More Related Content
Similar to FEDSM2013-16202
Similar to FEDSM2013-16202 (20)
FEDSM2013-16202
- 1. 1 Copyright © 2013 by ASME
Proceedings of the ASME 2013 Fluids Engineering Division Summer Meeting
FEDSM2013
July 7-11, 2013, Incline Village, Nevada, USA
FEDSM2013-16202
EFFECTS OF ADDING A CENTRAL GROOVE TO THE SQUEEZE FILM
DAMPER LANDS
Praneetha Boppa
Mechanical Engineering Dept.
Texas A&M University
College Station, Texas USA
Boppa.pranitha@gmail.com
Aarthi Sekaran
Mechanical Engineering Dept.
Texas A&M University
College Station, Texas USA
Aarthi.sekaran@neo.tamu.edu
Gerald Morrison
Mechanical Engineering Dept.
Texas A&M University
College Station, Texas USA
gmorrison@tamu.edu
ABSTRACT
Squeeze film dampers (SFDs) are used in the high
speed turbo machinery industry and aerospace industry as a
means to reduce vibration amplitude, to provide damping, to
improve dynamic stability of the rotor bearing system and to
isolate structural components. Past studies have not included
effects of variation of the stator geometries in a squeeze film
damper. A central groove added to the squeeze film land is
hypothesized to provide a uniform flow source which theory
predicts will result in forces less than one fourth of that seen in
SFDs without a central groove. In the present study, 3D
numerical simulations of SFDs with different size central
grooves on the squeeze film land are performed to predict the
variation of the dynamic pressure profiles. The numerical
model and method have been validated via comparison to
experimental data for a SFD without a central groove. When a
central groove is added to the squeeze film land, the pressures
generated are reduced to half of that generated when run
without a central groove on the land. The amount of reduction
in pressure values depends on the volume of the groove, not on
the aspect ratio of the groove. Addition of a central groove
reduces the pressures, rigidity developed in squeeze film land,
and forces generated by squeeze film damper.
INTRODUCTION
Squeeze film dampers are bearings whose main
purpose is to provide viscous damping in Turbomachinery
systems. A squeeze film damper consists of a fixed journal
(bearings on the outer layer of the journal) and housing. The
clearance between the journal and housing is filled with
lubricant that is supplied at constant external pressure. The
whirling motion of the journal, due to imbalance forces,
generates pressure inside the lubricant film which, in response
to these pressure forces, produces the damping forces that
reduce the vibrations of the rotor. The lubricant film in between
the journal and housing is squeezed to produce this damping
reaction force and hence the name - squeeze film dampers.
Design and cavitation parameters are of particular
importance while analyzing SFDs and predicting their
performance – peak performance is obtained when there is no
cavitation in the lubrication film. There are different kinds of
cavitation that can occur in SFDs. A comprehensive history of
SFDs including the detailed design and application parameters
is described by Zeidan, San Andres and Vance (1996). Zeidan
and Vance (1990a, 1990b) have also studied cavitation in SFDs
in detail identifying five different cavitation regimes which
were distinguished by studying the pressure curves during
operation.
A squeeze film damper has a non-rotating journal
mounted on ball bearings and a stationary outer housing
surrounding it. The rotating motion of the journal is constrained
by a squirrel cage and the gap between the journal and housing
is filled with lubricant as shown in Figure 1. As the journal
whirls or oscillates due to forces of imbalance, reaction forces
are developed in the lubricant film. The annular squeeze film,
typically less than 0.250 mm, between the journal and housing
is filled with a lubricant provided as a splash from the rolling
element bearing lubrication system or by a dedicated
pressurized delivery. In operation, as the journal moves due to
dynamic forces acting on the system, the fluid is displaced to
accommodate these motions. As a result, hydrodynamic
squeeze film pressures exert reaction forces on the journal and
provide for a mechanism to attenuate transmitted forces and to
reduce the rotor amplitude of motion (San Andres, 2010).
Figure 1(a) shows the location and configuration of the
central groove, more generously known as the feed grove. The
function of this groove is to ensure a continuous supply of
lubricant to the squeeze film land. A groove is thought to
provide a uniform flow source with constant pressure around
the bearing circumference. A central groove, in particular, also
- 2. 2 Copyright © 2013 by ASME
divides the flow region into two separate squeeze film dampers
working in parallel, i.e. the reaction forces from each land add
up. Studies from Arauz et al. (1997) and Childs et al. (2007)
also show that although according to theory, SFDs with grooves
would have smaller forces than those without grooves,
experiments shows larger forces than what was predicted. This
led to a detailed analysis of these SFDs and the conclusion that
central groove does not isolate the adjacent film lands, but
rather interacts with the squeeze film regions
For the present study, CFD simulations of two general
configurations were carried out. The first was a journal with a
smooth surface as investigated experimentally by Delgado, A.,
(2008) and depicted in Figure 2(a). This SFD was modified to
place a groove circumferentially around the journal as depicted
in Figure 2(a). Several grooves cross sectional areas and aspect
ratios were also considered. Central groove clearances varied
from 0.5mm to 4mm and their axial lengths varies from
6.35mm to 12.7mm. Simulations were run to determine the
influence of dimensional parameters on SFD performance. The
parameters of the seal are as listed in the table below.
Table 1. Squeeze Film Damper Dimensions.
Parameter Value
Rotor radius 63.5mm
Stator radius 63.627mm
Eccentricity between
stator and rotor 0.074mm
Inlet groove clearance 78c
Length of Inlet groove 6.35mm
Discharge groove
clearance 31c
Length of Discharge
groove 4.1mm
where c is the clearance between the stator and the
rotor and is 0.127mm
Since the use of experimental techniques to study
intricate details of a flow field is time consuming and
expensive, the use of numerical techniques for studying SFDs
is particularly attractive. Prior numerical studies have, however,
been limited to employing unsteady and computational
intensive models to solve the time dependent flow field of these
squeeze film dampers. An important numerical study was the
work done by Guo et al. (2005) who applied the Moving Grid
method and conjugate heat transfer for analysis of damper
characteristics and squeeze film dampers. Though these
simulations proved to be time consuming, the techniques were
seen to be efficient for complex geometries. A recent study was
also carried out by Khandhare (2011) who applied the Dynamic
Meshing technique in conjunction with User Defined Functions
(UDFs) to study the flow fields in SFDs. The results obtained
were compared to the experimental data from Delgado (2005)
but pressures predicted in the cavitation region were slightly
higher than the experimental values. In addition to this, the
Dynamic Meshing technique used was still computationally
expensive and time consuming.
The present study investigates the effect of adding a
central groove to the squeeze film land. Different groove
dimensions were tested at two different speeds of operation and
the pressure. Liquid fraction and forces generated are analyzed
for each condition. The technique used for these 3D simulations
is a steady state solver the feasibility of which was validated in
Boppa et al. (2013).
NOMENCLATURE
c clearance = 0.127 mm
D journal diameter = 127 mm
L film land length = 25.4 mm
Pin oil supply pressure = 31 kPa gage
density = 800 kg/m3
journal eccentricity = 0.075 mm
azimuthal angle,positive counter clockwise, origin at
minimum clearance
oil dynamic viscosity = 0.0016 kg/m sec
METHODOLOGY
A full three dimensional model was initially
constructed based on the geometry of the SFD used by Delgado
(2005). A cross-sectional view of this model is shown in Fig. 2
while a representative mesh of the 3D model is shown in Fig.
3(a). The different dimensions of grooves tested are as shown
in Table 2.
Table 2. Maximum and Minimum Pressures For Various
Groove Sizes for a 50 Hz Orbit.
Groove Size
Groove
Area(m^2) Pmax(kpa) Pmin(kpa)
0.5-0.5 6.35 45 -15
1-0.25 6.35 40 -25
1-0.5 12.7 40 -10
2-0.25 12.7 35 -10
2-0.5 25.4 35 0
4-0.25 25.4 35 0
Without Grove 0 90 -50
The meshes were generated using GAMBIT 2.4.6 with
hexahedral structured elements. A grid independence study
was carried out using a number of two-dimensional models
(this was done considering the high density mesh in three-
dimensional models which made them computationally
exhaustive). Based on the results of the study, a grid size of
1440 X 45(in the azimuthal and radial directions
respectively) is used for the simulations with bi exponential
compression over all the grid in the radial direction to
capture phenomenon close to the wall (boundary layer
phenomenon). For grid independence in the Z (axial)
direction, an adaptive gridding method is used with Y plus as
target. Adaptive gridding is done until a Y plus of 300 is
- 3. 3 Copyright © 2013 by ASME
achieved. This resulted in a total of 10 million nodes for a
three dimensional model (a representative mesh is shown in
Fig. 3). This high density mesh is necessary when using
the steady state solver to ensure accurate solutions for
dynamic pressures in three dimensional squeeze film
dampers.
As mentioned earlier, a steady state solver in an
absolute frame of reference was used to produce whirling
motion of the rotor in this study. The Moving Reference Frame
is a technique used to model and transform an unsteady state
problem in stationary reference frame to a steady state problem
in moving reference frame. In the moving reference frame
model, the equations of motion include additional acceleration
terms which occur due to the transformation from the stationary
to the moving reference frame. For modeling SFDs using the
steady state solver, an absolute velocity formulation is used. In
this formulation Coriolis and centripetal acceleration terms are
collapsed into a single term. For pressure based solvers both
relative velocity and absolute velocity formulations can be used
while for density based solvers only absolute velocity
formulations can be used. The governing equations of fluid
flow with absolute velocity formulation are as follows –
. 0 (1)
. . ̿ (2)
. . ̿. (3)
The working medium used was VG 2 Mobil oil with
a density of 800kg/m
3
a n d a dynamic viscosity of
0.0016kg/m s. An inlet pressure of 31kpa and circular
centered orbiting speeds of 50 and 100Hz were used as
boundary conditions. The mixture model with 3 percent
dissolved air in the lubricant was used along with the Singhal
cavitation model to model multiphase flow with cavitation
while a standard wall function is used to resolve near wall
effects. Pressure-velocity coupling for the continuity equation
was achieved via SIMPLEC. For the pressure discretization,
the PRESTO scheme was chosen. The equations for density,
momentum and kinetic energy were discretized with a second
order upwind scheme while vapor and dissipation rate were
discretized with a first order upwind scheme. A convergence
criterion of 1e-5 was used for all the parameters listed above.
RESULTS
Once grid independence was established, a final grid
size of 1440 X 45 X 100 (in the azimuthal, radial and axial
directions respectively) nodes was found to be optimal. The
solver was set up as mentioned in the previous section with a
pressure of 31Kpa at inlet (shown in the Figure 2). Both the
stator and rotor rotate at 314 rps in the clockwise direction
which is equal to journal’s orbit speed. This motion of both
rotor and stator simulates a centered circular orbit with the fluid
remaining stationary. However, in the stationary reference
frame, the fluid rotates in an anti-clock wise direction with a
speed of 314 rps. The pressure results obtained from the
simulation of this model were compared with the experimental
results obtained by Delgado (2005) at the center of the journal
land with no groove present. These results can be found in
Boppa et al. (2013) and helped validate the model and the
solution methodology used.
Figures 4 to 15 show that when the volumes of the
central grooves are the same, they have the same effect on the
liquid fraction and the pressure developed in the squeeze film
land is the same. A very interesting result is that the maximum
pressure present in the two lands is imposed upon the grooved
area. Hence, instead of the groove area bleeding fluid rapidly
in the azimuthal direction and acting like a low pressure
boundary for the lands, it essentially mimics the pressure
distribution in the land. The grooves do, however, reduce the
maximum and minimum pressures occurring. The grooves also
act to reduce the amount of vapor formed as the area increases
since the minimum pressure is higher. The highest pressure and
lowest pressure developed in the squeeze film land with a
central groove are half of that developed in a smooth SFD, with
pressure values decreasing slightly as the groove size increases
for the range of the dimensions tested. The pressure contours
are also indicative of the area occupied by the groove. From
Table 2 it can be seen that for the groove with smallest cross
sectional area of 6.35 mm2
, the maximum and minimum
pressures vary slightly with the dimensions of the groove. As
the area of cross section increases, the variation of pressure,
with dimensions of the groove, decreases. The maximum
pressure distribution remains relatively constant with increasing
cross sectional area. Pressure distribution varies along the axial
length of the SFD as the cross sectional area decreases.
Minimum pressure shows more axial dependence than
maximum pressure for all cases. The liquid fraction shows a
minimum value of 94% for all the cases with its lowest value
occurring at the groove. Figures 16 and 17 represent pressure
profiles at section 1 and 5 of the squeeze film damper (sections
are as shown in Fig. 3(b)) which are immediately after the inlet
(6.35mm from top) and before the discharge groove (4.1mm
from bottom) respectively. Figures 18 and 19 represent the
pressure and liquid fraction contours at midland section of the
central groove of dimension 0.5 X 0.5 in the moving reference
frame. These figures show that pressure and liquid fraction
have a slight radial dependence at section-1 while the radial
dependence of pressure and liquid fraction is more significant
at the midland due to the presence of the central groove. Figure
20 shows the pressure distribution for the same configuration in
a three dimensional view.
Comparing azimuthal pressure contours for the above
configurations in the central groove (Fig.21) which represent
the largest pressure excursions present, it can be seen that the
flow is moving from low pressure regions to high pressure
regions and thereby being pushed into the central groove.
Therefore, instead of developing pressure in the lands above
and below the central groove, the pressure increases in the
central groove region itself. The extra volume of the central
- 4. 4 Copyright © 2013 by ASME
grooves reduces the magnitude of the pressures developed in
that region and thus the magnitude of forces generated in the
Hence the addition of the central groove becomes an
effective method for reducing the pressures generated in the
squeeze film land. From Figure 21(a), it can also be seen that
when a central groove is added, the lowest and highest
pressures generated at the midland are reduced. Further, there
is a slight change in the position of occurrence of highest and
lowest pressures depending upon the size of the central groove.
But in general, it can be seen that for the range of
configurations simulated, the pressures generated are reduced
to half the values generated when run without any central
groove in it. This plot also confirms that the lines of equal
volumes overlap each other i.e. the amount of reduction in
pressure values depends on the volume of the groove, not the
length and the depth of the groove.
A similar trend is seen when comparing the volume of
fluid (vof) values, as seen in Figure 21(b) – the case without the
central groove shows larger variation of the vof than the cases
with central grooves. As in the previous case, there is little
variation among the cases with the groove with the same
groove volume (this is shown by the complete overlap of the
2X0.5 and 4X0.25 grooves). Hence, for better visibility, one
groove of each volume is plotted. Among the grooves of
different volumes, another subtle trend is seen in that the
groove with the highest volume, shows the least variation of the
vof, indicating that the presence of the groove contributes to
having a more well distributed fluid in the SFD.
Table 1: Comparison of forces generated for different
groove configurations of SFD.
Frequency
(HZ)
Groove
Depth
(mm)
Groove
Length
(inch)/Fx
Cross‐
sectional
Area
(mm^2)/
Fy Resultant
50 1 0.5 12.7
pressure
forces ‐85.33 55.04
Viscous
forces 0.15 ‐0.16
Total Force ‐85.18 55.03 101.41
Angle 147.2104
50 2 0.25 12.7
pressure
forces ‐83.26 57.6
Viscous
forces 0.19 ‐0.14
Total Force ‐83.07 57.46 101.01
Angle 145.4019
50 0.5 0.5 6.35
pressure
forces ‐89.5 85.14
Viscous
forces 0.14 ‐0.22
Total Force ‐89.36 84.92 123.27
Angle 136.5286
50 1 0.25 6.35
pressure
forces ‐98.41 83.9
Viscous
forces 0.18 ‐0.19
Total Force ‐98.23 83.71 129.06
Angle 139.6337
50 2 0.5 25.4
pressure
forces ‐65.9 31.9
Viscous
forces 0.14 ‐0.11
Total Force ‐65.76 31.79 73.04
Angle 154.2779
50 4 0.25 25.4
pressure
forces ‐62.15 39.6
Viscous
forces 0.13 ‐0.11
Total Force ‐62.02 39.49 73.53
Angle 147.5887
Table 3 shows and confirms the predicted results that
equal volumes have equal effect on force generated; grooves of
the same volume generate almost equal magnitude of forces
irrespective of groove dimensions.
To study the effect of variation of orbit speed, central
grooves of the same dimensions were tested at a whirling
frequency of 100 Hz. Figures 22 and 23 present the pressure
and vof data for the SFD operated at 100 Hz, which has central
groove dimensions of 0.5mm x 0.5inch. The pressures
developed and vapor generated in the squeeze film land are
reduced by 30% compared to that generated in the smooth SFD
run at the same speed. The spread of the low pressure and high
pressure region remains same in magnitude and position.
Rigidity of the SFD when run at 100Hz is reduced by adding
the central groove on the squeeze film land and effect of
cavitation is also reduced. The contours are similar to smooth
SFD except the pressures remain relatively constant across the
groove. The sudden collapse of the bubble is still present and is
- 5. 5 Copyright © 2013 by ASME
shown by sudden change in both pressure and vof values. The
area along the axis that is subjected to maximum pressure is
larger for the grooved SFD.
Table 2: Comparison of forces generated for different
configurations of SFD
Frequency
(HZ)
Groove
Depth
(mm)
Groove
Length
(inch)/Fx
Cross sectional
Area(mm^2)/Fy Resultant
50
pressure
forces ‐56.7 172.77
Viscous
forces 0.22 ‐0.22
Total
Force ‐56.48 172.55 181.56
Angle 108.1794
100
pressure
forces ‐75.6 539.01
Viscous
forces 0.69 ‐0.94
Total
Force ‐74.91 538.07 543.26
Angle 97.97543
100 0.5 0.5 6.35
pressure
forces ‐162.19 426.3
Viscous
forces 0.5 ‐0.8
Total
Force ‐161.69 425.5 455.19
Angle 110.863
From Table 4, It can also be seen that the addition of the central
groove, overall, reduces the magnitude of forces generated. At
higher speeds, the magnitude of force generated is also larger,
causing rigidity to some extent. This can be alleviated by
adding a central groove on the squeeze film land.
Figures 24 and 25 show curve fit for variation of
magnitude of force (at whirling rate of 50Hz) and line of action
with cross sectional area. Magnitude of force fits a linear curve
while line of action fits a logarithmic curve.
CONCLUSION
The feasibility of applying steady state solver (MRF) in
Fluent 12.1 to solve for pressure fields inside SFD was studied.
Steady state solver in stationary (absolute) reference frame was
used to generate whirling motion of the rotor. The results from
CFD simulations were compared against experimental results
obtained for the same case by Delgado (2008). Once the model
was validated, central grooves of different dimensions were
added to study their effect on dynamic pressure profiles of
squeeze film dampers. Upon addition of a central groove to the
squeeze film land, the pressures generated were reduced to half
the values of that generated when run without any central
groove on it. It was also observed that the amount of reduction
in pressure values depend on the volume of the groove but not
the length and the depth of the groove. Addition of the central
groove further reduces the rigidity developed in the film land
when run at higher speeds.
REFERENCES
1. Zeidan, F., San Andres, L., and Vance, J., 1996, "Design and
Application of Squeeze Film Dampers in Rotating Machinery,"
Proceedings of the Twenty-Fifth Turbomachinery Symposium,
Turbomachinery Laboratory, Texas A&M University, College
Station, TX, pp. 169-188.
2. Zeidan, F., and Vance, J., 1990a, "Cavitation and Air
Entrainment Effects on the Response of Squeeze Film
Supported Rotors," ASME Journal of Tribilogy, Vol. 112, pp.
347-353.
3. Zeidan, F., and Vance, J., 1990b, "Cavitation Regimes in
Squeeze Film Dampers and Their Effect on the Pressure
Distribution," Tribology Transactions, Vol. 33, No. 3, pp. 447-
453.
4. San Andres, L., 2010, "MEEN 626 Course Notes,"
Mechanical Department,Texas A&M University,
http://rotorlab.tamu.edu/me626/College Station, TX.
5. Arauz, G., San Andreas, L., 1997, “Experimental Force
Response of a Grooved Squeeze Flim Damper”, Tribilogy
International , Vol 30, pp.77-86.
6. Childs, D.W., Graviss, M., Rodriguez, L.E., 2007, “The
Influence of Groove Size on the Static and Rotordynamic
Characterstics of Short, Laminar-Flow AnnularSeals” , ASME
Journal of Tribilogy, Vol. 129, pp. 398-406.
7. Delgado, A., 2008, "A Linear Fluid Inertia Model for
Improved Prediction of Force Coefficients in Grooved Squeeze
Film Dampers and Grooved Oil Seal Rings," Ph.D, dissertation,
Texas A&M University, College Station, TX
8. Guo, Z., Hirano, T., and Kirk, G., 2005, "Application of
CFD Analysis for Rotating Machinery- Part I: Hydrodynamics,
Hydrostatic Bearings and Squeeze Film Damper," Journal of
Engineering for Gas Turbines and Power, Vol. 127, pp. 445-
451.
9. Khandhare,M., 2011, "Numerical Simulation of Flow Field
inside a Squeeze Film Damper and the Study of the Effect of
Cavitation on the Pressure Distribution., " M.S, thesis, Texas
A&M University, College Station, TX.
10. Boppa, P., Morrison, G.L., Sekaran, A., 2013, “A
Numerical Study of Squeeze Film Dampers”, Proceedings of
ASME Turbo Expo 2013, San Antonio, Texas, USA.
11. Delgado, A., 2005, "Identification of Force Coefficients in
a Squeeze Film Damper with a Mechanical Seal," M.S, thesis,
Texas A&M University, College Station,TX.
- 6. 6 Copyright © 2013 by ASME
Figure 1. Squeeze film damper configuration.
Figure 2. Grooved Squeeze Film Damper.
Inlet
Groove
Central
Groove
Section-1
Section-5
Discharge
Groove
- 7. 7 Copyright © 2013 by ASME
Figure 3. Computational Domain, 3D Squeeze Film Damper with Groove
.
- 8. 8 Copyright © 2013 by ASME
Figure 4: Liquid fraction distribution on rotor with
central groove of 0.5 X 0.5 operating at 50Hz
Figure 5: Pressure distribution on rotor with central
groove of 0.5 X 0.5 operating at 50Hz.
Figure 6: Liquid fraction distribution on rotor with
central groove of 1 X 0.25 operating at 50 Hz
Figure 7: Pressure distribution on rotor with central
groove of 1 X 0.25 operating at 50Hz.
Figure 8: Liquid fraction distribution on rotor with
central groove of 1 X 0.5 operating at 50Hz
Figure 9: Pressure distribution on rotor with central
groove of 1 X 0.5 operating at 50Hz
Figure 10: Liquid fraction distribution on rotor with
central groove of 2 X 0.25 operating at 50Hz
Figure 11: Pressure distribution on rotor with central
groove of 2 X 0.25 operating at 50Hz
- 9. 9 Copyright © 2013 by ASME
Figure 12: Liquid Fraction distribution on rotor with
central groove of 2 X 0.5 operating at 50Hz
Figure 13: Pressure distribution on rotor with central
groove of 2 X 0.5 operating at 50Hz
Figure 14: Liquid fraction distribution on rotor with
central groove of 4 X 0.25 operating at 50Hz
Figure 15: Pressure distribution on rotor with central
groove of 4 X 0.25 operating at 50Hz