Friction and wear behaviour of rolling–sliding steel contacts

Author’s Accepted Manuscript

Friction and wear behaviour of rolling-sliding steel contacts A. Ramalho, M. Esteves, P. Marta

www.elsevier.com/locate/wear

PII: DOI: Reference:

S0043-1648(12)00436-X http://dx.doi.org/10.1016/j.wear.2012.12.008 WEA100453

To appear in:

Wear

Received date: Revised date: Accepted date:

7 September 2012 5 December 2012 8 December 2012

Cite this article as: A. Ramalho, M. Esteves and P. Marta, Friction and wear behaviour of rolling-sliding steel contacts, Wear, http://dx.doi.org/10.1016/j.wear.2012.12.008 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Friction and wear behaviour of rolling-sliding steel contacts A. Ramalho; M. Esteves; P. Marta CEMUC – University of Coimbra, Portugal [email protected]

Abstract Improving the total life cycle costs and safety of trains are current research topics that hold great interest for those who build, maintain and operate trains. The maintenance interval for both wheels and rails has become a major issue in reducing costs and increasing safety and has encouraged the development of new tools for predicting the evolution of wear in order to establish a convenient maintenance schedule. These new tools require the synergy of dynamic analysis and the development of suitable wear models. Rail/wheel wear depends on the material properties resulting from a competition between contact fatigue and sliding wear. Therefore, all the contact conditions affecting the contact stress distribution will determine the wear behaviour. The current research paper investigates the effect of contact conditions on friction and wear behaviour of carbon and low alloy rail/wheel steels. Two-disc rolling-sliding tests were done to study the effect of the creep ratio, contact pressure and tangential speed on the resulting traction coefficient and amount of wear. Scanning and optical microscopy and the evolution of micro hardness were used to observe crack growth beneath the contact surface and to evaluate the strain-hardening effect, induced by contact stresses. Further rolling-sliding and unidirectional sliding tests were performed to assess the wear of the wheel rims. Results were analysed and discussed, comparing the effectiveness of a classical approach based on Archard’s equation, with models involving the energy dissipated by friction.

Keywords: Rolling-sliding; wear; contact-fatigue, Rail-wheel.

Symbols a Ae b E E* Er Ew F N H k Kc kE l P q R

Major axis of the wear particle (mm). Mean size wear particle area (mm2). Minor axis of the wear particle (mm). Energy dissipated by friction (J). Equivalent Young’s modulus of the contact pair (Pa). Young’s modulus of the rail material (Pa). Young’s modulus of the wheel material (Pa). Friction force (N). Normal force (N). Hardness (HV). Specific wear rate (m3/N.m). Wear coefficient. Energy dissipated specific wear rate (m3/J). Contact width (m). Contact Pressure (Pa). Distributed load (N/m). Equivalent radius of the contact pair (m).

Equivalent radius of the of the mean size wear particles (mm). Rail specimen radius (m). Wheel specimen radius (m). Creep ratio. Time (s). Sliding speed (m/s). Wear volume (m3). Poisson of the rail material. Poisson of the wheel material. Rail specimen angular speed (Rad/s). Wheel specimen angular speed (Rad/s). Sliding distance (m). Time interval (s). Friction coefficient between specimens.

r Rr Rw S t v V Ɣr Ɣw Wr Ww x t f

1.

Introduction

The railway industry is one of the most utilised means of transportation on land and increasing competition requires an evolution in service times, costs and comfort. These variables are directly dependent on rail/wheel maintenance, in which the major concerns are wear and rolling contact fatigue, typically induced by friction and cyclic overstressing of the rail and wheel materials [1, 2]. Some authors [3-6] suggest using Archard’s wear model, which relates the proportionality of the wear volume with the sliding distance times normal force divided by the hardness of the material, equation (1), mainly because it is widely used and very accurate.

V

Kc

Nx H

(1)

Considering the linear Coulomb´s friction, the normal force can be expressed as a function of the friction force and thus, it could be concluded that Archard’s model implicitly correlates wear volume with the energy dissipated by friction. Therefore, wear volume can only be considered in the slip zone of contact because the sliding velocity is null in the adhesion zone, justifying why this model is not always followed by users for reasons of simplicity. Others recommend energy approaches, considering wear as an explicit function of the energy dissipated in contact by friction, such as Zobory’s model [5, 7]. This assumes that the contact area is divided into two zones: the adhesion and sliding zone and that the wear is highly dependent on sliding speed and therefore, the major part of wear is located in the sliding zone of contact. Another energy approach is Pearce and Sherratt’s model [5, 8], which is focused on the prediction of wheel flange and rail thread wear, estimating the wear loss as the area of the cross-section removed by distance covered (mm2/km). As these types of contact involve the competition of two failure mechanisms, the applicability of these models is a difficult matter. Therefore, some studies predict better wear behaviour, based on Archard’s wear models, whilst others have a good performance in prediction of rolling contact fatigue [9]. There are a small number of works directed to quantifying both phenomena, as presented in [10], in which the

authors combine wear and rolling contact fatigue rail/wheel prediction models, after determining the contact conditions for predicting wear volume and the prospect of crack initiation. More recently, these wear models have been associated with simulation modelling of the dynamic behaviour of the train/track, in order to establish a complete and independent analysis [10]. A variety of parameters can influence the behaviour of the wheel-to-rail contact. Therefore, we study some of the most important, such as: the contact pressure (dependent on the applied load), the linear speed, creep ratio, friction force, contact geometry and for various reasons, the materials properties. Defining a relationship between the parameters and establishing a model that can be used in the evaluation or prediction of wear damage is the purpose of this work. There are only a few studies addressing the classification of wear regimes, where an energetic approach has been applied to laboratory and field data and the change of damage mechanisms identified as the leading cause of wear regime transition in various wheel and rail materials [11] and the influence of the temperature variation in the contact of the specimens, which alters some of the mechanical properties of the steels in both the wheel and rail, significantly changing the rate of wear [12]. In the process of selecting test conditions, some studies regarding real working scenarios were taken into consideration; however, some adaptations were made due to equipment compatibility as well as the contact and geometry of the test samples (described and justified later in this work). An evolution of contact pressure with real contact area, studied in [13], indicates the nominal values of contact pressure that should be applied. The slip ratio in this work, denominated creep by some authors, is defined for two situations: rolling/sliding, as indicated in [14] and pure sliding that by definition establishes total creep. The linear speed range was selected due to tribometer specification and in order to conform with [15]. The contact stress between surfaces was analysed using the Hertzian contact stress theory for two bodies in line contact [16], as well as the influence of friction in the distribution of tangential surface stresses. The analysis of wear volume against energy dissipated by friction will be used to link the data obtained in every trial, as well as the severity, Normal Force (FN) versus Sliding Distance (X). Finally, the results and conclusions will be compared with the reviewed bibliography for validation purposes, reviewing the influence of steel microstructure and heat treatments, as well as the application of different wear models.

2.

Experimental 2.1.Specimens and contact geometries

A Roller-on-Roller model was selected to simulate the contact between the wheel (lower specimen) and the rail treads (upper specimen) and to assure the compatibility of the results with the real working application, Figure 1 a). The need to establish a uniform contact between the specimens and to control the contact pressure constrained the design of specimen geometries and the test equipment. As mentioned above, the cylinderon-cylinder with parallel axes establishes good contact geometry and controls whenever the entire considered thickness is in contact; otherwise, the contact pressure could vary for the same load.

Rolling-Sliding tests were used to investigate the contact of wheel and rail along straight tracks; contact case shown in Figure 1 a). For this type of test, the wheel specimens present a cylindrical disc shape (Figure 2), whilst the rail specimens adopt a cylindrical disc shape with a fillet (Figure 3), in order to reduce misalignment and to allow high contact pressure tests under reasonable normal loads. Unidirectional cylinder-cylinder sliding contact with parallel axes was used to simulate the contact between the wheel flange and rail tread (Figure 1 b)), representing the pure sliding in rail curves. This type of contact solely involves a sliding movement, so that the wheel (lower specimen) is animated with a rotating motion and the rail (upper specimen) is fixed (Figure 1 b)). In this case, both specimens present a cylindrical shape but the rail specimen adopts the geometry of a disc with the entire thickness reduced, for the purpose of the calculation of wear volume (Figure 4).

The volume loss was estimated by weighing both specimens before and after the tests. Profilometry, integrating the area of the track wear, was performed to assess the wear on the wheel specimens (Figure 5) and to compare and validate the results. Specimen geometry limits the use of profilometry to just the wheel specimens. In the sliding tests, the resulting wear is predominantly on the rail specimen, because it is a contact geometry of overlapped cylinders with parallel axes and the rail specimen is stationary; hence, the small volume loss of the wheel specimen when compared with that of the rail. The wear volume on the rail specimen is determined by the geometric relations of the contact and by weighing before and after the tests, for data validation purposes.

2.2. Materials The material selection process was based on the composition of the hypoeutectoid steel alloys, as well as the ferrite-perlite microstructure of the typical heat-treated low alloy steel used in the railway industry [2, 9]. Quenched and tempered steel AISI 4140 was considered for the train wheel specimens, while a normalised AISI 1055 was selected as the test rail steel material (Table 1). The microstructure of both alloys is presented in Figure 6 and shows in Figure 6-a) that the AISI 1055 with equiaxed grains of perlite (dark) and a proeutectoid ferrite matrix (white) conformed with [17]. In addition, the AISI 4140, shown in Figure 6-b) a mixture of bainite (dark) with martensite (white) can be identified, which conforms with [17]. The performance of bainitic and pearlitic steels is analysed by [18], where the increase of hardness was shown to improve wear and rolling contact fatigue behaviour and that the microstructure influences the evolution of deformation. The roughness of the test specimens presented in Table 2, results from the machining process. These values interfere little in the results because of the long duration of the tests and similarity of the specimen’s hardness.

2.3.Tribometer The test equipment used was a Multi-Purpose Friction and Wear Tester model TE 53 of PLINT Tribology Products, with slight modifications on the loading mechanisms (Figure 7). An AC geared motor coupled to the shaft and gears was used to induce movement on the tribometer. The slip ratio was determined from equation (2), always maintaining the mean distance between centres as 60 mm (Rw + Rr=60).

S

RwWw  RrWr RwWw

(2)

The contact pressure was applied via a system composed of a pneumatic spring and a load cell, to manage the amount of force supported on the structure (Figure 7), which replaces the dead mass system of the original design. The normal force is transmitted to the contact by the load lever multiplying the load applied by the pneumatic spring. In order to assess the friction force, the rail specimen shaft is mounted on a floating bearing, which allows minor movement on the plane of the friction force load cell (Figure 7). There is also an accelerometer on the load lever to control vibration, which is one of the limiting conditions.

2.4.Rolling-Sliding test conditions Several works [13, 15] indicate contact pressure values within the range of 0–1500 MPa. In the present study, the pressure values ranged from 0 to 1000 MPa, because values close to 0 MPa present very low wear and those higher than 1000 MPa cause excessive wear for the selected contact configuration. The creep ratio lies between 0% and 15.23% to establish the connection between rolling-sliding and pure sliding. Speed range was adjusted from 0.5 to 1.85 m/s, according to the specifications of the tribometer and in agreement with the practical range [15]. For a better understanding on the influence of each condition on the friction and wear results, all the test conditions were cross-matched (Table 3 and Table 4), with common parameters: 500 MPa, 1 m/s and 1% slip. All the tests were performed for 100,000 cycles with the exception of 1000 MPa and 15.23% creep ratio, which applied significant wear at 30,000 cycles. Figure 8 maps the test conditions in a 3D plot with regard to pressure, creep and speed range. LabView software was used to develop data acquisition and control software. The contact pressure was determined by the Hertzian contact stress theory for cylinder-on-cylinder with parallel axes, considering an even load distribution on each contact surface on the specimens. The theory translates to equations, (3) to (6), allowing the quantification of the contact stresses as being: R the relation between the radius and q the contact load per contact length l.

R

( Rw Rr )  ( Rw  Rr )



(3)

q

N  l



(4)

1 E*



(1 Q w 2 ) (1 Q r 2 )  Ew Er

(5)

1/2

P

§ qE * · ¨ ¸ © SR ¹

(6)

2.5.Pure sliding test conditions Sliding tests were performed for better understanding of total slip situations, under the same range of contact pressure and linear speed, as presented in Table 5. The low value of sliding distance was selected because of the high severity present in pure sliding, which resulted in excessive wear volume.

2.6.Experimental procedure Both types of test were performed with the same procedure, concerning the same preparation, control and variable measurement. During the tests, the variables measured were friction force, noise levels and acceleration due to the vibration in the contact. The last two parameters were only used as limiting conditions and to investigate the regime transitions on friction. The analysis of results was performed by resorting to a group of parameters that allow the characterisation and classification of each evolution. The linear equation of Archard (7) was used to determine the specific wear rate k as a function of the wear volume V and the contact severity FN.X, [19]. The higher the value of k is, the lower the wear resistance is.

V

kNx

(7)

A similar analogy can be made with regard to the friction force Ff, obtaining the specific wear rate based on the energy dissipated by friction kF (equation (8)).

V

k E Fx

(8)

The work of the friction force translates into the energy dissipated by that force in the contact E, and can be determined by equation (9) [20].

E

³Fdx

t

G

¦ F v Ct

(9)

0

After every test, the wear debris generated by both wheel and rail specimens was gathered and observed using an optical Mitutoyo Toolmaker's Microscope TM-505 and measured using a Nikon Stereo Photo SMZ – 10 optical microscope with a digital photographic camera. The typical geometry observed on wear particles is outlined in Figure 9. The particle area was calculated considering an equivalent elliptical shape. The shape classification of the debris is made according to the equivalent radius of the particle (equation (10)), where Ae is the area of the mean particle size.

r

Ae

S

(10)

3. Results and Discussion 3.1.Rolling-Sliding tests In order to quantify the singular influence of each test variable, the results will be presented separately in relation to the effect of contact pressure, creep ratio and linear speed.

3.1.1.Effect of contact pressure Figure 10 a) shows the growth of the transversal area of track wear measured in the wheel specimens with the increase of contact pressure. Figure 10 b) shows a similar evolution of the wear volume on both rail and wheel; however, the rail specimen tends to show a little less wear, especially for the highest values of contact pressure. The increase of pressure also increases friction force (Table 6), which results in higher values of energy dissipated by friction and therefore, both Archard’s and energetic specific wear rates display the same evolution, Figure 11 and Figure 12, respectively, with small variations between 300 and 500 MPa and a significant increase for 1000 MPa. This behaviour suggests an alteration in wear mechanism from 500 MPa to 1000 MPa. In fact, observing the wear debris (Figure 13 and Table 7), one can conclude that the dimension and shape of the debris changed significantly in the test done under contact pressure of 1000 MPa. Nevertheless, after spectrum analysis of the particles, the presence of oxidation was noticed for the 1000 MPa test, the result of which was agglomeration of wear debris (Figure 14).

Figure 14 shows that the typical particle size is below the micron but that a more compact interlayer was generated by the agglomeration of these small oxidised particles. The effect of oxidation in the stabilisation of the third body has already been identified for different materials [21, 22].

3.1.2.Effect of linear speed With regard to the influence of the variation of linear speed on wear, Figure 15 and Figure 16 show that the wheel always suffers more damage than the rail. Maximum wear was observed for those tests done at 0.5 m/s and minimum wear occurred for 1 m/s. The friction coefficient revealed a similar variation to wear (Table 8). Figure 17 and Table 9 show the debris of the tests performed at different sliding speeds. The test done with the lowest speed produced the largest particles and the test done at 1.85 m/s generated a debris distribution with a much smaller tail. The appearance of the debris also changed; the particles had an elliptical shape except for the test at the highest speed, which produced debris of a more rounded shape. The wear mechanism for the lowest sliding speed should be adhesion because of high friction, whilst oxidation and abrasion are the prevailing mechanisms present at 1.85 m/s. A combination of both wear mechanisms was found at a sliding speed of 1 m/s, which resulted in the lowest wear. The obtained results reveal that the maximum value of wear was obtained for the test done at a sliding speed of 0.5 m/s. This agrees with the map proposed by Jendel [3], which estimates significantly higher values of wear at 0.5 m/s compared with sliding speeds equal to and above 1 m/s. The evolution of the wear debris shows a decrease in size and an approximation to rounder shapes with an increase of sliding speed, due to the change of wear mechanisms. EDS analysis was performed on the particles revealing oxidation, especially for higher sliding speeds (Figure 17-e)).

3.1.3.Effect of creep ratio Figure 18 shows the variation of the wear volume with creep ratio. For tests done with pure rolling contact (zero creep ratio), the amount of wear was similar in both specimens. For tests with a creep ratio of 0.45, 1 and 1.75 the wear was higher in the wheel compared to that of the rail. However, for tests with the highest value of creep ratio 15.23%, the rail displayed greater wear than the wheel. The specific wear rate of both rail and wheel is shown in Figure 19 and Figure 20 for both Archard’s and the energetic approaches, respectively. For creep ratios from 0.45 to 1.75%, the evolution of k and kE are similar for both the wheel and rail with approximately constant values. Table 10 and Figure 21 summarise the morphology and analysis of the particles for tests done with different values of creep ratio. The dimensions of the debris were approximately constant with an equivalent radius of 0.2 mm, except for the pure rolling test where the particle size was biggest and the test with the highest values of creep, which generated the thinnest particles.

All the results presented previously allow clarification of the difference in predominant wear mechanisms. For low creep ratio (0% and 0.45%), wear volume on both the rail and wheel specimens is similar with prevailing rolling contact fatigue. This mechanism was well identified by the cracks observed in the subsurface beneath the contact areas ((Figure 22 a), b) and Figure 23) located at a depth of less than 50 m below the surface, which agrees with the results of other authors [7, 12] and corresponds to a location around the maximum shear stresses, according to the Hertzian contact stress theory. The cracks are more dominant for those tests performed with a creep ratio of 0%, indicating contact fatigue as the main failure mechanism. For creep ratios of 1% and 1.75%, the wear is produced by two mechanisms: rolling contact fatigue and plastic deformation, which causes more wear in the hardest specimen, the wheel, which deforms less. As the wheel specimen is a bainitic steel and the rail specimen a pearlitic steel, the evolution for creep ratios from 0% to 1.75% is expected because the rail’s pearlitic steel offers more strainhardening ability, which increases the wear resistance [18]. In fact, the micro hardness of both specimens was assessed beneath the contact for both wheel and rail specimens (Figure 24 and Figure 25). The most severe condition, 15.23% creep, induces higher temperatures in the contact zone because the energy dissipated by friction is significantly higher, which leads to oxidation growth and the creation of iron oxides (Fe2O3) [23]. This explanation is well supported by the evolution of the friction coefficient (Table 11). In addition to plastic deformation, this tends to increase wear significantly causing more damage to the rail specimen because of the susceptibility of its microstructure to oxidation (30% ferrite and 70% perlite). The EDS analysis of wear particles found oxidation for every variation of creep ratio. Figure 26 shows a lack of agglomeration of the debris, where the fracture line of particles is visible on Figure 26 b), although there was no increase of oxidation detected for higher creep ratios. Additionally, the thickness of the particles remains below the depth of the maximum shear stresses, as shown in Figure 26 b), typically half the depth with a constant section. In fact, the increase of creep ratio directly increases with sliding speed, elevating the temperature on the contact zone [11], which promotes oxidation [15, 24]. This phenomenon alters the wear mechanisms and particle morphology [25], reducing the debris size for higher sliding speeds. The evolution of the wear debris morphology, presented in Table 10, supports this variation of wear mechanisms, although a more exhaustive analysis of the wear particles and influence on the failure mechanisms is desirable.

3.2.Pure sliding tests Pure sliding tests were performed for a contact pressure of 500 MPa because of the nature of high severity of this test and because of the difference between the contact pressures of the wheel flange/rail thread and wheel/rail thread. Figure 27 presents the evolution of the friction coefficient through the sliding distance and the sudden rise of μf represents the presence of galling because of high adhesion (Figure 28), followed by a steady state friction regime with moderate values of friction coefficient around 0.37. The steady state regime occurs by the oxidation of the contact track, which generated thin wear debris with an equivalent radius of 50 Pm.

3.3.Wear model comparison From the wear maps presented above, two types were selected at baseline to apply and compare the results obtained in this study: one based on Archard’s model and adapted in [9] and the other based on a typical energetic approach, such as that presented in [7, 12]. The applicability of both wear maps was tested successfully; however, because of the small amount of tests performed, the use of the energetic approach is not suitable due to little definition of the limits of the wear regimes. The Archard`s model was more appropriate to define the initial test conditions allowing their prior separation among different regimes (Figure 29), with regard to contact pressure and sliding speed. The energetic approach should be a better model for application to a continuous variation of the test conditions, because it uses the energy to obtain non-dimensionalised results, reducing the effect of the test scale and allowing a more direct comparison of laboratory results and real contact scenarios. The results presented in this work are located in the third and fourth regions of the Jendel’s type map of Figure 29; the first region being unachievable with the current experimental equipment due to very high contact pressure and the second, because of the low sliding speed. Nevertheless, there are some gaps in the range of test conditions, which should be addressed in future work.

4. Conclusions The influence of the contact pressure, sliding speed and creep ratio variation was analysed is this study. The results and conclusions were compared with previous work; specifically the evolution through the variation of each condition. The main conclusions that can be drawn from the results are: 

An increase in contact pressure increases the friction coefficient and the wear volume.



The increase of creep ratio induces a variation of failure mechanisms; specifically rolling contact fatigue for low creep ratios, a combination of plastic deformation and rolling contact fatigue for intermediate creep ratios and plastic deformation and oxidation for high creep ratios. This behaviour benefits pearlitic steels for intermediate creep ratios and bainitic steels for high creep ratios, although for both steels, wear volume increases with increasing creep ratio.



Lower tangential speed induced the highest wear loss on the specimens due to adhesion, decreasing for intermediate sliding speeds and then with a slight increase for the higher sliding speeds due to the introduction of oxidation induced by increasing temperature.



In general, pearlitic steel performs at least as well as bainitic steel through its strain-hardening ability, despite its lower hardness.



The increase of temperature affects the wear regime variation for creep ratio tests and sliding speed tests differently. Sliding speed variation solely increases wear resistance and creep ratio variation decreases wear resistance, both due to increasing oxidation.



The pure sliding test results establish a connection for the results found on the creep ratio variation, mainly to justify the wear mechanisms, although it is necessary to complement this study with lubrication tests, regularly used in this contact condition.

Acknowledgements The authors thank Portuguese Foundation to Science and Technology for funding this work, Project nº PTDC/EME-PME/115491/2009 which is co-funded by the program COMPETE from QREN with coparticipation from the European Community. 

References [1] R. Stocka and R. Pippan, RCF and wear in theory and practice - The influence of rail grade on wear and RCF. Wear, Vol. 271 (2011): 125-133. [2] M. Clarke, Wheel rolling contact fatigue (RCF) and rim defects investigation to further knowledge of the causes of RCF and to determine control measures. Wheel Steels Handbook. Rail Safety & Standards Board. 2008. [3] T. Jendel, Prediction of wheel profile wear - comparisons with field measurements. Wear, Vols. 253 (2002): 89-99. [4] R. Enblom and M. Berg, Simulation of railway wheel profile development due to wearinfluence of disc braking and contact environment. Wear, Vols. 258 (2005): 1055-1063. [5] J. de Arizon, O. Verlinden, and P. Dehombreux, Prediction of wheel wear in urban railway transport: comparison of existing models. Vehicle System Dynamics, Vol. 45, No. 9, September 2007: 849-874. [6] J. Tunna, J. Sinclair, and J. Perez, A Review of Wheel Wear and Rolling Contact Fatigue. Development of a wheel wear and rolling contact fatigue model. Rail Safety & Standards Board. 2006. [7] I. Zobory, Prediction of Wheel/Rail Profile Wear. Vehicle System Dynamics, Vol. 28 (1997): 221-259. [8] T.G. Pearce, and N.D. Sherratt, Prediction of wheel profile wear. Wear, Vol. 144, 1-2, 20 (April 1991): 343-351. [9] M.C. Burstow, Whole Life Rail Model Application and Development for RSSB - Continued Development of an RCF Damage Parameter. Rail Safety & Standards Board. September 2004. [10] B. Dirks, and R. Enblom, Prediction model for wheel profile wear and rolling contact fatigue. Wear, Vol. 271 (2011): 210-217. [11] R. Lewis, and U. Olofsson, Mapping rail wear regimes and transitions. Wear, Vol. 257

(2004): 721-729. [12] R. Lewis and R.S. Dwyer-Joyce, Wear mechanisms and transitions in railway wheel steels. Proceedings of the Institution of Mechanical Engineers, Part J: Journal of Engineering Tribology, 2004, 218(6), 467-478. [13] M. Pau, F. Aymerich and F. Ginesu, Distribution of contact pressure in wheel - rail contact area. Wear, Vols. 253 (2002) 265-274. [14] U. Olofsson. and R. Lewis, Tribology of the Wheel - Rail Contact. Handbook of Railway Vehicle Dynamics. Edited by R. Lewis and U. Olofsson. Taylor & Francis Group, LLC, 2006. [15] L. Deters and M. Proksch, Friction and wear testing of rail and wheel material. Wear, Vols. 258 (2005) 981-991. [16] V.L. Popov, Contact Mechanics and Friction - Physical Principles and Applications. Springer, 2010. [17] ASM Handbook. Metallography and Microstructures. Edited by George F. Vander Voort. ASM International, ASM Handbook Committee. Vol. 9, 2004. [18] P. Pointner, High strength rail steels - The importance of material properties in contact mechanics problems. Wear, Vol. 265 (2008): 1373-1379. [19] H. Czichos, Tribology: A Systems Approach to the Science and Technology of Friction, Lubrication, and Wear. Elsevier, 1978. [20] A. Ramalho and J.C. Miranda, The relationship between wear and dissipated energy in sliding systems. Wear, Vol. 260 (2006): 361-367. [21] I.L. Singer, S. Fayeulle, P.D. Ehni; Friction and wear behavior of TiN in air: The chemistry of transfer films and debris formation. Wear,Vol. 149 (1991) 375–394. [22] A. Ramalho, S. Fayeulle e M.T. Vieira, “The Role of Interfacial Films in the Friction and Wear Properties of W-Co-C Sputtered Coatings”, Thin Solid Films 253(1995) 131-138. [23] M. Vidaud and W.-J. Zwanenburg, Current situation on rolling contact fatigue - a rail wear phenomenon. 9th Swiss Transport Research Conference, Monte Verita / Ascona. September 9-11, 2009. [24] U. Olofsson and T. Telliskivi, Wear, plastic deformation and friction of two rail steels - a fullscale test and a laboratory study. Wear, Vol. 254 (2003): 80-93. [25] J. Sundh and U. Olofsson, Relating contact temperature and wear transitions in a wheel–rail contact. Wear, Vol. 271 (2011): 78-85.  



Figurecaptions  Figure 1 a)Contact of wheel and rail treads, RolleronRoller Contact Schematics; b)Contact of wheel flangeandrailtread,UnidirectionalRollerRollerSliding. Figure2–Wheelspecimendrawing. Figure3–RailspecimendrawingforRollingSlidingtests. Figure4–RailspecimendrawingforpureSlidingtests. Figure5Schematicsofthewearzonefortherollingslidingwheelspecimen. Figure6–Microstructure,etchingwith2%Nital:a)AISI1055steel;b)AISI4140steel. Figure7SchematicsoftheTribometer. Figure83Dmapofthetestconditions. Figure9–Typicalgeometryofwearparticles. Figure10–a)Transversalareaoftheweartrackmeasuredinthewheelspecimensfordifferentcontact pressure;b)Representationofwearvolumeasafunctionofcontactpressure. Figure11–Contactpressureagainstspecificwearrateandfrictionspecificwearrateforthewheel. Figure12Contactpressureagainstspecificwearrateandfrictionspecificwearratefortherail. Figure 13 – Wear debris for rollingsliding tests of: a)300 MPa; b)500 MPa; c)1000 MPa (agglomerates). Figure14–Weardebrissurfaceofthe1000MParollingslidingtest. Figure 15  Transversal area of the wear track measured in the wheel specimens for different sliding speed. Figure16Representationofwearvolumeasafunctionofslidingspeed(onrightforthewheel,onleft fortherail). Figure17Weardebrisforrollingslidingtestsof:a)0.5m/s;b)1m/s;c)1.5m/s;d)1.85m/s;e)EDS analysisofparticlesfor1.85m/s. Figure18Representationofwearvolumeasafunctionofcreepratio. Figure19–RepresentationofspecificwearrateusingArchard’sandenergeticapproachesagainstcreep ratioforthewheel.

Figure20RepresentationofspecificwearrateusingArchard’sandenergeticapproachesagainstcreep ratiofortherail. Figure21Weardebrisforrollingslidingtestsof:a)0%;b)0.45%;c)1%;d)1.75%;e)15.23%. Figure22TransversalsectionoftherailwearzonebySEMfor:a)railwith0%creep,b)wheelwith0% creep;c)railwith1.75%creep;d)wheelwith1.75%creep. Figure23–Contactzoneofa)rail;b)wheel. Figure24Transversalmicrographoftherailwearzoneandhardnessprofilefor1.75%slide. Figure25Transversalmicrographofthewheelwearzoneandhardnessprofilefor1.75%slide. Figure26–Rollingslidingwearparticlesfor0%creepbySEM. Figure27–Frictioncoefficientasafunctionofslidingdistance. Figure28Wearzoneforpureslidingwith500MPaofthe:a)wheelspecimenandb)railspecimen. Figure29Wearchartforwearrateresortingtoslidingspeedandcontactpressure(hardness)[3]. Table 1 – Chemical composition of tested steels.

Alloy AISI 4140 (wheel) AISI 1055 (rail)

C(%) 0.37-0.43 0.50-0.60

Si(%) 0.15-0.40 0.15-0.40

Mn(%) 0.60-0.90 0.60-0.90

Cr(%) 0.85-1.15 ----

    Table 2– Typical roughness parameters of the test specimens.

Ra (m) 2.78

Rq (m) 3.41

Rz (m) 14.98

   Table 3 – Cross-reference table of the contact pressure and the slip ratio (slide %).

Contact Pressure

300 MPa 1%

Mo(%) 0.15-0.25 ----

HV30 311±4 206±3

500 MPa 500 MPa 0% 0.45 %

Linear Speed = 1 m/s

500 MPa 500 MPa 500 MPa 1% 1.75 % 15.23 % 1000 MPa 1% Creep ratio %

   Table 4 - Cross-reference table of the contact pressure and the linear speed.

300 MPa 1 m/s 500 MPa 500 MPa 500 MPa 500 MPa Contact Pressure 1 m/s 1.5 m/s 1.85 m/s 0.5 m/s 1000 MPa 1 m/s Creep = 1 % Linear Speed     Table 5 – Pure sliding test conditions.

Contact Pressure (MPa) 500

Sliding distance (m) 94

Linear Speed (m/s) 1

  

Table 6 – Friction coefficient for contact pressure values.

f  

300 MPa 0.15

500 MPa 0.11

1000 MPa 0.39

 Table 7 – Wear debris size and shape for rolling-sliding tests and different contact pressure.

a (mm) b (mm) r (mm)

300 MPa 0.42 0.26 0.17

500 MPa 0.5 0.29 0.19

1000 MPa (agglomerates) 1.26 0.65 0.45

 

Table 8 - Friction coefficient for sliding speed values.

f

 

0.5 m/s 0.31

1 m/s 0.11

1.5 m/s 0.20

1.85 m/s 0.20

Table 9 - Wear debris size and shape for rolling-sliding tests and different linear speed.

a (mm) b (mm) r (mm)

0.5 m/s 1.02 0.53 0.37

1 m/s 0.50 0.29 0.19

1.5 m/s 0.50 0.25 0.18

1.85 m/s 0.25 0.19 0.11

  

Table 10 - Wear debris size and shape for rolling-sliding tests and different slip %.

a (mm) b (mm) r (mm)

0% 0.77 0.43 0.29

0.45 % 0.48 0.26 0.18

1% 0.50 0.29 0.19

1.75 % 0.5 0.39 0.22

15.23 % 0.33 0.21 0.13

1% 0.11

1.75% 0.24

15.23% 0.27

 

Table 11 - Friction coefficient for creep ratio values.

f   

0% 0.05

0.45% 0.11

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6 a)

6 b).tif

Figura 7.tif

Figure 8.tif

Figure 9.tif

figure 10 a).tif

figue 10 b).tif

Figure 11.tif

Figure 12.tif

figure 13 a).tif

figure 13 b).tif

figure 13 c).tif

Figure 14.tif

Figure 15.tif

Figure 16.tif

figure 17 a).tif

figure 17 b).tif

figure 17 c).tif

figure 17 d).tif

figure 17 e).tif

Figure 18.tif

Figure 19.tif

Figure 20.tif

figure 21 a).tif

figure 21 b).tif

figure 21 c).tif

figure 21 d).tif

figure 21 e).tif

figure 22 a).tif

figure 22 b).tif

figure 22 c).tif

figure 22 d).tif

figure 23 a).tif

figure 23 b).tif

figure 24 a).tif

figure 24 b).tif

figure 25 a).tif

figure 25 b).tif

figure 26 a).tif

figure 26 b).tif

Figure 27.tif

figure 28 a).tif

figure 28 b).tif

Figure 29.tif