Go to the complete Version in German
This article asserts that the applied ultrasonic testing method does not effectively recognize all possible defects of the wheel. For that reason a task optimized the technique which showed first good results and plans for the necessary modifications are underway for a long term investigation at the service station in Hamburg.
Our question mark: Are those investigations completed and modifications applied??
Follow up this question in the currrent Forum Discussion

Abstract

Wheel surfaces of high speed trains have to withstand great stress between wheel and track. Of course, that causes typical wear and tear damage. Let's take a look at how this happens. With NDT methods it is possible to detect such damage. The results are acceptance criteria for those wheels. Furthermore, the complete refurbishing strategy can be optimized.

The eddy current method and the ultrasonic method were used for testing 34 wheel pairs at intervals of 240.000, 360.000 and 480.000 km operation time. At the end of the investigation the wheels were tested using destructive methods.

Table of contents
1. Visual testing
2. Surface testing with eddy current testing
3. Ultrasonic testing with surface waves
4. Operation of the method
5. Conclusion
Rolf Diederichs' questions prompted Further Discussion

Visual testing

After 60 Tkm, all surfaces already showed more or less significant surface damage. Typically, two damage zones develop. (Fig 1). The zone of greater damage lies between 30-60 mm and zone of less damage lies between 80-90 mm from the outer edge of the wheel. At the early stage of damage development, the damage is visible only with the help of glasses or chemical surface treatment. The small cracks in zone 1 are mostly transversally oriented. In zone 2 no preferences of orientation are visible. After 360 Tkm the damaged zones are doubled in width and are clear visible to the naked eye. After 480 Tkm, some of the wheels were already damaged enough to meet rejection criteria. The rest did not develop higher damage levels. The number and size of the cracking increased clearly.

Fig1. Position of the damaged zones on the wheel surface

Surface testing with eddy current

Multi channel eddy current test equipment was used and 3 sensors were placed at 10 mm, 47 mm and 85 mm distance from the wheel edge (Fig.2). After 240 and 360 Tkm three sensors showed a higher noise level, which is a sign of damage. After 360 and 480 Tkm high noise together with single indications were measured.

Fig2. Position of the ECT probes

Ultrasonic testing with surface waves

The ultrasonic testing was performed with EMATs which generated Rayleigh surface waves (Fig.3). The waves propagate in two directions of the probe in a circular pattern on the wheel surface. Multiple echoes based on each wheel's circle are visible on the screen. Scattering and absorption cause a decrease in amplitude. Transversal cracks are visible between the first and second circulated echo. (A pre-condition is that no other damage exists.) On a wheel without any damage a test crack of 3 mm depth and 30 mm length is detectable. The same test crack is not indicated any more if the sound has to path a damaged section, thus the sound is completely absorbed and cannot continue the circulation. This effect with its decline in echo amplitude of the circulated echo can be used to characterize the degree of damage in the damage suspected zone 1.


Fig 3. Principle of EMAT generated ultrasound

At all wheel pairs the echo amplitude was recorded during the operation time intervals. Measures indicated that until the 360 Tkm interval the echo amplitude declined and from that point no change was measured. Fig 4 shows the relation between UT and ECT.

Fig 4. Relation between UT and ECT measurement

Operation of the method

The method was integrated into test intervals for trains and performed every 3 to 5 days. 410 kHz EMUS probes with no coupling medium generated waves which propagated 7 mm depth. The probes were mounted at the track so that an automated test could be possible by driving the train at a speed of 1,5 to 8 km/h over the probes (Fig. 5).

Fig 5. Installation of the UT probes

The experience of two units showed that ultrasonic testing with this configuration could not be used. After a relatively short operation time the wheel material characteristics changed by hardening in such a way that sound was already absorbed and no evaluation could be performed for future inspection. For that reason a task together with the FhG IzfP optimized the technique. The optimization recognized:

• Probe position so that it works in an area which is not damaged by normal wear
• New probes with one transmitter coil and two receiver coils
• Higher inductive current
• Change of probe construction (metallic coil body)

Conclusion

The first test showed good results and plans for the necessary modifications are underway for a long term investigation at the service station in Hamburg. The evaluation of the results along with characterization of changes in properties is opening the door to interaction with the materials and future design of wheel-train systems.

Further Discussion

After Mr. Hintze finished his presentation Rolf Diederichs asked some more questions which we summarize here.
Can that method using Rayleigh surface waves find all suspected cracks? The answer was very simple - - Yes!
Diederichs mentioned that in some refurbishing shops conventional UT pulse echo techniques [1] are successfully installed and are still offered by manufactures like Inspection Research Technology IRT or Krautkraemer. Examples include the Deutsche Bahn AG shop in Neumuenster (IRT) and the Austrian railroad.
Does this mean that this more labor-intensive application (disassembling ect.), could be replaced?
Mr. Hintze concluded that there are some types of cracks that will still need the old established method.
He also mentioned the new online high-speed testing by a special train. In Germany it is called SPE Schienenprüfexpress which moves at 100 km/h and performs the ultrasonic test. The advanced UT electronics were delivered by IRT. The theme is topical and very new, that means more information will be available later on.

Related information:
1. Railway testing in the Ultrasonic Testing Encyclopaedia
2. H.-J. Salzburger, Saarbrücken; H. Hintze, Kirchmöser Tiefenbestimmung von Laufflächenfehlern an Eisenbahnrädern unter Nutzung linear polarisierter Transversalwellen - Verfahrens-Optimierung und -Erprobung. DGZfP Jahrestagung 1996.

Author

Dipl. Ing Hartmunt Hintze
Since 1993 Head of Department Nondestructive Testing at
Deutsche Bahn AG,
Kirchmöser, Brandenburg

The paper was presented at the German-French Workshop in Aachen in April '97. Published in Berichtsband 57 ISBN 3-931381-13-7: Vorträge 2. Deutsch-Französisches Seminar Automatisierung und Modellierung für die Zerstörungsfreie Prüfung. Page 105 - 112. DGZfP Deutsche Gesellschaft für Zerstörungsfeie Prüfung e.V. Berlin 1997.