1、05FTM19The Application of Very Large, Weld Fabricated,Carburized Hardened however, the real loading spectrumis not usually obvious and is seldom simple. Actualtorque measurements have shown that the torquetransmitted by each gear mesh is not well repre-sented by the torque equivalent of the motor na
2、me-plate power and speed. Because of the very high in-ertia in the entire system and the dynamicconditions that occur as the rolls bite into the steelslab and proceeds to change both its shape andstructure, Figure 6, the torque that is applied to eachgear mesh in the system is different and varies s
3、ig-nificantly as the bar progresses through the rollingprocess.Figure 6. The Basic Steel Rolling ProcessOne would, of course, expect a difference in torqueon the motor to main gearbox and main to mill piniongearboxes due to the ratio of the main gear drive.This is true; however, strain gage measurem
4、ents,such as those shown in Figure 7, show that the dif-ference in torque is often not proportional to the gearratio alone. This plot shows torque measured onvarious shafts in the system as a function of time.Figure 7. Torque Measured on Various Shaftsin a Steel Rolling Mill (strain gagemeasurements
5、)Note that the torque variation is particularly highduring entry of the bar into the rolls and then itsettles down to a more stable “variation” pattern fi-nally dropping off as the bar exits the mill. Motorpower, as measured by motor current, also variesbut to a far lesser extent as the system inert
6、iainsulates” the motor somewhat from these verywide load variations. Because of this, measure-ments of motor power or motor current are generallypoor indicators of the torque applied to the gearmesh and very often under-predict the gear meshloading. This can lead to premature gear failure andbearing
7、 life problems.Considering these factors, from this data plot, itshould be obvious that the torque measured is notjust a function of the motor power and the gear ratiobut also of the system dynamics. Note that thetorque experiences significant variation as the barenters and exits the mill. Even duri
8、ng the actual roll-ing process the torque alternates significantly inmany cases.While we are dealing with the main gear drive here,it is also interesting to note that the torque carried by4the upper and lower spindles (and thus the torqueabsorbed by the top and bottom rolls) is not equal.Our design
9、requirement for this torque split is a60/40 split to either shaft. This split represents theusual expected “worst” case but we have observedsplits as high as 70/30 to these two shafts.This variation of torque is often referred to as theTorque Amplification Factor (TAF). The TAF is bestunderstood by
10、considering the simplified torqueload schematic shown in Figure 8.Figure 8. Simplified Torque SchematicIn order to properly design the gear and bearingsys-tem, it is necessary to have a reasonable knowl-edge of the TAF for any system. The TAF can be cal-culated by creating a dynamic model of the ent
11、iremill system or it can be measured through the use ofstrain gages applied to the various shafts. It canalso be estimated based on experience but such es-timates are fraught with the possibility or error aseach system is quite different. In any case, the TAFmust be properly addressed. The discussio
12、n ofloading is especially important in our context be-cause, unlike most gear applications, the design ofthe gearing itself can significantly impact the loadingthat the gear set will be required to transmit. Many,older mill gears were manufactured with relativelylight webs and stiffeners and equally
13、 thin rims underthe gear teeth, as the failed gear in Figure 9 shows.Note that the gear shown in Figure 9 is a cast gearbut it is also representative of the configuration ofmany weld fabricated mill gears as well.The teeth on such gears were often hobbed or,sometimes, shaper cut to AGMA Quality leve
14、ls inthe range of 6 to 8. At these quality levels, the com-bined pitch line runout of the gear and its mating pin-ion is such that the backlash in the gear set was usu-ally quite high, even considering the relativelycoarse diametral pitch typical of mill gears (pitch isgenerally in the range of 0.75
15、 to about 2.0). Typicalbacklash for these gear sets would be in the rangeof 0.060 to over 0.100 inch, depending on qualityand pitch. When the bar enters the mill and beginsits journey between the rolls, as is clear from Fig-ures 7 and 8, impact loading is applied to the entiresystem. This loading is
16、 fed back through the gearsystem. Backlash in the gear system exacerbatesthe effect of the impact loading thus minimizingbacklash is a desirable. In order to allow the back-lash to be minimized, the gear quality level must beincreased. A typical set of ATG mill gears, of a sizesimilar to the set sho
17、wn in Figure 5 would have abacklash in the range of 0.025 to 0.040 inch. Thislower backlash greatly reduces the TAF thus thegear tooth design itself can reduce the appliedloading.Figure 9. Failed Roughing Mill GearIn addition to providing dramatically improvedstrength and stiffness (discussed in mor
18、e detail be-low), the greater overall mass, thicker gear rims,heavier webs the cost per minute ofunexpected down time is correctly stated!).Figure 15. Scored Finishing Mill GearIn the case of these gears, scoring occurred veryshortly after start up. The gear set was replaced witha virtually identica
19、l set which also scored in a verysimilar manner in a very short period of time. Carefulinvestigation of the system indicated that misalign-ment was at work and certainly exacerbated the sit-uation. Unfortunately, because of the overallconstruction of the gearbox, deflections within thehousing made i
20、t very difficult to maintain an accu-rate alignment of the gears set under all load condi-tions. The most effective solution would be to re-place the gearbox in its entirety with a stiffer, higherload capacity unit.After considering all options, however, the clientconcluded that replacing the entire
21、 gearbox was theleast desirable solution. This being the case, it wasnecessary to improve the load capacity of the gearset, especially its scoring capability, so that it couldsurvive these less than ideal operating conditions.As a starting point, we calculated the basic strengthand durability load c
22、apacity of the gear set to pro-vide a basis for redesign to improve capacity. Theresults of this analysis, Figure 16, indicated that the7basic strength and durability service factors for theOEM design (configuration 1 in Figure 16) were, atthe actual operating conditions, rather low.Figure 16. Servi
23、ce Factors for Finishing MillGear System at Normal Operating ConditionsIn a steel mill application, we would generally rec-ommend at least a 2.00 service factor. In order toimprove the load capacity of this gear set towardthis goal, within the confines of the existing housing,we evaluated a number o
24、f alternatives starting withan upgrade to an ATG single helical carburized/through hardened gear set (configuration 2 in Fig-ure 16). While this did provide some improvement incapacity, it was not sufficient. We next consideredan ATG carburized/through hardened double heli-cal set and an ATG carburi
25、zed/carburized single he-lical gear set which provided yet another set im-provement in capacity. The final configuration wasan ATG double helical carburized/carburized con-figuration.Since the initial failure mode observed on this gearset was scoring, it is important to address the scor-ing load cap
26、acity as well. As Figure 17 shows, thescoring failure probability of all of the ATG systemsconsidered is significantly reduced as compared tothe original gear set design.Figure 17. Scoring Probability Comparisonfor Finishing Mill Gear SystemBased on this design development, a new ATGdouble helical c
27、arburized/carburized gear set wasdesigned and manufactured, Figure 18, for this ap-plication. The overall design, manufacture, andquality control was carefully accomplished is accor-dance with a very detailed specification that was de-veloped by DST and very effectively implementedby the gear manufa
28、cturing organizations selectedto implement the project.Figure 18. New, ATG Fully Carburized,Replacement Gear Set at Mesh CheckIn this gear set, both pinion and gear teeth are car-burized and Maag Hard Cut (MHC). In order to ac-commodate the observed and calculated systemdeflections, this gear set in
29、corporates both profileand lead modifications. The lead modifications arenot simple crowning but rather a calculated offset,crowned lead curve that is different on the RH andLH helices. Because of the nature of the deflections,the contact pattern at no load (as observed duringthe mesh check shown in
30、 Figure 18) is small andhighly localized, Figure 19. While these contact pat-terns may seem surprising for such a large gear set,they are, in fact, exactly what is required to producea full, uniform, load distribution across the full facewidth, as Figure 20 shows.8Figure 19. No Load Contact Patterns
31、 at MeshCheck (check performed on roll stands asfinal manufacturing quality check, beforeinstallation in gearbox)Because of the highly localized nature of these con-tact patterns, a final no load, “hand” rotation speedcontact pattern mesh check is an essential, integralpart of the overall Advanced T
32、echnology Gear De-sign and Manufacture system. This final patterncheck has been a part of the overall process with thecarburized pinion through hardened gear set com-bination that was the initial implementation of theATG concept but it is even more important with thecurrent carburized pinion carburi
33、zed gear com-bination. This is especially true for the latter com-bination since the hard surfaces do not wear in at alland thus must be absolutely correct from the start.The very small, localized contact patterns shown inFigure 19 are one of the most difficult aspects of theATG system to explain to
34、 mill maintenance person-nel who are generally accustomed to seeing no loadpatterns that meet an “ at least 80% of the full facewidth ” at installation. When the correct lead andprofile modifications are applied, however, the 80%or better contact pattern requirement is more appro-priately applied to
35、 the full load contact patterns. Itshould be obvious from Figure 20 that this full loadrequirement is met for the finishing mill gear setshown. It is interesting to note that there are noareas of concentrated contact apparent in Figure20, even though the contact extends to both ends ofthe face width
36、. The lack of load concentration wasconfirmed after several years of operation (see dis-cussion below).Figure 20. Loaded Contact PatternsThis fully carburized ATG gear set was installed inthe spring of 2003 and has operated without inci-dent since then. Based on the last visual evaluation,the pinion
37、 and gear tooth surface condition, Figure21 remains excellent.Figure 21. Finishing Mill Pinion ToothConditionNote also that the tip relief on these teeth is not com-pletely used at the normal loading. This slightly ex-cess profile modification is purposeful so that peri-odic overloads that are chara
38、cteristic of rolling milloperations do not result in hard contact between thetips of the teeth and the mating points near the low-est contact points. This is one of the finer points as-sociated with the ATG concept.Roughing Mill Application The gears in thedouble reduction, double helical main mill
39、gear driveshown in Figure 22 had suffered several failures.9Figure 22. Large, Double Reduction, DoubleHelical Roughing Mill Gear DriveVisual evaluation of the gear system revealed se-vere damage to the tooth surfaces in the form ofwear, galling, and pitting as Figure 23 shows.Figure 23. Damaged Roug
40、hing Mill GearTooth SurfacesWhile gears teeth can generally sustain a consider-able amount of surface damage before tooth frac-ture becomes a real risk, in this case, both thestrength and durability service factors are rather lowthus the surface damage produced tooth fracturefailures in relatively s
41、hort order, as Figures 24 and25 show clearly. Fracture failures cause a cessationof function, of course, but they also carry the risk ofcausing consequential damage to the mill itself thustheir occurrence is especially ominous in a large millapplication.As part of the failure analysis of this gear s
42、ystem, weevaluated the strength and durability service factorsat the normal full load condition (about 10,000 HPmotor nameplate plus appropriate TAF). As Figure26 shows, both strength and durability service fac-tors are quite low.Figure 24. Cracked Roughing Mill Gear TeethFigure 25. Fracture Roughin
43、g Mill PinionTeethFigure 26. Roughing Mill Gear Set ServiceFactor ComparisonAs was the case for the finishing mill case study, thefirst consideration was a replacement ATG carbu-rized/through hardened gear set. After evaluatingthe improvement that could be obtained within thesame footprint as the or
44、iginal gearbox, however, it10was apparent that this configuration would not pro-vide the load capacity margin that was required bythis application.After evaluating several possible design alterna-tives, it was determined that a completely new gear-box would be required to meet the overall mill sys-t
45、em production requirements. The new gearboxwas structurally stronger and far more rigid than theoriginal gearbox. It was also necessary for the newgearbox to fit in the same position as the existinggearbox, using the existing mounting studs and thesame input and output shaft connections. This com-bi
46、nation of requirements resulted in a unique gear-box configuration. Since our focus here is the geardesign, however, we will not address the gearboxconfiguration here.Both HS and LS gear sets in the new replacementgearbox were composed of ATG carburized/carbu-rized pinions in mesh with carburized ge
47、ars. The LSgear, Figure 27, has a pitch diameter of approxi-mately eleven feet, with a gross face width of closeto four feet and an approximately 0.6 diametralpitch. The gears (HS and LS) were completely fabri-cated, rough machined, and tooth rough cut beforecarburization. After carburizing, the gea
48、r blankswere finish machined and the gear teeth hardfinished.Figure 27. Roughing Mill LS Double HelicalCarburized GearThis gearbox was recently completed and shippedbut has not yet been placed in service.Heat Treatment RequirementsHeat treatment of these large Advanced Technolo-gy Gears did not begi
49、n when the parts were ma-chined and ready to be placed in the carburizing fur-nace. Instead the process began when themanufacturer had developed preliminary engineer-ing drawings of the parts. The complete process re-quired review of the steel manufacturing techniqueand heat treatment, and discussion of the engineer-ing drawings as well as the details of the carburiz-ing/hardening process. All of this was accomplishedprior to cutting any chips. The review was accom-plished by a team which included the manufacturer,designers, metallurgists and heat treat facility per-sonnel. This affo
