SocietYof petroleumEndneera

SPE 27694

Field Measurement of Hydraulic Fracture Fluids: Case Study

B.M. Davidson, B.F. Saunders, and S.A. Holditch, S.A. Holditch & Assocs. Inc.
SPE Members

copyright 19s4, Scdety of PetroIeum Engmwa Inc.

This paper was prepared for presentation at the 1S94 SPE Permian Basin 011and Gas Recovery Conference held in Midland, Texas, 1S-1S March 1994.

This paper was selected for Pms8ntaticm by an SPE Program Committee following reviaw of information contained in an abatract submitted by the authw(s). Contents of the paper,
as preasntad, have not bean reviewwd by the Society of Petroleum Englneem and are eubjeti to mrrecticm by the authOt’(S).The material, as presented, dons not fracaswlly reflect
any position of the Society of Petroleum Engineers, its offirm, or members Papers [Link] at SPE meetings are subject to PubIicatbn review by Edtorial Committals of the S&MY
of Patmieum Engineer.%Permission to copy is restrkted to an abstraci of not more than SW words. Illustrationsmay not be [Link] abstract shcdd mntain mnspictmus acknmdadgment
of where and by whom the paper Is pres8ntad. Write Librarian, SPE, P.O. SON SSSS36, Richardson, TX 750SS-SSS6, U.S.A. Telex, 1SS245 SPEUT.

Introduction at conditions designed to simulate reservoir tcmpcralures in our

IQC work. We use lhc chemicals and addilivcs that will
We performed intense quality control (IQC) on 54 actually be pumped during the treatment. Using the actual
hydraulic fracture treatments during 1992-93. Our detailed chemicals is a critical part of lQC, and it is the mason that
fluid testing in the field has indicated the complex nature of testing must be performed at the wellsite. prior to the treatment.
fracture fluid systems currently used by the industry. This Blended samples arc coilcctcd during the fracture treatment
paper documents our experience with frach,rrc fluids blended in and tested as the treatmcnl is in progress. The test.. performed
the field and tested under in-situ conditions. Several problems during the trrxatmcot arc compared to the pilot tests to ensure
were encountered while performing intense quality control the fluid pumped down hole is performing as expected. We
operations on these wells, including overcrosslinkcd gel inventory all ariditivcs before. during, and after the treatment to
systems and contaminated additives. Many of these problems verify the fmchrre fluid systcm is blendccl accurately.
became apparent only after evaluating the fracture fluids at
bottomhole temperatures and shear rates. Many fluids thal Often. the SUCCCMof a st imutation treatment is
failed viscosity tests at bottomhole conditions looked to be incorrectly judged by the amount of fluid and proppant Icft over
adequate when tested al surface conditions. at the cnd of the treatment. If the viscous fluid properties of the
fluid pumped during the treatment are measurably different
The importance of quality control in stimulation than those used to design lhc treatment, then it is likely that the
operations has been investigated and its importance recognii!x!d propped fracture geometry will also diflcr. For example. a low
by the industry.’-4 lntcnse quality control is perfonncd on viscosity fluid system with poor proppant transport properties
location using equipment carried in a rheology van. The will allow proppant to settle into the bottom of the fracture.
rheology van contains two Farm Model 50 viscomcters that wc reducing the amount of proppant placed across the pay interval,
...- Ifi “.-”e*
-.-],,~f~ the frac~,lro ni Iw-mnnrhole lern~ralurcs
u,-- ,“ ...W..”” s.., ..fll]id~
----- -. . . ... ... ... .. .. potentially reducing fracture length and fracture conductivity.
and shear ralcs, and a complelc wet chemistry lab. Analyses of A fracture fluid that is too viscous can create a wide fracture,
fracture fluids with the rheology van has heco described in a thus reducing the propped fracture length. If the fluid is too
previous pape~ and GRl publication.” viscous, the amount of time for the fracture to “clean up” will
increase, especially if the fluid does not break properly, Two
Intense quality control is not a ncw concept. IQC is SPE publications clearly show the problems onc can have if the
routinely used in the evaluation of cement slurries. To test iracmrc fiuid does not break.7’X rwr - LaaII
‘--- L?iL%%iup . . . “:.....,
31EI111 ”*l,.*1.,
IVtUILIY
recommended mixtures of ccmcnt and additives. pilot tests are rcducc profit from the fracture treatment.
always performed on cement slurries prior to pumping to tesl
recommended mixtures of cement and additives. Results of the DUC to the complex nature of fracture fluids, it has
pilot and blend tests indicate if a particular ccmcnt mixture will been WCIIdocumented that lhe apparent viscosity can be a
perform M designed, Similarly. wc lest fraciurc fluid systerms strong function of the testing procedures. *-11 The fracture

References at end of paper,

631
2 FIELD MEASUREMENT OF HYDRAULIC FRACTURE FLUIDS - CASE STUDY SPE ~7694

fluids discussed in this paper were tested under specific

conditions that may vary from tests performed in other
I
laboratories. ‘We“Xiieve our tests provide a reaiisiic projcciimi
of fluid performance in terms of crosslink time, long term
stability and break time. These arc the mosl important
elements 10 consider when pumping a hydraulic fkacture
treatmcnt.
Summarv of Datasct —

Table 1 presents a summary of the data wc collected

and evaluated for 54 different fracture trea(mcnts. We
-.-l?..
pclIurmrxl
..~ UVG1
-..-.. .JLv
<?II l,,ulvluual
:..~:..:[Link]-w, MAoI .JV
1 am,,,,., --,
<II .JiuPnuitv todu nn
v..,””.7.., .V... . “..
eleven different types of fracture fluid systems. Test
125-130190-255255-232323-390
temperatures simulating in-situ conditions ranged from 214°F
Injeetlon Time (mIn)
to 355”F, The distribution of reservoir temperatures are
illustrated in Fig. 1. For the average treatment, approximately
300,000 gallons of fluid was pumped placing an average of
Fig. 2- Pumping time of treatments monitored.
over 1,000.000 pounds of proppant in the formation.
Treatment pumping times ranged ffom 60 minutes to 525
minutes
..-...—..— with an average pump time of 283 minutes. Fig. 2 Material Inventorv
shows the distribution of time required to pump the treatments
that are included in this paper. These treatments were Fracture fluid systems are a compicx bicnd of many
performed in several geographic locations, including south and chemical additives. .somctimcs with w many as eight different
east Tcxa.., Alabama, and Wyoming. additives. Additionally, with the advent of gel concentrates, it
is becoming mocc common to mix the polymer and chemical
16 I 1 additives “on the fly”. It is critical for all of [Link] additives to
be blended in lhc proper ratio for the optimum fracture
treatment to be pumped. To manage [Link] complex mixing
operations. service companies have implcmcntcd advanced
computer controlled metering and monitoring systems.
However, even with the most sophisticated equipment. there
appears to be diffmmces between design amounts and the
actual concentration of polymer, crosslinkcr. buffer. and
breaker that is pumped. Table 1 and Figures 3, 6.9 and 11
show the percent different bctwccn the actual additive volumes
used on the treatments and the design volumes. The problem
becomes compounded when the amount of certain additives,
including crosslinker. temperature stabiliiw. and breakers are
“ 214-235235-255255-275275-2s52S5-315315-235335-355
dcsignd-to be based on the polymer loading.
Ternpereture (F)
Fig. 1- Temperature of treatments monitored.
Table 1
Designed vs. Actual Differences
Material Inventory Baw Gel Viscosity
Pumping Pad Proppant
Temp Time Polymer Crosslinkcr Buffer Breaker Stage Stage
(“F) (rnins) (’%) (%) (%) (%) (%) (%)
Average 270 283 -1.o I .0 2. I 11.4 1.6 -2.6
Minimum 214 55 -26. I -39.5 -68.3 -84.7 -26.9 -29.2
Maximum 355 525 27.4 44.8 131.8 275 28.1 46
Absolute Average 6.9 11.4 22.4 26.7 10.0 11.7
Total Treatments 54
Polymer Types Guar. HPG, CMHPG
Crosslinkcrs Borate, Titanium. Zirconium

632
SPE 27694 B. M. DAVIDSON. B. F. SAUNDERS and S. A. HOLDITCH ‘3

We should warn against relying too heavily on

computerized monitoring equipment. As the previous 60%
mentioned figures have show. much of which the data were ~
gathered on treatments with the computer controlled o Absoluts
AvwaaoOlttomnco
= 10.C%
equipment, physical inventories of the additives must also be
performed. not only before and after the fracture treatment, but
during the treatment 10 ensure the additives are correctly
blended in the proper ratios. The prc-fmc and post-frac
inventories may indicate the appropriate volume ww used, but
way not mom! for variances during the treatment that. at
times, are quite significant. III most cases, however. the
sophisticated monitoring systems will help minimim these
variances.
~ .*,O ~
Mixin~ the Polvmer
Treatments
As can be seen in Fig. 3, the average differences
between what was supposed to be and what was actually Fig. 4- Base Fluid Vkosity Difference, Pad
pumped ranged from 6.9 to 11.4 pcrccnt. The average
difference may not sound significant but, for some addilivcs,
these differences can significantly affect fluid performance. _ 40% I
Aboluts Awrago0wuwlo9=[Link]

4@%
g *OhJtU AwragDOifhNI’ICO
= 6.S%
. 3w0

a -30% I
I
irwtrnmt$i

-30% I Fig. 5- Base Fluid Viscosity Difference, Proppant Stage

Treatments
Fig. 4 is iit~ base gd Vkxiby’ ineaswed ear~yt in the
Fig. 3- Polymer inventory Difference Actuahllesign !rca!mcnt during (hc pad stage with a Model 35 viscometcr.
Fig. 5 is the base gcl viscosity memured later in the treatment,
Monitoring base fluid viscosity during the fracture iypically at the start of the proppant stage. The base fluid
~fi$cti+~i~heti!~ ~ -m”;,m-.l
,,,”,,,,”,-”
mwwsnntlv Amino the treatment.
W“,,..M.....J “... ...= ...- .. . . ..... .. .
treatment is required to evaluate the proper polymer bicnding.
Figs. 4 and 5 indicate the diffcrcncc between the actual The baw gc} viscosity will atTcct [Link] fluid performance
measured viscosity of the fluid compared to the service which. in turn. dctcrmincs proppant transport capabilities.
company published viscosity values. The fluid samples were Testing fluid with a Model 35 viscometer should be performed
tested after sut%cient time in the work tanks to achieve full to ensure the proper calibration of in-line viscometers used to
hydration. continuously monitor base gcl viscosity. Testing the fluid
viscosity prior to the treatment will also help dctcrrnine
hydration time for the polymer. Proper work tank volume can
then be chcckcd ba@ on hydration times seen in the baw fluid
viscosity tests.

633
 .>

4 FIELD MEASUREMENT OF HYDRAULIC FRACTURE FLUIDS - CASE STUDY SPE 27694

Mixin~ The Crosslinker

The cross] inkcr concentration grcatl y influences

bottom hole viscosity of the fracture fluid. t30th
overcrosslinking and undercrosslinking of the fraclurc fluid
system can result in a lower viscosity fluid which will
ultimately be detrimental to the proppant transport properties of
the fluid. As illustrated in Fig. 6. it is dilTlculi to accumtcly
meter crosslinkcr during a treatment and errors do occur.

~ 60?/0~
AverageOlflomnce
ALwol@a = 11.4%

Time (rein)
b .1
Fig. 7- Crosslinker Concentration, Standard
Concentration, and zs~. Lower Concentration
(286°F 40 lb/m CMHPG With Zirconium Crosslinker)

This probicm was not unique to a single geographical

area or fluid system, Adjusting the crosslinkcr concentration
ww required on 22 of the 54 fracture treatments. This included
lhrcc fracture fluid systems in two geographic locations. This
Treatments
dra..tic viscosi(y ditTcmncc sem in the test prompted further
investigateion. Therefore, a series of tcsls were conducted 10
Fig. 6- Crosslinker Inventory Difference, Actual/Design cvaiuate the effect of crossl inker concentration on the viscosity
protilc. Fig. 8 illustrates the results of our tcsLs.
Fig. 7 is a comparison of fluid viscosity using the
originally recommended scrvicc company crosslinker
concentration and a zs~. reduction in crosslinkcr
concentration. The standard service company crossiinker
concentration on this treatment resulted in an ovcrcrosslinkcd
fracture fluid. The Farm 50 test of the ovcrcrosslinkcd sample
indicated a very high initial viscosily. followed by a low
viscosity throughout the remainder of the test when it was
subjected to bottomhole temperatures. The fluid sample wa!s
very lumpy and contained significant amounts of free water
when removed from the Farm Model 50 viscomctcr. indicating
a severely overcrossl inked fluid. This is a typical example of a
fluid that passed “conventional” surface crosslinked test!! at
surface conditions. [f test.. were run using conventional
methods, no problems would have been noticed. However. by
testing the fluid at botlom hole tcmpcraturcs using a Farm 50
viscometer, the actual performance of the fluid in the formation Time (mIn)
is evident.
Fig. 8- Influence of Crosslinker Concentration
(260”F, 35 lbiNl HPG With Zirconium Crosslinker)

AS i!!~s!~!~d in Fig, fl: a s~blc fluid is obtained using

a crosslinker conccnWation of 0.4 gal/1000 gallons. if more or
less than 0.4 gal/1000 gals of crosslinker is used, the fluid is
less stable. Concentrations Icss than 0.4 gal/i 000 gals resulted
in lower viscosity fluids. but they were still relatively stable.
However, overcrosslinking the fracture fluid, using 0.5
gal/1000 gals. rcsuits in a drzustically lower viscosity, Atler the

634
SPE 27694 B. M. DAVIDSON, B. F. SAUNDERS and S. A. HOLDITCH

test, a visual inspection of lhc overcrosslinkcd sample indicated

a very nonhomogeneous fluid that is lumpy wilh excess free
water. To put this crosslinkcr concentration diffcrcncc in
perspective, on a treatment pumped at 15 13PM. the 0.1
gal/1 000 gals difference amounts to just 0.063 gal/rein. In
summary, it appears that ovcrcrosdinking of the fluid may be
occurring more frequently than expected, either by metering
problems or, in some cases, by using the actual recommended
concentrations.
Mixing The Buffers

Buffering agents arc required in most fmclurc fluid

systems. Buffers typically control gel hydration, crossiinking, “o 10 20 30 40 50
and overall fluid stability. Buffer inventories determined before mm(*)
and after treatments indicalc substantial differences between Fig. 10- Effect of Buffer Concentration on Fracture Fluid
the designed and actual volumes for the 54 trcatmcnt!s we Viscosity (2SO”F, 3S lb/M CMHPG
monitored. However, many service companies make changes to With Zirconium Crosslinker)
designed buffer concentrations during the treatment. The
amount of buffer required can vary depending upon the pH of Fig. 10 is a compiirison of a fracture fluid with two
the base fluid. The final pH of the gel system will be affected by different butTcr concentrations. The final pi-l of these fluids
base mixing water, type of butTcr and ik+ concentration. mcwurcd 5.2 and 5.8, a difference of only 0.6 pH units. This
Therefore, the pH should be monitored continuous y to ensure illusktates how bufft!ring agents can affect fluid peflorrnance.
that proper levels arc maintained. The actual volume of buffer As can be seen in Fig. 10, the viscosity profiles are drastically
used during the treatment will commonly vary from the design different. The lower pH system. 5.2, results in a higher initial
volume by a small amount to achieve the proper PH. Prc-job viscosity but it decreases rapidly. The higher pH fluid, 5.8,
testing is critical to ensure that adequate butTcr inventory is resulted in a much more stable viscosity profile. The data in
available on location to meet gel system rcquircrncnls. Figure 10 arc rmt meant to su,ggesl that ci(hcr of these fluids is
better than the other. Specific treatment requirements should
dictate which viscosity profile is preferable. However. Figure
10 does illustrate lhc effect of pH on fluid viscosity.
MixinrzThe Breakers

The usc of breakers has become more widespread with

the advent of encapsulated breakerx. New research ha$
illustrated the detrimental cffecLs of unbroken gel on well
pcrforrnancc.7”” Ba$cd on inventory checks before and after
trctatmcnts, the average difference between the actual and
designed breaker volumes is 11.4 pxcenl and the absolute
value of the perccnl difference is 26.7 pcrccnt, On nine of the
Ab~hltO Avu8g0 OiftWOIICO = 22.4% fracture treatments monitored, no breaker was used. Typically,
-loo% when breaker was used. a subslantiall y larger volume was
Treatments pumped than wa.. originally designed. However, service
company field personnel .somclimes recommend an increase in
Fig. 9- Buffer inventory Difference, Actual/Design
lhc amount of encapsulated breaker during the treatment if the
treatment is proceeding properly.

635
 ..

6 FIELD MEASUREMENT OF HYDRAULIC FRACTURE FLUIDS - CASE STUDY SPE 27694

[Link] on a previous treatment that was pumped to completion.

However. the only sure way to obtain the proper breaker
Absolms Avsmge OHfersnce = 26.7%
g t conccrrlralion is to lest each fluid w ilh the actual chemicals on
location. Every test should be conducted at bolmmhoie

h’-
temperature. In many ca!scs. these tests can be done at the local
,scrvicc points of the fracturing companies. Such tcsLswill help
increase your chances of gelling the proper fluid break.
d Ncw brcxakcr technology is dcspcratcly needed for
fracture fluids. More accumlc metering techniques also need to
be developed. With current technology. it is difflcull to
accurate] y add the low concentrations (0.25 - 1.0 lb/1000
gallons) ofbreakers that arc usually rccommcndcd.
-200% I
Treetmsnts
Fig. 11- Breaker Inventory Difference, Design/Actual Farm Model SOViscositv Tests

The use of encapsulated breakers has become more Routine testing of fracture fluids for crosslink time.
common in high tcmpcraturc formations. The benefits of such m a test to determine vortex closure time. only indicates
breakers in well performance has been documented.’z” ‘“; that there is crosslinkcr in the sample that was tested. To
However, breaker dcsigrus are not cmwistcnt among scrvicc dctcrminc the performance of the fmcturc fluid in the fracture,
companies. Some service companies recommend a very the fluid must be tested at bottomhole temperature. The
aggressive breaker schedule, while others recommend a very industry has invested significant resources to develop a better
conservative breaker schedule. The effect of breaker on fluid understanding of fracture fluids at bottomhole [Link]-19
performance can be seen in Fig. 12. The initial viscosities are
very similar but, as the fluid is exposed to temperature, the To perform intense quality control. wc [Link] a Fam
... !,-, -. A–. – –J-AL. Mrvl14
,----- <fl viwnmetm
g“ . .... .... .. . led
.. ... in
... fraclIIre
.. . . .. . . fiilid
----- nn
--- incdnn.
---------- TO
encapsulalea rxcaiccr ocgm.. to aegraac me vLswsity.
ensure that our test procedures arc precise, wc have developed a
methodology that we consistcntl y use in all fluid testing. Test
~ 2,000 i rcpcatabilit y increases our confidence that fluid performance is
$! predictable. Figs. 13 and 14 are typical of the consistent test
q
~ 1,500
OJ I results we strive for on individual treatments. When consistent
results arc achieved. wc arc confident that any conclusions
bawd on Farm Model 50 viscosity testing. reasonably represent
the fluid behavior.
,000

t **: ~
Test 1 Teet 2 Test 3
—--- —

:~
o 20 40 60 60 100 120 140 160
Time (mIn)

Fig. 12- Viscosity of a Fracture Fluid Wtih and Without

Encapsulated Breaker
(31O”F, 60 lb/M HPG With Zirconium Crosslinker)
, , 1 , , t ,

25 50 75 100 125 150 175 2 0

As can be seen in Fig. 12. the concentration of breaker
mtne (mtn)
during this trealment was too high for the treatment pumping
time of 180 minutes. Based on current pub] ished breaker data, Fig. 13- Comparison of Fluid
the concentration required to completely break lhe fluid and (2S6”F, 30 lb~-CMHPG With Zirconium Crosslinker)
maintain adequate proppant transport properties during the
treatment is difficult to determine.14 Recommended
concentmtions are o!lcn baaed on field experience. The breaker
conccntnation is typically ba..ed on the maximum concentration
636
SPE 27694 B. M. DAVIDSON, l% F. SAUNDERS and S. A. HOLDITCH *7

.6,000
Fig. 16 indicates the quality control problems that can
bc cncuuntcrcd with encapsulated breakers. The viscosity
profile of a fluid without encapsulated breaker is compared to
the viscosity with twu ditTcrcnt types of encapsulated breakers.
The low viscosity of the fluid tested with encapsulated breaker
A was the rcsul~ of a product that. after further investigation,
did UO1appear 10 bc encapsulated. The viscosity profile of the
fluid with encapsulated breaker B rcsulls in a viscosity profile
that is gcncraily expected when encapsulated breaker is added.
51,000

I n
“o 20 40 60 80 100
Time (mIn) ?9nrxli- 1— WkhoutBro8kar

Fig. 14- Comparison of Fluid

~al Q°F
\- =., -* @ MM CMEgPG With Ztrconium Crosslinkcr) <

Contaminated Chemicals 1,000 t ~/ EnwPsuloIui ~ I

\ # B,”k~ B I

Eveo if the chemical additives arc blended in the

proper ratios, problems can occur if the additives are
contaminated. Fig. 15 shows the effect of a buffer
contaminated with a borate crosslinkcr on Ihe apparent “o 10 20 30 40 50 60
..:,.,.. _,:,., The dz+- ;*.K;,,
. le. Is illuct~tc
..... .. ItW
.. - !Mpor!~ncc 0[ using
vlsLm>lly. ,0 ,1, .- ... Ttme(mtn)
the actual chemicals on location in all testing. This buffer
sample was collcctcd from a storage tank that had previously
Fig. 16- Quality of Encapsulated Breaker
contained a borate crosslinkcr. If additive samples had bum
(256”F, 40 lb/M CMHPG With Zirconium Crosslinker)
obtained at the service companies’ facilities, this contamination
problem would not have been detected. The contaminated
sample passedsurfiicc crowdink tests. but when testing the fluid
Viscosifi Comparison Measured v. Published Data
at bottomhole tcmpcraturcs. the effect of the contamiwdtcd
buffer on the down hole viscosity of the tkacturc fluid was
Figs. 17 and I R present a comparison between the
obvious. More important 1y, this problem was corrcctcd before
scrvicc company published viscosify data compared. to the
pumping the treatment. To find these problems. testing in the
viscosity mcawcd on location. As can be seen. viscosity
field with location chemicals is mandatory. Also, tcsLs at
mcuurcd in field operations can be siguiticanll y different than
bottomhole temperatures are necessary 10 dctcrrrrinc if the
the pub[ishcd values. Industry ha.. recognized that some of
additive on location will work properly once in the fracture.
thc,sc diffcrcnccs in [Link] viscosity may be the result of
diffcrcnl bob/cup configurations andlor shear rates used 10 test
the fluids.’”t‘ However. wc believe our tests provide a realistic
i! projection of fluid performance. Our tests arc ~formed with
‘Q 2,000
k
i. Uncontaminmd Bulfu an extended B-2 bob at a shear rate of 37.5 SCC- after an initial
“\< high shear period. TCSLS arc always performed at bottomhole
a
b. tcmpcmture.
~l,500 - s.
●.,
b%
●.-
[l,m *
-..
----
s
- II

c o! # , , , I
o 102030405060 70
mme (mtn)

Fig. 15- Contaminated Buffer

!g0SOF9 g MM HPG With Zkconium Crosslinker)
637
 FIELD MEASUREMENT OF HYDRAULIC FRACTURE FLUIDS - CASE STUDY SPE 27694

500
i
4 or ‘ , 1 # I

o 40 80 120 160 200

lima (mIn)
Fig. 17- Comparison of Measured and Published Viscosity,
Fluid A (273”F, 50 lb/TMC-MHPG ‘with
[Link]
-.. . . . . . . . . . . (%nadinkm-\
-. “-.. . . ..-. ,
From-Different Service Companies

I
4I GNiCO Co~811y A
I
t / Wlbm)

k=
1%
-... . ...”----------------
, winmnw vmcosmy
mu P w r
L\ -, aowicocompany
(451bM)

mme (rein)
r---
1
40
,
80 120
,
160
>

1
200
lime (mh)
Fig. 18- Comparison of Measured and Published Viscosity,
Fluid B (SO lb/M CMHPG With Zirconium Crosslinker) Fig. 20- Viscosity of 40 lb/M CMHPG + Zirconium
@ 255°F From Different Service Companies
It is obvious that the dilTerences between acmal
viscosity and published data can be significant. These dard -2,000 ~ I
illustrate exactly why design engineers need to be careful when
choosing a fluid, and predicting fracture geometry, and
$!
q ‘#j Suvica
CompanyD
proppant tran!![Link] on the published viscosity. R 1,500 -:
a I
/’
“’~-~%; .-, *b%
%%cositvCommwison - Service Comoanies
~ ;

pm -1
Figs. 19, 20. and 21 compare similar fracture fluids
from diflerent service companies. Fig. 19 is a comparison of an 8 ‘i
HPG with a zirconium crosslinkcr, while Figs. 20 and 21
500 ;
compare a CMHPG fluid with a zirconium crosslinkcr. i -j

to ‘. , , t , ,
0 30 6080 120 150 180
mm (mIn)

Fig. 2 I - Viscosity of 45 lb/M CM HPG + Ztrconium

(i? 270”F From Different Service Companies

638
.,

SPE27694 B. M, DAVIDSON, B. F. SAUNDERS and S. A. HOLDITCH 9’

As can be seen. not all fracture fluid systems result in Recommendations

the same viscosity at bottomhole Temperature. These lcsts
indicate the variability in viscosity between service companies 1, Crosslinkcr concentrations arc typically very small. 0.3
--- 1 1-.. 4:..-.. S-,1 ,Aa A..”;,... a..,,:,. maw g~i/i~Qo ie ! .5 g~i!i
()()().
DiilJiing
Crossilnkem tO increase
for simfiar gei systems allu I(XdUllI&, @IIU LIIU UWl&I UIISIIIVUI
~h~ui~ _=----
aRain be aware of these differences when designing a required concentrations may be ctXectivc in improving
fracture treatment. mclcring accuracy.

Conclusions 2. Ovcrcrosslinking a fracture fluid is common and can be

more detrimental than undcrcrosslinking. Therefore,
1. Modern fracture fluid systems are ve~ complex with
crosslinker conccntraiion should be adjusted according to
viscous properties that are not fully understood.
fluid performance seen in pilot tests.

2. Variations in chemical additive concentrations occur in the

3. Standard testing procedures need to be established to
field. and can have significant effect on the viscosity
facilitate the comparison of fracture fluids tested in the
profile of hydraulic fkacture fluids at in-situ tcmpcramres field laboratory.
research laboratory and ihc
that may not be apparent unless intense quality control is
us~ io ie~i ihe fiui~s.
4. Operators should assist the service companies in achieving
the optimal fluid by providing clean frac tanks and water.
3. Due to the multitude of additives currently being injcctcd
They must also clearly communicate performance
during “on-the-fly” fracture treatments. strict inventory
rcquircmcnts cxpeclcd during fracture treatment
practices must be utilized before, during and after the
operations.
treatment.
Acknowledt!ments
4. Surface tests of crosslink time. such m vortex closure. are
unreliable indicators of fluid performance at bottomhole
We gratefully acknowledge .M Asbill. Mike George.
conditions. Typical surface tests only indicalc that the
and 13rown Howard. whose work in performing the sometimes
fluid sample being tested contains some crosslinking agent.
thanklyw job of intense quality control helped make this paper
possible. We also appreciate the management of S. A. Holditch
5. Pumping a treatment to completion does not guaranlcc
& [Link]. Inc. for lhcir support during the preparation of
succes..ful well stimulation. Low viscosity fluitks result in
this paper.
poor proppant transport properties which reduce fracture
length and conductivity in the pay zone. High viscosiiy Rcfcrcnces
fluids can result in a plugged fracture. increasing the time
required for the fracture to clean up. 1. Hall, B. E. and Larkin. S. D.: “On-Silt Quality Control
of Fracture Treatments.” Journal (?/” Petroleum
6. Significant impmvcmcnt in metering control and mixing Technology (May 1989).
procedures are required to minimize problems that occur
when trying to mix these complex fluid systems. 2. Decs. J. M. and Coultcr. G. R.: “Fracture Treatment
Quality Control: Results of a Three-Year Study.” paper
7. Breaker testing. prior to performing the fracture treatment. SPE 15512 presented at the i 988 Rocky Mountain
is neccsaary to utilize encapsulated breakers effectively. Regional Meeting, Cas~r. WY. May 11-13.
Breakers can have a significant impact on fracture fluid
performance. 3. Ely. J. W.. Arnold. W. T.. and Holditch. S. A.: “New
Techniques and Quality Control Find Success in
8. Fluid viscosity is easily controlled by adjusting crosslinkcr Enhanced Productivity and Minimizing Proppant
concentration. However, crosslinker concentration affects Flowback: paper SPE 20708 presented at the 65th
both the current viscosity level and its performance over Annual Technical Conference and Exhibition of the
time. Pre-treatment testing is essential to be snue the proper SOciely of Petroleum Engineers, Ncw CMeans, LA, Sept.
amount of crosslinker is used during a treatment. 23-26.1990.

9. Our testing indicates there is a significant variability in 4. Nagcl, N. B., Arnold. W. T. [Link].M. W.. Scott, S.
fluid performance on a day to day basis. This variability L..and Thrasher. T. S.: “An Integrated Team Approach
may be the result of fluid contaminants (in the base mixing for Improving Company-Wide Stimulation Design and
water or chemical additives) inconsistent field blending. or Qualiiy Control; paper SPE 26142 presented at the
a combination of both, Intense quality control is required to 1993 Gm Technology Symposium. Caigary. Aiiberta.
identify and minimize these variances. Canada. June 28-30.

639
 ,#

10 FIELD MEASUREMENT OF HYDRAULIC FRKT-uRE FiiUii3S - CASE STUDY sp~ 27694

5. Ely. J. W., Haskett. S. E., and Holditch, S. A.: “Field

Experience With the GRI Rheology Unit.” paper SPE 15. Royce. T. N.. Beck. L. M.. and Richards, A. R.:
17715 presented at the 1988 Gas TcchnoloLg “Rhcological Chsractcrislics of .Adjustable Crosslinked
Symposium. Dallas, TX. June 13-15. Fracturing Fluids.” paper SPE 13178 presented at the
591h Annual Technical Conference and Exhibition,
6. S. A. Holditch & Associates, Inc. (eds.): “Anaiyses of” HO1&!O~.TX* S~P!, i&i 9? 19M.

the PhysicsJ Properties and Viscous Behavior of

Fracturing Fluids - Laboratory Results and Field 16. Worlow, D. W. and Holditch. S. A.: “Rhcological
Applications; Report No. GRI/90-0315 prepared under McasuremcnLs of a Crosslinkcd Fracture Fluid Under
GRI Contract 5088-211-1734. December 1990. Conditions Expcctcd During a Fracture Treatment.”
presented at the SPE Joint Rocky Mountain
7, Soliman, M, Y. and Hunt, J. L.: “Effecl of Frsclure Regional/Low Permeability Reservoir Symposium and
G..h;h;tim~ I)mwt=r (T). March 6-8.1989.
Fluid And Its Cleanup on Well Performance.” paper S?E L,Atl, ”t,l”s . . v.. .-., --- .

14514 presented at lhe 1985 Easlcrn Regional Meeling,

Morgantown, WV, Nov. 6-8. 17. Cameron. J. R.. Gardner. D. C., and Veatch. R. W.:
“Insights on Theological Testing and Flow Behavior of
8. Voneiff, G. W., Robinson. B. M.. arid i+ohiiich. S. A.: ~iaycd Crmslinkcd FrSCUMC Fluids.” SPE Pmduciiorr
“The Effects of Unbroken Fracture Fluid on GSS Well Eugiweriug, May 1990.
Petiorrnance.” paper SPE 26664 [Link] at the 68th
Annual Technical Confcrcncc and Exhibition, Houston. I 8. Kraxr. J.. Prud’homme. R. K.. Norman. L. R., and
TX. Oct. 3-6, 1993. Sandy. J. M.: “Characteristics of Metal-Polymer
Interactions in Fracturing Fluid Systems.” paper SPE
9, Gdanski, R. D., Harris, P. C.. Lord. D. L., Rcidenbach. 16914 presenlcd at the 62nd Annual Technical
B., and Shah. S. N.: ‘Anaiysis of Farm ‘M(JJei50 foi CQnf~rCnCCand Exhibition. Dallas. TX. Sept. 27-30.
Fracturing Fluid Rheology.” paper SPE 21647 [Link] 1987.
at the Production Operations Symposium, Oklahoma
City. OK, April 7-9.1991. I 9. Fan. Yong and Holditch, S. A.: “Viscous and
Degradation Properties of Crosslinkcd HPG GCIS Some
! o. . . ..x ... ... ... . R: K., Ellis, S., Constien, V. G., and
%d’hnmme. Laboratory Observations.” presented at the 1993 Eastern
Knoll, S.: “Rcproduciblc Theological MeasuremcnLs on Regional Conference and ExliibiiiOti. Piitsb@, PA,

Crosslinked Fracturing Fluids: paper SPE 18210 NOV. 2-4.

presented at the 63rd Annual Technical Confcrencc and
Exhibition. Houston, TX, Oct. 2-5.1988.

11. Fan, Yon& “Measurementsof Viscous and 13cgradation

Properties of Fracturing Fluids at in-Siiu Conditions of
Shear Rate and Temperature Using Farm 50C
Viscometer/ Ph.D. Dissertation. Texas A&M
University, Dec. 1992.

12. Brandon. H. D. and Pulsinelli, R. J.: “Evaluation of the

Breaker Concentration Required to Improve the
pemlea~~ity of Proppant Packs Damaged by Hydraulic
Fracturing Fluids: paper SPE 19402 presented at the
Formation Damage Control Symposium. Lafayette. LA,
Feb. 22-23.

13. Elbel. J.. Gulvis. J., King. M. T.. and Maniere. J.:
“Increased Breaker Concentration in Fracturing Fluids
Results in Improved Gas Well Pcrfmmancc.” paper SPE
21716 presented at the Production Operations
Symposium. Oklahoma City, OK. April 7-9.1991.

1’4. Craig, D. P.: “The Degradation of HydroxypmpylGuar

Fracturing Fluids by Enzyme. Oxidative, and Catalpcd
Oxidative Breakers.” Master’s Thesis, Texas A&i%
University, December 1992.
640