GAS FOIL BEARINGS FOR OILFREE ROTATING MACHINERY

1
29th Turbomachinery Research Consortium Meeting
May 2009
Measurements of Drag Torque Lift-off Speed and
Identification of Structural Stiffness and
Damping in a Metal Mesh Foil Bearing
TRC-BC-3-09
Luis San Andrés Thomas Abraham Chirathadam
Tae-Ho Kim
Project title Metal Mesh Foil Bearings for
Oil-Free Turbomachinery Test Rig for prototype
demonstrations TRC Funded Project TEES
32513/1519 V2
2
TRC Project Tasks 08/09

• Construction of
• Metal Mesh Foil Bearing (MMFB) Test Rig
• MMFB performance characteristics
• Bearing drag torque
• Lift- Off Speed
• Top Foil Temperature

• Identification of force coefficients (Impact
  load tests) with and w/o shaft rotation
• Structural stiffness and
• equivalent viscous damping

Current research builds upon earlier work on
metal mesh dampers conducted by Prof. John Vance
and students
3
Metal Mesh Foil Bearing (MMFB)
MMFB COMPONENTS Bearing Cartridge Metal mesh
ring and Top Foil Hydrodynamic air film develops
between rotating shaft and top foil.
Potential applications ACMs micro gas turbines
turbo expanders turbo compressors turbo
blowers automotive turbochargers APU

• Large damping (material hysteresis) offered by
  metal mesh
• Tolerant to misalignment and applicable to a
  wide temperature range
• Suitable tribological coatings needed to reduce
  friction at start-up shutdown

Cartridge
4
Metal Mesh Foil Bearings (/-)

• No lubrication (oil-free). NO High or Low
  temperature limits.
• Resilient structure with lots of material
  damping.
• Simple construction ( in comparison to bump-type
  foil bearings)
• Cost effective uses common materials

5
MMFB assembly
Simple construction and assembly procedure
METAL MESH RING
BEARING CARTRIDGE
TOP FOIL
6
MMFB dimensions and specifications
PICTURE
7
MMFB rotordynamic test rig
(a) Static shaft
TC cross-sectional view Ref. Honeywell drawing
448655
Max. operating speed 75 krpm Turbocharger driven
rotor Regulated air supply 9.30bar (120 psig)
Twin ball bearing turbocharger Model T25
donated by Honeywell Turbo Technologies
Test Journal length 55 mm 28 mm diameter
Weight0.22 kg
8
Test Rig Torque Lift-Off measurements
Thermocouple
Force gauge
String to pull bearing
Shaft (F 28 mm)
Static load
MMFB
Top foil fixed end
Torque arm
Positioning (movable) table
Preloading using a rubber band
Eddy current sensor
Calibrated spring
9
Test procedure

• Sacrificial layer of MoS2 applied on top foil
  surface
• Mount MMFB on shaft of TC rig. Apply static
  horizontal load
• High Pressure cold air drives the ball bearing
  supported Turbo Charger. Oil cooled TC casing
• Air inlet gradually opened to raise the turbine
  shaft speed. Valve closing to decelerate rotor to
  rest
• Torque and shaft speed measured during the entire
  experiment. All experiments repeated thrice.

10
Shaft speed and torque vs time
Applied Load 17.8 N
Rotor starts
Rotor stops
WD 3.6 N
Manual speed up to 65 krpm steady state
operation and deceleration to rest
Iift off speed
Startup torque 110 Nmm Shutdown torque 80 Nmm
Once airborne drag torque is 3 of Startup
breakaway torque
Lift off speed at lowest torque airborne
operation
Top shaft speed 65 krpm
11
Varying steady state speed torque
Manual speed up to 65 krpm steady state
operation and deceleration to rest
61 krpm
50 krpm
37 krpm
24 krpm
Drag torque decreases with step wise reduction in
rotating speed until the journal starts rubbing
the bearing
57 N-mm
45 N-mm
2.5 N-mm
2.4 N-mm
2.0 N-mm
1.7 N-mm
Side load 8.9 N
WD 3.6 N
Shaft speed changes every 20 s 65 50 37 -
24 krpm
12
Startup torque vs applied static load
Top foil with worn MoS2 layer shows higher
startup torques
Worn MoS2 layer
Fresh coating of MoS2
Larger difference in startup torques at higher
static loads
Startup Torque Peak torque measured during
startup
Dry sliding operation
13
DRY friction coefficient vs static load
Friction coefficient f (Torque/Radius)/(Static
load)
With increasing operation cycles the MoS2 layer
wears away increasing the contact or
dry-friction coefficient.
Worn MoS2 layer
Enduring coating on top foil required for
efficient MMFB operation!
Fresh MoS2 layer
Dry sliding operation
14
Data derived from bearing torque and rotor speed
vs time data
Bearing drag torque vs rotor speed
Side load increases
WD 3.6 N
Steady state bearing drag torque increases with
static load and rotor speed
4.5
35.6 N (8 lb)
4
Rotor not lifted off
26.7 N (6 lb)
3.5
3
17.8 N (4 lb)
2.5
Bearing torque N-mm
8.9 N (2 lb)
2
1.5
Increasing static load (Ws) to 35.6 N (8 lb)
1
Dead weight (WD 3.6 N)
0.5
0
20
30
40
50
60
70
80
Rotor speed krpm
airborne operation
15
Friction coefficient vs rotor speed
Friction coefficient f (Torque/Radius)/(Static
load)
Friction coefficient f increases with rotor
speed almost linearly
Increasing static load (Ws) to 35.6 N (8 lb)
Dead weight (WD 3.6 N)
f decreases with increasing static load
airborne operation
16
Bearing drag torque vs rotor speed
Max. Uncertainty 0.35 N-mm
Bearing drag torque increases with increasing
rotor speed and increasing applied static loads.
Lift-Off speed increases almost linearly with
static load
17
Lift-Off speed vs applied static load
Side load increases
WD 3.6 N
Lift-Off Speed Rotor speed beyond which drag
torque is significantly small compared to
Startup Torque
Lift-Off Speed increases linearly with static
load
18
Friction coefficient vs rotor speed
Friction coefficient ( f ) decreases with
increasing static load
f 0.01
f rapidly decreases initially and then
gradually raises with increasing rotor speed
Rotor accelerates
Dry sliding
Airborne (hydrodynamic)
19
Top foil temperature (bearing outboard)
Room Temperature 21C
Top foil temperature measured at MMFB outboard end
Side load increases
Top Foil Temperature increases with Static Load
and Rotor Speed
Only small increase in temperature for the
range of applied loads and rotor speeds
20
Test Setup Impact Load Test
IMPACT HAMMER
TC
MMFB
Top foil fixed end
Force gauge
Journal (28 mm)
Flexible string
Eddy current sensor
Accelerometers (Not visible in this view)
(FRONT VIEW)
Positioning table
(SIDE VIEW)
TC
MMFB
Accelerometer
Eddy current sensor
Journal (28 mm)
(TOP VIEW)
21
Identification model
1-DOF mechanical system
Assembly mass M 0.38 kg
Impact along Y direction only
SHAFT STATIONARY NOT ROTATING
22
Impact force
Shaft not rotating
Time domain
Frequency domain averages of 10 impacts along
vertical (Y) direction
Frequency domain
23
Bearing displacement
Shaft not rotating
Time domain
Frequency domain averages of 10 impacts along
vertical (Y) direction
Y
Time s
Rapidly decaying amplitude shows large damping
from MMFB
Frequency domain
24
Bearing acceleration
Journal not rotating
Time domain
Y
TC shaft stub is flexible AY -2 Y
Frequency domain
25
Curve Fit Identifying Static Stiffness and Mass
Estimated test system natural frequency fn
(KY/M)0.5 89 Hz
Journal not rotating
Re (F/Y) Kest Mest Re (A/Y)
Critical damping Ccrit2(KYM)0.5 423 N.s/m
Kest 1.179 105 N/m Mest 0.379 kg
26
Identified MMFB structural stiffness
Journal not rotating
Structural stiffness decreases (10) initially (
50-85 Hz) but increases with further increase in
frequency.
KY Re (F - MestAy)/Y
27
Identified eq.viscous damping
Journal not rotating
Equivalent viscous damping decreases with
increasing frequency

• MMFB shows lots of damping making test system
  just below critically damped
• Ccrit2(KYM)0.5 423 N.s/m

28
Aluminum foam bearings
DONATED by CIATEQ A.C.
Aluminum foam is stiff brittle Not
recommended for use as structural support of foil
bearing




29
Conclusions

• TC driven MMFB rotordynamic test rig to measure
  bearing drag torque bearing displacements and
  acceleration. Operates up to 70 krpm
• Bearing startup torque increases with applied
  static loads. A sacrificial coating of MoS2
  reduces start up torque
• Bearing drag torque while bearing is airborne
  increases with static load and rotor speed
• Top foil steady state temperature increases
  with static load and rotor speed
• Impact tests shows MMFB has large damping its
  stiffness gradually increases with frequency
  except while traversing the bearing natural
  frequency

30
Backups
31
TAMU past work (Metal Mesh Dampers)
METAL MESH DAMPERS provide large amounts of
damping. Inexpensive. Oil-free
Zarzour and Vance (2000) J. Eng. Gas Turb.
Power Vol. 122 Advantages of Metal Mesh Dampers
over SFDs Capable of operating at low and high
temperatures No changes in performance if soaked
in oil
Al-Khateeb and Vance (2001) GT-2001-0247 Test
metal mesh donut and squirrel cage( in
parallel) MM damping not affected by modifying
squirrel cage stiffness
Choudhry and Vance (2005) Proc. GT2005 Develop
design equations empirically based to predict
structural stiffness and viscous damping
coefficient
32
08 TRC MMFB Research at TAMU
San Andres L. Chirathadam T.A. and Kim T.H.
(2009) GT-2009-59315
Static Load Test Setup




Eddy Current sensor


Stationary shaft
Lathe tool holder
Test MMFB
Lathe tool holder moves forward and backward
push and pull forces on MMFB
33
08 TRC MMFB Research at TAMU
MMFB wire density 20
3 Cycles loading unloading

Large hysteresis loop Mechanical energy
dissipation
Displacement -0.060.06 mm Load -130 90 N
34
08 TRC MMFB Research
MMFB wire density 20

During Load reversal jump in structural
stiffness
Max. Stiffness 4 MN/m
35
08 TRC MMFB Research
Dynamic Load Test Setup
Motion amplitude controlled mode
12.7 25.4 38.1 µm
Frequency of excitation 25 400 Hz (25 Hz
interval)
36
08 TRC MMFB Research
Structural stiffness decrease with increasing
motion amplitudes Stiffness increases gradually
with Frequency
37
08 TRC MMFB Research
Eq. Viscous damping decreases with increasing
motion amplitudes Damping decreases rapidly with
frequency