1 Switch design optimisation: Optimisation of track gauge and track stiffness Elias Kassa Professor, Phd Department of Civil and Transport Engineering, NTNU Trondheim, Norway E-mail: elias.kassa@ntnu.no Baneseminaret 2015, Multiconsult, 28 January 2015
2 Outline of presentation 1. Background 2. Train-turnout interaction 3. Track gauge optimisation 4. Track stiffness optimisation 5. Conclusions
3 Background Turnouts are composed of a switch panel, a crossing panel, and a closure panel The maintenance cost is high in comparison with plain line The demand for a novel turnout design is very high Facing move Trailing move
4 Common Damage Mechanisms Plastic deformation (lipping) Wear Fracture Rail Head Cracks
5 Common Damage Mechanisms Some remedies to reduce maintenance costs are: 1. reducing turnout population (# turnouts) 2. using more durable and advanced materials 3. adopting a preventative maintenance strategy instead of corrective maintenance 4. optimizing the geometry (layout), support stiffness (structure) and rail profiles
6 Rail profile optimisation [not part of this presentation] The rail cross-section varies in the switch and crossing panels Switch rail 9 10 8 7 6 5 4 3 2 1 0 switch rail stock rail
7 TRAIN-TURNOUT INTERACTION Track models Track receptance
8 Train-turnout interaction models Vehicle model Wheel/rail contact model Track model Three models Vehicle model Track model Contact model M q v v M q t t C q v C q t v t K q v K q t v t F v F t Q v
9 Train-turnout interaction The MBS code SIMPACK is used to model the train-track interaction: freight vehicle model, with two Y25 bogies F v track model with time-varying stiffness and damping values The track model is for the standard turnout type UIC60-760-1:15 (curve radius 760 m and turnout angle 1:15) Simulation of dynamics simulation IV
10 Wheel-rail contact Accurately solving the wheel-rail contact detection at S&C is the first step towards the optimisation exercise Challenges Simpack uses rail profile interpolation based on arc length along the profile curve. Requires very closely measured (sampled) profiles 0 S6300 and6600 0-0.005-0.01-0.005-0.015-0.02-0.01-0.025-0.03-0.015-0.035-0.04-0.02-0.045-0.05 0 0.02 0.04 0.06 0.08 0.1 0 0.02 0.04 0.06 0.08 0.1
11 Track model The track model is based on a simple moving mass-spring-damper system, with few degrees-of-freedom, that is coupled to each wheelset This is a common method in MBS software to account for the track dynamics It is sufficient for studies of vehicle ride dynamics It doesn t accurately predict impact loads in switches and crossings k bv c bv m t, J k by c by c bv k bv 3 degrees-of-freedom m r m r k pz c pz c pz m t, J k pz k by c by k bv c bv c bv k bv 7 degrees-of-freedom
12 Track model This modelling approach is useful in parameter studies and optimisation exercise as several simulation can be run in short time The frequency range of the model can be increased by extending the Applied wheel load track model with few more degrees-of-freedom (q) rail k y12 m r Rail pad m 1 m 1 m r k y12 csleep. y12 k pz k z12 Ballast c pz c z12 sleep. m t, J Ballast m 2, J c pz c z12 sleep. Ballast c y12 k pz k by k z12 k y2 sleep. Ballast c by c y2 k bv k z23 c bv c z23 c bv c z23 k bv k z23 m 3 m 3 k z3 c z3 c z3 k z3 Track receptance 9 degrees-of-freedom
13 TRACK GAUGE OPTIMISATION
14 Track gauge optimisation When a vehicle is running through a switch panel, significant lateral wheelset displacement develop leading to severe flange contact in the curved switch rail sometimes leading to flange contact in the straight switch rail Lateral wheelset displacement, diverging route
15 Track gauge optimisation There is an artificial gauge widening on the side of the switch rail (the wheel tries to follow the stock rail) contact point trajectories contact point jump
16 Track gauge optimisation There is a short transition with a two point contact situation for the through route In the diverging route, the contact with the switch rail starts earlier and there is a long transition for the load transfer The severe flange contact result in an increase in wear of the switch rails and on some occasions rolling contact fatigue problems through route
17 Track gauge optimisation The aim is to relieve the flange contact with the switch rail at the early stage by steering the wheel towards the other rail A continuous gauge variation (dynamic gauge widening) is applied at the switch entry to balance the artificial gauge increase contact point trajectories contact point jump The optimal solutions for the nominal case is validated by running a wider set of simulation cases both in the facing and trailing moves Identify design parameters (maximum gauge widening is the critical parameter) Identify optimal design for facing move for trailing move Evaluate the performance of the optimum design with other load cases
18 Optimisation process The geometry of the gauge variation is represented parametrically by: 1. Length L1 where radius is Rc R C d R Out STOCK RAIL 2. Rout (curvature after the jump) 3. LTotal (total length of gauge increase) The influence of Rout and LTotal is not significant as their effect comes after the contact point jump The variable L1 is directly related to the maximum amplitude of the gauge increase d The optimal design is obtained by varying the values of d and five levels are evaluated: d = {8 mm, 12 mm, 16 mm, 18 mm and 20 mm} L 1 L Jump Contact on diverging rail R C L Total Jump to straight SWITCH RAIL
19 Track gauge optimisation Results Facing move The wheelset displacement is reduced to 1.3-2 mm, when using dynamic gauge widening The 18 mm and 20 mm gauge amplitudes lead to a larger displacement (-6 mm) to the reverse side There is a reduction in wear index with gauge widening Increasing the gauge beyond 18 mm amplitude gives an adverse effect
20 Track gauge optimisation Results Trailing move Gauge widening designs with 12 mm and 16 mm amplitudes are examined in trailing move to identify the optimal design The 12 mm gauge widening amplitude leads to a significant improvement in the wear index The wear index for the 16 mm gauge widening is rather increased both in the 1st and 2nd contacts
21 Evaluation of the optimum design The performance of the optimal design is assessed by several load cases with respect to wear and RCF indices The load cases are based on 18 measured wheel profiles and 100 realizations generated using one Karhunen-Loève expansion Deterministic analyses were performed for each realization to determine the distribution of the response 100 realizations of wheel profile contact point location
22 Evaluation of the optimum design The peak wear index value of 207 N for the nominal case has reduced to 133 N The location of the peak value has shifted forward RCF index exceeded the limit at several locations for the nominal case compared to the gauge optimised geometry
23 TRACK STIFFNESS OPTIMISATION
24 Advanced track design Track stiffness and track inertia varies along the turnout Steel - Concrete Two Layered Track (Corus) [source: INNOTRACK] vertical track stiffness along the turnout offer a more consistent support offer a bridging support
25 Track stiffness optimisation Measured track receptances at three locations in the switch at 4.5 m, 9.1 m and 21.85 m from the front of turnout are used to extract input data to the track model The stiffness in the upper springdamper elements of the moving 7-dof track model represents the combined rail and rail pad stiffness (kp) The remaining flexibility (the structure underneath the rail pad) is represented by the lower spring-damper elements (kb) m r m r k pz c pz c pz m t, J k pz k by c by k bv c bv c bv k bv 7 degrees-of-freedom Track receptance at location 1
26 Track stiffness optimisation There is a 70% change in the value of kp from location 1 to 3 There is an 80% increase in the value of kb The aim is to optimise the vertical track stiffness in the switch panel A simple procedure is followed based on varying the measured track data parameters 26
27 Track stiffness optimisation Two alternative stiffness kp variation are developed This change is expected to be gained by adjusting the rail pad stiffness along the turnout In k p _v1, the value of k p is increased by 30% at location 1 not changed at location 2 reduced by 15% at location 3 The overall increase in the k p value is 11% In k p _v2, the value of k p is increased by 28% at location 1 reduced by 6% at location 2 reduced by 19% at location 3 This limits the stiffness increase to about 8%
28 Track stiffness optimisation The lower spring-damper element stiffness (k b ) is also adjusted The value of k b is increased by 20% at location 1, reduced by 10% at location 2 and reduced by 20% at location 3 The overall change in the value of k b has reduced to 15% This change is expected to be gained by adjusting under sleeper pads and ballast mats along the turnout
29 Track stiffness optimisation Results A slight reduction in wear index is obtained when only k p is adjusted Varying only k b reduces the wear index at the second contact point by almost a half The peak wear index is reduced by 50% at the first contact point and by 80% at the second contact point when using a combined kp and kb optimised values
30 Conclusions Several gauge widening amplitudes have been analysed Larger gauge amplitudes lead to larger displacements to the other side of the track which cause additional lateral excitation The geometry with 12 mm gauge widening amplitude resulted in an improved performance, both in the facing and trailing moves for the through route The main benefits are very significant reduction of wear and RCF indices at all times along the switch panel, and therefore improved behaviour in terms of wear and rolling contact fatigue Also, the optimised geometry showed more consistency in the results when using different wheel profiles
31 Conclusions With gauge optimisation the contact points near the gauge corner move towards the rail head and this relieves the flange contact Reducing the stiffness variation along the switch panel seems to improve the performance of the turnout There has been demonstration tests in Sweden with different rail pad stiffness to validate the technique