Optimisation of horizontal alignments for railways - Procedures involving evaluation of dynamic vehicle response

Abstract: This doctoral thesis summarises and deepens eightpublications on a project concerning railway alignments.The thesis presents an approach to definingequally costlyalignment alternatives. The approachrelates differences in construction costs for new (or changed)alignments to removal of obstacles (or alternative measures fordealing with obstacles) and is illustrated in clothoid - circle- clothoid combinations.A graphic method,analysis of curvature and slew diagrams, has beendefined to judge when a curve element needs to be divided (i.e.when single curves need to be replaced by compound curves).The thesis provides a new definition of the mostsuitable lengths of transition curveswhen the trackstandards define limits for cant, cant deficiency, rate of cantand rate of cant deficiency. The most suitable transitionlengths are those which give the alignment the highestpermissible speed for a certain set of obstacles.Variables and criteria for evaluation of dynamictrack/vehicle interaction are discussed, with the focus onapplications where the alignments instead of the vehicles arealtered. Methods for quantifying passenger comfort as afunction of vehicle response caused by the alignment areinvestigated. Comparisons are made with ISO (1997) and BritishStandards (1987) for evaluation of human exposure to whole-bodyvibrations, and it is found that passenger comfort on curvetransitionsPCT(CEN 1999) offers a function which isbelieved to be better correlated with passenger(dis)comfort.When track standards are to be re-evaluated, the thesissuggests a division of criteria for evaluating dynamictrack/vehicle interaction into an object function and boundaryconditions. For railways with passenger traffic, it issuggested that passenger comfort on curve transitionsPCTshould be used as object function, whilewheel/rail forces and climbing ratios define boundaryconditions.A simplified analysis, with the vehicle dynamics neglected,as well as a dynamic analysis with comprehensive vehiclemodels, illustrates how the object function may be applied toequally costly alignments and how the optimal transitionlengths depend on the positions of the obstacles and certainvehicle parameters. It is shown how track quality (i.e. thelevel of track irregularities) indirectly affects the optimaltransition lengths, since wheel/rail forces may disqualifyalignments that generate the lowestPCTvalues.The concept of equally costly alignments is used in anapplication where S-shaped superelevation ramps are comparedwith linear ramps. The dynamic vehicle response, evaluatedaccording to the principles in CEN (1999) and UIC (1999), isnot superior on the S-shaped ramps.Keywords: Railway, track geometry, alignment, cant,climbing ratios, clothoids, guiding forces, passenger comfort,PCT, superelevation ramps, simulation, trackshift forces, track/vehicle interaction, transition curves,vehicle dynamics, wheel/rail forces.