1、09FTM01AGMA Technical PaperInfluence of theResidual StressesInduced by HardFinishing Processes onthe Running Behaviorof GearsBy Dr. V. Vasiliou, C. Gorgels, andDr.F.Klocke,RWTHAachenUniversityInfluence of the Residual Stresses Induced by Hard FinishingProcesses on the Running Behavior of GearsDr.Vas
2、iliosVasiliou,ChristofGorgelsandDr.FritzKlocke,RWTHAachenUniversityThe statements 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.AbstractLow noise and high load carrying capacity are two
3、 important characteristics of competitive powertransmissions.Thechallengeinthedevelopment,designandmanufacturingofthesepowertransmissionsisto meet these requirements economically. One of the ways to meet both of these requirements is through aprocess known as hard finishing. There are various types
4、of hard finishing and it is important to know whichprocess produces which requirement.Theaimofthisresearchprojectistoinduceresidualstressesintheedgeoftheworkpiecesbydifferenthardfinishing processes and to analyze their influence on the durability of the gears. The tested gears aremanufactured by pro
5、file grinding, gear honing and generating grinding. The gear deviations and the finishquality have to be comparable. Through this the influence of the residual stresses on the durability can beanalyzed independent from the geometrical conditions. The presentation will show the results of the loadcar
6、rying capacity tests depending on the values of the residual stresses.Copyright 2009American Gear Manufacturers Association500 Montgomery Street, Suite 350Alexandria, Virginia, 22314September 2009ISBN: 978-1-55589-954-73Influence of the Residual Stresses Induced by Hard Finishing Processes on theRun
7、ning Behavior of GearsDr. Vasilios Vasiliou, Christof Gorgels and Dr. Fritz Klocke, RWTH Aachen UniversityMotivation and objectiveNoise behavior and load carrying capacity areimportant demands on power transmissions. Thechallenges in development, design and manufac-turing of power transmissions are
8、to meet theserequirements economically and efficiently. Theselection of a capable hard finishing process for aspecificgeardependsonthedesignofthetransmis-sion. Therefore it is important to know whichrequirements can be met by which hard finishingprocess.To achieve high durability, high surface quali
9、ty, lowgear deviations, high gear quality and high residualcompressive stresses are required. So far theinfluenceofgeardeviationsandfinishquality ontheload-carrying capacity of gears have beenanalyzed. However, theinfluenceofmanufacturinginduced residual stresses is yet unknown. Further-more there i
10、s a lack of knowledge regarding howhightheresidualstressescausedbymanufacturinghavetobetoinfluencetheloadcarrying capacity ofgears in a positive way.The aim of the research project described in thispaper is to analyze the influence of residualstresses induced by different hard finishingprocesses on
11、the durability of gears. The testingpinions arehardfinishedby gearhoningandgener-ating grinding. Both processes are set up tocomparable finish quality. Hereby the influence ofthe residual stresses on the durability can beanalyzed independent from the geometrical condi-tions. In this paper the change
12、s in residual stressand surface roughness during the test aredescribed.ApproachFigure 1 shows theapproach of the project. At firstresidual stress measurements are carried out atindustrially manufactured gears in a round-robintest. The intent is to determine the spread of theresidual stresses on gear
13、s manufactured by thesame hard finishing method. At the same timedifferences between the level of residual stressesresultingfromdifferenthardfinishingprocessescanbe determined. The hard finishing methods usedare generating gear grinding and gear honing.WZLFigure 1. Approach4To analyze the changes in
14、 residual stress andsurface structure of differently hard finished gearsduringtheloadcycles, testingpinions aremanufac-tured by using generating gear grinding and gearhoning, parallel to the industrial round-robin test.Thetesting pinions are mountedin aback-to-backgear test rig. At the beginning and
15、 after a definednumber of load cycles the surface is analyzed bymeans of residual stress, roughness and hardnessmeasurements. The testing gears used areexclusively manufactured by profile grinding.Round-robin test to determine residualstresses due to hard finishingDue to different workpiece material
16、s, toolspecifications and process parameters, differentresidual stresses are measured for workpiecesmanufactured by the same hard finishing process.Therefore the following questions result: How higharetheresidualstresses andthescatter ofindustri-al produced gears? Does a specific hard finishingproce
17、ss lead to a certain residual stress level?Furthermore it has to be answered whether theresidual stress generated in a testing pinion iscomparable to residual stress generated in gearsmanufactured in industry and whether the residualstresses of the testing pinion are representative forthe industrial
18、 manufactured gears.To answer these questions residual stresses aremeasured using X-ray diffraction at the testingpinions and at the industrial manufactured gears.These automotive gears are hard finished by gearhoning or generating gear grinding. The toolsconsist of corundum (aluminium oxide) and ar
19、eceramic bound. All in all, 25 gears are analyzed.The residual stresses are measured on the toothflank at the position of the pitch circle at a depth ofaB=4.5mm.Theresults arepresented inFigure 2. In this figurethe residual stresses are shown in the profiledirection (above) and flank direction (bott
20、om). Theresidual stresses induced by generating geargrinding are shown on the left side. Those inducedby gearhoningarepresentedontherightside. Thehelical striped show the residual stresses of thetestingpinionmanufacturedunderlaboratorycondi-tions.In all for both directions, profile and flank, onlyre
21、sidual compressive stresses are determined.Furthermore the residual compressive stresses inthe profile direction are higher than in the flank di-rection. The difference between the maximum andminimum residual compressive stress measuredfor ground gears is P= 237 MPa in profile direc-tion and F= 254
22、MPa in flank direction. Forhoned gears the difference is P= 274 MPa inprofile direction and F= 163 MPa in flank direc-tion. Noticeable is that the scatter for honed gearsin the flank direction is smaller.WZLFigure 2. Measurement of residual stresses in round-robin test5The scatter of gears which wer
23、e produced in thesame batch is significantly smaller. This is shownby analyzing gears produced in two differentbatches. The difference between maximum andminimum residual stresses in the profile directionwithin a batch is P1/2= 70 MPa for generatedgroundgears. Intheflankdirectionthedifferenceiswithi
24、n the fist batch F1= 143 MPa and within thesecond batch F2= 91 MPa. For the honedbatches the difference is within the fist batchF1= 47 MPa and within the second batchF2= 64 MPa in the flank direction. In the profiledirection the difference is within the fist batchP1= 52 MPa and within the second bat
25、chP2= 143 MPa.Summarizing,acleartendencybetweenhonedandground gears is noticeable. The honed gears havehigher residual compressive stresses than thosegears manufactured by generating gear grinding.This is valid in the flank direction as well as in theprofile direction. The average residual stresses
26、inthe profile direction in honed gears isP= 545 MPahigher thaningroundgears. Intheflank direction the average value of honed gears isF= 418 MPa higher than in the ground gears.The measured residual compressive stresses intesting pinions are within the range of dispersion.So the testing pinions can b
27、e used as a specimenfor further investigations on fatigue.Hard finishing of the testing gearsThe testing gears are designed according to truckgears as shown in Figure 3. The testing pinionshave z =25 teeth, a module of mn= 3.5 and anormalpressureangleofn=20,thehelix angleof = 19.3, a face width of b
28、 = 18 mm and a tipdiameter of da= 101 mm. The testing gears havez = 36teethandatipdiameterof da= 138mm. Thegearsaremadeofcase-hardenedsteel. Thegearssurface hardness is approximately 58 HRC with ahardening depth of Eht550HV= 1 mm.The testing pinions are hard finished by generatinggrinding and gear h
29、oning. To ensure the compara-bility of the two processes, the testing pinions aretaken from the same material batch and have thesamestock. The testinggears arehard finishedbyprofile grinding. The manufacturing process chainaswellastheprocessparametersofmanufacturingthe testing pinions are shown in F
30、igure 4. In bothprocesses, ceramic bound tools with 30% sinteredaluminum oxide are used.For comparable results it is required to havecomparable profile and flank deviations on theground and honed testing pinions. Figure 5 showsthat both processes are comparable in theirdeviations. In Figure 5 the me
31、asurements of thetwist at aright flank of agroundandahonedtestingpinion are shown.WZLFigure 3. Data of testing gear and pinion6WZLFigure 4. Process chain of testing pinionsWZLFigure 5. Comparison of the profile and tooth traces of the pinionsChange in state of the edge zone of thegears due to loadTo
32、describethechangeoftheworkpiecesurfacetheground and honed testing pinions are mounted in aback-to-back test rig with profile ground testinggears. After a defined number of load cycles thetesting pinions were analyzed. For each versionsurfaceroughness andsurfacestructures werede-termined before and a
33、fter mounting. To determinethe residual stresses, teeth of the testing pinionswereerodedandresidualstresscurvesweregener-ated. Figure 6 shows the test rig and testingparameters. The load cycle corresponds to thenumber of load changes of the testing pinion.7WZLFigure 6. Design of experimentalThe inve
34、stigations on constitutional changes underload takes place on a back-to-back test rig(centerdistance a = 112.5 mm) according to DIN standard51354 with a closed power circuit. Figure 7 showsperformancedataandthefunctionalprincipleofthetest rig. The testing pinion and testing gear aremountedontwoparal
35、lelshafts,whicharelinkedtoatransmissiongear withthesameratio. Thehydrau-lic twisting motor applies a defined torque. Theexperiments are carried out under splash lubrica-tion (lubricant ISP VG100/Spirax 80 W) with an oiltemperature of toil=90C.WZLFigure 7. Back to back test rig8Figure 8 shows the cha
36、nge of surface roughnessunder load. The upper chart shows the averagesurfaceroughnessRaandthelowercharttherough-ness depth Rzlogarithmic plotted against thenumber of load cycles. For each testing gear thevalues is determined at initial state and afterdismounting. The allocation between initial value
37、and final value is done through the marking of thedatacheckboxes. AsLeubealreadypointedoutthesurfaces smoothing depends on the initialroughness. 1Thetrendsofaveragesurfacefinishandroughnessdepth are comparable. Concerning an initialsurfaceroughness intherangeof Rz=3.5mmupto4.5 mm it is reduced by 11
38、% already after twohoursrun-in period with a low torque of M2= 200 Nm.After NL= 10.5 million load cycles, the roughnessdepth Rzof this versionis reducedby 50%. Incon-trast the constitutional change of testing gears withan initial surface roughness between Rz=1.3mmand 2.3 mm is not significant.Thecha
39、ngeofsurfaceroughnessdepends inthesecases not on the hard finishing processes, but onthe process parameters and the tools used. Ofcourse by also gear honing a lower surfaceroughness can be obtained. However it has to beensured that for upcoming load-carrying capacityinvestigations, the roughness has
40、 to becomparable.The influence of the number of load cycles on thesurfacetextureisexaminedusinglight-microscopephotographs and topographic measurements. Thechange of the surface texture after a number ofcycles of NL= 10.5 million are shown in Figure 9(generated grinded variant) and Figure 10 (honedv
41、ariant). The upper row shows the condition afterhard finishing and the lower row the condition afterNL= 10.5 million load cycles. On the left side arerepresented the light-microscope photographs ofthe flanks. The middle field represents the surfacetexture for a measuring field of 22mmas3D-topography
42、. On the right side a 2D-profile line isshown, which is taken along a profile line out of the3D-topography. Because different measurementsystems are used, the surface roughness showninFigure 8 is not comparable to Figures 9 and 10.InFigures9and10thecutstructures resultingfromthe different hard finis
43、hing processes can berecognized. Whilewithgeneratinggeargrindingthegraintrajectories runparalleltotheflankline, onthehoned flanks the nondirectional surface structurecanberecognized. BycomparingtotheinitialstateforbothvariantsaflatteningofthesurfacestructurecanbeseenafterNL= 10.5 million load cycles
44、.WZLFigure 8. Influence of load cycle on roughness9WZLFigure 9. Influence of load cycle on topography of generated gear grounded testing pinionWZLFigure 10. Influence of load cycle on topography of gear honed testing pinionFigure 11shows theresidualstressesof thetestingpinions measured in profile an
45、d tooth direction.First the residual stresses before the hard finishingprocess are measured. The circular mark showsthe residual stress before hard finishing. Theresidual stress after hard finishing is marked byrhombus while the square marks show the residualstress after NL= 10.5 million load cycles
46、.Comparing to the stress level after heat treatmenthoninginduces compressiveresidualstress up toadepthofaB=8mm,generatinggrindinguptoadepthof aB=4,5mm. After NL= 10.5 million load cycles,the residual compressive stress in the profiledirection of the honed pinion has been decreasedfromP= -1030toP= -6
47、18MPa. Thestresslevelin the ground pinion has maintained at aboutP= -600 MPa. In the flank direction, the stresslevel of the ground pinion increases to P= -448MPa, whilethe honedpinion maintains at this level.10In Figure 12 the residual stresses at a depth ofaB=4.5mm (left) and aB=8mm (right) for th
48、edifferent numbers of load cycles are represented.At a depth of aB=4.5mm, a change of the residualstresses canberecognizedafter theinitialphaseofNL= 259,000 load changes. In the profile directionthe residual stresses of the ground testing pinionsdo not change significantly, while the honed testingpi
49、nions show a decrease of approximately 40%. Intheflankdirectiontheresidualstressesofthehonedvariants reduce about only 8%. In the groundvarianttheresidualstressesincreaseaftertheinitialphase by approximately 52% and afterNL= 10.5106by 109% compared to the statebeforefinishing. Whetherthechangeswereduetodifferent surface roughness cannot be determinedclearly. Further investigation is necessary.However, past investigations of 2 show that achange of residual stresses and the surfaceroughness are independent.WZLFig
