This query refers to the calculation of the required "slewing" to
correct the curvature of curved track. The curvature is normally
measured in terms of "versines" measured over a series of overlapping
chords.
In the textbook "British Railway Track" published by the P.W.I (4th and
5th editions) three methods are given namely averaging, the plus/minus
method, and another one based on moments.
The first of these makes a correction to the curvature at one isolated
point on the curve by comparing the versine to an average calculated
over three adjacent ones.
The second makes a series of corrections at chosen points where the
curvature is not uniform. Each slew affects the neighbouring versines.
The method of moments starts at the end of the curve and works along it.
If you make the wrong assumptions it can lead you down the embankment.
We will not consider that one here.
My query is concerned only with the first two methods, which have very
good numerical stability. Having a good correcting algorithm for a
chosen point (the averaging method), why does the second method not
employ this to make its corrections rather than a sequence of arbitrary
guessed slews ?
Sincerely,
Geoff Bagley.
P Way Dept.
Gloucestershire Warwickshire Railway.
--
Geoff Bagley
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|