1、11FTM22AGMA Technical PaperBearing Contribution toGearbox Efficiency andThermal Rating: HowBearing Design CanImprove thePerformance of aGearboxBy A. Doyer, SKFBearing Contribution to Gearbox Efficiency and ThermalRating: How Bearing Design Can Improve the Performanceof a GearboxArmel Doyer, SKFThe s
2、tatements and opinions contained herein are those of the author and should not be construed as anofficial action or opinion of the American Gear Manufacturers Association.AbstractGearbox efficiency is a topic of rising interest amongst both OEM and end-users due to an increasedsensitivity to gearbox
3、 performance, reliability, total cost of ownership (energy cost), overall impact on theenvironment, and also anticipating future regulations.Agearboxisbynatureaquiteefficientasset:aparallelshaftgearunitissaidinliteraturetohavetypicallossesof 1 to 2% per stage. As an example, a single stage gearbox c
4、ould have a nominal efficiency of 98 to 99%.Anyhow, due to the high power/ torque transferred by the system, it could be interesting to minimize theabsolute losses as they could be significant (1% of 1MW is still 1kW), especially when existing technologyallows it at very reasonable cost and without
5、much complexity.In a gearbox, there are difference sources of losses: gear, lubrication, seal and bearing loss. The use ofmodern simulation tools makes easier the evaluation of losses in various load case conditions.It has been demonstrated that the contribution of bearing loss on the system efficie
6、ncy is dependant on theloadcases.Evenifthebearingisbyfarnottheprimarysourceoflosses,theoptimizationofthebearingsetcansignificantly improve the gearbox performance.Simulationofasinglestagegearboxusingtaperedrollerbearings,showsthattherunningtemperatureofthegearbox can be reduced up to 10C, by using l
7、atest bearing generation. Suchasavingcouldimprovethethermal rating of the gearbox by up to 30%.Experiment also demonstrated that different design of tapered roller bearing shows significant variation infriction performance.Having proper bearing design could significantly improve the performance of a
8、 gear unit: by a lower runningtemperature, by improving lubricant life, potentially simplified lubrication system, and consequentlyreducedrunning cost.Copyright 2011American Gear Manufacturers Association1001 N. Fairfax Street, 5thFloorAlexandria, Virginia 22314October 2011ISBN: 978-1-61481-022-33 1
9、1FTM23Bearing Contribution to Gearbox Efficiency and Thermal Rating: How BearingDesign Can Improve the Performance of a GearboxArmel Doyer, SKFIntroductionGearbox efficiency is a topic of rising interest amongst both manufacturers and end-users due to anincreased sensitivity to gearbox performance,
10、reliability, total cost of ownership (in relation to energy cost),overall impact on the environment, and also anticipating future regulations.A gearbox is by nature a quite efficient asset and as such, it has not been subjected to the same debateregarding energy efficiency as other machine component
11、s as e.g., electrical motors. However, due to theincreased awareness of environmental impact and the increased energy costs, the optimization of energy isbecoming a topic of greater importance also for industrial gearboxes. Looking at the high power/ torquetransferredby thesystem, itis ofinterest to
12、minimizethelosses interm ofabsolutevalues(1%of1MW isstill10kW). This is especially valid, when existing technology allows it at reasonable cost and without addingcomplexity.As there is a competitive advantage to give the maximum possible output mechanical torque in a given gearunit size, there will
13、be a growing competitive race for manufacturers to show the highest thermal rating for agiven size (Figure 1). Energy efficiency is increasing its importance as selection criteria.In this paper, the author will give:S a brief recap of the gearbox inefficiency sources,S an overview of the latest bear
14、ing friction modelS information on the latest tapered bearing technology development, andS how this can affect the gearbox performance trough an example of a single stage gearboxGearbox efficiency, inefficiency and thermal ratingAsmostoftechniciansandengineerslearnatschool,agearboxisbyitsnatureaquit
15、eefficientasset:aparal-lel shaft gear unit is said in literature to have typical losses of 1 to 2% per stage(1). As anexample, asinglestage gearbox could have a nominal efficiency of 98 to 99%.Figure 1. Mechanical and thermal power ratings of a single stage helical gear ratio 5 for varioussizes, for
16、 different gearbox manufacturers4 11FTM23The losses are of different types:S gear lossesS lubrication lossesS seal losses (when seals are present)S windage losses (for high speed gears)S bearing lossesMany authors have already described well some of those losses and the overall behavior of the gearb
17、ox (1,2and3). Thoseauthorsshowedalsothattheproblemisquitecomplex,especiallybecauseinagearboxisa system where losses interact/influence each other due to the thermal equilibrium/heat dissipation.The ISO technical report 4, published a decade ago, lists guidelines for calculating the gearbox thermal r
18、at-ing, which is another term to describe efficiency. The advantage of this rating is that it can be compared withthemechanicalratingofthegearbox,andthustheusercanquicklyseewheneitheracoolingsolutionneedstobe added or improved, or the gearbox size needs to be altered.Most people have in mind that th
19、e gear losses are the dominating ones. While this is true in many cases, itdoesdependon thegearbox designand loadcases (2,3). Nowadays, withengineering software,detailedanalysis of gearbox has become much more simple, faster and accurate than before, and it can helpdesigners (and users) to optimize
20、and better understand their system. Doing so, they will learn the relativeimportance of the different losses may vary significantly; and that losses other than gear ones cannot beneglected in the analysis.As an example, the authors studies have shown that the bearing loss can range from 30 to 50% of
21、 the totallosses (nearly equal to the gear losses), depending on the applied loading. When a gearbox is used at thelevel of its nominal mechanical rating, gear losses tend to be dominant, which is expected.Itisalsointerestingtonoticethatthesplitbetweenthedifferentshaftsisnotequal. Dependingonthegear
22、boxdesign, the gearbox ratio, the bearing load and speed will vary. As illustrated inFigure 2, onemay findcaseswheretheinputshaftpositionsaremajorsourceofbearinglosses; others,where theoutput andintermediatepositions are the ones generating the highest bearing losses.In order to optimize the relevan
23、t part of the gearbox, it is therefore important touse andunderstand thelatestknowledge and model regarding bearing friction.Figure 2. Examples of possible distribution of bearing loss per shaft in different gearbox setup(results of SKF investigations)5 11FTM23SKFfrictionmodelSKFimplementedanewbeari
24、ngfrictionmodelinitsgeneralcatalogue(latesteditionsee5),alreadyin2003.See Table 1. This model is based on four sources of friction:(1)M = Mrr+ Msl+ Mseal+ MdragwhereM is total frictional moment, Nmm;Mrris rolling frictional moment, Nmm;Mslis sliding frictional moment, Nmm;Msealis frictional moment o
25、f the seal(s), Nmm;Mdragis frictional moment of drag losses, churning, splashing, etc., Nmm.Thisnewapproach6identifiesthesourcesoffrictionineverycontactoccurringinthebearingandcombinesthem;inadditionthesealcontributionandadditionalexternalsourcescanbe addedas requiredto predicttheoverallfrictionalmo
26、ment. Sincethemodellooksintoeverysinglecontact(racewaysandflanges),changesofdesign and improvements of the surfaces can readily be taken into consideration, making the model moreable to reflect improvements in SKF bearing designs.This four sources model allows the designer to understand in details i
27、n which conditions the bearing ininternallyfunctions. ForexampleinFigure 3,thefoursourcesoflossareplottedasafunctionofspeed;itcanbe check where the bearing losses are driven by rolling or sliding sources.Table 1. Comparison of philosophy: old model load depend/load independent, and SKF newmodel (4 s
28、ources of friction)Old model New modelMb= M0+ M1+ M2+ M3M = Mrr+ Msl+ Mdrag+ MsealM0= 107fo(vn)23d3mLoad-independent part(mainly rolling)MrrRolling friction moment(raceways)M1= f1Pa1dbmLoad-dependent part(sliding correction)MslSliding and spinning frictionmoment (flanges, raceways)Ms= f2FadmCRB axia
29、lly loaded (slidingflanges)MdragOil bath, large bathM3= d + Df32+ f4Sealed bears MsealFriction moment due tosealsFigure 3. Example of 4 sources distribution in a spherical roller bearing,with oil bath and thick oil6 11FTM23Inaddition,asshowninFigure 4,thenewfrictioncataloguemodelfitswellbearingfrict
30、ionmeasurement(hereon tapered roller bearing), over the speed rangeUsing this new model will allow gear user and designer to have a better prediction and understanding of thebearing losses in a gear unit, over various loading conditions. Thus, it allowsan improvedoptimization ofthesystem, by having
31、for example a more accurate comparison between different bearing types and designs.SKF energy efficient tapered bearingItisthankstotheunderstandingoffrictionbehaviorand thebetterfrictionmodel,thatSKFwasabletodevelopa new generation of tapered roller bearing (7 and Figure 5), so called energy efficie
32、nt bearing. Those bear-ings generate 30% less friction than conventional tapered roller bearing designs in most loading conditions(Figure 6 and Figure 7).Figure 4. Model prediction and measurement in a tapered roller bearingFigure 5. SKF energy efficient tapered roller bearing7 11FTM23Axial load: 40
33、 kN/oil: ISO VG 320/T=80CFigure 6. Frictional moment of SKF energy efficient tapered bearing versus SKF standard designFriction torque test - size 32230Radial load 10 kN, axial load 6 kN, 80CFigure 7. Frictional moment of SKF energy efficient tapered bearing versus SKF standarddesign, in different o
34、il levelManybearingdesignparameterswerereviewedandoptimizedtoreachthisfrictionsaving,withoutcomprom-isingthefatiguelifeofthebearing. Forexample,thisnewdesignhassomespecificflangegeometry,reducedtherecessandanextendedinnerringraceway. Moreover,specialracewayprofilesandrollertopographiesinconjunctionw
35、ithreducedroughnessoftheringracewaysandflangehavebeen adopted. Aspecial cagewithreduced bore diameter preferably made of PEEK or, for special demands, of sheet steel has beendeveloped.The most visible change consisted to reduce the number of taper rollers. For a bearing type 32230 J2, therollersetha
36、sbeenreducedbyfour. Thankstothereducednumberofrollers,therotatingmasshasdecreasedby approximately ten percent.8 11FTM23The reduced number of rollers has also a major influence on lubrication. Fewerrollers meanless frictionandmechanical working in the lubricant. This leads to lower operating temperat
37、ures, which in return improve theseparationofthesurfacesinrollingcontactthroughbetterlubricantfilmformationandadditionallyextendsthelubricant life30%lessfrictionisaquantifiableimprovement,butthequestionforagearboxdesigneranduseris:whatdoesit imply for the gearbox performance, in terms of thermal rat
38、ing and life performance? In the next section, anexample of simulation on a gear unit is given.What is the impact on a gearbox? Example of a single stage gearbox:A single stage helical gearbox was selected for the analysis. This gearbox has a mechanical power of280 KW, and a thermal rating of about
39、50 kW. The reduction ratio is 4. This gearbox is equipped with 4identicaltaperedrollerbearings(borediameter60mm,series323). Bearings1and2arelocatedfacetofaceon the input shaft. Bearings 3 and 4 are located face to face on the output shaft.The analysis was performed in 2 steps.S a preliminary analysi
40、s, where the bearing losses and temperatures are calculated based on gear loadsand speeds effects (no other loss interaction);S a complete analysis where all the losses are taken into account (gears, bearings, oil splash).All the analyses take into account the preload/clearance case of the bearings
41、(as it has an important role onthe bearing friction itself).Preliminary analysisInthisfirststep,theimpactofthenewbearingdesignwasevaluatedonlybytakingintoaccounttheirownheat(not taking into account heat equilibrium from other losses). This first analysis is very fast to perform andallowsonetounderst
42、andthebearingbehaviortrends(withoutexternalinfluences). Italsogivesaroughindic-ationof theloss splitper shaftand potentialimpact ofa bearingdesign change. Inany case,it willnot giveanaccurate prediction of the real performance, as one doesnt take into account the overall heat equilibrium:such result
43、s will be more optimistic than the real case.The detailed results are presentedin Figure 8and Figure 9. Theeffect ofpreload isvery clearon eachshaft.Outer ring temperatureFigure 8. Detailed comparison of the outer rings temperature (energy efficient versus standard),depending on the preload, only wi
44、th bearing losses)9 11FTM23Total bearing power lossFigure 9. Comparison of the sum of the bearing losses (energy efficient versus standard) overthe preload range, only with bearing lossesInaverage,incomparisontothestandardbearingtype,theSKFenergyefficienttaperedrollerbearingdesign:S saves 60 to 120
45、Watts (power losses reduced by 13 to 15%);S runs4to10C cooler (each position runs cooler);S have a longer fatigue life in most cases, due to better lubrication conditions (higher Kappa): min lifeL10mn 100 000 hrs (following SKF rating life method).This first analysis indicates good trends: the 30% l
46、ess friction announced expected withSKF energyefficientbearingare convertedinto reducedouter ringtemperature andreduced friction,when appliedin thegearbox.It is time to analyze what it means when taking into account the complete system equilibrium.Complete analysis:The complete analysis was performe
47、d including:S the gear losses (according to ISO TR 14179 formulas);S the bearing losses (according to SKF advanced friction modeling tool);S the oil splash loss (according to ISO TR 14179 formulas).The full gearbox was modeled into SKF Orpheus tool (including the housing).Theresultsanalyzedwere:theb
48、earingfriction,thetemperatureonthebearing outerring andthe bearinglife.As discussed previously, it is interesting to notice that the gear loss becomes dominant in the highest loadcase studied. Below 66% full torque, bearing and oil splash represent still close to 60% of the loss (seeFigure 10).It ha
49、s to be noted that the gearbox is subjected to forced cooling when power exceeds 30% of the nominalload. SKF thermal simulation confirmed that a cooling was needed in such a case. Transient simulationshowedthatheatcanincreasetounrealisticvaluesifacoolingisnotapplied:at100%loadwithoutcoolingtheheat generated is so high that the simulation would predict calculated temperature 250C (see Figure 11).The housing thermal expansion leads to bearing additional load (depends on grounding). Under thecalculated assumption, 10 to 20 kN additional axial load ar
