1、91 FTM 7AVLow Cycle and Static Bending Strength ofCarburized and High Hardness ThroughHardened Gear Teethby: W. Pizzichil, Philadelphia Gear Corp.American Gear Manufacturers AssociationITECHNICAL PAPERLow Cycle Fatigue and Static Bending Strength of Carburized and High HardnessThrough Hardened Gear
2、TeethW. Pizzichil, Philadelphia Gear CorpTheStatementsandopinions containedherein arethoseof the authorandshouldnotbe construedas an official actionoropinion of the American Gear Manufacturers Association.ABSTRACT:A summary of the testing methodsemployed and the results generated for unidirectional
3、andreverse bending tests ofvery coarse and medium pitch gear teeth is presented. Actual measured stresses were comparedwith FEM theoreticalstresses and AGMA stress numbers.The purpose of this testing was to evaluate which type hardening method would yield a gear tooth that could carry thehighestload
4、 withoutcatastrophicbreakage failure ina single,or very low cyclicload application.This testing simulatedtheoutput pinion and aplanet gear for ajack-up geardriveused on oil drilling platforms.Therewere three separatetestsconducted over a period of time.The first testwas performed staticallyon an act
5、ual 12inchcircnlarpitchpinionloadedagainstarack. A second statictestutilized a half scale model of a 12CP gear tooth fillet loaded to failure in a single cycle and also cycled to failure atextremely high loads. The thirdtest was a reverse bendingtest simulating a planet of a planetary gearused in th
6、e finalreduction stage.The last two tests were accomplished simultaneouslyfor both through hardened and carburizedtest pieces. The resultsdemonstrated the benefits of therelatively ductilethroughhardenedgears and theirability tolocallyyield and dismbutestresses without catastrophicallyfailing in thi
7、s type of application.Copyright 1991American Gear Manufacturers Association1500 King Street, Suite 201Alexandria, Virginia_22314October, 1991ISBN: 1-55589-604-9LOW CYCLE FATIGUE AND STATIC BENDING STRENGTH OFCARBURIZED AND HIGH HARDNESS THROUGH HARDENED GEARTEETHWilliam Pizzichil, P.E.Executive Vice
8、 PresidentPhiladelphia Gear Corp.King Of Prussia, Pa., 19406INTRODUCTION:The particular application for thegear drives that were tested is toelevate off shore drilling rig platforms(jack-ups) above sea level once theyhave been floated to the desireddrilling location, see Figure I. Thedrives are fixe
9、d to the deck of thevessel. They have output pinions whichmesh with racks that are attached tothree moveable triangular legs at eachcorner of the vessel. The legs arelowered to the ocean floor and theplatform is raised to the desiredheight.COMMENTThis paper presents the results ofactual testing perf
10、ormed on large scale,costly test pieces. It is recognizedthat there are statistical variationsassociated with this type of testingthat make absolute conclusions un-realistic. There is no attempt made tocharacterize the metallurgical phenomenathat governs the behavior.LITERATURE SURVEYThe literature
11、reveals that work FIGURE i. OFFSHORE DRILLING RIGhas been carried out evaluating the PLATFORM (JACK-UPS)performance of through hardened racksfor jack-ups by Honda (1,2,3). Therealso is extensive work available TEST 1relating to the failure of case hardenedgears. References (4,5,6,7,8) represent SING
12、LE CYCLE BENDING FAILURE TEST (12 CPa sampling of this literature. The Actual Pinion).testing presented in this paper wasperformed to gather data on very large TEST OBJECTIVES AND METHODcarburized gears, where very little The main objective of this testingtesting has been performed, was to determine
13、 the single cycle,- i -Fultimate fracture strength of a 7 tooth,12 CP, 8.5 inch face width, carburized, IIspur pinion, loaded at the highest point |_of single tooth contact.This test was necessary to _determine the maximum holding capabilityof the pinion supporting the jack-uprig. A second objective
14、 was toevaluate the effects of load sharing inthe narrow band of two tooth contactswith the mating rack. Tooth pair loadsharing potential is evaluated by thetransverse contact ratio parameter. Thenominal, transverse contact ratio isapproximately i.i. The final objectivewas to determine if the mating
15、, throughhardened, rack had equal or betterultimate fracture capacity compared tothe carburized pinion, when loaded atits highest point of single tooth contact.L FIGURE3 _JPx 1o.TEST CFigure 2 shows the test stand thatwas designed to support the jack-up geardrive with its outboard, 7 tooth pinion V
16、I/B“ ELECTRICin a similar way that the jack tower DRILL DRIVEsupports the drive in the jack up rig .50:1 SELF -_ 1,_ L_OCKING 1%40LEB nw T OR(_ U E “-J ?TRANSDUCERL FIGURE4 _JFIGURE 2 atACK-UP GEM DRIVETEST STANDFigure 4 shows a schematic of thetest drive train. The input shaft ofAn actual segment o
17、f rack was installed the test gearbox was coupled to a Lebowin the test stand to mesh with the Torque Transducer which was driven by apinion, see Figure 3. This rack segment 50:1 ratio, self locking worm gearreacted the statically applied drive. The worm drive input was thentangential load and loade
18、d the test rotated by a large 1/2 inch drivestand housing in a closed loop industrial drill for system loading. Anconfiguration and self contained the input torque of 350 ib-ft at the jack-extremely large tangential force of up up gearbox generated the 2,000,000 lb.to 2,lO0,000 ibs. tangential outpu
19、t force via the five- 2 -parallel shaft gear meshes and one Fplanetary stage for a total 1765:1 gear HARDNESS TRAVERSE_-_ drive ratio. The output force takes 60-_,._into account the frictional mesh andbearing losses. The system winds uplike a large torsional spring requiring38 full turns of the inpu
20、t shaft onceall backlash is removed to obtain1,600,000 lbs. tangential force on the _ 50 k7 tooth pinion. The self locking worm _drive held the input torque at anydesired level and prevented unwinding _of the geartrain. The Lebow Torque Cellwas connected to a digital displaycalibrated to read input
21、torque directly m 40 in footpoundsThe 7 tooth pinion and racks wereinstrumented with strain gauges at, andvery near the calculated critical stressarea of the tooth root fillets. Fourgauges were laid at each edge of the 30 040face width so that any maldistribution 0 00 0 I0 0 20 0 30of load across th
22、e face width, due to DEPTH (INCHES)misalignment, could be assessed, seeFigure 5. The strain was measured with I FIG URE 6a Vishay Strain Recording Indicator.FSTRAIN GAGE LOCATIONSON PINION _ /i _ PINION LOADING, PREDICTED RATING ANDINBOARD SIDE -7 / _ _ /_ STRESS ANALYSIS_ TII_ / At the maximum requ
23、ired holding0U ;SIDE load of i, 600,000 Ibs. tangential, theAGMA 2001 9 pinion bending stress was145 KSI. With a life factor of 2.7 forless than 103 cycles, the predictedService Factor (SF) is 1.21. The AGMAcontact stress was 528 KSI. The/ calculated contact stress was notGAGES ARE 1.5_ IN / _ consi
24、dered realistic because there wasFROM EACH END / _ local yielding of the rack in the/ contact zone./ The pinion of the jack-up rig wasanalyzed using the “ANSYS“ finite/ element program to calculate the/ principal stresses at the root of/ tooth due to the normal force applied atthe highest point of c
25、ontact whenI _I_U_E 5 I meshing with the rack. Only one toothand a portion of the adjacent tooth weremodeled because of the minimal effectsof the other pinion teeth, see Figure 7.The pinion was modeled using the STIF42PINION MATERIAL, MANEFACTURING PROCESS - 2D isoparametric solid elements fromAND H
26、EAT TREATMENT the ANSYS library. This element is usedThe 7 tooth output pinion was made for two dimensional models of solidof 17crNiMo6, a popular European structures and can be used either as acarburizing steel, which has high biaxial plane element or an axisymmetrichardenability. The manufacturing
27、 ring element. The element is defined byprocess for the pinion consisted of four nodal points and has two degrees offlame cutting the basic profile with an freedom at each node. A two dimensionalallowance for finishing. The fillets model with a unit depth was consideredwere then bored and the final
28、profile adequate as there would be no out ofwas shaper cut. The pinion was then plane effects in the pinion with itscarburized, hardened and tempered. The 8.5 inch face width. Figure 8 presentsactual hardness traverse from a tooth is the calculated principal stress at theshown in Figure 6. root of t
29、he pinion tooth.- 3 -.063 in. on one edge of the tooth and.049 in. on the other edge. The pinionof the second tooth pair had two thinfoil strips laid parallel to each othernear the center of the face width,running from the pitch diameter to theroot diameter of the pinion. Thesestrips were connected
30、with wires to apower source and an indicator light so_ that when the second pair of teethcontacted the foil strips, they would beshorted together completing a series_ circuit and lighting the indicator lamp._ It was anticipated that when the second_ tooth pair contacted, a significant_ _ change in s
31、hape of the load versus_ strain curve would occur as the first_ tooth pair transferred a portion of the_ total load to the second tooth pair._“ _ _ PINION TEST RESULTSThe input torque was applied inincrements and complete sets of allFIGURE 7 - BULL PINION MODEL AND available data were recorded. Stra
32、inBOUNDRY CONDITIONS Gauges 2 and 7 , which were at the peakstress points (one on each edge of thepinion tooth), are plotted on Figure 9against output tangential force. Theindicator lamp lit up at 1,000,000 lb.load indicating a second tooth paircontact. Gauge 2 showed a reduction inthe slope of stra
33、in versus applied loadbeyond the 1,000,000 lb. load level.Gauge 7 showed less change in slope.This lower slope verified a sharing ofload between 2 tooth pairs. Thereduction in load of the first pair, dueto load sharing, was not as great as wasexpected. The strain curves continuedin a nearly linear f
34、ashion up to1,800,000 ibs. and then showed a slightincrease in slope. The load wasincremented up to a tangential load of2,100,000 Ibs. and held for 15 minutesat this point. As the highest appliedload was approached, there was acracking noise much like the sound ofice cracking. When the load wasrelea
35、sed, there was a high amount ofresidual tensile strain showing on thestrain gauges. Gauge 2 showed (1143microstrain) and gauge 7 showed (1557microstrain). It was hypothesized thatyielding of the core occurred whichFIGURE 8 - CALCULATED PRINCIPAL left the case in residual tension.STRESS AT THE ROOT O
36、F A second test was run on the sameTHE PINION TOOTH tooth, similar to the one above, exceptthat the initial gap of the second toothpair was set at .074 in. and .097 in.The load was incremented again andseveral loud cracking sounds were heardTRANSVERSE LOAD SHARING while holding at 1,650,000 Ibs. The
37、When one pinion tooth is loaded at strain in gauges 2 and 7 went to zeroor near the highest point of single because the cracks occurred righttooth contact, a second tooth pair is through the gauge length. Within a fewvery close to contact. The test was run seconds the tooth failed cata-with the pini
38、on and rack indexed strophically. The sound was much like arelative to each other in such a way loud gun shot. Sparks flew out of thethat a second pair of teeth had a gap test fixture. This violent failure wasmeasured between contacting surfaces of different from what was reported in- 4 -Ref.8. Moor
39、ess testing indicated a some _what gentle parting after the initial I Icracking. This discrepancy might be afunction of the size difference betweenv the test samples. Several other teethwere loaded up to 1,800,000 lb.i tangential load several times with nocracking heard or any significantresidual st
40、rain in the gauges. It wasestimated that the low cycle failureload was likely to be around 2,000,000ibs. based on the testing done. Themaximum strain recorded of 8693microstrain at this load corresponds toa tensile stress of 261 KSI. Thisstress is very close to the estimatedtensile strength of 275 K
41、SI for thecarburized case material.The FEM results at 1,175,000 ibs.load (just after second tooth pair Ill contact) gave 148 KSI. The measured /_ Istress at this load was 143 KSI, which Iagreed reasonably well with the FEM Jcalculations. At a tangential load of1,800,000 ibs., the FEM results gave as
42、tress of 228 KSI. The measured L F I_ URE l 0result at that load was 218 KSI. This described above for the pinion. Thedifference can be explained by the load rack segment tooth root fillet wassharing effect among the two teeth. The instrumented with four strain gauges asAGMA prediction of a 1.21 SF
43、at shown in Figure ii. The fourth gauge on1,600,000 ibs. load seems reasonable each side of the rack was laid on theconsidering that the failure load was end face of the rack rather than theexpected to be 2,000,000 ibs. ( i.e., root fillet. Gauges 2 & 6 were at the2,000,000/1,600,000 = 1.25 SF). cal
44、culated maximum critical stress area.i- -1 F -1LOAD SHARING TEST MODIFIED PINION STRAIN GAGE LOCATION_-_10000700080009000 _ ,_/.,_/,/_asO_aoo.3OO_ ON RACK /z 6000 _ v4000 # INBOARDSIDE= 3000a000 _/ -50 / _ _/ OUTBOARD SIDE0 “0 _ i -4ONEDGEOF TOOTH0 300 600 900 1200 1500 1800 2100(8ON INBOARDSIDE)LDA
45、B - EQUIVALENT KIPSo GAGE #2GAGE #7I FIGURE9 -J L FIGURE1t 3RACK MATERIAL HEAT TREATMENTAND MANUFACTURING PROCESSThe rack material was not known asRACK TEST it was customer supplied. The toothIn this test, the rack and the profile was flame cut only. Thepinion were again indexed. The highest hardnes
46、s readings of the rack averagedpoint of single tooth rack contact was 350 BHN on the tooth flank surface andobtained and this point meshed with the 280 BHN two inches below the flankdedendum of the pinion as shown in surface, measured on the side of theFigure i0. The test was conducted as rack.- 5 -
47、FRACK LOADING, PREDICTED RATINGAND STRESS ANALYSISThe AGMA 2001 bending and contact RACK CAPACITY TESTstress numbers and the FEM bending i0000 300stress at 1,600,000 ibs. tangential load 9080were: / -250AGMA Bending Stress = 158 KSI 8000 /AGMA Contact Stress = 528 KSI 7000 _ _ /_/ -200FEM Bending St
48、ress = 170 KSI z 6000The material yield strength at 350 BHN _was estimated at 165 KSI. _ 5000 150D 4000 w3000 -ioo2000 /I000 “_ 50/0 00 300 600 900 1200 1500 1800 2100LOAD - EQUIVALENT KIPSGAGE #2GAGE #6RES.LTSO,RACTEST FIGURE 12 _1The rack was loaded against thepinion tooth to a tangential load of2
49、,100,000 ibs. Failure of the rack hadnot occurred even though the tooth rootfillet strain exceeded the expectedyield point of 5500 microstrain or 165 TEST 2KSI stress.The rack was not loaded to higher VERY LOW CYCLE BENDING FAILURE TESTlevels because there was concern for the (6, CP)test stand and other elements within thegearbox. Figure 12 shows the plot of TEST OBJECTIVE AND METHODstrain and stress versus tangential The second test was conducted toforce. The non-linearity of this plot determine what load would causeis attributed to the local
