mirror of
https://github.com/MaSzyna-EU07/maszyna.git
synced 2026-11-21 09:44:59 +01:00
feat(track): unevenness waves shaped by the track spectra of FRA and ORE B176 (#426)
* feat(track): unevenness waves shaped by the track spectra of FRA and ORE B176 The waves were 5-42 m sideways and 8-83 m vertically, with the long ones cut down, so a short wavelength carried a large share and the ride and the rails looked busy. They are now 7-90 m sideways and 8-110 m vertically, about the bands D1-D3 of EN 13848, each wave standing for a band of the spectrum S ~ Wc^2 / (W^2 (W^2 + Wc^2)), Wc = 0.8245 rad/m, the shape the FRA and the ORE B176 (German) spectra share: the longer a wave, the higher it is. The "uneven" value keeps its meaning, the peak deviation of a rail on a 10 m chord as PLK Id-14 measures it, so the values in sceneries stay as they are. Since the chord sees the long waves less, the amplitudes are calibrated again, and apart for the vertical and the sideways waves: stray_per_mm becomes stray_up_per_mm 0.00229 and stray_side_per_mm 0.0031, and cross_share 0.35. A track at "uneven 18" measures 18 mm vertically, 17 sideways, 15 of twist, +-10 of gauge, 21 of cross level and J 5.7 against the Id-14 limits for 80 km/h. A rail now goes up to about 5 cm off its design over the long waves, so the bound of the test goes from 40 to 60 mm; the short waves the cab shakes on are smaller, the sideways bend under the axles about 30% less. Every scenery with unevenness looks and rides differently after this. No release has unevenness yet. The grading tool follows the vertical waves; on Baltyk the pantograph keeps within 0.6 cm of the wire up to the product 1150 of the vertical value and the speed (925 before). * test(track): bound the sideways offset of an uneven track off its design The longer waves move the track about 4.5 cm sideways off its design at "uneven 18" (2 cm before), and nothing checked or stated it. The Id-14 test now bounds the peak sideways offset of the middle of the track below 5 cm, next to the vertical one, so a later change of the shares can't double it unnoticed. The doc gives the figure, and names platform edges and level crossing panels among what a rail moves away from. --------- Co-authored-by: maj00r <maj00r@host>
This commit is contained in:
@@ -20,13 +20,16 @@ http://mozilla.org/MPL/2.0/.
|
||||
#include <glm/gtc/constants.hpp>
|
||||
|
||||
// Track unevenness: a displacement field over the plan of the scenery, so everything at one place moves together: joined tracks, the
|
||||
// paths of a switch, sleepers, axles. Waves 5-42 m long, shares after the German track spectrum with the long waves cut down; up and down to 83 m.
|
||||
// The "uneven" value is the peak deviation of a rail on a 10 m chord, as in PLK Id-14 appendix 1; see test_track_uneven.cpp.
|
||||
// paths of a switch, sleepers, axles. Waves 7-90 m long sideways and 8-110 m up and down, about the bands D1-D3 of EN 13848. Their amplitudes
|
||||
// follow the shape of the track spectra of FRA and ORE B176, S ~ Wc^2 / (W^2 (W^2 + Wc^2)) with Wc = 0.8245 rad/m: the longer a wave, the
|
||||
// higher it is. The "uneven" value only sets the scale: the peak deviation of a rail on a 10 m chord, as PLK Id-14 appendix 1 measures it;
|
||||
// see test_track_uneven.cpp.
|
||||
namespace track_uneven {
|
||||
|
||||
inline constexpr float most { 100.f }; // largest value of the attributes, mm
|
||||
inline constexpr double stray_per_mm { 0.00145 }; // field amplitude per mm of "uneven", m
|
||||
inline constexpr double cross_share { 0.53 }; // share of the field amplitude the rails settle apart by
|
||||
inline constexpr double stray_up_per_mm { 0.00229 }; // amplitude of the field up and down per mm of "uneven", m
|
||||
inline constexpr double stray_side_per_mm { 0.0031 }; // amplitude of the field sideways per mm of "uneven", m
|
||||
inline constexpr double cross_share { 0.35 }; // share of the vertical amplitude the rails settle apart by
|
||||
inline constexpr double fade { 20.0 }; // length over which a track goes from the value at its end to its own, m
|
||||
inline constexpr double dip_share { 0.4 }; // joint dip depth per mm of "uneven", mm
|
||||
inline constexpr double dip_reach { 3.0 }; // reach of the dip on each side of a joint, m
|
||||
@@ -182,12 +185,12 @@ wave_set<Count> make_waves( double const Turn, std::array<double, Count> const &
|
||||
return result;
|
||||
}
|
||||
|
||||
inline constexpr std::array<double, 8> wave_lengths { 5.3, 7.1, 9.7, 13.1, 17.9, 23.9, 31.7, 41.9 };
|
||||
inline constexpr std::array<double, 8> wave_shares { 0.0721, 0.0997, 0.1344, 0.1703, 0.1549, 0.1386, 0.1226, 0.1074 };
|
||||
// each wave stands for a band of the spectrum, the lengths in a constant ratio, so its share goes as the square root of S(W) * W
|
||||
inline constexpr std::array<double, 8> wave_lengths { 7.0, 10.1, 14.5, 20.9, 30.1, 43.4, 62.5, 90.0 };
|
||||
inline constexpr std::array<double, 8> wave_shares { 0.0435, 0.0617, 0.082, 0.1045, 0.1294, 0.1579, 0.1909, 0.23 };
|
||||
|
||||
// fewer short ones than sideways: the rails come near the limit in the vertical more often, with the peak still at it.
|
||||
// The long ones hardly count on the 10 m chord: the body rises and falls on them, slowly enough for a pantograph to keep up
|
||||
inline wave_set<8> const up_waves { make_waves<8>( 0.0, { 8.3, 11.9, 16.7, 23.9, 37.1, 47.3, 61.9, 83.3 }, { 0.14, 0.19, 0.22, 0.22, 0.23, 0.23, 0.23, 0.23 } ) };
|
||||
// the same shape. The long ones hardly count on the 10 m chord and the cab hardly feels them: the body rises and falls on them
|
||||
inline wave_set<8> const up_waves { make_waves<8>( 0.0, { 8.0, 11.6, 16.9, 24.6, 35.8, 52.0, 75.6, 110.0 }, { 0.0761, 0.1058, 0.1393, 0.176, 0.2175, 0.2651, 0.3215, 0.3887 } ) };
|
||||
inline wave_set<8> const x_waves { make_waves( 1.1, wave_lengths, wave_shares ) };
|
||||
inline wave_set<8> const z_waves { make_waves( 2.3, wave_lengths, wave_shares ) };
|
||||
// the tilt along each axis of the plan. Long waves only: with the short ones the cross level reaching its limit would break the limit of twist
|
||||
|
||||
Reference in New Issue
Block a user