·Home ·Table of Contents ·Terrestrial and Marine Transportation

A dynamic and quantitative method for measuring wheel flats and abrasion of trains Qibo Feng, Jianying Cui, Yan Zhao, Ying Pi, Yongping Teng (Dept. Of Physics, Northern Jiaotong University, Beijing 100044,P. R . of China) Contact

Abstract

Wheel flats and other kinds of irregularities or defects on any wheel tread of trains are one of the main causes that threaten safety of trains. There are several methods at present to measure wheel flats. However, almost all these methods are qualitative, and can not get the values of the wheel flats, not to mention to get the values of the wheel abrasion. A totally new method to measure dynamically and quantitatively wheel flats is put forward in this paper, and both the values of wheel flats and abrasion in any tread of trains can be got when a train passes through the measuring device. The measurement system mainly consists of several sets of parallelogram mechanisms, displacement transducers and data collecting and processing system, and it was installed in the railway site. Through a lot of in-situ experiments, the accuracy of about 0.2mm in the measurement of both wheel flat and abrasion can be obtained by the system when the moving speed of a train is less than 15km per hour.

Key Words : Wheel flat, Abrasion, dynamic measurement and train wheels.

Introduction

The true running status of wheels in a wheel/rail system plays an important role, as any defects in treads of wheels will lead to significant vibration and noise. Especially, wheel flats will cause unacceptable additional varying loads on both wheel set bearings and rail tracks, and on some occasions, this great load can lead to rail fractures or even break rails [1~2]. So, how to measure correctly and quantitatively the value of wheel flats is a key problem that must be solved in the development of high-speed and heavy-load trains, and it has not been solved properly up to now. Currently, there are several methods of measuring wheel flats [3~5], such as using the accelerometers or gauges to measure the vertical impulse acceleration or force caused by the wheel flat. However, these methods are qualitative, they can not get the values of wheel flats. Moreover, the impulse force or acceleration impacting on the rail is influenced greatly by many other factors such as the speed, the weight of the train and the type and the quality of the rail etc., which makes it impossible to get exactly the values of wheel flats. How to quantitatively measure both the wheel flat and wear is a common concern among researchers in the field of railway safety technology, and it is the main objective of this paper.

The Measurement Principle

The measurement principle is schematically shown in Fig.1, and the system setup is shown in Fig.2. The whole system mainly consists of four parts. These are the parallelogram mechanisms, the displacement transducer system, the data collecting and processing system, and the computer and communication system. When a train passes through the measurement system, the flange top of the wheel will press down the long beam in the parallelogram mechanism. Any vertical movements will be detected by the non-contact displacement transducer, which faces to the long beam. If there are no defects on the tread of the wheel and no manufacturing errors in the parallelogram mechanism, signals from the displacement transducer are somewhat like a ladder wave shown in Fig.3 A. The output signals will be different ones if there exists a flat spot or an abrasion. In practice, signals are quite different from ideal ones because of various manufacturing errors and assembling errors. Fig.3B and Fig.3 C are typical signals in practical applications got by an X-Y recorder and by the computer respectively.

Fig 1: Schematic illustration of measuring the wheel flat and abrasion
Fig 2: Configuration of measuring wheel flat and abrasion of a train
Fig 3: A. Ideal signals. B. Signals got by an X-Y recorder. C. Signals got by the computer


To avoid two wheels acting on the same long beam, the following conditions must be met. First, at least two sets of parallelogram mechanisms must be installed on the each inner side of the two parallel rails shown in Fig.2. Second, the length of the long beam in the parallelogram mechanism must be less than the minimum distance between axles. Otherwise, even if we can find out the flat wheels, we can not get the information about locations of these flat wheels, which is meaningless.
Some significant advantages of this measurement system are, on one hand, the values of both wheel flat and wear on the tread can be measured quantitatively using the flange top as the measurement base, which is exactly the same situation when the wheel flat is measured in the factory shop. On the other hand, as the whole measurement mechanism is fixed with the rail, the vibration of the rail has no influences on the measurement results, thus greatly improving the measurement accuracy.

Fig 4: Ideal signals when a vehicle passes through two transducers


When a train passes through the measurement mechanism, two successive sensors in each side get a series of signals. Fig.4 typically shows the signals when one vehicle tank passes trough these two successive sensors. The computer can collect all these signals, and calculate the number of both tanks and axles through a simple analysis of the signals. At the same time, types of the tank for both vehicles and engines can be got and the tanks for locomotive can be removed from the measurement results based on analyses of the axial distance d, the distance of the bogie L, and the ratio of L to d.

Experiment results and Conclusions

Table one is a typical measurement report form given by the system when a train passes through the measurement device. In fact, due to adverse circumstances, the signals are greatly different from ideal ones, which produces a little difficulty in the data processing. The main tasks for data processing are to find the flat spot and calculate the values of both wheel flats and abrasion, to give the correct positions where all these defects are located. Finally, the computer will print out a report from to the people on duty, giving the locations and the values of both wheel flats and abrasion that exceed the limits.

Date:02/02/99, Time:05:07, Train Code:2129, Average speed:24Km/h, Total number of vehicle tanks:62, Total number of engine tanks:2, Recorded file:E2020507.dat
Number of car tanks	Number of wheels	Number of wheels	Flat Values(mm)
			By our system	By people
10	2	South	1.0	0.8
14	1	North	0.7	0.6
28	3	North	0.5	0.3
51	2	South	0.7	0.4
53	4	South	0.8	0.6
60	3	North	0.9	0.7
Table 1: Experiment results

This system works very well when the moving speed of a train is less than 15Km per hour. Theoretically, the lower the speed of a train, the higher the measurement accuracy. By comparing the results given by our system and the measurement results given by the technician using a traditional special micrometer, it is found that the measurement errors for both wheel flats and abrasion is 0.2mm. However, there is a great impulse force acting on the long beam at the moment the flange top collides with the beam if the speed is high, which causes a vibration in the parallelogram and makes the flange top not contact with the long beam. In such cases, the measurement base is lost, and the wheel flat can not be found out. To avoid this, some vibration absorber and damping system must be added to the parallelogram mechanism, and this is under development.

References

1. Kigawa, Takehiko, Kimoto. Influences of skidding pattern upon the occurrences of skid damage to railway wheels. Wear, 1993,167(2): 143-154.
2. S.G.Newton ,R.A. Clark. An investigation into the dynamic errors on the track or wheel flats on railway vehicles. Journal mechanical engineering science, 1979, 21(4)287-297.
3. Product news, Flat wheel detectors pass rails. Rly, Gaz, Inter,1988,143(12):835.
4. Investigation of wheel tread irregularities. Rly. Age. 1989, 190(12):51~52
5. Ge linfu, Yin Zhiben, Zhu Huaifang. Microcomputer intelligence detection of locomotive wheel tread irregularities. Journal of the China Railway Society. 1993, 15(2):35-39[In Chinese].

Qibo Feng received his both BS and MS in ME from Hefei University of Technology in 1983, 1986 respectively, his PhD in 1993 in optical instrument from Tsinghua University. Since 1993, He has hold a research and teaching position in Northern Jiaotong University and became an associate professor in 1995. During the period from 1998 to 1999, he joined in the University of North Carolina at Charlotte as a visiting professor. His main research interests include metrology, laser measurement and application, and testing technologies for the safety of the railway. He has published over thirty papers and received five Chinese patents.
Jianying Cui received his both BS and MS in ME from Tsinghua University in 1987,1989 respectively. Since 1989, he has held a position in Northern Jiaotong University. His main research field is NDT and mechanism design.
Yan Zhao graduated from Northern Jiaotong University in 1988,and became an engineer in 1994. Her main interests are experiment research and NDT
Ying Pi received her BS in physics from Hebei Teaching University in 1998. At the same year, she became a graduate student at Northern Jiaotong University. Her major is optics and applied physics.
Yongping Teng received his both BS and MS in physics from Beijing Normal University in 1986,1989 respectively. In 1989, he joined in Northern Jiotong University, and became an associate professor of physics in 1998. His main interests are NDT and data processing.

© AIPnD , created by NDT.net

|Home|    |Top|