Parameters Affecting PSJB Performance.pdf

7/29/2019 Parameters Affecting PSJB Performance.pdf

1/14

A STUDY OF PARAMETERS THAT AFFECT PIVOTED SHOEJOURNAL BEARING PERFORMANCE IN HIGH-SPEED TURBOMACHINERY

byScan DeCamillo

Manager, Research and DevelopmentKingsbury, Inc.

Philadelphia, Pennsylvaniaand

Keith BrockwellSenior Research Officer

National Research Council of CanadaInstitute for Aerospace Research

Ottawa, Ontario, Cauada

Scan DeCamillo is Manager ofResearchand Development for Kingsbury, Inc., inPhiladelphia, Pennsylvania. He is respon-sible for design, testing, analysis, anddevelopment of Kingsbury fluid filmbearings for worldwide industrial andmilitary applications. He began work inthis field in 1975 and has since providedengineering support to industry regardingproblem solving, design improvement,application, and operation of hydrodynam-ic bearings in various machinery.During the course of his work, Mr. DeCamillo has developedpelformance and structural tools for bearing analysis, establisheddesign cri ter iafor power, naval, and nuclear applications, and hascontributed information to ASM, AISI, EPRI, NASA, and API forvarious publications and specifications. He is an inventor and

author of several technical papers.Mr. DeCamillo received his B.S. degree (Mechanical Engineer-ing) from Drexel University. He is a registered ProfessionalEngineer in the State of Pennsylvania and a member of theVibration Institute, ASM, ASME, and STLE.

Keith Brockwell is a Senior ResearchOfficer with the National Research Councilof Canada's Institute for AerospaceResearch, in Ottawa, Ontario, Canada. Hismajor research interests are tribology, fluidfilm lubrication, and bearing technology.He has authored and coauthored close to50 technical papers and reports dealingwith these topics.

Mr. Brockwell received his postgraduatetraining from Imperial College andNewcastle Polytechnic, England. Previous UK appointmentsincluded the Glacier Metal Company and Michell Bearings wherehe was Manager of the Design Services Department.He is a Chartered Engineer (UK), a Member of he Institution ofMechanical Engineers, and a Fellow of the Society of Tribologistsand Lubrication Engineers. Mr. Brockwell is a member of theSTLEIASME Tribology Conference Planning Committee, and wasformerly an Adjunct Professor at the University of BritishColumbia, where his teaching interest was tribology.

9

ABSTRACTJournal bearing pad temperatures, oil flow requirements, and

power losses can impose limitations on the design and operationof high-speed turbomachinery. Over the past few years, theauthors have conducted extensive tests and studies of parametersthat affect pivoted shoe journal bearing performance. A special,high-speed test rig is described, which was designed and built forthe purpose of measuring bearing steady-state performance characteristics under light to moderately heavy loads, and to very highoperating speeds that are being approached in new turbine andcompressor designs. Instrumentation includes a detailed array ofpad temperature detectors and the direct measurement of frictional torque.Data are presented that compare the effects of pivot offset, oilflow, load orientation, method of lubrication, and oil dischargeconfiguration on 6 inch diameter pivoted shoe journal bearing performance. The parameters are shown to significantly influencebearing pad temperature and power loss, particularly at high loadsand speeds. Pad temperature profiles, isotherms, torque, and oiloutlet temperatures are compared and evaluated. Discussionsaddress the prediction and application of these parameters, andhow they may be used to improve the capacity and performance ofhigh-speed turbomachinery. The data and discussions are intendedto provide useful information to engineers, programmers, andpersonnel involved with the study or operation of pivoted shoejournal bearings.INTRODUCTION

Turbomachinery operating speeds and loads have increased overtime resulting in increased bearing temperatures and power losses.Gardner and Ulschmid (1973) addressed the concern of journalbearing limitations at turbulent operation, documenting a dramaticincrease in pad temperature and power loss of a 17 inch pivotedshoe bearing at 3600 rpm (267 fps surface speed). The pad was acenter pivot design, which has the advantage of being able tooperate in either direction of rotation. There are many technicalpapers that study center pivot journal bearings. Offset pivotsimprove journal bearing pad temperature limitations, but are not aswell documented in literature.Large steam and gas turbine designers are presently considering 22 inch and larger diameter journal bearings for powergeneration where surface speeds at 3600 rpm exceed 330 fps.Designers of steam turbines and compressors have intentions foroperating speeds approaching 400 fps. Such conditions are wellinto the turbulent regime where conventional bearing losses and


2/14

Test Rig 6 inch Test Shaft Speed to 16000 rpm Loads to 5500 Ibs

~ c o . e l1li__ _a_a _BII_--.Jo 'Figure Test Rig SchemCltic..

10 PROCEEDINGS OF THE 30TH TURBOMACHINERY SYMPOSIUM

temperatures become so high that options must be considered toaddress limitations. Direct lubrication is one solution that hasbeen successfully applied in thrust bearings for many years, andalso in special journal bearing applications dating back to themid-sixties. It is only recently that direct lube journal bearingshave been seriously considered for general turbomachinery applications because high surface speeds now warrant such aconsideration.As in the case of offset pivots, technical papers on direct lubejournal bearings are sparse. Data are published by Tanaka (1991)and Tanaka and Mishima (1989) comparing 100 mm (3.94 inch)diameter designs to 137 fps; Harangozo, et a1. (1991), tested 5 inchdiameter bearings to 152 fps; and Fillon, et a1. (1993), report on100 mm (3.94 inch) diameter data to 70 fps. The authors'(Brockwell, et aI., 1992, 1994; Dmochowski, et aI., 1993) tested3.88 inch diameter, leading-edge-groove (LEG) designs to 270 fps;DeCamillo and Clayton (1997) provide data on an 18 inch LEGgenerator bearing at 283 fps; and Edney, et a1. (1996), report on 5inch diameter LEG steam turbine bearings running to 312 fps.In assessing bearing limitations, there are many parameters thataffect results including geometry, operating conditions, and eveninstrument location. For example, DeChoudhury and Barth (1981)show that the drop in oil outlet temperature between the bearingand drain line can lead to significant differences in thermal balancecalculation of power loss. Pinkus (1990) describes peculiarbehavior between laminar, transitional, and high-speed turbulentregimes of bearing operation. Pettinato and DeChoudhury (1999)note high edge temperatures from misalignment in ball-in-socketpivot geometry. Wygant, et a1. (1999), show differences in steadystate and dynamic performance attributed to restriction of padmotion by friction in sliding contact pivots. Conventional floodedjournal bearings were used in these references. The flooded journalbearing design has been studied for many years, and information isavailable for a wide range of operating loads and speeds.In contrast, the few published papers on direct lube journalbearings cited earlier mostly report on low speed operation, in thelaminar to transitional range of operation. There is some disagreement regarding the magnitude of the benefits of direct lubrication.This may be due to laminar/transitional influences, but there arealso differences in the method of direct lubrication, as well as thetype of pivot, instrument location, etc. In general, all referencesagree that direct lubrication provides pad temperature benefIts thatappear to improve with load and speed. There is also agreementthat direct lube power loss is lower, although most authors reportthat thermal balance methods are of insufficient precision to allowan accurate assessment. Benefits are typically attributed to eliminating seal losses and reducing churning losses, but these arehypothetical because there are little data available for confirmation.In order to address such issues, a special test rig was designedand built to measure steady-state performance under light to moderately high loads, and to very high operating speeds that are beingapproached in new turbine and compressor designs. An importantfeature of the rig is the direct measurement of frictional torque,which provides a more precise measurement of power loss thanthermal balance techniques. Over the past few years, the authorshave conducted extensive tests and studies of parameters that affectthe performance of pivoted shoe journal bearings. The purpose ofthis work is to improve the capacity and performance of high-speedturbomachinery by extending bearing speed and/or load limitationsand improving efficiency.TEST RIG

Tests were performed on a new rig designed to investigatesteady-state performance under high operating speeds and light tomoderately heavy unit loads. The test apparatus is illustrated inFigure 1 and consists of a test bearing, the shaft and drive system,the loading system, and the rig supporting structure.

The test shaft is driven by alSO hp variable speed DC electricmotor. A belt and pulley system connects the motor to the test shaftand provides a 4.5: 1 speed step-up giving a maximum shaft speedof just over 16,000 rpm. The DC motor is linked to an electroniccontroller that ensures speed control to within 1 percent accuracy.The test shaft is supported by two 3.5 inch diameter pivoted shoejournal bearings spaced approximately 28 inches apart. The testshaft journal is 6 inches in diameter with circularity within 0.0005inch. Shaft speed is measured using a slotted optical switch in conjunction with a shaft-mounted disk containing a number of drilledholes. This was found to be a robust system capable of reliableoperation even at the top speed of the test facility.An important feature of the rig is the direct measurement of frictional torque. This is accomplished by the design of the testhousing (Figure 2). The two-piece housing is positioned on the testshaft, midway between the support bearings. The bottom face hasbeen accurately machined and sits in a spherical hydrostaticbearing that eliminates friction between the housing and theloading device. The housing is held against rotation by a 100 lbcapacity load cell mounted to a torque arm. Power loss is determined from the measured frictional force, and radii of the torquearm and bearing bore. The spherical hydrostatic bearing alsoprovides good alignment between the test bearing and shaft. Asecond plane hydrostatic bearing supports the spherical hydrostatic bearing. This provides lateral freedom and allows the load lineto be adjusted in relation to the test bearing. Proximity probesmounted on both ends of the test bearing housing measure the horizontal and vertical displacements of the test bearing with respectto the position of the shaft. These probes also provide an indicationof the level of alignment between the test bearing and the shaft.A pneumatic cylinder located between the test housing and baseof the rig (Figure 1) generates the static load. This has a capacityof 5500 lb. Three load cells located between the top of the loadapplicator and the bottom surface of the plane hydrostatic bearingmeasure the vertical load applied to the test bearing. The load cellsare arranged in such a way that they each share an equal proportion(one-third) of the load applied to the bearing. In calculating the netload on the test bearing, the combined weight of the test bearing,housing, and the hydrostatic bearing system is accounted for.A pump with a capacity of 20 gpm and a maximum supplypressure of 300 psig delivers oil to the test bearing from alSOgallon capacity tank. The flowrate is measured by a turbine typeflowmeter with a linear flow range of 2.5 to 29 gpm. Feed oil temperature is measured before it enters the test bearing, and iscontrolled by a water-oil heat exchanger to 1C (33.8F). Anindustrial type pressure transducer with a pressure range of zero to


3/14

Measurementof Frictional Torque

(Power Loss)

Figure 2. Test Housing and Torque Measurement System.

Table 1. Test Rig Measurement Uncertainties.Measurement Tyj>e of sensor Limit of error of sensorTemperature Type T thermocouple I 'c or 0.75%(whichever is greater)Shaft speed Optical switch 5 rpmBearing load 0-2000 lb. load cells (x 3) 5 lb. at full scaleFriction force 0-100 lb. load cell 0.1 lb. at full scaleDisplacements Eddy current proximity probes 0.00004 inOil flowrate Turbine flow meter 0.5% of readingOil supply pressure Pressure transducer 0.6Iblin"Leading edge groove Miniature pressure transducer 0.75 Ib/inoil pressureDrive motor power Power cell 10 hpconsumption

45 Type T thermos .020" below surfaceLoad

c'cw rotationII I

Figure 3. Babbit t Thermocouple Locations.% of arc length

75 8966 82 967 30 82., ., . 75 89667 t 9630 .. ,

o

c'cw rotationFigure 4. Loaded Shoe Instrumentation Locations.

A STUDY OF PARAMETERS THAT AFFECT PIVOTED SHOEJOURNAL BEARING PERFORMANCE IN HIGHSPEED TURBOMACHINERY 11

60 psig measures oil supply pressure to the test bearing.Thermocouples measure the oil inlet temperature, and the oil outlettemperature from each side of the test bearing. Flowrate and inletand outlet oil temperatures are used in thermal heat balance calculations of power loss for comparison with values obtained frommeasured torque.A special data acquisition system was developed for recordingthe data. In this system, process signals from the test facility arerecorded using a Windows graphical users interface program.This program provides an interface between the analog/digitalhardware and the operator, permitting real-time display of processconditions, the acquisition of test data, and signal processing suchas data trending, frequency analysis, and digital filtering. Signalsprocessed and displayed by the data acquisition system include testand support bearing temperatures; oil supply and outlet temperature, supply pressure, tank temperature, and flow; shaft speed,bearing load, test beating friction force, and X and Y displacements. The data acquisition hardware cards are rated at 1 MHz andthe display screen is updated every second. Measurement uncertainties of the equipment are listed in Table 1.

TEST BEARINGS AND OPERATING CONDITIONSThe test bearings are 6 inches in diameter, have a 0.437 lengthto-diameter ratio, a .0045 inch assembled radial clearance, and a.25 preload. The test bearing pads are of the same geometry, consisting of five pivoted pads with 60 degree pad angles. The pivot isa rolling contact design with curvature in both the axial and circumferential directions. This allows for axial alignment capability,and freedom of movement under changes in operating conditions.The test bearing pads are instrumented with an alTay of 45 typeT thermocouples with the tip of each thermocouple located .020inches below the babbitt surface (Figure 3). The loaded shoes aremore heavily instrumented along the centerline, and have additional thermocouples along the edges of the pads (Figure 4).

Tests were performed over a range of loads and speeds tocompare the effects of pivot offset, oil flow, load orientation, methodof lubrication, and oil discharge configuration. A description of thespecific parameter being tested is detailed in the associated sectionof this paper. The lubricant used in this series of tests was an ISOVG 32 turbine oil, with a measured viscosity of 32.76 centistokes at400 e (l04F), and 5.41 centistokes at lOOoe (212F). All tests wererun with an oil supply temperature of 49e (120.2F).CONVENTIONAL FLOODEDBEARING TESTS AND DISCUSSIONS

In hydrodynamic lubrication, the journal surface drags oil into thespace between the journal and bearing pads. The terms "conventional" and "flooded" are terms used to describe the method oflubrication that provides oil to the journal surface. The conventionalmethod is to restrict the hot discharge oil by sealing the sides of thebearing around the shaft such that the bearing cavity is flooded. Thisprovides oil to the journal surface by filling the spaces between thepads. There are a couple of variations. Floating seal rings aresometimes used with tight shaft-to-seal clearances. Oil discharge isthen restricted by oil outlet holes, sized to provide a slight backpressure to flood the housing. Another discharge configuration useslabyrinth seals where oil discharge is controlled by the gap betweenthe seals and shaft. Labyrinth seals were used for the flooded bearingtests in this report. RefelTing to Figure 5, oil is supplied to an outerannulus around the bearing, through radial oil supply holes betweeneach pad, and is restricted by seals on each side of the bearing.Center Pivot Versus Offset Pivot Tests

This series of tests was performed with conventional, floodedpivoted shoe journal bearings, to compare the effect of pad pivotlocation on bearing performance. The pivot locations tested are


4/14

Seals Seals

1.==.'1Oil outlet 4 ~ ~ ~ t ~ Figure 5. Conventional (Flooded) Bearing Lubrication.

50% offset 60% offset(center pivot) (offset pivot)

Figure 6. Center and Offset Pivot Geometry.

Center Pivot GEl Offset Pivot130 } - - - -Coad r}120

!JBearing Load On \':, +: 110 PadeE 100 l

LOP .. ()?

= : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : = = = = : : : : : : : : : : : ~ ~ = = - - .. Rotation'"oE 90

800..70 r

45 90 135 180 225 270 315 360position (deg.)

Figure 7. Pad Temperature Profiles, LOp, 200 PSI, 3600 RPM.

shown schematically in Figure 6. The center pivot design has thepivot located in the center of the pad, 50 percent of the pad arclength from the leading edge in the direction of rotation. The offsetpivot design has the pivot located at 60 percent of the pad arclength from the leading edge. Other than the location of the pivot,the bearing geometry for the two sets of pads is identical.

Pad Temperature ProfilesPad temperature profiles are obtained by plotting the centerline

thermocouple temperatures against the relative angular position ofthe detector in the bearings as depicted in Figure 7. This figure isfor a load-on-pad (LOP) orientation where the load is directedtoward a single pad. The loaded pad data are in the center of thefigure. As speed and load are increased (Figures 8 and 9), theloaded pad temperature increases and will eventually limit theapplication. Figure 10 compares the loaded pad temperatureprofiles plotted according to position along the pad arc length.

130

120: lIDe" 100"oE 90"0 80f1.

Center Pivot

70 .. . . . -

60 \rr""

I'J Offset Pivot

45 90 135 180 225 270 315position (deg.)

Figure 8. Pad Temperature Profiles, LOP, 200 PSI, 9000 RPM.

140

130

120: 110e" 100"oE 90

80

7060

50

Center Pivot I'J Offset Pivot( ) ? ~ ~ + L ~ t::t::::

/ f" .J7 po45 90 135 180 225 270 315

position (deg.)Figure 9. Pad Temperature Profiles, LOP, 320 PSI, 9000 RPM.

Center Pivot il l Offset Pivot140 ~ - ~ - - - - - - - - - - - - ~ - - - - - -130120

: 110e" 100"oE 90-g 800..

70605 0 L - _ ~ _ - L _ ~ __ L - _ ~ _ - L _ ~ __ ~ _ - L _

o 10 20 30 40 50 60 70 80 90 position (%arc)loaded pad

Figure 10. Loaded Pad Temperature Profiles, LOp, 320 PSI, 90RPM.


Pad SUiface IsothermsBefore any comparisons are made, it is useful to look at tentire pad surface profile. This is generated by curve fitting dfrom the array of thermocouples in the loaded pad surface. Fig

II is an isometric view of the loaded, center pivot pad. Figureis the same data looking directly at the babbitt face. Figures 13 a


5/14

A STUDY OF PARAMETERS THAT AFFECT PIVOTED SHOEJOURNAL BEARING PERFORMANCE IN HIGH-SPEED TURBOMACHINERY 13

14 are similar plots for the loaded, offset pivot pad under identicalconditions of load, speed, and oil flow. The surface temperatureprofiles are important in that they show if the pad is aligned withthe shaft or if there is edge or skew loading that can distort comparisons and lead to erroneous conclusions. The symmetricpatterns of Figures II through 14 indicate good axial alignmentbetween the loaded pad and shaft, which is attributed to thealignment capability of the rolling contact pivot design. Anotherobservation is that there is a fairly substantial axial temperaturegradient noticeable in the figures. Temperature drops 15 to 20 De(59 to 68 DF) from the center hot spot to the sides of the pad for theoperating conditions shown.

ilil-130:135!i i! 125-130;'11!l120-125:L'i! 115-120

110-115:0105-110;

':=i100-105iiD95-100 I!O 90-95if ] 85-90iD80-85iEl75-80,I ii! 70-751l11li65-70:1I!I60-65I1 l1li 55-60i 1150-55

Figure 11. Center Pivot Pad Suiface Isotherm, Isometric, LOP, 320PSI, 9000 RPM.

:-D130-135-!1II1II 125-1301I ill 120-1251!1J115-120!IIi!l 110-1151

SE I iJ 105-110!f! I 1 0 0 ~ 1 0 5 i ICi 95-100 '8. 1090-95E I1085-90'0 ;080-85&. li!il 75-80

:Gill 70-75i l1li65-70iGill60-65iIGill 55-60, 'i _ ~ __~ O : . 2 __J

Figure 12. Center Pivot Pad Surface Isotherm, Plan View, LOp,320 PSI, 9000 RPM.Pad Temperature Comparisons, Center Versus Offset Pivot

A study of pad surface isotherms indicated that the pads werealigned with the shaft, and that peak temperatures were locatedalong the circumferential centerline for the conditions tested. Withthis assurance, centerline pad temperature profiles can be used toprovide an accurate indication of the highest pad temperatures.Figures 15 through 17 are pad temperature profiles for a loadbetween-pad (LBP) orientation where the test load is applieddirectly between the two bottom pads. LBP pad temperatureresponse is similar to the LOP data of Figures 7 through 9 in thatas speed and load are increased, the loaded pad temperatures also

iI!!l130--;35i135 !1iII125-130!

Iliil120-125ijill! 115-1201ilill110-115 1!Cci 105-110[: i!::l100-105!e 'IU95-100'1090-95 !: ] 85-90Ie'] 80-85I - ,'t l ii2l75-80&. I ll! 70-7511I!III65-701111160-65 iIU55-60 ILiII 50-55 _J

Figure 13. Offset Pivot Pad Suiface Isotherm, Isometric, LOp, 320PSI, 9000 RPM.111130-135l11li125-130:!ill 120-125!Jl! 115-120;

SE :&\l110-115!:Ell 105-110'-: 100-105.ar: i:J 95-1000- :090-95Eil l [] 85-90I -

"0 80-85


6/14


Center Pivot [gjJ Offset PivotJ40 ~ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - , 130120

: 110:lf 100"oE 90

80

LBPLoad

BetweenPads

Bearing Load

I. F = = = = = = = = = = : @ ~ ~ Rotation70605 0 ~ ~ 45 90

... ...

135 180 225 270 315position (deg.)

Figure 15. Pad Temperature Profiles, LBP, 200 PSI, 1800 RPM.+ Center Pivot III Offset Pivot

140130 fl )J.120

~ + c O : 110:l

100"oE 90 jJ-"0 80". 70 ..0

5045 90 135 180 225 270 315

position (deg.)Figure 16. Pad Temperature Profiles, LBP, 200 PSI, 9000 RPM.

Center Pivot III Offset Pivot

360

360

140 r - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~

7 0 ~ . / 60

45 90 135 180 225position (deg.)

270

2ndloadedpad

315

Figure 17. Pad Temperature Profiles, LBP, 320 PSI, 9000 RPM.360

Differences between center and offset pivot temperatures arealso noticeable in Figures 7 through 9 and 15 through 17. In allcases, the offset pivot runs with cooler overall pad temperatures,particularly noticeable on the loaded pads. Figure 10 for LOP andFigure 18 for LBP focus on the hotter, loaded pad. For the centerpivot, the pad temperature is noticed to increase to a maximum ata position of 82 and 89 percent of the arc length for LOP and LBP,respectively, after which it falls to a lower level. For the offset pivot

Center Pivot III Offset Pivot1 4 0 r - - - - - - - - - - - - - - - - - - - - - ~ - - , _ - - - - - - - - - - - - - - -

130120

10 20 30 40 50 60 70 80 90 position (%arc) 2nd loaded pad

Figure 18. Loaded Pad Temperature Profiles, LBP, 320 PSI, 9RPM.design, the pad temperatures reach a maximum close to the trailedge of the pad. These observations are consistent with 3.88 idiameter results in earlier work by the authors (Brockwell, et 1994). For the range of loads and speeds tested, the center pimaximum temperature varied between the 65 percent and percent location. The offset pivot maximum temperature varbetween the 82 percent location and trailing edge. The speclocation of the maximum temperature varied with speed and lo

To further reduce the data for generalized comparison, it is helpto view a single location. The 75 percent position is a typical locaused in industry (Nicholas, 1994) and is recommended in many spsuch as American Petroleum Institute (API). With the entire promeasured in detail, and data showing different maximum tempeture locations for the center and offset pivot designs, some authchoose to compare maximum measured temperatures (Simmons Lawrence, 1996). Figures 19 through 22 compare 75 percent locaand maximum temperatures for LOP and LBP orientation at percent of recommended flow. The speed range for these compisons was limited because the lubrication system could not provthe recommended flow for higher speeds. However, tests were aperformed with 50 percent flow, shown in Figures 23 and 24.

140,----------------------------------------

130120ui 110

::s 100(])CoE 90(])I --g 80c..

7060

o

- Center Pivot........ Offset Pivot

0320 psio 200 psi

2000 4000 6000 8000 10000 12000 14000 1Shaft rpm

Figure 19. 75 Percent Location Pad Temperatures, LOPThe change in slope and inflection noticeable in Figures 23

24 indicate that the speed range covers laminar, transitional, turbulent operating conditions. Figures 19 through 24 show thatoffset pivot design has lower 75 percent location and lower mmum temperatures than the center pivot design for all conditi


7/14


140 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~ 130120u 110e::l..... 100 -

QlC.E 90 -QlI -"C 80 c.E 90 -C1>I -

"C 80


8/14

16 PROCEEDINGS OF THE 30TH TURBOMACHINERY SYMPOSIUM80 ~ - - - ~ - - - . - - - - - - . - ~ - - ~ - ~ - - ~ - ~ ,

706050

I/)I/).3 40o 30a.

20 -


o 2000 4000 6000 8000 10000 12000 14000 16000Shaft rpm

Figure 25. Center Versus Offset Pivot Power Loss, LOP, 100Percent Flow.80,----------------------,

70

6050

I/ )I/ )0 40 -..J..Q):: 30 -0a.

2010

- Center Pivot........ Offs et Pivot

o 200 psi

o 2000 4000 6000 8000 10000 12000 14000 16000Shaft rpm

Figure 26. Center Versus Offset Pivot Power Loss, LEP, 50 PercentFlow.1 4 0 , - - - - - - - - - - - - - - - - - - - - - ~ 130120

Q: 110l!!100

Q)a.

- Center Pivot, 100% Flow........ Center Pivot, 50% Flow

90--g 80a.

7060

0320 psio 200 psi

o 2000 4000 6000 8000 10000 12000 14000 16000Shaft rpm

Figure 27. 75 Percent Location Pad Temperatures, LOp, 50 Versus100 Percent Flow.with flow at high speeds. These appear to coincide with the laminar,transitional, and turbulent film regimes. And so it would seem thatthe observed phenomenon is associated with film turbulence.

Be that as it may, the purpose of the work is to improve bearlimitations. Flow reduction for the center pivot design is inapppriate because pad temperatures (Figure 27) are already hiHowever, the lower temperatures of the offset pivot design mallow for an improvement in efficiency with reduced oil flow.With center pivot pad temperatures as a baseline, Figure compares the effects of decreasing the offset pivot bearing oil flfor a LOP orientation at 320 psi projected load and 10,660 rpm. a 50 percent reduction in flow, the offset pivot loaded pad tempatures increase fairly evenly from leading to trailing edge, such tthe entire pad is running approximately lOoe (50F) hotter. W50 percent flow, the offset pivot 75 percent location is still coothan the center pivot baseline, but the maximum pad temperaturehotter. The question as to which comparison should be usedjudge the two designs becomes more relevant.

... center pivot 100% flo140 r-______ ~ - - ~ - - - - - - - - - ~ ~ o T f f ~ s ~ e ~ t P ~ i v ~ o ~ t ~ 5 ~ O ~ % ~ f ~ l o ~ w130120

liDe!" 100"oE 90-g 80Q.

70

60

i ... qffset pivot 100% flo----_-'----1 - - - - ~ - " - - - - - - . - ! . - ..---------

i i- - . ~ - - . - - ---I ----- - - - - - - 1 - - - - - - - - + - - - ------- .. ~ - - - - - i - - - - - - - - - - - - ~ - _ + 7 " " _ _ r

50 L-__ ~ _ ~ __ _ _ L _ ~ __ _ _L___ L _ _ ~ _o 10 20 30 40 50 60 70 80 90

position (%arc) loaded padFigure 28. Effects of Flow on Loaded Pad Temperature ProfilLOp, 320 PSI, 10,660 RPM.

In the authors' experience, both are important. Maximum pad teperature is of concern in regard to the oil, where additives and bstock can break down or deposit on the pad surfaces causing beartemperatures to increase over time. In regard to the bearing, thepercent location is a general area known from experience to sudistress under adverse conditions. This is typified by Figure 29, whshows mechanical fatigue of the babbitt near the 75 percent locatof a center pivot pad. The 75 percent location is only a rule-of-thumThe actual location changes with load and speed and is also inenced by other parameters. A technically valid comparison requian analysis of pressure and temperature related to mechanical critfor the babbitt. One method used by the authors is as follows:

rFigure 29. Mechanical Fatigue on a Center Pivot Pad.


9/14

A STUDY OF PARAMETERS THAT AFFECT PIVOTED SHOEJOURNAL BEARING PERFORMANCE IN HIGH SPEED TURBOMACHINERY 17

Figure 30 is a plot of yield point versus temperature based ondata from ASTM-B23 for grade 2 babbitt. In the normal range ofbearing operating temperatures, the yield point is approximately4000 psi at 70C (I58F), and reduces to 2000 psi at 130C(266F). The purpose of the figure is to establish a materialreference point to judge the two designs.

60005000

Q.- ; 4000...til 3000el. 2000

1000

ASTM 8-23 Grade 2 White MetalYield Point (O.125% of Gage length)

o + . - - - - . - . - . , - - - - ~ . o 20 40 M W 100 1 ~ 1 ~ 1M 1WTemperature (C)

Figure 30. Babbitt .125 Percent Offset Yield Point.Figure 31 plots the calculated pressure profile and curve-fittedmeasured pad temperatures for the center pivot pad of Figure 28.Notice that the peaks occur at different locations such that at peaktemperature, the pressure on the babbitt is very low and at peakpressure, the temperature is low. Figure 32 is a similar plot for theoffset bearing with 50 percent flow. By comparing the pressure andtemperature to the babbitt yield point, the condition along the padarc can be evaluated. For example, from Figure 31 at the 52 percentlocation, the film pressure and temperature are 1400 psi and 91C(195.8F), respectively. From Figure 30, the yield point at 91C(195.8F) is 3200 psi. The ratio of the film pressure to yield pointis 1400/3200 psi = .46 for this location. By repeating thisprocedure for each position along the pad, a map of the pressure toyield ratio can be generated as shown in Figure 33. The closestapproach to the yield point is found at the 62 percent location for

the center pivot design, and at the 82 percent location for the offsetpivot design. The highest ratio for each is approximately .47 or, inother words, both have a worst case pressure location that is 47percent of the babbitt yield point.

140

130

120

Q: 110100'"

ell 90C.Eell80I -70

60

50

Pressure (psi) II1II Temperature (C)

PressureProfile Temp,Profile 17501500

1250

1000 i1l750 D.

500

250

~ - - , - - - - - . - - - - - - - . - - - _ _ _ . - - _ _ _ . - - - - - . - - - - - . - - - - - . - - _ _ _ , - - ~ O 0 10 20 30 40 50 60 70 80 90 100position (% arc) loaded pad

Figure 31. Center Pivot Loaded Pad Pressure and TemperatureProfiles, 100 Percent Flow.

The comparison shows that the offset pivot bearing running with50 percent flow and the center pivot bearing running with 100

ill Pressure (psi) r:il Temperature (C)140 , - - - - - - - - - - - - - - - - - - - - - - - - - . - - - - - - - - - - - - ~ 2000130

120

Q: 110.a 100

90E 80

7060

PressureProfile1750

1500

1250

1000 i1l

750 D.500

250

5 0 ~ - - . _ _ _ - - . - - - _ _ _ . . _ _ _ _ _ _ . - - _ _ _ . - - - - - . - - _ _ _ , - - _ _ _ , - - _ _ _ , - - ~ 0 o 10 20 30 40 50 60 70 80 90 100position (% arc) loaded padFigure 32. Offset Pivot Loaded Pad Pressure and TemperatureProfiles, 50 Percent Flow.

1.000.90

1: 0.80'8. 0.70'0]! 0.60>.- 0.50'wC.OAOE'i: 0.30g 0.20ns!r 0.10

0.00

I. JI -+- Center Pivot, 100% flow 1:_- ~ - - -Offfiet P;"'tj='n--f - - .

--_. : L -- 'L 1 -- . ' ~ ... - "'-.K ',/ /1-'"' 1\

A'"


10/14

]8 PROCEEDINGS OF THE 30TH Tl'RBOMACHI:\ERY SYMPOSIUM

706050

(/j(/j0 40 -...Jt300a.20

10

o


~ , ~ ' -,"If"-

,.0-'0320 psi

2000 4000 6000 8000 10000 12000 14000 16000Shaft rpm

Figure 34. Power Loss-Center Pivot 100 Percent Flow VersusOffset Pivot 50 Percent Flow.9 0 - , - - - - - - - - - - - - - - - -- - - - - - - - - - - - - - -- - - - - - - - - - - ,

8580

Q: 75:::>

70Qla.E 65 -

60 -55

o


0320 psi

2000 4000 6000 8000 10000 12000 14000 16000Shaft rpm

Figure 35. Drain Temperature-Cellter Pivot 100 Percent FlowVersus Offset Pivot 50 Percellt Flow.Figure 36 compares temperature profiles fo r LOP and LB Porientations for the center pivot bearing at identical operatingconditions. It is noticed that the hottest pad temperature for

each orientation reaches similar levels, just above lOOoe(212F). In other words, there is no t a significant reduction inpad temperature going from LOP to LB P orientation for thecase shown. Figure 37 compares pad temperature profiles ofthe LOP loaded pad with the LB P second loaded pad. There islittle difference in pad temperature profile between the two orientations.

Figure 38 plots center pivot 75 percent location pad temperatures, and Figure 39 plots offset pivot power loss measurementscomparing LOP and LBP performance. These figures were chosento represent general observations of a study of all operating conditions tested that include center and offset pivots at various flowsand over a range of operating loads and speeds. LBP orientation inmost cases ran with lower pad temperatures than the LOP configuration, but not as much as would be expected. This is attributed tothe effects of ho t oil canyover. In general, the temperature benefitsof LBP orientation improve with load and speed. In regard topower loss, there was no measurable difference between LOP andLBP orientation (Figure 39).

140130

120110

" 100".E 90"--g 80a.

70

60

50

LOP Q L B P

LOP LBP2ndloaded loadedpad pad

l- f !srP 1/~ A r 0-~ . 45 90 135 180 225 270 315position (deg.)

Figure 36. LOP Versus LBP Pad Temperature Profiles, 207760 RPM, 50 Percent Flow. LOP ( 1 st pad) Q L B P ( 2nd pa

1 4 0 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

130

1202: 110e" 100E 90,!

80a.7060

10 20 30 40 50 60 70 80 9position (%arc)

Figure 37. LOP Versus LBP Loaded Pad Temperature ProfilePS1, 7760 RPM, 50 Percent Flow.

Q::::>....r:Ql0-EQlI -"0coa.

1 4 0 , - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

130 - LOP (1st pad)........ LBP (2nd pad)120 -110100908070 - 0320 psi

o 200 psi!!, 50 psi

60

o 2000 4000 6000 8000 10000 12000 14000Shaft rpm

Figure 38. 75 Percent Locat ion Loaded Pad Temperatures, Pivot, 50 Percent Flow.FLOODED VERSUS DIRECT LUBETEST DATA AND DISCUSSIONS

This section compares the effects of direct lubricatiodischarge configuration. As mentioned earlier, the term "flo


11/14


80 . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ~ 706050

enen.3 40 -o 30c..

2010

- Offset Pivot, LOP........ O ffse t Pivo t, LBP

0320 psi

o 1 ___ ~ ~ ~ ~ ~ : : ~ - _ . - - ~ - - ; ~ = - ~ 5 ~ O ~ P : S I ~ -Jo 2000 4000 6000 8000 10000 12000 14000 16000Shaft rpm

Figure 39. Power Loss-Offset Pivot, LOP and LBP, 50 PercentFlow.refers to a method of lubrication. For the flooded design tested, oilis provided to the bearing between the pads and the discharge isrestricted by labyrinth seals, which provides oil to the journalsurface by flooding the spaces between pads (Figure 40). Leadingedge-groove is a method of lubrication that provides oil directly tothe journal surface at the entrance to the oil film (Figure 41). Inboth of these methods, the principles of hydrodynamic operationare the same. Specifically, the journal drags oil into the spacebetween the journal and bearing pads.Flooded

Oil outlet

Evacuated

OpenedSeal

Clearance

Figure 40. Discharge Configuration, Flooded Versus Evacuated.Conventional

j\/ \60% offset / '

(offset P i V / " ~ _ 6 ~ _ .\/-'. - j ./ c : t ~ J / < ~ ~ 0 Oil inlet ' " ~ i - ' ; ~

Pivot

LEG

PivotFigure 41. Method of Lubrication, Conventional Versus DirectLube.

In the case o f direct lubrication, it is desirable to allow the hot oilto exit freely from the bearing to reduce parasitic losses and efficiently remove heat. There are a few different methods forevacuating the bearing cavity. For test data reported in this paper, thelabyrinth seal clearances were enlarged (Figure 40). This is a mostefficient method because the side leakage (exiting oil) is dischargedbetween the end seals and shaft without being captured and recirculated within the bearing. Note that in these comparisons, twoparameters are changed-the method of lubrication and thedischarge configuration. Otherwise, the flooded and LEG bearingsare the same. The pads have 60 percent offset pivots (Figure 41), anddata are compared under identical operating conditions and oil flow.Conventional (Flooded) Versus LEG (Evacuated)

Figure 42 compares conventional, flooded bearing pad temperature profiles to direct lube, evacuated pad temperature profiles atconditions of 320 psi and 10,660 rpm. The direct lube design hasnoticeably lower pad temperature levels, especially at the leadingedges of the pads. This is attributed to providing cool oil directlyto the leading edge, which affects the entire loaded pad (Figure 43).Comparing 75 percent location temperatures for 200 psi in Figure44, and 320 psi in Figure 45, this benefit is noted to improve withload and speed, up to 16C (60.8F) cooler at extreme test conditions. It is also noticed that the differences begin around theturbulent transition, evident by the inflection in the curves. Thisseems to tie in with the behavior of pad temperature and flow withturbulence discussed in the reduced oil flow section of this paper.

[J Conv. flooded LEG evacuated140 ,------------------------____130120

: 110e:J 100QlC.E 90

80a..

60

.0'-0

50 L___ ___ L____ ___ ___ L _ _ ~ __ L___o 45 90 135 180 225 270 315 360

position (deg.)Figure 42. Pad Temperature Profiles, LOp, 320 PSI, 10,660 RPM,50 Percent Flow.

140130120

: llO!i 1008.E 90{!.80

70

6050

Conv. flooded LEG evacuated

-'l-V) ----1 / /

------y V

---- l-----' -----Vf - - - J--t- L----...I----

10 20 30 40 50 60 70 80 90 100position (%arc) loaded padFigure 43. Loaded Pad Temperature Profiles, LOP, 320 PSI, 10,660RPM, 50 Percent Flow.


12/14


13/14


80 II0 4- Conv. flooded60 G LEG flooded50


14/14

22 PROCEEDINGS OF THE 30TH TURBO MACHINERY SYMPOSIUM

labyrinth seals. The LEG evacuated discharge configuration hadlarge clearance end seals. The conclusions are: Part of the reduction in LEG pad temperatures is due to thedirect lube design, and part is due to the evacuated discharge configuration. A significant portion of the power loss reduction is attributableto evacuation of the housing, and a smaller portion to the methodof lubrication.ADDITIONAL COMMENTS

Data contained in this report show the necessity to considerother parasitic losses (such as churning and seal losses) incomputer models. Parasitic influences are discussed by manyauthors (e.g., Booser, 1990), and are required to predict the higherlosses in flooded designs, and the lower losses in evacuateddesigns. Other phenomena mentioned throughout the paper need tobe better understood to develop accurate computer models. A studyof pad temperature profiles can give valuable insight i nto hydrodynamic behavior, and is necessary for programming models toaccurately predict performance. The surface isotherms showingaxial temperature gradients are also useful data for program development. It is hoped that the information and high-speed datacontained in this paper will enhance the study and prediction ofbearing performan ce in turbomachinery.REFERENCESBooser, E. R., 1990, "Parasitic Power Losses in Turbine Bearings,"

STLE Tribology Transactions, 33, (2), pp. 157-162.Brockwell, K., Dmochowski, w., and DeCamillo, S., 1992,"Performance Evaluation of the LEG Tilting Pad Journal

Bearing," I Mech E Seminar Plain Bearings-PlainBearings-Energy Efficiency and Design, MEP, London,England, pp. 51-58.

Brockwell, K., Dmochowski, W., and DeCamillo, S., 1994,"Analysis and Testing of the LEG Tilting Pad JournalBearing-A New Design for Increasing Load Capacity,Reducing Operating Temperatures, and Conserving Energy,"Proceedings of the Twenty-Third Turbomachinery Symposium,Turbomachinery Laboratory, Texas A&M University, CollegeStation, Texas, pp. 43-56.

DeCamillo, S. and Clayton, P. J., 1997, "Performance Tests of an18-Inch Diameter, Leading Edge Groove Pivoted Shoe JournalBearing," Proceedings of the Second International Conferenceon Hydrodynamic Bearing-Rotor System Dynamics, Xi ' an,China, pp. 409-413.

DeChoudhury, P. and Barth, E. w., 1981, "A Comparison of FilmTemperatures and Oil Discharge Temperature for a Tilting-PadJournal Bearing," ASME Journal of Lubrication Technology,103, pp. 115-119.DeChoudhury, P. and Masters, D. A., 1983, "Performance Tests ofFive-Shoe Tilting-Pad Journal Bearing," ASLE Transactions,27, (1), pp. 61-66.

Dmochowski, w., Brockwell, K., and DeCamillo, S., 199Study of the Thermal Characteristics of the Leading Groove and Conventional Tilting Pad Journal BearASME Journal of Tribology, 115, pp. 219-226.

Edney, S. L., Waite, J. K., and DeCamillo, S. M., 1996, "PrLeading Edge Groove Tilting Pad Journal Bearing for Load Operation," Proceedings of the TwentyTurbomachinery Symposium, Turbomachinery LaborTexas A&M University, College Station, Texas, pp. 1-16.

Fillon, M., Bligoud, J. C., and Frene, J., 1993, "Influence oLubricant Feeding Method on the ThermohydrodynCharacteristics of Tilting-Pad Journal Bearings," Proceeof the Sixth International Congress on Tribology, BudaHungary, 4, pp. 7-10.

Gardner, W. W. and Ulschmid, 1. G., 1973, "Turbulence Effects inJournal Bearing Applications," ASME Paper No. 73-LubHarangozo, A. v., Stolarski, T. A., and Gozdawa, R. J., 1991,Effect of Different Lubrication Methods on the Performana Tilting Pad Journal Bearing," STLE Tribology Transac34, pp. 529-536.Nicholas, 1. c., 1994, "Tilting Pad Bearing Design," Procee

of the Twenty-Third Turbomachinery SymposiulIl, Tmachinery Laboratory, Texas A&M University, CoStation, Texas, pp. 179-194.

Pettinato, B. and DeChoudhury, P., 1999, "Test Results of KeSpherical Pivot Five-Shoe Tilt Pad Journal Bearings-PPerformance Measurements," STLE Tribology Transac42, (3), pp. 541-547.Pinkus, 0., 1990, Thermal Aspects of Fluid Film Tribology,York, New York: ASME Press.Simmons, J. E. L. and Lawrence, C. D., 1996, "PerfonnExperiments with a 200 mm, Offset Pivot JournalBearing," STLE Preprint No. 96-AM-7B-4.Tanaka, M., 1991, "Thermohydrodynamic PerformanceTilting Pad Journal Bearing With Spot Lubrication," AJournal of Tribology, 113, pp. 615-619.Tanaka, M. and Mishima, N., 1989, "Thermohydrodyn

Performance of a Spot-Fed Tilting Pad Journal BeaProceedings of Eurotrib89, 3, pp. 198-203.Wygant, K. D., Flack, R. D., and Barrett, L. E., 1999, "Influen

Pad Pivot Friction on Tilting-Pad Journal BeMeasurements-Part I and Part II," STLE TribTransactions, 42, pp. 210-215 and 250-256.

ACKNOWLEDGEMENTThe authors sincerely thank all personnel at Kingsbury, Incthe National Research Council of Canada, who contributed work documented in this paper. The authors are gratefKingsbury, Inc., and the National Research Council of Canadpermission to publish this information.

Documents

Parameters Affecting PSJB Performance.pdf