Comparison of Field and Laboratory-Simulated Drill-Off Tests

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Comparison of Field and

Laboratory-Simulated Drill-Off Tests


...·c. Bourdon, *
Sedco-Forex Technical Centre; G.A. Cooper, * * SPE, D.A. Curry, t SPE, and
D. McCann,* SPE, Schlumberger Cambridge Research; and B. Peltier,! Sedco-Forex Technical Centre

Summary. Field drill-off test results are compared with data from laboratory simulations. A simple theory for analyzing drill-off tests
is developed. The weight-on-bit (WOB) decay with time is close to exponential, but large threshold WOB's, resulting from poor weight
transmission downhole, are sometimes observed in field tests.

Introduction
The drill-off test was devised nearly 30 years ago as a practical normally be run with 5-in. [12.7-cm] drillpipe. The compliance of
means of estimating and optimizing drilling performance, but its the bottomhole assembly (BHA) is negligible compared with that
use in the field has been limited. Two factors seem to limit its value: of the drillpipe. Therefore, the following equation gives the length
the quality of data recorded with conventional rig equipment and of formation drilled:
uncertainty in data interpretation. A joint program was undertaken
OLb "'" 6.2 X 1O-9Ldp oFh . ............................ (2)
by Sedco-Forex and Schlumberger Cambridge Research to address
these limitations and hence to increase the information actually ob- The rate of change of WOB with depth is
tained from a drill-offtest. The work involved two aspects: the de-
oFhloL b =1.6X 1081Ldp ' . ., . ., .. .,.,.,.,., . ., . .,.,., (3)
velopment and implementation of a rig-site drill-off test data
acquisition system and the laboratory simulation of drill-off tests. Initially regarding 5,700 lbf/in. [1 kN/mm] as a reasonable up-
This paper describes the progress to date, concentrating on the com- per limit for the rate of change of WOB and assuming 600 ft [200
parison of field and laboratory drill-off test data and on the resul- m] of drill collars, we calculated a minimum drillstring length of
tant improvements in our interpretation of drill-off tests. 3,000 ft [1000 m]. Thus, the maximum length of formation drilled
during the drill-off will be 8 in. [20 cm]. As a result of these calcu-
Procedure lations, we decided that it was unlikely that valuable information
The drill-off test is a simple, practical procedure, proposed by Lu- would be derived from drill-off tests performed at depths of less
binski,1 for determining the relationship between rate of penetra- than 3,000 ft [1000 m], and we restricted our investigations ac-
tion (ROP) and surface WOB. The driller builds WOB to a cordingly.
predetermined maximum and then sets the brake. As the bit drills,
Interpretation Theory
the WOB decays at a rate determined by the ROP and the drill-
string compliance. The surface WOB is recorded as a function of Laboratory and field studies 4 have shown that it is reasonable, in
time. The drillstring compliance is determined from a knowledge many circumstances, to consider the ROP, R, proportional to the
of the drillstring composition and is used with the recorded WOB WOB, W. Thus,
to compute the distance drilled as a function of time. Differentia- R=KWf(v). . ..................................... (4)
tion yields the ROP as a function of time, and crossplotting gives
the dependence of ROP on WOB. Ref. 2 describes an alternative During a drill-off test the WOB is linked to the bit ROP by the
procedure for the analysis of drill-off test data based on Vidder' s3 elasticity of the drill string .
procedure. The times taken to drill off successive increments of -oR=oFh =EAdp(oLhILdp)' ......................... (5)
4,000 lbf [18 kN] WOB are recorded, and the mean ROP over each
increment is computed from the drillpipe compliance. (Note that aw EA dp aL h EA dp
length changes in tool joints and drill collars can be neglected.) Hence, - = - - - - - = - - - K W f ( v ) . .,.,.,.,., .,(6)
A log-log plot of ROP against WOB reveals the exponent relating at L dp at L dp
ROP to WOB. If there is a significant threshold WOB-i.e., a WOB Integration gives the following expression for WOB as a func-
below which the bit will not drill-then the exponent can be deter- tion of time during a drill-off test.
mined by subtracting different threshold WOB's from the record-
ed values until a straight-line fit is obtained. Regardless of which W = Wie( - EAdpILdplKf(vlt. . .......................... (7)
analysis is adopted, variations in rock drillability and oscillations We found in some cases that the WOB decays not to zero, but to
in the hook load can combine to restrict the meaning of the derived a finite offset or threshold WOB, WOo If this threshold WOB is in-
relationship between ROP and WOB unless some method of smooth- troduced into the expression relating ROP to WOB, the depend-
ing the raw data is used. The next section presents one way of over- ence of WOB on time becomes
coming the problem.
When a drill-off test is designed, both the distance drilled and Ws = ~o + Wier - EAdplLdplKf(vlt
the rate of change of hook load must be considered if the test is =W + Wie- kt . ................................. (8)
o
to generate results that are representative of the formation being
drilled. Assuming that no change occurs in hydraulics during the Three valuable pieces of information can be obtained by fitting this
drill-off, the change in length of the drillstring is related to the expression to WOB data recorded during a drill-off test: the quali-
change in WOB by ty of fit, which indicates whether the assumption of linearity be-
tween ROP and WOB is valid; the value of Wo , whose significance
is discussed below; and the value of K, which characterizes the ROP.
oL b =!:.. OFh =(LdC + L hw + LdP ) OFh . ............. (1)
Note that the curve-fitting scheme adopted to determine Wo is
A E Adc Ahw A dp E equivalent to the trial-and-error procedure previously mentioned 3
With an 8\h-in. [21.6-cm] roller-cone bit, the maximum change because both rely on finding the value of Wo that gives the. best
in WOB is restricted to about 45,000 lbf [200 kN], and the bit would fit between the data and the assumed relationship between ROP and
WOB. Here, however, we work with data recorded at a much higher
• Now at Foramer/Forasol. frequency and consequently gain enhanced resolution .
"Now at the U. of California.
tNowat IntI. Drilling and Downhole Technology Centre. Experimental Details
*Now at Sedco·Forex Technical Centre.
§Now at Schlumberger Cambridge Research. More than 50 drill-off tests were performed in vertical wells and
Copyright 1989 Society of Petroleum Engineers in wells deviated up to 40°. All were on land in regions with well-
SPE Drilling Engineering, December 1989 329
TABLE 1-DRILLING PARAMETERS RECORDED
DURING FIELD DRILL-OFF TESTS
140
Bit: 6" 5-3-7
Parameter Sensor RPM: 80
120
Hook load Gauge on dead-line anchor
Traveling block altitude Encoder on fast sheave of crown
block
Rotary speed Proximity detector in Z
:!.
transmission :iic:: 80- Beginning of
Drill-off
Mud flow rate Proximity detector in 0
transmission E 60
Standpipe pressure
Rotary torque
Pressure gauge on standpipe
Ampmeter on DC motors f
40

characterized lithologies. Tests were carried out with 8 1h- and 6-in. 20
[21.6- and 15.2-cm] -diameter roller-cone bits at depths ranging
~rom 3,650 to 6,700 ft [1100 to 2050 m]. Mechanical and hydraul-
~·5±0~0--~MO~0--~8~700=---~88'00-----89~00-----90~00-----9T'00--­
IC parameters were monitored during field drill-off tests wjth ex-
Time(s)
isting and specially installed surface rig sensors, and their values
were recorded with a microcomputer-based data acquisition sys-
tem. Table 1 lists the drilling parameters that were recorded and Fig. 1-WOB recorded during field drill-off test.
the corresponding sensors. Special attention was given to the de-
sign of the data acquisition system to ensure good resolution and Results
to avoid aliasing effects. Once recorded, the field drill-off test data
were transferred onto our engineering computer for reformatting Field Data. Two tests were performed with a 6-in. [l5.2-cm] in-
calibration, and analysis. ' sert bit, IADC Code 5-3-7, at a depth of 6,861 ft [2091 m] in a
The design, specification, and function of the drilling test machine well de~iated by about 15°. The formation was predominantly lime-
(DTM) are described in Ref. 5. The test control computer attached stone with some shale layers, and an oil-based mud was used. The
to the DTM is used to simulate drill string elasticity and hence to initial WOB was 29,2251bf [130 kN] in both cases, and the rotary
perform a drill-off test in the laboratory. An initial WOB and the speed was 80 rev/min. Fig. I shows the WOB recorded during one
required drillstring compliance are specified. The change in pene- test as a function of time, together with the best-fit exponential curve.
tation depth from that at the start of the drill-off test is monitored, The agreement between the data and the fitted curve is clearly very
and the corresponding reduction in WOB is computed. This is sub- good. Correlation coefficients of 99.91 % were given for both tests.
tracted from the initial WOB to update the demanded weight, and Threshold WOB's, wo , of2,855 and 2,520 Ibf[12.7 and 11.2 kN]
the process is repeated at a frequency of 50 cycles/sec [50 Hz] were found, and the corresponding drilling response coefficients,
throughout the test. k, were 1.68 and 1.51. Fig. 2 shows a crossplot ofROP vs. WOB.
All laboratory drilling tests were performed with 8Vz-in. Its apparent form supports the relationship between ROP and WOB
[21.6-cm] -diameter roller-cone bits. Several rock types were assumed in the drill-off test analysis. Fig. 2 also shows the ROP
drilled, including Bolton Wood sandstone, Carrara marble, Port- computed during steady-state drilling immediately before the two
land limestone, and a limestone aggregate concrete. A low-solids drill-off tests. It is within 5% of the ROP indicated by the drill-off
~ests for that WOB. This discrepancy is less than the uncertainty
bentonite/lignosulfonate mud was used throughout the tests. The
In the computed drill string compliance and suggests that there are
sample stresses were an overburden of 3,625 psi [25 MPa], a con-
fining pressure of2, 175 psi [15 MPa], and a pore pressure of 14.5 no large transient effects on the drilling response during drill-off
psi [0.1 MPa], with an annular mud pressure of 1,450 psi [10 MPa]. tests.
The rotary speed was 60 rev/min, and the mud flow rate was 330 . A number of drill-off tests were performed in a vertical well with
gal/min [1250 Llmin]. Three 1O/32-in. [7.9-mm] -diameter noz- an 8V2-in. [21.6-cm] IADC Code 1-3-5 bit at depths ranging from
zles were used. A drillstring compliance equivalent to that of 4,570 3,719 to 4,023 ft [1134 to 1226 m]. The formations over this in-
ft [1400 m] of 5-in. [12.7-cm] drillpipe was simulated. Unless other- terval were mainly limestones with some interbedded shales. Three
wise stated, these standard conditions applied for all tests described tests were run immediately before the bit was pulled; these all
in this paper. Several rock samples were drilled with bits that had showed WOB offsets of about 6,750 lbf [30 kN]. Severe overpulls
been used for field drill-off tests. In these cases, the drillstring com- were experienced during the early stages of the subsequent trip-out.
p~iance .computed for the field tests was adopted for the laboratory
simulatIOns, and the laboratory rotary speed (80 rev/min) and Laboratory Simulations. The laboratory drill-off tests all showed
hydraulics matched those for the field tests. more or less exponential decay of WOB with time. The form of

200

Bit: 6" 5-3-7 Bit: B1A?" 1-3-6 (TO)


RPM: 80 Rock: Carrara marble
ROP from exponential fit
I
z
~ 120
~
5
E
'"
~ 80

40

otO--~~-'30~-----'6'0-------9~0-------'~20------~'50 ~j50~------~1000~------~1~'5~0------~'3~OO--------,T45-0~
Weight on Bit (kN) Time{s)

Fig. 2-ROP as a function of WOB during field drill-off test. Fig. 3-WOB recorded during laboratory drill-off test.

330 SPE Drilling Engineering, December 1989


Bit: B1f:;!" 1-3--6 (TO)
Rock: Carrara marble
TABLE 2-LABORATORY DRILL-OFF TEST PARAMETERS
..,
k Wo =
§. 6
Bit> (seconds -1)
Rock ~ c
0
Carrara marble 1-3-6 TO 0.0044 -3.2 '"t
Limestone concrete 1-3-6 T7 0.0050 -5.6 c
4
Portland limestone 1-3-6 TO 0.0053 -11.8 "c.
'0
Portland limestone 1-3-5 T8 0.0056 4.3
Portland limestone 1-3-5 T8 0.0022 15.8 ~
II:
Portland limestone 1-3-5 T8 0.0060 9.7
Portland limestone 1-3-5 T8 0.0112 4.8
Portland limestone 5-1-7 T1 0.0124 20.3
Portland limestone 5-1-7 T1 0.0218 10.5
Portland limestone 5-1-7 T1 0.0106 19.0 50 100 150 200 250

Portland limestone 5-1-7 T1 0.0311 10.4 Weight on Bit (kN)

Portland limestone 5-1-7 T1 0.0155 0.9


• lADe code, Fig. 4-ROP as a function of WOB •

WOB decay was not strongly influenced by test conditions, such sample that corresponded with the departure from linearity in
as the bit type, bit wear state, or fluctuations in rock drillability. penetration rate. The ROP computed from the coefficients of the
Fig. 3 is an example of the laboratory results from a simulated drill- exponential fit gives a reasonable representation of the average drill-
offtest on Carrara marble with an IADC Code 1-3-6 bit in TO-BO-I ing response over this interval of rapidly varying drillability, but
condition. It shows the filtered WOB plotted as a function of time the quality of fit clearly is not as good as it is in more uniform rock
and the exponential curve fitted to these data. Table 2 gives the samples.
drill-off test parameters determined for this test and for other labo- Five simulated drill-offs with a field-worn insert bit, IADC Code
ratory tests. Fig. 4 shows the ROP vs. WOB relationship comput- 5-1-7, were run in Portland limestone, again covering a variety of
ed for the marble test and the corresponding reduced ROP data mud flow rates and BHP's. These gave Wo values that ranged from
determined during steady-state drilling of the same rock sample. 200 to 4,560 lbf [0.9 to 20.3 kN] with a mean of 2,740 lbf [12.2
The relationship between ROP and WOB was clearly the same dur- kNJ. In one rock sample, the results of two drill-offtests with simu-
ing the drill-off test as during steady-state drilling in this imperme- lated drillstring compliances equal to those of 2,225 and 4,450 ft
able rock. Furthermore, the relationship computed from the [678 and 1356 m] of 5-in. [12,7-cm] pipe were compared with re-
exponential fit is very close to the directly measured relationship. sults of steady-state drilling. Fig. 7 shows the ROP plotted against
A laboratory drill-off with a 1-3-6 bit in T7 condition also showed WOB. The drill-off and steady-state data are clearly similar, with
good agreement between the fitted exponential and the observed data from the shallower drill-off giving an ROP slightly lower than
WOB decay. Note the lack of any large threshold WOB with this that of steady-state drilling.
well-worn bit (see Table 2). Again there was no detectable differ- To investigate the influence of hydraulics on drill-offtests, three
ence between the ROP's at any given WOB during the drill-offtest successive tests were run in Portland limestone with a new lADC
and during steady-state drilling. Code 5-2-6 bit. The first was run under standard conditions, the
The 1-3-5 bit with which field drill-off data were gathered was second with a much reduced mud flow rate of 92 gal/min [350
used for four laboratory drill-off tests in Portland limestone, with Llmin], and the third with the full mud flow rate but a reduced
various mud flow rates and bottornhole pressures (BHP's). Table differential pressure across the workfront-Le., 435 psi [3.0 MPa]
2 lists the resultant drill-off parameters. as opposed to 1,450 psi [10 MPa]. Fig. 8 shows both the measured
Fig. 5 shows the decay of WOB with time for a drill-off testTn and the computed ROP's as functions of the WOB. As expected,
Portland limestone. The exponential curve seems to give a good the highest ROP for any given WOB was observed at the low dif-
fit to the data, but some divergence between the two sets is notice- ferential pressure and the lowest ROP was observed at the low mud
able above - 13,500 lbf [60 kN] WOB. When the measured ROP flow rate. For all three drill-off tests, the relationship between ROP
for this test is plotted as a function of WOB, it becomes clear that and WOB was not linear. Instead, all three tests indicated a con-
their relationship is nonlinear (see Fig. 6). Coring and mechanical tinuously rising gradient of the ROP/WOB relationship, with no
testing subsequently established the presence of large local fluctu- evidence of any reduction in gradient with increasing weight. The
ations in strength (2,200 to 10,150 psi [15 to 70 MPa]) in this rock effect of poor hydraulics was apparently to reduce the ROP by a

2~.----------------------------------------;
Bit: aw'1-3-6 (TO)
BIt: BW' 1-3-6 (TO) Rock: Portland limestone
Rock: Portland limestone 16
200

Z
.!. 150 Measured RQP
iii J
C
0
:E
'"
';; 100
~
....... ROP from exponential fit

50

O+-~----'-------'-------~-------r------~
o 50 100 150 200 250

0 Weight on Bit (kN)


1300 1400 1500 1600 1700
Time(s)
Fig. 6-ROP as a function of WOB in limestone sample with
Fig. 5-WOB recorded during laboratory drill-off test of lime- large strength fluctuations.
stone sample with large strength fluctuations.

SPE Drilling Engineering, December 1989 331


20

25 Bit: 8W'5-2-6

Bit:8112"5-1-7
RPM,80
•• 16
Rock: Portland limestone

'l:"
Rock: Portland limestone
'l:"
~
20
___ Measured ROP
ic
1c
t
ROP from exponential fit 12

.2 15
e
;;
c
!
c '0
!
~
10
'0 a:
~
a:

0
0 50 250
O~C-----.-------~------~------~------~
o 50 100 150 200 250 Weight on Bit (kN)

Weight on Bit (kN)

Fig. 8-ROP as a function of WOB at different mud flow rates


Fig. 7-ROP as a function of WOB. and BHP's.

fixed proportion irrespective of the WOS in a manner that would Winters and Warren 7 showed that natural diamond bits display
seem analogous to an increase in rock strength. The same effect threshold WOB's as a result of "pumpoff" effects. The mud pres-
was observed in other tests with both milled-tooth and insert bits. sure drop across the bit causes significant upward forces that must
be overcome before the bit face can be brought into contact with
the rock. The return flow area available with roller-cone bits, typi-
Laboratory vs. Field Drlll·Off Tests cally 15 % of the total bit area, is much greater than that on natural
The first point to emerge from this study is that the drilling response diamond bits. Consequently, pumpoff effects are likely to be negligi-
during a drill-off test is similar to that during steady-state drilling. ble with roller-cone bits, so any threshold WOB cannot be attributed
The laboratory data indicate clearly that there are no significant to this cause.
differences when impermeable rock is drilled. Small differences Further support for the proposal that the offset WOB corresponds
were observed in the shallow simulated drill-off test of Portland to the weight lost between the surface and the bit is provided by
limestone (permeability about 10 md) when the WOS was decreas- the effect of changing hydraulics during drill-off tests. Appendix
ing rapidly-i.e., when transient effects were at their greatest. A A shows that the hook-load variation accompanying changes in mud
theoretical analysis of transient effects on pore-pressure elevation flow rate will depend on the boundary condition of the bottom of
resulting from dynamic filtration beneath the bit was presented by the drillstring. The mud flow rate was deliberately varied near the
Peltier and Atkinson. 6 This indicates that the pressure gradients end of a drill-off test for which a threshold WOB of 12,360 Ibf
will depend on the prevailing ROP and that changes in the ROP [55 kN] was determined. Fig. 9 shows the hook load and mud flow
will be followed by a transient period before the pressure gradients rate as functions of time during the test.
attain their new steady-state values. During a drill-off test, the WOS If the drillstring were freely hanging, then the hook-load varia-
is continuously decreasing so that under some conditions the ROP tions would have resulted from nozzle-jet-reaction changes, which
might be expected to be less than that for steady-state drilling at were calculated to be at most 5S5 Ibf [2.6 kN]. The observed vari-
the same WOS. The experimental evidence is inconclusive here; ations in hook load were clearly much larger than this. The hook-
the only disparities observed between drill-off and steady condi- load changes for the case of a fixed drill string were calculated with
tions were not large in comparison with rock variability. Further Eq. A-S. The calculated and measured values are given in Table
experiments are planned to pursue this point. 3, where the letters A, B, C, and D refer to the points on Fig. 9.
Overall, encouraging similarity was found between data gathered The calculated and measured values are obviously in good agree-
during field drill-off tests and those from our laboratory simula- ment, suggesting strongly that the bottom of the drillstring was fixed.
tions. There was one significant difference, however, that concerned This confirms our interpretation that the occurrence of a large-offset
the value of the threshold WOS indicated by the exponential fit. WOS, wo ' indicates weight loss between the surface and the bit.
The laboratory simulations showed modest Wo values that could The magnitude of Wo could be a valuable indicator of hole condi-
be either positive or negative. These seem to reflect nonlinear ROP tion and could even provide early warning of stuck drillstring
response to WOS because of the bit action, particularly insert-bit problems.
action, and because of fluctuations in rock drillability. Interesting- It is generally held that there is an optimum WOS above which
ly, the milled-tooth bits did not give large threshold WOB's when the ROP of a roller-cone bit increases only slowly with increasing
worn even to TS condition. In the field, however, some drill-off weight. This WOB corresponds to the onset of bit floundering-
tests gave positive threshold WOB's that were significantly larger i.e., when the cones roll either on the workfront directly or on a
than any observed in the laboratory. The occurrence of a large bed of cuttings, which prevents the attainment of higher ROP's.3.8
threshold WOB is interpreted as indicative of poor weight transfer Bit floundering and the consequent insensitivity of ROP to WOB
downhole rather than tooth wear. This explanation is supported by should be manifest during a drill-off test as an initial WOB decay
the observation of overpulls when tripping immediately after drill- rate less than that observed subsequently. No examples of this were
off tests that showed significant threshold WOB's. seen, either in the field drill-off tests or in the laboratory simula-

TABLE 3-HOOK-LOAD CHANGES RESULTING FROM HYDRAULIC EFFECTS

OPsp OPdP Error


Interval (MPa) (MPa) Computed Measured (%)
AB -7.6 -0.7 33.0 35.5 7.6
BC -3.6 -0.3 15.4 16.5 7.2
CD 10.3 1.0 -45.0 -34.0 -24.6

332 SPE Drilling Engineering, December 1989


tions. In the case of drill-off tests performed with poor bit hydraulics,
the lack of any obvious optimum weight is slightly surprising, par-
ticularly because significant reductions in ROP accompanied the ~
reduced hydraulics. The obvious conclusion here is that a drill-off ~
test will not necessarily show an optimum WOB even if the ROP _0
I:~
is limited by poor hydraulics. Therefore, repeated drill-offtests with e:! zm 0

~
different hydraulics are needed to determine the adequacy of the
hydraulics:
.
1ii
~o
~

"~
Both the field and the laboratory data show that a good represen- ~g
0- "'~
o~
0: 0"
tation of a drill-off test can always be obtained by use of the ex- :I:

ponential relationship indicated by the simple theory discussed ""


::Ii

earlier. Fitting an exponential curve to the WOB recorded during ~ $


a drill-off test provides a representative average drilling response
for the rock and bit in question that is not sensitive to fluctuations
in drillability during the drill-off test. This procedure seems to over- § ~
7740 7800 7860 7920 7980 6040
come some of the data-reduction problems previously associated Time(s)
with the drill-off test. It would be a simple matter to use the drill-
off-indicated average drilling response in conventional drilling op- Fig. 9-Hook load and mud flow rate during field drill-off test.
timization routines with the incidental benefit that the weight loss
between the surface and downhole is indicated and automatically
incorporated into the ROP analysis.
All drill-off tests reported here were performed with soft forma-
Po = external pressure
tion bits drilling rocks of low and medium compressive strengths Psp = standpipe pressure
when the assumption of a linear relationship between ROP and WOB !:J.PbOO = pressure drop in bit and BHA
was clearly justified. It is possible that, in some cases, the ROP !:J.p dp = pressure drop in drillpipe
might not be directly proportional to WOB. 4 ,9 Then, if the form q = mud flow rate
of the ROP response were known, it would be, in principle, straight- ri = inner radius
forward to substitute it for Eq. 4 in our analysis to derive an ex- r0 = outer radius
pression for the decay in WOB that is more accurate than Eq. 8. R = ROP
This could be used to analyze the drill-off test data in a manner t = time
directly analogous to that adopted here with the exponential WOB v = rotary speed
decay. W=WOB
Wi = initial downhole WOB
Conclusions Wo = offset WOB
I. With good sensors and data logging, the drill-off test can give Ws = surface WOB
valuable information about drilling conditions. €z = axial strain

2. The WOB decays during a drill-off test in a more or less ex- JL = Poisson's ratio
ponential manner. The exponential decay constant determined from p = mud density
the recorded WOB data gives a good representation of the drilling a r = radial stress
response of the rock and bit in question. az = axial stress
3. No significant transient effects on ROP response were observed a 0 = hoop stress
in either field drill-off tests or laboratory simulations.
4. The presence during a drill-off test of a large threshold WOB
below which the bit does not drill indicates poor weight transmis- Acknowledgments
sion between the surface and the bit, and could provide a useful We thank Esso REP for permission to carry out the field experi-
guide to hole condition. ' ments necessary for the development of this project. We also thank
5. Poor bit hydraulics that lead to reduced ROP will not neces- the Sedco-Forex and Schlumberger Cambridge Research staff whose
sarily be revealed by the presence of an optimum WOB during a assistance and cooperation made this work possible.
drill-off test.
References
Nomenclature
1. Lubinski, A.: "Proposal for Future Tests," The Petroleum Engineer
A = effective steel section area of drill string (Jan. 1958) B50-B52.
A de = steel section area of drill collars 2. Bourgoyne, A. T. et al.: Applied Drilling Engineering, Textbook Ser-
A dp = steel section area of drillpipes ies, SPE, Richardson, TX (1986) 227-29.
A g = internal section area of gooseneck 3. Vidder, A.: "Chevron Drill-Off Test (DOT)," Chevron Oil Co., New
Ahw = steel section area of heavy weights Orleans.
4. Speer," J. W.: "A Method of Determining Optimum Drilling Tech-
Ai = internal section area of drillpipes
niques," Drill. & Prod. Prac., API, Dallas (1958) 130-47.
An = bit-nozzle area 5. Cooper, G.A. and Peltier, B.: "Advanced Techniques for Laboratory
E = Young's modulus Full-Scale Drilling Tests," paper SPE 14783 presented at the 1986
f = function of rotary speed in drilling model IADC/SPE Drilling Conference, Dallas, Feb. 10-13.
F = drillpipe tension 6. Peltier, B. and Atkinson, C.: "Dynamic Pore Pressure Ahead of the
F boo = total force between formation and BHA Bit," SPEDE (Dec. 1987) 351-58.
7. Winters, W.J. and Warren, T.M.: "Laboratory Study of Diamond-Bit
Fh = hook-load force Hydraulic Lift," SPEDE (Aug. 1986) 267-76; Trans., AIME, 281.
k = drilling-response coefficient 8. Outmans, H.D.: "The Effect of Some Drilling Variables on the In-
K = drilling-model coefficient stantaneous Rate of Penetration," JPT (June 1960) 137-49; Trans.,
L = effective drillstring length AIME,219.
Lb = length of formation drilled in drill-off test 9. Cunningham, R.A.: "An Empirical Approach for Relating Drilling Pa-
Ldc = drill collar length rameters," JPT (July 1978) 987-91.
10. Winters, W.J. and Warren, T.M.: "Field Application of Diamond-Bit
L dp = drillpipe length
Hydraulic-Lift Principles," SPEDE (Aug. 1986) 277-87; Trans., AIME,
Lh = hole length 281.
Lhw = heavy-weight length 11. Timoshenko, S.P. and Goodier, J.N.: Theory of Elasticity, McGraw-
Pi = internal pressure Hill Book Co., New York City (1982).
SPE Drilling Engineering, December 1989 333
Appendix A-Hydraulic Effects During the following expression for the variation of drillpipe tension, of,
a Drlll·Off Test resulting from a change of hydraulics.
Winters and Warren lO presented an analysis of the changes in hook (oFIAdpE) + (oLdpILdp) =0. . ....................... (A-5)
load and WOB that accompany changes in drillpipe pressure dur-
ing drill-off tests with natural diamond bits. Here, similar princi- Combining the two previous expressions gives
ples are applied to the case of changing flow rate with a roller-cone of= -Ail-2j.toPbha + 'h(1-2j.t)OPdp]' ............... (A-6)
bit when the pressure drop across the bit is not influenced by WOB.
If the drillstring hangs freely, the hook load changes as a result Forces acting directly on the top support must now be added.
of changes in the jet reaction force: These forces result from the variations of hydraulic momentum in
the gooseneck (negligible) and of standpipe pressure acting at the
OFh = -(pIAn)(qt -q6) -2(PIAg)(qt -q6) . ........... (A-I) top end of the drillstring. Thus, the hook-load change resulting from
When both ends of the drillstring are' fixed, as when the bit is on changes in hydraulics is
bottom or when the BHA is stuck to the formation, the system is OFh = -Ail(l-2j.t)opsp + 'h(l-2j.t)OPdp]' ............ (A-7)
statically indeterminate, and changes in the drillstring length will
affect the hook load and the WOB. Timoshenko and Goodier ll F bha experiences the same variation because of drillpipe elon-
show that the axial strain of the drillstring can be expressed as gation and also changes as a result of variations of the bit nozzle
jet force and of the pressure drop in the BHA.
€z = (lIE){ U z -2j.t[(Pir l-p o r;)/(r; -rlm . ...... .' ... (A-2)
oFbha = -oF+AioPbha -(PIAn)(qt -q6) . ............. (A-8)
Variations of annulus mud pressure can usually be neglected in
comparison with those inside the drillstring. Thus, the variation With no friction between the BHA and the formation, the WOB
of axial strain of the drillstring is related to axial stress and mud is equal to F bha .
pressure variations by
51 Metric Conversion Factors
O€z =(lIE){ouz -2j.t[rlop/(r; -rlm . ............... (A-3) bar x 1.0* E-Ol MPa
Assuming that there is neither buckling of the drillstring nor ax- ft x 3.048* E-Ol m
ial friction between the drillpipe and the formation and that the com- in. x 2.54* E+OO cm
pliance of the drill collars can be neglected, this expression can lbf x 4.448222 E+OO N
be integrated to find the virtual elongation of the drillstring. psi x 6.894757 E+OO kPa
oL=(LdplE)(A/A dp )[ -2j.t0Pbha + 'h(l-2j.t)OPdp]· ...... (A-4) • Conversion factor is exact. SPEDE
Original SPE manuscript received for review March 15,1987. Paper accepted for publica·
Invoking the principle of elastic superposition, the boundary con- tion May 5, 1989. Revised manuscript received Dec. 12, 1988. Paper (SPE 16162) first
ditions of no displacement at either end of the drillstring lead to presented at the 1987 SPE/IADC Drilling Conference held in New Orleans, March 15-18.

334 SPE Drilling Engineering, December 1989

You might also like