Track Model Differences 1995 - 1998

Each time a new inclinometry dataset is converted into a new track model, a comparison is made with the previous model. This is done quantitatively on a channel -by- channel basis, with the rms scatter in the difference being recorded. The track model uses only channels 1, 2, 5, 8 (TX, TY, LY, RY), so a value for the model difference is obtained by adding the individual rms scatters in quadrature.

A 4-year history of these differences is shown below, and is compared with the symmetry characteristics :

The symmetry data include vertical lines showing the dates of major engineering events such as central bearing adjustments (green lines) and track welding (blue lines). A correlation exists between sudden increases in the 'difference' quotient and c.b. adjustments as is the case also with symmetry changes. Track welding seems to have no impact upon symmetry, but coincides with noise in the differences (early 1997).

It was my impression while compiling these data that the larger the interval of time between consecutive models, the larger the measured 'difference'. The 1998 data are plotted below to illustrate this effect :

. . . but the tendency is minimal.

Perhaps more interestingly, the behaviour of LY and RY during 1997-8 :

. . implies that their contribution to the track model changes were essentially equal until the CBA of July 1997, the right A-frame 'suffered' most immediately following that CBA, but recovered to become the 'silent partner' thereafter.