Sie sind auf Seite 1von 5

}f, ...

,------
m.
.. .
t
.. . . .. .... . .. . . . .................... ...__,
___
-- ..-~.\4Jw!..
-------
,
-- &sFr-

Effect of Dynamic Bit Forces on Bit Bearing Life


kS50 PRODUCTION RESEARCH CO.
D. W. DAREING HOUSTON, ~EX.
U.OF ILLINOIS . .
E. 1, RADZIMOVSKY URBANA, ILL.
,

ABSTRACT BACKGROUND
This paper presents theory wbicb is usdd to Manufacturers of- bearings have conducted
predict rock-bit bearing life reduction due to simple Laboratory tests which show how bearing life is :?-
harmonic variations in bit forces. The theory is affected by steady bearing load. 1 The well-known
based on. the premise that if u const~nt load acts empirical ~quation which defines bearing life
for a certain fraction of the life which the bearing follows from bearing test data. It is:
would have i{ that constant load were applied /or
-~=c9
its etrtire life, then the scme fraction of the [ife of
the bearing
related
is consumed. Total
to simp Ze hzrmorzic variations
bearing Ii/e, is
in bearing
()7..... 1)
where L = bearing life, hr
life associated with instantaneous bearing loads.
P = bearing load,- lb
The results show, that rock bit bearing lije is not
C = constant .
ajjected appreciably by jorces cornrnordj impressed
y = exponent (constant): ;
upon a bit, but it is ajjected appreciably by
relit iue,ly large dynamic bit jorces occasionally-. The effect of environment on bearing performance
..
encountered during drilling. In addition, the results is manifested through the expontint q and the
show that the !ije oj sea[ed bearings is rnor~ constant, C. King, 2 in simiJar tests con&cted on

sensitive to bit jorce variations than is the lije oj roller bearings used in rock bita, showed that, in
utiealed bearings. addition to the load, the environment (abrasive
,, materials, chemical components, etc.) greatly
INT~ODUCTION. , ~~~ affects bearing life.
. His data, which show how the Iife of roller
Roller bearings are important cori$onents in bearings in a 7-7/8-in. W7R bit rotating at 60 rpm
roller cone. rock bits because they transmit a Iarge is affected by bit weight and fluid environment, are
portion of the drill .coIlar weight to. the formation displayed on a log-log plo~ by broken lines made up .
being driHed. Unfortunately,
. they have a i-eladvely . of stniight line segments. ~ach broken line
short life under down -hole drilling conditions. corresponds to a given environment. One ,segment
- Roller bearings are usually the first bit components .- (q= 1.4) of the broken line corresponding to a fhid
to fail, and their failure leads immediately to environment of bsrite mud is shown in Fig. 1along
. failure of the entire bit. Short rolter bearing Iffe,
an-i? thus short bit life, is due mainly to high @
preferences givenat end of paper,.

f-till:. WE
MT!
am111111
Ioads traiismitted by the bit and to the adverse
environment in which the bearings operate. The

.,
results of the study discussed in this report show
how dynamic bit forces also affect roller bearink life.
1%~ purpose of thi~ paper is to discuss ;heory
~ A 7-7/1 W?R
PRESENTED IN Tcn*s
~lTaROTATtH9
O*
aT S0 MPU

._deveIoped , to pre_4ict the effect of different


maggitude~ of peghdic ;bit forces on roller bearing
- life. Theoretical results are used to gompare the
~ life of rock bit roller bearings subjected, to.steady
forces to fhe life 0$> rock bit roHer bearings
si. -
subjected to unsteady forces similar to the type of
unsteady forces measured at the bit with a specidl
--~: downhal+reording:insmmekc.~ .-. . .. A--L - Ia.Rkm&An
m.21
t ! I 1 r 1 1 1 1I I 1 1 t. t 1 t t
.- 4W WL3
[?0 *D m 60 w !90 too
. .. .
~.Origlnaf rnanusefipt receivedin Soc.iet y of Pefraieum Ewineers i RO&LSfl SCMIIIO. Ll$$ - KRS
-. ... ..-+
office June W, 196% Revised manus@pt received Oct. 22, 1965.
Paper (WE 1261) presented at SPE Fall bfeetiag held Ott, 3-6, FIG, EFFE~T
1 - OF WEIGHT AND ENVIR@hlENT -. -
196S, in Denver, Coio. ON ROCK.BET BEARING P13RFORM~NCq.
. .. . L--.
!. SOCIETY OF PETROLEUM EAGINliEftS JbURNAL
Z7Z .-
-. - ... .,, . . . .
with a segment (q = 2.8) of the broken line load is represented by rhe force vector P. With no
corresponding to t clean environment of forced air. load applied, the contacg between a roller and the
The upper load limits of these two segments, 60,000 inner race is a line (into the page) indicated by
lb (g = 1;4) and 100,000 lb (q = 2,8), could be arrow a; under load, elastic deformation causes the
extended slightly and still. remain within the litie of contact to become an area of width 2A.
statistical accuracy of the test data. Because of the; geometry of the bearing elements,
While the tests recorded in Fig. 1 give an upper contact stresses along a are higher than contact
lit-nit of about 2.8 for q, roller bearing fatigue. tests stress es between the roller and th~ outer race along
conducted by manufacmrers of roller bearings, give the line indicated by arrow c. The number of stress
an upper Iimit,.of 3.3 for a bearing which is properly cycles experienced by the inner race is al~o higher
lubricated throughout its Iife. The difference may than the number experienced by the outer racti.
be $ thi air and oil lubricants. Consequently, bearing fatigue failure caused by
,The results of the study discussed in this report dynamic or steady bit loading shouId start some-
show, how dynamic bit forces alao ,affect b~aring where aIong line a. Rock-bit bearing failures confirm
;
life. NO one has preciously described the effects this.
of dvnamic
. bit forces on bit bearing life either , 13e~ause the load on a roller i~ not distributed
experimentally or mathematically, poss~bly tiecause uniformly along its length, failure wi 11 start at one
of a lack of quantitative data giving the magnitudes ~ end of the line of contact. 4 This nonuniform load.
and frequencies of the dynamic forces at the bit? dist~.bution is ,caused by misalignmetit resulting
Recently we measured bir forces and motions from Initial, c~earance and eventual wear in the nose
wi tb a dow~ -hole recording instrument. Tests made bearing (slldlng or journal bearing).
in severa 1 different wells under various operating As the center Iine of the roller passes line a, the
conditions showed that amplitudes of bit weight stresses at this point change rqpidly and approach
variations were usuaHy about 25 to 50 per cent of zero asymptotically (Fig. 2). The time & indicates
the average indicated bit weight. OccaaionaHy, the interval during which the contact area in the
bit-weight variations were relatively large, with inner race is subjected to high stresses as a roller
amplitudes of the variations reaching 100 per cent passes. The magnitude of i% can be determined by
of the average indicated bit weight. kittematic and Hertz defection the&ies, 5-7 ,
We have assumed these variations to be typical The velocities of different components in a roller
and have used the theorv. develoDed in this report
s. bearing are det~rmined (assuming a pure rolling
to determine their effects on bearing life. No motion) by treating each bearing component as a
exDekimental work hss been done to check the component of a planetary gear rrain. III this case,
re&.dts, but the approach is generally accepted by the inner race repres~nts a fixed sun gear whiIe the
experts in the bearing industry. 3 Followiri~ sections outer race represents a ring gear and the rollers
of ~ this re~ort describe the kinematics of roller represent planet gears. The angular velocity of an
bearings, ~efine the mathematical form of measured- imagi~ky link connecting the center of, the inner
bit fo~ces and evaluate the life expectancy of race an,d the center of q roller is:
bearings. Finally, the bearing life reduction caused
by dynamic bit forqes is determined. P/t =:JD J. . . . . . . . . . .
!,
. . (2)

KINEtiATICS OF , 1+~
o
ROLLER BEARING COMPONENTS
where N! = angular velocity of imaginary link
A schematic of a cylindrical roller bearing used rela~:ve to pin, rpm
in rock bits is shown in Fig: Z. The radial bearing
N = Angular velocity of cone relative to
.,
.. . ~ /-coNs . pin, rpm
Dj = outside diameter of inner race, in.
Do = inside diameter of oute{,race, in.
The linear velocity (in./sec) with which a roller
moves over the inner race is:
,.,
!,
n-DiNp. ... -..?
_
. . i3J ----
-60..-.-
.x
., Th& width of the contact swip between the roller
and the inner race is d&eirained froni Hertz fo~uli:
,
) /1 l\

=--r= (A}
-
FI13, .2 . ROLLER BEARING C030M13TRY. ~.,
-:-n- .
,
.

. *
where A = half of the width of contact between corresponding to an instantaneous value of the
two cylinders pressed together (see unsteatiy external force. This approximation is
Fig. 2, in. necessary to integrate the load variation during ;
v = Poissons ratio, dimensionless one cycle to determine bearing lifd. .-
For our. discussion, the frequency of the load
.,
e= length ofroI1er, in.
variation and the frequenc~ of- he str~ss cycle ~
P . steady force transmitted by bearing, lb
: repetition at a may be assumed to be two independent
R.l = radius of roller,in.
quantities. According to the law, of. probability,
El = modulus of elasticity ofroller, psi over a long period of time the: (mapitudes of the
. R2= radius of inner race, in. various instantaneous forces rrrksniitted through a
,, E2 = modulus ofelasticity ofinnertace, pai. rolIer at point a can be arranged to produces function
Using these last tw~ eqqatio.ns, we find that for in time having the same generaI sha& as the peri-
, [
a 7-7/8-in. bit loaded with 40,000 lb and r~tating at dic external force function. This observation Ieads
80 rpm the increment of time during which high to a convenient rqethod, which will be used in a later
stresses act between the rollers and inner bearing section, of evaluating the life expectancy of
race (pin) is 0.003 sec. This increment is small bearings in rock bits subjected to variable loacilkg.
compared with the period of about 0.25 sec for
actual bit- load cycles. This means the external DEFINING THE MATHEMATICAL FORM OF 5.
load acting on the, bearing during any single high- MEASURED BIT FORCES
stress increment may ke considered as a constant, In this analy~is, the external load acting on the
as shown in Fig. 3. bearing cohaists of twb component~: a constant
~ other words, as a roller moves past a, the (mean) component (Pm ) and a variable (alternating)
resulting contact stresses can be approximated by component. The alternating component is described
those produced by a constant external force by a ,sine function having an ampIitude of .pv. The.
instantaneous load on the hearing during a loading
D
cycie is expressed by the foHowing equation:
,,
2 mt
,Pi= Pm+ Pvsin--j7. ..j. . . ...(5)

This equation is illustrated by the curve in Fig. 3,


The field-observed cwves of bit-weight variation
do not agree exac~y with the assumed function
defined by Eq. 5. Nonetheless, the final rewlts
are probably within engineering accuracy, since the
form- of the load cycle on a bearing influences
bearing life only slightly.
There are many load functions which will produce
the same effect ,ori bearing life as long as the -
magnitudes of the maximum and minimum forces in
the- 6YcIes are he same. However, if the load ,
t
TIME-
T

SBCON03
4 function deviates radically from the aiwumed
function (Eq. 5), applying resuits of this study
become questionable. In, that case, the general
FIG. 3.- LOAD CYCLE FOR RQLLER BEARING.
approach given hete can be used to develop new
bearing-life data. .!..

EVALUATING LIFE EXPECTANCY

.. . . . .

, ,.

---
.

....
.>
,,, -
..
..---------
1 :.
,
load that will cause fatigue failure in the bearing cal methods using a computer. Fig. 4 shows the
after the same number of revolutions (hours, if resuIts of this computation. - ~
speed is constant) as. a given unsteady load As mentioned earlier, P@ (or @ represents the
defined by Eq. 5. equivalent steady load. that will produce the same :
Derivation of the eqtiivalent load is based on the bearing life as an unsteady load. Eq. 14 describes
premise that if. a constant load acts for a ce~airi how the equivalent load depends on the load
fraction of the life which the besring would have if variation components. The ratio, ~, which is equal
thati ~onstant load were applied for its entire life, to or greater than It not only depends on load
then the same fraction of the life of the bearing is variations but also on the exponent q. King is tests 2
consumed, even though other loads are applied show that practical values of q for unsealed rock-
during the remaining portion of the bearings life. bit bearings range between 1 and 1.5, while &
For another. load, having a .tifferent magnitude, practical value of, q for a roller bearing which is
only the remaining part of the beating life can be properly lubricated throughout its life is about 3.3.
consumed. s The above statement is expressed The better the envirorirnent in which the bearings
mat~ematically by the following equation: operate, the higher the q value. According to Fig.
4, ~ increases with q as well as with a
~+=1, ...:-...;......(;) Fig. 4 shows that when bit-force variations are -
j=~ Li about 2S per cent of the wean there is littIe
where ti = duration that load pi acts incentive for eliminating the dynamic compone~t ,to
x&low an increase in bit weight withour shortening
Li = bearing life corresponding to a constant bit bearing life. To show this, let us assume the ~
Ioad Pi. alternating component of the bearing force is 25
Ift1=t2= t3=. ... =tn=8t, then per cent of the average bearing force (a= 0.25) and
g= 1.4. According to experimental data given
,$ fi= l........:;-:....(s) etklier, these numbers are fairly realistic. The
~=1 equivalent load factor, #, corresponding to these
Substituting fur LJ, we have: numbers is 1.01, indicating that if the average bit
force is 40,000 lb, bir force can be increased 400
lb if che dynamic component is eliminated. This
;J$,yat, =1 . ...(9)....(9) increase. is so small that it could hardly be detected
by present rig instmrnenta~ion.
Taking the limit of this equation as tlt approaches
,,
O, we have:, DETERMINING BEARING LIFE REDUCTION
DUE TO DYNAMIC BIT FORCES
r
L
,,
P9(t)df=c9 ,. . . , . . . . - .(10)
Another way to Iook at the effect of dvnamic
do
lcading~ on ~earing performance is go c~mpare
or . , besring !life corres~onding to dynamic loading with
nT
bearing life corresponding to no dynamic Ioadlng. :
p9(t)-dt=nj-p:, . .. , . ~ . , . ;(ll) Eq. 15,; which follows from Eqs. 1 and 6, makes . I
/0.
this comparison:
- where T = period of one load cycle K=l,/f19, . . . .:. . :. .. .: ... . (15)
nT =~L, total bdaring life, where K is the ratio (bearing life corresponding to
from which: dynamic loading)/(bearing life -corresponding to fio
dynamic loading). K is called the life reduction
coefficient. Values. o.f K vs a and g are given in -
Fig. 5. .,

vq
Substituting Eq. 5 into Eq. 12, we have: Consider the example given above (a = 0.2>~ q =
T 1.4, and average bit for$e = 40,000 lb). FJE. 5
Pv
P;=+
[/( o 1 ~-
.
In

277t q
T
Clt . .(13),
)1
shows that the life reductton ratio corresponding
these condititms is about 0.99. This means bearing
life is decre~sed by I per cent when the bit
to

This equation is more conveniently expressed in


-experiences: bit weight v@.ati?ns .of. 25 per cent. ..-
the dini@ionIess fo-iin: - - ,- .
The experiment@. data in Fi& 1 show a ~fe
exPecmncy Of 22 hours for a rolle~ bearing in a-bit
loaded at 40,900 I&in bsrite mud.- Tfte d&cresseNin ,
bit life (1 per cent of 22 hours, or 13 minutes) due
P Pv
to dynarriic forces at the bit (a = 0~25) wouId: hardly
where ~= --J& a= .. ~ . .be noticed. This exsmple sho~s that the life of .. ___

1
.-A %~ , ... beaFh@-u&d-hiWE& sttiadard-utiset&Fbe=ri&g-bEr--=--
n
. .
.> .- (g= 1.4) is rather insensitive to magnitt!des o! load --
a whole number, as variation of 25 per cent or-less. - .-
shown by the bearitig tests tneri#oned earlier, the

l
- Now qonsider the effe& of larger dynapic bIt
. integral in this equation ,.was. determined by numeri- farce weight variation,,of 100 per cent on unsealed . . ;,, :i
DltCEMBER, 196S . , .,, g7s
-,
. ...2LA. . ?. ....... . . . . . . . ......:-:.:. -.. ~~ . .. . . ... . ....
- ,.4
...
-.. . . ..
.. .. . . .. . .. . . ....-. .,
,. ..

.
bit-bearing life. Assume the mean bit load in this subjected during drilling~- A typical example showed
case remains around 40,000 lb, while the ratio of a decrease in bearing life from 22 to- l$)-hours.
the amplitude of the load variation to the mean load 4.. The life of aeakd-bit bearings is more sensi-
(a) is increased to one. The Iifp-reduction coefficient tive to ,dynamic forces than is the Iif e of wnaealed
corresponding to q = 1.4 and a = 1.0 is 0.88 (Fig. 5). bit bearings., This conclusion suggests that as
Therefore, the predicted life of unsealed bit bear- bit - bearing seals are improved, there is more
ings subjected to the larger load variations is about incentive for decreasing large dynamic bit forces.
19 hours, a significant decrease from ., the 22-ho~lr
life urider steady load. NOMENCLA~URE
Recently; rock bits with sealed bea~ings have ,
a = Pv/Pm, dimensionless
been placed on t!!e market. Field data indicate that
the bearing life of ~ese bits is about 25 per cent 6 = ./%s equivalent load factor; dimensionless
greater than the life of unsealed bits. In some C = bearing wear constant
cases the bearing seal fails before the roller at = time increment, sec ,
bearing has had time to realize ik full potential in A = half of the width of contact between two ,.
a clean eirvironment. cylinders pressed together (Fig. 2), in.
Since the partially effective seaLs anti lubricant Di = outside diameter of. inner race, in.
place the sealed bearings environment, somewhere Do = inside diameter. of outer race, in.
between that of an unsealed bit bearing (q = 1.4)
El = modulus of elasticity of roller, ps;
and that of a bearing which is properly lubricated
E2 = modulus of elasticity of inner race, ~si
throughout its Iife (q = 3.3), it is reasona~le to
assume that the bearing wear exponent (q) for a : K = life reduction coefficient; dimensionless
sealed-bearing bit is greriter than 1,.4 and less thx f? = length of rolIer, in.
313. Ultimately, as seals continue to be improved, f. = bearing life, hr
the bearing-wear exponent for sealed-bearing bits Li = bearing life corresponding to loa~ Pi, hi
could approach 3.3. N ~ = angular velocity of imaginary link reIative to
As the bearing-wear exponent is increased by pin, rprfl
improvements in bearing seals, the effect of
N = arrgrdar velocity of cone reIative to pin, rpm
vibration on bearing life will become more important.
u = Poissons ratio, dimensionless
This is illustrated by a hypothetical example in .,
which a sealed bearing has a wear ex~otrerx (q) Of Pe = equivalent be&ing load, lb J ~
2.5 and a 50-hour life under a steady load of 40,000 Pi = ins~antaneous load on roller bearing, lb
lb. The effect of severe down-hole vibrations (a = Pm = mean bearing load, lb
1.0) would be to shorten bit life (life reduction PV ~ variable bearing ~oad component, lb
coefficient, K, of O.52), from 50 to 26 hours. . P = steady force transmitted by bearing, lb-
~. ,.
p -_, _coNcL-uslf3Ns q = bearl~g
.. weir, exponent, dimensionless
l?l = radius of roller, in. , .
1. The
life of unsealed bit bea;ings ii not affected
appreciably by the kind of dynamic bit loading (25 R2 = radius ofin~er race, ih.
per cent variation) to which rock bits are usually ~ b = contact stress, psi
subjected during drilling:- t = tim~, sec .,
2. Environment has a much greater effect on the &i = duration that ~oad Pj ~cts
life of unsealed bearings than does-dwiic bit T = period of load cycle, sec
loading.
V = linear velo-ci~ of roller surface relative to
3. The Iife of unseaIed bit bearitigs is Iow&ed
imaginary link; in. /see
appreciably bY the larger dynamic Ioading (100 Per
cent variation) to which rock bits are occasionally REFERENCES

1. Besring Catalog No. 425, SKF Indtistries Inc.,


Philadelphia, Pa, (1960), , ,
2. King, G, R: [%f feet of Fluid &vironment on Rock -13it
Bearing Performancetz, AAODC! Paper (1959).
--3. .Prdr&ren,. A.: Ball aa~ .:RoIler Beairkg Engineering,. .
,.
~. 3rd Ed., SKF Indu+ries Inc., Philadelphia, % (1959).
4. Dareing, IL W. and RAdzimovsky, E, L; Misaligned .-._.r
Roller Bearipgs -How Faak Will They Fai~; h!acbisw - ... . . ..
Design (Feb. 13, 1964). . .

6.
-l%:
5, Ham, C. W., Crsrte, E. J, and Rogers, W. f+ Mechanics
oj Macbfaefy, 4th Ed., McGraw-Hill Book Co,, Inc.,
!
..5 e. . - , , I L
N. Y. (1458). ~ -
.,

e%
...-* k - L- -. .-. . . -2:---- .. . ---- .ti -- -
,~. , Gesarmnefte Werke, Lei~z\g (1885) Vol. 1. .
% 0.: 0s 0.s 0,4 0s on O.r 0,8 0,s :.0

l
-w
7. Radztiovsky, E, 1,: Wires ~ Distri&ition arid S~reng&
- L*D AMPUTLIIWRA?1O - 01 Mi3t81t#Ltss
Condition of Two Rolii.ng C!yllnders Pressed Togetflert$, .,- :. . ,m
FIG, 5-- B13ARfN~ LI@E REDUCTIOiJ DUE TO UN-
. . Engr. Experiment StUtiO.n Brill,, -s0, No, 408, U. of -
STEADY LOADING. Illirrdis (19s3), .* *-*

Das könnte Ihnen auch gefallen