of Texas, Arlington (1977) .
Jl.p = differential pressure, psi [kPa) 10. Prasuhn, A.L.: FU1Idatnentals ofFluid Meclwnics, Prentice Hall, En-
glewood Cliffs, Nl (1980) 508-13.
q = total flow rate, bbl/min [m 3 /min) 11. Shah, S.N.: "Correlations Predict Friction Pressures of Fracturing
p = fluid density, Ibm/gal [kg/ml) Gels," Oil cI: Gas J. (Jan. 16, 1984) 92-98.
12. Saucier, R.J. and Lands, J.F. Jr. : "A Laboratory Study of Perfora-
Ackno~edlntent. tions in Stressed Formation Rocles," JPT(Sept. 1978) 1347-53; Trrms.,
AlME, l6S.
We express our appreciation to Jim Giddens for assistance in prepar- 13. Haynes, C.D. and Gray, K.E.: "Sand Particle Transport in Perforated
ing this paper. We also thank Jet ReSearch Center, Arlington, TX, Casing, " paper SPE 4031 presented at the 1972 SPE Annual Meeting,
for technical advice on perforations and Halliburton Services for San Antonio, Oct. 8- 11.
the opportunity to publish this paper. 14. Shryock, S.G., Dunlap, R.G., and Millhone, R.S.: "Preliminary Re-
sults From [Link] Gravel Packing Studies," paper SPE 7123 present-
Reference. ed at the 1978 SPE California Regional Meeting, San Francisco, April
12-14.
I. Nolte, K.G. 8IId Smith, M.B.: ..Interpretation of Fracturing Pressures,"
15. Cramer, D .O.: "Application of Limited-Entry Techniques in Massive
JPT (Sept. 1981)" 1767-75.
Hydraulic Fracturing Treatments," paper SPE 16189 presented at the
2. Conway, M. W. et aI.: "Prediction of Formation Response From Frac-
1987 SPE Production Operations Symposium, Oklahoma City, March
ture Pressure Behavior," paper SPE 14263 presented at the 1985 SPE
8-10.
Annual Technical Conference and Exhibition, Las Vegas, Sept. 22-25.
3. Lord, D.L. and McGowen, 1.M.: "Real-Time Treating Pressure Anal-
ysis Aided by New Correlation," paper SPE 15367 present~ at the
81 Metric Conversion Factor.
1986 SPE Annual Technical Conference and Exhibition, New Orleans, bbl x 1.589873 E-Ol
Oct. 5-8. ft x 3.048* E-Ol m
4. Shah, S.N. and Lee, Y.N.: "Friction Pressures of Proppant-Laden gal x 3.785412 E-03 m3
Hydraulic Fracturing Fluids," SPEPE (Nov. 1986) 437-45. in. x 2.54* E+OO cm
5. Hannah, R.R., Harrington, L.l., and Lance, L.C.: "Real-Time Cal- Ibm x 4.535924 E-Ol kg
culation of Accurate Bottomhole Fracturing Pressure From Surface
Measurements with Measured Pressures as a Base," paper SPE 12062
psi x 6.894 757 E+OO kPa
presented at the 1983 SPE Annual Technical Conference and Exhibi-
tion, San Francisco, Oct. 5-8. 'Conversion factor Is exact. JPT
6. King, G.E., Anderson, A.R., and Bingham, M .D. : "A Field Study
Original SPE manuscript received for review Oct. 5, 1986. Paper accepted for publication
of Underbalance Pressures Necessary To Obtain Clean Perforations Aug. 17, 1987. Revised manuscript received Oct. 23. 1987. Paper (SPE 15474) fIrst~­
Using Tubing-Conveyed Perforating," JPT(June 1986) 662~; Trans., eel 8\ tile 1986 SPE Annual Technical Conference and exhibition held in New Orleans.
AIME,281. Oct. 5-8.