Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Md~H@
SPE 18047
An reproved Modeling Program for Computing the Torque and
Drag in Directional and Deep Wells
by H-S. Ho, NL Petroleum Services
SPE Member
CW@ht
ffrls paper waa prapered for preeantatiorr et the S2rd Annual Technical Conference and Exhibition of the Society of Petroleum Engineers held In
Hou$ton, TX, October 2-5, 1SS8.
This pepw waa eelected for preeerrtetkm by an SPE Program committee following review of Information contained in an ebstraot submitted by the
dvx(a). Oontenfe of rho paper, ee preeemed, have not beerr reviewad by the society of Petroleum Errgirrearaend re subjeof to eorreofiorr by the
euthor(s). The meteriai, u preeented, doee not necaaearily refleot eny position of the Society of Petroleum Erwineers, its officara, or mambers. PaPars
preeenfed
tSt%
meefinw
are
subject
an increasing
urgency to reduce the total
drilling,
cmpletion, and production cost of a
-11 in order to develop a reservoir more
economically. This will reduce the influence
of price fluctuation
of the worlds crude oil
on the exploration aml devalqment programs of
the ognxating coqanies. Directional drilling
is now regarded as one way to achieve this
objective.
is increasingly
Directional
drilling
being regarded as an effective means to minimize overall
development and production cost
of an oil field,
particularly
for the following situations:
(1) Drilling
multiple directional wells
from the same platform or rigsite,
particularly
in offshore and arctic
areas, to reduce rig cost;
(2)
(1)
(2)
I, INIROLWCrION
Challenges In Directional
Drilling
horizontal
wells to improduction
drainage,
avoid
prove
water coning, and develop very thin
reservoirs.
A.
comprehensive drilling
progrartttrtust therefore address the following issues in an integral manner:
Drilling:
AS oil
(1)
expensive
*-.
&fling/prediction/control
of
well
AN IMPROVEO
FKWELING
PROGRAM
FORCONPUTING
Tt
(2)
Avoidancefirediction/advisory
drilling troubles
(3)
of
It(1) is imperative,
since the target
must be reached within certain offsets if the
wall is to be effective.
In general, the drilling trajectory
does
not conform to the
plennad well path. Mglegs cauaed by unwanted
deviationa are the major source of drilling
trouble, and require costly course corrections
causes of these deviation tenand trips.
dencies need to be well understood, in order
to plan for khem if at all pzsible,
and to
monitor and account for their effects on the
drilling o~ration.
During wdl planning,
optimm~ll
~th
should be de:ined to minimize total drilling
cost. Conventional well-path planning merely
defines the trajectory
connecting tmpoints,
mostly asauming 2-D trajectory.
This should be
expanded to include the anticipateti deviai.ion
caused by the BHA end the formation; the
generated torque end drag, end the ensuing
implications
to
drillstring/casing
design
requimmmts.
while drilling,
constant monitoring is
needed to maintain proper well trajectory,
taking corrective actions
if needed. If@UWt3d
cmtrol am3 predictive
capabilities
will require fewer cxrective
actions, and result in
major coat savings.
Item (2) deals with the many potential
troubles during drilling,
such as: fluid pressure control (kick or lost circulation);
insufficient cuttings transport
and hole cleaning; drillstring
failure; severe hole crookedness which may cause key-seating
and stuck
pipe . Severe drilling
troubles may require
side tracking or even hole abandonment. Most,
if not all of these problems, become more
acute andmre difficult
to resolve in directional wells. JWS it is essential to understand the causes of these drilling troubles,
and develop capabilities
to monitor, intt?rpret, control and predict them.
Item (3) deals with the optimization of
cost for the entire wsll, by considering trade-offs
between conflicting governing parameters.
TMs task is considerably
more difficult
in directional drilling since
more parameters are present. Therefore better
understanding is needed about the effect of
variation of each parameter on the overall
drilling cost. k example of trade-off is the
choice of drilling
mud, Lubricated muds can
reduce borehole friction
but are much more
expensive and mere difficult
to dispose, while
the water-based muds are cheaper but will
cause higher torque and drag.
the total
TORQUE
AWDDRAGIN 01RECTIONAL
ANDQFFP WLLS SPE ~
tion),and casing ruming/eettin9
~rati~-~.
They also affect
the completion/ cenmting
ration.
Their
accurate predic;ons
are
x refore very iqnxtant if the well is tobe
successfully
and economically drilled
and
coepleted.
planning
IIIprWed well
reduce torque and drag;
design
to
(c) ll~~lfte
trouble-spotting
using diagdrilling
(tripping)
logs:
Ccmparisonof measured and predicted
torque and drag
can help s~t
potential troubles;
(F)
(G)
Aid in depth
asurements;
(H)
Aid in jarring
operation: Identification of the free point and the
overpull needed to activate jarring,
since both are affected by drag;
(I)
correlation
of tIWOme-
At present,
applications
of torque and drag
are mostly limited to (A), (B), rind some limited cases of (C) and (D), Increased usage of
such models will be possible
only with improved reliability
and sophisticationof
such
models, along with improved surface measurements of hook load and torque.
Some directional
drillers
use steerable
&&hole motors to change the drilling direcrotary drilling,
motor
Ccnnpred to
driliing generates much higher drag in the
drillstring,
because frictior
it generated by
sliding along the axial directiun,
rather than
the circumferential
direction,
of the drillstring. For horizontal
and for extended reach
directional wells, there may be adepth l?.mit
beyond which one can no longer apply proper
w to the motor,
Recently, the
voted considerable
tal understanding
drilling
industry has deeffort to develop fundamenand analysis software pro..
,
Mm
eumu
M
-----nnm-annn
nu
[ 16047
. .
internal arsl external
flting
fluids will be
grto a&lress the rnjor iumes in direcpresented in a later paper.
timal
drilling
-timed
in the previous
section. In particular,
several BHA(bottose
B. Limitations Of lhe Soft-string
Model:
hole aaaed31y) analysis
programs [1-51 and
acm predictive
bit-rock
interaction swdels
Aa described
in Appendix B, the aoft[1,2,6-10] have been proposed anrVordevelstring n&al neglects
the stiffness
of the
heae programs, if proven in the field,
drillstring,
and is independent of the clearY llpeve thewy for a true predictive direcante between the drillstring
and the borehole
ticwl drilling program that canba usedbdh
wall. AS a result,
effects
of tignt holes and
in the office as a planning aid, and in the
severe local hole crookednesses cannot be
fielda$
a mnitoring
ml advisory tool. BY
easily detected by such a model. AS a result,
coupling such a predictive
drilling prowun
the soft-string
nmdel generally underestimates
with a trcuble analysis program which accounts
the torque and drag, or overestimates the
for theeffectofh oledeviatibnon
torque and
friction coefficient,
drag, we have the basic
elements of a directional drilling sinulator
that will enable us
These limitations
are clearly illustrated
to drill swell ona ccmputer.
in Figs. 3 and 4 to be presented in the following section. They will severely limit the
usefulness of the soft-string
model as a rigsite monitor/advisory
tool for trouble-spot11. PREVICUB
~E0RCOf4PUTIN0
IWVEW
ting.
~
The derivationof
(1)
(2)
(3)
for
A mOre complete
*Ai.
discussion
AII in-house
developed soft-string
sudal program, lURORA-O=,coded with
a very stable numerical integration
technique.
(2)
A BHAanalysis
program for the stiff
collar section. This is modified from
DIDRIL-1~ (a finite-difference
based
large
program using
defornution
theory [5]) to account for the drag
generated wh+le tripping.
rhis improved torque-drag program,Toritm&Ism, can handle top drives when the drillstring is rotated while tripping. It is also
being modified to allow the computation of
stiffness effect
in more than one segment of
if
needed, Itarrently
conthe drillstring
tains the following options:
the effective
To handle the fluid effects,
tension Tewill be used in place of the true
tension T and the sulxnerged weight density
will rep!ace the drillstrings
dry weight
~da;t~i$as
in [12]. The relation between Te
Te=T+po*~-pi
(1)
infinitesimal
drillstring
For an
element, it violates
momentequiliFOr
brium in the lateral direction.
any finite
drillstring
segment, the
assumed torque transfer is incorrect.
Ironically,
when viewed
defornwtion theory [51, the
neous torque transfer equation
a correct simplification,
as
(A-17) in [5].
IMPROVED
rCewE-DRAG
MODEL
Model
~d:~~~~i~
~~
1 ng premises:
III.
(1)
Soft-string analysis
analysis bypassed;
only, BHA
(1)
of the effects
of
Am
NJ INPROVED
MODELING
PROGRAM
FORCOMPUTING
THETORQUE
ANDDRAGIN DIRECTIONAL
ANDOEEPHEELS SPE 18047
changed, since the soft-string
raodel is independant of the borehole diameter.
(2)
c.
(1)
lherefore,
when using MWDsurvey, care
must be taken to reasonably snwoth the data by
the following:
(a)
(2)
(1)
(b)
Stiffness effect
is very significant
when passing
in
collar
section
severe local hole crookedness. For
curve section
example, when the
length is 50, tension at point B is
greater
than that
about 8 kips
computed from the soft-string
model.
(c)
(d)
Such effect
Iessens dramatically for
HWDP,and is negligible
for drillpipe .
IV. SIMPLIFIEDMODEL
FORDEEPVERTICAL
WELLS
In theory,
there will be no torque and
drag in truly vertical
wells.
In reality,
comnonly observed hole spiraling will lead to
continuous wall contacts generating torque and
drag. The problem
rows in exponential manner
as the well depth ? ncreases, and become~ serious for very deep vertical wells. Once hole
spiraling occurs, a vertical well behaves much
like a directional well, and needs to be handied as such. Causes of hole spiraling are:
B,
~rthermore,
in Fig, 4, when borehole
clearance is reduced for the curved length at
100, the improved mcdel shows drematic increase in the effective
tension at point B,
whereas the soft-string
model remains un-
(1)
.Am
41U
Bit tilt
from an i;,itial vertical
position, due to buckling of the
collar when ~ is applied,
HiJA-SHAN
HO
SPE 18047
B.
5
TOr~e-DMg CKMIpUtatiOn
For A Spiraled
(2)
(3
Up-dip drilling
deviation
tendency
[1,9,101 when drilling
into anisetropic formation with non-horizontal
badding.
~s.
(7,8) show that torque and drag are
uniquely determined by the following quantities: the equivalent
curvature k the coefficientof
friction f, andt hewal Y depthS.
A.
Simplified
Vertical
Relations
Spiraling is periodic
pitch. This implies:
d(~)/ds
= O)
d(~)/dS
with constant
= const.= kz
(2)
(2)
The kale
constant
deviation
but
vary
,angle,
small,
~,
is
so that
td=(~-1)/(fkb6);
(3) me tOrwe factor ~,
definedas
the ratioof
over-torque fthe difference between the surface and the drnm
hole torque, if any) over the torque
generated by a slanted well with
constant deviation angle ~:
sin(~)
situations,
as:
~-kb
equs. (A-6
-Tkb=
(4)
T Sin(Ad)
kz;
dT/dS-fTkb=-gcos(Ad)
For tripping
out
equ. (6) leads to:
Nb= 0;
-..9;
condition,
T(S) =Toexp(fk
S) +g [eXp(f kbS)
-11 /tf kb);
Similarly,
to:
integration
Mt(S)=~+rfkb(.5g
of
equ.
(5)
(6)
integration
of
(7)
(A-13) leads
S2 -TOS]
sin(Ad)).
The table
shows dramatic exponential
increases in torque and drag as the well depth
increases,
and illustrates
the significant
impact of deep hole spiralingon
drillstring
design and surface equipment selection.
We
observe:
s/(2
(3)
N-Nnm
Well
sensitive
(2) These values are very
the curvature of the spiral.
to
(3) Variation
from the perfect
spiral
assumed herein
tends to increase
the surface torque and drag. Thereof
the
fore, accurate
definition
~ot~le
trajectory
is very impor-
(8)
,.4
Qll
AN IMPROVED
NODELING
PROGW4 FOR CONIWTIN(i. THE TORQUE ANLY!DRAG IN DIRECTIONAL
TABLEI
EFFBCTOF~CAL?iOW
lURQUE& IRIP*
10.
SPIRALING
DRAG
(1)
kb -
0.S
(2I kb - 1.0
(3) kb= 2,0
dm/100t
dw@OO~l
d*q/100~)
Ml
bawd
on f = 0.2/
Ad * 2 d,q,
wells
should be
walls.
(5)
(6)
the
(7) With improved understanding,
improved torque-drag model can be
used for many other applications to
help the drilling process.
VI. ACxNmmGSMENTE
The autho~~ishes
to express his apprefollowing
persons: John
ciation
to
Fontenot for his guidance while he was at NL;
Paul stodney and Vik ~o for their sustained
suppxt and encouragesumts; Steve Sawaryn of
Britoil snd James Dech of Arco for their well
dataandhelpful
discussions.
rhenksalsogo
to Cathy Connally, nary Spencer, andmary
Pouts for their rush t ings~rt.
Lest but
s NL management for
not least, the author L
the penaissiont opblishthispepar.
VII. ~IURS
Ai:
Drillstring
section area
defined by inner diameter
Di
~:
Drillstring
section area
defined by outer diameter Do
~!
Deviation angle
As:
Asimuth angle
E:
V. CCNCLUSICNS:
In conclusion,
vertical
hendled as directional
we note:
(Youngs) mdulus
(fi;,~~,~~):
(~,~~l~~)!
~;
Frincipal
~~t
Binormel direction;
--4
Et:
i$
. .-
Elastic
normal direction;
Tangential direction,
positive uphole;
Resultant force vector at
seetion of drillstring
------
UU&-WAU
Ilwn-avnuvWIw
16947
f:
t::
(F1,F2,F3):
g %
Friction
coefficient
Te:
To:
Sticking
force (effective)
t:
Components of resultant
vector force P at a section
in global coordinates
~:
~:
~:
%/-%
sulxnerged
- gfj
weight density;
gs:
Drillstrings
density
dryweiqht
Fluids=@tde@W
(Vn,Vb,T):
Physical c~~ne
resultant
curvilinear
I: Moment of inertiaof
drillstring
= n (D04-
(x, Y, z):
section
M4)
64
kb:
kn:
Wstural tortuosityof
drillstring
centerline
k~:
Rate of change of
azimuth angle: ~dS
I
VIII.
tit
*:
Mt~ Drillstring
(O,Mb,-Mt)\
t>rque
Components of fiin
curvilinear coordinates
Ss
~re
fluid pressure
Torque-generating
drillstring
radius of
I
I
lengthof borehole/
drillstring
centerline,
positive going uphole
Arc
I
1
A*.
ts of
force $ in
coordinates
RE~
_...
,.
nil
brlrn,
*wb-**.
..-.
-.
v..
. .
. . . . .
----
.--m
-A -- . ... .
. .,--
----
#h
---
.,-
[10]
[11]
[121
[13]
[14]
.-m.-
(A-4)
d~/dS_-kb~+kn~;
(A-5)
co*
E;+ coaAd-33;
(A-6)
-5;
-*3
(~dssinf@l
(A-7)
kb2= (d@8)2+
(d?@)2
kb >0.
(Sill%d)2;
Therefore:
DERIVATI@4
OP BOF%STRING
_
IN NAIURAL
-IIWES
of
le=T+po&-p~Ai
(A-1 )
g~, by
9V= 9s - 9fr
(A-2 )
/dS+fi
(A-3)
contact
(A-9)
we obtain
dT/dS
- fN+g~~*~~=O;
(A-it))
Nnm-(TKb+g~~*~)t
(?-11)
Nb=-g~@~.
(A-12)
/dS+
is described by~
f
rNii~=
O.
(A-13)
-fN~:+g~@.
dNt/dS=frN.
Along the ~ direction,
we have:
(A-14)
egu, (A-13) impliesl
SPE
NliA-SMN
18047
w:l?%z~~~==$=
~kb=O,
fiisviolatss
equilibrium, unless kb= O.
lhxthermore, uhen any finite lengthof the
drillstring
is takenas a free bcdy, overall
maent aquilibrhas is clearly violated in all
directions,
unless the borehole is straight.
a$jds+z;+gi$.of
lb illustrate,
Fig. 5a isa finite seg~ntof
the drillstring
with constant (2-D)
curvature kb subjected to torque Mtl andmt~
at both ends, end en asawaecl constant distrlbutad torque, t, for ease of illustration.
To
consider masent equilibrium, w do not need to
include all the fyces acting on the free
body, since there is in general no force couple. W can therefore consider mcmentequilibri~abouta
point
on the line of action of
the resultant total force.
Length CD = LOn9th
( froa (A-14) ) I
AS + Arc
(s-1)
&ds+iQx$+t
E;=o;
(B-2)
where
Fig. 5b is a ge~tric
constmction of
the total moment acting on the free i.edy by
the applied torque. The straight lines Asand
CDdanotethe torque at band c, i.e., ~land
%2 respectively,
whereas the curved (circular
arc) section BCdsnot~ the integration of the
distributed
torque t ~. Note the following:
(a)
M)
t.
iLr@g+NbEJ;
fr
@J;
(s-3)
and the resultant bending momant, ~, isdefinedby the kxxeholes flexural curvature,
kb, by:
Mb=kb*EI.
length SC
Noting that:
(b) Ve&rcCD
is
&ds=ayas+E;xi,
where Iij is the natural
vector of the borehole:
(B-4)
total
curvature
(1) Momentequil,
(M$/ds - t;
Therefore, the total resultant moment for
this section is the vector AD, and not zero,
rhis implies that the section is not inmcment
equilibrium,
in fi~ direction:
t=frN.
(B-5)
in~~ direction;
d/dS (T + Mb2/(2EZ))
-fN+g~~*~;=()!
(B-6)
APP~IX (B)
RI~RWS DSRIVATIWOF
CONSTRA2NSD
DRILLSTRING
MODEL
in~direction$
-d2M~dS2 + kn (kbhlt+kn
Mb)+1Kb
+Nn+g~~*-<.Oj
(4) Force equil,
_d(kbMt+knMb)/dS~
(B-7)
in~~direction~
kndM~dS+Nb
In additiorh
we have:
(s-9)
Note that the asmxnption of zero 13tif fness by
the soft-string
mdel implies Mb= O. However,
we cannot therefore assume zero shear force ~
as does the soft-string
model, because of the
teKmkb~.
ThiSerKOK Will lead to inCOKKeCt
normal contact force.
THI
SPE ~
(2) ~tothe
above condition, the
drillstring
twist is different from the borehole tortuosity and not zero, and does contribute to the tortuosity of its centerline (also
discuesed in [5]). rherefore significant error
exists in the c~tation
of the contact force
N.
(2) -ring
(s-7) to (&8) in ~ting
the contact force ~,
the norsul ~tof
w see that the soft-stringmdel
(A-8) raiaaes
the first tw terms. Aaatming planar curves
(8s is the caee with moat aurvey~in~otion
mrthoda), then the tortucdty
Therefore, if themsent
(or hoe
? curvature)
varies linearly~ no rror is irwolved. @herwiee, substantial error will occur in the *stiaate of%. Note that real boreholes doposSeSS *VSniShi~
kn.
(3) comparing (B-8) to (A-9) in conqmting
the binormal conponentof the contact force
Nb, undrx the assumption of zero tortuosity,
we see that the soft-string
mdel misses the
term:
kbdMt/dS
+Mtdkb/dS.
The second term vanishes if the circular arc method is used, but the first term is
m-kbti(frN),
curvatures.
IYAsis amajorassumption
and the
source of Significant error. It completely
of hole clearance,
ignores the effect
drillstring,
M@+5s)
.. ... .
T(S
+ 0S) /
1
-
SIU19
~=
..
L
/
...
S8nd -
%%
.....*...L~/
---
kMm#t2d
1
#
......
Suid...
.......%..
.
sluh
_
_.
......
......
SMd
/sufml
St8ii&
Fig.2-kNalan4htmp01alMMp@.5.
B./AT
i\
AT
40K
+,
w
~.
%..
-.
-.-
e-
40
Tatd s-zm
%titmha,
--*--*
D12rn0-u m2u
~-e-
m22mrc
--K:
Mo2u
a-m?2e/ft
EI-52XXPMI?
HWDP
9 =3s.24
b/ft
E1-8xlfP2Ai
A-
~.+.
2.5
R---------____
--
DPfOD.+.
9 = IS.21 Wfi
so
FIG. 3 EFFECT
OF S*EP
W.3
--
X d b-$12)
1
209 &ffu
o
*
w
*,,
1!
II
,1
l!
18047