Beyond Heat Checking: Frictional

Heating Causes Drillpipe Failure in an

Extended-Reach Well
R.J. Vernon, SPE, CNR Intl. (U.K.) Ltd. and E.H.R. Wade, SPE, MetalEcosse Ltd.

Summary The failure was characterized by a region of high hardness that

Failure of a 5-in. proprietary hole-cleaning drillpipe joint occurred extended 2 in. from the fracture. In this region the hardness was in
during platform drilling of a North Sea extended-reach well. The the range of 572 to 599 HV10 (54 to 55 HRC). This hardening and
joint was run in heavyweight drillpipe above the bottomhole as- the associated microstructural changes were uniform around the
sembly (BHA). Its failure resulted in the loss of the BHA, includ- circumference and through the thickness.
ing high-value rotary-steerable tools, at 23,250 ft measured depth The fracture surface on the pipe end had been damaged exten-
(MD). The failure occurred when the string was being rotated out sively after the fracture.
through a tight spot. Topdrive stallouts and intermittent circulation The laboratory concluded that the pipe had been hardened by
were experienced. The maximum overpull applied was 50,000 lbf. quenching from a temperature above 1,292F. Rotary frictional
The appearance of the recovered pipe indicated ductile failure, despite heating was suggested as the cause of heating.
a zone of abnormally high hardness adjacent to the fracture.
Investigation established that local frictional heating had heated Rig Activities at Time of Failure
the drillpipe above 1,441F and that ductile failure had occurred The time of failure was identified from drilling records. These
when hot, although the string had not been overloaded in tension or showed a coincident loss in measurement-while-drilling (MWD)
torsion. After the rupture, the mud-flow quench hardened the hot communications and an unexpected fall in the mud-circulating
pipe to above 572 Vickers of 10kg load Hardness (HV10) [ap- pressure. Immediately before these events, the string was being
proximately 54 Rockwell Hardness on a C scale (HRC)Note: rotated out through a tight spot in a 61 tangent section between
hardness values are given as measured; conversions are approxi- 23,246 and 23,261 ft MD. During this period, topdrive stall-
mate and for information only]. outs and intermittent circulation were experienced while work-
Heat checking (i.e., downhole surface hardening of friction- ing the string with a maximum overpull of 50,000 lbf. Before these
ally heated drillpipe) is an established phenomenon that is de- difficulties, rotational speeds were kept below 50 RPM, and mud
scribed in the literature. The case reported herein represents an flow was maintained at 350 gal/min. Unexplained excursions to
extreme consequence of frictional heating that has important im- 175 RPM occurred after the onset of irregular, reduced mud-
plications for the management of topdrive drilling operations. flow rates. Logs indicate that the rotational speed was below 90
RPM the time of failure, which occurred approximately 500 ft
Introduction above the bit.
Two indicative calculations were performed to assess stresses
A drillpipe failure occurred during the drilling of an extended-
at the failure in the 5-in. body pipe under rotating-out conditions:
reach well from a North Sea platform. The intended total depth
If the full reported maximum overpull of 50,000 lbf is ap-
was 26,000 ft MD at 12,800 ft true vertical depth (see Fig. 1). The
plied to the 5-in. pipe, the nominal tensile stress is 7.1 ksi. This is
failure occurred when the well had been drilled to 25,468 ft MD
only 6% of the room-temperature yield strength of 124 ksi given
with an 8.5-in. bit. This section was drilled with a topdrive and
in Table 2.
12.8-lbm/gal oil-based mud. While rotating out, a joint of a pro-
Similarly, if 10% of the 45,000 ft-lbf reported stallout torque
prietary 5-in. drillpipe failed when the bit was at 23,250 ft MD in
is applied to the 5-in. section, the maximum shear stress is 466 psi,
a Kimmeridge clay interval. The joint was one of five, run with the
or 0.7% of the room-temperature yield strength in shear. (Note:
heavyweight drillpipe immediately above the BHA. These joints
10% is an arbitrary value that recognizes that maximum torque is
were equipped with integral, spiral-bladed, hole-cleaning pro-
applied at the surface, and that torque at the failure location was
files. Details of the BHA and the failed joint are given in Table 1
much lower. The outcome of the calculation shows the accuracy of
and Fig. 2, respectively. The failure resulted in the loss of the
the assumption to be unimportant.) On this basis, the cause of
rotary-steerable BHA. An understanding of the cause of failure
failure was not a simple mechanical overload under normal down-
was required to justify further expenditure on this well and future
hole conditions.
similar wells.

Heating and Hardening

Examination of Failed Pipe
The basic metallurgy of the thermal cycle required to harden the
Rig photographs of the parted joint are shown in Fig. 3. These
AISI 4145 drillpipe is undeniable. The high and uniform hardness
show the outside diameter (OD) of the pipe tapered smoothly
and the microstructure reported could be obtained only by heating
toward the fracture. The pipe parted in the 5-in. body pipe ap-
the steel sufficiently to transform it to the austenite phase, followed
proximately 94 in. above the pin end, as marked in Fig. 2.
by rapid cooling. The pipe manufacturers data, given in Table 4,
A laboratory examination of the recovered piece identified the
show that during heating, the ferrite-to-austenite transformation
base pipe as a quench-and-tempered AISI 4145 low-alloy steel
occurs between 1,375 and 1,441F. Rapid cooling below 410F is
with an average hardness of 315 HV10 (32 HRC). This was con-
then required to achieve the complete martensite transformation
sistent with the mill-certified composition and properties given in
observed adjacent to the failure. During manufacture, the same
Tables 2 and 3.
austenitize-and-quench sequence was used to harden the pipe be-
fore tempering.

Copyright 2004 Society of Petroleum Engineers. Downhole Frictional Heating. A search of the SPE literature for
Original manuscript received for review 20 August 2003. Revised manuscript received 16
causes of drillpipe heating found three references describing heat
September 2004. Paper 86562 peer approved 28 October 2004. checking.13 Heat checking is the creation of superficial hard spots

December 2004 SPE Drilling & Completion 249

 Fig. 1Well profile.

on tool joints (and other parts) that experience rotary frictional On this basis, frictional heating, when stuck rotating out with
heating downhole. The heating and cooling cycle is as described interrupted circulation, was judged capable of heating the pipe to
earlier for hardening low-alloy steels. Heat checking is of concern a temperature (above 1,292F) at which it would have little use-
because the hard spots created may crack and cause failure. The ful strength and exhibit ductile failure at low applied loads. After
SPE papers identify side loads in doglegs as the primary cause of the failure, restored mud circulation rapidly cooled and locally
frictional heating. Simple calculations show that high side loads hardened the pipe. The extent of hardening indicates that the pipe
are easily capable of causing sufficient heating. Cooling (to circu- was heated above 1,441F; at this temperature, it would have been
lating mud temperatures) then causes hardening. red hot.
Key points from the references are as follows:
The frictional heat energy generated is proportional to side Hardening During Manufacture. The possibility that the abnor-
load, coefficient of friction, joint OD, and rate of rotation and mal hard zone originated from the manufacture of the drillpipe was
duration. A sample calculation is provided by Eaton.2 considered. The manufacturing process involved quench-and-
The use of topdrives to rotate out of tight intervals and key- temper treatments, tool-joint welding, and post-weld heat treat-
seats is identified as a growing concern that can result in frictional ment. The mill certificate for the central part of the failed joint
heating of drillstring components. shows that it was water quenched from 1,607F and tempered
Fig. 5 in the paper by Altermann and Smith1 shows a failure at 1,130F.
that appears to be similar to the one currently reported. It occurred Manufacturing is not considered a realistic explanation for the
while using a topdrive to rotate the BHA up through a bridge. final hard condition of the failed joint. The primary reason is that
the failure was essentially ductile. The pipe was locally, smoothly
The failure currently reported occurred in a tangent section of necked (by approximately 0.6 in. in diameter) with corresponding
the hole where side loads resulting from doglegs should have been wall thinning. The smoothness of these features would have been
low; however, the resistance encountered while rotating out of the disturbed had the hard zone been present at the time of failure.
hole indicates that other causes of frictional loading exist. Also, a brittle failure from the hard zones would have been more
The calculation given by Eaton2 was used as a basis for esti- likely than a ductile overload.
mating the conditions required to cause sufficient frictional heating Additionally, nothing in the manufacturing history indicates
to enable the observed hardening of the recovered pipe. A sum- hard zones in the as-supplied pipe. In particular, the 5-in.-OD
mary is given in the Appendix. This indicates that (ignoring heat section of pipe that failed was machined from a 7.07-in. mother
losses) local friction equivalent to 10% of the reported stall- tube after final heat treatment. A rogue tube with hardness of 599
out torque could provide the necessary heat in less than 1 minute. HV10 (55 HRC) would not be expected to go unnoticed during
These conditions would not be detectable as abnormal on the machining. Noise, tool damage, and a local change of size and
drill floor. surface appearance would all attract attention.

250 December 2004 SPE Drilling & Completion

 Fig. 2Schematic of failed joint.

Subsequent Drilling Practice At this temperature, when red hot and above conventional
Modified drilling practices were subsequently implemented to stress-relief temperatures, the pipe had no useful strength in
sidetrack around the lost BHA and drill the well to the planned tension or torsion.
depth. The modified practices included elimination of the propri- 3. Ductile-overload failure resulted from a combination of the lo-
etary hole-cleaning joints and precautions to prevent extreme fric- cal tensile and torsional loading where the steel temperature was
tional heating. Guidance was given to avoid high rotary speeds greatest and its strength, therefore, least. Only trivial loads were
whenever the drillstring was packed off in cuttings beds until full required to cause failure at or above 1,441F.
circulation had been re-established. The previous approach of try- 4. Heating probably occurred when circulation was interrupted.
ing to back-ream through tight spots was discontinued. Then, separation of the joint restored mud flow, causing rapid
cooling and local hardening of the drillpipe immediately above
the failure.
Conclusions 5. Subsequent working of the string caused mechanical damage to
The circumstances of the reported failure cannot be determined the hardened free end during contact with the separated part.
precisely, but the drilling records and the metallurgical examina- The original fracture face was thus modified by wear, cracking,
tion indicate the following: and material loss before the laboratory examination.
1. Frictional heating of the drillpipe occurred in the tangent section 6. Frictional heating may cause rapid drillpipe failure when stuck
of a deviated hole while stuck when rotating out with intermit- pipe is back reamed with intermittent circulation.
tent circulation and rotational speeds up to 90 RPM.
2. Heating above 1,441F occurred local to the point of failure:
Acknowledgments
Without cooling, this could occur in less than 1 minute with-
out causing abnormal indications on the drill floor. The authors thank CNR Intl. (U.K.) Ltd. for permission to publish
A temperature above 1,441F was required to achieve the this paper. The initial laboratory investigation was performed by
uniform microstructure and high hardness observed. OIS plc, Aberdeen, U.K.

Fig. 3Rig photographs of failed drillpipe; two views of the 5-in. section of the drillpipe as retrieved.

December 2004 SPE Drilling & Completion 251

 In the present case, doglegs were not a concern, but stallouts
and intermittent circulation indicate the potential for frictional
heating and poor cooling. For the 5-in.-OD pipe, rotating at 75
RPM, the frictional heat that could be generated by 10% of the
45,000-ft-lbf stallout limit was 47,125 ft-lbf/sec (86 hp). This is 33
times the rate given previously for dogleg side loading. With no
heat loss, this rate of heating would raise the temperature of a
12-in. length of the 5-in.-OD pipe from 212 to 1,441F in 57
seconds. A 12-in. length is six times the hardened length recovered
from above the fracture.

SI Metric Conversion Factors

References bbl 1.589 873 E01 m3
1. Altermann, J.A. and Smith, T.B.: Heat Checking/Quench Cracking ft 3.048 E01 m
Tool Joints, paper IADC/SPE 23846 presented at the 1992 IADC/SPE ft-lbf 1.355 818 E03 kJ
Drilling Conference, New Orleans, 1821 February. ft-lbf/s 1.355 818 E03 kW
2. Eaton, L.F.: Tool Joint Heat Checking While Predrilling for Auger F (F 32)/1.8 C
TLP Project, paper SPE/IADC 25776 presented at the 1993 SPE/ gal 3.785 412 E03 m3
IADC Drilling Conference, Amsterdam, 2325 February. hp 7.460 43 E01 kW
3. Choi, H.J.: A Deeper Understanding of Heat Checking, An Old But in. 2.54* E+00 cm
Frequently Overlooked Failure Mechanism, of Downhole Drilling ksi 6.894 757 E+03 kPa
Components, paper SPE/IADC 37651 presented at the 1997 SPE/ lbf 4.448 222 E+00 N
IADC Drilling Conference, Amsterdam, 46 March.
lbm 4.535 924 E01 kg
4. Rotary Drill Stem Elements, second edition, API Spec. 7 (March 2002). psi 6.894 757 E+00 kPa
*Conversion factor is exact.
AppendixSummary of
Frictional-Heating Calculations
Uncertain heat losses make meaningful calculation of potential Roger Vernon joined the oil industry in 1972 as a trainee driller
frictional-heating rates difficult. The rate of heat generation is with Shell Intl. He has subsequently worked for major operators
proportional to the frictional torque and the speed of rotation. and drilling contractors in many parts of the world, playing a
Some indication of relative heating rates can be obtained by com- leading role in offshore drilling operations. Now based in Ab-
parison with the calculation presented by Eaton.2 He calculated a erdeen, he works as a Consultant Drilling Superintendent, spe-
heating rate of 1,423 ft-lbf/sec (2.6 hp) for a 7.25-in.-OD tool joint cializing in cold-start subsea projects. He holds an MSc degree
rotating at 150 RPM with a side load of 2,000 lbf from steel-on- in drilling engineering from the Robert Gordon U. Ed Wade is a
consultant metallurgist and corrosion engineer based in Aber-
steel contact in oil-based mud. The calculation was made at the deen. He worked for Marathon Oil U.K. Ltd. for 13 years before
maximum dogleg side load then recommended by API RP7G for becoming a consultant in 1996. Since then, he has worked for
avoidance of heat checking. Without heat loss, the heat generated MetalEcosse Ltd., specializing in oilfield materials. He holds a
was sufficient to heat 1 lbm of steel to 1,888F in 2 minutes. BSc degree in metallurgy from Nottingham U. and a PhD de-
Obviously, heat loss is critical for the prevention of heat checking. gree in mechanical engineering from Edinburgh U.

252 December 2004 SPE Drilling & Completion