A S L E Transactions

ISSN: 0569-8197 (Print) (Online) Journal homepage: http://www.tandfonline.com/loi/utrb19

Performance of Tapered Land Thrust Bearings for

Large Steam Turbines

A. Missana , E. R. Booser & F. D. Ryan

To cite this article: A. Missana , E. R. Booser & F. D. Ryan (1971) Performance of Tapered
Land Thrust Bearings for Large Steam Turbines, A S L E Transactions, 14:4, 301-306, DOI:
10.1080/05698197108983255

To link to this article: http://dx.doi.org/10.1080/05698197108983255

Published online: 25 Mar 2008.

Submit your article to this journal

Article views: 22

View related articles

Citing articles: 3 View citing articles

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=utrb19

Download by: [University of Arizona] Date: 13 June 2016, At: 11:00

 Performance of
Tapered Land Thrust Bearings
for Large Steam Turbines
A. MISSANA, E. R. BOOSER, ASLE, and F. D. RYAN
General Electric Company, Schenectady, New York

surface velocities are similar in nature to those recently

Downloaded by [University of Arizona] at 11:00 13 June 2016

Bearings of 350-, 363-, and 450-square-inch area and ranging

up to 31.5-inch outside diameter were tested at speeds up to 5000 reported for steam turbine journal bearings ( 6 ) .
rpm and at loadc up to 1000 psi. Laminar-turbulent transition This paper reviews performance characteristics of ta-
conditions are analyted as well as performance relationsfor both pered land thrust bearings under test conditions simulat-
regimes. ing operation in steam turbines. Included are observa-
tions on power loss, oil flow, film thickness and
temperature rise. Since many of the tests involve opera-
INTRODUCTION tion within the turbulent regime, discussion is included
on the relation of the laminar-turbulent transition to
The recent rapid growth in size of steam turbine- geometric and operating factors. Also covered are obser-
generators for electric power production has been accom- vations on the relation of temperature rise to oil feed
panied by a corresponding growth in bearing size. The rates. Influence of taper design variables are considered
tapered land thrust bearing plates commonly used in along with observations on a marked tolerance for taper
these units now range in outside diameter u p to 31.5 wear.
inches a t 3600 rpm with losses u p to 1000 horsepower
per bearing.
This trend to larger thrust bearings has introduced a TEST APPARATUS
variety of new design factors. T h e bearing oil film has The thrust bearing test machine utilized for the studies
shifted to turbulent operation with a need to modify the reported in this paper is shown schematically in Fig. 1.
laminar analyses previously employed (1, 2). Much T h e test bearing assembly is fixed rigidly to the casing
greater power loss has correspondingly increased the and loading is achieved through an opposed, axially
quantity of oil required for cooling. As the limiting moveable, thrust bearing. A maximum load of 450,000
babbitt metal temperature is more closely approached pounds is applied by a set of hydraulic cylinders con-
under high transient loadings ( 3 ) ,closer attention is also nected in parallel. A variable speed steam turbine drive
required to land geometry and to use of more thermally is capable of 1500 HP output and speeds u p to 5000
conductive backing materials (4, 5). Many of these char- rpm. A separate lube oil feed and drain system for the
acteristics involved in operation of thrust bearings at high test bearing assembly allows determination of bearing
losses by the energy balance method. A second method
Presented at the 26th ASLE Annual Meeting in Boston, of determining test bearing losses uses a torque meter
Massachusetts, M a y 3-6, 1971 in the steam turbine drive to measure total power loss

NOMENCLATURE K = turbulent fluid friction coefficient [See Eq. 11

f = turbulent fluid friction fractor, & N = rotational speed, rev/sec
pUL/2g r, = inside radius of lands, in.
g = gravitational acceleration, 386 in./sec2 r, = outside radius of lands, in.
h , = maximum oil film thickness at entrance to taper, in. re = effective radius [See Eq. 41
- = minimum oil film thickness at trailing edge of taper, in.
h2 Re = Reynolds Number, 7 i ( 2 ~ r ~ N ) p / ~ g
h = area average oil film thickness, in. U = runner surface velocity, in./sec
H = power loss, in. lb/sec p = oil density, lb/in3
J = laminar power loss coefficient [See Eq. 51 p = oil viscosity, Ib sec/in2
k = fraction of total area between rl and r, included in bearing 7 = friction stress on surface, Ib/in2
lands
30 1
 TAPERED-LAND BALL-SEAT HYDRAULIC S L I D I N G MOUNT PHANTOM PORTION
TEST B E A R I N G MOUNTING CYLINDERS FOR LOAD BEARING TO BE REMOVED
\ 1 / / /SLIDING SUPPORT PADS

0.009 IN.
Fig. 1 -Thrust bearing tester.
Fig. 2-Schematic diagram of typical tapered land thrust bearing.

in the test machine. From this total is then subtracted

the windage, seal, journal bearing and load thrust bear- 10 -
ing losses, to yield the test bearing power loss. - BEARING
Thrust bearing babbitt metal temperatures were -- A
B x
O

measured with thermocouples embedded in the babbitt -

Downloaded by [University of Arizona] at 11:00 13 June 2016

with the sensing tip positioned within x2

inch from the
i
-
0 -
C a
D 0

-I'
V)
active surface. Sufficient thermocouples were used to 2 TURBULENT
x
allow determination of the temperature field on a typical
land. Oil film pressures were measured by means of taps
5 I:
d"
XI
drilled to the active surface and connected to Bourdon
tube type gages. Capacitance film thickness gages were
employed.
T h e bearing sizes tested are given in Table 1. Figure
2 is a schematic of the tapered land type thrust bearing.
A petroleum turbine oil of 150 SUS viscosity a t 100 F SURFACE VELOCITY, FT/MIN.
was en~ployedthroughout the test series. T h e oil feed
temperature was held a t 115 F for the data reported here. Fig. 3-Power loss vs. surface speed at 200 psi load.

POWER LOSS
2.1 hp/sq in. for a total friction loss of about 950 hp.
T h e rapid rise in power loss encountered with higher While power loss rises rapidly with increasing diameter,
surface speeds is depicted in Fig. 3. For laminar opera- the larger bearing area involved does lower the loading
tion with velocities u p to about 13,000 ft/min a t the pressure. It is readily concluded that, for any specific
effective radius, power loss per unit area rises as the 1.5 thrust area required by turbine design, a minimum effec-
power of surface velocity. For turbulent operation a t tive bearing diameter is desirable to limit friction loss.
higher speeds, increase in friction is more pronounced Power loss in the turbulent regime can conveniently
and rises approximately as the cube of velocity. Increas- be described by a dimensionless friction factor f as shown
ing bearing diameter has an even more pronounced in Fig. 4. This fluid friction factor represents the ratio
inlluence. In addition to bringing increased surface veloc- of shear stress a t the bearing oil film surface to the fluid
ity for a given rotational speed, the bearing surface area kinetic energy term U2p/2ga n d is related as follows to
also increases as the square of the diameter. With the the Reynolds Number for the bearings in this test series:
450 square inch test bearing a t 3600 rpm giving a surface
velocity 24,000 fpm, the power loss from Fig. 3 becomes

TABLE1-TAPERED LAND TEST BEARINGS

TAPER
TEST AREA T1 No. AT rl FRACTION
BEARING SQUARE
IN. IN. LANDS IN. FLAT
 Performance of Tapered Land Thrust Bearings for Large Steam Turbines 303
-

BEARING I TABLE
2-TURBULENT LOSS COEFFICIENTS
K FOR TURBULENT
AREA FRICTIONFACTOR,
BEARING SQUARE
INCH f =K/RCO.~~
A 350 0.027
B 350 0.027
C 363 0.025
D 450 0.024
-
Avg. 0.026

0.0011 , I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
100 1000 10.000
REYNOLDS NUMBER. Re typical of turbulence with power loss occurring primarily
in two surface boundary layers whose properties depend
Fig. 4-Turbulent friction factor vs. Reynolds Number.
on Reynolds Number. Since the fluid friction factor
varies inversely only as about the 0.25 power of Re,
viscosity would also have only a minor influence on
power loss; and this was the case for varying either oil
Downloaded by [University of Arizona] at 11:00 13 June 2016

The characteristic film thickness used in this Reynolds feed rate or oil feed temperature in test. Table 2 gives
Number was the average for a land. Viscosity was taken the turbulent friction factor-Reynolds Number relation
at the average temperature for oil discharging from the for the individual test bearings. The average value of
land (drain temperature when the oil feed rate matched K of 0.026 compares with the 0.016 typical for journal
hydrodynamic pumping capacity). This combination of
bearings (6).
viscosity and film thickness is believed to describe ade- In laminar operation at lower Reynolds Numbers, the
quately the average degree of turbulence in the bearing
friction factor can be expressed as
oil film and gave a satisfactory correlation of the test
data.
The characteristic bearing radius to be combined with
rotational speed for the velocity term in Reynolds Num-
Laminar coefficient J is unity for a narrow flat land and
ber follows from the defining equation for friction fac-
increases with increasing taper. J also increases with
tor$ smaller minimum film thickness (increasing load) as in-
dicated in Fig. 5. Data from the present test series with
nearly square lands (radial breadth = circumferential
length) commonly range about 20 percent above analyti-
cal values given by the solid line. The analysis assumes
that the center-line pressure distribution in the direction
of motion is given by the infinite width solution of
Reynolds Equation and that the radial pressure distribu-
An effective radius re is then selected which gives the same tion drops sinusoidally to zero at the land radial bound-
power loss when combined with the bearing area in the aries (I). A similar result is obtained by a two-
relation dimensional finite-difference numerical solution (2). The
higher loss in test than by laminar analysis may reflect
additional pumping and churning effects, or simply some
deficiency in the analysis itself. In general,J ranges from
about 1.5 with large minimum film thickness to 2.5 with
small film thickness for typical tapered land designs.
For equality of power loss in Eq. [2] and [3]

For a design in which rl = 0.5r2, effective radius re is

0.80r2ratio.
r2/rl and this value varies very slowly with changing

The turbulent friction factors found in test are plotted

in Fig. 4 for loads ranging from 200 to 800 psi, for speeds
from the lower end of the turbulent range up to 26,000
3
c
w

E
a
u
z
1 2

I
-4
2
0

4
xQ)

6 8 10 20 40
F I L M THICKNESS RATIO. H,/H2
60 100
ft/min, and for various oil feed rates. The minor influence
of load and other factors influencing film geometry is Fig. 5-Laminar loss coefncient vs. film thickness ratio.
 LAMINAR-TURBULENT T R A N S I T I O N 81 I
I
T h e power loss in the vicinity of the laminar-turbulent TURBULENT
transition region always appeared to be that of the regime 0 6-
CALCULATION

dictating the higher friction. With increasing speed, for 8

instance, the power loss curves indicate that turbulent 5
Y
operation is initiated as soon as more energy is consumed -I 4 -
-
/x
x--'T-----
0
in turbulence than would have been the case with lami-
LAMINAR
nar flow.
CALCULATION
With this relatively simple definition, the influence of
various factors on the laminar-turbulent transition point
can be judged. Higher thrust load, for instance, will I I
increase laminar power loss but will have little influence 0 1000 2000 3000 4000
SPEED. RPM
on turbulent power loss; hence, higher thrust load will
delay the onset of turbulence when going to higher
Fig. 6.-Increasing minimum film thickness with rising speed. Beoring
speeds. The same is true of higher oil viscosity. 8, 200 psi load.
T h e transition speeds obtained from trend lines in test
plotted as in Fig. 4 and 5 are given in Table 3 for the
test bearings operating a t various loads with the light
petroleum turbine oil. T h e results show no typical
Downloaded by [University of Arizona] at 11:00 13 June 2016

ating temperature multiplied by the ratio of the turbulent

lieynolds Number for the transition. No single value may power loss to the loss which would be expected for larni-
be adequate to cover this transition for fluid conditions nar conditions. For a given percentage increase in power
in an oil film encountering wide range of film thicknesses, loss resulting from turbulence, load support then increases
surface velocities and temperatures. Turbulence in a just as much as with the same percentage increase in oil
given bearing, for instance, is likely first encountered a t viscosity in laminar conditions. Although finite difference
the outside radius along the leading edge of a land. T h e numerical computer solutions of Reynolds and energy
turbulent zone then progresses both radially inward and relations give more direct results for tapered land bear-
circumferentially along the taper over a range of increas- ings, this use of a n effective viscosity appears useful in
ing shaft speeds. extending laminar theory to turbulence for both journal
A unique rise in film thickness is encountered on (6, 7) and various thrust bearings. This in turn allows
reaching turbulence with rising speed in Fig. 6. T h e ready inspection of the influence of various factors on
measured values of film thickness are seen to be con- performance, a procedure that often supplements nicely
siderably greater in turbulence than would be expected the finite difference numerical solutions.
from laminar calculation ( I ) . For the case shown, exten-
sion of laminar theory to higher speeds even gives a dip
O I L F L O W A N D TEMPERATURE RISE
in calculated film thickness because of decreasing viscos-
ity with high power loss. As an approximation for ex- Oil commonly feeds a tapered land thrust bearing a t
tending laminar film thickness analyses into the turbulent its inside diameter, flows radially outward in distributing
regime, an "efTective" viscosity can be introduced which grooves, and then is pumped circumferentially onto the
is simply the actual oil viscosity a t the bearing land oper- tapered lands by the rotating surface of the thrust runner.

TABLE
3-LAMINAR-TURBULENT TRANSITION CONDITIONS
TRANSITION
AREA LOAD
BEARING SQUARE
INCH PSI RPM RE
A 350 100 1290 580
200 1660 660
600 2570 840
U 350 100 1350 720
200 1788 850
600 2878 1180
 Performance of Tapered Land Thrust Bearings for Large Steam Turbines

Hydrodynamic action developed over a land finally 1.4

causes oil to exit at both the inside and outside diameters,

and a t the trailing edge. Total oil flow depends both on
the hydrodynamic characteristics of the lands and on the
flow resistance of the grooves and any groove orifices.
Oil flowing through the bearing experiences two tem-
perature rises. The first involves mixing in the oil dis-
tributing grooves of the cool feed oil with hot oil recircu-
lating from the lands. This groove temperature, which
experimentally reflects a simple heat balance for oil flows
into and out of a groove, then becomes the oil feed 0.8 -
temperature for the bearing lands. The second tempera-
ture rise in the bearing is normally the major one and
reflects hydrodynamic heating over the bearing lands 0.6 I I
0 20 40 60
themselves. An analytical expression has been previously PER CENT FLAT
developed for the hot spot temperature for a tapered land
with either laminar or turbulent operation ( 3 ) . Fig. 8-Variation in load capacity with per cent flat. Bearing B, 3600
Although oil feed in large bearings can be varied over rpm, 200 F maximum babbitt temperature.

a considerable range, the optimum rate found in test was

Downloaded by [University of Arizona] at 11:00 13 June 2016

approximately that which matched the hydrodynamic

- -

pumping characteristic of the tapered lands. With lower

net oil flow out from the lands, while a larger circum-
flow, a higher maximum temperature is encountered as
shown in Fig. 7. With higher feed rates, a drop in oil ferential taper at the inside diameter improves the hy-
groove temperature results along with a slightly lower drodynamic load capacity. For high surface speeds in the
overall bearing temperature; friction loss and oil flow on largest sizes where turbulence generally provides ade-
individual lands are relatively unaffected. quate oil film thickness, a relatively large one-
dimensional taper is desirable for inducing a large flow
of cooling oil to minimize the bearing metal temperature.
INFLUENCE OF TAPER DESIGN VARIATIONS Influence of the proportion of flat area is of special
A common design involves a taper of approximately importance. It determines not only the characteristics
0.001-0.002 in./inch at the radial center line for 90 under initial operating conditions when a bearing is new,
but also governs tolerance to wear. Figure 8 shows the
percent of the land length. The final 10 percent at the
test results for the 350-square-inch bearing B of Table 1.
trailing edge of the land is then flat. This flat not only
Load capacity is expressed as the ratio of the load that
provides location and load capacity during starting, but
could be carried with the indicated percent flat as com-
actually provides an increase in load capacity over that
pared to the loading possible with no flat, in all cases
possible with a full-length taper.
using 200 F maximum babbitt temperature for defining
Within this general practice, a number of variations
the load limit. Load capacity in test is seen to rise as the
can be made to match specific operating and size re-
land area was machined flat up to about 20 percent. Only
quirements. For smaller bearings with low power loss and
after the lands were machined flat to about 30 percent
little temperature rise, for instance, a double taper is often
did the load capacity again drop to the level found for
used. The smaller taper at the outside diameter minimizes
the normal initial flat area of 10 percent. These results
indicate that tapered land bearings are relatively insensi-
tive to wear, and that a small amount of wear will increase
their load capacity.

4 CONCLUSIONS
1) Turbulence can usually be expected in operation of
tapered land thrust bearings when surface velocity
m
exceeds about 12,000 ft/min at the effective radius.
2) Power loss per unit area of bearing surface increases
as approximately the 1.5 power of surface velocity in
laminar operation, as the cube in turbulence.
3) Tapers should be increased in turbulence to induce
a greater flow of cooling oil across each land.
I I I I 4) Load capacity of tapered land bearings is only mildly
I4O0 40 80 120 160 200
OIL FEED. GPM influenced by the percent of flat area provided; a
small amount of wear during normal operation will
Fig. 7-Maximum temperature vs. oil feed rate. Bearing B, 1800 rpm. actually give some increase in load capacity.
 REFERENCES pered-Land and Pivoted-Shoe Thrust Bearings for Large Steam
Turbine Application," Trans. ASME, A, 81, 208-214, (1959).
(I) Wilcock, D. F., and Booser, E. R., "Bearing Design and Appli- (5) Bahr, H. C., "Recent Improvements in Load Capacity of Large-
cation," McGraw-Hill Book Co., New York, 1957, 321-329. Steam Thrust Bearings," Trans. ASME, A, 83, 130-134, (1961).
( 2 ) Pinkus, O., "Solution of the Tapered-Land Sector Thrust Bearing," (6) Booser, E. R., Missana, A,, and Ryan, F. D., "Performance of Large
Ili.nt~s.ASME, 80, 1510-1516, (1958). Steam Turbine Journal Bearings," ASLE Trans., 13, 262-268,
(3) Booser, E. R., Linkinhoker, C. L., and Ryan, F. D., "Maximum (1970).
l'emperature in Hydrodynamic Bearings under Steady Load," Lub. (7) Duffin, S. and Johnson, B. T., "Some Experimental and Theoret-
Eng., 26, 226-235, (1970). ical Studies of Journal Bearings for Large Turbine-Generator Sets,"
(4) Brandon, R. E., and Bahr, H. C., "Load Capacity Tests on Ta- f i t . IME, 181, 89-97, (1966).
Downloaded by [University of Arizona] at 11:00 13 June 2016