SPE 89598

Modeling Coiled Tubing Drag Forces – Small Factors Have a Big Impact
A.R. Terry, Mahesh. Mahajan, Lance Portman, BJ Services

Copyright 2004, Society of Petroleum Engineers Inc.

• Small changes in down hole conditions are not
This paper was prepared for presentation at the SPE/ICoTA Coiled Tubing Conference held in reflected significantly on the surface weight
Houston, Texas, U.S.A., 23-24 March 2004.
indicator, when the coil is an appreciable distance
This paper was selected for presentation by an SPE Program Committee following review of
information contained in an abstract submitted by the author(s). Contents of the paper, as
away from its maximum attainable depth.
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
• The same small changes in down hole conditions
position of the Society of Petroleum Engineers, its officers, or members. Papers presented at reflect significantly on the surface weight
SPE meetings are subject to publication review by Editorial Committees of the Society of
Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper indicator, when the coiled tubing nears its final
for commercial purposes without the written consent of the Society of Petroleum Engineers is maximum attainable depth.
prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300
words; illustrations may not be copied. The abstract must contain conspicuous • Refining a computer simulation model using field
acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O.
Box 833836, Richardson, TX 75083-3836 U.S.A., fax 01-972-952-9435. data from shallow depths does not permit accurate
estimation of what the maximum attainable depth
will be.
Abstract
A significant limitation in the application of coiled tubing • Estimation of chemically derived drag reduction
technology is the ability of the coiled tubing to push. Push effects, from weight readings at depths distant
may be required to move a heavy bottom hole assembly from the maximum attainable depth is inaccurate.
(BHA) along a highly deviated or horizontal well. Push may
simply be required to push against the drag forces induced by Overview of the behavior of Coiled Tubing
the coiled tubing’s own weight, as it is run into a well. Either in Compression
way, it is often important to know how far coiled tubing can It is useful to have a simplified model in mind when
be pushed into a well, or what weight/length of BHA can be considering what is happening when coiled tubing (or any
pushed to a certain depth. Accurate modeling is required to other tubular) is pushed or placed under compression. Perhaps
avoid a costly mis-run with coiled tubing that fails to get to the the simplest scenario to envisage is a perfectly straight section
necessary depth. of coiled tubing, being pushed along a perfectly straight,
Recent projects carried out offshore Brunei illustrate how horizontal well bore. This, of course, is a totally improbable
critical accurate force modeling is and how small well factors scenario but is still valid for the purpose of creating a picture
can greatly affect the depth that coiled tubing can reach. This of the generic behavior of coiled tubing.
paper will also demonstrate how difficult it is to infer down Initially, as the coiled tubing enters the straight well, it
hole drag conditions from surface readings unless the coiled simply runs along the low side of the well, the drag on the
tubing is nearing its point of maximum reach. coiled tubing being equal to its own weight multiplied by the
coefficient of friction between the coil and the well, µ. As the
Introduction coil extends deeper into the well, the total weight of coil in the
An in-depth, theoretical engineering treatment of the buckling well increases and so, therefore, does the force required to
of tubulars supported within larger tubulars is not the subject push it. At a critical point, the coiled tubing will buckle; it will
of this paper. Other works have shown the theory behind this go from being perfectly straight, to form a spiral. In the same
behavior1,2,3,4,5,6. This paper focuses more on the observed way that a small diameter rod will bow if a large compressive
response of coiled tubing seen during specific campaigns of load is placed upon it, the coiled tubing will bow until it hits
work and makes comparisons with an established, proven the inside wall of the well. The coiled tubing is prevented
coiled tubing simulator. from bowing out a significant distance because the inside
The campaigns involve running long lengths of perforating surface of the well cannot move.
guns into newly drilled and completed horizontal wells. The So our theoretical length of coiled tubing suddenly flips
guns were all run on coiled tubing, the perforating conducted from straight to spiraled where the compressive load exceeds
in the balanced or underbalanced condition, using down hole the critical buckling load for the coil. In this scenario, only the
formation isolation valves to deploy the guns in and out of end being pushed will spiral as the compressive load at the
the wells. free end remains under the critical compressive load.
Computer simulation modeling results are used to illustrate The buckling of the coil actually increases the contact
and quantify the effects seen during the actual work programs. force between the coil and the well. The contact force is now
The observations will show the following: - not only the weight of the coil, it is also the force exerted by
2 SPE 89598

the well to stop the coil bowing out uncontrollably. This two scenarios is only about 300lbf. On the graph above,
means that the drag force increases over this section. More 300lbf looks like a large difference, but in a real well with real
drag means more buckling force and so more contact forces equipment, 300lbf is very difficult to discern.
and so more drag. In other words, a vicious circle is initiated This illustrates how difficult it is to predict how far the
and the force required to push on the coil suddenly increases coiled tubing will reach, based only on running weights at
very rapidly. This can have the appearance of some restriction shallower depths. In this example, a 300lbf change in weight
being encountered down hole, but is actually the natural indicator reading at 1,000ft signifies the difference between
behavior of the coil in even a straight, uniform well bore. the coil being able to reach a maximum depth of 5,500ft,
Figure 1 illustrates this behavior. The leading edge (right) or 2,800ft!
of the coil string remains straight. Back from the leading edge, This exponential type behavior can be even more marked
at some point the compressive load exceeds the critical when heavy bottom hole assemblies are run on the end. Slight
buckling load and the coil goes into a spiral. The spiral results increases in the drag on the BHA result in the entire coil string
in more contact force and even more drag, so producing even being placed under more compression, with the corresponding
tighter spirals. The force required to push the coil into the well increases in spiraling, normal contact force, and drag. An
suddenly starts to increase rapidly. additional load on the BHA causes more of the coiled tubing
to go into a spiral, acting like a force multiplier. A small
Normal Contact Force

increase in BHA load can result in a large increase in push

required at the top end of the coil. This is shown in pictorially
in Figure 3 below.

No end load
Small end load applied

Distance along Horizontal Section

Figure 1 Coiled Tubing Buckling Pattern and the resulting normal

contact force as the coil is pushed along a horizontal well.

Figure 2 shows a computer generated prediction of the Distance along Horizontal Section
force required to push a length of 1½” coiled tubing along a Figure 3 Response of coiled tubing when a small end load is
horizontal well. Two curves are shown, one representing a added
friction coefficient of 0.2, the other 0.4. Both curves show a
greatly increasing push required once the coil gets to a certain There are two distinct regimes in this simplified example.
distance along the horizontal section. The first regime is when the coiled tubing is straight. In this
condition, small increases in drag result in small increases in
0 1,000 2,000 3,000 4,000 5,000 6,000 push required at surface. The second regime is when the coiled
0 tubing has gone into a spiraled condition, caused by a
significant compressive load. In this regime, a small increase
Push Required (lbf)

-1,000 in drag can result in a very large increase in the push required
at surface. In a real well, when the coiled tubing is distant
-2,000 from its maximum attainable depth, it is in the “straight
regime”. As the coiled tubing reaches its maximum attainable
-3,000
depth, it enters the “spiral regime”. It should be noted that the
-4,000 terms “straight” and “spiral” are not literally exact. Coiled
tubing is rarely exactly straight as it has a natural curvature to
-5,000 it, imposed by the bending it undergoes as it is run into the
Distance in Horizontal section (ft) well head. However, in terms of force transmission, the coiled
tubing behaves similarly to perfectly straight coiled tubing in
0.2 0.4 Coefficient of Friction the “straight” regime, and exhibits more substantial buckling
in the “spiral” regime.
Figure 2 Circa™ simulation of 1½” coiled tubing pushed inside To illustrate the magnitude of this effect, using the
7" casing.
scenario shown in Figure 2 with a friction factor of 0.2, the
effect on the required push on surface from an additional
Notice that the coil can reach twice as far when the 200lbf drag on the BHA is 200lbf when the BHA is at 2,000ft
coefficient of friction is halved. Also note, that on route to and 1,600lbf when the BHA is at 4,500ft.
bottom, the difference in push required is relatively small. At
1,000ft along the section, the difference in push between the
SPE 89598 3

Weight Indicator Reading Considerations 30,000 10

As demonstrated in the previous section, any matching of field 29,000
data and computer generated models requires that the weight 28,000 8
27,000
indicator is very accurate. This involves not only ensuring that
26,000 6
the calibration factor for the load cell is correct, but also

Dogleg (deg/100ft)
25,000
involves ensuring that the indicator is zeroed correctly.

Weight (lbf)
24,000 4
Zeroing a coiled tubing weight indicator is actually more 23,000
complicated than might be thought at first. The load cell on a 22,000 2
coiled tubing injector does not only measure the pull (or push) 21,000
on the pipe in the well, it also measures: - 20,000 0
19,000

• The friction of the pipe in the stuffing box. 18,000 -2

• The force exerted by any wellhead pressure trying to 17,000

16,000 -4
blow the coil out of the well. 4,100 4,300 4,500 4,700 4,900 5,100 5,300
• The pull required to pull the pipe off the reel (which Depth (m)
POOH RIH without Drag Reducer
is deliberately held back to keep the coil on the drum RIH with Drag Reducer Dogleg severity
neatly spooled).
• The force required to straighten and then bend the Figure 5 Tubing Weight Indicator Readings during Dummy Run
pipe plastically as it exits the reel and enters the on MLJ-105
gooseneck, as presented by Y.S. Yang7.
Figure 5 shows that there was a benefit attributable to the
Each of these forces can be several hundred pounds. When
drag reducing chemical. This benefit was evident by a running
running in the well, a significant push reading will typically be
weight reduction of about 1,000lbf. This is actually a
seen even if the hanging weight of the coiled tubing itself is
significant reduction in drag down hole as the coiled tubing is
zero. The only way to practically account for all these forces is
in the straight regime at this point. Even with this significant
to use a computer simulation program that accounts for each
reduction, notice that the reduction is of the same order of
and every one of these factors, and zero the weight indicator
magnitude as the noise in the data (collected by a computer
before the coiled tubing is stabbed into the injector.
data acquisition system).
To be useful, this field data next needed to be processed to
Case Study – MLJ-105
derive an effective coefficient of friction reduction attributable
Well Profile
to the chemical and hence predict the maximum length of
The first well used to demonstrate the issues surrounding
perforating guns that could subsequently be run to what
force modeling of coiled tubing in horizontal wells is MLJ-
maximum depth.
105. The profile of this well is shown below.
0
Effects of Drag Reducer Chemicals - Computer Modeling
-2000

-4000
Computer simulation modeling was conducted to derive
-6000
the effective coefficients of friction seen during the dummy
TVD (ft)

-8000 run described above. The running in and pulling out data was
-10000 looked at in order to achieve a best fit set of modeling
-12000
parameters. The best overall fit was achieved with a friction
-4000 -14000
-2000 -800
-100
00 coefficient of 0.18 in the completion and 0.27 in the liner. The
0 0
0
-600
0 matches using these figures are shown in Figure 6
200 -40
00
-20 and Figure 7.
No

0
400 00
rth

0
in
g

) 31,000
600
0 (ft
(ft

g
)

20 in
00 st
Ea
30,000
Weight Gauge Reading [lbf]

00
80 40
00
29,000

Figure 4 MLJ-105 Well Profile 28,000 Actual Data

27,000 Best fit
Effects of Drag Reducer Chemicals – Field Data
The first run into this well offered an opportunity to isolate 26,000

the effect of running a particular drag reducing chemical into 25,000

the well bore. The run involved pushing 174ft of 27/8” dummy
24,000
guns weighing 1,100lb in air. This BHA was run into the well,
3,000 3,500 4,000 4,500 5,000 5,500
the well being full of brine. The BHA was then pulled back up
BHA Depth [m]
2,800ft while spotting a drag reducing chemical, and then ran
back into the well. Comparison of the two runs should isolate Figure 6 Actual and Simulation Data Match for Pulling out of MLJ-
the effects of the drag reducing chemical. 105 – Best Fit
4 SPE 89598

The final step in the study was then to model the run with
20,000
the drag reducing chemical, to see what the best fit
coefficients of friction are with the chemical. This then will
Weight Gauge Reading [lbf]

19,000
give an indication of effectiveness of the drag
reducing chemical.
18,000
Actual Data Figure 10 compares several simulation scenarios against
17,000
Best fit the actual field data for the run with the drag reducing
chemical in place. The three scenarios shown are a friction
16,000
coefficient of 0.18 in the completion and 0.16 in the liner, 0.18
in the completion and 0.27 in the liner (best fit figures from
15,000
the previous runs without drag reducer), and 0.144 in the
3,000 3,500 4,000 4,500 5,000 5,500 completion and 0.216 in the liner.
BHA Depth [m] It can be seen that picking the most appropriate simulation
result is difficult. Each of the three simulations is accurate to
Figure 7 Actual and Simulation Data Match for Running into MLJ- within about 1,000lbf of the actual results. Consequently,
105 – Best Fit exact determination of the benefit of the drag reducing
chemical is not possible. However, approximations can be
These fits are good. However, it should be noted that these made, and the effectiveness of the drag reducing chemical
numbers are open to a great deal of interpretation. Figure 8 appears to be of the order of 20% in terms of friction
and Figure 9 show the same field data as compared to the coefficient reduction.
simulation results for different friction coefficients. It can be
seen that a large range in modeled friction coefficients can be 20,000
accommodated within an error band of plus or minus 1,000lbf.
Weight Gauge Reading [lbf]
The figures below assume a coefficient of friction of 0.2 in the 19,000

completion and vary the liner coefficient of friction between Actual Data
0.12 and 0.28. 18,000
0.16 in liner
Best fit
31,000 17,000
80% x
30,000
Weight Gauge Reading [lbf]

16,000
29,000
Actual Data
28,000 15,000
0.28 in liner
3,000 3,500 4,000 4,500 5,000 5,500
27,000 0.2
0.12 in liner BHA Depth [m]
26,000
Figure 10 Actual Data as compared to Simulation Results while
25,000 running into well MLJ-105 after a drag reducing chemical has
been placed
24,000
3,000 3,500 4,000 4,500 5,000 5,500
BHA Depth [m]
Case Study – MLJ-204
Well Profile
Figure 8 Actual Data as compared to Simulation Results while The second well used to demonstrate the sensitivity involved
Pulling out of well MLJ-105 with force modeling is MLJ-204. The profile of this well is
given below.
20,000
0
Weight Gauge Reading [lbf]

-1000
19,000 -2000
-3000
-4000
Actual Data -5000
18,000
TVD (ft)

0.28 in liner -6000

-7000
0.2 -8000
17,000 -9000 -800
0.12 in liner -10000 -700 0
-11000 -60000
-500
-12000 0
-400
16,000 000 0
-11 000 -300
-10 000 0
t)

-9 -200
ng (f

00 0
-80 -100
15,000 00
Easti

-7 0 0
00
3,000 3,500 4,000 4,500 5,000 5,500 -60 0
0
00
-5 100
BHA Depth [m] No 00 0
r th - 40
ing 0
(ft 00 200
) -3 0
Figure 9 Actual Data as compared to Simulation Results while
0
00
-2

Running into well MLJ-105 30

00

00
-10

Figure 11 MLJ-204 Well Profile

SPE 89598 5

Well Completion likely caused the drop in weight seen. As the coiled tubing is
The well is completed as follows: - pushed further and it enters its spiraled regime, the same down
hole condition might lead to a very large impact on the depth
Table 1 Completion tubular sizes the perforating guns can be run to.
Top Bottom Outside Weight Inside
depth (ft) Depth diameter (lb/ft) diameter Fifth Run
(ft) (in) (in) The fifth run made into this well demonstrates how a
0 11,063 5.5 17.0 4.890 minor discrepancy in running weights on route to bottom
11,063 11,309 5.5 17.0 4.912 greatly impacts on the maximum reach of the coiled tubing.
11,309 16,634 4.5 16.9 3.74 This run involved 651ft of guns to be run down to 17,388ft.
16,634 17,881 4.5 18.8 3.64 The predicted and actual field weights are shown below in
Restriction at 16,634 3.615 Figure 13.

It should be noted that the tubulars were all new and

circulated clean prior to the coiled tubing runs. 17500

Weight Gauge Reading [lbf]

15000
First Run 12500
The first run analyzed was made with 164ft of dummy
guns. The dummy guns were 27/8” diameter, weighing 10000

12.0lb/ft in air. The purpose of the run was to correlate 7500

computer simulation runs, so that accurate predictions could 5000
be made for the future perforating runs.
2500
A total length of some 2,400ft of perforating guns were
required to be run. Obviously, the longer the length of gun run 0
8000 9000 10000 11000 12000 13000 14000 15000
each time, the fewer runs would be required and therefore the BHA Depth [ft]
less rig time would be consumed. If one gun run could not Figure 13 Actual (light) and predicted (dark) weight indicator
reach depth, however, a rerun would be required costing a readings when running in hole with 651ft of dummy guns
great deal of time and money. Accurate modeling was critical
to the overall cost of the perforating operation. From Figure 13, it can be seen that the actual running
The purpose of the dummy run was to determine which weights were a little higher than those predicted, although well
coefficients of friction should be used in the computer within a 1,000lbf error band. However, at a depth of 15,400ft,
simulation, to best fit actual data. These same coefficients the predicted run in weight starts to decline rapidly, whereas
would then be used to model longer lengths of real guns in the actual data did not show that trait. This is because the
order to determine how many gun runs to make, of what simulation model was assuming slightly more drag than that
length, per run. seen in the real well, and therefore switched into the spiral
Figure 12 shows the actual and predicted tubing weight regime sooner. This example demonstrates again how small
values, modeled using coefficients of friction of 0.2 in the changes in drag conditions can have a huge effect on the
production string and 0.23 in the liner. As can be seen, the two overall job. If the guns were required to be pushed to 16,000ft,
compare well, generally well within 1,000lbf. the computer simulation would suggest that it was not
possible, yet in reality it is. This is despite the fact that the
simulation model is accurate to within 500 to 1,000lbf during
15000
the entire trip into the well.
Weight Gauge Reading [lbf]

12500
Conclusion
10000
Computer simulation modeling provides for very accurate
7500 prediction of the surface running weights to be expected
during a coiled tubing operation. Depsite this, however, it is
5000
possible for the same computer software to make significant
2500 errors when predicting the maximum depth the coiled tubing
can be run to. This is because once the coiled tubing gets close
0 to its maximum reach, it switches to a spiraled regime, which
8000 9000 10000 11000 12000 13000 14000 15000 16000 17000
BHA Depth [ft] is very sensitive to small changes in load down hole. These
Figure 12 Actual (light) and predicted (dark) weight indicator small changes may, for example, be attributable to the
readings when running in hole with 164ft of dummy guns
condition of the well tubulars or small amounts of debris left
in the well.
The actual data shows a drop in surface weight of about
These small changes in downhole loads will typically not
1,000lbf after 13,000ft, weight which re-appears after about
reveal themselves as the coiled tubing is run into the well, so
16,500ft. Surface readings such as this raise concerns as to
aborting a run early will generally not be possible. Further, the
what is going on down hole. In this scenario the coiled tubing
exact down hole conditions typically cannot be inferred from a
is in the straight regime, so a relatively large downhole force
previous (dummy) run, unless that dummy run took the coiled
6 SPE 89598

tubing most of the way towards its maximum attainable depth,

in other words into the spiral regime. This is because small
changes in down hole loads lead only to small changes in
surface wieght readings when the coiled tubing is in its
straight regime. These small changes are masked in the
general “noise” in the entire system.
In instances such as those described in this paper, where a
BHA has to be broken into shorter lengths and deployed over
several runs, it is advisable to conduct the dummy run with a
BHA length and weight that will take the coiled tubing into
the spiral regime. This will then permit much more accurate
predictions of the maximum BHA lengths that can be run to
given depths in that well.

Acknowledgments
The authors wish to thank the management of Total E&P
Borneo B.V. and of BJ Services for their permission to publish
this paper.

References
1
He, X. and Kyllingstad, A.: “Helical Buckling and Lock-Up
Conditions for Coiled Tubing in Curved Wells,” SPE 25370,
presented at the SPE Asia-Pacific OGCE, Singapore, February, 8-11,
1993.
2
Qiu, W., Miska, S., and Volk, L. : “Drill Pipe/Coiled Tubing
Buckling Analysis in a Hole of Constant Curvature,” SPE 39795,
paper presented at the 1998 SPE Permian Basin Oil and Gas
Recovery Conference, Midland, TX, 25-27 March, 1998.

3
Miska, S., Qiu, W., and Volk, L. : “ An Improved Analysis of Axial
Force Along Coiled Tubing in Inclined/Horizontal Wellbores,” paper
SPE 37056 presented at the 1996 SPE International Conference on
Horizontal Well Technology held in Calgary, Canada, November 18-
20, 1996.

4
Qiu,W. : “Theoretical and Experimental Study of Buckling
Behavior of Coiled Tubing and Axial Force Transfer Modeling in
Coiled Tubing Drilling,” Ph.D. Dissertation, The University of Tulsa,
1997.
5
Qiu, W., and Miska, S. : “Prediction of Unit Contact Force for Drill
Pipe/Coiled Tubing,” SPE 51092, paper presented at the 1998 SPE
Eastern Regional Conference, Pittsburgh, PA, U.S.A., November 9-
11, 1998.
6
Weiyong Qiu, S.Z. Miska, L.J. Volk, “Effect of Coiled-
Tubing Initial Configuration on Buckling Behavior in a
Constant-Curvature Hole”, SPE55682, published in the SPE
Journal, March 1999.
7
Yong S. Yang, “Bending Torque and Strain Energy in
Reeling Coiled Tubing”, paper SPE 38459, presented in
Montgomery, Texas, February 26-29, 1996.