Rtiwari RD Book 06b
Rtiwari RD Book 06b
Rtiwari RD Book 06b
in)
0.6 0.7. Journal bearing can have a significant effect on a machines vibration characteristics. The
oil film behaves like a complicated arrangement of springs and dampers and so influences the
machine critical speeds and imbalance response. Moreover, bearing fluid film forces can cause rotor
instability, which results in serious levels of self-excited vibration (e.g. oil-film whirl). The most
common forms of journal bearing are given in Figure 6.12.
Plain cylindrical
3 lobe
Circumferential
Partial arc
4 lobe
Offset halves
Lemon bore
Tilting pad
288
Partial-Arc Bearings
In another variation of the plain cylindrical bearing the lower part of the bearing bush bore is relieved.
Typically, 60 1200 of included arc of surface are left to carry the load whose direction is virtually
constant. The advantage obtained by relieving the bush bore is to reduce friction losses in the bearing
whilst retaining most of the bearing load-carrying capacity.
289
x = u0 + k xx x + k xy y + c xx x + c xy y + m xx x + m xy y
y = v0 + k yx x + k yy y + c yx x + c yy y + m yx x + m yy y
where
x
k xy =
y
( u 0 , v0 )
x = f x mx
(1)
x
c xy =
y
( u 0 , v0 )
y = f y my
x
m xy =
y
( u 0 , v0 )
and analogously the remaining bearing dynamic coefficients may be defined. In the matrix form of
equation (1) all diagonal terms are called direct coefficients and off-diagonal terms are called crosscoupled terms. The latter terms arise due to the fluid rotation within the bearing. is the reaction
force of fluid film on the journal, f is the external excitation force on the journal, m is the journal
mass, u0 and v0 are the static equilibrium position of the journal from the bearing center, x and y are
the displacements of the journal from its static equilibrium position, x and y are the instantaneous
journal center velocities and x and y are the instantaneous journal center accelerations, in the vertical
and horizontal directions respectively. The dot indicates the time derivatives and kij, cij and mij (i, j =
x, y) are bearing stiffness, damping and added-mass (also termed the virtual fluid-film mass or inertia)
coefficients respectively. The indices of the stiffness, damping and added-mass coefficients have the
following significance: the first index gives the direction of loading which produces the elastic
(damping/inertia) force and the second index gives the direction of the displacement
(velocity/acceleration). Because (u0, v0) is the equilibrium position, then u0 equals the static load, W.
The equilibrium position depends on a unique value of the dimensionless Sommerfeld number
( S = ( RL / W )( R / cr ) 2 ( L / D ) 2 ), where is the lubricant viscosity, is the journal rotational
speed, D is the bearing bore, R is the journal radius, L is the bearing length and cr is the bearing radial
clearance. The Sommerfeld number, S, defines the operating conditions (speed, lubricant viscosity,
static load and geometry). The dynamic coefficients are evaluated for a particular static equilibrium
position, which is a function of the Sommerfeld number, S, which means that for a given application,
they are functions of rotor speed. Moreover, bearing dynamic coefficients could be external excitation
frequency, , dependent.
It should be noted that equation (1) is a complete form of linearised fluid-film dynamic equation and it
contains twelve stiffness, damping and added-mass coefficients. Consistent with the assumptions
inherent in reducing the Navier-Strokes equations to the Reynolds equation, the conventional laminar,
thin film lubrication theory ignores the inertia forces in the fluid-film (Pinkus and Sternlicht, 1961;
Schlichting, 1960). This is theoretically justified for small values of the Reynolds number (of the
order of 1). On the other hand, the assumption of laminar flow ceases to be valid when there is
transition to either Taylor vortex flow or to turbulent flow which, for fluid-film cylindrical journal
290
bearing, occurs at a Reynolds number of approximately 1000 to 1500. Thus, there is an intermediate
range, say for values of Reynolds number of the order of 100, where added-mass effects may become
noticeable (several times the mass of the journal itself) without affecting the assumption of laminar
flow. The added-mass coefficients represent the mass of the bearing fluid-film (Reinhardt and Lund,
1975) but are significant only in exceptional cases and in most analyses the added-mass of the bearing
film are ignored. The stiffness and damping coefficients can be obtained by a finite difference
solution of the perturbed Reynolds equation (Lund and Thomsen, 1978).
From lubrication theory (ignoring inertia effects) damping coefficients are symmetric but stiffness
coefficients are not. Therefore principal directions do not exist (although this was assumed by Hagg
and Sankey (1956) and Duffin and Johnson (1966-67)), and in the experimental determination of the
coefficients, it is necessary to obtain two independent sets of amplitude-force measurements. Lund
(1987) emphasised the experimental measurement of the bearing coefficients and established more
uniform agreement with analytical calculations by considering the influence of thermal and elastic
deformations and practical problems of manufacturing and operating tolerances of bearing geometry,
clearance and lubricant viscosity. Although the load-displacement characteristics of a journal bearing
is evidently non-linear, the concept of linear dynamic coefficients is still used for modern rotor
dynamic calculations for unbalance response, damped natural frequencies and stability since
experience has demonstrated the usefulness of the coefficients. Clearly the applicability of these
techniques is dependent on small amplitudes of vibration at the journal (relative to the bearing
clearance).
For squeeze-film bearings the governing equation for the fluid-film reaction force has a similar form
to equation (1) with negligible stiffness coefficients and no static load.
x = m xx x + m xy y + c xx x + c xy y
y = m yx x + m yy y + c yx x + c yy y
(2)
Mittwollen et al. (1991) showed theoretically and experimentally that hydrodynamic thrust bearings,
which are often treated as an axial support, might affect the lateral vibration of a rotor-bearing system.
If no axial force is present, then the resulting reaction moments of a thrust bearing can be written in
terms of stiffness and damping coefficients (Jiang and Yu, 1999 and 2000)
x = k x + k x + c x + c x
y = k y + k y + c y + c y
(3)
where kx, etc. represent moment dynamic coefficients of the thrust bearing, and and are angular
displacements (slopes) in x and y directions, respectively.
All of the bearing models discussed thus far are linearised models. A few researchers have considered
non-linear bearing models and these will be described in appropriate places.
291
h3 p h3 p
+
= 6 U ( h) + 2 ( h)
x x y y
t
x
(6.23)
where x locates the distance around the bearing circumference of the point under consideration
(measured from some arbitrary datum), y is the position in the axial direction, is the lubricating
density, is the dynamic viscosity, h is the film clearance, p is the lubricant pressure, U is the
tangential velocity of the journal surface and t is the time. The film clearance is very small compared
with the journal diameter and so the effects of curvature can be ignored for the purposes of evaluation
of the lubricant pressure variation. For the purpose of mathematical modeling, the lubricant film may
Film thickness, h
be considered as through it were unwrapped from around the journal as indicated in Figure 6.8.
R
J
Axial position
Angular position,
2R
Circumferential position, x
h
= 0 ) we get from equation (6.23)
t
292
3 p 3 p
h
h
+ h
= 6 U
x x y y
x
(6.24)
Above equation describes the variation of lubricant pressure in both the axial and circumferential
directions. An approximate solution (closed form) may be obtained by using the short bearing
(Ocvirk, 1952) approximation (where pressure variation in the circumferential direction is assumed to
be negligible compared with that in the axial direction and so
p
is set to zero). The eight linearized
x
stiffness and damping coefficients depend on the steady state operating conditions of the journal, and
in particular upon the angular speed. For the short bearing, the dimensionless bearing stiffness and
damping coefficients, kij = kij cr W , cij = cij cr W , and i, j = x, y, as a function of the steady
eccentricity ratio, , of the bearing are given as (Smith, 1969, Journal Bearings in Turbo Machinery)
k xy =
k yx =
cxx =
c yy =
2 2 2 2 (16 2 ) 4 Q ( )
1
k yy =
2 + ( 32 + 2 ) 2 + 2 (16 2 ) 4 Q ( )
(1 ) {
+ 2 ( 2 8 ) 2 Q ( )
4 2 + ( 32 + 2 ) 2 + 2 (16 2 ) 4 Q ( )
(1 )
2
k xx = 4 2 2 + (16 2 ) 4 Q ( )
2 2 + 2 ( 24 2 ) 2 + 2 4 Q ( )
(6.25)
(1 2 )
with
Q ( ) =
{ (1 ) + 16 }
2
cxy = c yx = 8 2 + 2 ( 2 8 ) 2 Q ( )
12
To determine stiffness and bearing coefficients of a short bearing, the Sommerfeld number
293
S=
DLN r
W
cr
(6.26)
is first determined, where W is the load on the bearing, r is the bearing radius, D is the journal
diameter, L is the length of bearing, is the viscosity of lubricant at operating temperature,
(= 2 n) the angular speed of journal, n is the number of revolutions per seconds and cr is the
radial clearance. The eccentricity ratio of the journal center defined as, = e c r , where e is the
journal eccentricity. We can then determine the eccentricity ratio under steady state operating
conditions by
2
1 2 )
(
L
S =
D 2 (1 2 ) + 16 2
2
(6.27)
For the long bearing approximation (where the pressure variation in the axial direction is assumed
to be negligible compared with that in the circumferential direction and so
p
is set to zero). This
y
approximation also enable closed-form solution of equation (6.26) to be obtained, provided that
appropriate boundary conditions (Figure 6.9) are selected to enable evolution of the constants of
integration. (Cameron, 1981; Pinkus and Sternlickt, 1961 and Rao, 1993).
Lubricant pressure, p
Reynolds condition
dp
= 0 when p =0
d
294
Real bearings are neither infinitely long nor infinitely short. Most bearings have a length to diameter
ratio in the range 0.5 to 1.5, so although the approximate solutions described above may be used in
preliminary design calculations, final design studies will often require a more realistic analysis. For a
more general solution of the Reynolds equation it is necessary to use numerical method. Most
common numerical technique used in hydrodynamic journal bearing analysis is the finite difference
method, because of its relative simplicity and because of the ease with which it can be adopted to suit
most bearing geometries. In the first instance the problem should be reduced to a two dimensioned
one by considering the oil film to be unwrapped from around the journal. The oil film is then divided
into a number of sections of finite size by describing it in terms of mesh nodes, as indicated in Figure
6.10. In between each nodes described, a number of further points are considered, for example those
surrounding the node i, j are shown in Figure 6.11. Since the Reynolds equation describes the
behavior of the lubricant at any location in the fluid film, it can be written for every specific node
contained within the finite-difference mesh.
i, j+1
i, j
Axial
position
i, j+1/2
i-1, j
i+1, j
Further points
considered
i, j-1
Circumferential position
Figure 6.11
pi , j +1 pi , j
i, j +
x 2
1
2
pi , j pi , j 1
i, j
1
2
x 2
+h
1
i+ , j
2
pi +1, j pi , j
y 2
1
i , j
2
pi , j pi 1, j
y 2
h
= 6 U
i, j +
1
2
i, j
1
2
pi , j = k1 + k2 pi +1, j + k3 pi 1, j + k4 pi , j +1 + k5 pi , j 1
where h
i, j +
1
2
(6.28)
is the film thickness between (i, j) and (i, j + 1) nodes and k1, k2 etc. are constants whose
values are known for every nodes. Equation (6.28) may then be written for every node in the mesh,
except for those where the lubricant pressure is already known (for example at the nodes representing
295
the ends of the bearing where the lubricant pressure must be equal to the ambient pressure). These
equations can then be solved simultaneously. Alternatively, an iterative method of solution may be
employed such as successive relaxation where the value of the pressure at each node is determined
successively according to the most up-to-date values of the terms on the right-hand side of the
equation (6.28) (initially, all pressure might be set to zero). This evaluation of pressure is repeated for
all nodes several times over until the change in the value of lubricant pressure at any node is no
greater than a small fraction of 1% of the lubricant pressure. At first consideration it may appear to be
difficult to impose the correct Reynolds conditions, because of the uncertainty of where the lubricant
pressure becomes zero. It appear that only the half-sommerfeld conditions can be imposed by setting
any negative pressure to zero, after solution has been found. However, if a second iteration procedure
or solution of simultaneous equations, is carried out with these pressure set to zero then different
nodes will be found to have sub-zero pressure (negativeve pressure). This process may be repeated
until there is no change in the lubricant film edge position. In the case of iterative procedures the
process can be speeded up by assigning any sub-zero pressure to be zero as and when the arc
evaluated. When this process is completed it is found that, because the Reynolds equation is a
continuous function, the final pressure distribution corresponds to the Reynolds boundary conditions
with the constraint of
p
= 0 at the trailing edge of the lubricant film automatically catered for. Once
the lubricant pressured have been evaluated for a particular journal attitude angle and eccentricity (i.e.
then film thickness will be known), the corresponding forces provided by the lubricant on the journal
may be evaluated by integrating numerically all of the element forces associated with each node. For
example, each node pressure is considered to contribute to force on the journal over an area X &
pi,j
Y
Various
node
positions
Some node
position around
the bearing
circumference
Figure 6.12 A closer view of the grid and node around circumference
296
Thus, if the node is situated at some angle to the horizontal direction as shown in Figure 6.12, then
the lubricant force on the journal in the horizontal direction (from left to right) is
Fh = pij cos ij xy
(6.29)
j =1 i =1
Fv = pij sin ij xy
(6.30)
j =1 i =1
where i is the axial node positions (Total n such nodes) and j is the circumferential node position
(total m such nodes). The above equations for lubricant forces acting on the journal should also
include the tangential friction forces, in most cases however, the friction forces are small when
compared with the pressure forces. For bearing designed to carry vertical loads only (for example
gravity loads) the relationship between eccentricity ratio and journal attitude angle may be
determined by investigating different values of for a given value of until the value of Fn is found
to be zero. This trial and error method enables corresponding value of , and Sommerfeld number
S to be found. The steady-sate journal loci for plain cylindrical bearings and their relationship with
sommerfeld number are shown in Figure 6.13. Where L is the bearing length, D is the bearing
diameter, W is the load on the baring, R the bearing radius, C is the radial clearance and N the shaft
rotational speed in rev./sec.
Eccentricity ratio
900
L/D=0.5
Eccentricity ratio
=e/c
0.5
L/D=1.0
600
L/D=1.0
0.25
10-1
Sommerfeld number S =
10
2
DLN R
W
cr
297
Figure 6.13 Typical variation of (a) eccentricity ratio with Sommerfeld number and (b) eccentricity
ratio with journal altitude angle for plain cylindrical bearings for L/D = 0.5 and 1.0
6.2.3 Friction
Friction forces in fluid film bearings arise out of the energy dissipation associated with the shearing
of the fluid film. From the elementary fluid mechanics the shear stress at some point in a fluid layer is
given by
du
dz
(6.31)
du
is velocity gradient
dz
across the fluid film. In a hydrodynamic the velocity gradient term arises partly from the variation in
pressure around the journal causing lubricant flow and partly as result of the journal rotation
dragging lubricant around the clearance space. Thus, for a point at the journal surface equation
(6.31) becomes
1 dp h U
+
dx 2 h
(6.32)
The first term in the bracket is the pressure-induced term and the second term is the velocity-induced
term. The total drag force on the journal is then given by ( x = Rd )
Fd =
2
+ d dp h
U
Rd
+
L 0 Rd 2
0 h Rd dy
2
2
(6.33)
where pressure-induced component is assumed to be present only when the lubricant pressure is
positive i.e. over the region 0 < < + d. The velocity-induced component is present wherever there
is lubricant, for example all around the clearance space (with reasonable approximation). It is possible
to develop closed-form expressions for the integral in equation (6.33) in the case of the infinitely long
and infinitely short bearings (Pinkus and Sternlicht, 1961). In the case of finite bearings the
integration may be carried out numerically and expressed as
298
n l
n m
hij dp
KRU
RU
xy +
xy
xy +
j =1 2 dx ij
i =1 j =1 hij
i =1 j =l hij
l
Fd =
i =1
(6.34)
where l is the node corresponding to the edge of the pressurized section of the lubricant film
(circumferential node position). Equation (6.34) is a slight variation of equation (6.33) in that the last
term of (6.34) now allows for the break up of the lubricant film, where K is a constant whose value is
less than unity. The magnitude of K may be determined empirically, and is the concern of current
research. This feature of real lubricant film is not allowed for in equation (6.33).
Q=2
h 3 dp
Rd
12 dy
(6.35)
For the finite bearing, using a trapezium rule for integration for the finite difference mesh shown in
Figure 6.10, we have
h 3 ij dp
x
j =1 12 dy ij
m
Q = 2
(6.36)
299
L
= 0.25
D
L
= 0.5
D
L
= 1.0
D
Eccentricity Ratio, = e / c
Figure 6.14 Variation of flow rate with eccentricity ratio
Fx = f1 ( x, y, x , y )
and
Fy = f 2 ( x, y, x, y )
(6.37)
where x and y are both zero. When there are changes in these displacements and velocities as
described above, the new values of lubricant film forces will be Fx + dFx and Fy + dFy respectively.
300
These forces may also be expressed using equation (6.37) as a four-variable Taylor series, neglecting
small terms, as
Fx + dFx = f1 ( x, y, x , y ) + dx
Fx
F
F
F
+ dy x + dx x + dy x
x
y
x
y
(6.38)
and
Fy + dF y = f 2 ( x, y, x, y ) + dx
Fy
x
+ dy
Fy
y
+ dx
F y
x
+ dy
Fy
y
(6.39)
Fx
F
F
F
+ dy x + dx x + dy x
x
y
x
y
= k xx dx + k xy dy + cxx dx + cxy dy
dFx = ( Fx + dFx ) Fx = dx
(6.40)
and
Fy
F
+ dy y
x
y
x
y
= k yx dx + k yy dy + c yx dx + c yy dy
dFy = ( Fy + dFy ) Fy = dx
Fy
+ dy
Fy
+ dx
(6.41)
where kij and cij are known as the oil-film stiffness and damping coefficients respectively, i and j can
take values of x and y. It is these quantities, which must be known in order to calculate the overall
system response to imbalance and stability. Values of kxx and kyx may be found by calculating the
lubricant film forces Fx1 and F y1 when a small displacement dx is imposed on the journal (from its
static equation position) with dy , dx , and dy all set to zero. The process is then repeated with a
displacement of dx to calculate Fx2 and Fy 2 . The stiffness coefficients kxx and kyx are then given by
k xx =
Fx1 Fx2
2dx
and k yx =
Fy1 Fy2
2dx
(6.42)
Similarly, values of the other stiffness and damping coefficients may be evaluated by imposing small
variations in dy , dx , and dy on the journal, and investigating their effect on the lubricant film
forces, as
301
k xy =
Fx3 Fx4
2dy
k yy =
Fy3 Fy4
2dy
cxx =
Fx5 Fx6
2dx
;
(6.43)
c yx =
Fy5 Fy6
2dx
cxy =
Fx7 Fx8
2dy
c yy =
Fy7 Fy8
2dy
It should be noted that when proceeding in the manner described above care should be taken when
formulating Reynolds equation (6.24) in finite difference form, as the quantity
h
is no longer zero
t
when dx or dy are non-zero, and appropriate value of U is also dependent on dx and dy . Solutions
for several types of bearing have been carried out (Lund et al., 1965; Homer, 1960, Someya, 1988). In
addition to the spring and damping coefficients introduced above, some authors have suggested the
use of inertia coefficient and moment coefficients. Although such coefficients usually have only a
small effect on bearing dynamics. Inertia coefficients can be defined as
and
F y ( x, y, x , y , x, y) = 0
(6.44)
Hence
dFx = +
Fx
F
dx + x dy = + mxx dx + myx dy
x
y
(6.45)
Fx ( x, y, , , x, y ,, ) = 0
(6.46)
Hence
dFx = +
Fx
F
.d + x .d + = + kx d + kx d +
(6.47)
dM x = +
M x
M x
.d +
.d + = + k d + k d +
(6.48)
and
302