INTRO: Rail grinding to remove corrugations helps to ensure a high quality of track, but it is important to ensure that the process does not introduce its own irregularities. Advanced measurement techniques will help to refine acceptable quality standards
BYLINE: Stuart L Grassie, J D Smith, and Martin Saxon *
THE PRINCIPAL REASON for grinding rails - at least in Europe - is to remove corrugations. Grinding reduces both rolling noise and dynamic loads, leading to reduced damage to vehicle and track components. On main lines, short wavelength corrugation is generally within the range 30 to 100mm. For metros and slower routes different mechanisms are active, and the wavelength range is broader (30 to 300mm).
In ’acoustic grinding’, where reduction of noise is the main concern, irregularities in the range 10 to 30mm are also of interest. Irregularities with wavelengths of less than 10mm are commonly regarded as small scale surface roughness.
Because dynamic loads and noise both increase with the amplitude of longitudinal irregularity, it is clearly desirable that the rail is as smooth as practically possible after grinding. It is critically important when removing pre-existing corrugation that the grinding process introduces longitudinal irregularities which are as small as possible.
This is not a trivial matter: main line rail grinders almost exclusively use grinding modules which rotate about an axis normal to the rail. For typical rotational speeds of 50 to 60Hz and grinding speeds of 5 to 10 km/h, the distance progressed in one turn of the grinding stone is in the range 23 to 56mm. Inevitably, some longitudinal irregularity (which may be extremely small) is left at this pitch. Horizontal axis machines have similar problems.
Stringent specifications
Increasingly tough specifications have been set by European railways to limit the longitudinal irregularities which can remain after grinding. Limiting amplitudes specified by one railway are 0·01, 0·02, 0·02 and 0·13mm for the 10 to 30mm, 30 to 100mm, 100 to 300mm and 300 to 1000mm wavelength ranges respectively.
Measurements of rail vibration under a grinding train indicate that the rail itself vibrates with peak-to-peak amplitudes of this order in the corresponding frequency ranges, even if the train is simply rolling over the rail at a speed of about 8 km/h. If the train is grinding, vibration amplitudes are much greater. To measure tiny irregularities on a component which is moving with greater amplitudes than have to be measured is a major challenge. The fact that residual irregularities are commonly less than these amplitudes is a tribute to modern rail grinding machines.
The present specifications are open to some ambiguity. A most stringent interpretation is that no irregularity can remain after grinding anywhere on the ground portion of the rail or within the grinding site which exceeds the specified limits.
An easier interpretation is that a moving average of peak-to-peak amplitudes of irregularity measured along a specified line on the rail cannot exceed these limits. Both interpretations may be relaxed to admit isolated exceedences at features such as joints, bad welds, ballast spalls and wheel burns, which could only be ground out at considerable cost.
For the 10 to 300mm wavelengths, no equipment exists on present grinding trains which can reliably measure such tiny amplitudes of irregularity, although the principal contractors can grind finished rails to meet these specifications. Indeed, to the best of our knowledge, no satisfactory and consistently validated equipment is available which can measure rail continuously for several km.
For wavelengths over 300mm it is difficult to undertake sensible datum measurements, because many instruments are limited by the length of the straight edge on which they are based (commonly 1m). Whilst longer measurements are possible by superposing data, this must be at some cost in accuracy.
Profile measurement trolley
One attempt to resolve some of these measuring difficulties is Loram’s Corrugation Analysis Trolley. This is portable, as mounting sufficiently accurate equipment on a train presents even more formidable challenges.
Our first objective was to develop an instrument which could:
