11

ASME GT2013-94244

Luis San AndrésMast-Childs Professor, Fellow ASME

Texas A&M University

This material is based upon work supported by the TAMU Turbomachinery Research Consortium (TRC) and NASA GRC.

2013 ASME Turbo Expo Conference, June 3-7 , 2013, San Antonio, TX

On The Failure Of A Gas Foil Bearing: High Temperature

Operation Without Cooling FlowKeun Ryu

Assistant Professor

Hanyang University

2

To develop a detailed, physics-based computational model of gas-lubricated foil journal bearings including thermal effects to predict bearing performance.

GFB research at TAMU

Since 2003, supported by NSF, NASA, Capstone and TRC

Bearing cartridge

Rotor

Ω

Foil spot weld

Top foil

Bump strip layers

In support of oil-free systems reduce overall system weight, complexity (low footprint) increase system efficiency due to low drag power losses extend maintenance intervals.

Objective

3

Gas foil bearings (+/-)• Increase reliability & good load capacity (< 20 psi)• Dispense with oil lubrication• Reduced weight & number of parts• High and low temperatures • Tolerate high vibration and shock load.

• Load capacity: less than rolling or oil lubricated bearings

• Endurance: wear during start up & shut down (lift off speed)

• Thermal management for high temperature applications (gas turbines, turbochargers)

• Predictive models lack validation for GFB operation at HIGH TEMPERATURE

4

Year Topic2008-13 Metal Mesh Foil Bearings: construction, verification of lift off performance and load

capacity, identification of structural stiffness and damping coefficients, identification of rotordynamic force coefficients

2008-10 Performance at high temperatures, temperature and rotordynamic measurements. Extend nonlinear rotordynamic analysis

2007-09 Thermoelastohydrodynamic model for prediction of GFB static and dynamic forced performance at high temperatures

2005-07 Integration of Finite Element structure model for prediction of GFB static and dynamic forced performance Effect of feed pressure and preload (shims) on stability of FBS. Measurements of rotordynamic response.

2005-07 Rotordynamic measurements: instability vs. forced nonlinearity?

2005-06 Model for ultimate load capacity, Isothermal model for prediction of GFB static and dynamic forced performance

2004-09 Measurement of static load capacity, Identification of structural stiffness and damping coefficients. Ambient and high temperatures

TAMU research on foil bearings

5

Objectives & tasks

Temperature and rotordynamic measurements of a heated rotor supported on gas foil bearings without cooling flow

• Measure temperature of bearings and rotor and the motions of rotor for increasing shaft temperatures• Identify post-test condition of test rotor and bearings after sudden failure• Estimate bearing clearance and top foil temperature to until bearing seizure occurs• Compare the experimental results to predictions from an in-house computational program

GT2013-94244

6

Thermal effects in GFB SystemsGas bearings (when airborne) are nearly friction

free, with small drag power and temperature rise.

However, the material structure of a foil bearing tends to heat up quickly since the lubricant, a gas, has low density and low thermal conductivity.

Rises in temperature change material properties (solids and gas) and the bearing clearance…..

Applications demand large cooling flows to control thermal growth of components and to remove efficiently mechanical energy from the rotor mainly.

7

Cooling effectiveness

Heater set temperature = 100ºC

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500Cooling flow rate [L/min]

No rotor spinning10 krpm20 krpm30 krpm

[ºC/L

/min

]Te

mpe

ratu

re ri

se C

oolin

g flo

w ra

te

T 1~4 -T amb

Cooling flow rate increases

TrFE

Te

TrDE

EnclosureTh

Bearing housing

T1~T4 T6~T9

Free end bearing

0

0.2

0.4

0.6

0.8

1

1.2

0 100 200 300 400 500Cooling flow rate [L/min]

No rotor spinning10 krpm20 krpm30 krpm

[ºC/L

/min

]Te

mpe

ratu

re ri

se C

oolin

g flo

w ra

te

Tr FE -T amb

Rotor free end

GT2012-68074Best Paper Award

Cooling flow rate increases

-Even a little flow can cool the bearings & rotor. Too large cooling flows stop being effective. -The cooling effect of the supplied flows is most distinct when the rotor is hottest and at the highest rotor speed.

8

No cooling flow?

The current experimental results show poor operating procedure and ignorance

on the predicted system behavior.

The following details the test procedure, temperature measurements, and

discusses the possible causes leading to the failure.

Inadequate thermal management caused a sudden system failure (bearing seizure) while operating at a rotor speed of

37 krpm with the shaft OD temperature at ~250 °C.

9

The test rig

10

Hot rotor-FB test rig

Drive motor (max. 50 krpm) Instrumentation for high temperature.

Significant temperature gradient along rotor axisHeat source warms (unevenly) rotor and its bearings

Drive motor

Infrared thermometersCooling air supply for coupling

Infrared tachometer

Flexible couplingGFB support housing

Cartridge heater

Fiberoptic displacement sensors

Hollow rotor

Heater reference temperature at 600°C

Reference thermocouple to control heater temperature

Test GFBs

11

Hot rotor-FB test rig Dimensions

44.4

5

62.23

19.69

153.

04

144.78

15.9

0

17.9

0

254

50

180

200.66

Hollow rotorCartridge Heater (1.6 kW)

FB support housing

Free endGFB

Drive endGFB

Enclosure

5

Front view

Te

Th

Hollow rotor (AISI 4140): 1.31 kgLength: 200.7 mm, OD 36.46 mm and ID 17.90 mm. NO coating on rotor surface

12

Hot rotor-FB test rig Instrumentation

Foil bearings

Cartridge heater Drive motor

Heater temperature

controller

~208 V AC

Temperature digital panel meter

Thermocouples (K-type)

Displacement sensors

FFT Analyzer

Oscilloscopes

Signal conditionerTachometer

Data acquisition board

PC

Infrared thermometer

Infrared thermometer

~460 V AC

Motor controller

Thermocouples: 1 x heater, 1 x Bearing housing enclosure

2 x 4 FB outboard, 2 x Bearing housing2x Drive motor, 1 x ambient

+ infrared thermometers 2 x rotor surface (Total = 17)

13

The test bearings

14

Test foil bearings

Top foil

Cartridge sheet

Bumps

FB nominal dimensions

Parameter [Dimension] Drive end Free endCartridge inner diameter [mm] 37.98 37.92Cartridge outer diameter [mm] 44.64 44.58Axial bearing length [mm] 25.40 25.40Number of bumps 24× 3 24× 3

Bearing radial (assembly) clearance [mm] 0.076 0.042

Generation II FBThree (axial) bump strip layers, each with 24 bumps.Patented solid lubricant coating (up to 800°F) on top foil surface.

Other data proprietary

15

Oblique view

Thermocouples in test FB

Four (4) thermocouples placed in machined axial slots.

uncoated (Gen II)

TrFET1~T4

Th GFBs

Free End (FE)

Drive End(DE)

TrDET6~T9

Thermocouple fixed at mid-span bearingFree end

T1

T2

T3

T4

Ω

Drive end

ΩT8

T7

T6

T9

16

Test Procedure

17

Test procedure

Rotor spins at 37 krpm

Th is set at 200°C, 400°C, 600°C(~60 minute intervals)

050

100150200250300350400450500550600650

0 20 40 60 80 100 120 140 160 180 200

Tem

pera

ture

[ºC

]

Time [min]

Heater (Th)

Heater(Th)

Heater off Ths=200ºC Ths=400ºC Ths=600ºC

Failure at ~190 minute of elapsed

test time

RBS failure

Heater reference temperature at 600°C

Reference thermocouple to control heater temperature

18

Temperature measurements

TrFE

Te

TrDE

EnclosureTh

Bearing housing

T1~T4 T6~T9

19

Rotor OD temperature

050

100150200250300350400450500550600650

0 20 40 60 80 100 120 140 160 180 200

Tem

pera

ture

[ºC

]

Time [min]

FE rotor (TrFE)DE rotor (TrDE)Heater (Th)

Heater off Ths=200ºC Ths=400ºC Ths=600ºC

Significant axial thermal gradient from the rotor free end towards its drive end

TrFE >> TrDE

Max. 251ºC

Free end rotor OD(Tr FE)

Drive end rotor OD (Tr DE)

Heater(Th)

RBS failure

20

Bearing OD temperature

0

20

40

60

80

100

120

0 20 40 60 80 100 120 140 160 180 200

Tem

pera

ture

[ºC

]

Time [min]

T1

T2

T3

T4

Heater off Ths=200ºC Ths=400ºC Ths=600ºC

FB sleeve OD temperatures are different depending on their circumferential location; a result related to the heater warming the shaft and bearings unevenly, even without rotor spinning.

Max. 114ºCT4

T2T1

T3 RBS failure

21

Steady state temperatures

TrFE

TrFE

TrFE

TrFE

T4

T4

T4

T4

Td

Td

Td

Td

T7T7

T7T7

TrDETrDE

TrDE

TrDE

0

50

100

150

200

250

300

1 2 3 4Heater set temperature, Th [°C]

Tem

pera

ture

[°C

]

Th=200°C Th=400°C Th=600°CHeater off

TrFE TrDET4 T9Te

TrFE

TrDET4 Te T9

TrFE

TrDET4Te

T9

TrFE

TrDETe

T9

* * *

* Steady state temperature (~60 minute heating)

T4

Transient temperature (~10 minute heating)

TrFE TrDET4 T9Td

Te

Large thermal

gradient in the rotor

81 ºC

127 ºC

Te and TrDE rise above the temperatures in the bearing sleeves (T4, T9) as the heater temperature increases

None of the temperature measurements gives any evidence of an impending failure!

22

Rotor motion measurements

FV

GFBs

DV

FH DH

g

23

Rotor motion components at its free end show no distinctive differences as the rotor temperature increases.

Waterfall of rotor response37 krpm

Free end (Vertical)

FV

GFBs

DV

FH DH

g

Time increases

Th=200ºC (56 min)

Th=400ºC (63 min.)

Th=600ºC (10 min.)

Heater off (58 min.)

1X 0.5X 0.25X

FE Test bearing failure occurs

FE rotor OD temp: ~250ºC

No cooling flow into bearings

24

Synchronous rotor response

Rotor temperature does not affect the size and orientation of the FE rotor orbits.

0

50

100

150

200

250

300

350

0 20 40 60 80 100 120 140 160 180 200Time [min]

Tem

pera

ture

[ºC

]

Test bearing failure occurs

Free end rotor OD(Tr FE)

Heater offTh=200ºC Th=400ºC

Th=600ºC

FV

GFBs

DV

FH DH

g

37 krpm No cooling flow into bearings

FE bearing temperatures difference is largest

along horizontal plane Produce uneven

thermal growth leading to bearing ellipcity!

25

Post-test conditions

26

Rotor OD surface conditions

HT on FE rotor OD renders a much darker color!

Rotor surface where the FE bearing is held shows considerable wear and noteworthy scratches.

Drive end bearing location Free end bearing location

Before operation

AFTER incident

NO protective coatings on Rotor surface

Smears of top foil protective

coating

27

Drive end foil bearing after incident

Minor wear marks on the inboard top foil

Wear marks

Wear marks

Bearing still functional!Outboard

Inboard

Outboard

Inboard

28

Free end foil bearing after incident

Protective coating evaporated, and distinct sections of the top foil melted, in particular at the location where the top foil contacts the bump foil crests

Galled coating

Ripple traces

Melted top foil

Melted top foil

Beyond repair!Melting temperature of the foil material is 1,430ºC!

Outboard

Inboard

Outboard

Inboard

29

Close up view of FE bearingMost metal melted right at the line contacts with the crest of the bump foils Bearing seizure!

Bump

Flexible top foil

Pressure Pressure

“Sag”

Large amount of energy generated, producing an increasing temperature, first melted the protective coating (max. 400°C) and later the top foil.

30

Thermal expansion of test rig components

31

Predicted thermal expansion

Thermal expansion of the rotor OD is more pronounced since it is hotter than the bearing sleeve

FE Rotor (AISI4140) THERMAL EXPANSION

20 50 80 110 140 170 200 230 2600

20

40

60

80

100

Rotor temperature [C]

Ther

mal

exp

ansi

on [u

m]

Heateroff

Th=200º C

Th=400º C

Th=600º C

Rotor temperature [ºC]

20 40 60 80 100 1200

2

4

6

8

10

Bearing temperature [C]

Ther

mal

exp

ansi

on [u

m]

Heateroff

Th=200º C

Th=400º C

Th=600º C

Bearing cartridge temperature [ºC]

FE Bearing sleeve (Inconel 718) THERMAL CONTRACTION

32

Estimated bearing clearance

Free end FB

-20

-10

0

10

20

30

40

50

20 120 220 320 420 520 620Bea

ring

radi

al c

lear

ance

[µm

]

Heater temperature (Th) [ºC]

35 59 85 111 138 164

28 42 56 70 84 98

FE rotor (underneath bearing) temperature [ºC]

FE bearing (mean) temperature [ºC]

Th=400ºC Th=600ºCTh=200ºC

c=42 µm at cold condition(21°C of bearing & rotor temperature)

FE bearing eventually loses its clearance, followed by sudden bearing seizure and system failure!

C=Cr-ST-SC-BTCr: radial bearing clearance (ambient)ST : rotor radial thermal expansionSC : rotor centrifugal growthBT : bearing sleeve thermal expansion

FE bearing clearance is nil or disappears (turning negative)!

33

Damage on inboard side of bearing?

Trapped air in the enclosure became a thermal sink causing FE bearing inboard temperature to be higher than that on outboard side of the bearing!

Melted top foil

Outboard

Inboard

TrFE TrDET4 T9Td

Te

Inboard side of the bearing sleeve shrunk more than outboard side or at the mid-span.

The enclosure was closed, not even permitting the natural

convection of hot air into the environment.

34

Temperature predictions versus test data

35

TAMU GFB TEHD model

San Andrés and Kim (2008)

Gas filmReynolds eqn. for hydrodynamic pressure generation Energy transport eqn. for mean flow temperatureVarious surface heat convection modelsMixing of temperature at leading edge of top foil

Top foil & underspring

Thermo-elastic deformation eqns.Finite Elements and discrete parameter for bump strips.Thermal energy conduction paths to side cooling flow and bearing housing.

Bearing clearance

Material properties (gas & foils) = f (Temperature)Shaft thermal and centrifugal growthBearing thermal growth

Software licensed by TAMU

36

Recap: test rotor and FB

Outer flow stream Top foil

Bearing housing “Bump” layerzx

PCo , TCo

Bearing housing

PaThin film flow

z=0 z=LT∞

Ω RSo

Heatsource

Th

Coolingstream

Hollow rotor

Hot air (out)

ambient

View of rotor , heater cartridge + side cooling stream

Natural convection on

exposed surfaces of

bearing OD and shaft ID

37

predictions & testsBearing temperature

20

40

60

80

100

120

20 40 60 80 100 120 140

Bea

ring

tem

pera

ture

[ºC

]

Rotor temperature [ºC]

FE bearing-TEST

DE bearing-TEST

FE bearing-PREDICT

DE bearing-PREDICT

Predictions (DE bearing) agree well with test data!!

X

Y

Ω

45º

Θ

Load

Discrepancies are due to large temperature

gradient along heater axial length FE -TEST

DE -TEST

FE - PREDICT

DE - PREDICT

Thermal mixing coefficient λ=0.65 37 krpm

38

Estimation of the temperature raise in the top foil

39Dry friction contact generated large amounts of energy that could not be convected quickly to the bearing sleeve or removed by the gas film since there was no forced cooling flow. Hence, the temperature increased very rapidly.

Estimation of top foil temperature)( Ω=℘ SOfriction RWµ

)( Ω=℘ SOfriction RWµ

Generated Mechanical power due to rubbing

( ) ( )Foilp Foil frictionFoild T

C M Ah T Tdt ∞ ∞

+ − =℘

Top foil temp. increase

Foil specific heat and mass

Contact area

Heat convection coefficient (air)

130 W (with 6 N and 37 krpm)

( ) 1

t tfrictionFoil t o

air

PT T e T e

Ahτ τ− −

∞ = + + −

τ= ( )P FoilC M

Ah∞

Solution is valid for TFoil < 1,430 oC (Tmelt)

40

Rapid growth of the foil temperature to reach its melting point in ~14 s.

Predicted top foil temperature

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 180

300

600

900

1200

1500

1800

time (s )

Top

foil

tem

pera

ture

rise

(C)

Tmelt

*

Top foil coating temperature limit

Rubbing begins

41

Unusual operating condition, w/o cooling flow and too large increment in rotor temperature (up to 250°C) led to the incident which destroyed one of the test bearings. Bearing clearance decreases notably as rotor temperature increases until seizure occurs. Upon contact between the rotor and top foil, dry-friction quickly generated vast amounts of energy that melted the protective coating and metal top foil. In spite of the loss of one foil bearing, the other system components (rotor, 2nd bearing and instrumentation) remain functional, showing little damage.

Conclusions ASME GT2013-94244

Predictive tool validated & benchmarked to reliable test data base.

42

Even small quantities of air could be effective to promote the evacuation of hot air. The qualitative assessment for GFB system thermal management requires considerable experience! A cautious predictive design and conservative operation of the GFB system, along with an adequate thermal control strategy, could have prevented the bearing seizure and the rotor bearing system failure

Recommendations ASME GT2013-94244

Both (a) engineering thermal management with adequate cooing scheme and (b) adequate assessment of thermal effects prior to actual operation are mandatory for high temperature operation.

43

AcknowledgmentsThanks to the support of• NASA GRC (2007-09)• TRC (2004-10)

Learn more http://rotorlab.tamu.edu

Questions (?)

On The Failure Of A Gas Foil Bearing: High Temperature Operation Without Cooling FlowSlide Number 2Gas foil bearings (+/-)Slide Number 4Objectives & tasksThermal effects in GFB SystemsSlide Number 7No cooling flow?The test rigHot rotor-FB test rigHot rotor-FB test rigHot rotor-FB test rigSlide Number 13Test foil bearingsSlide Number 15Test ProcedureSlide Number 17Temperature measurementsSlide Number 19Slide Number 20Slide Number 21Rotor motion measurementsSlide Number 23Synchronous rotor responsePost-test conditionsRotor OD surface conditionsDrive end foil bearing after incidentFree end foil bearing after incidentClose up view of FE bearingThermal expansion of test rig componentsPredicted thermal expansionSlide Number 32Damage on inboard side of bearing? Temperature predictions versus test dataTAMU GFB TEHD modelSlide Number 36Slide Number 37Estimation of the temperature raise in the top foilSlide Number 39Slide Number 40Slide Number 41Slide Number 42Slide Number 43