Abstract
This paper builds on earlier work [1] and describes a full-scale programme of tyre tests using a purpose-built tyre test rig. The tyre being studied here is a specialized tyre intended only for use with Road-Rail Vehicles (RRVs). As the name suggests these vehicles can operate on both the road and railway tracks. When operating on railway tracks, they can be used for a number of purposes. One of these is to carry out shunting operations moving heavily loaded waggons around a railway yard. In the earlier publication [1] for this work the design, manufacture and commissioning tests for a bespoke tyre test rig that could operate to test pneumatic tyres on rails was described. That rig is summarized again here along with the description of the tyre tests that consider variations in load, variations in speed and testing in wet and dry conditions. An initial study is also described where thermal imaging has been carried out to look at the temperatures generated in the tyre when operating on a rail head. For all of the tests the focus was on the generation of longitudinal tractive forces as a function of slip, this being the most important characteristic of an RRV tyre when considering the pulling power of an RRV shunting engine. The results of this study provide new insights into the friction acting between an RRV tyre and railway track and form the basis for an ongoing study to develop a new predictive tyre model to be used in simulations that can support the design of an RRV.
Introduction
Background
This paper builds on work describing the build and commissioning of a bespoke tyre test rig used to test pneumatic tyres fitted to Road-Rail Vehicles (RRVs). That study 1 was the first and to date only work carried out to understand the tyre forces generated in the tyre to rail contact patch.
As the name suggests RRVs are specialized vehicles that can operate both on the road and on railway tracks. They have been in use by rail operators since the middle of the last century when it was realized that a pneumatic tyre could generate significantly more traction on a rail head than a traditional steel rail engine wheel. Figure 1 shows some early examples of these vehicle where it is clear the development approach was to take an existing road vehicle and convert it so that it can have the additional capability to drive on railway tracks. The evolution of RRVs occurred in both Europe and the USA where the vehicles are more commonly referred to as ‘Hi-Rail’.
The use of RRVs for passenger transport was introduced on the Paris rail network in 1951 and has evolved since then. The latest version of those trains, shown in Figure 1 on the left, was brought into service in 2013 as driver-less trains. The standard rail track of this specific metro line is modified in the acceleration and deceleration sections close to and in the stations. The modification consists of a flat section running along the track at the same height as the tracks top flange. The upper face of this flat section is plated with a sandpaper type coating. The metro train is fitted with rubber tyre road wheels at either end of each railroad axle, sitting next to the rail wheels. These road wheels roll on the sandpaper type flat section providing acceleration performance similar to road vehicles, thus allowing for shorter train sequences and higher passenger frequency on those metro lines.
Another important application of RRVs is to support the construction, maintenance and any emergency operations on a rail network. This is important, particularly in remote or mountainous region where the ‘last mile’ can only be accessed using the rail network. An example for this type of RRV is shown on the right of Figure 2 for a fire engine that is used on the Swiss rail network.
Paris metro with pneumatic drive for maximum acceleration and deceleration (left) and Fire service RRV on a Swiss mountain railway (right).
As mentioned earlier RRVs are used to support shunting operations where the improved traction generated by a pneumatic tyre allows more waggons or larger loads to be moved around a rail yard. An example of this application is shown in Figure 3 for a commercial RRV shunting engine.
Rotrac RR24 RRV engine with a shunting capability of 2400 tonnes.
The use of RRVs for shunting purposes has increased over the last decade as the transport logistics sector is under pressure to adopt a more economical and eco-friendly approach. This has led to the classic, dated, heavy and expensive shunting engines with big diesel engines being replaced by cheaper purpose-built RRVs. Shunting RRVs are guided on the rails via traditional steel rail wheels, but the traction relies on the rubber road tyres sitting on the rail’s head. As the tyre to rail friction is higher than that for a steel wheel an RRV can be much lighter than a normal shunting engine pulling the same loads.
Despite these clear advantages the design and development of RRVs has lagged significantly behind the methods used in the mainstream road and rail industries. RRVs are often purpose-built one-offs designed to meet customers’ needs. To a potential customer, the highest priority is to know whether a certain train or set of waggons with a certain mass can be confidently hauled between two or more dedicated places on clearly specified tracks and under what weather conditions the shunting duties can be performed. Providing this information before a vehicle build requires the use of computer models for simulation during the design stages.
The provision of a model that can predict the performance of an RRV requires a mathematical representation of the intended vehicle and the environment that it can interact with. For the vehicle itself it is necessary to represent mass and inertial effects, springs, dampers and other compliant elements. Within the automotive, aerospace and rail industries the use of a Multibody Systems (MBS) approach has become widely established over the last 50 years. In addition to the modelling of the physical components, these models also need to represent how the vehicle interacts with the road, terrain, runway or railway track. In many of these applications this requires a mathematical representation (tyre model) that determines the forces generated at the interface between the tyre and relevant surface. For the work described in the rest of this paper the focus is on a programme of tyre tests that could be used to understand the behaviour of the tyre that would enable future work to develop a model that can represent the interaction of a tyre with a rail track and enable the simulation of an RRV shunting operation.
Tyre modelling and test methods
There is no existing literature or any form of study that academically examined the traction forces in the contact patch between a rubber tyre and a rails top flange. But there is a large body of work covering testing and modelling when it comes to rubber friction and tyres running on different surfaces for operations associated with road vehicles, race vehicles, off-road vehicles, agricultural vehicles or aircrafts.
Friction models are mathematical representations that describe the behaviour of friction forces between surfaces. They allow for a prediction of how rubber interacts and moves relative to another surface under various conditions. Rubber friction is considered to have four components: adhesive friction, hysteretic friction, viscous friction and abrasive friction. Of these the viscous and abrasive components are very small and are ignored in most mathematical models. The key references in this area are the work of Kummer, Grosh, Persson and Lorenz. Refs4–6 have all contributed to the understanding and simulation of tyre-road interactions by integrating realistic friction characteristics into tyre models, leading to more accurate predictions of tyre behaviour under various driving conditions.
Adhesive friction refers to microscopic interaction forces and chemical bonding between two surfaces. The energy required to break these bonds is one contribution to the overall frictional force. 7 investigated the role of adhesion in tyre-road interactions and its importance for grip.
A second contribution to the overall friction force is the hysteretic friction. It results from the energy dissipation within a viscoelastic material, like rubber. The energy dissipates as it deforms and recovers while being moved over a rough surface. The energy loss during these deformations contributes to the overall frictional forces. 8 undertook foundational research on the viscoelastic properties of rubber and emphasized the importance of hysteresis in rubber friction. Lorenz et al. expanded on the hysteresis models by exploring how temperature 9 and frequency 10 affect energy dissipation in viscoelastic materials. He provided experimental evidence to support these theories.
A third contribution to the overall friction force is the viscous friction. It arises from the resistance of fluid layers between surfaces in motion. This type of friction is proportional to the relative velocity and depends on the fluid’s viscosity. 11 especially contributed to the area be examining lubricated patches between rubber and the road surface that are sealed in by contact between rubber and surface. Persson et al. 11 called it the sealing effect.
The fourth contribution to the overall friction force is the abrasive friction. It occurs due to the mechanical interaction and removal of material when one surface wears or scratches another. It is influenced by the hardness and roughness of the surfaces. 7 already highlighted how surface texture and material properties contribute to abrasive wear also in tyre applications.
In summary, the primary contributors to rubber friction are adhesive friction for 50%–60% and hysteretic friction for 30%–40%. Adhesion is playing a significant role through microscopic interactions and chemical bonding. Hysteresis arises from energy dissipation within the viscoelastic rubber material as it is loaded and unloaded while travelling over abrasive surfaces. As a result, for the scope of this work on rubber tyre on rail friction, the focus will be on these two components, with the viscous and abrasive components considered to be negligible.
Within the automotive industry there are tyre models for different kinds of vehicle dynamics simulations such as ride, handling or durability. There are empirical tyre models, such as Pacejka’s well stablished Magic Formula, 12 the Fiala tyre model 13 or the Harty tyre model. 14 These models are empirical and use mathematical functions to represent the physical behaviour of the tyre based on observations and measurements taken during a programme of tests on a physical tyre. These models are traditionally used by the automotive industry to simulate the vehicle dynamics handling tests performed at a vehicle proving ground. For these applications the Magic Formula Tyre model is the most widely used.
The second class of models makes use of a physical representation of the contact patch of the tyre carcas and tread material. The most detailed of these are finite element models. These are mainly used by tyre manufactures to support the design of automotive 15 and aircraft tyres. 16 These models are however computer intensive and require more modelling effort. As such, a more efficient class of semi-physical tyre models has evolved that use masses and force elements that can capture the mechanical performance of the tyre. Typical representatives of this type of model currently in use are the Ftire, 17 CDTire 18 and TameTire 19 tyre models. Within the automotive industry these models are mainly used for ride and durability studies that can include events such as kerb strikes or pothole impacts. The most advanced tyre models can also include the effects of speed and temperature, providing more precise predictions of forces and moments under different conditions.
It is important to note that although the work involved with testing, modelling and understanding tyre behaviour is well established in the automotive and aerospace industries the work presented here is one of the first to look at tyres operation on rails to support the design of RRVs.
In the original paper associated with this study 1 a review was provided of existing methods to test tyres, so a summary only is provided here. Laboratory testing using test rigs such flat bed and drum machines provides a controlled environment to support repeatability. Flatbed test rigs are most useful for generating measurements that can be used to parametrize empirical tyre models such as the Magic Formula Tyre model. Capturing the tyre behaviour over the wide range of conditions that can be experience in operation on a vehicle can be an extensive and expensive process and involves technical specialists within the industry. Trying to obtain the optimal information in the most efficient manner was a challenge addressed by, 20 using the Calspan tyre test facility in the USA.
Drum tyre test machines have also been used to obtain measurements for empirical models but have a disadvantage due to the curvature effect at the contact point between the tyre and the drum. This can be compensated by using a drum with a large diameter but cannot completely represent a tyre running on a flat surface. In the past researchers have also looked at developing routines that account for this effect when generating Magic Formula tyre model parameters. 21 Some drum test machines offer the advantage of testing the tyre rolling inside the drum to investigate the effects of contamination by water or sand on the behaviour of the tyre. Drum testers can be equipped with cleats at different angles to generate modal responses in tyres in order to generate parameters for semi-physical tyre models such as FTire, 17 CDTire 18 and TameTire. 19
Testing outdoors can offer the advantage of more real-world conditions, albeit with challenges such as changing weather. The usual methods involve using a trailer or a truck such as the facility used by, 22 formerly known as TNO. Jaguar Land Rover has developed its own vehicle-based tyre test rig. 23 The Vehicle Based Objective Tyre Testing (VBOTT) is a Range Rover equipped for tyre testing. It offers realistic evaluations on actual road surfaces, capturing detailed data on traction, wear and handling in varied conditions.
The most comparable of existing methods studied would be that involving a trailer. Figure 4 shows the trailer-based tyre test rig developed by Mavros at Loughborough University, which provided useful direction for the development of the rail-based rig developed in this study.
Trailer based tyre test facility.
In the next section the method used to test tyres on rails is summarized together with a full-scale test programme used to develop an understanding of the tyre behaviour and generate parameters for a new empirical tyre-rail model.
Road rail vehicle tyre testing
Test rig and programme
The purpose-built tyre test rig designed and used in this project was built and commissioned at the Müller site in Frauenfeld, Switzerland making use of its own section of rail track within its premises that also connects with the main Swiss rail network. The design, build and commissioning of this test rig was described in the original paper. 1
To obtain a full data set for an automotive tyre can involve recording data for a wide range of controlled tyre states including variations in load, slip angle, camber angle and slip ratio. This study focuses on maximizing traction force. Therefore, the test rig was designed to measure longitudinal force and longitudinal slip. The handling properties of an RRV tyre are of minor importance as the vehicle’s trajectory is dictated by the guide wheels and rails. By measuring vertical load, longitudinal load, test wheel speed and the speed of the real forward motion, wheel slip and effective rolling radius can be calculated.
The design of the test rig is illustrated by the graphic shown in Figure 5 accompanied by a photograph showing the test wheel sitting on the rail-head. An important feature of the design is the capability to vary the load on the tyre through the use of a hydraulic ram.
Tyre test rig design and positioning of the test tyre on the rail-head.
Photographs of the test rig together with the shunting engine used to pull it along the track are shown in Figure 6. As can be seen, the test rig needs to be ballasted with enough load to ensure that the hydraulic ram will not lift the test rig structure from the rails. This is essential to ensure that the maximum normal load can be applied and maintained at a constant value during each test run. For ease of transportation, the ballast cubes can be transported separately.
The RRV tyre-rail test rig in operation.
In all of the tests described here, the rig was driven in a straight line and the tyre was braked to vary the slip ratio between 0 for a free rolling tyre, and 1.0 for a fully locked and sliding tyre. The tyre was also inflated to its rated pressure, and the tests were performed on clean rails, at zero slip angle. The range of tests performed is shown in Table 1. The intervals chosen for the speed range may appear random but were dictated driven by the need to maintain a steady constant velocity and to drive the shunting engine in a selected gear throughout the test. This is required to maintain a constant speed. In terms of understanding the dependency of the tyre behaviour on speed and developing a tyre model this has no detrimental effect. The combination of the condition, load and speed parameters resulted in a programme of 18 tests.
Parameters and test conditions.
For every test condition, wet (test 1–9) or dry (test 10–18), a test sequence at all three loads in combination with all three speeds was run. Each test sequence consisted of three test runs to ensure repeatability. To further maximize the data recording each test run allowed for two to five testing events, where the tyre was brought to a standstill and then the run was repeated until the length of available test track was covered. The number of events within each test run was limited by speed, with more events possible at lower speeds for the fixed length of available test track. The overall objective was to take the opportunity to record as much data as possible using the facility.
The purpose-built test rig used in this study was designed to brake the test wheel but not to accelerate it, the assumption being that the force generating characteristics for a braked or driven wheel are similar. The tests were carried out at low speeds representative of those in normal shunting operations. The thermal behaviour of the tyre was observed in this study but the dependency of forces on temperature has not been considered in the modelling.
The tyre used in this study, shown on the left in Figure 7, was a Continental HGV radial tyre with a dimension of 10R22.5 fitted with a Pneu Egger tread-less bandage specifically designed for road-rail use. This is a common method used to modify a standard tyre for road-rail use and ensures a smooth area of tyre where the tyre meets the rail head to ensure the maximum contact area of rubber with steel where the tyre meets the rail head. As with motorsport and the use of ‘slicks’ this works well when the conditions are dry but is more affected when conditions are wet. Hence, in this study an important consideration was to test in both the dry and the wet.
Continental 10R22.5 HGV tyre with tread-less bandage.
The way that a road-rail tyre sits on the rail head is completely different from any other application with a tyre as shown on the right in Figure 6. Due to the need to avoid any contact with fixed installations along the side of the railway track, the top flange of the rail does not sit in the centre of the tyre but is slightly offset depending on the tyre’s actual width. For most applications, the maximum tyre width is also restricted to 210 mm. There is a very limited choice of suitable tyres, and Continental provide the tyre with the highest load rating for road use (There is no specific load rating for rail use). For this reason, most shunting RRVs use the above-mentioned tyre type with or without tread.
Test results
In this section, a selection of graphs is provided plotting braking force against slip ratio. As stated in 1 it should be noted that in order to generate a model that can predict the performance of the RRV on the rails the emphasis is on the amount of traction that can be generated and hence the overall pulling power of the vehicle. This can be understood from inspection of Figure 8 below and consideration of points A, B and C. Point A is at zero, point B is where the braking force initially peaks and point C defines the force generated for a slip ratio of 1.0. The behaviour is similar to a tyre on the road. Between point A and B the tyre response is dominated by elastic behaviour. Between point B and C the tyre response is dominated by sliding behaviour.
Main data points needed to define the tyre-rail tyre model.
The gradient measured at point A defines the longitudinal stiffness of the tyre. This is critical for a road vehicle tyre model but is not so significant for a model that predicts the overall pulling power of slower moving RRVs. It should also be noted that in these tests braking force was measured. For this work it is assumed that there is a symmetric relationship between traction force and braking force. As such, the graphs are labelled on the y-axis as Traction Force.
Figure 9 shows the results for braking force versus slip ratio for a tyre load of 25.5 kN and a speed of 4 kph. For comparison, the results obtained on dry and wet tracks are plotted on the same graph.
Braking force versus slip ratio (test load 25.5 kN, test speed 4 kph).
The results show, as expected, that the traction force is greater in the dry than in the wet. The peak values for traction force in the dry is 12 kN and in the wet is 8 kN. Both values peak at a slip ratio of 0.2. This is similar to the behaviour of tyres on the road and would be the target operating point for an anti-lock braking system (ABS) or traction control. The results also indicate that the longitudinal stiffness of the tyre is greater in dry than wet conditions. This is also behaviour that would be expected for tyres on the road, but as stated earlier is not of immediate importance for a future RRV tyre/rail model.
Dividing the peak tractive force by the tyre load gives maximum coefficients of friction for the dry and the wet of 0.47 and 0.31 respectively. These initial results are already of use in RRV design. Simple calculations for an RRV of known mass, and when driven steadily at a constant speed, would provide a prediction of both the pulling power of the RRV on the flat and the ability to drive on an incline. Extending this to a tyre model would allow more complex multibody system simulations for RRVs in real world operating scenarios such as when RRVs are pulling a train traversing a bend or different inclines at one time.
These initial values indicate that the friction values for the tyres on a rail are less than 50% of the values that would be obtained for tyres on the road. The friction generated between a tyre and a test surface has two major components, adhesion and hysteresis. The hysteresis component relies on both the macro-texture and the micro-texture of the road or rail surface. The adhesion component of friction relies on molecular bonding between the two surfaces. In wet conditions, the film of water contaminant between the two surfaces severely compromises the generation of adhesive friction, and the tyre is very reliant on the hysteretic component.
The head of the rail is relatively smooth when compared with a granular road surface, and so it can be assumed that there is no macro-texture contribution to the hysteretic component. This will be a contributing factor to the indication from this test that tyres on rails appear from the initial test to generate far less friction as tyres on roads in both wet and dry conditions. The real situation will be more complex than that. The tyre also has a narrower contact patch compared with a tyre on the road. This will lead to the generation of shear forces over a smaller area for an equivalent load and slip ratio. A full understanding would require an analysis of the shear stress distribution throughout the contact patch and the transition between elastic and sliding behaviour in the contact patch.
Another interesting aspect of this initial test is the friction generated in the wet. It might be imagined that the rail-head is very smooth and polished surface. If that was the case, then the tractive force would be close to zero in the wet. The fact that a friction coefficient of 0.31 was achieved indicates hysteresis is in play and that the rail has a micro texture that is contributing to its generation. 8
A final feature of this initial test is the tractive force generated in both the wet and the dry when the slip ratio exceeds 0.2. In this region the friction is dominated by sliding behaviour. In the dry the value of tractive force is maintained at a constant. Similar behaviour has been noted experimentally 10 albeit just at very low velocities and when studying rubber on glass in dry conditions. 24 As sliding increases, the rubber heats up and the temperature increase benefits the adhesion. For the initial test here, it can be seen that in the wet there is a slight drop in tractive force at higher slip ratios. This could be due to the cooling effect of the water and the fact the hysteretic friction does not benefit as much from temperature rise as adhesive friction.
In Figures 10 to 12 the traction force is plotted against slip ratio, for dry conditions at the three different speeds. In each figure the graph is plotted for the three test loads.
Test speed 4 kph, dry (varying test load).
Test speed 8.4 kph, dry (varying test load).
Test speed 16.3 kph, dry (varying test load).
The graphs show that as with tyres on the road the traction force increases with load and that the increase is not linear. At the lowest speed the traction force remains approximately constant after reaching the initial peak at a slip ratio of about 0.2. For the higher speeds this behaviour changes and the traction force, and hence the coefficient of friction, continues to increase steadily until the wheel is fully locked at a slip ratio of 1.0. This can be due to increase of kinetic and viscous friction with temperature10,24 or local wear. In Figures 13 to 15 the traction force is plotted against slip ratio, for dry conditions at the three test loads. In each figure, the graph is plotted this time for the three test speeds.
Test load 18.5 kN, dry (varying test speed).
Test load 25.5 kN, dry (varying test speed).
Test load 30.5 kN, dry (varying test speed).
For these three graphs it is evident that speed has little influence on the tyre behaviour at low slip ratios up to and just beyond the initial peak at a slip ratio of about 0.2. After this point it can be seen, as before, that the traction force continues to increase with speed up to a slip ratio of 1.0.
In Figures 16 to 18 the traction force is plotted against slip ratio, for wet conditions at the three different speeds. In each figure, the graph is plotted for the three test loads.
Test speed 4 kph, wet (varying test load).
Test speed 8.4 kph, wet (varying test load).
Test speed 16.3 kph, wet (varying test load).
The graphs show that, as with the tests in the dry, the traction force increases with load and that the increase is not linear. In fact, the maximum traction force is heavily compromised for higher loads. The assumption being that when the tyre is heavily loaded, it overhangs the top flange of the railway track and the tyre bulges in the centre of the top flange. Thus, making way for the water being concealed between tyre and top flange leading to a seal effect similar to the one described by. 11 The fact that the effect is even more pronounced for higher speeds supports this theory.
Dividing the peak braking force by the tyre load gives maximum coefficients of friction for high and low load at 4 kph test speed of 0.31 whereas the maximum coefficient of friction at medium load is 0.35.
It can also be seen that at low speeds, the braking force stays approximately constant after reaching the initial peak at a slip ratio of about 0.2.
For the higher speeds, this behaviour changes, and the traction force, and hence the coefficient of friction, starts to decrease steadily until the wheel is fully locked at a slip ratio of 1. This behaviour is different from the behaviour in the dry, where the traction force steadily increases after the initial peak.
The slip ratio at which the maximum brake force occurs seems to be affected by the test speed. The following graphs have been plotted to further investigate the influence of speed on the maximum traction force and the related slip ratio. In Figures 19 to 21 the braking force is plotted against slip ratio, for wet conditions at the three test loads.
Test load 18.5 kN, wet (varying test speed).
Test load 25.5 kN, wet (varying test speed).
Test load 30.5 kN, wet (varying test speed).
In all three graphs, the influence of speed is relevant when the tyre is running in the wet. One feature is that the maximum tractive force decreases with increasing speed, more notably with lower loads than with higher loads. Another, probably less obvious feature is that the slip ratio at which the maximum tractive force occurs is smaller at higher speeds than at lower speeds.
It has already been seen that after reaching the peak value, the tractive force decreases steadily. The decrease is more pronounced at higher speeds and loads.
In a previous section it could be seen that in the dry, the rubber heats up as sliding increases. The temperature increase benefits the adhesive component of friction that relies on molecular bonding between the two surfaces. The adhesive friction might be compromised by the cooling effect of the water and the film of water contaminant between the two surfaces. Both effects benefit from higher speeds. Due to the faster forward travel, the tyre gets in touch with more water per time unit and experiences more cooling effect. The mass inertia of the water lying on the track leads to a thicker film of water between the two surfaces at higher travelling speeds.
Figure 22 to Figure 24 show wet and dry traction curves for the same load and speed respectively.11,25 report that the friction is reduced by 20%–30% from a dry to a corresponding wet road. On dry asphalt, the coefficient of friction typically ranges from 0.7 to 0.9, whereas on wet surfaces, it decreases to approximately 0.4–0.6.
Braking force versus slip ratio (test load 18.5 kN, test speed 4 kph).
Braking force versus slip ratio (test load 25.5 kN, test speed 4 kph).
Braking force versus slip ratio (test load 30.5 kN, test speed 4 kph).
Table 2 shows the friction coefficients of the tested road tyre on the rail at a slip level of 0.2 for various loads under dry and wet conditions. It can be noted that the grip levels in the dry are approximately 35% lower than those expected for a road tyre on the road.
Friction coefficients for rubber tyre on wet and dry railway track.
Even more important is the fact that the reduction in the friction coefficient from dry to wet is even more pronounced for a tyre running on rails than for a tyre running on the road. This being 34%–38% for a tyre on a wet railway track compared to 20%–30% for a tyre on a wet road, according to.11,25
Furthermore, it should be noted that the measured and calculated coefficient of friction peaks at the rated load of the tyre. For higher and lower loads than the rated load, the coefficient of friction is smaller in the wet and in the dry and the reduction between dry and wet is higher as well.
Thermal effects
Heating effects, as described by10,24 seem to be one explanation for the behaviour of the tyre on the rails at higher slip ratios above 0.2. If this were true, it should be possible to provide evidence of a substantial temperature increase at the tyre surface. As shown by, 26 thermal imaging can be used to get a basic understanding of the behaviour.
Figure 25 shows the set-up of the thermal imaging (top left) and then a sequence of test results showing the thermal behaviour of the tyre’s surface (top right to bottom centre). Point one marks the reference temperature of the tyre’s surface; point two was set to a zone where the tyre’s surface was in contact with the rail’s top flange and point three measures the temperature of the rail’s top flange. All temperatures were measured in degrees centigrade; the test rig was travelling at 8.4 kph within an ambient temperature of 26°C.
Thermal imaging setup and results.
The tyre’s reference temperature was between 37°C and 41°C. The surface in contact with the rail gradually heated up to 124°C. The demonstrated temperature increase from 83°C to 87°C seems significant in this respect.
Ref. 26 states that tyres must be brought to their usual operating temperature before testing. The fast-running automotive tyre heats up during a quick manoeuvre like a lane change. The measuring sequence in Figure 25 shows that this is not the case for slow-moving RRV. After a more extended test sequence, the tyre’s contact surface to the rail has only heated up to the temperature of the rail and no more. It can be assumed that the slow movements and the huge tyre and rim combination absorbed the local temperature increase at the tyre’s surface (bottom right).
It can be concluded that there was a significant local heat increase at the surface where the tyre was in contact with the rails’ top flange, supporting the theory that heat effects lead to an increasing friction force at higher slip ratios in the dry.
Conclusions
In order to get a good understanding of the tractive behaviour of a rolling tyre on a rail, load and speed sensitivity have been analysed for three different loads and three different speeds, all of which are applicable to the typical operation of RRVs. All nine combinations of load and speed have been tested using the purpose-built test rig in dry and wet conditions. This led to a total of 18 combinations of tests and has provided a good initial understanding of the influence of load and speed in wet and dry on the achievable traction force.
A unique tyre test rig has been designed and built to perform a series of tests to measure the forces generated between a pneumatic tyre and a rail track. The results provide data that could be parameterized to populate a mathematical tyre/rail model to support the computer simulation and design of RRVs.
Tests were successfully performed to measure braking force as a function of tyre slip ratio for a range of conditions relevant to the performance of RRVs during typical operations when on a rail track. It is assumed that the generation of the tractive force used to drive an RRV has the same relationship as braking force with slip ratio, taking into consideration any tyre coordinate system used in a future tyre model.
The operating parameters varied during the tests included tyre load and vehicle speed. The tests were repeated for both wet and dry conditions at the rail-head. The results indicate a behaviour that includes an elastic and a sliding region as usually observed when testing tyres on surfaces representative of roads. The friction coefficients estimated from the tests are approximately 35% of those obtained on the road in both dry and wet conditions.
Most notable is the fact, that friction drops a lot more from dry to wet for a tyre running on a railway track than for a tyre running on a road, approximately 34%–38% compared to 20%–30% respectively.
The results obtained from this test programme have provided data supporting the development of a new empirical tyre/rail model to permit design simulation and optimization of future RRVs.
Footnotes
Appendix
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
