Académique Documents
Professionnel Documents
Culture Documents
This article presents a report on an investigation into the per be the case in an upgrade or when different applications (e.g., a
formance characteristics of a steadily loaded pivoted shoe jour gearbox and compressor) share a common lube system. In such sit
nal (PSI) bearing that is lubricated with ISO VG 32 and VG uations, the changes in bearing performance need to be assessed.
68 oils. The article describes a testing machine on which the ex A number of authors have presented steady-state results from
experimental work on PSJ bearings. This includes studies of bear
perimental investigation was performed. Measurements of shaft
ing and bulk oil temperatures, over a wide range of bearing
torque, pad temperature distributions, oil inlet and outlet tem
sizes, speeds, and unit loads (Brockwell, et a1. (1), (2); Simmons
peratures, oil flow rate, and eccentricity have all been recorded
and Dixon (3); Simmons and Lawrence (4); DeChoudhury and
as functions of load and speed. The experimenteil results from
Masters (5); DeChoudhury and Barth (6). Others have mea
both test oils are presented in graphical form and are compared sured shaft-to-bearing displacements, also over a wide range of
with theoretical predictions obtained from the author's com operating conditions (Brockwell and Kleinbub (7); Pettinato and
puter model of the PSI bearing. These results showed that the DeChoudhury (8); Tripp and Murphy (9). All of the studies men
thicker ISO VG 68 oil provided thicker oil films. However, it tioned are based on tests using light turbine oils.
also had higher bearing temperatures and power losses. A good The purpose of this article is to fill a gap in the literature by pro
correlation between the theoretical and experimental results has viding information from PSJ bearing tests comparing bearing per
been found. Theoretical analysis of the bearing friction losses formance using ISO VG 32 and ISO VG 68 oils. Shaft torque, bear
indicate that shear losses predominate and churning losses ac ing and bulk oil temperatures, and shaft displacements have been
recorded for various loads and speeds. Comparisons are made be
count for approximately 20% of the total losses.
tween the experimental results and theoretical predictions.
KEYWORDS
480
An Investigation of the Steady-State Performance of a Pivoted Shoe Journal Bearing with ISO VG 32 and VG 68 Oils 481
Test
--..
Housing
& Bearing
Scale _____ I
o I tr
the bearing housing to be used as a low-friction torque-sensing Proximity probes arranged in two mutually perpendicular axes
device. This provides a more precise determination of bearing at either end of the bearing housing measure shaft eccentricity.
power loss than the thermal balance method. A balance arm, in These probes measure the horizontal and vertical displacements
conjunction with a load cell, is used to record the friction torque of the test bearing with respect to the shaft's location. They also
directly. provide a check of the alignment between the bearing and shaft.
Signals from probes mounted directly opposite each other were
combined to give the true relative position because this setup can
celed out any shaft bending effects. Measurements from both ends
test bearing housing of the bearing housing were then averaged to determine the dis
placements (in both orthogonal planes) at the midplane of the
bearing. Steady-state eccentricities were measured with respect
to an equilibrium position, which was determined at 72 mis with
zero load on the bearing after thermal equilibrium had been es
upper spherical
tablished in the bearing.
hydrostatic bearing
A pump with a capacity of 1.26 x 10-3 m3/s and a maxi
mum supply pressure of 2.1 MPa delivers oil to the test bear
lower hydrostatic ing from a 0.6 m3 capacity tank. The flow rate is measured by a
bearing turbine-type flow meter with a linear flow range of 0.16 x 10-3
to 1.83 X 10-3 m3/s. Feed oil temperature is measured as it
enters the test bearing and is controlled by a water-oil heat ex
changer to within 1C. An industrial-type pressure transducer,
with a 0 to 0.41 MPa operating range, measures the oil supply pres
sure to the test bearing. Thermocouples monitor the oil inlet tem
perature and the oil outlet temperature from each side of the test
bearing.
Measurement uncertainties are listed in Table 1.
air bag
TEST BEARINGS
TEST CONDITIONS ature is constant through the thickness of the oil film. The third
boundary condition is calculated on the basis that hot oil from the
Tests were performed over a range of speeds and loads.
preceding pad affects the oil temperature at the leading edge of
Applied bearing loads were 3.51, 14, and 22.24 kN (0.34, 1.38,
the next pad.
and 2.19 MPa unit loads).
Oil flow rates used in the tests are listed in Table 2.
Tests were performed with the "load between pads."
TABLE 2-ThST OIL FLOW RATES
Table 3 lists the oil properties of the ISO VG 32 and VG 68
turbine oils. Load, kN (lb) Speed, rpm Flow rate, m3/s (US gpm)
Oil supply temperature was 49C.
3.51 (788) 1,800 5.7 x 10-5(0.9)
3,600 1.4 x 10-4 (2.2)
5,810 2.9 x 10-4 (4.6)
Oil inlet 7,750 4.8 x 10-4 (7.5)
9,040 6.5 x 10-4 (10.2)
10,650 9.1 x 10-4 (14.4)
14 (3150) 1,800
3,600
7.3 x 10-5 (1.15)
1.7 x 10-4 (2.7)
5,810 3.4 x 10-4 (5.4)
i
----- 1 ----- 7,750 4.9 x 10-4 (8.5)
Seal Seal 9,040 7.2 x 10-4 (11.3)
1 !
10,650 9.9 x 10-4 (15.6)
011
outlet
1.\ 5,810
7,750
9,040
3.7 x 10-4 (5.8)
5.7 x 10-4 (9)
7.5 x 10-4 (11.8)
10,650 1 x 10-3 (16.2)
Fig. 3-Flood-lubricated PSJ bearing.
An Investigation of the Steady-State Performance of a Pivoted Shoe Journal Bearing with ISO VG 32 and VG 68 Oils 483
Temperature distribution at the pad surface is calculated from
'Ii
VI
.2 30 I v
j
the Laplace heat transfer equation, assuming a constant temper
ature in the axial direction. The boundary conditions for the heat 0..
transfer equation correspond to an equality of heat fluxes at the 20 v
/:
boundary between the oil film and the pad surface, and heat con
vection on all remaining surfaces.
Approximate pad deflection is calculated from the one
10
/
dimensional equation for a beam. Shear forces, bending moments,
and local differences in temperature across the thickness of the
pad are taken into consideration. In the axial direction, the load
distribution is assumed constant and is calculated by averaging o
35
In this study, measurements of shaft torque, pad temperature ....... Kinetic losses
distributions, oil inlet and outlet temperatures, oil flow rate, and ....... Chumlng losses 1-----fL--I
___ Shear losses
30
eccentricity have all been recorded as functions of load and speed.
Performance comparisons for the two oil grades are based on iden +----------;
25
tical operating conditions of load, speed, oil flow, and oil inlet
temperature. In addition, the experimental results from both test .J----+--f
oils are compared with theoretical predictions obtained from the
author's computer model of the PSJ bearing. f20
...
+--------- ,
15
Q)
40 Tr=========----' 150
"'-2.19 MPa, 68 oil (calculated)
o 10. 5% temp., 0.34 MPa
t:. 2.19 MPa, 68 oil 140 o 95% temp., 0.34 MPa
o 0.34 MPa, 68 oil
5% temp., 2.19 MPa
130
95% temp., 2.19 MPa
30 +------1---1
-+-95% temp., 2.19 MPa (calculated)
120
o
(.)
o
o.
e
110
Ii ::::I
III
100
/
20+_-------- ell
.. a.
E 90
Q.
80 I
/
u
0
10+------7=---r 0
70
I' ()
0
0;
60 ..
t:. 10.
o 10. 10.
10.
50
0+----.,---,---....----.,,.--1 o 20 40 60 80 100
o 20 40 60 80 100 Shaft speed, m/s
Shaft speed, m/s
ing edge for both oils and for all operating conditions tested.
speeds and between 80 and 90% at higher speeds. Churning losses
Figures 8 and 9 show the effect of shaft speed on the sec
make up most of the balance, with the kinetic losses accounting
. ond loaded pad leading- (5% location) and trailing- (95% lo
for only 1-2% of the total losses.
cation) edge temperatures for the ISO VG 32 and 68 oils,
Figure 7 shows the relationship between bearing power loss
and shaft speed for the ISO VG 68 oil. With this thicker oil, losses 150
10. 5% temp, 0.34 MPa.
are larger than those recorded with the ISO VG 32 oil, but not
o 95% temp, 0.34 MPa
significantly so. For example, with a shaft speed of 7 2 mls and 140 5% temp, 2.19 MPa
/V/
a unit load of 2.19 MPa, the measured losses with the ISO VG o 95% temp, 2.19 MPa
-+-95 % temp, 2.19 MPa (calculated)
32 and 68 oils were 23 and 27 kW, respectively. This smaller-than
130
expected difference is attributed to much higher pad temperatures
encountered using ISO VG 68 oil (compared in the next section).
The higher temperatures significantly lower the oil film viscosity
()
120
o
/0
and, therefore, the shear losses within the bearing, which account
0_ 110
for 80-90% of the total losses. I!!
::::I
I
Q) 0
a.
Both the loaded and unloaded pads were heavily instrumented
90
/
with thermocouples, as shown in Fig. 4. Particular attention was 0
I-
paid to the trailing half of the two loaded pads, since it is within 0
these regions of the bearing that the maximum operating tem
80
perature occurs. Also, on both loaded pads, thermocouples were
i 0
0
130 -
Leading edge (32 011) o 2.19 MPa, 68 oil
Trailing edge (32 oil)
/-,
140
..... 2.19 MPa, 68 oil (calculated)
120
... Leading edge (68 all)
_ Trailing edge (68 Oil / A 0.34 MPa, 68 oil
130
<>
110
2.19 MPa, 32 oil
i 90 +----f'/L-7''-/ -----
_ 110
CIl
...
80 +--+-1-+-/
-----1
100
CIl
Q.
'L
S 90
70
CIl
....
/ .
+---1-----------1
60+-_T===L--
80
50+-------,-----
70 o
o
o 20 40 60 80 100 60
o
Shaft speed, m/s
50
0 20 40 60 80 100
Fig. 10-Pad circumferential temperature rise as a function of shaft
speed-2.19 MPa unit load. Shaft speed, m/s
130 ------
for ISO VG 68 was of the order of 1 30C, compared to only
1 06C for ISO VG 32.
Both speed and load have a considerable effect on the increase -+-0.34 MPa
in temperature from the leading to trailing edge. For example, 120 -1---------
- 1------
...... 2.19 MPa
---
with the ISO VG 68 oil and a unit load of 2.19 MPa, the increase
110+-------
at 7 2 mls is of the order of 7 0C (Fig. 10).
Speed, load, and even oil viscosity have only a small effect on
second loaded pad leading-edge temperature. For all of the
test cases considered in this study, recorded leading-edge pad 100 -j-----f-----f
temperatures were of the order of 60 5C. e
::;,
E(1) 90 +--------7---
Figure11 shows the effect of shaft speed, load, and oil viscosity 0-
S
on the second loaded pad 75% location temperature. Generally, all
80
of these independent variables have a large influence on tempera
ture at this location in much the same way as they do on maximum
pad temperatures. It is noted that the 75% location temperatures
are somewhat lower than the maximum pad temperatures with
differences of as much as 20C depending on load, speed, and oil
type. Even so, the 75% location is an important indicator because
babbitt integrity is also dependent on pressure, and the 75% lo
130 40
() 100
0
i
;:,
-
e
Q)
90
/
Q.
E 6.
Q)
... 80
70 10 oj
60 5 /
o
o 45 90 135 180 225 270 315 360 o 20 40 60 80 100
Circumferentia\\ocation,o Shaft speed, m/s
FIg. 13-Pad temperature profiles, 72 mIs-ISO VG 68 011. Fig. 14-Bulk oil temperature rise as a function of shaft speed.
In the example shown in Fig. 13 at 2.19 MPa, the increase mea where P is unit load, Cp is pad clearance, R is shaft radius, Lis
' bearing width, U is shaft surface speed, and /1- is oil viscosity.
sured about 20C on the unloaded pads, 45C on the first loaded
pad, and almost 7 0C on the second loaded pad. The range of Sommerfeld numbers covered in Fig. 15 is based
Unlike the loaded pad temperatures, which rise with increas on a pad clearance-to-shaft ratio of 0.002, shaft speeds that ranged
"-
Bulk Oil Temperature Rise 0
80
4.
Figure 14 shows the effect of load, speed, and oil viscosity on .-
E
bulk oil temperature rise. Oil outlet temperatures are significantly ::L . '"
c '"
agreement. Q)
U
U . . '"
w -. 0
Bearing Displacements
40
Measurements showing eccentricities in PSJ bearings are rare. 'I '"
Tripp and Murphy (9) presented steady-state eccentricities for two l 0
.ISOVG 6a all
0 ...ISOVG 32 all
five-pad PSJ bearings with 0.0 and 0.2 preloads.
20
Figure 15 shows shaft eccentricity as a function of Sommer 0 o ISOVG 68 all {calculated}
feld number for the two different oils tested in this study. The '" ISO VG 32 oil (calculated)
Sommerfeld numbers corresponding to the measured and theo
retical data are based on the measured and calculated viscosity of o
the oil at the bearing outlet, respectively, and on the pad radial o 2 4 6 8
clearance: Sommerfeld Number
120 With the thicker ISO VG 68 oil, losses are larger than those
recorded with the ISO VG 32 oil, but not significantly so. For
100 . example, at 72 mls shaft speed, the difference in power loss was
only of the order of 4 kW.
...
The small difference in power loss is attributed to lower oil vis
80 '"
... cosity resulting from significantly higher loaded and unloaded
<> ... ...
-E pad temperatures measured in the case of ISO VG 68.
Q)
E 60 Calculated losses and pad temperatures are in good agreement
Q) <>
()
III with the measured data. Calculations suggest that shear losses
Ci
en 0 <> <> <> predominate (>80%), with churning losses making up most of
i:5 40 n
the balance, and kinetic losses accounting for only 1-2% of the
losses.
20 II! ii I!!I
In the case of the offset pivot bearing, loaded pad 75% loca
tion temperatures can be somewhat lower than maximum pad
between 14 and 85 mIs, and unit loads that ranged between 0.34 and CLOSURE
2.19 MPa. Bearing operating conditions such as these result in a
Some assessment can be made using Figs. 11 and 16 when con
range of Sommerfeld numbers that are typic.al of journal bearings
sidering a change in oil viscosity grade. The data indicate that a
in field applications. The higher Sommerfeld numbers represent
switch from ISO VG 68 to ISO VG 32 in a comparable bearing will
highly loaded/low-shaft-speed operating conditions.
give a substantial decrease in pad temperature for surface speeds
Comparison of calculated and measured eccentricity are only
above 45 m/s. Eccentricity is already low at the higher speeds,
in fair agreement, which can be attributed to uncertainty in es
and a 20% increase using a lighter oil does not impose any se
tablishing the true zero position of the shaft. For these measured
rious consequence. Conversely, changing to an ISO VG 68 oil at
data, the zero position was based on 72 mls and zero load oper
higher speeds may result in unacceptably high pad temperatures,
ation. On the other hand, relative movement was very precisely
depending on the magnitude of the load.
measured by the instrumentation and methods described earlier,
Pad temperatures for both oils are low at the lower speeds.
and comparison shows very good correlation between calculated
Here, a change to ISO VG 68 can be used to increase film thickness
and measured trends.
with only a small increase in pad temperature and power loss.
Measured steady-state shaft displacements for the two test oils
It is not desirable to switch to ISO VG 32 at the lower speeds
are plotted in Fig. 16. The data represent the measured position of
because of the high eccentricities (thin oil films), unless the load
the shaft in the bearing clearance space for various operating con
is substantially lower than 2.19 MPa.
ditions relative to the 72 mls equilibrium position. The larger dis
placements represent a closer apprqacn to the loaded pad surface;
ACKNOWLEDGMENTS
that is, a larger displacement indicates a smaller film thickness.
The maximum measured displacement was 100 f../,m. This oc The authors are grateful to Kingsbury, Inc., and the National
curred at 14 mls with ISO VG 32 oil with a unit load of 2.19 MPa. Research Council of Canada for permission to publish the re
sults of this study. The authors particularly want to thank Brian
At the same operating condition, the ISO VG 68 shaft displace
ment is 87 f../,m. For all data, the ISO VG 68 oil ran with 8-33% Liko and Tim Breithaupt of the National Research Council of
Canada, without whose efforts this work would not have been
lower displacements, approximately 20% lower on average. The
lower displacements are a measure of thicker oil films in the case possible.
of ISO VG 68 oil.
REFERENCES
CONCLUSIONS
(1) Brockwell, K., DeCamillo, S. and Dmochowski, W. (2001), "Measured Tem
Tests have been performed on a 152-mm-diameter fiood perature Characteristics of 152 mm Diameter Pivoted Shoe Journal Bear
lubricated 60% offset conventional PSJ bearing, comparing per ings with Flooded Lubrication," Tribol. TrailS., 44, pp 543-550.
(2) Brockwell, K., Dmochowski, W. and DeCamillo, S. (2002), "Effect of Oil
formance using ISO VG 32 and ISO VG 68 oils. Key conclusions Flowrate on the SteadyState Performance of Pivoted Shoe Journal Bear
arising from these tests are as follows: ings," Tribol. Trans., submitted for publication.
488 K. BROCKWELL, W. DMOCHOWSKI AND S. DECAMILLO
(3) Simmons,J. and Dixon, S. (1994), "Effect of Load Direction, Preload, Clear But
ance Ratio, and Oil Flow on the Performance of a 200 mm Journal Pad
Bearing," Tr ibal. Trans., 37, pp 227-236. CM = 0.982(1ogRe)-2.S8
(4) Simmons, J. and Lawrence, C. (1996), "Performance Experiments with a
200 nun, Offset Pivot Journal Pad Bearing," Tribal. Trans., 39, pp 969-973.
(Dorfman (12)) where Re = Reynolds number = r2wlv and v =
(5) DeChoudhury, P. and Masters, D. (1983), "Performance Tests of Five-Shoe
Tilting-Pad Journal Bearing," Tribal. Trans., 27, pp 61-66. kinematic viscosity.
(6) DeChoudhury, P. and Barth, E. (1981), "A Comparison of Film Tempera For disk, power loss:
tures and Oil Discharge Temperature for a Tilting-Pad Journal Bearing,"
ASME J. Lub" Technol., 103, pp 115-119.
(7) Brockwell, K. and Kleinbub, D. (1989), "Measurements of the Steady State
Operating Characteristics of the Five Shoe Tilting Pad Journal Bearing,"
For cylinder, power loss:
Tribal. Trans. , 32, pp 267-275.
(8) Pettinato, B. and DeChoudhury, P. (1999), "Test Results of Key and Spher
ical Pivot Five-Shoe Tilt Pad Journal Bearings-Part 1: Performance Mea
surements," Tribal. Trans., 42, pp 541-547.
For part of a cylinder, power loss:
(9) Tripp, H. and Murphy, B. (1984), "Eccentricity Measurements on a Tilting
Pad Bearing," ASME Trans., 28, pp 217-224.
(10) Dmochowski, w., Brockwell, K., DeCamillo, S. and Mikula, A. (1992), PCH = Pc(2rr - nrp)r/2rrr
"A Study of the Thermal Characteristics of the Leading Edge Groove
and Conventional Tilting Pad Bearings," ASME J. Tribal., 115, pp 219- where n = number of pads and rp = angular extent of pads.
226. Hence,
(11) API Standard 670 (1986), "Vibration, Axial-Position, and Bearing Temper
ature Monitoring Systems," American Petroleum Institute, Washington, PCH = 0.982(1- (ncp/2rr)) x Low3r4 x (logRe)-2.s8
DC.
(12) Dorfman, L. (1963), Hydrodynamic Resistance and the Heat Loss of
Rotating Solids, Oliver and Boyd, Edinburgh, London. Kinetic Losses
Kinetic energy:
APPENDIX A
EK = mV2/2
Churning Losses