·Home ·Table of Contents ·Terrestrial and Marine Transportation

Comparison Between the Magnetic Properties of Ball Milled Nanocrystalline Perlitic Steel and the Running Surface of Rails in High Speed Railway Tracks N.Takacs, D.L.Beke, L.Harasztosi Dept. of Solid States Physics; University of Debrecen, Hungary; Gy.Posgay, P.Molnar Metalelektro Ltd. Budapest, Hungary; Contact

Summary

It is well known that the magnetic Barkhausen noise (MBN) depends among others on the grain size. This dependence was investigated in detail in many earlier works, but only on microcrystalline materials. However, it is not known whether these relationships are also valid or not in nanocrystalline state (when the grain size is less then 100 nm). This can be very important in the understanding of the formation of surface corrugation, caused by the surface loading and can be related to the formation of a nanocrystalline layer. In our measurements ball milling of rails was carried out in order to model the formation of the nanocrystalline layer and to enable the investigation of the relationship between the grain size and the magnetic properties. On-site measurements on railway tracks were also carried out and will be discussed.

The milling

For our investigations we applied a new UIC-54 type rail (made by Steel Company Diosgyõr, Hungary). From the upper part of the running surface we removed a tenth millimeter thick layer with mechanical disposition. The produced shaving was milled in a ball mill for different periods of time. The mill reduced the grain size of samples by mechanical loading. We hoped that the loading of the mill is similar to the loading of railway wheels on the rail surface.


Fig 1: Grain size changing with milling time

The change of grain size with the milling time is plotted in Figure 1. We measured these grain size values by X-Ray diffraction method. It is visible that the grain size decremented from about 50nm to about 10 nm. It can be observed that when the milling time was between 10 and 20 hours, the produced sample was strongly textured.


Fig 2: TEM pictures. Left to right: Original sample (x12k mag), after 50 hour milling (x100k mag) and after 350 hour milling (100k mag)

We took TEM pictures of the samples. Some of these are in Figure 2, where the fraction of the initial perlitic state can be seen during the milling.
The produced powder state sample was sticked to a cylindrical shape plexy specimen holder. With this method we could avoid the rotation of powder parts caused by the external magnetic field during the MBN measurement.

The MBN measurement device

The block diagram of the measurement setup is shown in Figure 3. The specimen is placed coaxial with the excitation and detector coils. The output of the external signal generator is connected to a power amplifier. The produced sinusoidal signal with 30Hz frequencies and about 20 Watts power is connected to the excitation coil. The Barkhausen signal, which is indicated in the detector coil, is connected to the filters and amplifiers.


Fig 3: Measurement setup

After these units the signal is connected to RMS (Route Mean Square) circuit or to a 'one channel amplitude analyzer' unit. In the first case the RMS unit integrates the Barkhausen signals for some excitation period.
In the second case the instrument counts how many elementary Barkhausen signals are in a voltage window. If we slip this window, we can plot the amplitude distribution of elementary Barkhausen signals. From this distribution function we can conclude the total number of the pinning centrums and the pinning capability of these.

Results

We measured both the RMS value and the amplitude distribution of MBN. The dependence of changes of RMS value on grain size and milling time can be seen in Figure 4 and Figure 5. In double logarithmic plotting a linear function can be fitted to the data, so the equation between the grain size and the MBN rms value is similar to the one below:
where the exponent factor is about 2. It can be well seen that the MBN rms values of textured samples do not fit this curve, because their signals are larger than the fitted value. This is an important experience for the future because in the next measurements it must always be investigated whether the material is textured or not.

Fig 4: Changes of MBN rms value depend on milling time

Fig 5: The change of MBN rms value depends on grain size

Measurements on used rails

We wanted to check that the results of the ball milled material are in accordance with the results obtained on used rails. For the comparison we carried out a series of measurements on railway tracks at the site. For our investigations we chose a new track, which we had been able to observe since it was laid. We measured the MBN rms value on the running surface of rails at different times.
These results can be seen in Figure 6. In this figure there are 3 lines, which accord with the 3 parts of the railway track measured by us.


Fig 6: Changing of MBN rms on the running surface of a rail built into a railway track

The first part was before the curve of the track, the second was in the curve, and the last part was after the curve. In all the three cases the similarity to the results of the milled samples in Figure 4 can be observed. The comparison between the two sets of results can only be qualitative.
It can also be seen that the second line (which was in the curve of the track) saturated to a lower value than the other two. It accords with the lower grain size caused by the higher mechanical loading in the curve.

Conclusion

We found a good correlation between the grain size and the MBN rms value in the results of the ball-milled materials. These results also show that the MBN rms values are very sensitive to the textured structure of material.
We also investigated a used rail, and recorded the variation of the MBN rms value at different times, and we found correlations similar to the ones on the ball-milled materials.
Consequently MBN measurements may be utilized to investigate the formation mechanism of surface corrugation in rails.

© AIPnD , created by NDT.net

|Home|    |Top|