1、07FTM07Grinding Induced Changes in ResidualStresses of Carburized Gearsby: R. LeMaster, B. Boggs and J. Bunn, University of Tennesseeand C. Hubbard and T. Watkins, Oak Ridge National LaboratoryTECHNICAL PAPERAmerican Gear Manufacturers AssociationGrinding Induced Changes in Residual Stresses ofCarbu
2、rized GearsRobert LeMaster, Bryan Boggs and Jeffrey Bunn, University of Tennessee andCamden Hubbard and Thomas Watkins, Oak Ridge National LaboratoryThe statements and opinions contained herein are those of the author and should not be construed as anofficial action or opinion of the American Gear M
3、anufacturers Association.AbstractCarburizing is a commonly used method for increasing the strength and wear resistance of gearing. Asignificantbenefitof the carburization process is thatcompressive residualstresses are developed near thesurface due to phase transformations that occur during the post
4、 carburization heat treatment steps. Aftercarburization and heat treatment it is necessary to finish the gear by processes such as grinding or skiving.These finishing processes develop the precise geometric form required while improving the surface finish.Finishing processes change the residual stre
5、ss imparted by carburization and subsequent heat treatment.Residual stresses change during the finishing process due to the removal of material and stresses inducedfrom the machining operation.Thispaperpresentstheresultsofastudyperformedtomeasurethechangeinresidualstressthatresultsfromthe finish gri
6、nding of carburized gears. Residual stresses were measured in five gears using the x-raydiffractionequipmentintheLargeSpecimenResidualStressFacilityatOakRidgeNationalLaboratory.Twoofthe gears were hobbed,carburized,quenched and tempered,butnotfinished.The remaining three gearswere processed similarl
7、y, but were finish ground. The residual stresses were measured at 64 differentlocations on a tooth from each gear. Residual stresses were also measured at fewer points on other teeth todetermine the tooth-to-tooth variation. Tooth profile measurements were also made of the finished andunfinished gea
8、r samples.Theresultsshowafairlyuniformandconstantcompressiveresidualfieldinthenonfinishedgears.Therewasasignificantreductionintheaverageresidualstressmeasuredinthefinishedgears.Additionally,therewasasignificantincreaseinthevariabilityoftheresidualstressthatwasintroducedbythegrindingprocess.Largevari
9、ationswereobservedinboththelateralandlongitudinaldirectionsonatoothsurface.Analysisofthedatasuggests a linear relationship between the change in average residual stress and the amount of materialremoved by the grinding process.Copyright 2007American Gear Manufacturers Association500 Montgomery Stree
10、t, Suite 350Alexandria, Virginia, 22314October, 2007ISBN: 978-1-55589-911-01Grinding Induced Changes in Residual Stresses of Carburized GearsRobert LeMaster, Bryan Boggs, and Jeffrey Bunn, University of Tennesseeand Camden Hubbard and Thomas Watkins Oak Ridge National LaboratoryIntroductionCarburizi
11、ng is a commonly used method for in-creasing the strength and wear resistance of gear-ing. A significant benefit of the carburizationprocess is that compressive residual stresses aredeveloped near the surface due to phase trans-formations that occur during the post carburizationheat treatment steps.
12、 After carburization it is nec-essary to finish the gear by processes such asgrinding or skiving. These finishing processesdevelop the precise geometric form required whileimproving the surface finish. Finishing processeschange the residual stress imparted by carburiza-tion and subsequent heat treat
13、ment processes.These changes are due to the removal of materialand the associated rebalancing of the residualstresses and the introduction of near surfaceresidual stresses by the machining operations.A grinding allowance is used to specify the amountofmaterialtobeleftonamachinedgearpriortoheattreatm
14、ent. This excess material and any materialassociated with a geometry change during heattreatment are removed by the finishing process.The magnitude of this grinding allowance will affectthe strength, fatigue life, and wear resistance of thefinished gear because of its relationship tochangesin the re
15、sidual stresses. Removal of the excessmaterial will also remove any retained Austenitelocated in the layer removed.This paper presents the results of a study per-formed to measure and quantify the change inresidualstressthatresultsfromthefinishgrindingofcarburized gears. Residual stresses weremeasur
16、ed in five gears using x-ray diffractionequipment in the Residual Stress User Center atOak Ridge National Laboratory. Two of the gearswerehobbed,carburized,quenched andtempered,but not finished. The remaining three gears wereprocessedsimilarly,butwerefinish ground. There-sidual stresses were measure
17、d at sixty-four loca-tions on a randomly selected tooth from each gear.Residual stresses were also measured at fewerpoints on other teeth to determine the tooth-to-tooth variation. Tooth profile measurements werealso made of the finished and unfinished gearsamples.Test gearsThe three finish ground s
18、amples were designatedas Finished 1, 2, and 3. The two remaining unfin-ishedsamplesweredesignatedasUnfinished1and2. Each gear had twenty-five teeth, a diametralpitch of 4 teeth/inch, a pressure angleof twenty de-grees, full radius fillets, no addendum modification,and a face width of 0.75 inch. The
19、gears were flatwithnoribs,rims,orotherweightreductionfeatures(Figure 1).Figure 1. A typical finished gear. This gear isdesignated Finished 1.Themeasurementofresidualstressesingearteethusing x-ray diffraction is complicated by the curva-ture ofthe involuteand trochoidgeometries andthepotentialforinte
20、rferenceoftheincidentordiffractedbeambyadjacentteeth. Thesize ofthegearsusedin this study was chosen so that the residual2stresses could be measured over most of the toothsurface.Most of the residual stresses measuredwere in thelongitudinal direction of the gear tooth (Figure 2). Afew residual stres
21、s measurements were also madeinthelateraldirection. Thesixty-fourlocationsonatooth from each sample where residual stresseswere measured are shown in Figure 3. There areeight lateral locations associated with each radius.The lateral locations are spaced 0.079 inch apart.Residual stress measurements
22、were not made atthe critical bending stress location in the fillet. Thiswas due to the high curvature in the fillet area andinterference with the incident or refracted beampath from adjacent teeth.Key1 Dedendum2 Clearance circle3 Pitch circle4 Addendum circleFigure 2. Longitudinal and lateral toothd
23、irections on tooth.The gear blanks for each sample were taken fromthe same length of 8620H bar stock. The stepsused in the fabrication of the samples are listed inTable 1. The time at temperature for the normalize,stress relief, and defuse steps was basedon 1hourper inch of thickness. The carburizat
24、ion step wasdone using an 80-90% natural gas derived endo-thermic gas atmosphere. Test slugs were pulledduring the carburization step to verify an effectivecase depth of 0.030 inch. The final surface hard-ness was determined to be within the range of58-62 HRC. The finish grinding was done using avit
25、rified alumina grinding wheel on a CNC grinder.Figure 3. Radial and lateral locations wherethe residual stress measurements were made.Table 1. Fabrication steps.1. Rough machine2. Normalize (1740 F)3. Stress relief (1250 F)4. Finish machine5. Carburize (1650 F)6. Defuse (1550 F)7. Oil quench (135 F)
26、8. Temper (450 F)9. Finish grindTheprofilesofeachsampleweremeasuredusingaCoordinate Measuring Machine (CMM) at OakRidge National Laboratory. The profiles weremeasuredatthreelaterallocationsoneachsample.There was no discernable difference between thethree lateral measurements for each sample whenthe
27、measurements were superimposed. Theprofiles measured at the first lateral location on thefive samples are compared in Figure 4. There wasvirtually no discernable difference between themeasured profiles of the samples designatedUnfinished1andUnfinished2. Therewas alsolittlenoticeable difference betwe
28、en the samples desig-natedFinished1andFinished2. However,thegeardesignatedFinished3hadnoticeablymorematerialremoved at the tip than did gears Finished 1 andFinished 2. At the pitch circle all of the finishedsamples were virtually the same.3Figure 4. Comparison of the measurementprofiles of the finis
29、hed and unfinished gears.Figure 5 shows the grind depth versus radius foreach of the finished samples. The grind depthreported is the perpendicular distance from the un-finishedprofiletothefinishedprofile(Figure6). Theincreased tip relief observed in the sampledesignated Finished 3 is quite noticeab
30、le. With theexception of the tip relief found in Finished 3, thegrind depth is similar for all radii greater than theform radius. The material removed at the pitchcirclebygrindingrangedfrom0.0082to0.0085inch(0.208 to 0.216 mm).X-ray diffraction measurementsResidual stress measurements were made usin
31、gthe Model 1600 TEC diffractometer in the ResidualStress User Center at Oak Ridge National Labora-tory. The residual stress measurements involvedmeasuring the inter-atomic spacing (d-space) be-tween atoms for the (211) crystal plane at differentx-ray beam incident angles () 1. The measuredd-spaceist
32、heaveragevalueforagroupofproperlyoriented grains near the irradiated surface. The re-sidual stress was determined using the sin2 tech-nique 2. In this method d-space is plotted as afunction of sin2. The y-intercept of the plot wastaken as the unstrained d-space (d0) withtheslopebeing proportional to
33、 the residual stress. A 2 mm(0.079 inch) diameter collimator, vanadium filter,andKradiationfromachromiumx-raytargetwereused. Figure 7 shows a picture of a portion of thediffractometer and one of the samples mounted inthediffractometer. Onaverage,ten-angleswithatwo-degree oscillation were used at eac
34、h radiallocation. As seen in Figure 7, black electrical tapewas used to cover neighboring teeth to eliminateany radiation scattered from them. Figure 8provides an example of a typical d-space versussin2 plot.Figure 5. A comparison of the grind depthversus measurement radii for the threefinished gear
35、s.Figure 6. Grind depth is the differencebetween the unfinished and finished toothsurfaces measured normal to the finishedgear tooth surface.4Figure 7. (a) X-ray source and detector portion of TEC model 1600 diffractometer; (b) Unfinished2 mounted in x-ray diffractometer.Figure 8. An example sin2 pl
36、ot used to determine the residual stress.The slope is proportional to the residual stress.5Residual stress data for unfinishedsamplesTheresidualstresscomponentactinginthelongitu-dinal direction for the gears designated asUnfinished1andUnfinished 2is shownin Figures9and 10. Each line is associated wi
37、th a specific ra-dius. There are eight data points per line. Table 2givestheaverageandstandarddeviationofthelon-gitudinal residual stress measured on therandomlyselectedtoothforeachgear. Theaverageandstan-dard deviation values given in Table 2 are for thesixty-four measurement locations on the tooth
38、.The average residual stress for all sixty-fourlocations on the tooth from sample Unfinished 1 is185 ksi and 150 ksi for Unfinished 2. The 35 ksidifference in the average longitudinal residualstressledtothemeasurementofthelongitudinalre-sidual stress at a similar location on several addi-tional teet
39、h of gear Unfinished 2. The tooth-to-tooth variation in the longitudinal residual stress(measured at a radius of 3.244 inch (82.4 mm) andlaterallocationof0.315inch(8mm)isshowninFig-ure 11. The tooth-to-tooth variation ranges from138 ksi to 191 ksi with an average value of 160 ksi.The average longitu
40、dinal residual stress measuredon the tooth from both Unfinished 1 and Unfinished2fall withinthis tooth-to-toothvariation. Variationsin residual stresses can be caused by non-homo-geneous chemistry and microstructurein the mate-rial as well as non-uniform furnace heating, carbonpotential, and quenchi
41、ng rates.Table 2. Statistical properties of longitudinalresidual stress measurements.GearAveragelongitudinalresidualstress (ksi)Standarddeviation oflongitudinalresidualstress (ksi)Unfinished 1 185 13Unfinished 2 150 7Finished 1 116 15Finished 2 108 15Finished 3 110 22Figure 9. Longitudinal component
42、 of residual stress measured in gear Unfinished 1.Average = 185 ksi, Std. Dev. = 13 ksi6Figure 10. Longitudinal component of residual stress measured in gear Unfinished 2.Average = 150 ksi, Std. Dev. = 7 ksiFigure 11. Tooth to tooth variation in the longitudinal residual stress measured at a radius
43、of3.244 inches and a lateral position of 0.315 inches. Average 160 ksi, Std. Dev. = 17 ksi.7The residual stress in the lateral direction was alsomeasured ata fewlocations onthe unfinishedsam-ples. The residual stresses were in all cases com-pressive and approximately equal to the residualstress meas
44、ured in the longitudinal direction. Thisindicates a biaxial stress field in which the normalstresses were approximately equal. No shearstresses were measured. An exampleof thestressstate for one of these points is shown in Figure 12.Figure 12. The unfinished gears exhibited abiaxial stress state wit
45、h equal lateral andlongitudinal compressive stresses.Residual stress data for finished samplesThe longitudinal residual stress measured on arandomly selected tooth on gears Finished 1,Finished 2, and Finished 3 is shown in Figures 13,14, and 15. A comparison of thesefigures with Fig-ures 9 and 10 sh
46、ows that the finished gears havemore variation than the unfinished gears. Table 2givestheaverageandstandarddeviationofthelon-gitudinalresidualstressforeachgear. Theaveragelongitudinal residual stress in the finished gears isapproximatelythesamewithamaximumdifferenceof 8 ksi. The standard deviations
47、of the data areapproximately the same for gears Finished 1 andFinished 2. The standard deviationof the longitudi-nal residual stress in Finished 3 is 22. ksi, whichis considerably greater than that for Finished 1( 15.)andFinished2( 15.). Thislargerstandarddeviation is due in large part to the lower
48、residualstresses measured at locations having a radius of3.322 and 3.283 inches (bottom two lines of data inFigure 15). These lower residual stresses areexpected due to a larger amount of material havingbeen removed at the tooth tip.Figure 13. Longitudinal component of residual stress measured in ge
49、ar Finished 1.Average = 116 ksi, Std. Dev. = 15 ksi.8Figure 14. Longitudinal component of residual stress measured in gear Finished 2.Average = 108 ksi, Std. Dev. = 15 ksiFigure 15. Longitudinal component of residual stress measured in gear Finished 3.Average = 108 ksi, Std. Dev. = 22 ksi9The residual stress in the lateral direction was alsomeasured at a few locations on the finished sam-ples. Unlike the unfinished gears, the lateralresidual stresses measured in the finished gearswere different thanthose meas
