You are on page 1of 14

I Societyof PetroleumEngineers I

SPE 24599

A New Approach to Preventing Lost Circulation While Drilling


Giin-Fa Fuh, Nobuo Morita, P.A. Boyd, a n d S.J. McGoffin, Conoco Inc.
SPE Members

Copyright 1992, Society of Petroleum Engineers Inc.

This paper was prepared f6r presentation at the 67th Annual Technical Conference and Exhibition of the Society of Petroleum Engineers held in Washington, DC, October 4-7. 1992.

This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author@).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 positionof the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publicationreview by Editorial Committees of the Society
of PetroleumEngineers. Permissionto copy is restricted to an abstract of not more than 300words. Illustrations may not be copied. The a b s t r a c t s
of where and by whom the paper is presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836 U S A . Telex, 730969 SPEDAL.

ABSTRACT INTRODUCTION

This paper presents a new concept, theoretical formulations, The idea of lost circulation prevention is not only achievable,
and field test results of the loss prevention material (LPM*) but also practically feasible. Needless to say, lost circulation
developed for use in many possible applications such as high- prevention is far more cost effective than any remedial action
angle and horizontal well drilling, drilling through severely taken after the occurrence of lost circulation. Especially with
depleted formations (to optimize casing requirements), drilling oil muds, losses can be very expensive, unsafe (underground
through highly tectonically active areas (i-e., regions of over- blowout), and time consuming. It is under this premise that
thrust faults and salt dome structures) where the in situ rock this new LPM (loss prevention material) product was devel-
stresses are high and directionally unequal. Both open-hole oped for use in "lost-circulation-prone"zones or in high-angle
microfractures and 5000-ft open-hole leak off tests using LPM hole drilling when a higher mud weight is required to counter
were successfully conducted in our wells in Newkirk, Oklaho- potential borehole collapse, stuck pipe, tight hole, or other hole
ma, and Ventura, California. LPM is a specially selected instability problems. Adding LPM to the drilling fluid will pro-
(material strength and specific gravity) and narrowly sized vide the extra margin of borehole protection for mud weights
granular material which, after being fully mixed in oil or water used and will reduce the degree of uncertainty involved with
based mud at an optimum concentration, can provide good any type of work or operation concerning natural geologic for-
protection to formations against lost circulation while drilling. mations.

Higher formation fracture resistance resulting from the use of Based on our laboratory test results using large Berea sand-
LPM in drilling fluids can reduce or prevent the occurrence of stone blocks, an optimum LPM concentration can be found
lost circulation while drilling. This paper will show that a and it varies with mud densities. According to our experimen-
remarkable increase of 8.0 ppg in botehole breakdown pres- tal investigation and analytical verification: formation fracture
sure was achieved in one of our tests in the Newkirk well. An initiation and propagation during well drilling can be prevented
increase of fracture propagation pressure in the range of 3.0 or inhibited by including an effective concentration of LPM (a
to 6.0 ppg was recorded in the field tests. These results are in specially sized granular material) in the drilling fluid. Our
general agreement with our laboratory findings also discussed laboratory findings had determined a specific size range of
in this paper. LPM particles that were most effective in inhibiting the

References and illustrations at end of paper.

*The subject of LPM technology has been filed and pending


a U.S. patent application by G. F. Fuh and N. Morita of
Conoco Inc. entitled "Method for Inhibiting the Initiation and
Propagation of Formation Fractures While Drilling!"
G. F. FUH, N. MORITA,
SPE 24599 P. A BOYD, AND S. J. McGOFFIN

a bridging material can seal a fracture tip with We, where We small. Even if it fails to induce screen-out for a short
(Figure 1) is the critical width which can sustain fracture fracture, it can still stabilize fracture propagation with a
pressure without breaking. Assume that the pressure in the higher fracture resistance than a standard drilling fluid.
fracture is Pf from fracture inlet to the bridging point, and Po
from the bridging point to fracture tip, where P, is the (c) It serves as a plugging material for large pores and natu-
formation pore pressure as shown in Figure 1. Then, the par- ral fractures.
ticulate screen does not break until the fracture pressure
reaches However, the above analysis also shows the LPM material has
the following limitations:

(a) A certain degree of filtrate loss is required to cause


particulate screen-out. Hence, it cannot be applied to an
where impermeable shale or limestone. However, the imperme-
able shales or limestones are normally strong or have high
in situ stress. Therefore, unless they have natural frac-
tures, lost circulation may not be induced. Note that
many so-called shale is actually mudstone or sandy silt-
stone, hence, it can form mud cake if fracture is induced.

Equation 2 shows that the fracture propagation resistance, (b) The fracture propagation resistance is low for low
because of LPM screen-out, increases with elastic modulus of Young's modulus zones or high pore pressure zone.
formation and net in situ stress, but reduces with fracture size. However, fracture may not be induced in the low Young's
Figure 2 shows the fracture pseudo-toughness calculated using modulus zone or high pore pressure zone since these
Equation 3. Normally, the range of a, - Po is several hundreds zones may have high in situ stress due to high pore pres-
to several thousands psi; the range of W, is .05" to .2"; and sure, or they may easily cause screen-out due to high
El(1-u2) is .5x106 to 5x10'. Then, the range of fracture pseudo- formation permeability.
toughness may become 5000 to 30000 psi%". Substitutingthis
value into Equation 2 shows that the fracture resistance gained (c) Fracture propagation resistance becomes smaller with
by LPM screen-out ranges several hundreds to several thou- fracture size, hence, theareen-out must be induced as
sands psi, depending upon fracture size. quickly as possible while fracture size is still small.

Although many particulate materials seal mud fluids, the above (d) Once fracture surface is covered with mud cake, the LPM
analysis shows that an effective LPM material should have the screen-out is not induced until fresh fracture surface is
following characteristics: created with additional fracture propagation. Hence, it is
not effective to cure an induced fracture (prior to LPM
(a) A relatively large and uniform particle size, as specified application), although it can stabilize further fracture
in our patent application, can prevent mud leaks after propagation.
being packed; a large and uniform particle size does not
significantly change the mud properties, and it can be Because of the above limitations, the LPM is not always effec-
easily removed and recycled. tive. However, statistics show that approximately 50 percent
of lost circulation occurs in permeable sand zones. In addition,
(b) The specific particle sizes are not buried in mud cakes, the rest of the 50 percent induced in shale zones may, in fact,
hence, their concentration increases with fracture be partially induced in permeable mudstone formations rather
propagation as mud fluid is lost because of cake accu- than perfectly impermeable shale. Hence, the LPM should sig-
mulation and filtrate. nificantly reduce the possibility of lost circulation problems.
Careful applications after examining formation lithology would
In addition to the above properties, the following properties significantlyenhance the effectiveness of the LPM applications.
are desired for other considerations.
LPM FIELD TEST NO. 1: GRUBB LEASE
(c) Similar density to base mud. NO. 388 WELL,VENTURA, CALIFORNIA

(d) Not abrasive and not easily crushed while drilling. Prior to setting the 9 5/8" casing at about 9,775 ft TVD, there
are 3 major pressuredepleted formations located at 6,386 ft
The effects of the LPM upon lost circulation are: (First Grubb), 7,981 ft (Second Grubb), and 9,158 ft (Thud
Grubb) TVD, respectively. Mud weights in the 12 114" hole
(a) The magnitude of fracture propagation resistance have been kept below 10.5 ppg since the formation pressure in
increases with formation elastic moduli and effective in each of these zones was estimated to be less than 11.5 ppg.
situ stress. The actual setting depth for 9 518" casing is governed by these
depleted zones as the formation pore pressure will continue to
(b) Fracture propagation resistance is relatively high with a increase and require a high mud weight of approximately
high LPM concentration since a high concentration of 13.5 ppg in the 8 lfl hole below the shoe. It was decided to
LPM induces screen-out while an induced fracture is still run a 50004 open hole leak-off test when the 12 114" hole
A NEW APPROACH TO PREVENTING
LOST CIRCULATION WHILE DRILLING SPE 24599

initiation and propagation of the fractures induced while mud, which in turn prevented leaks of base fluid such as base
drilling. The LPM having this size range was also found to be oil or water. These two chain processes, i.e., (1) a large
optimum in providing an effective sealant for a typical drilling particulate material prevents leak of mud particulate drilling
fluid in a series of "modified" API mud filtrate tests in which fluid and (2) mud particulate material prevents leak of base
the conventional filter paper was replaced by a 1/2" LPM bed. fluid, significantly enhances fracture propagation resistance for
As an example of one of the laboratory test results, an increase induced fractures and also prevents fluid loss into larger
of 1,700 to 3,300 psi fracture extension pressure resulted from natural fractures and pores.
the addition of an optimum amount of LPM to a 16-ppg LVT
oil mud. This increase in fracture propagation pressure is The following four conditions should be satisfied to enhance
equal to an EMW of 3.2 to 6.3 ppg at a drilling depth of fracture propagation resistance:
10,000 ft. The LPM was shown to be effective in both oil and
water based mud systems. (a) The concentration of particulate material should be
increased with fracture propagation such that it packs
LPM is designed to provide a dual-protection mechanism in around the fracture tip.
the borehole. Firstly, it stabilizes the borehole by plugging
minute surface flaws and gravel-size pores. If this protection (b) The initial concentration of the particulate material is
system fails, the second mechanism will initiate as soon as the sufficiently high to induce screen-out while an induced
borehole breaks down and fractures start to form. The LPM fracture is still small.
in the drilling fluid will then act like a fracture propagation
inhibitor through a "screen-out" or self-condensation at frac- (c) The packed particulate material should not allow drilling
ture tip. This process is accomplished by the LPM which falls mud to leak through itself.
in a proper size range, since these particles are large enough
to cause a continuous solid screen-out and provide an excellent (d) The packed particulate material has sufficient strength
fluid sealing property that induces a very large increase in which can plug a sufficiently wide fracture tip region.
formation fracture resistance. This narrow particle size range
is also good for a shale shaker screen. A proper selection of To clarify the factors involved in the screen-out phenomena,
the screen size will not only prevent screen plugging but also a simple analysis is shown below. First, an analysis is made for
allow recovery of LPM for continuous use. condition (a) and (b) considering the material balance around
the fracture tip shown in Figure 1. Suppose the LPM particu-
The preferred LPM materials are nut shells or calcined petro- late material is large such that it will not accumulate along the
leum coke, both of which are environmentally acceptable and fracture surface as a part of mud cake. Then, concentration
do not sigmficantly alter the rheology of drilling fluid even of LPM particulate increases with fracture growth as mud fluid
when introduced at concentrations up to 100 pounds per bar- volume is reduced around the fracture tip. The mud fluid loss
rel in the drilling fluid. As some cuttings have sizes that are occurs partly due to mud cake accumulation and partly
similar to LPM particles and cannot be easily separated from because of filtrate loss. As shown in Appendix A, the condi-
each other while drilling, two LPM circulation schemes have tion that LPM particulate material creates sufficient size of
been used: (1) recirculate both cuttings and LPM until a lost packed screen is:
circulation zone is drilled through; and (2) add a proper
amount of LPM while removing both larger and smaller
cuttings (than the LPM particles) with shale shakers corre-
sponding to LPM size ranges for continuous drilling. The
cuttings with similar sizes to LPM are removed by cyclone or
settlement in mud pits. In this paper, we will describe and This equation shows that: (1) a high fluid loss drilling fluid
discuss two recent field test results using LPM in Conoco wells tends to increase LPM's concentration, (2) a high Young's
located in Ventura, California, and Newkirk, Oklahoma. In modulus and a low fracture toughness creates a narrow frac-
both cases, ground calcined coke was used as the LPM. Other ture tip, resulting in a quick screen-out, and (3) a higher
field tests using walnut shells gave very similar findings to those concentration of LPM, i.e., a lower C,, also enhances LPM
reported here. screen-out.
THEORETICAL ASPECT OF LPM: FRACTURE TIP To search for the particle size satisfying condition (c), an
SCREEN-OUT MECHANISM OF PARTICULATE experiment was conducted using a modified version of a stan-
MATERIAL dard API mud filtrate tester by replacing the filtrate paper
with a layer of LPM particulate screen (about 1/2"thick) while
The standard mud is a composite of two materials with signifi- varying the mesh size of LPM for optimum results on mud
cantly different sizes, where mud (intermediate size particle) sealing. It was found that mud fluid, if Bentonite is included,
prevent. leak of base fluid (fine particle) by forming a dense seals the particulate screen with a specific mesh size, regardless
colloidal particulate cake. This same mechanism was applied of the mud weight or base fluid type. (Note that other types
to the LPM mud system with an order of magnitude larger of fluids, such as a hydraulic fracturing gel, require a finer
scale by extending the composite to a compound of three mesh particulate to seal.)
materials with significantly different sizes, i.e., large LPM,
intermediate size mud particle, and fine base fluid. A properly The LPM particulate material that bridges around a fracture
sized particulate material was searched to prevent leaks of tip will break if fracture fracture pressure increases. Suppose
A NEW APPROACH TO PREVENTING
LOST CIRCULATION WHILE DRILLING SPE 24599

reached the depth of 8,348 ft TVD (or 8,448 ft MD). This pressure gradients include well pressure during LPM testing,
interval covered the first two depleted zones mentioned above. pore pressure, normal mud weight used, and estimated pres-
The objective of this test was to find out if the formation in sure limit for several depleted Grubb zones. As shown in the
the test interval could withstand a pressure of 13.0 ppg EMW diagram, the mud weight equivalent for the well pressure gra-
or higher using the LPM in the existing mud system. dient during LPM testing varied linearly from 15.7 ppg at the
casing shoe point (3,500 ft) to 12.6 ppg at the bottom hole
A water based mud (CYPAN), having a mud weight of 9.8- (8,348 ft). As shown in Figure 4, the pressure gradient was
9.9 ppg, was used. The mud was mixed and spot treated with already far greater than the ordinary mud weights, i-e., 9.0 to
the LPM as planned. The mud weight was increased to 10.3 ppg, used to drill this interval. However, the LPM
10.6 ppg after treatment. The volumes of mud injection and pressure gradient could be much higher if it was not limited by
corresponding wellbore pressure were carefully monitored and the sudden rupture at the shallowest and the weakest casing
recorded during the leak-off test. Pumping was maintained at shoe point at 3,500 ft (where the maximum pressure gradient
a level of 0.5 barrel per minute. The test was terminated after was encountered). Figure 3 also indicated that the resulting
borehole breakdown. A total injection of 30.5 barrels was well pressure gradients in the First Grubb and Second Grubb
achieved. were 13.4 and 12.8 ppg, respectively. These were much
greater than the original limit of 11.5 ppg in the pressure-
The plot for the injection volume (bbl) vs. injection pressure depleted Grubb zones. The pore pressure and normal mud
(psi), as recorded during the open hole leak-off test, is shown weight gradients were shown in Figure 4. The transitional
in Figure 3. To facilitate discussion, the curve was divided into increase of pore pressure starting around 9,800 ft dictated the
five different stages with explanations given as follows: location of the 9 5/8" casing point.

Stage 1: The injection pressure started to bend, indicating The stable growth of fracture and "screen-out," as observed in
a small fracture propagation in the First Grubb the two Grubb zones (Figure 3), demonstrated that the LPM
zone (from 12.2 to 12.6 ppg). The fracture propa- is effective in the pressure-depleted intervals likely to induce
gation remained very stable until it reached the lost circulation problems. In general, the lostcirculation-prone
next stage. zones have surface flaws (cracks or irregular borehole) and
abnormally high stress concentrations because of hole inclina-
Stage 2: A new fracture propagation suddenly occurred tion andlor thermal stress effects, etc. The stress concentra-
(12.2 ppg at the Second Grubb zone). tions induced by these factors would initiate a stable fracture,
but the fracture propagation would be prevented by the sealing
Stage 3. The fracture formed in Stage 2 was arrested while effect of the LPM. (This phenomenon usually occurs when
the first fracture repropagated stably in this stage. the fracture length is less than several feet.) Another effect of
The fracture diameter was estimated to be 32 to the LPM is the screen-out, which will occur at a relatively
41 ft, based on our analysis. longer distance (around 20 to 30 ft) after fracture propagation.
This will most likely occur in a potential-circulation zone in
Stage 4: The fracture propagated slowly at the injection which the pore pressure has been depleted. As shown in
rate of 0.5 bbl per minute. The LPM screen-out Figure 3, the LPM screen-out did occur in Stage 4, which
occurred in the pressure-ascending portion of the demonstrated the LPM effectiveness in preventing fracture
stage. Normally, it would require a pressure of propagation.
more then 2,000 psi to break the solid screen-out
at the fracture tip. The final fracture size was However, the LPM has a limited capability to improve a for-
estimated to be an ellipsoidal shape with height mation that is very tight, impermeable, and flawless, as shown
and length around 32 to 40 ft and 55 to 70 ft, by Stage 5 in Figure 3. The breakdown pressure at the 13 518"
respectively. casing point could well exceed 15.7 ppg if the formation were
more permeable. The leak-off pressure determined previously
Stage 5: An abrupt borehole breakdown occurred at some at the same location was 12.7 ppg (without LPM). Neverthe-
other formation (injection pressure of 920 psi), i.e., less, an impermeable shale zone like this is naturally strong and
most probably at the 13 5/8" casing shoe point at will rarely induce lost circulation problems. The average
3,500 ft. The maximum borehole pressure gradi- borehole Young's modulus calculated from the highest slope
ent reached was 15.8 ppg at the peak of injection. -
(near Stage 5) in Figure 3 was in the range of 0.5x106 1.0x106
The very unstable fracture propagation suggested psi. Using these values in our theoretical model, we were able
that the borehole surface was nearly flawless and to calculate the critical fracture aperture at which the unstable
the formation permeability should be very small fracture was induced in Stage 5 (Figure 3). It showed that the
(s0.01 md). According to the log and drilling critical fracture aperture, We, was in the range of 0.007" to
data, the formation at the casing shoe was a very 0.02" when the maximum effective well pressure was 1,320 psi
tight shale or siltstone. The effectiveness of LPM (injection pressure of 920 psi plus an effective hydrostatic
is not significant in an impermeable formation. stress of 400 psi derived from a 10.6 ppg mud). This value
(W,) was approximately equal or slightly better than that
To highlight the effect of LPM on the formation capability to achieved by water based mud without LPM? This suggested
take additional load because of increased fracture resistance, that the LPM would not increase the breakdown pressure
Figure 4 shows various pressure gradients vs. depth starting much further at the 13 3/8" casing shoe because of the for-
from the 13 318" casing shoe at 3,500 ft to 12,000 ft. The mation's tightness and relative impermeability. However, the
G. F. FUH, N. MORITA,
SPE 24599 P. A BOYD, AND S. J. McGOFFIN

most desirable LPM screen-out mechanism did occur in other during the injection (including flow back) test. The flow rate
locations of higher formation permeabilities, e.g., the First and was maintained at a level of 3 to 4 gpm, and a total injection
Second Grubb zones, as mentioned above. of 15 gallons for each step. The injection pressure and flow
rate versus time of testing were plotted (two fracture reopen-
LPM FIELD TEST NO. 2: CONOCO 33-5 WELL, ing tests were also included). All surface and bottom hole
NEWKIRK, OKLAHOMA pressure recordings agreed with each other very consistently.
However, only the bottom hole recordings will be shown for
Four open hole microfrac tests were conducted in several for- each test result in order to eliminate any slight ambiguity
mations using LPM. Microfracture testing was particularly and/or possible error from flow friction associated with surface
chosen to determine the effectiveness of LPM to increase for- pressure monitoring.
mation fracture resistance in a small and selective interval (6
to 8.5 ft). Microfracturing is a test originally designed for In order to establish baseline data, Test 1 (as shown in
determining the in situ rock stresses in the f~rmation?~The Figure 8) was conducted to determine fracture propagation
test is usually performed at relatively low injection rates of 3 pressure in a sandstone using only the drilling fluid without any
to 25 gpm over a short period of time, such as 3 to 10 min. LPM additive. The plot clearly indicated that there was an
Our objective was to run tests in several formations of similar existing closed fracture in the tested interval, and the fluid
rock types and mechanical characteristics in order to establish injection only induced a fracture reopening at a constant level
a good benchmark for direct comparison of fracture initiation of approximately 756 psi. One more injection was carried out
and propagation pressures caused by drilling fluids with and in the same interval (shown in Figure 9) confirmed the exis-
without LPM additives. In all tests, the water based drilling tence of a closed fracture. Again, the same fracture reopened
fluid (WBS-200) in the 33-5 well was used. at a pressure of about 760 psi, even though a higher pressure
(777 psi) was achieved at some point, indicating the same frac-
A microfrac test set up is shown in Figure 5, where only one ture surface was slightly extended before the pressure dropped
packer was used for the zone isolation and the microfracture back to about 758 psi at the end of injection. The test results
was created around the bottom of the well. A slight variation were still useful since it provided a direct measurement of the
of this setup using two packers (and an anchoring device at the least principal in situ stress, i-e., 760 psi. The formation frac-
bottom) for zone isolation in between were used to test those ture gradient, at a mean depth of 1477 ft, was therefore in the
formation intervals located further away from the bottom. range of 9.9 to 10.1 ppg equivalent. The magnitude was what
The selection of test zones was aimed at formation homoge- we had anticipated prior to conducting the test.
neity and some degree of rock permeability. As a result, shale
zones were excluded and the Gamma log from Well 33-6
In Test 2, we conducted 4 different injections in the sandstone
located nearby (Figure 6) was used to determine the potential formation. The first 2 injections had no LPM, while an LPM
test zones. There were six zones chosen initially. However, it
was used in the last two injections. As shown in Figure 10, an
turned out that only four test intervals were possible because
initial breakdawn pressure of 880 psi (or 11.6 ppg equivalent)
of tremendous difficulties encountered by the anchoring device
designed to provide a firm grip to the wellbore when two was recorded. However, a somewhat constant pressure of 785
to 795 psi (or 10.3 ppg equivalent) was maintained throughout
packers were used to seal off the test zones. Also shown in
Figure 6, four tests were successfully carried out at the fol- the test, including those injections with LPM in the drilling
fluid. The results were not surprising as we have always stated
lowing depths based on test sequence:
that the LPM will not be effective once a lost circulation had
Test 1 (1473 to 1481 ft, sandstone), occurred and an open fracture present. (The fracture surfaces
already covered by mud cake or other solids render LPM
Test 2 (1451.5 to 1463 ft, sandstone), ineffective.) LPM is never intended for a remedial treatment
since the fundamental requirement to trigger the needed LPM
Test 3 (1431.5 to 1443 ft, limestone), plugging mechanism is to have a freshly created fracture
surface for simultaneous filtration and selfandensing (or
Test 4 (1416.5 to 1428 ft sandstone). screen-out) action. The fracture reopening pressure was quite
consistent to the one obtained in Test 1 above.
All the tests were conducted in the 8 ln" hole below the
13 318" casing shoe set at 1364 ft. The injection rate was There were 3 LPM injections made in Test 3, as shown in
maintained in the range of 3-5 gpm for better comparison Figure 11. This was a limestone formation. The maximum
among tests. The addition of LPM increased the mud density injection pressure achieved was 870 psi, or 11.6 ppg equivalent.
to 9.fppg from its original 8.7 to 8.8 ppg. According to the The fracture reopening pressure was in the range of 765 to
mud engineer, the change in mud rheology caused by LPM 790 psi (or 10.2 to 10.6 ppg equivalent) indicating that LPM
addition yas not significant. was not effective in this formation. The results did puzzle us
during the tests. However, we presumed that the limestone
There were two pressure monitoring transducers: one was the formation was too impermeable to trigger any LPM plugging
surface unit which recorded only the injection pressure, the or screen-out mechanism. The latest permeability measure-
other the downhole unit (located in the test interval) which ments, based on cores remered from the 33-1 well in the
recorded the total pressure consisting of injection pressure and same depth interval, has confirmed our belief that the lime-
the hydrostatic pressure column of the drilling fluid. Figure 7 stone was indeed very impermeable. The permeability was
shows a typical recording for the surface pressure and flow rate determined to be much smaller than 0.01 md.
A NEW APPROACH TO PREVENTING
LOST CIRCULATION WHILE DRILLING SPE 24599

After many failed attempts to use the double-packer and an Conoco and sixteen other companies through the Drilling
anchoring device for the next test, we were finally forced to Engineers Association (DEA-13 project) in 1985. The experi-
use the same setup (i.e., a single packer without bottom mental procedures and test results from DEA-13 have been
anchorage unit, as shown in Figure 5) for Test 4 with some published previously in an SPE paper.' In the following
modifications. The tail unit of the test apparatus was fully paragraphs, we will briefly present three figures to support the
extended to the bottom of the well through the use of correct fact of LPM effectiveness in both oil and water based muds.
length of drill pipes and other fittings. The whole process was For simplicity, only the portion of fracture propagation curves
both tedious and time consuming, but there were no other op- will be given and compared with other baseline results
tions at the time. The test turned out to be the best success established earlier by DEA-13 for all types of muds.
for LPM. As shown in Figure 12, there were 4 successful
LPM injections in the test. The initial breakdown pressure was Figure 15 shows the results of a very si@cant increase in
1440 psi, which was 19.5 ppg equivalent. This was a net fracture propagation pressures for the 16-ppg LVT mud after
increase of nearly 8.0 ppg as compared to the results (i-e., mixing with LPM, as compared with the 16-ppg LVT mud
11.6 ppg) obtained from Tests 2 and 3 in which either there with no LPM. The pressure increase was about 1500 to
was no LPM used or LPM effect was not significant. During 4000 psi, which is large enough to stop fracture growth and
fracture propagation in the first injection, the pressure was thereby prevent lost circulation from occurrence during drilling.
impressively high in the range of 15.0 to 17.2 ppg before the The data also compares extremely well with the test results
injection was stopped. The most desirable "screen-out" obtained from the joint-industry DEA-13 project on lost circu-
plugging mechanism was started right after the initial break- lation. The DEA-13 results shown here include: 16-ppg WB
down during our close monitoring. To confirm this result, (water-based mud), 16-ppg DOB (diesel oil based), 16-ppg
three additional injections were performed. The fracture MOB (mineral oil based), 10-ppg WB, 10-ppg DOB, and 10-
reopening pressure remained very high, i.e., in the range of ppg MOB. It should be noted here that the best performer
15.4 to 18.0 ppg, further indicating an excellent LPM sealing established earlier in the DEA-13 test results, i.e., the 16-ppg
in the fracture initiated. The fracture reopening pressure, or WB, was outperformed very significantly (800 to 2300 psi
commonly called the secondary breakdown pressure, was 5.0 increase in fracture propagation pressures) by the 16-ppg LVT
to 7.0 ppg higher than the previous tests. Obviously, the LPM oil mud mixed with our new LPM product (i-e., the higher
worked very well in this permeable sandstone. Based on the pressure increase is the better formation fracture resistance).
recent core test results on permeability, the sandstone in this
interval had a value in the range of 500 to 800 md, which was Figure 16 shows the results of very significant increase in frac-
very similar to the sandstone formation located in the Test 2 ture propagation pressures for the 10-ppg LVT oil mud after
interval (400 to 800 md). The sandstone permeability in the mixing with LPM, as compared with the 16-ppg LVT mud
Test 1interval was determined to be 24 to 86 md. Figure 13 without LPM. A pressure increase exceeding 3000 psi is
shows the same but a complete test curve, using a different shown. The same results from DEA-13 experiments, as ex-
scale, that is equivalent to Figure 12 for Test 4 described plained in Figure 15 above, are also included here for com-
abwe. A hydrostatic head at the level of about 670 psi (indi- parison.
cating a correct 9.1-ppg mud weight equivalent) is clearly
shown in this diagram. Figure 17 shows the results of a very significant increase in
fracture propagation pressures for 10-ppg WB (water-based
To highlight the LPM effectiveness to increase formation frac- mud) after mixing with LPM, as compared with the 10-ppg
ture resistance, Figure 14 was replotted to compare the 2 case WB with no LPM. A pressure increase exceeding 4000 psi is
results (Tests 2 and 4 described abwe) in which the LPM was shown in the figure. The same results from DEA-13
used or left out. The difference was a very lopsided contrast. experiments, as explained in Figure 15 above, are also included
As mentioned earlier, the borehole breakdown pressure was here for comparison.
increased by nearly 8.0 ppg because of LPM effect. The frac-
ture propagation pressure was also increased very impressively The successful results shown above (Figures 15 through 17)
by a range of 5.0 to 7.0 ppg. Obviously, the use of LPM in formed the basis of our drive to field test and apply the LPM
this type of formation will not have any likelihood to induce product to our drilling operations.
lost circulation while drilling; even a very high mud weight
must be applied (because of high angle build or some abnor- The concentration of material used as LPM will vary with the
mally high pore pressure zones located in the adjacent areas, drilling fluid used and the conditions of use. Since the
etc.) for borehole stability or well control reasons. effective LPM plugging of a propagating fracture is funda-
mentally related to the volume of the material used, the con-
FURTHER DISCUSSIONS ON LPM centration of LPM required is thus dependent on the specific
gravity of the material used. Our laboratory findings (given in
The two LPM field tests described abwe were conducted with Figures 15 through 17) clearly shows that the need for the
water based mud. However, we confirm here that the LPM LPM additive will coincide with the need for a heavier drilling
will work equally well in both water and oil based muds. This fluid. So the mud density or mud weight is an important
assertion has been validated and supported by our laboratory factor on deciding the concentration of LPM to be applied.
lost-circulation experiments on a number of large Berea sand- As a principle, the concentration must be great enough to
stone blocks (30" cube) using various oil and water muds provide an effective inhibition of fracture initiation and
injected through a drilled wellbore in the rock samples. The propagation, but should not be so high to change mud rheolo-
lost-circulation research project was originally initiated by gy significantly. Based on the laboratory and field testing
G. E FUH, N. MORITA,
SPE 24599 P. k BOYD, AND S. J. McGOFFIN

results, the most preferred concentration of LPM can be circulation occurrences (if there are no large natural
empirically represented by an equation presented and fractures or vuggy zones).
explained fully in our patent application?
LPM can be applied to benefit drilling operations in a
Based on our experimental and field test results, we have number of ways. By effectively increasing formation frac-
verified that there is a critical size range of the LPM for ture resistance as demonstrated in the field tests, LPM is
optimum fracture inhibition. A similar concept, but using a capable of decreasing the incident of lost circulation,
combination of different materials, was recently applied by which can reduce drilling fluid costs and the time required
Amoco in a water based drilling fluid to successfully drill to drill a well. Since LPM is capable of significantly
through multiple, severely depleted, unconsolidated sandstone increasing borehole breakdown and fracture propagation
reservoirs? The LPM concept, as introduced here, may sound pressures, it is feasible that the application of the tech-
a little unusual or out of our normal paradigm today; however, nology will allow an increase casing setting depths, thus
it is very possible that the LPM technology would soon become eliminating the need for an intermediate casing string.
a part of drilling routine to counter many difficult formation (We have recently accomplished this important objective
problems and/or borehole instabilities associated with the ever- in two of our Midland wells in which the requirements for
increasing demands for high-angle and horizontal wells world- a casing string were eliminated while the previously
wide. reported lost circulationswere prevented completely while
drilling.) By eliminating one or more casing strings, we
SUMMARY can significantly reduce casing and cementing costs while
allowing smaller holes to be drilled. This will allow
This paper provides the theoretical aspect of the LPM and additional cost advantages, as smaller hole sizes require
documents two LPM field tests successfully conducted in smaller volume of drilling fluid and generate fewer drill
Conoco wells in Ventura, California, and Newkirk, Oklahoma. solids requiring costly disposal.
The field results, based on microfracturing and long open hole
leak-off tests as conducted, have confirmed that LPM can sig- The LPM should be effective in both oil and water based
nificantly increase fracture resistance in permeable formations. muds. Its application is most promising in formations that
A remarkable increase of 8.0 ppg in borehole breakdown pres- are likely to induce lost circulation. It will be effective for
sure was achieved in one of our tests in the Newkirk well. An wells l&ted in tectonically stressed zones, depleted
increase of fracture propagation pressure in the range of 3.0 formations, high-angle or horizontal well drilling, highly
to 6.0 ppg was recorded in our field tests. These results are in permeable sand (or shale) zones, or strong formations
agreement with our laboratory findings and are also discussed with highly directional in situ stresses. Since the LPM is
in this paper. Higher formation fracture resistance, resulting aimed at inhibiting any extensive fracture growth in fairly
from the use of LPM in drilling fluids, can reduce or prevent compact but permeable formations during drilling, it will
the occurrence of lost circulation whiie drilling. LPM can be not be effective in preventing lost circulation in zones
continuously mixed in oil or water based muds and will not sig- having vugular structure, large open fracs, or shallow gas
nificantly change mud rheology. This paper also presents zones that are loosely consolidated with gravel and/or
several LPM applications in high-angle and horizontal well large pores.
drilling, drilling through severely depleted formations, and
drilling in highly tectonically active areas where the in situ Prevention of lost circulation is far more cost-effective
stresses are high and unequal. than remedial action taken after the occurrence of lost
circulation. Especially when oil base drilling fluids are
CONCLUSIONS used, mud lost to the formation can be very expensive,
unsafe (underground blowout), and time -consuming.
1. The theoretical aspect of the LPM concept is now for- Thus, adding the LPM as a fracture inhibitor to the
mulated and presented in this paper. Our field test drilling fluid will provide a margin of fracture protection
results and laboratory findings have supported this new for high mud weights used and reduce the degree of
concept and product development. uncertainty involved with drilling operations. Drilling
through highly depleted formations without losing muds
2. The field test results have shown that LPM is a very represents a perfect candidate for immediate LPM appli-
effective inhibitor of fracture initiation and propagation cations.
(or growth) in permeable formations. A remarkable
borehole-breakdown pressure increase of 8.0 ppg was The mechanisms of LPM to prevent lost circulation
achieved in one of our field tests using LPM in a water requires that an effective amount of granular material (as
based mud. In most cases, the fracture propagation determined by the equation in our patent application) in
pressure was increased by 3.0 to 6.0 ppg. These findings a specific size range. Out of this size range, a preference
were in agreement with our laboratory results using for at least 75 percent of the additive should be in a
LPM in sandstone fracture experiments. Although LPM certain narrow mesh cut (as disclosed in the patent appli-
is not as effective in formations with extremely low cation) to ensure an optimum fracture inhibition and fluid
permeability because of high shale and limy contents, sealing. Both nut shells and calcined petroleum coke are
such formations generally have either high in situ stress our primary choice for LPM as they performed equally
or high rock strength, and thus not likely to have lost- well in the field tests. However, any other inert granular
A NEW APPROACH TO PREVENTING
LOST CIRCULATION WHILE DRILLING SPE 24599

material with 1.2 to 2.0 specific gravity would also be an 4. Fuh, G. F., and Morita, N. et al.: "Method for Inhibiting
appropriate candidate for LPM.* the Initiation and Propagation of Formation Fractures
While Drilling," Conow Patent Application (Oct. 1991).
NOMENCLATURE
5. Ali, A , Kalloo, C. L., and Singh, U. B., "A Practical
Fluid loss coefficient Approach for Preventing Lost Circulation in Severely
Liquid concentration of drilling fluid Depleted Unconsolidated Sandstone Reservoirs," Paper
Mud fluid concentration with respect to total mud IADCBPE 21917 presented at the 1991 SPE/IADC
(mud +LPM) Drilling Conference (1991), pp. 189-202.
Moduli of rock
Fracture height APPENDIX A: FACTORS AFFECTING SCREEN-OUT
Rock fracture toughness
Fracture pseudo-toughness Factors forming packed particulate screen around a fracture
Propagation pressure resistance induced by tip are analyzed in this appendix using simple assumptions.
screen-out
Fracture pressure Suppose a large fracture is propagating with a constant velocity
Pore pressure v. Consider a fracture tip section r, (distance r, from fracture
Flow rate tip). Only a unit height is considered (Figure A-1).
Total filtrate loss
Radial distance Suppose fluid loss occurs with
Fracture tip region
Fracture pressure
Spurt loss
Time
Fracture tip velocity The total fluid lost from fracture surface around the fracture
Fracture volume tip is
Fracture width
Fracture width at sealing point when fracture has
stopped (approximately equal to cake thickness if
We is large compared to rock grain size)
Time initiating filtrate loss
Porosity of packed LPM
Horizontal in-situ stress (negative)

ACKNOWLEDGEMENTS

We thank the management of Conow Inc. for permission to


publish this paper. Halliburton Services assisted us in carrying
out the field tests as presented.

REFERENCES

1. Morita, N., Black, A. D., and Fuh, G. F., "Theory of Lost


Circulation Pressure," Paper SPE 20409 presented at the where Spis spurt loss and t = (r,-r)k is the time passed since
65th Annual Technical Conference and Exhibition of the the fracture tip crossed at r,.
Society of Petroleum Engineers (1990) pp. 43-58.
Approximate half fracture tip width is given by
2. Kuhlman, R. D., "Microfrac Tests Optimize Frac Jobs,"
Oil and Gas Journal (Jan. 1990) pp. 45-49.

3. Daneshy, A. A , Slusher, G. L., Chisholm, P. T., and


Magee, D. A , '!In-Situ Stress Measurements During
Drilling," Paper SPE 13227 presented at the 59th Annual where k; is pseudo-fracture toughness.
Technical Conference and Exhibition of the Society of
Petroleum Engineers (1984). Then the volume of fracture tip becomes

*As stated earlier, the LPM technology developed and dis-


cussed in this paper has been filed and pending a U.S. patent
application by G. F. Fuh and N. Morita of Conoco Inc.
576
G. F. FUH, N. MORITA,
SPE 24599 P. A. BOYD, AND S. J. McGOFFIN

Note that the fluid loss coefficient Co and S, are those for
drilling fluid traveling a certain distance through fracture gap,
hence, they should be si@cantly smaller than those measured
with the API specified condition.
Drilling fluid volume, which can be removed from the fracture Equation A-7 shows the screen-out region r, becomes wide if
tip, is the left term approaches to zero or negative. Hence, the con-
dition to stop fracture propagation is
3
-

Since the total mud fluid lost from the tip region is greater or
equal to the mud fluid volume to be removable from the
region to induce screen-out, we have
cO/c,2 ( I - u 2 ) kc, (C" - 4) fi
f i E

Although Equation (A-9) is derived based upon simple


assumptions, it suggests that fracture propagation stops if fluid
loss coefficient Co is high, fracture velocity is low, Young's
modulus is high, pseudo-fracture toughness is low, and mud
fluid content C, adjacent to fracture tip is close to irreducible
Equation (A-6) is simplified as mud fluid content 4.
0 5 10 15 20 25 30
Injection Volume (bbl)

Fig. 3--5000-ft open-hole leak-off test using LPM in the Grubb No. 388 Well.

Fig. 1--LPM screen-out at fracture tip.

LPM Field Test


Well Pressure Gradient Est. Pressure Limit for
during LPM Testing lst,2nd,3rd Grubb Zones
. -.. .-..-. .-..- .--*--.
15 ..-...
. -.
Normal Mud Weight
............. Pore Pressure Gradient
--c-
14 . .

. -
.......................*...........* 0
_------_---
............."...........
_---
............................. -_.---I

C - _--- 1st Grubb ( 4 1.5 ppg)


2nd Grubb (<I 1.5 ppg)
8
4 t
4,000 5,000 6,000 7,000 8,000 9,000 10,000 11,000 12,000
Wc*E*(a,,-Po)/(l-U' ) @si**2*in) Depth (ft)

Fig. 4--Comparisonof LPM pressure gradient with pore pressure gradient and normal
Fig. 2--Fracture pseudo-toughness. mud weight used in the Grubb Lease wells.
I OPEN HOLE MICROFRACTURING I

TEST 4

TEST 3

TEST 2

--
1 4 8 1 TEST

TEST 1: NO L P M

INJ
1 NO LPM
7 NO LPM
INJ

u INJ
3 LPM
INJ 4 LPN
TEST 3 : LPM
T t s i 4 : LPM
Fig. 5--Field set-up for open-hole microfracture tests (Ref.
Fig. 6--Selection of four open-hole microfracture tests using
LPM (based on Gamma log data) in the Conoco 33-5 Well.

Fig. 7--?lpical recordings of surface pressure and flow rate data during open-hole
microfracture tests in the Conoco 33-5 Well.

579
Presswe Vs Time
Pressue Vs Time

Fig. &-Test 1: fracture opening test (no LPM used) in a


sandstone formation (with closed fracture). Fig. 9--Test 2: fracture reopening test (no LPM used) in a
sandstone formation.

Pressure Vs Time Pressue Vs Time

(n-

Fig. 10--Test 2: fracture opening test (four injections) in a Fig. 11--Test 3: fracture opening tests (three injections) using
sandstone formation (LPM used in the last two injections LPM in a limestone formation.
only).
Pressure Vs Time Pressue Vs Time
2.1

2.0

*^
n^
Q 1.2 ' 16

- .
5
;
i f 1;"
0.9
I 0.8

0.4

0.6 0.0

B 8
Fig. 12--Test 4: fracture opening tests (four injections) using Fig. 13--Test 4: complete downhole pressure curves showing
LPM in a sandstone formation. .four injections using LPM in a sandstone formation.

800-
-M
1-ma

700-

-
B m-

I
P
S
=-
1spW LYT (with LPM)

I ,-
300-
Hydrostatic Haad = m pal (9.1 ma)

I n j d o n Time, seconds

- 10 Fig. 15--Comparison of fracture propagation pressures for


different mud types with or without LPM. (The test for 16-
ppg LVT oil mud using LPM is shown.)
4 8 12 16 20
lnjeclion Volume, Gals

Fig. 14--Comparison of well injection pressures using LPM


(Test 4) and without LPM (Test 2) in sandstone formations.
Fracture cross section \
01 I I I I
120 240 360 480 8
Injection The, seconds

Fig. 16--Comparison of fracture propagation pressures for


different mud types with or without LPM. (The test for 10-
ppg LVT oil mud using LPM is shown.)

\ Fracture tip

lsppg WB (with LPM)

Fig. A-1--Material balance around fracture tip.

lo-ppg WB

0 I I I I
120 240 380 480
Injection lime, seconds

Fig. 17--Comparison of fracture propagation pressures for


different mud types with or without LPM. (The test for 10-
ppg water based mud using LPM is shown.)

You might also like