1、07FTM13Influence of Grinding Burn on the Load CarryingCapacity of Parts Under Rolling Stressby: C. Gorgels, F. Klocke, and T. Schrder, RWTH AachenUniversity of TechnologyTECHNICAL PAPERAmerican Gear Manufacturers AssociationInfluence of Grinding Burn on the Load CarryingCapacity of Parts Under Rolli
2、ng StressChristof Gorgels, Fritz Klocke and Tobias Schrder, Laboratory for MachineTools and Production Engineering (WZL), RWTH Aachen University ofTechnologyThe statements and opinions contained herein are those of the author and should not be construed as anofficial action or opinion of the America
3、n Gear Manufacturers Association.AbstractThedemandforcontinuousimprovementconcerningeconomicefficiencyofproductsandprocessesleadstoan increasing cost pressure in manufacturing and design of power transmissions. Therefore the powerdensityofgearshastobeincreasedwhichleadstoademandforhighergearquality.
4、Inmoreandmorecasesthiscanonlybeachievedusinghardfinishingprocesses.Fromamanufacturingpointofview,processcostshave to be decreased as well which means that also a higher productivity is needed.The demand for higher gear qualities leads to an increased use of gear grinding. The necessary increase ofpr
5、oductivity takes the process closer to the risk of thermaldamages such as grinding burn on the gear flank.Theinfluenceofthermaldamageonthesetinbehaviourisneverthelesshardtojudgesothatdamagedgearsare often scrapped. This leads to increasing failure costs.Especially the lack of knowledge of the effect
6、 ofgrinding burn on the load carrying capacity ofgears leads tothe point that the same degree of damaged is judged differently by different companies. Therefore it isnecessary to do trials with thermally damaged parts in order to know how much a certain degree of thermaldamage influences the load ca
7、rrying capacity.The investigations described in this report are aiming to determine the load carrying capacity of parts underrolling stress. Therefore thermally damaged rollers are employed on a roller test rig, since with this analogyprocess the part geometry is easier to describe and easier to dam
8、age reproducibly using this analogyprocess.Copyright 2007American Gear Manufacturers Association500 Montgomery Street, Suite 350Alexandria, Virginia, 22314October, 2007ISBN: 978-1-55589-917-21Influence of Grinding Burn on the Load Carrying Capacity ofParts Under Rolling StressChristof Gorgels, Fritz
9、 Klocke, Tobias Schrder, Laboratory for Machine Toolsand Production Engineering (WZL), RWTH Aachen University of TechnologyIntroductionThe demand for a continuous improvement of theeconomic efficiency of products and processesleads to an increasing cost stress in design andmanufacturing of power tra
10、nsmissions. Thereforethe power density of each gear inside a transmis-sion has to be increased leading to a demand forhigher gear qualities. For more and more gearsthese quality demands can only be met applyinghard finishing processes.Therefore the use of gear grinding processes hasincreased over th
11、e past years. Especially in geargrinding the efforts taken for a rise in productivitylead to an increasing risk of inducing thermal dam-ages to the external zone of the gear flanks 1, 2,3, 4. The consequences of such thermal dam-ages ontheperformanceof agear are today hardlypredictable. Therefore ev
12、en slightly damagedgears are scrapped in many cases resulting in highcosts.Due to the lack of knowledge concerning the influ-ence of thermal damages on the performance of agear each company has different rules on the ac-ceptable limit of a thermal damage. Thereforebaselineknowledgeisneededinordertoh
13、aveaba-sis deciding how to treat gears with grinding burn.Today there is a good knowledge about the influ-enceof grindingburnonthetoothroot loadcarryingcapacityofgearswithgrindingburn4,5. Thede-termination of the influence on the flank load carry-ingcapacity thoughis alot moredifficult. Theprob-lem
14、is, that, e. g. in gear profile grinding, thermaldamages only occur locally. Therefore it cannot beensured that the damaged area is also the highlyloaded one. The interpretation of test results andespeciallytheirgeneralizationisproblematic6,7.Thereforethisreportaimstodeterminetheloadcar-rying capaci
15、ty of parts under rolling stresscompared to the degree of thermal damage. Theperformance of rollers reproducibly thermally dam-aged using a laser shall be determined on a rollertest rig. This kind of trial set up ensures the highlyloaded area on the part to be also the damagedarea. Theloadingandload
16、distributionontherollersis comparable to the situation of a gear tooth flank.Previous investigations: round robin teston nital etchingPrevious to the investigations described in thisreport WZL carried out a round robin test on nitaletchingtogether withthemembercompanies oftheWZL Gear Research Circle
17、. This test aimed to in-vestigatethereliabilityofthenitaletchingprocedureconcerning the detection of grinding burn.Different samples from two case hardened materi-als were sent to the participating companies. Oneach sample laser damaged tracks with a differentdegree of damage were applied. The task
18、of thecompanies was to undergo the samples their stan-dardnitaletchingprocedure,toevaluatethedegreeofgrindingburnandtodecideifagearwithsuchde-gree of damage would still be used or scrapped.Theresultshaveshown,thatgrindingburncouldbedetected using the nital etching procedure by all fif-teenparticipat
19、ingcompanies reproducibly. Stillsig-nificantdifferenceswerefoundconcerningtheeval-uation of the etching result. Especially concerningthe slightly damaged areas the opinions about thepossible use of the samples differed significantly.Thisshowsthatthereisahighuncertaintyabouttheinfluenceofdifferentdeg
20、reesofthermaldamageonthe performance of a gear.ObjectiveIn using gear grinding processes thermal damagessuch as grinding burn occur from time to time. Theimportant question after the detection of grindingburn on a gear is, if these damages have asignificant influence on its performance.The objective
21、 of this paper is to determine the loadcarryingcapacityofthermallydamagedpartsunder2rolling stress. Since the carrying out of such ex-aminations using real gears is problematic, rollersare chosen as an analogy. This way it can be as-sured that the highly loaded area is also the dam-agedarea. Thether
22、maldamageisinducedintotherollers using a laser. This ensures a reproducibleinduction of different degrees of thermal damage.The examined degree of thermal damage reachesfrom non-damaged parts as a reference up to arehardening zone as the worst case. Also twodegrees of a tempered zone have been exami
23、ned.Workpieces and execution of the trialsRoller test rig and geometryTherollergeometryisderivedfromthecontactcon-ditionsbetweentwogearflanks. Theroller radiusistaken from the curve radius of the gear flank invo-luteprofileinthecontact. Inthiscasethepitchcirclewaschosenasthebasisforthedeterminationo
24、ftheroller radius. The outer diameter for both rollers isD1=D2=42mm. Thetestroller,whichistobether-mallydamaged,hasacylindricalshape. Thecontraroller is bossed in order to prevent wear in the areaof the roller edges. In order to prevent edge pittingon the contra roller, a diameter in axial direction
25、 ofD3= 166 mm is applied for crowning. The furtherdata can be taken from Figure 1.Figure 2 shows the test rig and the technical data.The motor drives a transmission. Thedrive gear ofthe transmission is coupled with test roller and thedrivengearisconnectedtothecontraroller. Gohritzfound for this gear
26、 set that a pitting would be mostlikely to occur in the area of a negative slippage ofs = -24 % 8. Therefore a slippage between thetwo rollers of s = - 24 % is realized through agear-ing between the two mating rollers. The revolutionspeed on the drive side is n1= 2850 min- 1.Theoilissuppliedbyaninje
27、ctionsystemwithatem-perature of oil=55C and a tolerance ofoil=3 C. A typical oil for industrial gear boxapplications with typical additives is used. Thecontraforceinorder torealizetheHertzianstressisrealized by a hydraulic pressure system.Itisknowntodatethatarollertest stilldoes notfullyrepresent an
28、 exact analogy to a real gear flank.Therefore typically higher Hertzianstresses canbeapplied to rollers compared to gear flanks due togeometricaleffects. Stilltherelationbetweendiffer-ent roller e. g. for materialtestingstillcanbetrans-ferred from roller tests to real gears. Therefore areference wit
29、hin the roller tests is needed 9, 10.Figure 1. Geometry of gear flank and test roller3Figure 2. Twin roller test rigExternal zone of the rollers before set inWithintheexaminationsdescribedinthisreport,be-sides a non-damaged reference, three test rollersareused withdifferent degrees of thermaldamage.
30、Thethermaldamageisrealizedusingalaser. Inor-der toeliminate theinfluence of theoxide layer afterlaser damaging, a very small amount of stock hasbeen removed from the roller surface by grinding.The surface of the thermally damaged samples,after the laser process and after grinding and nitaletching, c
31、an be seen in Figure 3.Four different variants are to be examined:1. non-damaged(reference)2. slightly tempered zone (PLaser= 700 W)3. strong tempered zone (PLaser= 800 W)4. re-hardening (PLaser= 900 W)Figure 3. Test rollers - laser damaged and nital etched4One important point in the analysis of the
32、 laserdamaged rollers is to make sure, that the induceddamage is comparable to typical grinding burn in-ducedbyagearprofilegrindingprocess. Inordertoanalyzethecharacteristicsofthedamagethemate-rialstructure(Figure4),thehardnessprofile(Figure5) and the residual stress profile (Figure 6) havebeen anal
33、yzed.Theanalysis shows that theoccurrenceof thether-mal damage shows a good correlationto therealityon a gear flank. The structural damage goes into ahigh depth from surface. The tempered sampleshave a decrease in hardness close to the surfaceand tensile residual stresses. The re-hardenedsample show
34、s a very high hardness close to thesurface layer decreasing inside the tempered zonebelow the re-hardened area. Also very high com-pressive stresses can be found near the surfacechanging into very high tensile stresses in adistance from surface of z = 200 mm.Figure 4. Material structure of damaged t
35、est rollersFigure 5. Microhardness profile of selected test rollers5Figure 6. Residual stress profile of selected rollersIt can be concluded, that the results of the thermaldamage using a laser show a very goodcorrelationtothermallydamagedgears. Thereforeagoodcor-relationbetweentheloadcarryingcapaci
36、tyfromtheroller tests and real gears can be expected.Test rig trials using thermally damagedrollersIn order to determine the load carrying capacity ofthermally damaged rollers test rig trials have beencarried out. The objective of the examinations is todetermine the sustainable load cycles of damage
37、drollers under high loading compared to a non-damaged reference. The focus is on the sustain-able load cycles as well as on the occurring failuremechanisms.Running time of rollers depending on thedegree of thermal damageIn order to examine the load carrying capacity ofthermally damaged rollers, dama
38、ged and non-damaged rollers are tested on a twin disc test rig.The cylindrical test roller on thetest rigis incontactwith the bossed contra roller. The normal force be-tween the rollers is Fn= 8708 N resulting in a Hert-zianstressofHertz= 2800 N/mm. Theinjectedoilis atypicallubricant for gear box ap
39、plications (ShellSpirax 80 W) at a temperature of l=55C with atolerance of = 3C. The test roller is rotatingwith a rotational speed of n = 2850 min- 1and thesliding between the two rollers amounts tos=-24%.In order to determine the fatigue life of the differentrollers the maximum load cycles dependi
40、ng on thedegree of thermal damage are shown in Figure 7.After reaching N = 50,000,000 load cycles the limitfor the fatigue strength is reached. In order toachieve the goal the examinations have beencarriedoutintheareaoffatiguelife.Thiswayacom-parison between the differently thermally damagedrollers
41、can be carried out.The non-damaged reference has an average sus-tainablenumberofloadcycles,untilapittingoccurs,of N = 38.8106andis thereforeclosetothelimitforfatigue strength. Thefailure mechanism is apittingontheslidingsurface. Therollers witha slight tem-pering (PLaser= 700 W) fail after an averag
42、e ofN = 35.1106load cycles. The results show that aslight tempering only leads to a small reduction ofsustainable load cycles.6Figure 7. Fatigue life depending on thermal damageThe two variants with a stronger structural thermaldamage show a drastic reduction of fatigue lifethough. The rollers damag
43、ed with a laser power ofPLaser= 800 W with a strong tempering are failingafter anaverageof N = 5.4106load cycles. Thisisa reduction of fatigue life of more than 80 %comparedtothenon-damagedreference. Alsothefailuremechanism is changing. Apittingoccursnotonly on the test roller but also on the contra
44、 roller.The rollers damaged with a laser power ofPLaser= 900 W witha re-hardeningare failingafteran average fatigue life of only N = 3.3106load cy-cles. The failure mechanism is changing againcomparedtothenon-damaged reference. Thepit-tingonthetest roller does not occur inthemiddleofthe sliding surf
45、ace but on its side. In this area thethermal damage changes from re-hardening totempered zone.Analysis of the wear mechanismsThe wear phenomenon to be found on the test roll-ers after set in is displayed in Figure 8. The nondamagedroller andthetwo rollers with atempering(PLaser= 700 W and 800 W) sho
46、w a pitting in themiddle of the contact area. The roller with a re-hardeninginsteadshowsapittingontheedgeofthecontact area.Figure 8. Comparison of pitting depending on degree of thermal damage7Anexplanationforthemovementofthepittingoutofthemiddle canbe givenif Figure9istaken intoac-count. The materi
47、al structure in the area of the pit-ting is shown. The middle of the contact surface ismarked where the hardness profile was measured.The hardness profile shows a high hardness nearthe surface and the area with a low hardness (tem-pered zone) below. The pitting instead occurs onthe side of the conta
48、ct surface.Thelineshowingthestartof theannealingzonebe-low the re-hardening shows that the middle of thepittingis intheareaof thechangeover from there-hardening to the tempered zone on the surface.Thisshowsthatthere-hardenedareabreaksoffthetempered zone below due to the high tensilestresses in this
49、area (cp. Figure 6).InFigure10scanningelectronmicroscopepicturesof the pitting of all four variants are shown. Thenon-damaged and the slightly damaged variantshow a pitting without cracks around. The stronglytempered roller (PLaser= 800 W) shows a crackaround the pitting. In this area the pitting wouldhaveenlargedifrunningmoreloadcycles. Inoppo-site to this below the pitting on the roller with a re-hardened zone cracks transverse to the sliding di-rectioncanbefound. Thisshowsthatthesurfaceisstrongly affected and shatte
