Validation of CAT

How is the CAT validated?

When the first version of the CAT was produced in the mid-1990s there was no accepted method for validating or demonstrating the accuracy of equipment to measure rail corrugation.  We developed a novel calibration technique that could be used with any corrugation equipment.  This work was published openly in an issue of the Journal of Sound and Vibration that contained proceedings of a seminar on corrugation at TU Berlin in 1997 and became the basis of Annex A (“Procedures to verify reference instruments”) of the European Standard on reprofiling (EN13231-3).

The calibration procedure that we proposed involves manufacture of a reference beam with small surface irregularities in all wavelength ranges.  This is measured in a Coordinate Measuring Machine (CMM) to give a reference measurement.  The beam is then measured with the test instrument (CAT or bi-CAT) and the measurements compared in the various wavelength ranges of interest (10-30mm, 30-100mm, 100-300mm are undertaken as standard for all CATs and bi-CATs; 100-1000mm is requested by some clients and is possible with our current beam).

CAT validation
CAT validation
CAT validation

Correlation of datum and CAT measurements of a reference beam, showing filtered displacements in 10-30mm, 30-100mm and 100-300mm wavelength ranges and also one-third octave spectra.  The wavelength range for the CAT measurement is limited to 6.3mm by the sampling length of 2mm used here.

RMS amplitudes of CAT and datum measurements in the different wavelength ranges are shown below with the limits that we demand for an instrument to be acceptable.

  10-30mm 30-100mm 100-300mm
datum (µm) 5.8 10.5 25.4
CAT (µm) 5.9 10.5 25.8
difference (µm) 0.1 0.0 0.4
acceptable limit (µm) 1 1 2

Our current calibration beam provides a comparison of one-third octave spectra and is therefore some validation of the equipment for measuring acoustic roughness.  Another method of giving some confidence that the equipment gives results that are at least comparable to those obtained with other instruments that have been made for this purpose is to conduct a well-controlled comparison or benchmark test.  Tests of this type were part of a so-called “road test” of the European Standard on roughness measurement, EN15610, which was undertaken in 2007.

road test near rail
road test far rail

Measurements made on two rails at one test site, from an independent “road test” of EN15610 conducted in 2007.  The CAT is instrument H, the original Cambridge University trolley is instrument G.  All other measurements are from instruments based on straight-edges, in some cases the same type of instrument with different measuring teams.  There is as little difference between the CAT and other instruments as there is between measurements made using the same type of straight-edge instrument.