43rd Turbomachinery & 30th Pump Users Symposia (Pump & Turbo 2014)

September 23-25, 2014 | Houston, TX | pumpturbo.tamu.edu

AXIAL SUBSYNCHRONOUS VIBRATION

Scan DeCamillo
Manager, Research and Development
Kingsbury, Inc.
Philadelphia, PA 19154 USA

Scan DeCamillo is Manager of Research INTRODUCTION

and Development for Kingsbury, Inc.,
responsible for design, analysis, and The axial SSV under consideration is a discrete
development of Kingsbury fluid film subsynchronous axial vibration of the rotor of relatively high
bearings for worldwide industrial and amplitude, sometimes traveling the full set clearance of the
military applications. He began work in thrust bearings. This type of axial SSV has been witnessed
this field in 1975 and has since provided during mechanical tests as well as in field applications for
engineering support to industry various types of machines from small compressors to large
regarding the application and steam and gas turbines. The occurrence is usually unexpected
performance of hydrodynamic bearings. and can be dangerous because axial vibration is not typically
monitored in industrial applications.
Mr. DeCamillo has developed performance and structural
bearing analysis tools during his career, establishing design Three personal experiences are given here for the purpose
criteria used in many publications and specifications. He has of discussion. The first was an application for a 200 MW gas
patents and has authored several papers on bearing research, turbine that encountered an unexpected SSV during
which is currently focused on advancing hydrodynamic bearing commissioning. The SSV appeared suddenly as the rotor speed
technology for high-speed turbomachinery applications. passed 2700 rpm and disappeared above 3150 rpm. The
frequency was 30 percent of the rotating speed with a 26 mil
Mr. DeCamillo received his B.S. degree (Mechanical (0.66 mm) peak-to-peak amplitude. The axial SSV coincided
Engineering, 1975) from Drexel University. He is a registered with an operating thrust reversal from forward to aft at 2900
Professional Engineer in the State of Pennsylvania and a rpm.
member of STLE, ASME, and the Vibration Institute.
Another experience was with a small high-speed steam
ABSTRACT turbine that encountered a 4 mil (0.10 mm) peak-to-peak
indication at a frequency of 30 percent of the rotating speed as
Severe axial subsynchronous vibrations (SSV) can be the shaft speed passed 10,500 rpm. The SSV, also coincident
encountered in turbomachinery. A major concern is the with a thrust load reversal, disappeared after 11,500 rpm with
potential for danger from the high levels of vibration acting on no trace up to 14,500 rpm.
the rotor. Longer term implications are fretting and fatigue
damage of the rotor components and support structures, which The third experience involved a discrete sub-synchronous
include the thrust bearing pivot contact areas. Although axial vibration during no-load spin tests of a high-speed compressor.
SSV is a concern in turbomachinery applications, little Amplitudes on the order of 3 mils (.08 mm) peak-to-peak at a
published literature can be found on the subject. A series of frequency of approximately 60 percent of rotating speed were
laboratory tests were therefore designed and performed to encountered as the shaft speed passed 10,000 rpm and persisted
investigate the phenomenon in more detail. up to the design operating speed of 12,000 rpm.

This paper presents experimental data from a series of This type of axial SSV is known to be sensitive to no-
pivoted shoe thrust bearing tests investigating the influence of thrust-load situations, which is the case in the above examples.
operating conditions and other parameters on axial SSV. Indications tend to disappear under load and so a common
Parameters include thrust load, shaft speed, oil flow, inlet industry technique is to “preload” the thrust bearings by
temperature, axial clearance, pivot offset, and direct and reducing the axial clearance. The solution, however, was not
flooded lubrication. Test results are used to evaluate techniques successful or repeatable in all of the above situations.
that have been used when attempting to resolve axial vibration
issues. Additional data from thrust shoe load cells and thrust A literature search finds little published research on the
shoe proximity probes provide unique and valuable insight that subject of axial shaft SSV in turbomachinery. Gardner (1998)
allowed development of a more robust solution. reports on an experimental study of thrust pad flutter

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station

investigating pivot damage of the inactive side thrust pads. The For tests investigating the influence of external thrust load,
paper is often used as a reference but does not actually measure the test bearings were installed in positions A and B and a
or comment on axial shaft vibration. Other techniques (gathered single thrust bearing was installed in position D to apply the
from personal experience and in conversations with others) load. For tests investigating parameters other than external
include increasing the thrust bearing oil flow, modifying the thrust load, the sliding housing was bolted securely to the
balance piston, reducing the pivot offset, and creating baseplate and the test bearings were installed in positions A and
asymmetry between loaded and slack bearings. These have also B or positions D and C, depending on the test. In all cases the
been applied with varied levels of success and failure, being test bearings were carefully shimmed to set the axial clearance
based mostly on anecdotal evidence. and to hold the position the shaft so that no axial force is
transmitted through the coupling.
In order to better understand the nature of axial SSV, the
author’s company, Kingsbury, Inc., designed and performed a
series of pivoted shoe thrust bearing tests to gather
experimental data and investigate the influence of various
parameters. Results are presented in this paper. The intention is
to provide information of value to those involved with
hydrodynamic bearings and vibration in turbomachinery.

TEST RIG AND TEST BEARINGS

Kingsbury’s high-speed test rig is described fully in

reference Wilkes, et al. (2000). A brief description is given here
to explain how the axial vibration tests were configured.
Figure 2. Test Rig Schematic.
The test rig shaft is driven by a variable speed gas turbine
Four thrust bearing designs were tested during the series
with a rated output of 1100 horsepower (820 kW) and a
including conventional flooded center and offset pivot 6-shoe
controllable test speed range of 4,000 to 17,000 rpm. The
bearings, a 6-shoe 60% offset pivot leading edge groove
turbine is connected to the test shaft by a flexible coupling.
(LEG®) direct lube bearing, and an 8-shoe 65% offset pivot
Two identical housings external to the turbine enclosure
LEG® direct lube bearing which is shown in the photograph of
contain the bearing components (Figure 1). Two tilting-pad
Figure 3. All bearings have a babbitt outside diameter of 10.5
journal bearings support the test shaft.
inches (267 mm) and a total bearing area of approximately 55
square inches (35484 mm2). Parametric tests of various
operating conditions and other geometry modifications were
studied over the course of the series, which are explained in
following sections. All tests were run with ISO VG 32 turbine
oil.

Figure 1. Test Rig Housings and Bearings.

The existing test rig was designed to test two double thrust
bearings under load. Referring to Figure 2, the fixed housing
adjacent to the turbine is firmly secured to the foundation while
the sliding housing is restrained but is free to slide axially.
Axial load is applied by means of an external hydraulic system
which transmits force to the sliding housing. As a result, thrust
bearings A and D are loaded against each other while thrust
bearings B and C remain unloaded. A load cell mounted on the
end of the sliding housing measures the applied force.
Figure 3. 8-shoe 65% Offset Pivot Direct Lube Test Bearing.

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station

 A National Instrument LabVIEW data acquisition system
was used to record the test data. Steady-state measurements
include shaft speed, applied thrust load, oil flow, oil supply
pressure, oil supply temperature, oil drain temperature, and pad
temperature. Load cells were mounted under a thrust pad on
either side of the collar to monitor and record hydrodynamic
film forces.

High-speed waveform data were recorded for seven

vibration probes. Radial probes were mounted inboard of each
of the fixed and sliding housing journal bearings, 45 degrees off
top dead center, and an axial probe monitored the end of the
shaft as indicated in Figure 4. Proximity probes were also
mounted under a thrust pad on either side of the collar to
monitor pad vibration (Figure 5). An FFT is performed to
generate frequency spectrums and waterfall plots as shown in
example Figures 6 and 7. The FFT and waterfall amplitude
scale denotes peak rms in mils (thousandths of an inch). The
backplane of the waterfall plot displays a projection of the data,
Figure 6. Example Wave Form and FFT.
and the record number is an indication of time in seconds.

Axial Probe
Radial
Probes

104 F
14 krpm

10 gpm
100 F

20 gpm
92 F

Figure 4. Fixed Housing Proximity Probes.

Figure 7. Example Waterfall Plot.

REPRODUCING AXIAL SSV ON THE RIG

Many operating conditions and parameters were

investigated in an effort to understand the nature of axial SSV
and to find a suitable method or solution to address axial SSV
issues. Test rig operating conditions cover a speed range from
4,500 to 16,500 rpm and oil flow capacity up to 60 gpm (227
l/m) per bearing. Inlet temperatures from 80°F to 150°F (27°C
to 66°C) and axial clearances between .004 and .019 in (0.10
and 0.48 mm) were also investigated. Representative data from
the test series are provided in the following sections to relay the
main points of the tests and observations.

Finding an Axial SSV Signature

The first series of tests were performed with 8-shoe 65%

offset pivot direct lube bearings in fixed housing positions A
and B, and a single 6-shoe 60% offset pivot direct lube bearing
Figure 5. Thrust Pad Proximity Probe Location. in sliding housing position D, backed far enough away from the

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station

collar so as not to interfere with the no-load dynamics of 8-SHOE BEARING NO-LOAD TESTS
bearings A and B. Initial tests were run under no-load
conditions while adjusting the inlet temperature, shaft rpm, oil The thrust bearing in position D was removed, the sliding
flow, and axial clearance in attempt to find an axial SSV housing was bolted securely to the baseplate, and the test
indication. bearings were run under no-load conditions for the remainder
of the test series.
Many hours of test time were spent varying the parameters
and the search began to seem futile until a discrete Influence of Oil Flow
subsynchronous vibration suddenly appeared at 14,000 rpm,
indicated by an arrow in the associated waterfall plot of Figure Operating conditions for the 8-shoe 65% offset pivot
7. Oil flow, oil inlet temperature and axial clearance at the time bearing were again set close to those that excited the axial SSV
were 10 gpm, 104°F, and .005 in (38 l/m, 40°C, and 0.13 mm), above and the oil flows to sides A and B were varied
respectively. The indication was an important aspect of the tests independently. In this test the axial SSV displayed a strong
as axial SSV was successfully produced on a laboratory rig sensitivity to small changes in flow. Figure 9 contains data for
without many of the common sources of excitation found in operating conditions of 15,000 rpm, 114°F (46°C) oil inlet
turbomachinery. The success allowed continuation of a test temperature, and .005 in (0.13 mm) axial clearance. The figure
series to investigate how operating conditions and other begins at record 800 with 10 gpm (38 l/m) to bearing A and 10
parameters influence the behavior of axial SSV. gpm (38 l/m) to bearing B. Arrows illustrate the results of a 3
gpm (11 l/m) increase first to bearing B and then to bearing A.
Influence of Load The disappearance of the 165 Hz subsynchronous frequency is
noticeable in the figure.
Beginning with the conditions that excited the axial SSV
above, the influence of thrust load was next investigated by
applying load using thrust bearing D. The results were fairly
straight forward and consistent with field experience in that
axial SSV indications disappeared when load was applied.
Example results are shown in Figure 8.

13 gpm A

10 gpm
A&B

13 gpm B

10 gpm
No load A&B

Figure 9. Influence of Oil Flow, 8-Shoe 65% Offset Bearing,

Load 15,000 rpm, 114°F Inlet, .005 Inch Clearance.
applied

Influence of Axial Clearance

No load
The axial clearance for the 8-shoe 65% offset pivot
Figure 8. Application and Release of External Load. bearings was incremented over the course of the test series to
investigate the influence on axial SSV. The clearance was
Another important observation from Figure 8 is a shift in changed to a more typical value of .012 in (0.30 mm) for the
subsynchronous frequency between record numbers 5640 and next series of tests.
5740. The load was released very slowly during this interval to
investigate the onset of axial SSV. A 160 Hz indication Operating conditions were again set to those that excited
appeared as the bearing unit load passed below 20 psi (0.14 the axial SSV above, but the indication did not return. At first it
MPa) and drifted to 125 Hz as the thrust load approached zero. appeared as though the increase in clearance had eliminated the
The results indicate that axial SSV can occur not only in no- SSV. Further investigation found that it had shifted to a
load or load reversal situations but also in the case of low loads, different set of operating conditions. By varying speed while
which can have a significant influence on SSV frequency. maintaining the 10 gpm (38 l/m) oil flow, a 45 Hz SSV was
eventually encountered at 13,000 rpm. This 45 Hz SSV is
indicated by arrows at the beginning and ending records of
Figure 10.

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station

 sensitive condition. The goal to assess negative or positive
influence was abandoned and subsequent tests varied the
parameters in more refined increments to acquire a more
detailed map of the response.

Intermediate Assessment

10 gpm Assessment of the test results from a resonance point of

44 gpm view allowed for a more consistent generalization of the test
results. Parameters that tend to increase the stiffness of the
44 to
60 gpm bearing (i.e. lower inlet temperature, higher oil flow, tighter
46 gpm axial clearance) act to increase the frequency of the discrete
axial SSV indications. The axial SSV is sensitive to a particular
set of operating conditions which also changes with adjustment
10 gpm of the parameters. Resulting amplitudes therefore may be
noticed to grow larger or smaller or to appear and disappear
depending on proximity to the sensitivity.

6-SHOE BEARING NO-LOAD TESTS

Figure 10. Flow Ramp (10 to 60 to 10 gpm), 8-Shoe 65% Offset
Bearing, 13,000 rpm, 114°F Inlet, .012 Inch Clearance. The axial vibration series continued with tests of 6-shoe
60% offset direct lube thrust bearings installed positions D and
Continuing with the parametric study, oil flows to bearings C. Figure 11 plots axial vibration data for four speed ramps
A and B were varied from 10 to 60 to 10 gpm (38 to 227 to 38 performed with 10, 20, 30, and 40 gpm oil flows (38, 76, 114,
l/m). As a reference, 40 to 50 gpm (150 to 190 l/m) would be a and 151 l/m) as indicated in the figure. The data is for operation
suitable flow range for field applications of this bearing size with an axial clearance of .012 in (0.31 mm) and an inlet
and speed. A second SSV indication at 113 Hz was encountered temperature of 130°F (54°C).
near 44 gpm (167 l/m), also noticeable in Figure 10. The
occurrence of two distinct axial shaft SSV frequencies is An axial SSV sensitivity to 30 gpm (113 l/m) is noticeable
similar to Gardner’s (1998) inactive side pad vibration in the data for this particular bearing configuration. The
observations. indication came in as the shaft speed passed 13,000 rpm and
persisted up to the maximum test speed of 16,500 rpm.
The low frequency indications (below 12 Hz in Figure 10) Referring to Figure 11, the SSV signature can be seen to
represent random shifts in the shaft position between the two increase in frequency and amplitude from the slightest
bearings. The movement, commonly referred to as shaft shuttle, indication in the 10 gpm speed ramp to the more obvious
was consistent through the axial SSV and to the highest test indication in the 30 gpm ramp. No indication is apparent in the
flow. Tighter clearances (Figure 9 for example) tended to 40 gpm data.
reduce the low frequency amplitudes and to increase the
frequency of any discrete axial SSV indications.

Other Parameters and Comparisons

Initial investigations of axial SSV compared broad changes

in parameters such as high versus low oil flow. The goal was to
assess if the parameter had a negative or positive influence on
axial SSV. Generalizations became difficult as comparisons
seemed to produce conflicting results. High oil flow, for 40 gpm
example, appeared to eliminate SSV in some cases, and low oil
flow in others. Similar results occurred with oil inlet 30 gpm

temperature, axial clearance, and flooded versus evacuated

20 gpm
comparisons. The test producing the results in Figure 10 helped
clarify the situation. 10 gpm

Figure 10 data shows a strong SSV sensitivity at 44 gpm

(167 l/m) for this particular bearing configuration and operating
conditions. It is apparent that the SSV indication can be
eliminated by increasing or by decreasing the flow. Behavior is Figure 11: Speed Ramps, 6-Shoe 60% Offset Direct Lube
indicative of the excitation of a natural frequency where an Bearing, 130°F Inlet, .012 Inch Clearance .
increase or decrease in a parameter can cause an increase or a
decrease in response depending on the proximity to the As a side note, the increase in synchronous amplitude at

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station

275 Hz is a false indication traced to the natural frequency of of operating conditions tested. The axial SSV indications for
the axial proximity probe bracket. There is also a 45 Hz the center pivot bearing were of very low amplitude in a
indication in the data attributed to a structural resonance. The frequency range between 20 and 75 Hz. Figure 13 depicts four
indications are present in all 6-shoe vibration figures. speed ramps for the bearing performed with 10, 20, 30, and 40
gpm (38, 76, 113, and 151 l/m) oil flows.
Influence of High-Speed Tapers

The test series included investigation of a leading edge

taper (noticeable in Figure 3) described in detail by Wilkes, et
al. (2000). In summary, high-speed turbomachinery spin tests
and certain field conditions can encounter a situation of high
pad temperatures at low load. The high-speed/low-load
phenomenon is attributed to conditions where friction forces,
which act to close the leading edge film, dominate over 40 gpm

hydrodynamic film forces. The high-speed taper acts to keep

30 gpm
the leading edge of the pad from closing under these conditions
(Wilkes and DeCamillo, 2000).
20 gpm

Figure 12 contains vibration data results from tests of the

10 gpm
6-shoe 60% offset direct lube thrust bearings with high-speed
tapers at conditions equivalent to Figure 11. Comparisons of
the figures indicate an increase in axial SSV frequency and
amplitude. Considering that the tests are under no-load
conditions, the increase in SSV frequency suggests an increase
in film stiffness, which is consistent with the function of the Figure 13: Speed Ramps, Center Pivot Flooded Bearing.
tapers.
To better understand the low indications of axial SSV in
the center pivot bearing, attention was given to the thrust shoe
load cell data. The data was collected from two load cells, one
mounted under a thrust shoe (Figure 14) in the bearings on
either side of the collar. In the case of these no-load tests, the
pad load data represents a measurement of the hydrodynamic
film force acting on the pad.

40 gpm

30 gpm

20 gpm

10 gpm

Figure 12: Speed Ramps, 6-Shoe 60% Offset Direct Lube

Bearing with Tapers, 130°F Inlet, .012 Inch Clearance .

Conventional 6-Shoe Bearing Tests

The axial vibration series continued with tests of 6-shoe

center and 60% offset pivot conventional flooded bearings. As
above, tests include the study of the influence of axial
clearance, speed, inlet temperature, and oil flow as well as the
influence of high-speed tapers on the offset pads.
Figure 14: Example Load Cell Position (Thrust Pad Removed).
In the case of the center pivot bearing, it was only possible
to get the faintest indication of axial SSV over the broad range

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station

Hydrodynamic Film Forces or when it was at a relatively high level. In the case of Figure
17, for example, SSV indications were high at pad loads
In viewing pad load data for the 6-shoe center pivot between 75 and 150 lbs, indicated by the shaded area in the
bearing, it was noticed that the film force on either side of the figure.
collar dropped out at a relatively low speed. Figure 15 plots
speed and pad load data corresponding to the axial probe
10 gpm 20 gpm 30 gpm 40 gpm
vibration data of Figure 13. Figure 16 plots the same data
against speed where the variation of pad load with speed and oil
flow is more apparent.

Pad Load (lbs)

10 gpm 20 gpm 30 gpm 40 gpm

Pad Load (lbs) Figure 17: Offset Pivot Direct Lube Bearing, Pad Load vs
Time.

Figure 15: Center Pivot Flooded Bearing, Pad Load vs Time.

40 gpm
30 gpm
20 gpm
10 gpm
40 gpm
30 gpm
20 gpm
10 gpm

Figure 18: Offset Pivot Direct Lube Bearing, Pad Load vs

Speed.
Figure 16: Center Pivot Flooded Bearing, Pad Load vs Speed.
O-Ring Damper

Basic hydrodynamic theory predicts an increase in film The pad load data suggested that preloading the bearing by
force with speed, contrary to the test results. The data, however, reducing axial clearance would eliminate SSV if the
actually supports the high-speed/low-load phenomenon hydrodynamic force could be kept sufficiently high. This was
described by Wilkes and DeCamillo, 2000. Figures 17 and 18 possible in many cases, but was difficult to achieve over the
contain pad load data for the offset 6-shoe high-speed taper broad range of operating conditions tested.
bearing corresponding to the vibration data of Figure 12. Figure
18 also shows a decreasing trend in pad loads at the higher An idea was conceived to impose a relative high preload
speeds although some level of hydrodynamic film force persists by means other than axial clearance, which is described in
which is attributed to the influence of the tapers. detail by Wilkes and DeCamillo (2006). Figures 19 and 20
show a version of the concept in the form of an O-ring placed
While comparing and evaluating pad load data against between thrust bearing side C and its casing. The design was
vibration results, a trend was noticed in that conditions exciting installed in the 6-shoe offset direct lube tapered thrust bearing
the axial SSV to significant levels were accompanied by a and subjected to a duplicate series of tests.
certain level of hydrodynamic load. Axial SSV was not
apparent in cases where the film force dropped out completely

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station

 Although the original intention of the O-ring was to raise
the hydrodynamic force, the significant reduction in amplitudes
over the entire range of frequencies suggests a high level of
damping, derived from oil in the gap between the bearing and
housing under no-load conditions. The device is therefore
referred to as an O-ring damper.

Figure 19: Half Bearing Showing the O-Ring Design.

Pad Load (lbs)

Figure 22: Offset Pivot Direct Lube with O-Ring, Pad Load vs
Time.

Figure 20: Assembled Bottom View Showing the O-Ring

Design.

As an example, the axial vibration results of Figures 21, 40 gpm

30 gpm
22, and 23 with the O-ring duplicate the conditions of Figures 20 gpm
10 gpm
12, 17, and 18 without. Comparisons show relatively high
levels of pad load and a very low amplitude axial vibration
signature with no indications of axial SSV.

Figure 23: Offset Pivot Direct Lube with O-Ring, Pad Load vs
Speed.

Pad Vibration

40 gpm Pad vibration data were recorded during the course of the
axial SSV test series, which is of importance as it concerns the
30 gpm potential for fretting and fatigue damage of the thrust bearing
pivot contact areas. Figure 24 contains pad vibration data for
20 gpm the 6-shoe offset pivot bearing corresponding to the axial shaft
vibration data of Figure 12. The frequency scale is extended to
10 gpm
750 Hz to capture the range of super-synchronous indications
noticed in pad vibration test results.

Figure 25 is a record of duplicate conditions performed

with the O-ring damper corresponding to the axial shaft
Figure 21: Speed Ramps, 6-Shoe 60% Offset Direct Lube vibration data of Figure 21. A significant reduction of all pad
Bearing with O-Ring, 130°F Inlet, .012 Inch Clearance. vibration amplitudes is obvious in the comparison.

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station

 The main problem is that the techniques have not proved to be
reliable or repeatable:

1. Increase the oil flow

2. Tighten the axial clearance
3. Modify the balance piston
40 gpm
4. Reduce the pivot offset
5. Create asymmetry
30 gpm

The axial SSV test results suggest a resonant type

20 gpm behavior, which can explain the apparent lack of consistency
and help provide a better understanding of the intended purpose
10 gpm
of the technique and possible outcomes when troubleshooting
axial SSV issues.

Increase the Oil Flow

Increasing the thrust bearing oil flow has proven to be a

Figure 24: Pad Vibration Speed Ramps, 6-Shoe 60% Offset viable solution in past experience. A common mistake is to
Direct Lube Bearing, 130°F Inlet, .012 Inch Clearance. assume axial SSV is a sign of starvation, which is true in low
oil flow situations. If there is sufficient flow, however, an axial
SSV indication is likely a system resonance. In this case,
adjustments to flow can worsen the response. The flow has to
be increased (or decreased) a sufficient amount to move the
resonance out of the range of operating conditions. Of all the
techniques, this is the most feasible option and easiest to adjust.
40 gpm
The greatest challenge is to assure the axial SSV does not
appear over the required range of operating conditions
(min/max inlet temperature, supply pressure, speed range, etc.).
30 gpm

20 gpm
Tighten the Axial Clearance

10 gpm
This is another viable solution proven successful in past
experience. The task is more involved as the machine needs to
be opened to measure and adjust the bearing clearance.
Inconsistent results (higher amplitudes for example) may
appear if the endplay does not push the resonant frequency
sufficiently high. Another variable is that axial thrust bearing
Figure 25: Pad Vibration Speed Ramps, 6-Shoe 60% Offset clearance is difficult to set precisely and repeatedly. If
Direct Lube Bearing with O-Ring, 130°F Inlet, .012 Inch clearance is used to address axial SSV symptoms, it is
Clearance. important to account for possible changes in clearance over
time. Opening the clearance is another possible option.

Similar comparisons of shaft and pad vibration data in Modify the Balance Piston
duplicate tests with and without the O-ring show that the design
eliminates the axial shaft and pad SSV as well as the super- This is a suitable option in many applications, again
synchronous pad vibrations, and generally lowers all addressing the fact that hydrodynamic thrust bearings do not
synchronous and low frequency broadband indications. The run well under low or no-load conditions, especially in high-
design in Figures 19 and 20 and variations of the design have speed machines. It is prudent to design for the least amount of
been successfully applied to resolved axial SSV issues in time in no-load, low load, or load reversal operation when
various turbomachinery. possible. Situations where a balance piston modification was
not successful may be due to an insufficient increase in bearing
DISCUSSIONS load. The bearings and configuration in the axial SSV test
series actually showed an increase in SSV frequency and
Many of the axial SSV tests were performed to assess amplitude with load before it dropped out above 20 psi.
techniques that have been used when attempting to resolve
axial vibration issues. The techniques are based on personal Reduce the Pivot Offset
experience as well as solicitations of colleagues as such
information was unavailable from published literature. The five Test results indicate that geometry that acts to decrease the
techniques tabulated below have had some measure of success. bearing stiffness tend to produce lower frequency and lower

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station

amplitude SSV. Reducing the pivot offset has been successful grow larger or smaller or to appear and disappear.
in eliminate axial SSV in field applications. Changing the pad
offset is involved as it requires machining modifications or The test results were used to critique five industry
replacement of the pads or bearing. The solution to reduce techniques that have been used when attempting to resolve
offset also has consequences of higher bearing temperatures axial vibration issues. The results explain the apparent lack of
and lower load capacity, and may not be a suitable option in consistency and help provide a better understanding of the
high speed machinery. intended purpose and possible outcomes of the techniques.

Create Asymmetry Pad load cell data displayed a drop in hydrodynamic film
force with speed, contrary to basic hydrodynamic theory. Based
This technique has mostly been used as a troubleshooting on the observations, a device was designed in the form of an O-
tool. The theory was to break the symmetry of the system or to ring damper that produced high film forces over the entire
bias the film forces by using different main and slack bearings speed range, and very low amplitude shaft and pad vibration
(removing every other shoe from the slack side bearing, for signatures with no indications of axial SSV. The device and
example). The technique was successful in one application, and variations of the device have been successfully applied to
unsuccessful in another. Based on test results, it is unlikely that resolved axial SSV issues in various turbomachinery.
asymmetry is a valid theory to confront axial SSV issues. The
arrangement is still a system with resonant frequencies and CLOSING REMARKS
associated sensitivities. This technique is not recommended as a
viable solution. The inference of a natural frequency raises a question as to
the source of the excitation. Gardner (1998) suggests periodic
O-Ring Damper pressure pulsations in the discharge annulus, which is not a
possibility in the axial SSV test rig configuration. A separate
The main focus of the axial SSV test series has been on study of the shaft and pad vibrations of Figures 12 and 24
shaft SSV. Discussions of the preceding techniques above are displayed an intricate interaction between the supersynchronous
from a shaft vibration point of view and do not necessarily pad vibration and subsychronous shaft vibration that may hold
indicate the same behavior and results regarding pad vibration. a key.

The O-ring damper is the result of the SSV tests series goal There is certainly a need for more research and
to understand the nature of axial shaft SSV and provide a more development, experimental and theoretical, on the subject of
robust solution. It is the only technique investigated that axial SSV. In the meantime, the author hopes that the tests and
confirmed a reliable and significant reduction in pad vibration. data presented in this paper will provide a useful reference for
topics related to axial vibration in turbomachinery.
The design has proved to be reliable and consistent in
eliminating axial SSV. The O-ring version of the device REFERENCES
requires periodic inspection and replacement, and the bearing
design and assembly requires extra steps and considerations. Gardner, W. W., 1998, “An Experimental Study of Thrust Pad
Flutter,” ASME Journal of Tribology, 120, pp 577-582.
CONCLUSIONS
Wilkes, J., DeCamillo, S., Kuzdzal, M. and Mordell, J., 2000,
Axial SSV was successfully produced on a laboratory rig “Evaluation of a High Speed, Light Load Phenomena in
without many of the common sources of excitation found in Tilting Pad Thrust Bearings,” Proceedings of the 29th
turbomachinery. This allowed a series of pivoted shoe thrust Turbomachinery Symposium, Turbomachinery Laboratory,
bearing tests to be performed to investigate the influence of Texas A&M University, College Station, Texas, pp. 177-
operating conditions and other parameters on axial SSV. 185.

The test results support spin test and field experience that Wilkes, J. and DeCamillo, S., 2000, “Thrust Bearing,” United
axial SSV is eliminated under conditions of applied load. The States Patent No. 6,089,754, Jul. 18, 2000.
data more specifically indicate that the load must be of ample
magnitude. Loads above 20 psi (0.14 MPa) were required to Wilkes, J. and DeCamillo, S., 2006, “Thrust Bearing,” United
eliminate the SSV indication for the bearing size and States Patent No. 7,063,465 B1, Jun. 20, 2006.
configuration tested.
ACKNOWLEDGEMENTS
The axial SSV indications displayed a resonant type
behavior. Parameters that tend to increase the bearing stiffness The author would like to thank Kingsbury, Inc., for
(i.e. lower inlet temperature, higher oil flow, tighter axial permission to publish the information in this technical report as
clearance, pivot offset, and high-speed tapers) act to increase well as for all of the tables and figures containing data that have
the frequency. The axial SSV is sensitive to a particular set of been accumulated from the company’s engineering and
operating conditions which also changes with adjustment of the research and development laboratories.
parameters. Resulting amplitudes therefore may be noticed to

Copyright© 2014 by Turbomachinery Laboratory, Texas A&M Engineering Experiment Station