PAPERS
Increase in Tangential Force by Ceramic Particles under Low Adhesion Conditions
2025 Volume 66 Issue 1 Pages 37-43
Details
2025 Volume 66 Issue 1 Pages 37-43
Railway wheel spinning and sliding caused by leaves on the line in autumn affect operational safety and regularity. This paper reports on results of braking tests using a vehicle equipped with apparatus for applying ceramic particles to rails to enhance adhesion. To ensure uniform conditions along the rails and minimize differences between test trials, low adhesion conditions were simulated by applying paper tape to the top of the rail. The tangential coefficient at the beginning of the wheel slide increased with the amount of ceramic particles applied. The larger the area of holes made by the particles in the paper tape, the higher the tangential force during sliding. These findings will be useful in designing new adhesion enhancers and evaluating their performance.
This paper is published in Wear Vol 530-531 (2023) 205001 “Penetration of contaminated film on rails by ceramic particles and increase of traction force,” © 2023 Elsevier B.V., doi.org/10.1016/j.wear.2023.205001.
Railway wheel spinning due to low adhesion with rails can cause train delays or cancellations. Sliding while braking can result in overrunning, which is a safety concern. Furthermore, spinning and sliding can cause damage to components, for example rail burns and wheel flats.
The application of sand or other hard particles from vehicles is typically used to mitigate low wheel/rail adhesion. In Japan, sand particles of a few millimetres in diameter have been applied to the rails using a principle similar to that of gravitational force since the era of old steam locomotives. However, during high-speed running of the vehicle, the proportion of sand actually landing on the wheel-rail interface was low because most of it was blown away by wind, or bounced off the rail surface as it fell during high speed running. [1]. In Japan, another process was developed to overcome the problems of using sand. This process employed hard ceramic particles (e.g., alumina) with a diameter of approximately 0.3-0.4 mm spread in small quantities (e.g., 30 g/min) using compressed air [1]. This technology has been widely used in Japan since the 1990s for conventional locomotives and diesel trains, as well as for high-speed trains such as the Shinkansen. This method has also been applied to trains on Taiwan's high-speed rail network [2, 3]. Furthermore, low-pressure injection systems have been developed for vehicles without compressors and are now installed on many conventional rail vehicles in addition to vehicles without compressors [4]. These particle applications are not only used in high-speed railways but also widely employed across Japan's conventional railways network, particularly in sections with steep gradients. Figure 1 shows a photograph of the particle applicator. The hard particles are primarily composed of ceramic particles, such as alumina, with a diameter of approximately 0.3-0.4 mm, which is significantly smaller than natural sand used in older sanding devices where particles have a diameter of several millimeters.
Low adhesion can be caused by several factors. On high-speed railways, which operate mainly on viaduct tracks, water films caused by rain and condensation lead to low adhesion. These films are typically of the order of 1 µm or less [5], and a slight increase in roughness can be considered to contribute to improving adhesion. However, the thick films that form on rails in mountainous sections of the track can be a more serious problem in the autumn. Leaves blown onto lines by passing trains, are crushed by rails and wheels, and stick to the railhead surface. Moisture from morning dew causes repeated reactions between the leaves and rail, resulting in the gradual formation of a tough film. The thickness of this film has been reported to be of the order of several tens of micrometers [6], and particles significantly thicker than water film may be present. These conditions call for suitable adhesion enhancers.
One paper offers a review of the studies which have been carried out on materials applied to the wheel/rail interfaces to recover adhesion and methods used to apply these materials [7]. For example, Lewis et al. investigate the effects of different hose type, mounting angle, and nozzle on the amount of material fed to the interface using full-scale rigs [8]. Ban et al. investigate the effect of hard particle size on wheel/rail friction using a full-scale friction machine [9]. They conduct tests under low adhesion conditions using paper tapes simulating the leaf layer and reported a higher tangential force with larger particles than with smaller particles. Skipper et al. investigate the effects of three different sand types under dry, wet, and leaf extract-contaminated conditions using a high-pressure torsion test [10]. The results of their study show that under low adhesion conditions, adhesion increased with particle size. In addition, when contaminated with leaf extract, larger particles are more likely to cause breakage of the leaf film [11]. Arias-Cuevas et al. investigate the effect of sand particle size on wheel-rail adhesion using a wide range of particles with particle sizes varying from a few hundred micrometers to a few millimeters [12, 13]. Their results of field tests show that the coefficient of friction decreases with increase in particle size. The reason for this is not only that particles are more likely to be trapped at the wheel/rail interface, but also that extremely fine sand can be incorporated into the contaminant and lose [13].
As mentioned above, although the differences between the properties and effects of different adhesion enhancers have been investigated from various perspectives, many aspects remain unknown. In this study, the effects of differences in the amount of adhesion enhancer and particle size were investigated using a simple, uniform film of paper tape simulating a thick contaminant, such as leaves on the line. In the tests, an actual vehicle was driven on a test track, and braking tests were performed while the adhesion enhancer was fed from an onboard injection system. Paper tape was stuck to the surface of the rails to recreate low adhesion and a uniform film. After the test, the condition of the paper tape was observed to determine its effect on the friction properties.
Figure 2 shows a photograph of the test train. Tests were conducted on a test track at the Railway Technical Research Institute (RTRI). The test vehicles were the R291 series (two vehicles) owned by RTRI. The second bogie, one of the four bogies in the traveling direction, was equipped with motors.
Figure 3 shows a schematic diagram of the mounting positions of the test equipment on the test train. An apparatus for applying ceramic particles between the wheel and rail was installed in front of the bogie with motors.
Figure 4 shows the photographs of the test devices mounted on the second bogie. The nozzles for injecting the ceramic particles were mounted at an angle to allow them to inject the ceramic particles into the wheel/rail contact interface (Fig. 4(a)). A photoelectric sensor (Fig. 4(b)) was used to confirm that the vehicle was in the test section. Two water spray nozzles (Fig. 4(c)) were mounted on the rear axle of the bogie to clean the wheel tread outside the test section. Photoelectric sensors were installed to project light vertically and detect the light reflected from the targets (reflectors) placed on the track in the test section.
Figure 5 shows the test site for the test train. Paper tape was stuck to the top of the rail to ensure uniform conditions along the rail and minimize differences between test trials (50 mm wide × 0.14 mm thick). The use of paper tape on rails to recreate low adhesion conditions during tests on actual tracks has been reported by the Rail Safety and Standards Board in the UK [14]. The actual thickness of the leaf layers was reported to be between 5-100 µm [6, 15], and the test conditions in this study were made to be slightly more severe compared with these values. However, the present study focuses on the effect of the quantity and size of the applied particles on adhesion. In order to emphasize these parameters, other parameters have to be as uniform as possible. Therefore, in this study, paper tape was used for simplicity, as has been done by other researchers in the past.
The test train was operated using only the electric brake of the powered bogie. The test train in this study used a special program that increased the braking force linearly after braking was initiated. This made it possible to calculate the train's acceleration and deceleration using the motor current, motor torque, and vehicle dynamic model. When the braking force increases, there comes a point where the wheels start to slide. The program is designed to release the brakes when deceleration falls below a certain threshold for a specific period of time.
Owing to the low conductivity of the paper tape applied to the rail, arcing could occur between the wheel and rail because of interruptions in the return current path when the train ran on the rail covered by the tape. Therefore, in this study, the characteristics of the R291 test train which can be operated in battery mode were utilized to prevent the return current from flowing from the wheel to rails. The contact pressure calculated based on the design geometry of the wheel and rail was approximately 650 MPa.
The tangential force Ft on the outer circumference of the wheel was calculated by the following equation:
| (1) |
where Tm is the motor torque, α is the gear ratio, r is the wheel radius, ηg is the gear efficiency, and ηi is the inverter efficiency. Tm was calculated by multiplying the measured motor current by a factor obtained from the calibration tests. The tangential force coefficient was defined as Ft/Fn, which was value by dividing the tangential force Ft by the static load Fn of the wheelset. Table 1 lists the test conditions used in this study.
| Parameter | Value |
| Initial speed (target) | 30 km/h |
| Injection pressure | 500 kPa |
| Amount of ceramic particles applied (target) | 60, 130, 200 g/min |
| Average diameter of ceramic particles (Median diameter) | 383 μm (regular size); 487 μm (large size) |
The ceramic material used was alumina with two different particle sizes. Figure 6 shows photographs of the ceramic particles.
Figure 7 shows an example of the volume-based size distribution of the ceramic particles. Measurements were performed using a laser diffraction and scattering particle size analyzer (MicrotracBEL, SYNC-ST01). Three trials were conducted for each type of particle and the size distribution of the ceramic particles was obtained. Consequently, the average particle diameters of three trials were 383 µm and 487 µm (median diameter) for the regular and large ceramic particles, respectively. The coefficient of uniformity was 1.64 for regular particles and 1.98 for large particles, indicating that the large particles have a wider distribution of particle sizes compared with the regular particles. Note that the calculation of the coefficient of uniformity was made under the assumption of uniform density, based on the volume-based particle size distribution.
Figure 8 shows a schematic diagram of the test procedure. Prior to the test, several conditioning runs were conducted to remove surface rust. The following steps (i)-(v) were repeated during the test.
(i) The test vehicle begins to travel and accelerate to a target speed (30 km/h).
(ii) Application of ceramic particles and braking are conducted before the motorized bogie enters the low-adhesion zone. At this time, only the electric brakes of the motorized bogie are operated (the braking force increases gradually and linearly).
(iii) Ft/Fn is obtained when sliding is detected in the low adhesion zone.
(iv) The test is considered as completed when the motorized bogie passes through the low adhesion zone.
(v) The paper tape is removed from the rails. The top of the head is then ground and cleaned using a rail grinder. Since a grinder with a dust collection function is used, little dust is generated. However, the presence of dust is checked and cleaned. The wheels are cleaned using a water spray outside the test zone when the vehicle comes back.
At the end of the test, the tape was removed and placed on a transparent polyethylene terephthalate film for storage. The film was placed on an LED panel where light was projected from below and it was photographed from above, to examine holes created in the tape by the application of ceramic particles. Figure 9 shows the locations of the paper tape samples collected for analysis. To minimize the effects of changes of the injection and vehicle running conditions, and the rail geometry on the creation of hole, to the extent possible, during each test the samples were taken from a total of six points on the left and right rails at three locations, namely, near the entrance, center, and exit of the test zone. The length of each sample taken at the six sampling locations was approximately 280 mm, and the total length of samples collected per test was over 1.5 m.
All photographs were trimmed to sizes of (1) 45 × 45 mm, (2) 45 × 90 mm, or (3) 45 × 135 mm in the sleeper and longitudinal directions. The length in the sleeper direction was 45 mm, because the width of the paper tape was 50 mm and was not strictly parallel to the camera when photographed. For the longitudinal direction, (3) was selected when there were no particular problems to minimize the influence of the distribution bias, sizes (2) or (1) were selected for paper tapes that tore when removed from the rail.
Figure 10 shows an example of the changes in each measured parameter during the test. The bottom line represents the values measured using a photoelectric sensor. The data between the first and second peaks from the left were obtained when the motorized bogie traveled in the low adhesion zone which was simulated by paper tape. The second line from the bottom (dotted line) shows the operating status of the particle applicator, with the higher and lower values representing the operating (ON) and non-operating (OFF) states, respectively. The application of ceramic particles and the braking action occurred before the motorized bogie entered the low-adhesion zone. Ft/Fn began to increase approximately 2 s before the motorized bogie entered the low adhesion zone and peaked at the middle of the test zone. A difference was observed in the velocity change between the non-motorized and motorized axles at the peak Ft/Fn. Only the motorized axle slid, since braking force was not imposed on the non-motorized axle. As mentioned in section 2.2, the braking force was programmed to decrease when deceleration fell below a threshold value for a certain period of time. Therefore, the tangential force decreased and the velocity difference reduced, so that Ft/Fn peaked. The peak of Ft/Fn is the tangential coefficient at just before the start of the slide and will be hereafter referred to as the (Ft/Fn)peak.
Figure 11 illustrates the relationship between the amount of ceramic particles applied to the wheel-rail interface per minute and the (Ft/Fn)peak. For both particle sizes, the (Ft/Fn)peak increased with the amount of ceramic particles applied. For a large particle, the (Ft/Fn)peak tends to be higher. On the other hand, for regular particles, the adhesion increased linearly with the number of ceramic particles applied. However, for the large particles, the increase in the amount of ceramic particles was not linear with respect to the amount of ceramic particles applied per minute.
Figure 12 shows examples of the photographed rail surfaces after the test: (a) before removing the paper tape and (b) after removing the paper tape. It is noted that the two locations are not the same. The tiny white particles scattered on the rail shown in Fig.12(a) are assumed to be crushed ceramic particles. This is because ceramic particles are white when crushed. The white particles shown in Fig.12(b) can be seen more clearly, with many particles having penetrated the paper tape. At the time of manufacture, the thickness of the paper tape was 0.14 mm. However, when the thickness was measured after the test, using an electromagnetic film thickness gauge (Elcometer, 456 Coating Thickness Gauge), the thickness was approximately 0.09 mm, indicating that the thickness had decreased due to contact between the wheel and the rail.
Figure 13 shows photographs of the paper tapes obtained after the tests, which were binarized for visual clarity. The paper tapes obtained from the tests in which (a) regular particles were injected and (b) large particles were injected, are shown together for comparison. Numerous holes were observed in both cases. The holes shown in (b) are clearly larger than those shown in (a). It is assumed that this is due to the holes created in the tape by the compression of the particles applied between the wheel and rail.
Figure 14 shows the relationship between the amount of ceramic particles applied per minute and the number of holes made per meter of paper tape. The number of holes per meter of the paper tape was obtained by counting the number of holes within a specified range (e.g., 45 × 135 mm), based on binarized data, and extending it to a length of 1,000 mm. Six paper tapes were obtained near the entrance, as well as at the center and exit of both the left and right rails of the low adhesion zone to determine the number of holes. Average values per meter were obtained and plotted. For the regular particles, the number of holes increased linearly with the amount of ceramic particles applied, whereas for the large particles, the number of holes increased nonlinearly. This difference can be attributed to the presence of smaller particles, which are more likely to be homogeneously dispersed between the wheel and rail.
Figure 15 shows the relationship between the number of holes per meter of the paper tape and the (Ft/Fn)peak. For the regular particles, the (Ft/Fn)peak tended to increase linearly with the number of holes per meter of the paper tape. A similar trend of the (Ft/Fn)peak increasing with the number of holes was observed for the large particles, although not as clearly as that for the regular particles.
In Figs 11 and 14, the (Ft/Fn)peak and number of holes increase with the amount of particles applied respectively, suggesting a possible correlation between them. Figure 11 shows a good linear relationship between the (Ft/Fn)peak and the amount of regular particles. On the other hand, the same figure shows large enough deviation from the linear relationship for the large particles: (Ft/Fn)peak in particular rises when the amount of ceramic is applied at 60 g/min. A similar trend to this was observed in the relationship between the feed rate of the ceramic particles and the number of holes, as shown in Fig. 14. Regular particles are relatively small and numerous and can be distributed uniformly across the wheel-rail interface, resulting in a linear relationship between the number of holes in the tape and the amount applied. In contrast, large particles may be relatively nonuniform, resulting in an irregularly large number of holes in the tape, such as at 60 g/min. Another factor that may contribute to this variation is the difference in the uniformity of the particle size distribution. From the size distribution of particles in Fig. 7, it can be seen that the size distribution in large particles is wider than that in regular particles. Also, the coefficient of uniformity of large particles was larger than regular particles (regular particles: 1.64 and large particles: 1.98). These can result in increased variability in particle distribution within the contact area between wheel and rail.
Figure 15 indicates that the tendency for (Ft/Fn)peak to increase with the number of holes in the tape is steeper for large particles than that for regular particles. The reasons for this difference in effect depending on particle size are discussed below. Figure 13 shows that the holes formed on the paper tape when larger particles are applied are larger than those formed on the paper tape when regular particles are applied. Figure 16 shows the relationship between the approximate hole area and (Ft/Fn)peak. Assuming that the increase in tangential force when ceramic particles are crushed at the wheel rail interface is proportional to the projected area of the particles, the hole area was determined by multiplying the number of holes made by the projected area per particle (the area was calculated using the mean radius on the basis of the assumption of the particles as spherical particles). The method of simply reading the sum of the white areas from the binarized data was not chosen. This is because that the tape could have been torn during penetration, and subsequent enlargement of the hole could have occurred when it was removed from the rail.
We compare the relationship between the number of holes and the (Ft/Fn)peak shown in Fig. 15 with the relationship between the penetrated area per meter and the (Ft/Fn)peak shown in Fig. 16. It was found that the relationship in Fig.16 more than the relationship in Fig.15 is closer to a linear relationship regardless of the size of the particles. This could mean that the ceramic particles that penetrated the paper tape film mainly transmitted the tangential forces between the wheel and rail. The effect of hard particles on increasing the (Ft/Fn)peak could be due to several factors such as roughening of the wheel or rail and cleaning of the surface. However, Fig. 16 also suggests that the penetration of ceramic particles through the low adhesion coating is one of the factors that increase the tangential force between the wheel and rail. Therefore, efficiently breaking through the low adhesion coating is considered an effective approach to increasing the (Ft/Fn)peak. Figure 17 shows the ratio of the number of tape-penetrating ceramic particles to the number of regular and large particles applied respectively. The number of particles applied was determined by dividing the overall weight by the average weight of one particle. The average weight per particle was obtained by dividing the weight of dozens of particles by the number of particles counted actually. The values are the averages of test results, and the error bars indicate the maximum and minimum values. This suggests that the large particles may have penetrated the paper tape used in this test more easily than regular particles. Related to the trend described in the above, Skipper et al. investigate the effect of particle characteristics on adhesion and surface conditions with high pressure torsion testing [16]. They suggest that small particles could lead to a reduction in leaf-cleaning efficiency due to the reduced area acting on leaves.
As shown in Fig. 16, traction performance depends on the area of holes made in the paper tape, and large particles tend to have larger hole areas than regular particles. As shown in Fig. 13, there was a clear difference in the size of the holes formed in the paper tape between the regular particles and large particles. It is possible that the alumina particles, which are incompressible, were crushed, remained on the rail, and penetrated the paper tape due to the pressure created between the wheel and rail. Considering the difference in diameter of approximately 0.4 mm and 0.5 mm, it may seem like a small difference, but simple calculations show that the projected area increases with the square of the diameter (radius) of the particle. It is estimated that the projected area per particle of large particle is approximately 1.5 times to that of a regular particle. Furthermore, the volume increases in proportion to the cube of the sphere's diameter, so that the difference in lateral spreading during compression may become even greater. In addition, the particle size has a certain degree of spread, so the effect of particle size on penetrating the paper tape may be even greater than the average size. However, as mentioned above, it cannot be simply said that the only increase in the particle size increases the tangential force. This is because large particles may lead to uneven action at the interface. In the future, more systematic studies are required to examine the relationship between film thickness and particle size, as well as particle behavior during compression.
The experimental data here demonstrates the importance of applying particles with properties that resist film thickness, which inhibits metal contact between the wheel and rail, in order to increase the tangential force. These findings will be useful not only in designing new adhesion enhancers, but also in evaluating their relative performance. This is because measuring traction requires a significant amount of effort, whereas observing films is relatively easy.
This study investigated the tangential force during braking with the application of an adhesion enhancer. A vehicle equipped with an apparatus for applying ceramic particles as adhesion enhancers was used for the tests. Low-adhesion conditions due to thick films formed by leaves on the line in autumn were simulated to recreate a simple but uniform film using paper tape. The following conclusions were drawn from the test results:
1. The larger the amount of ceramic particles applied, the higher the tangential coefficient at the beginning of the wheel slide. For regular size ceramic particles (mean particle size of 0.4 mm), the tangential force increased linearly with flow rate. For large particles (mean particle size of 0.5 mm), the tangential force tended to deviate from linearity, when a small amount (60 g/min) of large ceramic particles was applied, compared with regular particles, due to the increased tangential force.
2. A linear relationship was observed between the number of holes formed through the paper tape after the test and the tangential coefficient at the beginning of the wheel slide.
3. It was assumed that the product of the number of holes formed in the paper tape and the projected area per particle could be used to estimate the penetrated area of the paper tape. The relationship between the approximate penetrated area and the tangential coefficient at the beginning of the wheel slide was closer to a linear relationship regardless of the size of the particles.
4. In the presence of a thick film on the rail, it could be important to ensure that hard particles, such as adhesion enhancers, penetrate the film sufficiently.
|
Shinya FUKAGAI, Ph.D. Senior Researcher, Frictional Materials Laboratory, Materials Technology Division Research Areas: Friction Management between the Wheel and Rail, Tribology |
|
Takemasa FURUYA, Ph.D. Assistant Senior Researcher, Traction Systems Laboratory, Vehicle Technology Division Research Areas: Traction Control Systems, Power Conversion Systems |
|
Ryo TAKANO
General Manager/Technical Manager, TESS Co., LTD |
