On The Failure Of A Gas Foil Bearing: High Temperature Operation Without Cooling Flow

Transcription

1 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 Flow Keun Ryu Assistant Professor Hanyang University Luis San Andrés Mast-Childs Professor, Fellow ASME Texas A&M University ASME GT This material is based upon work supported by the TAMU Turbomachinery Research Consortium (TRC) and NASA GRC. 1

2 GFB research at TAMU Objective To develop a detailed, physicsbased computational model of gas-lubricated foil journal bearings including thermal effects to predict bearing performance. Rotor Ω Foil spot weld Bump strip layers Top foil Bearing cartridge 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. Since 2003, supported by NSF, NASA, Capstone and TRC 2

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 3

4 TAMU research on foil bearings Year Topic 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 Performance at high temperatures, temperature and rotordynamic measurements. Extend nonlinear rotordynamic analysis Thermoelastohydrodynamic model for prediction of GFB static and dynamic forced performance at high temperatures 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 Rotordynamic measurements: instability vs. forced nonlinearity? Model for ultimate load capacity, Isothermal model for prediction of GFB static and dynamic forced performance Measurement of static load capacity, Identification of structural stiffness and damping coefficients. Ambient and high temperatures 4

5 Objectives & tasks GT 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 5

6 Thermal effects in GFB Systems Gas 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. 6

7 Cooling effectiveness GT Best Paper Award 1.2 Rotor free end 1.2 Free end bearing [ºC/L/min] Temperature rise Cooling flow rate No rotor spinning 10 krpm 20 krpm 30 krpm Tr FE -T amb [ºC/L/min] Temperature rise Cooling flow rate No rotor spinning 10 krpm 20 krpm 30 krpm T 1~4 -T amb Cooling flow rate [L/min] Cooling flow rate [L/min] 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. Cooling flow rate increases Heater set temperature = 100ºC T1~T4 Te T6~T9 Th Enclosure Tr FE Bearing housing Tr DE 7

8 No cooling flow? The current experimental results show poor operating procedure and ignorance on the predicted system behavior. 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. The following details the test procedure, temperature measurements, and discusses the possible causes leading to the failure. 8

9 The test rig 9

10 Hot rotor-fb test rig Instrumentation for high temperature. Drive motor (max. 50 krpm) Fiberoptic displacement sensors Infrared thermometers Cooling air supply for coupling Infrared tachometer Cartridge heater Drive motor Heater reference temperature at 600 C Hollow rotor Test GFBs GFB support housing Flexible coupling Significant temperature gradient along rotor axis Heat source warms (unevenly) rotor and its bearings Reference thermocouple to control heater temperature 10

11 Hot rotor-fb test rig Dimensions 254 Free end GFB T e Drive end GFB Cartridge Heater (1.6 kw) Hollow rotor T h Enclosure FB support housing Front view Hollow rotor (AISI 4140): 1.31 kg Length: mm, OD mm and ID mm. NO coating on rotor surface 11

12 Hot rotor-fb test rig Instrumentation Thermocouples (K-type) 208 V AC ~ Infrared thermometer Infrared thermometer Heater temperature controller Cartridge heater Drive motor Motor controller Displacement sensors Foil bearings ~ 460 V AC Temperature digital panel meter Signal conditioner Oscilloscopes FFT Analyzer Data acquisition board PC Tachometer Thermocouples: 1 x heater, 1 x Bearing housing enclosure 2 x 4 FB outboard, 2 x Bearing housing 2x Drive motor, 1 x ambient + infrared thermometers 2 x rotor surface 12 (Total = 17)

13 The test bearings 13

14 Test foil bearings Cartridge sheet FB nominal dimensions Top foil Bumps Parameter [Dimension] Drive end Free end Cartridge inner diameter [mm] Cartridge outer diameter [mm] Axial bearing length [mm] Number of bumps Generation II FB Three (axial) bump strip layers, each with 24 bumps. Patented solid lubricant coating (up to 800 F) on top foil surface. Other data proprietary Bearing radial (assembly) clearance [mm]

15 Thermocouples in test FB uncoated (Gen II) Oblique view Free end T4 Thermocouple fixed at midspan bearing Drive end T9 T1 Ω T3 Th Free End (FE) GFBs Drive End (DE) T8 Ω T6 T2 T7 Tr FE T1~T4 Tr DE T6~T9 Four (4) thermocouples placed in machined axial slots. 15

16 Test Procedure 16

17 Test procedure Heater off Ths=200ºC Ths=400ºC Ths=600ºC Heater (Th) Heater reference temperature at 600 C Heater (Th) Temperature [ºC] RBS failure Reference thermocouple to control heater temperature Time [min] Rotor spins at 37 krpm Th is set at 200 C, 400 C, 600 C (~60 minute intervals) Failure at ~190 minute of elapsed test time 17

18 Temperature measurements Th T1~T4 Te Enclosure T6~T9 Tr FE Bearing housing Tr DE 18

19 Rotor OD temperature Temperature [ºC] Heater off Ths=200ºC Ths=400ºC Ths=600ºC FE rotor (TrFE) DE rotor (TrDE) Heater (Th) Heater (Th) Free end rotor OD (Tr FE) Drive end rotor OD (Tr DE) Time [min] Max. 251ºC RBS failure Tr FE >> Tr DE Significant axial thermal gradient from the rotor free end towards its drive end 19

20 Bearing OD temperature Heater off Ths=200ºC Ths=400ºC Ths=600ºC T4 T1 T2 Max. 114ºC Temperature [ºC] T3 T1 RBS failure 20 0 T2 T3 T Time [min] 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. 20

21 Steady state temperatures Temperature [ C] Tr FE * Steady state temperature (~60 minute heating) TrFE T4T T7TrDE 4 T Td 9 T e Tr DE Tr FE TrFE Transient temperature (~10 minute heating) T4 4 Td e T9 T7 TrDE Tr DE TrFE 81 ºC T4 Td T T e 4 TrDE T7 9 Tr DE TrFE FE T4 Td T 4 T e T 9 T7 127 ºC Tr DE TrDE Large thermal gradient in the rotor 0 * * * Heater 1 off Th=200 C 2 Th=400 C 3 Th=600 C 4 Tr FE T 4 T ed T 9 Tr DE Heater set temperature, Th [ C] T e and Tr DE rise above the temperatures in the bearing sleeves (T 4, T 9 ) as the heater temperature increases None of the temperature measurements gives any evidence of an impending failure! 21

22 Rotor motion measurements FH GFBs DH g FV DV 22

23 Waterfall of rotor response 37 krpm No cooling flow into bearings T h =600ºC (10 min.) T h =400ºC (63 min.) 0.25X 0.5X 1X FH FV Time increases Free end (Vertical) GFBs DV DH FE Test bearing failure occurs FE rotor OD temp: ~250ºC g T h =200ºC (56 min) Heater off (58 min.) Rotor motion components at its free end show no distinctive differences as the rotor temperature increases. 23

24 Synchronous rotor response 37 krpm No cooling flow into bearings 350 Heater off Th=200ºC Th=400ºC 300 Temperature [ºC] FH FV GFBs DV DH g Time [min] Free end rotor OD (Tr FE) Test bearing failure occurs Rotor temperature does not affect the size and orientation of the FE rotor orbits. Th=600ºC FE bearing temperatures difference is largest along horizontal plane Produce uneven thermal growth leading to bearing ellipcity! 24

25 Post-test conditions 25

26 Rotor OD surface conditions Before operation NO protective coatings on Rotor surface Smears of top foil protective coating AFTER incident Drive end bearing location Free end bearing location HT on FE rotor OD renders a much darker color! Rotor surface where the FE bearing is held shows considerable wear and noteworthy scratches. 26

27 Drive end foil bearing after incident Inboard Wear marks Inboard Wear marks Outboard Bearing still functional! Outboard Minor wear marks on the inboard top foil 27

28 Free end foil bearing after incident Melted top foil Inboard Outboard Galled coating Melted top foil Inboard Outboard Melting temperature of the foil material is 1,430ºC! Ripple traces Beyond repair! Protective coating evaporated, and distinct sections of the top foil melted, in particular at the location 28 where the top foil contacts the bump foil crests

29 Close up view of FE bearing Most metal melted right at the line contacts with the crest of the bump foils Bearing seizure! Pressure Flexible top foil Bump Sag Large amount of energy generated, producing an increasing temperature, first melted the protective coating (max. 400 C) and later the top foil. 29

30 Thermal expansion of test rig components 30

31 Predicted thermal expansion FE Rotor (AISI4140) THERMAL EXPANSION FE Bearing sleeve (Inconel 718) THERMAL CONTRACTION Heater off 100 Th= 200º C Th= 400º C Th= 600º C Heater off 10 Th= 200º C Th= 400º C Th= 600º C Thermal expansion [um] Thermal expansion [um] Rotor Rotor temperature [C] [ºC] Bearing Bearing cartridge temperature temperature [C] [ºC] Thermal expansion of the rotor OD is more 31 pronounced since it is hotter than the bearing sleeve

32 Estimated bearing clearance Free end FB Bearing radial clearance [µm] c=42 µm at cold condition (21 C of bearing & rotor temperature) Th=200ºC Th=400ºC Th=600ºC FE bearing clearance is nil or disappears (turning negative)! C=C r -S T -S C -B T C r: radial bearing clearance (ambient) S T : rotor radial thermal expansion S C : rotor centrifugal growth B T : bearing sleeve thermal expansion FE rotor (underneath bearing) temperature [ºC] Heater temperature (Th) [ºC] FE bearing (mean) temperature [ºC] FE bearing eventually loses its clearance, followed 32 by sudden bearing seizure and system failure!

33 Damage on inboard side of bearing? Tr FE T 4 T ed T 9 Tr DE Melted top foil Inboard Outboard The enclosure was closed, not even permitting the natural convection of hot air into the environment. Inboard side of the bearing sleeve shrunk more than outboard side or at the mid-span. Trapped air in the enclosure became a thermal sink causing FE bearing inboard temperature to be 33 higher than that on outboard side of the bearing!

34 Temperature predictions versus test data 34

35 TAMU GFB TEHD model Gas film Reynolds eqn. for hydrodynamic pressure generation Energy transport eqn. for mean flow temperature Various surface heat convection models Mixing 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 growth Bearing thermal growth San Andrés and Kim (2008) Software licensed by TAMU 35

36 Recap: test rotor and FB Bearing housing Outer flow stream x z Bearing housing Bump layer Top foil P Co, T Co Cooling stream Thin film flow Ω R So P a Hot air (out) T ambient Hollow rotor Heat source T h z=0 z=l Natural convection on exposed surfaces of bearing OD and shaft ID View of rotor, heater cartridge + side cooling stream 36

37 Bearing temperature predictions & tests Thermal mixing coefficient λ= krpm Bearing temperature [ºC] FE bearing-test DE bearing-test FE bearing-predict DE bearing-predict DE - PREDICT FE - PREDICT FE -TEST DE -TEST Θ 45º Load Discrepancies are due to large temperature gradient along heater axial length X Ω Y Rotor temperature [ºC] Predictions (DE bearing) agree well with test data!! 37

38 Estimation of the temperature raise in the top foil 38

39 Estimation of top foil temperature = µw ( RSOΩ) friction dt ( ) Foil C M Ah ( T T ) p Foil Foil friction dt + = Foil specific heat and mass Contact area Heat convection coefficient (air) Top foil temp. increase friction Generated Mechanical power due to rubbing = µw ( R Ω) SO 130 W (with 6 N and 37 krpm) t Pfriction t TFoil() t To e τ T 1 e τ = + + Ahair Solution is valid for T Foil < 1,430 o C (T melt ) τ= ( C ) PM Foil Ah Dry 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. 39 Hence, the temperature increased very rapidly.

40 Predicted top foil temperature Top foil temperature rise (C) Rubbing begins T melt Top foil coating temperature limit time (s) * Rapid growth of the foil temperature to reach its melting point in ~14 s. 40

41 Conclusions ASME GT 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, dryfriction 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, 2 nd bearing and instrumentation) remain functional, showing little damage. Predictive tool validated & benchmarked to reliable test 41 data base.

42 Recommendations ASME GT 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 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. 42

43 Acknowledgments Thanks to the support of NASA GRC ( ) TRC ( ) Questions (?) Learn more 43