Sie sind auf Seite 1von 5

Loads on Drillpipe

During Jarring Operations


Thor V. Aarrestad, SPE, Statoii. and Age Kyllingstad, SPE. Rogaland Research

Summary

Jarring implies heavy loads on the drill string. The highest load on
the drill pipe before jarring is at the rig floor. This paper discusses
loads on drillpipe before, under, and after jarring. We show that for
most situations, the shock wave from the jar impact does not imply
additional load on the' drillpipe compared with static load. The
theoretical results are confirmed by measurements of a jarring operation with stuck point at = 1200 m measured depth. Loads on the
drillpipe can be a limiting factor in jarring operations because fear
of possible additional loads from jarring dynamics may restrict the
trip force (overpul!) on the jar. Our main conclusion is that dynamic
jar forces do not give additional loads on drillpipe. This information
can be used to set an optimal trip force on the jar.
Introduction

Stuck pipe is an expensive event in drilling operations. Operation


companies put a lot of effort into avoiding stuck pipe. Whenever a
stuck-pipe situation occurs, a precise and adequate reaction from the
driller will increase the probability of freeing the pipe.
Publications on jarring operations are scarce if patents are not
counted, with only about 15 papers published on jarring connected
to drilling operations from 1975 to 1992; we cite nine of these. 1-9
The main concerns in these papers are dynamics, jar performance,
and effect on stuck point. Two papers 7,9 discuss the acceleration
phase in detail and show how speed and force change step-wise. No
papers discuss the dynamic loads transmitted up the drillpipe. This
paper addresses these loads.
Deep, highly deviated or horizontal wells may result in operating
conditions near the tensional limit of drillpipes. Stuck pipe in such
wells will therefore give additional concerns about loads during jarring. Possible addition of a shock-wave amplitude on the static load
may impose operational limits on the trip force of the jar. However,
the jarring operation can be divided into phases in which the acceleration and postimpact phases determine the possible added load on
the drill string by the impact. Operational limits from the jar dynamics can be avoided by the use of optimized drill-collar lengths and
heavyweight drillpipes (HWDP's) described in this paper.
Jarring Process

Our analysis will be restricted to up-jarring because it normally represents higher loads than down-jarring. The jarring cycle can be divided into the following. five phases.
1. Loading. In this phase, string is stretched to store strain energy
and the jar is exposed to a (preset or operator-determined) tension
force. This phase typically lasts only a few seconds, but it can take
minutes when hydraulic jars with a long delay time are used.
2. Acceleration. This phase, also called the preimpact phase, is
the time from jar release until jar impact and typically lasts from 50
to 200 milliseconds. In this phase, the strain energy is converted to
translation energy. The higher the speed at impact, the higher the impact force will be at the stuck point.
3. Impact. The inside of the jar is often pictured as consisting of
two parts: the hammer and the anvil. The impact phase is the short
time interval, typically 10 to 50 milliseconds, when the jar hammer
hits the anvil. In this phase, the bottomho1e assembly (BHA) is exposed to high impact forces.
Copyright 1994 Society of Petroleum Engineers
Original SPE manuscript received for review Nov. 16. 1992. Revised manuscript received
May 16. 1994. Paper accepted for publication June 28, 1994. Paper (SPE 24970) first presented at the 1992 SPE European Petroleum Conference held in Cannes, Nov. 16-18.

SPE Drilling & Completion, December 1994

4. Postimpact. This phase lasts until the string has come to a complete rest again.
5. Resetting. During this phase, the string is lowered and the jar
is put into a small compression force to reset the jar reset for a new
jarring cycle.
Fig. 1 illustrates the jarring process phases.
Assumptions and Basic Equations

For simplicity, the analysis will be restricted by the following assumptions:


1. The stuck string can be divided into a few uniform sections
with different cross-sectional areas but with equal material properties: the section below the jar (often called the fish section), the jarring-mass section (drill collars above the jar), the HWDP section
(optional), and the drillpipe section.
2. The drillpipe section is so long that reflections from the surface
can be neglected during the acceleration and impact phases.
3. Jar release is instantaneous, with no transition time between the
loading and acceleration phases.
4. Friction forces are neglected, both internally in the jar and externally along the drillstring.
5. Residual vibrations in the fish section at the time of impact are
neglected.
Assumption I implies that jar accelerators are excluded. A jar accelerator placed at the top of the jarring-mass section normally increases the impact force across the jar and at the stuck point. Because it transmits shock waves poorly, it dampens the dynamic loads
on the drill pipes. Consequently, the conventional jarring assemblies
without an accelerator represent the worst cases for drill pipe loads.
Jar release is usually not achieved instantaneously (Assumption
3), However, Assumption 3 simplifies the mathematics significantly. Because this assumption affects both relaxation of the drill string
and the impact force, the simplifications should not influence the
conclusions of the paper. A further discussion of this assumption is
given later.
As Fig. 1 clearly shows, Assumption 4 is a simplification. If the
internal jar friction force is constant, an effective jar pull equal to actual overpull minus internal friction force can be defined. The wellbore friction can be handled similarly, and the two friction forces
may therefore be regarded as static forces adding to the gravitational
force. Therefore, Assumption 4 is not a serious limitation.
The requirement of continuous force and speed across the drillstring intersections implies that a wave propagating in Section i is
reflected by a relative amplitude of

rij = (Ai - AJ/(A i + A;) ...... , ................... (I)


at the boundary to neighboring Sectionj. The reflection coefficient,
rij, refers to the speed amplitude. rij for the corresponding force
is - rij. The speed amplitude transmission coefficient is 1+ rij,
whereas the force transmission coefficient is

Kij

1 - rij

= 2A;/(Ai + A;). . ...................

(2)

The sudden jar release, from the static jar pull, F0, to zero force, represents a negative force wave propagating upward in the string. The
same wave may also be described as a speed wave with an amplitUde
equal to the free-contraction speed of the jarring-mass collars:
Vc

= (va/EAc)Fo . ................................ (3)


271

SGGG

; Acxele .. ~ PoslllD....,1

LodiDc pM.. ~ p.....:

4000

p.....

3200 kN

JGGO

....
.s zooo

..

Il

.s

..,li

690kN

1000

1000

0.1

0.2

0.3

11me In second.

Fig. 1-Jar-test measurements illustrating phases of jarring operations (from Ref. 1).

Jarring Without an HWOP Section


With only one section between the free jar end and the drill pipes, dynamic behavior is quite simple. Recalling that the rij = 1 for a free
end and representing the reflection coefficient at the top of the drillcollar section by rep' the speed at the free end is 9

Vn

= vc( 1

+2

= 3.75 t,

i>c~), .......................... (4)


1=1

where n = number of reflections (round-trip periods) in the jarringmass section. The corresponding force at the interface or at the lower end of the drill pipe section is

Fn = (rcXFo . .................................. (5)


Note that for large n, Fn-+O while the speed approaches (A"IAp)ve,
the free-contraction speed of the drill pipes.
Fig. 2, illustrating the decay or relaxation of the force, shows the
force distribution in the jarring collars and in the lower part of the
pipe section as time increases from jar release. The time is given in
terms of the acoustic round-trip time in the jarring-mass section.

t =4.75 t,

<:>

O.~_

t=

_ _

5.75 t,

O.______Q
. . . . . . . . . . . . . . . . .. .

Fig. 2-Relaxation of drillstring with direct drill-collar/drillpipe


connection. Arrows indicate direction of front.

te = 2Ldclva. . .................................. (6)


In Fig. 2 the arrows indicate the direction of the wave propagation.
Assume that the jar hammer hits the anvil after an acceleration
time corresponding to n reflections. From this time of impact, the
lower end is not free but solidly connected to the upper end of the
fish. With Assumption 5 and the continuity of force requirement, the
strength of the primary impact force is 9

F/ = 'Y/(vlI/vc)Fo, ..................... . ........ . .. (7)


where 'Y/ is defined by
'Y/ = Aj/(A j

+ Ac) . ...............................

(8)

This factor equals 0.5 if the cross-sectional areas of the fishing section and the jarring-mass section match; i.e., if Af=A".
When this impact wave is transmitted to the drillpipe section, its
amplitude is reduced to

Eq. 10 can be rewritten as


c:I> = 'Y/(l

+ rcp-2rc;+1)/(I-rc; ) '

...................

(II)

We can easily show that


c:I> n:::.~ 'Y/[2A,j(Ac

+ Apl]. .. ...................... (12)

The number of reflections needed to make c:I> less than unity depends on the area ratio between the jarring mass and drill pipes and
on the factor 'Y/. Table 1 lists the dynamic force ratio, c:I>, explicitly
calculated for different cases; the reflection coefficients are calculated relative to 12.7-cm [5-in.], 33.6-kg/m drillpipes. In all cases,
'Y/ = 0.5; i.e., the collars above and below the jar are assumed to have
equal sizes. Table I and Fig. 3 show that the relaxation will be sufficiently large to compensate for the shock wave if the jarring mass
is short enough to allow for at least four reflections before impact.

Jarring With an HWOP Section


To compare the dynamic peak force in the drillpipes with the static
prejarring force Fo, we use the force ratio
c:I> = KcpF//(Fo-rc;Fo) . .......................... (10)

The denominator represents the relaxation force during the acceleration phase. If this ratio is equal to or less than unity, the peak force
will not exceed the prejarring force. This argument is fulfilled even
better higher up in the string because friction is not accounted for
here.
272

An HWDP section is commonly used between the jarring mass and


drillpipes. The length of this section typically is 100 m so that reflections from the cross-sectional discontinuity at its top cannot be ignored. These reflections significantly increase the complexity and
the number of waves that have to be tracked during the acceleration
phase (Fig. 3).
The force distribution in the upper BHA is given at discrete points
of time during the acceleration phase. Note that the force at the lower boundary (between the drill collars and HWDP) does not decrease monotonously as time elapses. The force rapidly approaches
SPE Drilling & Completion, December 1994

TABLE 1-<1> FOR DIFFERENT VALUES OF n

Case 1
Case 2
Case 3

Case Parameters
Drill Collar Diameter
(em [in.])
'Ji
16.51 [6 1/ 2 ]
0.595
20.32 [8]
0.734
24 .13 [9 1/ 21
0.812

!Sa

0.405
0.266
0.188

<I>

Case 1
1.095
0.908
0.851
0.826
0.814
0.807
0.797

1
2
3
4
5
6
00

Case 3
1.311
1.087
1.014
0.978
0.957
0.944
0.906

Case 2
1.234
1.022
0.954
0.921
0.903
0.891
0.870

t = 2.75 t,

zero in the first part and increases again when the reflected wave
from the upper boundary returns. Also note that the force in the drillpipes (or at the upper boundary) decreases more slowly than in the
case with no HWDP section.
The speed at the top boundary can be seen to approach the speed
of the jar hammer after a few round-trip times in the HWDP section;
i.e., it approaches the jar-hammer speed after a few multiples of
th

= 2(Lh /v a ).

.. . . .. . . . . .

= Fo

t =4.75 t,

:
:

1.

- [(vp-vn)/vp]Fo

=3.75 t,

(13)

However, the speed of the upper boundary approaches the freecontraction speed of the drillpipes more slowly than it does for a
drillstring without an HWDP section. From the general relation between force and speed for progressive waves, we can see that the
force reduction in the drillpipe section at impact is

l'lFdp

= 5.75 t,

(v,,/vp)Fo ........ .... (14)

because Vn is an approximation for the real speed at the drillpipe


boundaries. Combining the primary impact force wave given by Eg.
7 with the transmission coefficient through two boundaries,
(15)

KchKhp = (1 - rchl( 1 - rhP)'

Fig. 4-Drillstring reiaxation with HWDP section included. Arrows indicate direction of front.

gives

<I> =

Kch KhpF[
l'lFdp

= ."

4AhA c
(A c + Ahl(Ah

+ Ap) .

... ....... ..

(16)

Comparing Eqs. 12 and 16 shows that <I> is slightly higher with an


HWDP section than for a simple jarring assembly without HWDP.
The main reason is that the transmission coefficient is higher for tapered strings than for direct connection between drill collars and
drillpipes; i.e.,

KChKhp> Kcp ... ... .. . .. ...... ... ... .. . . ... . ... (17)
provided that Ac >Ah >Ap.

From this discussion, we can deduce that an HWDP section reduces the jarring efficiency. For jarring collars of the same size and
length, the jar-hammer speed increases more slowly and may not
reach the upper limit, vp , before impact. This is clearly seen in the
extreme case when the acceleration time is shorter than the acoustic
round-trip time in the HWDP section. This may happen even in
practical operation if a long HWDP section is combined with a high
pull force. In this case, the impact speed, Vn , is approximately equal
to the free-contraction speed of the HWDP section. In comparison,
Vn without the HWDP section would be close to the free-contraction
speed of the drillpipes. The corresponding <I> equals the cross-sectional ratio AplAh and typically is 0.5. Ref 10 discusses the effect of
HWDP sections on jarring efficiency further.

1.5 -----------,---------~

l
~

0.5

+ - - - - - - + - - - - -+-- - - - - t - -- - ---j

0~1~---~~------~2-0.-3c-m-(-8.~)DC---~--+-~--~~~
10

15

20

Number 01 reftecHons. n

Fig. 3-Dynamic force factor for different numbers of reflections.


SPE Drilling & Completion. December 1994

Comparison With Real Data


In Nov. 1986, a 1220-m-long string was accidentally stuck in Well
U2 at Ullrigg (Rogaland Research's full-scale research drilling rig
in Stavanger). The BHA of this string included 18.7 m of 20.32-cm
[8-in.] drill collars and 55.8 m of 12.7-cm [5-in .] HWDP above a
19.7-cm [7%-in.] mechanical jar. The rest of the drillstring was
12.7-cm [5-in.], 33.6-kg/m drillpipe. The jar was fired repeatedly to
try to free the string. Before jarring ended, strain gauges were applied to the kelly body to record jarring forces accurately and at a
high data rate.
Figs. 4 and 5 show the results from two recordings. Fig. 4 shows
the kelly tension over a wide time span, 20 seconds, including loading through postimpact phases. The data clearly show a lower amplitude in the dynamic part of the jarring cycle. We also see that the
driller uses several seconds to stop the upward motion of the traveling block after detecting the jar firing. The final tension of the string
273

JARRING OPERATION
TIME-SPAN 500 ms

JARRING OPERATION
TIME-SPAN 20 s
1400

1000

.5

=
=
:I
0

'r;;

./1

600

--

---

200

550

=
'"
=
:I

400

-200

.S

./

700

12

16

r--

1'\

.5

V J

Q;

L
~

.--

20

250
100

Time in seconds
Fig. 5-Kelly tension recorded during jarring operation.

is therefore higher than before the jar starts to move, but this is clearly not a result of the jar dynamics. This sequence is typical with use
of mechanical jars.
The data in Fig. 5 are captured at a higher data rate (i.e., 0.5-millisecond sampling interval) and cover only the most-interesting parts
of the jarring cycle. The first part of the trace represents the acceleration phase. There is a smooth transition from the loading phase at the
start of the trace to full release at 100 milliseconds. Consequently,
our Assumption 4 is not fulfilled very well. This finite release time
is probably the main reason for the rather smooth and continuous
force decrease in contrast to the discrete steps predicted by the simple theory described earlier. We discuss other effects that partly explain the smooth behavior later.
The kelly force is measured close to the heavy traveling mass,
"., 19 000 kg. This lumped mass acts as a nearly fixed end for highfrequency waves. The recorded drop in the tension force before impact and the impact peak are therefore nearly twice as high as the
corresponding incident-wave amplitude. The ratio is, however,
maintained during reflection.
The trace shows a relaxation force of 480 kN and a following impact-force pulse of"., 385 kN. Therefore, a rough estimate for <I> is
""0.80. The real <I> value for the incident wave (before reflection)
is even lower when the constant hoisting speed and the finite traveling mass are accounted for. The static hoisting speed gives an extrapolated increase of "., 60 kN for the reference force. The acceleration of the finite traveling mass can cause a tension increase of "., 40
kN relative to the infinite mass case. AccQrdingly, the corrected experimental force ratio becomes <I> ""0.65. The corresponding
theoretical value is 1.08 when Eq. 16 is used with 1] = 0.5.
The pronounced discrepancy between theoretical and experimental values for <I> is probably caused by smoothing effects, which
broaden the primary impact force peak and and make it less sharp
then predicted by simple wave-propagation theory. One such effect
is the multiple reflections inside the jar caused by the complex internal geometry with many cross-sectional discontinuities. Also viscous damping because of drillstring/mud interactions tends to
dampen high-frequency components more than low-frequency
components. Finally, the regularly spaced tool joints in the drillpipe
section represent a dispersion of longitudinal waves that smooths
spikes and rapid force transients.
The results in Fig. 5 can also be used to estimate downhole impact
forces and speeds. From the relaxation force amplitude of 290 kN
(corrected for reflections and traveling-mass motion), Eq. 14 predicts Vn = 1.9 mls. The corresponding primary F[ is = 1100 kN. If
F[ is reflected 100% at the stuck point, the peak force at the stuck
point can reach twice that value.
This example illustrates that high-frequency surface measurements combined with a thorough understanding of jarring dynamics
can be used to evaluate downhole jarring forces. The analysis could
274

100

'"

.1\

'w

AlJ r

200

300

400

500

Time in milli seconds


Fig. 6-High-rate data of kelly tension during jarring operation.

be extended by carrying out simulations for the kelly forces to find


the prejarring pull, Fo, that gives the best match. This is, however,
beyond the scope of this paper.
Conclusions

1. Jarring-induced dynamic forces in the drillpipes may be regarded as unloading (relaxation) followed by a rapid reloading. The
dynamic peak force in the drill pipe normally will not exceed the
static prejarring force.
2. The peak load in the drill pipes can exceed the static force if the
collars below the jar are larger than the collars used as jarring mass
above the jar or if the stuck point is just below the jar.
3. An HWDP section above the jarring-mass collars increases the
force ratio between the dynamic peak force transmitted to the drillpipes and the preceding relaxation force. In addition, a long HWDP
section is detrimental to jarring efficiency because it slows down the
jar-hammer speed and thereby affects the forces transmitted to the
stuck point.
4. Smoothing effects, such as scattering, viscous damping, and
tool-joint dispersion, will decrease the amplitude of the sharp force
transients propagating up the string.
5. Kelly tension data recorded during real jarring events support
the theoretical conclusions.
6. Measurements and theory both confirm that the dynamic
forces induced in the drillpipes will not be a limiting factor in jarring
operations.
Nomenclature

Ai = cross-sectional area of drillstring Section i (collars,


fish, HWDP. and drillpipe), L2, m2
E= elasticity modulus, mlLt2, GPa
F0 = overpull force on jar, mUt2, kN
f1Fdp = force reduction in lower part of drillpipe section after
n reflections
F[ = impact force, mL/t2 , kN
Fn = remaining tension in drillpipe after n reflections,
mUt 2, kN
Kij = transmission coefficient between Sections i andj
(collars, HWDP, and drillpipe)
4tc = jarring mass = length of drill-collar section between
the jar and the HWDP (or drillpipe) section, L, m
Ln = length of the HWDP section, L, m
n = number of reflections in acceleration phase
rij = reflection coefficient between Sections i andj,
(collars, HWDP, and drillpipe)
t = time, t, seconds
SPE Drilling & Completion, December 1994

Ie = round-trip time for a stress wave in the jarring mass,


th =
Va

Vi =
Vn

11 =

4> =

t, seconds
round-trip time for a stress wave in the HWDP
section, t, seconds
acoustic velocity of steel, Llrl, mis-I
free-contraction velocity of drillstring Section i
(collars, HWDP, drillpipe), LIt-I, m1s- 1
speed of hammer after n reflections, Llrl, mis-I
factor accounting for area difference between fishing
and jarring masses
impact force transmitted to drillpipes/relaxation force
of drillpipe ratio

Acknowledgments

We thank Statoil for permission to publish this paper. Some results


are based on research performed by Rogaland Research sponsored
by AlS Norske Shell and Statoil. We thank these sponsors for the
permission to use these results.

8. Schmid, J.T. Jr.: "Designing BHA's for Better Drilling Jar Performance," World Oil (Oct. 1982) 195, No.5 , 100.
9. Friedman, M.B., Skeem, M.R., and Walker, B.H.: "Drill string Dynamics During Jar Operation," lPT (Nov. 1979) 138\.
10. Aarrestad, T.Y. : "Improving Hydraulic Jars Efficiency By Use of Drag
Calculations," Oil & Gas 1. (March 29, 1993) 65.

General Reference
Gabolde, G. and Nguyen, J-P.: Drilling Data Handbook, sixth edition, IFP
& Editions Technip (l99\) 47- 104.

SI Metric Conversion Factors

ft x 3.048*
ft 2 x 9.290 304*
in. x 2.54*
Ibf X 4.448 222
Ibm x 4.535 924
psi x 6.894 757

E-Ol
E-02
E+OO
E+OO
E-OI
E+09

Thor Viggo Aarrestad is a staff engineer. drilling technology, with


Statoi!. He has worked at Statoil since 1988 in drilling technology
development. Before that he was at Rogaland Research,
where he also worked in drilling technology. Aarrestad holds MS
and PhD degrees in applied mathematics from the U. of Bergen. Age Kylllngstad is head of research in the Drilling & Well
Technology Dept. at Rogaland Research. His primary research
interest is problems related to drillstring dynamics. He holds a
PhD degree in physics from the U. of Trondheim.

Aarreslad

SPE Drilling & Completion, December 1994

=N
=kg
=GPa
SPEDC

Conversion factor is exact.

References
1. Kyllingstad, A. and Halsey, G.W.: "Performance Testing of Jars," paper
SPE 20001 presented at the 1990 IADC/SPE Drilling Conference,
Houston, Feb. 27-March 2.
2. Kyllingstad, A. and SI0rdal, L.H.: "Performance Testing of Jar Accelerators," paper SPE 19557 presented at the 1989 SPE Annual Technical
Conference and Exhibition, San Antonio, Oct. 8-11.
3. Wang, J.K . et al.: "A Practical Approach to Jarring Analysis," SPEDE
(March 1990) 86.
4. Askew, W.E. : "Computerized Drilling-Jar Placement," paper SPE
14746 presented at the 1986 IADC/SPE Drilling Conference, Dallas,
Feb. 10-12.
5. Lerma, M.K.: "A Study of How Heavyweight Drillpipe Affects Jarring
Operations," paper SPE 14326 presented at the 1985 SPE Annual Technical Conference and Exhibition, Las Vegas, Sept. 22-25.
6. Brown, M.e. : "Using Bumper Subs, Jars and Intensifiers," Drilling
Contract (April 1985) 41, No. 4.
7. Kalsi, M.S., Wang, J.K., and Chandra, U. : "Transient Dynamic Analysis
of the Drillstring Under Jarring Operations by the FEM," SPEDE
(March 1987) 47; Tran s., AIME, 283.

=m
=m 2
=cm

Kyllingslad

275

Das könnte Ihnen auch gefallen