[email protected]@generic-67fc6ada12dd.pdf
TRANSCRIPT
7/23/2019 [email protected]@generic-67FC6ADA12DD.pdf
http://slidepdf.com/reader/full/100000tribologyasmedigitalcollectionasmeorggeneric-67fc6ada12ddpdf 1/5
J .W.Lund
Department of Machine Elements,
The Technical U nivers i ty of Denmark,
2800 Lyngby, Denmark
Review
of the
C oncept
of
Dynamic Coefficients for
Fluid Film Journal
Bearings
The developmen t of the concept of spring and dampin g coe fficients for journal
bearings is briefly reviewed. Metho ds for compu ting the coefficients are described,
and their use in rotor dynamics calculations (unbalance response, stability) is
discussed. The limitations imposed by nonlinearities on the application of the coeffi
cients is illustrated by examples.
Introduction
Th e id ea o f r ep resen t in g t h e d y n am ic r esp o n se
characteristics of a journal bearing by means of stiffness and
damping coefficients originates with Stodola
[ ]
and H umm el
[2].
Their aim was to improve the calculation of the crit ical
speed of a rotor by including the flexibility of the bearing oil
film. Concurre ntly, New kirk [3, 4] described the phenom enon
of bearing induced instability, which he called oil whip, and it
soon occurred to several investigators that the problem of
rotor stabili ty could be related to the properties of the bearing
coefficients as can be seen from numerous references in the
reference list.
In the initial stage, one of the difficulties was that the only
available solution of Reynold's equation was that of Som-
merfeld for the infinitely long bearing, sometimes modified by
a suitable end leakage factor, and tha t the only calculations of
rotor dynamics was that of the first critical speed, using the
methods of either Rayleigh or Stodola.
In the late forties, more advanced rotor dynamics calcula
tion methods were developed, and with the advent of the com
puter in the fifties, it soon became feasible at little cost to ob
tain numerical solutions of Reynold's equation and to per
form more elaborate rotor calculations.
Even so, the concept of bearing coefficients was not im
mediately accepted, probably because the load-displacement
characteristic of a journal bearing is so evidently non-linear.
Experience, however, has demonstrated the practical
usefulness of the coefficients, and modern rotor dynamics
calculations are firmly based on the concept.
Today the emphasis is on experimental measurements of the
bearing coefficients and establishing more uniform agreement
with calculated values. The need for some improvements or
refinements in the theory seem indicated, b ut the basic calcula
tion method appears sound.
The appended reference list gives some, but far from all of
Contributed by the Tribology Division of THE
AM ERICAN SOCIETY
O F
M ECHANICAL ENGINEERS and presented a t the A SM E/A SL E Joint Tr ibology
Conference, Pittsbu rgh, Pa. , Octob er 20-22 , 1986. M anuscript received by the
Tribology Division M arch 14, 1986. Paper N o. 86-Trib-48.
BEAR ING
CENTER
STAT IC EQU IL IBR IUM
.LOCUS CURVE
• x
F i g . 1 C o o r d in a t e s y s t e m
the papers that mark the development of the concept of bear
ing coefficients. L it t le attem pt has been made of a historical
review, and there is no systematic scheme behind the selection
of the papers. The cited works, however, are sufficiently
representative to allow compiling a com prehensive l ist by com
plementing the present reference list by those of the papers.
Analys is
In an x-j '-coordinate system with origin in the bearing center
and the x-axis in the static load direction (see Fig. 1), the reac
tion forces from the lubricant film are given by:
/• LII |> e
J
LI I
Jfl
cos(T]
\RdQdz
sinflj
(1)
where /? is the pressure in the fi lm, R is the journal radius, L is
the axial length, z is the axial coordinate, and 9 is the cir-
Journal of Tribology
JANUARY 1987, Vol. 10 9/3 7
Copyright © 1987 by ASME
wnloaded From: http://tribology.asmedigitalcollection.asme.org/ on 10/27/2014 Terms of Use: http://asme.org/terms
7/23/2019 [email protected]@generic-67FC6ADA12DD.pdf
http://slidepdf.com/reader/full/100000tribologyasmedigitalcollectionasmeorggeneric-67fc6ada12ddpdf 2/5
cumferential angular coordinate, measured from the negative
x-axis. The film extends from 0, to d
2
where both angles may
be functions of z.
The film pressure derives from Reynold's equation:
dp
h* dp
Rd d Vl2 / i Rdd J dz V 12^7 dz
+
U
) -
ViQ-
dh dh
~W
2)
where / is t ime, n is the lubricant viscosity, Q is the jo urna l
angular speed of rotation, and h is the film thickness:
h = C + xcosB + ysind
3)
C is the radial clearan ce, and x and y are the coordinates of the
journal center.
It appears that the reaction forces are functions of x and y,
and of the instantaneous journal center velocities, x and y
( dot indicates t ime derivative). Hence, for small amplitude
mot ions,
Ax
and
Ay ,
measured from the static equilibrium
position (x
0
, y
0
), a first orde r Taylor series expa nsion yields:
+ K
xx
Ax+K
xy
Ay + B
xx
Ax+B
xy
Ay
F
y
= F
yo
+ K
yx
Ax+K
yy
Ay + B
yx
Ax+B
yy
Ay
4)
where the coefficients are the partial derivatives evaluated at
the equilibrium position:
K
r]
EL)
dy Jo
B
r
\ dy Jo
5)
and analogously for the remaining coefficients.
Because (x
0
, y
0
) is the equilibrium position, then F ̂ = 0
while F ̂ equals the static load, W.
In the literature the coefficients are often computed directly
by numerical differentiation. While this may be adequate for
practical purposes, the inherent numerical inaccuracy of the
method can be eliminated by employing a perturbation solu
tion. Thus, equation (3) may be written as:
where:
h = h
0
+ Ah
h
0
= C+x
0
cos8+y
0
sinf?
Ah = A xcos0 + A ysinfl
dh
~dT
= A icosS + AysinO
(6)
(7)
(8)
(9)
The perturbation in film thickness gives rise to a similar per
turbation in the fi lm pressure:
P =Po +
A
P
(10)
where:
A/7
=
p
x
'Ax+p
y
-Ay+p
x
-Ax+p;'Ay (11)
By substituting equation (6) to (11) into Re ynold's equatio n,
equation (2), and by retaining first order terms only, five
equations are obtained:
R{ p
0
) = V2Q
dh
0
dd
••
1
/2O (-;c
o
sin0+.)>
o
cos0)
R{ p
x
h
0
dd
, K dp
0
12ji Rdd
d / cos0\
Rdd \ h
n
J
VittsmO
(12)
(13)
R{ p
y
} = -1-
sin0
^7
<Ati
/ i
3
0
dh
0
dd
dPo
12/t Rdd Rdd \ h
f s i n 0 \
\~hTJ
+ ViUcosd (14)
R [p
x
} =
cosO
R {p'
y
} = sin0
where the left-hand side operator is:
d ( h\ d \ d
R[
/ dz V 12u dzJ
(15)
(16)
(17)
Rd6\l2n Rdd/ dz V 12/i d z /
The boundary conditions are that the pressure is zero at the
edges
{p
0
=
Ap
= Q a t
z
= ±
ViL)
and equals the feeding
pressure at any supply grooves.
When film rupture occurs in the divergent fi lm region, the
location of the trail ing end bounda ry is governed by the condi
tion of zero pressure gradient, requiring some iterative pro
cedure to determine the actual boundary curve. If (0
O
, z
0
) is a
point on the curve, the pressure in a neighboring point is, to
the first order:
p(6
0
+ A6,z
0
+ Az) =p(6
0
,z
0
) + ( - ^ - )
o
A8
( -^ - )
Q
A Z
=Po(.e
Q
,z
0
) + Ap(d
0
,z
0
)+ (-jf)
Q
Ad+ (-j£-)
0
A z (18)
With p
0
(6
0
, z
0
), (dp
0
/dd)
0
, (dp
Q
/dz)
0
being zero, and by
assu m in g th a t t h e p ressu re o n t h e b o u n d ary , p(8
0
+
A0,Zo + A z), is zero also und er dynamic conditions then the
boun dary con dition for the perturbed pressu re is that A p is
zero on the statically determined c urve. In case of film ruptu re
(film contraction) ahead of the leading edge of the film, the
location of the boundary curve is governed by flow
equilibrium such that the pressure gradient is not zero. Even
so ,
the condition of Ap being zero on the boundary curve is
sufficiently accurate for most purposes.
By combining equations (1), (4), (10), and (11) it is found
that:
K„
K
v
,
f
L /2 j
e
2
J -111 J 8,
X
cosf?
sing
VRdddz
(19)
and analogously for the remaining coefficients.
Because they derive from self-adjoint operators, the cross-
coupling damping coefficients are equal:
B
yx=
B
xy (20 )
The damping matrix, therefore, is symmetric and possesses
principal directions. This is not true for the stiffness matrix w h
ere the cross-coupling stiffness coefficients are unequal:
K
yx*K
xy
(21)
Thus, the stiffness matrix is nonconservative as is readily
demonstrated by considering a closed whirl orbit with angular
whirl frequency co:
Ax= \Ax
I cos pit + 4>
x
)
(22)
Ay= \Ay\sin(wt
+ 4>
y
)
where <j>
x
and <j>
y
are phase angles. The energy dissipated over
one cycle is:
Ediss = ™(Bxx\Ax\
2
+B yy\Ay\
2
- {B
xy
+ B
yx
) I Ax I • IA^ I sin(<k - 4>
y
)
- ir K
xy
-K
yx
) \Ax I . I Ay
I
cosfo , - </>,)) (23)
With K
xy
po sitive, the cross-c oupling stiffness coefficients
may provide negative damping or, in other words, they can
supply energy to the motion. The last term may also be written
as:
ir K
xy
- K
yx
){ - I Ax j I
2
+ I Ax
b
\
2
) (24)
where the ell iptical whirl orbit has been decomposed into two
circular orbits, one with forward whirl direction and
amplitude \Ax
f
\, and one with backw ard whirl direction and
3 8 / Vo l . 1 0 9 , JA NUA RY 1 9 8 7
T r a n s a c t i o n s o f t h e A S M E
wnloaded From: http://tribology.asmedigitalcollection.asme.org/ on 10/27/2014 Terms of Use: http://asme.org/terms
7/23/2019 [email protected]@generic-67FC6ADA12DD.pdf
http://slidepdf.com/reader/full/100000tribologyasmedigitalcollectionasmeorggeneric-67fc6ada12ddpdf 3/5
amplitude \Ax
b
I. Thus i t is the forward whirl com pone nt
which is responsible for the negative damping.
It is customary to analyze the stability of a bearing by
assigning a mass M to the journal such that the equations of
motion become:
MAx
+
B
xx
Ax
+
K
xx
Ax
+
B
xy
Ay
+
K
xy
Ay
=
0
MAy + B
yy
Ay + K
yy
Ay + B
yx
Ax + K
yx
Ax = 0
These equations admit of solutions for Ax and Ay of the form
e
5
' where:
s = \ + iu (i = ̂ l ) (26)
For a nontrivial solution, the eigenvalues s are the roots of
the determina nt of the coefficient matrix. A t the threshold of
instability, X is zero whereby equation (25) takes on the form:
of frequency. For a plain journal bearing B
mm
is always
positive, while B
min
is negative for frequencies smaller than co
0
(for o>
= oi
0
,
B
min
equals zero and
K
min
equals
K
0
).
Hence, if
the lowest natural frequency of the system, co
nat
, is less than
co
0
, self-excited whirl can be sustained. The speed, fi
0
, at which
instability sets in, may then be expressed as:
Q
n
35)
( o )
0
/ Q )
0
A typical value for the whirl frequency ratio , (co
0
/Sl)
0
, is
around 1/2 from which stems the rule-of-thumb that a rotor
becomes unstable when the speed reaches twice the first
critical speed (~co
na t
) .
(Z
xx
-u
2
M)
where:
(Z
yy
-u
2
M)
--K
rr
+ iuB
x
Ax
Ay
= 0
(27)
(28)
and similarly for the other impedances.
For a chosen static equilibrium position (Sommerfeld
number), the values of the coefficients are given, and setting
the determina nt equ al to zero determines that value of M , call
ed M
crit
> which makes the specified eq uilibrium position
unstable. The solutions for M
crk
and the associated whirl fre
quency,
O
= CO
0)
are:
6
M
c r i t =
*^xx yy **^yy**xx *^xy yx ^yx&x.
^ xx • **yy
i
K
xx — « o ) ( ^ v v
_
*o ) - K
xy
K
K
0
0 =
R R
^xx^yy
' ^xy^yx
(29)
(30)
In order to establish on which side of the threshold the be ar
ing becomes unstable , the determ inant, A , may be expande d
around the threshold value as:
A s A o + f — ) >ds+(-4^r) -dM (31)
0
Vidco/o \ dM Jo
With A and A
0
equal to zero, the equation may be solved:
(-
dM
( -
dX
dM
dw
~ d ~
/ o
=
\ dw )ol\ dM / o
from which
dX
ViufcBxx + B
yy
){B
xx
B
yy
-B
xy
B
yx
)
<ft\ Viu%(B
xx
+
B
yy
\B
xx
B
yy
-
\dM Jo ufcB
xx
+ B
yy
)
2
K
2
0
+ {K^K^-
K
xy
K
yx
-K\f
(33)
J
yyf
,v
0 ' \**xx -yy **xy
x
*yx
This value is positive which means that an increase, dM , in
journal m ass above the crit ical value results in a positive value
of X or, in other wor ds, causes the bearing to become unstab le.
In equation (27) the term - w
2
M represents the rotor im
pedance at the journal . It may be thought of as being opposed
by some equivalent bearing impedance Z which, then, is a
solution to the determinantal equation:
Z= a ( Z „ + Z
y
,)±J—(Z„ - Z
yy
)
2
+
Z
xy
Z
yx
•^max ~~ ^m ax •
(
™ m a i
= K
mm
+ iwB„
(34)
The effective stiffnesses, K
mm
and -K
min
, and the effective
damping coefficients, B
max
and B
min
, are of course functions
Dis c us s ion
The preceding stability analysis applies only to a rigid rotor
which is symmetric and supported in two identical bearings.
Hence, the results are of limited usefulness, and in general it
can be quite misleading to rank various bearing types for
stability on the basis of the critical mass.
A s seen from equation (29), the crit ical mass value is a
measure of the stiffness of the bearing, but an increase in bear
ing stiffness does not ensure that the lowest natural frequency
is raised to improve the instabili ty threshold speed.
Similarly, equation (30) suggests that bearings with a low
whirl frequency ra tio may be more stable. For a flexible roto r,
however, it is found that it is the whirl frequency itself, equal
to the lowest natural frequency, which stays unchanged as the
rotor becomes unstable, not the whirl frequency ratio .
Furthermore it should be noted that associated with the
calculated instabili ty threshold is a characteristic whirl orbit ,
obtained as the solution for
Ay/Ax
from equation (27) with
the determinant being zero. For a flexible, non-symmetric
rotor, the modeshape will not conform to the characteristic or
bit, and thereby the effective stiffness and damping stability
properties of the bearing change.
For a practical rotor there is no real substitute for a stabili ty
calculation in which the roto r, with i ts bearings represented by
their coefficients, is modeled as close to the actual design as
feasible. Such calculations are today standard procedure when
designing high speed machin ery. The results are in sa tisfactory
agreement with experience, considering the uncertainties and
often poorly established tolerances on bearing geometry,
clearance, and pedestal st iffness.
A different type of calculation, bu t with the same rotor-
bearing-system model, is the one for the response to mass un
balance. It is used to check the unbalance sensitivity of the
rotor and, also, to determine the damped crit ical speeds.
It is inherent in the definition of the bearing stiffness and
damping coefficients t hat , mathe matically, they are only valid
for infinitesimal amplitudes. Comparisons with exact solu
tions, however, prove that for practical purposes the coeffi
cients may be employed for amplitudes as large as, say, forty
percent of the clearance.
A n example is shown in Fig . 2 which is for a two-axial
groove journal bearing with a length-to-diameter ratio of
X
A.
The journal is forced to whirl synchronously by a rotating un
balance force, equal to half the static load, while the inertia of
the journal is ignored.
The elliptical whirl orbit, calculated on the basis of the eight
coefficients belonging to the indicated static equilibrium posi
tion, is compared to the whirl orbit obtained from a true
nonlinear solution. It is found that the amplitudes, whith a
maximum value of 38 percent of the clearance, are in close
agreement, but, as expected, there are also deviations.
First , the center of the orbit moves closer to the bearing
center to satisfy overall dynamic equilibrium:
Journal of Tribology
J AN U AR Y 198 7, Vo l . 1 0 9 / 3 9
wnloaded From: http://tribology.asmedigitalcollection.asme.org/ on 10/27/2014 Terms of Use: http://asme.org/terms
7/23/2019 [email protected]@generic-67FC6ADA12DD.pdf
http://slidepdf.com/reader/full/100000tribologyasmedigitalcollectionasmeorggeneric-67fc6ada12ddpdf 4/5
-0.2 -0.1 R 0.1 0.2 0.3 O.t 0.5 0.6 0.7 OB 0.9 10
STATIC LOAD DIRECTION
F i g .
2 U n b a l a n c e w h i r l o r b i ts . C o m p a r i s o n b e t w e e n l i n e a r a n d
n o n l in ear an alysis .
F
r
ds=W
F
y
ds = 0
(36)
where s is the arc length along the path of the orbit . Second,
the shift in orbit center is accompanied by a change in orbit
orientation which affects the agreement in phase angle. Fin-
nally, it is evident that the non-linear orbit contains higher
order harmonics which of course are absent in the l inear
solution.
It should be emphasized, however, that in the matter of
greatest importance, namely the size of the amplitude, the
agreement is close. This holds true in the range of interest in
practice where amplitudes larger than, say, 40 percent of the
clearance would rarely be tolerated.
For even larger amplitudes, the various phenomena
associated with non-linear vibrations become evident. A s an
example, Fig. 3 i l lustrates the existence of multiple so lutions in
certain operating regions. The diagram is computed for the
idealized case of an unloaded, infinitely short journal bearing
where the jour nal , with mass M , whirls synchronously in a cir
cular orbit whose radius, normalized with respect to the radial
clearance C, is given by the ordinate.
The curves in the diagram are labeled by the dimensionless
unbalance mass eccentricity
p,
normalized by
C.
The abscissa
gives the dimensionless journal mass m = CMQ,
2
/oW where <r
is the short bearing Sommerfeld number:
f iURL / R
W
( 4 - ) ( 4 - )
37)
Hence, for a particular rotor and bearing, the abscissa is
simply proportional to the rotational speed.
It is seen that, for some given mass un balan ce, there will be
a speed beyond which more than one solution is possible and
where jump phenomena may occur . The l inear ized so lu t ion
with stiffness and damping coefficients agrees well with the
small amplitude orbit , but i t is of course unable to predict the
concurrent orbits with larger amplitudes.
Similarly the previously discussed stability analysis, based
on the liniearized coefficients, is concerned only with the
stabili ty of the static equilibrium and the inception of whirl ,
and it can not be used to study whether that whirl orbit may
actually be contained and stable (l imit cycle). Furthermore,
while the l inear analysis predicts that an unloaded journal
bearing is inherently unstable, the stabili ty curve in Fig. 3
demonstrates that stable synchronous whirl orbits may exist
10 10* 10'
t
DIMENSIONLESS JOURNAL MA SS : i^ -^ T?
F i g .
3 C i rcu lar u n b alan ce wh i r l o rb i ts fo r u n lo ad e d , sh o r t jo u rn a l b ear
ing with r ig id rotor
for p greater than 0.13 or, in other words, the bearing can be
stab i l ized by adding unbalance.
The nonlinear behaviour, however, is rarely of much con
cern in practice, and for practical purposes the instabili ty
threshold c ompute d on the basis of the l inearized coefficients
should be adopted as the design criterion. Past the threshold
speed, self-excited whirl takes place in one form or another,
and although numerous cases are on record of safe operation
in this region, at least as many cases of uncontrollable whirl
can be cited.
The threshold based on the coefficients proves in general,
therefore, to be a conservative estimate. There remain,
however, some basic questions about applying Reynold's
equation. The major problem is the physics of fi lm rupture
which has yet to be clarified, and which makes uncertain what
the proper boundary condi t ions are under dynamic load . A s a
curiosity, another example is the question of the thermal con
ditions prevailing under synchronous whirl when the orbit en
circles the bearing center.
Finally, the four stiffness coefficients can be derived direct
ly from the static equilibrium locus curve. It is unfortunate,
therefore, that experimental measurements of the curve all too
frequently deviate appreciably from the theoretical curve. The
measurements are admittedly not easy to make, and the ex
perimental tolerances are difficult to ascertain, but enough
results have been published to indicate that the discrepancies
are real .
Under those conditions, experimental measurements of the
coefficients (typically by applying a vibratory force and
measure the resulting response) can not be expected to agree
well with the theoretical values. Th e experiments are no longer
a test of the validity of the dynamic analysis; rather, they
become indirect evidence of discrepancies in the more fun
damental solution of Reynold's equation for static conditions.
Conc lus ion
The concept of stiffness and damping coefficients for jour
nal bearings has proven very fruitful, and modern rotor
dynamics calculations for unbalance response, damped
natural frequencies, and stabili ty are based on this concept.
The theoretical limitation of small amplitudes is of little im
portance in practice, but the analysis sti l l needs imp rovem ents,
for example in refinements in the model for fi lm rupture.
In recent years, extensive experimental programs have been
carried out to measure the bearing coefficients, not just in the
laboratory, but also for large industrial bearings. While many
results show good agreement with the theoretical values, cases
are sti l l reported of discrepancies that should be resolved.
4 0 / V o l .
109 , JAN U AR Y 1987
T r a n s a c t i o n s o f t h e A S M E
wnloaded From: http://tribology.asmedigitalcollection.asme.org/ on 10/27/2014 Terms of Use: http://asme.org/terms
7/23/2019 [email protected]@generic-67FC6ADA12DD.pdf
http://slidepdf.com/reader/full/100000tribologyasmedigitalcollectionasmeorggeneric-67fc6ada12ddpdf 5/5
The question of the influence from therma l and elastic
deformations needs further consideration as do also the prac
tical problems
of
man ufacturing and operating tolerances
on
bearing geometry, clearance, and lubricant viscosity.
References
1 S todola , A., K ristische Wellenstorun g infolge der Nachgiebigkeit des
O elpols te rs im L ager , Schweizerische Bauzeiting, Vol . 85, 1925, pp . 265-2 66.
2 Hum m e l , C, K r is t ische Drehzahlen als Folge der Nachgiebigkeit des
Schmiermittels im L age r, VDI-Forschungsheft 2 87, 1926.
3 Newkirk, B. L., S ha f t Whipp ing , General Electric Review, 1924, p.
169.
4 Newkirk, B . L . , and Taylor , H. D. , Shaf t Whipping due to O i l A c t ion in
Journa l Bear ings , General Electric Review, 1925, pp. 559-568.
5 Rober tson, D., Whir l ing of a Journa l in a Sleeve Bearing,
Philosophical Magazine, Series 7, Vol. 15, 1933, pp. 113-130.
6 Hagg, A . C . ,
'
'The Influence of O il-Film Jo urnal Bearings on the Stability
o f R o tat i ng M a c h ine s , A S M E Journal of Applied Mechanics, Vol. 13, 1946,
pp.
A-211 to A - 2 2 0 .
7 Ha gg , A . C, and Warner , P. C, O i l Whip o f Flexible Rotors , Trans .
A S M E, Vo l. 75 , No . 7, 1953, pp. 1339-1344.
8 Ha gg , A . C, and Sankey, G . O . , Some Dynamic Proper t ies of Oil-Film
Journa l Bear ings with Reference to the Unba lance Vibra t ion of Ro tors , A SM E
Journal of Applied Mechanics, Vol. 23, 1956, pp. 302-30 6.
9 Ha gg , A . C , and Sankey, G . O . , Elas t ic and Damping Proper t ies of O i l -
Film Journal Bearings for A ppl ica tion to Unbalance Vibra t ion C alcula t ions ,
A S M E Journal
of
Applied Mechanics, Vol. 25, 1958,
p.
141.
10 Por i tsky, H. , Contr ibu t ion to the Theory o f O i l W h i p , Trans. ASME,
Vol. 75, 1953, pp. 1153-1161.
11 P inkus , O . , No te on O i l Wh ip , A S M E Journal of Applied Mechanics,
Vol. 20, 1953, pp. 450-451 .
12 Pestel, E., Bei t rag zur Ermit t lung der Hydrodynamischen Dampfungs
und Federeigenschaften von Gle itlagren , Ingenieur-Archiv, Vol . XX II , No. 3,
1954,
pp. 147-155.
13 Cam eron, A . , O il Whirl in Bearings-Theoretical Deduction of a Further
Cr i te r ion, Engineering, L ondon, Feb. 25, 1955, pp . 237-239.
14 K orovchinski i , M . V. , S tabi l i ty of Position of Equilibrium of a Journa l
on an O i l F i lm, Friction and Wear in Machinery , Vol. 11, 1956, pp . 248 -305.
15 Tondl , A . , The M otion o f a Journa l in a Bearing in the Unstab le R egion
of Equilibrium Position of the Center of the Jou rna l , Ninth Inte rna t iona l
Congress o f A ppl ied M echanics, Vol . V, 1957.
16 Tondl , A . ,
Some Problems of Rotor Dynamics,
Publ ishing House
of
the
Czechoslovak A cademy of Sciences, Prague, 1965.
17 Hori,
Y., A
Theory
of Oil
W h i p , A S M E
Journal of Applied
Mechanics, Vol. 26, 1959, pp. 189-198.
18 K ramer , E., Der Einfluss des Olfilms von G leitlagern auf die
Schwingungen von M aschinenwellen, VDI-Berichte, Vol. 35, 1959.
19 Sternlicht, B., Elast ic and Damping Proper t ies of Cylindrical Jo urna l
B e a r ings , A S M E Journal of Basic Engineering, Vol. 81, 1959, pp . 101-108.
20 Capriz, G., O n the Vibra t ions of Shafts Rotatin g on L ubricated Bear
ings , Annali de Matema tica Pu ra
e
Applicata, IV, Ser. 50, 1960, p. 2 2 3 .
21 Lund, J. W. , and Sternlicht, B. , Roto r-Bearin g Dynamics with Empasis
o n A t t e nu a t i on , A S M E Journal
of
Basic Engineering, Vol. 84, No . 4, 1962,
pp. 491-502 .
2 2 L und , J. W., Sprin g and Damping Coefficients for the Tilting Pad J our
na l Bear ing, Trans. ASLE, Vol. 7, 1964, pp. 342-352.
2 3 L und , J. W., Rotor -Bear ing Dynamics Design Technology. Par t III:
Design Handbook for Fluid F i lm Type Bear ings , Repor t AFA PL -TR-65-45,
Par t I I I , A ir Force A ero Propuls ion L abora tory, W right-Pa tte r son A ir Force
Base , O hio, M ay 1965.
24 Holmes, R., O i l -Whir l Charac te r is t ics of a Rigid Rotor in 360 Journa l
Bear ings ,
Proced. of
the Institution
of
Mechanica l En gineers, Vol. 177, No .
11, 1963, pp. 291-307 .
25 Holmes, R . , Non-L inear Per formance of Turbine Bear ings , Journal of
Mechanical Engineering Science, Vol . 12 , No. 6, 1970, pp. 377-379.
26 Someya, T. , Stability of a Balanced Shaft Run ning in Cylindrical Jour
na l Bear ings ,
Proc. of the Second L ubrication and Wear Conference,
The In
stitution of M echanica l Engineers, L ondon , 1964, Paper No. 21 .
27 Warner , P. C, Sta t ic and Dynamic Propert ies of Partial Jo urnal Bear
i n g s , A S M E Journal
of
Basic Engineering, Vol. 85, 1963, pp. 247-2 57.
28 Warner , P . C , and Thom an, R . J. , Th e Effect of the 150-Degree Part ial
Bearing on Rotor -Unbalance Vibra t ion, A SM E Journal of Basic E ngineering,
Vol. 86, 1964, pp. 337-347.
29 Glienicke, J., Exp erimen tal Investigation of the Stiffness and Damp ing
Coefficients of Turbine Bea r ings ,
Proced.
of the Institution of Mechanical
Engineers, Vol. 181, Part 3B, 1966-67, pp. 116-129 .
30 Orcut t , F. K., and A r w a s, E. B., The Steady-State and Dynamic
Characteristics of a Full Circular Bearing and a Partial A rc Bearing in the
L a m ina r and Tur bu le n t F low R e g im e s , A S M E
JOURNAL
O F
L UBRICATION
TECHNOLOGY, Vol. 89, 1967, p. 143-153.
31 L und , J. W. , Ca lcula t ion of Stiffness and D amping P roperties of G a s
Bearings, ASME J O U R N A L O F L U B RI CA T I O N T E C HN O L O G Y , Vol. 90, 1968, pp.
793-803.
32 Smith, D. M . , Journal Bearings in Turbomachinery, Chapm an and Hal l
L td . , London, 1969.
33 O t a , H., and Inagaki, M., Elast ic and Damping Proper t ies of Journa l
Bear ins , Bulletin of the JSME, Vol. 17, No . 108, 1974, pp. 693-700 .
34 M or ton , P. G., The Der iva tion of Bearing Characteristics by M eans of
Transient Excitation A pplied Directly
to a
Rota t ing Shaf t , Dynamics
of
Rotors , IUTA M S ymposium 1974, L yngby, Denma rk. Springer-Verlag 1975,
pp.
350-379.
35 M alcher , L., Die Federungs- und Dampfungseigenschaften von
G leitlagern fur Turbom aschinen, Disse r ta tion, Fakul ta t fur M aschinenbau,
Technischen Hochsch ule Karlsruh e, 1975.
36 Reinhardt, E., a nd L und , J. W. , The Influence of Fluid Inertia on the
Dynamic Properties
of
Jou r na l B e a r ings , A S M E
JOURNAL
O F
L UBRICATION
TECHNOLOGY, Vol. 97, 1975, pp. 159-167.
37 L und , J. W. , and Thomsen, K . K . , A Ca lcula tion Method and Data for
the Dynam ic Coefficients of O i l -Lubr ica ted Journa l B ear ings , Topics in Fluid
Film Bearing and Rotor Bearing System Design and Optimization, A S M E ,
1978,
pp. 1-28.
38 K irk , R. G . , The Influence of M anufac tur ing Tole rances on M ult i -Lobe
Bear ing Performance in Turbomach inery, ib id . , pp . 108-129.
39 Hisa, S., M atsuura , T . , an d Someya , T., Exper iments on the Dynamic
Characteristics of Large Scale Journa l Bear in gs ,
Proced.
Second International
Conf. on
Vibrations in Rotating Machinery , Cambr idge , England, The Inst itu
tion of Mechanica l Engineers, 1980, pp . 2 23-230.
40 Nor dm a nn , R., and Schol lhorn, K., Identification of Stiffness and
Damping Coefficients of Journa l Bear ings by M eans of the I m pa ct M e thod ,
ibid . , pp . 231-238.
41 Diana, G., Borgese , D. , and Dufour , A . , Exper imenta l and A na lyt ica l
Research on a Ful l Sca le Turbine Journa l Bear ing , ib id . , pp . 309-314.
42 Hashimoto, H., W a d a , S., and M a r uk a w a, T., Per formance
Characteristics of Large Sca le Ti l t ing-Pad Journa l Bear ings , Bulletin of the
JSME, Vol. 28, No . 242, 1985, pp. 17 61-1767.
J ou r na l o f T r i b o l o g y
J AN UA RY 198 7 , V o l . 1 0 9 / 4 1