DRILL COLLAR WHIRL: A CAUSE OF TRICONE BIT FAILURE

Matt J. Meiners Roy Ledgerwood

Hughes Christensen Company Hughes Christensen Company
9110 Grogan’s Mill Rd. 9110 Grogan’s Mill Rd.
The Woodlands, Texas, 77380 The Woodlands, Texas, 77380

ABSTRACT technology must also increase to consistently lower cost per

Tricone bits are generally assumed to be robust enough to foot. Traditionally, little emphasis has been placed on the
withstand downhole vibrations. Downhole vibrations can be effect of bottom hole assembly (BHA) vibrations on roller cone
categorized as torsional (stick slip), axial (bounce) and lateral bit durability. This paper will demonstrate that strong lateral
(whirl). Drill strings with tricone bits are prone to axial and vibrations can significantly damage a roller cone bit in very
lateral vibrations. Consideration is rarely given to BHA design short period of time, thus demonstrating the importance of
and selection of drilling practices that minimize vibration of controlling these vibrations to optimize drilling performance (9).
tricone bits. In contrast, it is well documented that downhole
vibrations cause catastrophic damage to PDC bits. The BHA vibrations can be categorized as axial, torsional or
combination of anti-whirl bit design features, BHA design and lateral. Torsional vibrations are typical when (but not limited to)
optimized drilling practices have resulted in longer PDC bit life drilling with PDC bits and are instigated by the high reactive
and significantly lower drilling costs. Tests described in this torque of the PDC drilling process. In a severe case the bit
paper demonstrate instances where downhole vibrations also can actually “stick” on bottom while the drill string continues to
cause catastrophic damage to tricone bits. rotate and build energy. Once the energy is released, the bit
“slips” causing instantaneous bit RPM to be much higher than
Field testing of various bit designs show that downhole the average rotary speed. This chain of events is usually
vibrations are a significant cause of catastrophic damage to referred to as stick-slip.(2, 3, 12). Roller cone bits are more
rollercone bits. This finding is based on surface and downhole susceptible to axial vibrations that are instigated by the
instrumentation to detect the occurrence and magnitude of the formation of a tri-lobed bottom hole pattern (6). Thus, axial
vibration in addition to visual inspection of the bits after drilling. vibrations usually occur at a frequency of 3X rotary speed.
These tests indicate tricone bits can endure strong axial Lateral vibrations can occur with either PDC or roller cone bits.
vibrations but are damaged by lateral vibrations of equal With PDC bits, lateral vibrations are associated with bit whirl
magnitude especially when drilling certain lithologies. To test (1). When drilling with roller cone bits, lateral vibrations are
the extent of the damage in a controlled environment, new bits being generated by drill collar whirl (5,7) rather than by the bit.
were subjected to lateral vibrations while drilling 2-4 foot (0.6 –
1.2 m) intervals. These experiments revealed that breakage This paper is an overview of a series of field tests
begins in the inner rows and progresses through the remainder conducted at the Amoco Catoosa Facility, near Tulsa,
of the cutting structure. These studies indicate that lateral Oklahoma, whereby various downhole vibrations were induced,
vibrations are not always reflected in the vibrations at the rig measured and correlated to roller cone bit failures. A partial
floor, therefore downhole monitoring is critical. Testing also stratigraphic section at the Catoosa Facility is shown in Fig 1
revealed that tricone bits do not always cause these vibrations, (14). The particular formations of interest discussed in this
as is often the case in PDC applications. Rather, the vibrations paper are the Mississippi Limestone (1270-1570 feet, 387.1-
are generated by drill collar whirl. 478.5m) , Misener Sandstone (1600-1630 feet. 487.7-496.8 m)
and Arbuckle Dolomite (1630 feet. 469.8 m and below). The
The authors discuss how to monitor these vibrations in real- majority of these tests used several 7-7/8” (200 mm) IADC 417
time and how the presence of a downhole motor affects roller cone bits. The progression of catastrophic failure
vibrations. mechanisms, as a result of BHA vibrations was studied in
detail. Catastrophic failure is defined in this text as when the
INTRODUCTION majority of the bit’s cutting structure is extensively damaged in
Since the advent of tungsten carbide insert (TCI) bits in a short time period or approximately 5-7 feet (1.5 – 2.1 m) of
1951, significant advances in TCI bit technology have increased drilling. In contrast, non-catastrophic failure would be failure
overall bit performance. As the necessity for hours and footage that occurs while drilling a substantial amount of footage and is
drilled by a bit continues to increase, improvements in bit characterized by gradual wear, chipping, breaking and eventual
failure of the cutting structure.
 A typical surface chart recording with the surge pattern in
DOWNHOLE DATA ACQUISITION the WOB plot is shown in Fig. 3. The RPM was 70 and the
WOB varied between 15 KIPS (6.7 kdaN) to 40 KIPS (17.8
For these tests, a commercially available downhole kdaN). The WOB surge pattern repeats at a low period, which
recording tool was used. The tool is comprised of three is on the order of several seconds and varies as a function of
mutually orthogonal AC coupled accelerometers. The the WOB, from about 10 seconds at 15 KIPS (6.7 kdaN) to
accelerometers are rated at +- 200 g and are sampled at 1024 about 30 seconds at 30 KIPS (13.3 kdaN). The cycle begins
Hz with a resolution of 0.2g. The orientation of the with smooth WOB, smooth RPM and low torque. As the torque
accelerometers allows measurements to be taken in three slowly builds, the WOB begins to fluctuate, forming the wide
directions, namely X, Y and Z. The Z accelerometer is aligned portion of the surge pattern on the chart. At the time of highest
parallel to the borehole, thus being sensitive to axial torque, the WOB displays maximum variation; the rig floor
accelerations. The X accelerometer is aligned to be sensitive bounces and the RPM fluctuates. Then almost
to lateral and radial accelerations while the Y accelerometer is instantaneously, the torque drops and the vibrations at the rig
sensitive to lateral and tangential accelerations. Effectively, for floor smooth out. This pattern repeats itself.
lateral accelerations, X and Y measurements will be nearly the
same ( X-Y near 1g ). In the case of torsional accelerations Figure 4 shows the downhole vibration data that
there is a separation between X and Y that is greater than 1g. corresponds to the surface data shown in Fig. 3. In Fig. 4, and
The signal from each accelerometer is conditioned and the all such figures in this paper, the X-accelerations are shown in
results were processed to produce average, peak and normal line weight. Y-accelerations are shown in bold line and
instantaneous forms. In all tests, the tool was programmed to Z-accelerations are shown dotted line. The bit weight was
record a one-second, high frequency vibration sample every incrementally increased from time 9:35 to time 10:01. As the
four seconds. The tool kept a record of the average value and WOB increases, the intensity of the vibration increases. Note
peak value for each high frequency sample. For these tests, that the Y-acceleration is much higher than the X-acceleration.
only the average recordings were used. This reflects a large torsional component to the vibration. Note
also the increasing Z-acceleration with increasing WOB.
Most wells drilled in this investigation were vertical and
drilled with a “slick” assembly. A slick assembly contains only Referring again to Fig. 3, note the surge patterns in the
the bit and drill collars. The recording tool was placed surface data die out above 30 KIPS (13.3 kdaN), at time
immediately above the bit and below 24 drill collars. The 9:49:36. However, the YZ vibration measured downhole
collars used here were 6-1/8 inch OD and 2-13/16 ID weighing continues, and increases in intensity as WOB increases all the
80 lbs/ft (155.6 mm OD, 71.4 mm ID, 1.17 kN/m). This way to 45 KIPS (20.0 kdaN) at 10:01:08. It was not unusual for
configuration allows the vibration measurements to be taken as the downhole tool to indicate YZ vibration even when there are
close to the bit as possible. A non-scaled representation of the no indications at the surface. This is probably due to
BHA used for this study is illustrated in Fig 2. dampening effects and the dynamic transfer function across the
drill string.
Two major types of vibration modes were identified during
these tests. The first type was a coupled torsional/axial A physical explanation of the YZ vibration has been
vibration. The second was a pure lateral vibration generated by proposed (10). It may be the alternate formation and
drill collar whirl. Various operating parameters, formations, breakdown of a cam-lobed bottom hole pattern. As a cam
BHA design (including motors) were all studied to determine the lobed pattern begins to form, the torque rises and the WOB
effect these have on the vibration type and magnitude. fluctuates as the bit rides over the cam. This increases to a
Hereafter, the torsional/axial vibration will be referred to as “YZ” peak when the cam is at its largest. Then instantaneously the
vibration (High Y and Z vibrations recorded at the tool) and cam-lobed bottom hole pattern breaks down, perhaps due to
lateral vibrations will be referred to as “XY” vibrations (High X the impact loads; the drilling smoothes out and the cycle
and Y vibrations recorded at the tool). The next two sections of repeats.
this paper will describe these vibrations in more detail and
surface recordings will be presented that correlate to each type The YZ pattern of vibration is often stopped when WOB is
of vibration. increased. In these cases the torque changes from an
alternating value to a high steady value. An example of this is
TYPES OF VIBRATION shown in Fig. 5. The high torque that occurs when the YZ
vibration is eliminated results from backward drill collar whirl
YZ - Coupled Torsional/Axial Vibration (discussed in subsequent sections). The high friction between
The coupled torsional/axial vibration, or YZ vibration, was the drill collars and the wall of the hole dampens out the axial
the first vibration characterized. Its signatures are bounce on vibrations and kills the YZ vibration.
the rig floor, fluctuating surface RPM, periodic surge patterns
on the WOB plot and high average Y and Z accelerations with XY – Lateral Vibration-Drill Collar Whirl
low average X accelerations downhole. The signatures of the XY vibration are high surface torque,
smooth running at the rig floor and high average XY
accelerations with low Z-accelerations. The high surface torque
of the XY vibration is particularly easy to identify when the XY In the Misener Sandstone the weights and rotary speeds
vibration first initiates. Fig. 6 shows such a case. In the time that cause the XY vibration were mapped. Fig. 9 summarizes
interval shown, rotary speed was a constant 70 RPM, and bit this data. Each circle represents a weight and rotary speed
weight was being increased from 15 to 45 KIPS (6.7 – 20 combination run during testing. The shaded circles represent
kdaN). Note the torque is increasing slowly as a function of bit combinations where XY vibration occurred and the open circles
weight from 15 KIPS (6.7 kdaN) to 30 KIPS (13.4 kdaN). When represent combinations where it did not. The number in the
bit weight is increased to 35 KIPS (15.6 kdaN), there is a shaded circles represents the average intensity of the XY
dramatic increase in torque. The higher torque level observed vibration in g. None of the circles represent YZ vibration
from 35 KIPS (15.6 kdaN ) to 45 KIPS (20.0 kdaN) is typical of because it was never observed in the Misener Sandstone,
the XY vibration. despite numerous attempts. The XY vibration existed at rotary
speeds as low as 45 RPM and bit weights as low as 10 KIPS.
This transition torque is sometimes observed without any There appears to be a safe zone in the 20 KIP (6.1 kdaN)/60
change in drilling parameters. The chart recording in Fig. 7 RPM range and below. Note the intensity of the XY vibration is
shows such an incident. In this test, the bit weight is 25 KIPS almost totally a function of rotary speed. Once drill collar whirl
(8.9 kdaN). The rotary speed is 80 RPM at time 10:36 and then is initialized, it is fueled almost totally by rotary speed.
increased to 90 RPM at 10:38. During the 80 RPM portion,
there is a well defined surge pattern of the YZ vibration, with the A similar matrix was mapped while drilling the Mississippi
period of the WOB surge being about 10 seconds. When the Limestone, shown in Fig. 10. The filled circles represent test
rotary speed is increased to 90 RPM, the surge pattern points at which XY vibration occurred and the open circles
continues, but at a higher frequency of about 4 seconds, represent test points at which YZ vibration occurred. Note the
making the surge patterns difficult to see on the strip chart YZ vibration tended to exist at low rotary speeds and to some
recording. Five minutes later, at 10:41, with no change in degree at low bit weights. As found in the Misener, the intensity
drilling parameters, torque increases, indicating drill collar whirl of the XY vibration is almost totally a function of rotary speed.
and the transition to XY vibration. The surge patterns of the YZ When rotary speed and weight combinations exceeded a
type vibration are dampened out. This change from YZ to XY certain level, the drill collars began to whirl causing XY
vibration can also be seen in the downhole data in Fig. 8. At vibration. This is demonstrated by comparing Fig. 9 and Fig. 10.
10:41 where the vibration changes from the high axial (Z) and Note it was more difficult to initiate XY vibration in the
torsional (Y), to the high lateral (X and Y). This high XY is Limestone than it had been in the Misener Sandstone. In the
typical of lateral vibration. Limestone, XY vibration did not exist below 90 RPM. In
contrast to this, XY vibration was easy to start in the Misener as
The nature of drill collar whirl has been well documented in low as 45 RPM.
the literature (3,4,7,13) In these series of tests, the resulting
vibrations created by drill collar whirl were thoroughly studied.
In summary, the following observations were made. ROLLER CONE DAMAGE DUE TO LATERAL VIBRATIONS
• Drill collar whirl starts most easily when bit weight is
increased and the drill collars are buckled. Throughout the testing program, two bits were totally
• Once drill collar whirl begins, it can only be eliminated by destroyed while drilling the 30 ft (9.1 m) section of Misener
picking-up off bottom and completely stopping drillstring Sandstone. Both bits were IADC 417 type and had identical
rotation. cutter counts, rows, and tooth shape and carbide grade. The
• Simultaneous downhole vibration measurements at the only difference was the standard bit had a positive cone offset
bit and near the top of the drill collars show nearly and the second had a negative cone offset. Average downhole
identical vibration characteristics. accelerations for the first of these bits are shown in Fig. 11. In
• When drill collars are in an overgauge hole, drill collar this test, WOB and RPM matrices were conducted at the
whirl will did not occur. beginning of the test, between 9:30 and 11:10. At 11:12 drilling
• The transition into drill collar whirl is more easily began at 20 KIPS (8.9 kdaN) and 65 RPM and continued using
instigated by the excitation of the bit. However, the these drilling parameters until 11:42. At 11:30, without a
presence of the bit is not necessary to transition the change in drilling parameters, the downhole vibrations
collars into whirl mode. transitioned from nearly non-existent to XY vibration at 4-5 g
average. This corresponds to the time the bit encountered the
Misener Sandstone. The increase in vibrations at 12:00 result
EFFECT OF OPERATING PARAMETERS AND FORMATION from a rotary speed increased to 90 RPM and bit weight
ON VIBRATION MAGNITUDES decrease to 15 KIPS. A photo of this dull bit, next to a slightly
dulled bit of the same type is shown in Fig. 12. The second
Some tests in this investigation were dedicated to mapping IADC 417 encountered similar vibrations and was also pulled in
out the combinations of bit weight and rotary speed at which the the condition represented in Fig. 12.
XY and YZ vibration occurred. This data also show there is a
lithology factor affecting the initiation of XY vibration.
CUTTING STRUCTURE BREAKDOWN PATTERN AND inner row of the number 3 cone and one broken off at the cone
EFFECT OF TOOTH SHAPE. shell in the first inner rows of the number 1 and number 2
cones. The remaining inner row teeth had cracks and wear
Because of the excessive damage to the bits used to drill flats.
the 30 feet of Misener Sandstone, it was not possible to
determine where the damage began or how it progressed. In These tests indicate it is unlikely a more durable tooth
order to learn how and where the damage was initiated, three shape will improve a bit’s ability to resist breakage from XY
tests were conducted in which new bits were run with XY vibration. However, this is not to say metallurgical
vibration in the Misener for about 4 feet (1.2 m) of drilling each. advancements in carbide properties will minimize this kind of
These bits also used varying tooth shape to determine if more damage.
durable tooth shapes would resist the breakage. All three bits
were ten-row, IADC 417 with identical row and cutter layout. EFFECT OF BIT AGGRESSIVENESS (IADC CLASS) ON
The only differences were tooth shape and tooth projection. In BREAKAGE
all cases, damage initiated on the inner rows, especially on the
first inner row of the number 3 cone, usually at the wide busted An IADC 417 type bit is a more aggressive bit (a softer
pitch in that row. formation bit) than would normally be chosen to drill a rock as
hard as the Misener Sandstone. In the field, a less aggressive
The first IADC 417 drilled four feet (1.2 m) of Misener, from bit (harder formation bit) would normally be chosen. In order to
1607 to 1611 feet (489.8 –491.0 m) in 2-1/2 minutes. The bit determine what IADC class of bit is required to drill the Misener
weight was 45 KIPS (20.0 kdaN) and the rotary speed was 70 without tooth breakage, a variety of bits were tested. These
RPM. The bit experienced XY vibration of 4-5 average g for the were (in order) an IADC 437, 527, 627 and 637. Furthermore, it
2-1/2 minutes. Examination of the bit revealed three broken was not known whether a hard formation bit merely endures the
compacts on the first inner row of the number 3 cone and two XY vibration or whether a hard formation bit might not instigate
chipped teeth on the first inner row of the number 1 cone. Most the XY vibration..
inner row teeth had cracks and a small wear flat. Photos of the
damaged compacts are shown in Fig. 13. The mode of failure All of the bits were run with 45 KIPS (20.0 kdaN) and 70
for the damaged compact shown in Fig. 13 is shear failure. RPM and drilled about 7 ft (2.1m) feet of the Misener
Shear failures begin with a subsurface void generation resulting Sandstone. The IADC 437 drilled from 1614 to 1623 feet
from both excessive compressive and shear stresses. (491.9 – 494.7m). It had one broken tooth on the first inner row
Downward from the crack initiation site was a small sheared of the #2 cone with most other inner row teeth being cracked.
region. Areas beneath the sheared region exhibited a tensile The IADC 527 bit drilled an offset well from 1600 to 1607 feet
mode failure from the wedging effect of WOB pulling the crack (487.7 – 489.8m) and experienced average XY vibrations of 5-6
apart. Fig. 14 highlights and pinpoints the distinct regions g. It had one broken tooth on the second inner row of the #2
found typical of shear failure. cone and one broken tooth on the third inner row of the #3
cone. Other inner row teeth were cracked. The IADC 627 bit
Shear failures are not often directly observed in the field. If drilled from 1607 to 1614 feet (489.8 – 491.9 m) and again
the fracture surface is observed shortly after failure, it will experienced average XY vibration 5-6 g. It had one broken
exhibit a shiny appearance. Shear failure usually happens tooth on the second inner row of the #2 cone and one broken
abruptly and therefore any further damage such as chipping, tooth on the first inner row of the #3 cone. Other teeth in the
spalling and erosion of the remaining compact usually mask the first inner row of the #3 cone were cracked.
characteristics of shear failure.
The IADC 637 bit drilled and offset well from 1604 to 1607
The other two bits that were tested in the same manner feet (488.9 – 489.8 m), experiencing average XY vibration of 5-
had the same projection as the first, but more durable 7 g and was pulled without damage or broken teeth. This same
compacts were used. The second bit used compacts with a bit was used in several subsequent tests in which a variety of
larger cross section and blunter nose compact than those used weights and rotary speeds were used to map-out the effect of
in the bit shown in Figs. 12-14. The third bit used compacts drilling parameters on vibrations. To date, this same bit has
with shorter projection chisel shape. Both of these bits suffered drilled a total of 41 feet in the Misener Sandstone and 53 feet in
damage that initiated on the inner rows, like the first bit the Arbuckle Dolomite. In summary, all of the above bits
previously described. The bit with blunt compacts drilled from experienced the XY vibration and all of them, except the IADC
1611 to 1615 feet (491.9 - 492.3 m ) in five minutes at 45 KIPS 637 suffered catastrophic tooth breakage
(20.0 kdaN) and 70 RPM experiencing an average XY vibration
of 5-6 g. It had two broken compacts on the first inner row of There are several differences between an IADC 417 and
the number 3 cone and one broken compact on the first inner 637 roller cone bit. These include: cutter count, projection,
row of the number 2 cone. The remaining inner row teeth had pitch, shape and size, carbide grade, number of rows, pitch
shear cracks and wear flats. The bit with shorter chisels drilled schemes, cone profile, pin angle, cone oversize angle and
from 1615 to 1619 ft (492.3 – 493.5 m) in three minutes at 45 offset. It is not known at this time which combination of these
KIPS (20.0 kdaN) and 70 RPM, also experiencing an average features allows the IADC 637 to survive these vibrations. A
XY vibration of about 5 g. It had three broken teeth on the first thorough study of the design variations between these designs
and the breakage patterns on them may lead to insights. particular test, a pendulum assembly was used in an attempt to
Although the IADC 637 can drill this formation, it does so at a generate YZ vibrations in the Misener. Prior test experience
much slower ROP. Ultimately, if a softer formation bit, like the demonstrated that running a pendulum assembly and running
IADC 417 could be designed to endure the vibrations induced drilling parameters with low bit weight and low rotary speed
by drill collar whirl, or if drill collar whirl could be eliminated, a favors the YZ vibration. As a result, drill collar whirl and XY
much lower cost per foot would be realized on applicable vibrations occurred and the bit was pulled with typical compacts
drilling operations worldwide. failed due to shear.

EFFECT OF MOTOR ON VIBRATIONS THE BENEFIT OF REAL TIME MWD VIBRATION DATA

In an additional test a downhole motor was run, in a slick Without knowledge or insight about downhole vibrations,
assembly, with a vibration-recording tool on each end of it it’s natural to rely on surface parameters to draw inferences
simultaneously. The interval drilled was 1593-1610 feet (485.5- about bit dynamics. In some instances, this can be deceiving.
490.7m), focusing on the last ten feet of the Misener Again referring to Fig. 12, the bit on the right was subjected to
Sandstone. The first five feet of Misener was drilled in the severe YZ vibrations while the bit on the left experienced drill
sliding mode with the motor turning at 300 RPM. The last five collar whirl and strong XY vibrations. From the surface
feet was drilled in the rotating mode, with the drill string rotated vibrations one might think the opposite to be true since the bit
at 50 RPM and the motor at 250 RPM. In each case, the bit on the right experienced sever bit bounce at the rig floor while
rotary speed was 300 RPM. The target bit weight was 35 KIPS the bit on the left produced no vibrations at the rig floor. As well
(15.5 KdaN) in each case. depth increases, more of the downhole vibrations can be
dampened by borehole contact and mask any indications at the
The downhole vibrations from both tools are shown in Figure surface.
15. The vibrations on the bit side of the motor are shown in
solid lines and the vibrations on the drill collar side of the motor There are commercially available MWD tools that will
are sown in dotted lines. In the sliding portion of the test, record, calculate and scale downhole vibrations. These scaled
between 1600 and 1604 feet (487.7 and 488.9 m) an electronic vibrations can then be transmitted to the surface in terms of
ROP limiter (rig specific, 200 ft/hr, 61m/hr) was inadvertently vibration severity level via mud pulse technology. The use of
set, which prevented the bit weight from building up to the real time MWD downhole vibration monitoring has proven to be
target 35 KIPS. The bit weight was about 15 KIPS between a very successful practice in field applications (8, 11). In the
1600 and 1604 feet (487.7 and 488.9 m) and finally reached 35 scope of this paper, this technique was used as an example to
KIPS (15.5 kdaN) at 1605 feet (489.2 m). determine if and when drill collar whirl and XY vibration were
The bit side vibrations are of the XY type, coupled with an axial occurring on a deeper well.
(Z) vibration, and range from 4g to 8g average while the drill
collar side vibrations are between 1g and 3g. This indicates that An 12-¼” (311.2 mm) IADC 447 tri-cone bit on a packed
the main source of vibrations in this test come from the bit and assembly was used to drill a hard dolomite section near Beggs,
that the motor effectively filters out these vibrations at the drill Oklahoma from 2970 to 3105 feet (905.3 – 946.4m). The strip
collars. The bit used in this test was a 7-7/8 inch (200mm) chart in Figure 16 shows the lithology, surface parameters
IADC 417 and was pulled with shear cracks and broken (WOB, RPM, and torque), downhole lateral vibration (XY),
compacts on the inners rows of the bit. surface acceleration (measured at the rig floor) and scaled
MWD vibration severity. The downhole lateral vibration was
dumped from the tool after tripping out of the hole, whereas the
VIBRATIONS THAT HAVE NOT DAMAGED BITS. severity level was transmitted to the surface in real time while
drilling. Drilling initiated using 20 KIPS (8.9 kdaN) using 60-120
In all tests conducted, an IADC 417 type bit was used to RPM. Drilling torque varied slightly between 1500-2000 ft-lbs
drill the Mississippi Limestone section. While drilling this (2033 – 2711 N-m). From 2970 – 3010 ft (905.3 – 917.4 m),
interval, average XY vibrations as high as 10 g were recorded, drilling was smooth with surface vibration near zero g and
but no damage to the bit was ever observed. Also, no lateral severity level from the tool was also zero. At 3010 ft
catastrophic damage had occurred to any bit, which (917.4 m) WOB was increased to 50 KIPS (22.3 kdaN). This
experienced the YZ type vibration, though this vibration has caused the drill collars to whirl, indicated by the surface torque
often exceeded 5 average g and at times has been as high as jumping to 6000 ft-lbs (8133 N-m). Vibrations at the surface
10 average g. The non-damaged bit in Fig. 12 drilled the entire remained smooth, however lateral severity level from the
Mississippi Limestone section mostly with 5-10 g YZ vibration vibration recording tool were 4 to 5, correlating to a XY vibration
and strong bit bounce at the surface. magnitude of 3-8 g. As the well deepened the severity level
increased to a level 6, indicating lateral vibration of up to 15 g.
The YZ vibration never occurred in the Misener Sandstone. This behavior continued until 3105 ft (946.4 m) at which time
Therefore, it is not known if the YZ vibration would cause the rig experienced a twist off in the collars. The bit was pulled
damage to a bit drilling in the Misener Sandstone. In one (fished) with a dull condition of 1-4-BT-G-E-I-CT-DSF. In this
dull, nearly half of the heel row compacts were broken due to
shear failure. A picture of one of the broken compact is shown • Downhole motors filter out vibrations and may be
in Figure 17. beneficial, insulating the drill string from excitations of the
bit and insulating the bit from drill collar whirl (while in
This example shows the advantage of monitoring downhole rotating mode).
vibrations in real time. The lateral vibrations recorded in this
test were induced intentionally. Had these vibrations been
observed in a real field application, surface parameters could RECOMMENDATIONS FOR FUTURE WORK
have been carefully altered to minimize the vibration, thus
preventing catastrophic bit failure and a twist off, both costing It is not accelerations themselves that cause damage; it is
valuable rig time. the forces that arise from these accelerations. It is likely that a
roller cone bit compact could accelerate the relatively small
CONCLUSIONS mass of the bit itself at the levels of acceleration we observe
without breaking. But when the mass of several drill collars are
• The XY vibration is generated by the drill collar whirl. This added, forces on the compact become very high. This
lateral vibration has been identified to catastrophically suggests it may be possible for aggressive bits to be used in
destroy soft formation rolling cutter bits. hard formations by appropriately isolating the bit from the mass
• In a hard sandstone, XY vibrations on the order of 5 of the drill collars. Isolating the bit will also reduce the tendency
average g are sufficient to cause catastrophic damage to of the bit to initiate drill collar whirl. A motor also has the added
all bits softer than IADC 637, even when drilling as little as benefit of allowing the rotary speed of the drill string to be
4 ft (1.2m). reduced below critical values. Tests should be conducted with
• When a bit drills the Mississippi Limestone at Catoosa, XY lower speed motors to determine what operating parameters
vibrations on the order of 10 average g have not will allow a soft formation bit to drill hard sandstone without
catastrophically damaged an IADC 417 type bit, even when catastrophic damage. This offers the most immediate hope for
drilling as much as 170 ft (51.8m). a change in operating practices that allow aggressive bits to be
• Initiation of the XY vibration in hard sandstone and the applied in hard rocks.
damage it causes are very repeatable. It has occurred on Rock properties are some of the key factors governing
29 occasions, in 21 tests, on 3 different BHA’s in six catastrophic damage to bits. They are also the factors about
different wells which we know the least. A deeper understanding of what
• The occurrence of the XY vibration is not a strong function lithology and in-situ confinement states cause damage would
of bit type. All rolling cutter bits tested, seven types in all, yield valuable guidelines for bit selection and operating
ranging from IADC 417 to IADC 637 experienced XY practices operators might use to avoid catastrophic damage to
vibration. These all had similar vibration initiation rolling cutter bits.
characteristics and vibration intensities.
• Damage from the XY vibration (7 7/8”, 200mm bit) begins Since the XY vibration is generated by drill collar whirl, it is
on the inner rows and spreads outward. possible a higher lubricity mud, such as an oil base mud would
• Increasing the durability of the teeth may not be an inhibit drill collar whirl. Roller reamers might be necessary to
effective strategy for surviving XY vibration. stabilize the BHA without introducing another source of whirl as
would an integral blade stabilizer.
• XY vibration is likely to initiate when the drill collars are
buckled or when rotary speed of the drill string reaches a
Ultimately this sort of research will result in having the
critical value (due to eccentricity of the collars) high enough
capability to run more aggressive bits in harder formations;
to throw the collars out against the borehole wall.
increasing overall ROP and durability will translate into lower
• A rolling cutter bit appears to supply some forcing function
drilling costs by the operator.
which helps start drill collar whirl. There are also other
non-bit related factors, which can start drill collar whirl.
Once the process is started, it continues without the aid of
the bit. ACKNOWLEDGEMENTS
• A coupled torsional axial vibration (YZ) has not yet been
observed to cause catastrophic damage, even though its The authors of this paper wish to thank Hughes
magnitude has reached as high as 10 average g. This is Christensen Company for permission to present this paper.
repeatable and has occurred in 7 tests. Thanks to the BP Amoco Catoosa Facility for assisting in
• The YZ vibration has never occurred with high intensity in conducting vibration field tests.
the Misener Sandstone, despite numerous attempts. The
Misener sandstone tends to trigger the XY vibration REFERENCES
instead.
• Surface vibrations at the rig floor do not necessarily
correlate with vibrations downhole. The rig floor is smooth 1. Brett, J.F., Warren, T.M., Behr, S.M., “Bit Whirl: A New
when the XY vibration is occurring downhole. Theory of PDC Bit Failure,” 1990, SPE/IADC 19571, SPE
Drilling Engineering, December 1990, pp. 275-281.
 Protection,” 1999, ETCE99-6646, presented at the 1999
2. Brett, J.F., “The Genesis of Bit-Induced Torsional ETCE/ASME Conference, Houston, Texas, February
Drillstring Vibrations,” 1991, SPE/IADC 21943, presented 1999.
at the 1991 SPE/IADC Drilling Conference in Amsterdam,
March 4,1991. 10. Parfitt, S.H.L., “A Model for Shock Sub Performance
Qualification,” 1995, SPE/IADC 29354, presented at the
3. Dykstra, M.W., Chen, D.C., Warren, T.M., “Experimental 1995 SPE/IADC Drilling Conference in Amsterdam,
Evaluations of Drill Bit and Drill String Dynamics,” 1992, February 28-March 2
SPE/IADC 28323, presented at the 1994 SPE Annual
Technical Conference in New Orleans, Louisiana, 11. Shuttleworth, N.E. , Van Kerkoerle, E.J., Folmer, D.R.,
September 25-28. “Revised Drilling Practices, VSS-MWD Tool Successfully
Addresses Catastrophic Bit/Drillstring Vibrations,” 1998,
4. Dykstra, M.W., Chen, D.C-K, Warren, T.M., Azar, J.J.; IADC/SPE 39314, presented at the 1998 Drilling
“Drillstring Component Mass Imbalance: A major Source Conference, Dallas, Texas, March 1998.
of Downhole Vibrations” ; SPE/IADC 29350, presented at
the 1995 Drilling Conference, Amsterdam, February 28- 12. Warren, T.M., Oster, J.H., “Torsional Resonance of Drill
March 2, 1995. Collars With PDC Bits in Hard Rock,” 1998, IADC/SPE
49204, presented at the 1998 SPE Annual Technical
5. Jansen, J.D., “Whirl and Chaotic Motion of Stabilized Drill conference in New Orleans, Louisiana, September 27-30.
Collars,” 1992 SPE/IADC 20930, 1992 Europec 90, The
Hauge, Netherlands, October 22-24. 13. Warren, T.M., Oster, J.H.; “Shock Sub Performance Tests”
;SPE/IADC 39323, presented at the SPE/IADC Drilling
6. MacPhearson, J.D., Jogi, P., Kingman, J.E., “Application Conference; Dallas, Texas, September 27-30, 1999
and Analysis of Simultaneous Near Bit and Surface
Dynamics Measurements,” 1998, SPE/IADC 39397, 14. Winters, W.J.; Warren, T.M.; Onyia, E.C.; “Roller Cone Bit
presented at the 1998 SPE/IADC Drilling Conference in Model With Rock Ductility and Cone Offset” ; SPE/IADC
Dallas, Texas, March 3-6. 16696, presented at the 62nd Annual Technical
Conference & Exhibition, Dallas, Texas, September 27-30,
7. Mason, J.S., Sprawls, Mike ; “Addressing BHA Whirl – The 1999
Culprit in Mobile Bay” ; SPE/IADC 35034, presented at the
Annual Technical Conference & Exhibition, New Orleans,
La, March 12-15, 1996

8. McLellan, W., Hough, S., Ferguson, M., McKay, G.,

“Harding Field: A North Sea Success Story,” 1998, World
Oil, May 1998, pp. 113-120.

9. Meiners, M., Jacobsen, J., Kunning, J., “Understanding

Downhole Dynamics While Reaming Enhances Gauge
Figure 1: Partial Stratigraphic Section at Catoosa.
 Figure 2: Slick Assembly BHA Used During Catoosa Testing.

Figure 3: Surface Parameters Indicating YZ Vibration.

5 AvX
AvY
average g's

4
AvZ
3
1

0
9:35 A M 9:40 A M 9:45 A M 9:50 A M 9:55 A M 10:00 A M
T im e

Figure 4: Graph of Downhole Vibrations Indicating YZ Vibration.

 Figure 5: YZ Vibration Eliminated by Increasing WOB.

Figure 6: Surface Torque Behavior Indicates Drill Collar Whirl.

Figure 7: Onset of Drill Collar Whirl.

 4

3.5
AvX
3
AvY
Average g's

2.5
AvZ

1
1.5

0.5

0
1 0 :3 7 :0 0 1 0 :3 8 :0 0 1 0 :3 9 :0 0 1 0 :4 0 :0 0 1 0 :4 1 :0 0 1 0 :4 2 :0 0 1 0 :4 3 :0 0 1 0 :4 4 :0 0 1 0 :4 5 :0 0 1 0 :4 6 :0 0 1 0 :4 7 :0 0 1 0 :4 8 :0 0 1 0 :4 9 :0 0

T im e
Figure 8: Graph of Downhole XY Vibration from Drill Collar Whirl.

Figure 9: Vibration Magnitude Matrix in Misener Sandstone.

Figure 10: Vibration Magnitude in Mississippi Limestone.

 8

7 AvX
6 AvY
average g's

5
AvZ
4

0
9:30:00 10:00:00 10:30:00 11:00:00 11:30:00 12:00:00 12:30:00
Time
Figure 11: Graph of Downhole Vibration that Caused Catastrophic Damage to Bit.

Figure 12: Dull Bit (left) After Drilling 30 ft (9.1 m) of Misener Sandstone.
Figure 13: Shear Damage to Compact after 4 ft (1.2m) of Drilling.

Figure 14: Typical Compact Shear Failure.

 8

7 AvX Bot
6 AvY Bot
Agerage g's

5 Avz Bot
AvX Top
4
AvY Top
3
AvZ Top
2

0
1598 1600 1602 1604 1606 1608 1610
D e p th (ft)
Figure 15: Simultaneous Measurement of Vibrations Above and Below a Motor.

Figure 16: Drilling Data Using MWD Showing Downhole XY Vibration and Severity
Level Transmitted to Surface.
Figure 17: Broken Heel on 12 ¼ inch (311.2 mm) IADC 447 Tricone bit as a Result of Sever Lateral Vibrations.