英文版

AbstractThis paper presents the results obtained in the study of the cracking problems on the driving gearboxes of series 2600locomotives. These components were under big vibration and severe fatigue conditions. The cracks in these componentsappear in two distinct parts: in the upper zone of the cover and in the frontal central zone of the body of the housing. Theloading data was obtained in service, in two routine journeys: Lisboa–Porto intercity train, with maximum speed of160km/h and Entroncamento–Guarda freight train, with a maximum speed of 120km/h. Finally from a synthesis ofall the data obtained a set of conclusions are presented to prevent further failures in the driving gearboxes. 2006 Published by Elsevier Ltd.Keywords: Failure study; Housing of railway gearboxes; Fatigue analysis; Cracking problems; St22-31. IntroductionTrainsarecomplexmachines,withmanycriticalcomponents.Structural designofthetrainsisbasedonthedesign standards, the function of the structure and the environmental conditions. The service conditionincludes the forces, vibration, environment, etc. For the design, strength analysis and testing of the structure,the service condition has to be taken into account. The simulated external loads for the strength analysis and testing of the structure are determined according to its practical serviceconditions [1,2].Failure analysis in cast steel railway coupling of the freight trains (coal transportation) line has been conducted by the authors [3], where strain gauge data were analyzed and from those data fatigue cycles were derived. Calculation of the damage factors in several zones of the coupling and for the di?erent stages of the journey and loading conditions were made. And the stress concentration factors were calculated from the computed stress distribution and for the critical areas of the coupling.Correspondingauthor.E-mailaddresses:tmorgado@ipt.pt(T.L.M.Morgado),cmbranco@dem.ist.utl.pt(C.M.Branco),vinfan te@dem.ist.utl.pt(V.Infante).1350-6307/$ - see front matter 2006 Published by Elsevier Ltd.doi:10.1016/j.engfailanal.2006.11.052Housings of railway vehicle gearboxes are examples of components which are subjected to complex stresses during operation. Stresses are mainly generated by mass forces transmitted from motor via the flanged connection to the housing neck (Fig. 1) and also by loading resulting from lateral drive and starting and braking operations. Moreover, the loads acting on gear wheels generate bearing forces which must be taken up by the gearbox housing, [4,5].A significant number of failures have occurred in housings of railway vehicle gearboxes which led to great losses and disruption of service. The cracks appear to be fatigue cracks due to dynamic loads produced in the load wagons of freight trains or of the passenger trains. In this study, attention is devoted to analysis of thedamage of the housing gearboxes subjected to both operational and overspeed conditions and also to indicate critical areas from thepoint of view ofthestressanalysis. The additionalgoalof this analysis isto improve the safety and the reliability of these components.In this paper, measurement procedure of the displacements in service and results are presented. For the accomplishment of this work the following tasks were made:●Evaluation of the mechanical characteristics of the steels of thetwo di?erent parts of the driving gearboxes (cover and body),and determination of the elasticity modulus of the steels.●Measurement of the extensions in service of the housing(cover and body) of the box, using electric strain gauge/rosettetype, bonded in critical zones, where the cracks were detected.●Calculation of the stresses in the most highly requestedsupports of the driving gearbox.●Forecast of the fatigue damages at critical locations.●Final conclusions.2. Material characterization of the housing of the gearboxThe housing of the driving gearbox is of welded construction; the body of the box is a structural steel of low carbon, with 8 mm thickness, according to ARMCO (St22-3) specification. The steel of the cover is St37-2DIN specification. In Table 1 are indicated the chemicalcompositions in weight of these two steels, obtained in removed samples of one of the housings. The data showed in Table 1 gives a lower carbon content in the body of the housing ,but for the cover of the housing there is a higher amount of alloy elements .Therefore this gives an increase in the carbon equivalent (0.26% for body and 0.20% for the cover).Vickers hardness was obtained also for the two parts of the housing. In this procedure a load of 5 Kgf was applied during 15 s in polished steel samples. The equipment used was a digital macro hardness machine. Tenindentations were made. The results were following:Modulusofelasticity,E,andPoissonratio,m,wereobtained.Thetestswere madewithstandardspecimens,accordingtoASTME-8[6],on a servohydraulic machine. Those specimens were machined from the cover and body of the housing gearbox and were instrumented with two longitudinal strain gauges and two transversal lines, mounted in complete bridge. In each test six readings (corresponding to six loadings) of the strains were registered in a strain gauge static bridge. It was verified that the value of the coe?cient of Poisson is close for the two materials (mcover = 0.28 and mbody = 0.29), but the modulus of elasticity obtained for the steel of the body of the housing (Ebody = 198 GPa) is about 10% less of the modulus of the steel of the cover. Elasticity modulus of the cover is Ecover =218 GPa.Tensile tests were performed until rupture following ASTM E-8. In these tests the same servohydraulic machine was used (mentioned in the previous paragraph) and the extensions had been continuously registereduntil rupture, using an axial strain gauge directly connected to the specimen, with deformation length l0 = 25 mm. Nominal tension–extension curves until rupture were obtained, and the values are shown in Table 2. In this table it can be observed a big scatter of results for the steel of the housing. The steel of the cover of the housing (DIN St37-2) is tougher than the steel of the box body and the ductility of this steel is lower than the ductility of the cover (ARMCO St22-3).3. ExperimentalIn this work nine rosettes were bonded (45 rosettes – 3 linear strain gauges) in the following positions:rosette 1 in the central zone of the cover of the housing to measure the stresses in the cover; rosette 2 in a zone away from weld areas. In the box body, to measure the nominal stresses; rosette 3 in the weld toe of weld zone on the lower reinforcement box, place where fatigue cracks were started; rosettes 4–9 were bonded in the six supports of the housing to obtained the reaction forces (Figs. 2 and 3). The same procedure of surface preparation has been conducted by the authors in the study of cast steel railway coupling used for coal transportation [7]. The data was collected with a portable PC and the HPVEE processing system was used for treatment and analysis of the signals.The data was obtained in service, in the Lisbon–Porto Intercity passengers train, with a maximum speed of 160 km/h and in the Entroncamento–Guarda freight train, with a maximum speed of 120 km/h (Table 3). The program of the rosettes readings were made in order to get the best possible comparative information. Those stages are thecritical section of the journey, in what concerns speed, power and track oscillations.For the Lisbon–Porto journey in the Intercity, the following 6 stages were selected:●Entroncamento–Fa′tima: Acquisition time –894 s; Number of km –23.3 km;●Fa′tima–Pombal: Acquisition time –1180 s; Number of km –40 km;●Pombal–Alfarelos: Acquisition time – 1037 s; Number of km – 29 km;●Alfarelos–CoimbraB: Acquisition time – 900 s; Number of km – 18.7 km;4. Stress spectra in service and analysis of fatigue damage5. Numerical simulationFig.8 (a)shows a2D Finite Elements mesh of the model. The finiteelements code ANSYS was used and the type of elements was the 8nodes is oparametric. Near the support eye the mesh was refined. The boundary conditions of the computational model are identified in Fig. 8(b). The loading, p, as applied of form to obtain the values of principal and equivalent peak stresses, measured in service. Table 5 gives the details of the mesh that produced more consistent results.Since the support force, in operation, is revolving in order of time it was admitted that reaction resultant will have values to occur with a direction, h, changing in the plan of support joint. In this way, was obtained the supporting reaction force in support eye, R ,for h of 30in30 and for the direction angle of principal stress, 22.73 using the following equations: R 1/4 Ap; e11T A 1/4 2prl; e12TWhere p is the load applied in FE simulation, A is th econtact surface area of the bore with the pin, r is the bore radius and l is the support pin thickness.Table 6 presents the data values of the supporting reaction force, R, obtained for each angular position and for the principal and equivalent stresses measured in service (rosette 7). The results of supporting reaction forces are significantly independent of h until angle of 60 and are high for forces between 220 kN and 230 kN (experimental results).6. ConclusionsOf the values of the supports reactions and the fatigue behaviour of the driving gearbox, the following conclusions can be taken off:●The steel to carbon of the box body (St22-3) does not possess the level of resistance and the ductility adjusted for this application.●Although the steel of the box cover (St37-2) to possess better mechanical characteristics than St22-3 steel, is neither indicated for this type of application.●Tensions level measured in the driving gearbox with the electric strain gauges is raised, what originates problems of fatigue cracking, especially in the critical zones (weld zones).●Reactions in the superior lateral supports had given very high values.●The results of supporting reaction forces are significantly independent of h until angle of 60 and are high for forces between 220 kN and 230 kN.●Fatigue damage presented high values.●Calculations of safe life gave values very close to the life values obtained in service for the driving gearboxes where fatigue cracks were detected.AcknowledgementsThe authors are thankful to CP (Portuguese Railways Company) and to EMEF, S.A. (company of the CP Group).References[1] Weihua Zhang, PingboWu, Xuejie Wu, Jing Zeng. An investigation into structural failures of Chinese high-speed trains. EAnal 2006;13(3):427–41.[2] Gu Diankang. Test analysis of frame intensity of 209 serie bogie.Railway Locomotive Car 2002;22(4):11–2.[3] Infante V, Branco CM, Brito AS, Morgado TL. Failure Analysis of cast steel railway couplings used for coal transport Failure Anal 2003;10(4):475–89. Agosto.[4] Sonsino CM. Structural durability of cast aluminium gearbox housings of underground railway vehicles under variable loadings. Int J Fatigue 2005;27:944–53.[5] Sonsino CM, Kaufmann H, Foth J, Jauch F. Fatigue strength of driving shafts of automatic transmission gearbox operational torques. SAE paper 970706 (1997); SAE Transactions Section 5. J Mater Manufact, USA 1997, 635–48.[6] ASTM, Specification for tensile testing of cast ferrous materials. American Society for Testing and Materials, USA, E8 spe 1992.[7] Branco CM, Ascenc?a?o L, Infante V, Costa C, Morgado TL. A failure analysis study of cast steel railway couplings use transportation – Part I: Analysis of fatigue Damage and Stress Analysis. In: Proc Eighth Portuguese Conf on Fractur Portugal 2002, 449–61.[8] Ascenc?a?o L. Private report from UMAT/CP, Engineering Department, Lisbon, 2002.[9] Branco CM, Ferreira JA, Costa JD, Ribeiro AA. Design of Machine Elements. Lisbon: Calouste Gulbenkian Foundatio Portuguese].[10] Radaj D. Design and analysis of fatigue resistant welded structures. Cambridge, UK: Abington Publishing; 1994.[11] Maddox SJ. Fatigue Strength of welded Structures. second ed. UK: Woodhead Publishing; 1991.[12] Eurocode 3, Section 9, Regras de projecto a` fadiga de estruturas soldadas de ac?o. 2001.[13] SPSS Advanced Models 6.13, 2004 SPSS Inc.。