Validation

How to validate RCA

Instruments such as RailMeasurement’s CAT can be validated by taking a calibrated beam and comparing measurements made with the test instrument (the CAT) to those made with a “datum” instrument, a Coordinate Measuring Machine (CMM).  If the CMM has been calibrated relative to a national standard, this gives a so-called “traceable” calibration of the test instrument.  With the very much larger, vehicle-mounted RCA, such a direct calibration is very much more difficult.

RailMeasurement validate each RCA by undertaking tests in the field in the conditions in which the equipment will be used.  Ideally there is an extremely low level of irregularities on the rails that are measured as it is a greater challenge to measure small irregularities accurately than to measure significant corrugation.  The greatest challenge for the equipment is to show that track has been reprofiled to a high standard.  Consequently we try to undertake these proving tests, in particular the test of accuracy, on track that has been reprofiled to the extremely demanding requirements of Class 1 of EN13231-3:2006 (which is a very much more demanding standard than the 2012 revision).  Results are shown here from very well reprofiled track and from track with a significant level of irregularities.

Typically our tests include the following:

  • Repeatability: a comparison of measurements made over the same track under the same conditions (no grinding, same speed, same measuring direction)
  • Reproducibility: a comparison of measurements made over the same track under different conditions (grinding and no grinding, different speeds within a typical range of operating speeds, different measuring direction)
  • Accuracy: a comparison of RCA and CAT measurements over the same track. In principle, since the CAT has been calibrated relative to a CMM, this gives a “traceable” calibration of the RCA.

Clearly the last one of these is of critical importance: equipment may be extremely repeatable, but it may nevertheless give the wrong results.

(a) 10-30mm

(c) 100-300mm

(b) 30-100mm

(d) 300-1000mm

Reproducibility of RCA over 500m in a siding, different speed and direction of travel (5km/h, A cab leading, and 10km/h, B cab leading). There is almost perfect reproducibility of amplitudes e.g. in the 100-300mm wavelength range, 6m RMS and 7m RMS in the two runs on the left-hand rail, 13m RMS in both runs on the right-hand rail. Correlation of the percentage exceedence is excellent when the measurement is not close to the limit e.g. in the 100-300mm wavelength range, 1% and 2% on the left-hand rail, 7% for both runs on the right-hand rail.

Correlation of “raw displacements” measured for two runs with RCA over 500m of track.  The “raw displacement” contains all wavelengths within the nominal range of 10-3000mm measured by the equipment.  The longer wavelength irregularities, which are typically of very much greater amplitude, contribute more to the raw displacement than shorter amplitude irregularities.

Excellent correlation of CAT and RCA measurements on both rails over 500m of track on which irregularities are well within the requirements of Class 1 of EN13231-3:2012, despite large irregularities at the frequent welds.  Irregularities on this track are not within Class 1 of EN13231-3:2006, particularly in the 100-300mm and 300-1000mm wavelength ranges because of the effect of the welds.