Freight Bogie Design Measures to Improve the Lifetime Performance of Switches and Curves

Paper 125 Civil-Comp Press, 216 Proceedings of the Third International Conference on Railway Technology: Research, Development and Maintenance, J. Pombo, (Editor), Civil-Comp Press, Stirlingshire, Scotland. Freight Bogie Design Measures to Improve the Lifetime Performance of Switches and Curves M. Hiensch 1,2, M. Linders 2, N. Burgelman 2, W. Hoeding 2 M. Steenbergen 1 and A. Zoeteman 1,3 1 Railway Technology, Technical University Delft, The Netherlands 2 DEKRA Rail, Utrecht, The Netherlands 3 ProRail, Utrecht, The Netherlands Abstract Rail infrastructure in the Netherlands is facing problems of high capacity demands with 24/7 operation, resulting in a limited availability of maintenance windows. This extensive usage demands a robust transport system with track and vehicle interface managed within a systems approach. From the viewpoint of overall system performance, infra manager ProRail has an interest in the level of track friendliness of vehicles that have access to their tracks. Bogie design improvement related to freight transport and their impact on the Dutch track is of special interest here. The research studies, presented in this paper, consider the potential of proposed design measures regarding the improvement in curving behaviour and switch negotiation of freight bogies and related wheel-rail contact stresses. For this purpose a sensitivity analysis has been carried out by means of track-train simulations within the VAMPIRE multi body simulation software. The research shows that the standard Y25L freight bogie design displays rather good track friendliness behaviour. Two of the evaluated bogie design modifications, both targeting the PYS characteristics, show the potential to further improve track friendliness. One other design modification has been shown to reduce the lateral track forces; this will expectedly lead to a reduction of track geometry degradation and related maintenance effort and cost. Keywords: turnout, switch panel, frequency selective stiffness, track friendliness, Y25 bogie. 1 Introduction The Dutch heavy-rail network is one of the busiest networks in the world. Every day, some 1.2 million passengers and 1, tons of freight are transported on nearly 7, km of railway line. The situation on the network is characterised by separated infrastructure management and profit-driven operators, some being paid 1

by regional governments through tenders. Netherlands Railways (NS) is by far the largest operator (some 8% of passenger train km). ProRail is the infrastructure manager, responsible for constructing, managing and operating the infrastructure. Infrastructure use is charged on a variable cost basis, which includes the tonnage borne by the infrastructure. The tonnage borne is constantly monitored by Quo Vadis weigh-in-motion (WIM) systems which are located at key points on the network and therefore capture practically all traffic movements in real time. Revenues from track access charges as well as government budgets (still by far the major income) are used by infrastructure managers, including ProRail, for constructing, managing and operating the infrastructure. Maintenance cost associated with wear and fatigue damage at the wheel-rail interface are dominating the budget of many infrastructure manager and train operating company. It can be understood from the occurring wear and fatigue features that damage initiation at rail and wheels is strongly connected to track geometry, especially curved track sections. Consequently the curving behaviour of a vehicle plays an important role in controlling wear and fatigue loading at the wheel-rail interface. Apart from track with radii smaller than 3 m, vehicle curving behaviour is particularly challenged at the diverging route through a turnout. Recent developments within bogie design are aiming, among others, to reduce wear and fatigue loading of track and wheels. Bogies which fulfil these conditions are considered track-friendly. From the viewpoint of overall system performance, ProRail has an interest in the level of track friendliness of the vehicles that have access to the track. Bogie design improvements related to freight transport and their impact to Dutch track are of special interest. The presented study assesses a number of potential measures regarding the improvement of curving behaviour of a railway freight bogie and related wheel-rail contact stresses. For this purpose a sensitivity analyses has been carried out by means of track-train simulations within the VAMPIRE multi body simulation software. The research demonstrates the potential benefits of freight bogie design optimization and associated performance, thereby providing insight in the possible contribution to a more sustainable rail transport. 2 Vehicle curving behaviour When negotiating a moderate radius curve, the rolling radius difference, built up by the lateral displacement of the conical wheels towards the outside of the curve, will create steering forces. These steering forces, when larger than the yaw resistance of the wheelset, lead to a more radial position of the wheelset. However, in sharper curves in general the wheelsets of railway vehicles suffer from under-steer. Because of the angle of attack lateral slip will be generated at the wheel-rail contact patch, where the guiding wheel of the leading axle normally is dominant with respect to tangential stresses and slip levels. Lateral slip of both wheels is directed towards the inner side of the curve. The lateral slip forces are balanced by the resultant flange contact force. These contact forces and resulting stresses can lead to plastic deformation, wear and rolling contact fatigue damage (RCF) at the rail head and wheel tread. Consequently, demanded maintenance is high; e.g. inspection, grinding, 2

repair welding and rail replacement at track and wheel re-profiling and replacement at the workshop. Figure 1: Wheel tread with RCF crack damage Figure 2: Severely worn narrow radius track section; wear debris at ballast and collecting magnet 2.1. Freight bogie Y25L In Europe the most commonly used bogie designed for freight wagons is the twoaxle bogie type Y25L-UIC (L stands for the French word Lourd which translates in English to Heavy). The bogie Y25L was developed by the French National Railway SNCF and standardised by the (International Union of Railways) UIC. As described in [1] the bogie Y25L primary suspension consists of a set nested coil springs (with a bi-linear characteristic for tare/laden ride) and a single Lenoir link providing vertical and lateral frictional damping. The friction force depends on the vertical load. The longitudinal axle box clearance is 4 mm, unidirectional, allowing a certain yaw angle. After this small amount of longitudinal clearance has been exceeded, the longitudinal stiffness will rise steeply. Therefore the wheelsets can 3

steer themselves in curves as long as the 4 mm clearance is not fully consumed. For curves with a radius below 6 m this clearance is exceeded, and relatively poor curving behaviour as well as high lateral wheel-rail forces are to be expected [2]. From the viewpoint of efficiency, freight wagons are often loaded to their maximum, reaching the allowed track axle load which in the Netherlands is 22,5 ton. Although in absolute numbers the annual freight transport in the Netherlands is relatively low compared to passenger traffic (4 vs. 144 mio. train km), poor narrow curving behaviour in combination with high axle loads can, depending on curve radii distribution, cause a disproportional contribution of freight transport to track degradation. Achieving improved track friendliness of freight bogies therefore can significantly contribute to improving the whole system performance, resulting in a more sustainable rail transport system. Identification and quantification of performance-based design principles can further support this development. Figure 3: The bogie Y25L is admitted for axle loads up to 22,5 ton; the maximum running speed for a wagon fitted with these bogies is 12 km/h when empty and 1 km/h when loaded 3 Measures to improve track friendliness To improve track friendliness of freight bogies, over the years a vast number of bogie design innovations have been considered and reviewed by the industry. A small number have made it to the testing phase and fewer to implementation. The SUSTRAIL project [3] has studied freight bogie design optimisation of the primary suspension by the use of double Lenoir link primary suspension, allowing opposite longitudinal displacement at the wheelset s both primary suspensions, thus facilitating larger yaw angles of the wheelsets while still utilising standard components. Further assessment of wheelset guiding design optimisation included the benefit of linkages providing longitudinal stiffness between the axle boxes using a radial arm. In [4], a study of freight vehicle effects on rail surface damage is presented. Regarding the vehicle design it was concluded that there are significant differences in the resulting rail surface damage between the different types of design. In particular, the characteristics of Primary Yaw Stiffness (PYS) should be determined accurately. An overview of further design optimisation measures is presented in [5]. These include features such as: Inboard bearings, reducing the unsprung mass of the bogie significantly; 4

Secondary suspension extended by a rubber ring mounted between the bogie frame and the lower part of the centre pivot, providing a certain amount of elasticity in all translational and rotational directions; Elastic side bearings, soft in lateral direction, to absorb lateral accelerations and allow bogie rotation and self-centering; Application of cross anchors to improve radial steering. The vast number of Europe s 4. strong freight wagon fleet, however, is still fitted with classic bogies. Quantifying the claimed benefits of these innovations in bogie design to the overall rail transport will support decision making and could push implementation. 3.1. Research scope The research presented here focuses on the improvement of switch negotiation and curving performance and related wheel-rail contact stress levels by optimisation of the steering behaviour of Y25 freight wagon bogies. Additionally, a brief study was performed regarding the effect of reducing the bogie unsprung mass on track degradation in terms of RCF and wear. From the above identified design optimisation measures a parametric study has been set up, investigating the effects of several freight bogie design parameters on rail surface damage. A sensitivity analysis was carried out by vehicle dynamic simulations, performed with respect to the characteristics of primary and secondary suspension, unsprung mass and cross anchor application. For some of the evaluated bogie designs, the wheelsets are connected to the bogie frame by so-called radial arms (also called trailer arm). These radial arms are connected to the bogie frame by pivot bushes. For this type of suspension, the primary yaw stiffness (PYS) is dominated by the radial stiffness of the radial arm pivot bush. This rotational stiffness of the wheelset within the bogie frame strongly determines the radial setting of the wheelsets in curves, having a direct effect on the contact conditions, the level of creepage and the lateral forces at the wheel-rail contact patch. Recent developments in bogie design are, among others things, the application of new elastic components with a characteristic that is dependent on loading frequency. As was presented in [6] these so-called Frequency Selective Stiffness (FSS) elements or Hydro bush elements can, when applied at the radial arm pivot bush, lead to a significant reduction of rail and wheel surface damage. For this reason, application of Hydro bush elements at the Y25 bogie is included in the analysis. The impact on running behaviour and track loading of all proposed Y25 bogie design modifications will be discussed in this paper. 4 Assessment of track friendliness Assessment of the level of track friendliness is carried out in terms of the variables curving behaviour, expected wear and RCF loading. The key parameter in this part of the assessment is the parameter Tγ ( T-gamma ), in which longitudinal and lateral tangential forces and creepages are combined to calculate the energy dissipated at the wheel-rail contact patch. Note that spin moment and spin creepage 5

terms are not included. The Wear number Tγ is a direct output from the VAMPIRE multibody analysis. Tγ can be used as an output value indicating the expected damage development with respect to wear and RCF development. This damage development can be derived from the damage function as presented in [7,8]. The energy dissipation value Tγ is determined at the first outer wheel for curves with different radii and for switch panel negotiation. The running stability and the impact on track degradation have been assessed through simulations on a tangent track. Figure 4: RCF-damage function for rail grade R22 [7,8] 5 Dynamic simulations input 5.1. Bogie and vehicle models Based on data presented in [9] a model of the Y25L bogie was set up in VAMPIRE. The simulation results presented in [9] were used for the validation of the bogie model. Based on the validated Y25 bogie model, six further bogie vehicle models were set up, in order to assess the following design modifications: 1. Bogie with double Lenoir linkage 2. Bogie with trailer arms, fitted with conventional bushes 3. Bogie with trailer arms, fitted with Hydro bushes 4. Bogie with reduced secondary horizontal stiffness 5. Bogie with reduced unsprung mass 6. Bogie with cross-anchors In order to quantify the benefits of any new design, a benchmark vehicle has been selected based on Y25L bogies and a 6867 HHA coal hopper wagon. The wagon body, including everything except bogies, is defined in terms of mass, inertia and centre of gravity. Both for the tare and laden condition (7 and 22.5t axle load, respectively). The wagon body rests on the bogie s secondary springs. Simulations were carried out in tare and laden condition. The coefficient of friction between block braked wheels and rail is set to f =.45, in correspondence to [4]. 6

5.2. Track models The track input for curving behaviour assessment consists of a set of right-hand curves with radii 25, 5, 75, 1, 15, 2 and 25 metres. The cant is set to 5 mm for the radii from 25 m to 1 m, to 3 mm for 15 m, and to mm for 2 m and 25 m. For each simulation the operational speed is set to a value that results in a cant deficiency of 3 mm. No track irregularities were applied to the curving track models (rail inclination 1:4, track gauge 1435 mm). Runs were performed with the UIC54 E5 (anti-head check) rail profile and new (unworn) S12 wheel profiles. The assessment of switch loading was carried out using a track model of the diverging route through a switch panel, with vehicle speed set to 4 km/h. The most common type of turnout applied in Dutch track has been selected. This turnout has a crossing angle 1 over 9, and a switch radius of 195 m as presented in [1]. In this simulation, measured rail profiles were applied in combination with measured S12 wheel profiles. Assessment of the running stability has been carried out using the VAMPIRE library design file Stabilit2, a tangent track with irregularities specially designed to assess running behaviour. For the assessment of the impact of running behaviour to track degradation, 1 m of measured tangent track was used with consisting shortwave vertical and lateral track irregularities. The Power Spectral Density (PSD) of measured track irregularities as a function of wavelength are presented in figures 5a and 5b. 15 a) Wavelength [m] Lateral shift PSD [m 2 /m] 1 5 2 4 6 8 1 12 14 16 18 2 Wavelength [m] Vertical shift 1 8 b) PSD [m 2 /m] 6 4 2 2 4 6 8 1 12 14 16 18 2 Figure 5a, b: The Power Spectral Density (PSD) of measured track irregularities, lateral and vertical, as a function of wavelength 7

The simulated train speeds varied with a maximum of 1 km/h when fully laden and 12 km/h when tare. Runs were performed with the UIC 54 E1 rail profile, and new (unworn) S12 wheel profiles. 6 Results Over 2 simulation runs were performed for this study. All simulations have been executed in VAMPIRE Pro, Version 6.3. All simulation results have been compared with the simulation results for the Y25L reference bogie. Only those results that produced noteworthy differences are discussed in detail. 6.1. Primary yaw stiffness (PYS) In this section the effect of PYS design modifications is discussed with respect to switch negotiation, curving behaviour and running stability. For the Y25L bogie design the yaw stiffness is determined by the longitudinal stiffness between the axle boxes and bogie frame. When starting the yaw rotation, the first 4 mm of longitudinal play is consumed at a relatively low stiffness value of approx. 1 kn/mm. After this 4 mm displacement, a nominal longitudinal stiffness value of 1 kn/mm is assumed in accordance to [9]. The analysed PYS design modifications are: a) Double Lenoir linkage, b) Trailer arms with conventional bush with linear stiffness of 2 kn/mm, c) Trailer arms with Hydro bush (static stiffness 2,8 kn/mm, dynamic stiffness 15 kn/mm). 6.1.1. Switch negotiation Figures 6a to d present the Tγ development for the different PYS design variations at contact patch areas of the first wheelset during switch negotiation in diverging direction. Upon entering the switch, the wheel flange of the leading right wheel comes into contact with the switch rail at approximately,5 m behind the switch toe. This results in a steep increase of Tγ. When comparing the Tγ peak values, it can be seen that these are not influenced by the examined PYS design measures. Significant differences however can be observed at the wheel tread/ flange root contacting area of the switch panel (position 4 to 48 m) and at both the contacting area of flange and wheel tread/ flange root of the closure panel (position 48 to 63 m). For the design variant double Lenoir and trailer arm with Hydro bush flange contact at the closure panel becomes negligible, eliminating side wear loading. The application of trailer arms with conventional bushing leads to an increase of PYS compared to the reference situation, resulting in Tγ values entirely in the region of Wear. By applying the damage function as shown in figure 4, the expected effect of the individual modifications can be understood more clearly. Table 1 presents for each PYS modification the expected damage at switch and closure panel, based on the operational wear number from the guiding right wheel. 8

18 16 14 Tgamma leading wheelset, Double Lenoir, 1:9 turnout a) b) 1L 1R 1FL 1FR 18 16 14 Tgamma leading wheelset, Reference, 1:9 turnout 1L 1R 1FL 1FR 12 12 T [J/m] 1 8 T [J/m] 1 8 6 6 4 4 2 2 35 4 45 5 55 6 65 7 Position [m] 35 4 45 5 55 6 65 7 Position [m] 18 16 14 Tgamma leading wheelset, Trailer arm, 1:9 turnout c) d) 1L 1R 1FL 1FR 18 16 14 Tgamma leading wheelset, Trailer arm with Hydro-bush, 1:9 turnout 1L 1R 1FL 1FR 12 12 T [J/m] 1 8 T [J/m] 1 8 6 6 4 4 2 2 35 4 45 5 55 6 65 7 Position [m] 35 4 45 5 55 6 65 7 Position [m] Figures 6a to d: Tγ development for the different PYS design variations at the contact patch areas of the first wheelset during switch negotiation. The right wheel (R) is guiding, 1R/L: wheel tread/ flange root contact, 1FR/FL: flange contact Modification Switch Panel contact area Closure Panel contact area Flange Tread/ flange root Flange Tread/ flange root Y25 - reference Wear Wear RCF Wear Double Lenoir Wear Wear Below damage RCF threshold Trailer arm with Wear Wear Wear Wear conventional bush Trailer arm with Hydro bush Wear Wear Below damage threshold Wear Table 1: Rail damage development type, based on Tγ loading of the guiding right wheel 9

6.1.2. Curving behaviour Figure 7 presents the simulation results for the different PYS design variations and different curve radii. 14 Tgamma flange leading wheelset in curves, cant deficiency = 3 mm RMS Tgamma [J/m] 12 1 8 6 Y25 (reference) Trailer arms Trailer arms with cross anchor Trailer arms with hydro bushes Double Lenoir suspension Reduced unsprung mass Reduced secondary horizontal stiffness 4 2 5 1 15 2 25 3 Curve radius [m] Figure 7: Tγ development for the different design variations and curve radii It shows the quadratic mean (RMS) Tγ value of the outer wheel of the leading wheelset, determined in the full curve. These RMS Tγ values are plotted against the curve radius for each design modification. The results in figure 7 show that improved curving behaviour, compared to the Y25 reference bogie, only occurs for radii below 75m. This positive effect is only present for two of the evaluated design modifications; the double Lenoir and trailer arm with Hydro bush, and to a lesser extend for the reduced secondary horizontal stiffness. The best narrow radius curving performance is shown by the double Lenoir design, with resulting Tγ values below the damage threshold value down to the 5 m radius curve. As was the case with switch negotiation, it can be seen here that the trailer arm bogie design with conventional bush results in increased Tγ values due to the increased PYS. For the assessed operational conditions, the influence of the cross anchor is negligible. 6.1.3. Running stability Of the two design modifications with best curving behaviour, being the double Lenoir and the trailer arm with Hydro bush, running stability were evaluated. Figures 8a, b and c present the lateral accelerations for these design variations, laden at 8-9-1 km/h. Both the double Lenoir and trailer arm with Hydro bush show an improvement in stability compared to the reference situation. 1

4 3 a) Carbody lateral acceleration, Reference, Laden, Stabilit 8 km/h 9 km/h 1 km/h 4 3 b) Carbody lateral acceleration, Double Lenoir, Laden, Stabilit 8 km/h 9 km/h 1 km/h 4 3 c) Carbody lateral acceleration, Hydro-bush, Laden, Stabilit 8 km/h 9 km/h 1 km/h 2 2 2 Acceleration [m/s 2 ] 1-1 Acceleration [m/s 2 ] 1-1 Acceleration [m/s 2 ] 1-1 -2-2 -2-3 -3-3 -4 2 4 6 8 1 12 14 16 18 2 Position [m] -4 2 4 6 8 1 12 14 16 18 2 Position [m] -4 2 4 6 8 1 12 14 16 18 2 Position [m] Figures 8a, b, c: Lateral accelerations for reference and two design modifications 6.2. Reduced secondary horizontal stiffness The impact of a rubber ring mounted between the Y25L bogie frame and the lower part of the centre pivot was examined by introducing a lateral stiffness to the centre bowl set to.2 kn/mm; a limit value proposed in [11]. The resulting Tγ values for switch negotiation are not significantly influenced by this design modification, it has only a minor influence on curving behaviour. 6.3. Reduced unsprung mass The effect of reducing the unsprung mass of the bogies on track degradation is investigated by examining the forces between wheel and rail in lateral and vertical direction. Reduction of the unsprung mass is implemented in the vehicle model by reducing the mass per wheelset with 4 kg; additionally the mass per bogie frame is reduced with 2 kg. This results in a mass reduction of 1 kg for each bogie (approx. 2% reduction) and implies 2 kg of extra freight. As mentioned in section 5.2 the track section applied for this analysis is tangent track with measured track irregularities. The applied vehicle speed is 1 km/h, laden. For the laden condition 22,5 ton axle load is applied. The power spectral density of Tγ as a function of the wavelength is presented in figure 9a. It can be seen that a reduction of the unsprung mass results in a significant reduction in Tγ loading. The dominating wavelength of 9 m corresponds to the wavelength of the Klingel motion of a Y25 wheelset. The lateral forces power spectrum is presented in figure 9b. Reducing the unsprung mass predominantly has resulted in a significant reduction of lateral track forces and associated Tγ. Reduced lateral track forces can expectedly lead to a reduction in track geometry degradation and related maintenance effort and cost. For the vertical track forces, the observed reduction is much smaller. In the simulations the track geometry input data did not comprise lateral wavelengths below 6m and vertical wavelengths below 8m. It should therefore be noted that track geometry as registered by a measurement coach is not a fully adequate parameter to determine the effect of unsprung mass on track degradation. 11

T as function of wavelength 12 35 Reference Reduced unsprung mass 3 1 a) b) Lateral track force as function of wavelength Reference Reduced unsprung mass PSD [N 2 /s] 8 6 4 PSD [N 2 /s] 25 2 15 1 2 5 2 4 6 8 1 12 14 16 18 2 Wavelength [m] 2 4 6 8 1 12 14 16 18 2 Wavelength [m] Figure 9a,b: Tγ and lateral force as a function of the wavelength of the leading wheelset for the reference bogie and the bogie with reduced unsprung mass The unsprung mass has a significant effect on the vertical dynamic wheel-rail force for wavelengths that are shorter than or in the same order of magnitude as the wheel circumference, which is about 3 m. This implies that the unsprung mass plays a crucial role in the degradation velocity of points and other switch parts, of insulated joints and potentially of welded connections with insufficient straightness. These components are key assets in the performance of a railway network. For conventional train speeds, the first derivative of the trajectory of the gravity centre of the unsprung mass may exhibit a discontinuity, implying wheel-rail impact conditions (see ref. [12]). This is also true in lateral directions. It has been demonstrated in [12] that in these cases the dynamic contact force is proportional to the mass, which is consistent with both theoretical and field results for P1 peak forces from the literature [13]. The level and frequency content of the dynamic contact force have a direct relationship to the degradation rate of concerned railway structural assets, even apart from the parameter Tγ and induced wear/rcf. It is therefore useful to strive for minimum inertia of unsprung bogie parts. 6.4. Cross anchors A cross anchor is a pair of linkages applied diagonally between the wheelsets. These wheelsets, when yawed by the wheel-rail contact forces, will assume a radial position. Providing diagonal linkages between the wheelsets of the Y25L reference bogie is however not useful, since the combination of these linkages with the single Lenoir dampers would prevent all longitudinal movements of the wheelsets relative to the bogie frame. Therefore, in this research the cross anchor is limited to a combination with the design with trailer arms and conventional bush. The resulting Tγ values do not show a significant influence of this design measure on running behaviour; for the switch negotiation as well as for curving, the found Tγ levels are corresponding to those for the trailer arm without cross anchor. 12

7 Discussion To improve track friendliness of freight bogies, over the years a vast number of bogie design innovations have been considered and reviewed by the industry. The proposed modifications especially target curving behaviour, aiming to reduce track loading during curving and while negotiating a turnout in the diverging route. A parametric study has been carried out, investigating the effects of several freight bogie design parameters on rail surface damage. A sensitivity analysis was carried out by vehicle dynamic simulations, performed by varying the characteristics of primary and secondary suspension, unsprung mass and cross anchor application. 7.1. Switch negotiation The track loads resulting from simulations in the switch panel demonstrate wear to be the dominant damage type for the assessed 1:9 turnout. This is the case for they25 reference design, as well as for the assessed design modifications. Clearly the Tγ peak values occurring when entering the switch panel cannot be reduced by lowering the PYS value. At the closure panel, with curve radius 195 m, a significant difference is observed for two of the assessed design modifications, being the double Lenoir linkage and the trailer arm with Hydro bush. For these modifications, wheelrail wear at the flange (gauge face) becomes negligible. Simultaneously the Tγ level at the wheel tread/ flange root contacting area is reduced, shifting from wear into the RCF region for double Lenoir, where the Hydro bush loading level remains in wear. The other investigated design modifications (trailer arm with conventional bush, reduced secondary horizontal stiffness, reduced unsprung mass and the cross anchor application) did not result in a significant optimisation in switch loading. 7.2. Curving behaviour For the assessed operational conditions, the reference Y25 bogie displays excellent curving behaviour for all radii down to the 75 m radius, keeping operational Tγ levels below the RCF damage threshold. At curve radius 5 m Tγ values sharply increase, resulting in RCF damage development for the reference situation. At 5 m radius, only the double Lenoir design remains below the damage threshold. Although the trailer arm with Hydro bush design results in reduced Tγ values compared to the reference bogie, its Tγ level at 5 m radius lays within the RCF region. At 25 m radius, both double Lenoir and Hydro design modifications are experiencing RCF damage development. Of the other investigated design modifications, only the reduced secondary horizontal stiffness leads to improved curving, although minor. 7.3. Vehicle stability/ track degradation Running stability was examined for the two design modifications with the best curving behaviour, being the double Lenoir and the trailer arm with Hydro bush. 13

Both modifications show improved stability compared to the reference situation. The effect of reducing the unsprung mass of the bogie on track degradation, has been assessed by evaluating the forces between wheel and rail in lateral and vertical direction. It can be observed that reducing the unsprung mass predominantly has resulted in a significant reduction of the lateral track forces, and related Tγ. The observed reduction is closely related to the Klingel movement of the bogie. Reduced lateral track forces will expectedly lead to reduced track geometry degradation and related maintenance effort and cost. This potential will be quantified in further research. 8 Conclusion The potential of proposed design measures regarding the improvement of curving behaviour and switch negotiation of railway freight bogies and related wheel-rail contact stresses have been studied. For this purpose, a sensitivity analysis has been carried out by means of track-train simulations within the VAMPIRE multi-body simulation software. Curving behaviour of the standard Y25L freight bogie design, with respect to expected wear and RCF development, is shown to be rather good, keeping operational Tγ levels below the RCF damage threshold for a large range of radii. Only for the evaluated curve radii of 5 and 25 m, RCF development can be expected for the standard Y25L design. Two of the six evaluated bogie design modifications, being the double Lenoir linkage and the trailer arm with Hydro bush, deliver a distinctive contribution to track friendliness of the Y25L freight bogie. For the assessed conditions the double Lenoir modification expands the curve radii range with Tγ operational levels below the RCF damage threshold to 5 m. Both double Lenoir and trailer arm with Hydro result in improved switch negotiation, especially for the closure panel at which for the assessed situation flange wear loading is almost eliminated. For the assessed operational conditions, both design modifications also show improved stability compared to the reference situation. Other design modifications, among which the application of cross anchors, did not significantly contribute to improving track friendliness under the evaluated conditions. The research shows that the standard Y25L freight bogie design possesses rather good track friendly behaviour. For two bogie design modifications, both targeting the PYS characteristics, the potential is identified to provide further improved track friendliness. Reducing the unsprung bogie mass has shown to reduce the lateral track forces, which will expectedly lead to a reduction in track geometry degradation and related maintenance effort and cost. The potential benefits of the identified freight bogie design optimisations and associated performance however require further research, providing a more detailed insight in the possible contribution of new bogie designs and their impact on sustainability of railway assets. This insight can then help to determine whether these new designs can result in an improved situation for the whole system. 14

Acknowledgements The authors wish to acknowledge the support of ProRail in funding this study. References [1] A. Orlova, Y. Boronenko, Handbook of railway vehicle dynamics, the anatomy of railway vehicle running gear, 26. [2] P-A. Jönsson, Dynamic vehicle track interaction of freight wagons with link suspension, KTH Docteral Thesis, 27. [3] S. Iwnicki et.al, The SUSTRAIL high speed freight vehicle: Simulation of novel running gear design, IAVSD13-3, 2-ID49. [4] T. Tunna, S. Clark, C. Urban, A Study of Freight Vehicle - Effects on Rail Surface Damage, Office of Rail Regulation, UK TTCI(UK), Ltd.31 May 26 [5] D. Scholdan, Freight bogie overview RWTH-IFS-Seminar, Aachen, 214. [6] M. Hiensch et al., Improving track-friendliness of rolling stock, Proceedings of International Heavy Haul Association (IHHA 215), Australia, 215. [7] M.C. Burstow, Whole life rail model application and development for RSSB development of a RCF damage parameter AEATR-ES-23-832 Issue 1 October 23 [8] M.C. Burstow, Whole life rail model application and development for RSSB continued development of an RCF Damage Parameter, AEATR-ES-24-88 Issue 2 September 24 [9] M. Molatefi, M. Hecht, M.H. Kadivar, Critical speeds and limit cycles in the empty Y25-freight wagon, Porc. IMechE Vol. 22 Part F: JRRT67, 26. [1] M. Hiensch, P. Wiersma, Improvement of structural performance of the switch panel by enhancing track friendliness of trains, Proceedings of Contact Mechanics (CM215), Colorado, USA, 215. [11] P. Shackleton, Suspension simulation of a rail freight vehicle for LDHV goods, EU Seventh Framework Programme SPECTRUM, presentation Paris, 8th April 215. [12] M. Steenbergen (27). The role of the contact geometry in wheel rail impact due to wheel flats. Veh. Syst. Dyn. 45 (12), 197-1116. [13] H. Jenkins, J. Stephenson, G. Clayton, G. Morland, D. Lyon (1974). The effect of track and vehicle parameters on wheel/rail vertical dynamic forces. Railway Eng. J. 3 (1), 2 16. 15