QUIET-TRACK: Track optimisation and monitoring for further noise reduction dr.ir. Geert Desanghere Akron, Belgium geert.desanghere@akron.be www.akron.be
Quiet-Track: EU-project: Consortium QUIET-TRACK: Track optimisation and monitoring for further noise reduction 2
Work packages 1 Monitoring of rail roughness, track dynamic properties and average wheel roughness 1.1 Monitoring of rail roughness 1.2 Monitoring of track decay rate 1.3 Investigation of Track Decay Rates (TDR) of embedded rails 1.4 Average wheel roughness determination 2 Rolling contact model enhancement in the existing rolling noise models 2.1 Rolling contact model enhancement 2.2 Model for low frequency noise emission 3 Concepts and tools for noise related track maintenance 3.1 Concepts for acoustic rail grinding 3.2 Concepts for rail profile correction 3.3 Development of a noise related track maintenance tool QUIET-TRACK: Track optimisation and monitoring for further noise reduction 3
Work packages 4 Development and validation of high performance solutions for reduction of track related noise 4.1 Combination of existing track solutions 4.2 Innovative solutions based upon reduced rail roughness growth rate 4.3 Acoustical embedded rail 4.4 Rail type and hardness selection for optimal acoustic performance and wear 5 Development of noise management tools 5.1 Procedure for performance measurement of mitigation measures 5.2 Noise management tool for track maintenance activities 5.3 Noise management tools for noise mitigation solutions at the track level QUIET-TRACK: Track optimisation and monitoring for further noise reduction 4
Modelling Enhancement: New multi-point W/R contact model W/R NOISE program: is a further development of the original wheel-rail noise emission models developed by Paul Remington further developed than TWINS includes the possibility to import measured track and wheel impedances (horizontal, vertical and cross) or calculated impedances by a very precise dynamic finite element model. source code of W/R Noise in Matlab, Enhancement: introduction of realistic multi-point wheel-rail contact conditions in curves (and in some worn tangent track sections). existing models use a Hertzian single point contact with a roughness wavelength filter which is only related to vehicle speed and not to the real wheel-rail contact conditions which are influenced by rail wear, curving, presence of defects. QUIET-TRACK: Track optimisation and monitoring for further noise reduction 5
Modelling Enhancement: New multi-point W/R contact model Validation: a complete numerical noise analysis with the enhanced W/R Noise model or two selected reference track systems. a tangent track with discretely supported rails on sleepers in a ballasted bed a similar curved track, considering the same vehicle at the same speed and the same rail type. input parameters for the noise calculations will be measured: TDR s, rail roughness and average wheel roughness. track and wheel impedances will be measured and computed. rail profiles (influencing the wheel-rail contact conditions) will be measured. The numerical results obtained with the different sets of input parameters will be compared with measured pass-by noise measurement results. QUIET-TRACK: Track optimisation and monitoring for further noise reduction 6
Modelling enhancement: Model for low frequency noise emission Procedure for calculating the low frequency noise emission below 250 Hz and propagation based on the use of deterministic models (acoustic finite elements and boundary elements). will be integrated as a specific module within the enhanced W/R Noise software. Validation: A complete numerical noise analysis below 250 Hz with acoustic boundary element analysis and above 250 Hz with the existing W/R Noise software will be carried for two selected reference track systems one embedded track one track with discretely supported rails on sleepers in a ballasted bed, same approach compared with measured pass-by noise measurement results. QUIET-TRACK: Track optimisation and monitoring for further noise reduction 7
Existing Wheel-Rail software Normal rolling noise on straight track Absence of wheel and rail imperfections or discontinuities leading to impact noise (wheel flats, rail joints, ) No curving noise (squeal) Primary generating mechanism Wheel and rail roughness Broadband frequency spectrum 250 to 5000 Hz QUIET-TRACK: Track optimisation and monitoring for further noise reduction 8
Structure of the W/R program Analytical model Developed by Remington Modified to include lateral rail radiation QUIET-TRACK: Track optimisation and monitoring for further noise reduction 9
Wheel / Rail interaction Roughness: roughness on wheel and rail running surfaces: w and r Contact stiffness: to account for local deformations of wheel and rail under vertical load (FV): KCR and KCW Y Y Displacements at the point of contact WR WA Y RV RH w F / K r F K Y / V CW Derivation Velocities at the point of contact. Y WR., Y WA., Y RV., Y RH V CR Definition of wheel & rail impedance QUIET-TRACK: Track optimisation and monitoring for further noise reduction 10
Wheel / Rail interaction Definition of wheel and rail impedance [N/ m/s] Wheel Radial impedance: ZWR Axial impedance: ZWA Cross-impedance: ZWVH Rail Vertical impedance: ZRV Horizontal impedance: ZRH Cross-impedance: ZRVH Y Y Y Y WR WA RV RH F F F V V ( F V / Z / Z V / Z WR WVH / Z RVH RV F ) F H H / Z / Z WA RH Y RH QUIET-TRACK: Track optimisation and monitoring for further noise reduction 11
Example: description of a test site Vehicle characteristics Light rail car 6 axles 60 tons Speed : 50 km/h Track characteristics Girder rail NP 4 am Rail directly fastened every 60 cm to a concrete slab A pavement allows the rail to radiate only in the vertical direction Measurement Distance: 7.5 m from track centerline QUIET-TRACK: Track optimisation and monitoring for further noise reduction 12
Wheel / Rail interaction at test side Roughness of wheel and rail running surfaces Typical roughness spectra Trued wheel Welded rail Total contact stiffness Wheel and rail material Elasticity modulus Poisson ratio Wheel and rail geometry Radius of curvature Vertical load forcing the wheel against the rail Roughness spectra (50km/h) wheel rail QUIET-TRACK: Track optimisation and monitoring for further noise reduction 13
Wheel / Rail interaction at test site Contact area filter Degree of transverse correlation Hypothesis: wheel and rail roughness are well correlated for two parallel paths in the direction of rolling High degree of transverse correlation Contact area radius Hertz theory QUIET-TRACK: Track optimisation and monitoring for further noise reduction 14
Rail behaviour (example) Rail impedance (vertical, horizontal, cross) Measurement of the vertical impedance on the test track Development of a finite element model of the test track Tuning of the model by comparing measured and calculated vertical impedance Calculation of the horizontal impedance and cross-impedance Load sensor Deformation 1700Hz Vibration sensor QUIET-TRACK: Track optimisation and monitoring for further noise reduction 15
Rail behaviour (example) Rail impedance - Results Comparison between measured and calculated vertical impedance Calculated horizontal and crossimpedance Admittance = 1 / Impedance horizontal vertical cross calculated vertical - measured QUIET-TRACK: Track optimisation and monitoring for further noise reduction 16
Track decay rate (test site) Vibration sensors 1m Measurements on the track Vertical and lateral excitation on the rail with an impact hammer Measurement of the rail vibration at different points in vertical and lateral direction => Vertical and lateral rail vibration spectra at various distances from the point of excitation Impact hammer 1m Point of excitation QUIET-TRACK: Track optimisation and monitoring for further noise reduction 17
Track decay rate (example) Example: Decay of vertical rail vibration for each frequency band -2dB/m -4dB/m QUIET-TRACK: Track optimisation and monitoring for further noise reduction 18
Wheel behaviour (example) Wheel impedance (radial, axial) Measurements on a wheel of the test vehicle Radial or axial excitation on the wheel with an impact hammer Measurement of the wheel vibration in radial or axial direction => Measured radial and axial admittance axial radial QUIET-TRACK: Track optimisation and monitoring for further noise reduction 19
Sound radiation and propagation Sound radiation Average vibration response Radiating areas Radiation efficiencies Wheel: radial or axial Rail: vertical Sound propagation Ground reflections Sound level cancellation at wavelengths equal to twice the path length difference At higher frequencies (> 250Hz) Direct sound and reflected sound tend to add as incoherent sources and thus increase the sound level A ground effect term is introduced in the propagation equations wheel rail QUIET-TRACK: Track optimisation and monitoring for further noise reduction 20
Sensitivity Analysis Metro vehicle A reference metro vehicle has also been selected and its characteristics have been used to define the reference values of the input parameters. This reference vehicle is composed of six cars of 18,2 m each and is partly illustrated in figure below. Parameter Reference value Alternatives Length of the vehicle [m] 109 - Number of axles 24 - Axle load [T] 12 10 11 13 14 Vehicle speed [km/h] 80 60 100 Type of wheel solid wheel - QUIET-TRACK: Track optimisation and monitoring for further noise reduction 21
Wheel roughness QUIET-TRACK: Track optimisation and monitoring for further noise reduction 22
Data related to the track Parameter Reference value Alternatives Sleeper Elasticity modulus [MPa] 18 000 - Poisson s ratio 0,3 - Density [kg/m³] 1 000 - Loss factor [%] 8 - Dimensions [mm] 2 600 x 240 x 140 - Distance between sleepers [m] 0,60 - Ballast Loss factor [%] 60 - Dynamic stiffness [kn/mm/sleeper] 25 - Rail Type RATP 52 EB50T, Np4am, 35G Loss factor [%] 2 - Rail pad Loss factor [%] 25 10 50 Dynamic stiffness [kn/mm] 150 - QUIET-TRACK: Track optimisation and monitoring for further noise reduction 23
Finite Element Model QUIET-TRACK: Track optimisation and monitoring for further noise reduction 24
Results Sleeper Rail Wheel TOTAL Vehicle speed 60km/h 68,25 73,23 73,90 77,18 80km/h 71,92 77,84 77,86 81,38 100km/h 74,48 81,10 81,01 84,52 Axle load 10T/axle 72,15 78,23 78,29 81,77 11T/axle 72,03 78,03 78,06 81,57 12T/axle 71,92 77,84 77,86 81,38 13T/axle 71,80 77,65 77,67 81,20 14T/axle 71,69 77,48 77,49 81,03 Rail pad loss factor 10% 74,36 80,62 77,77 83,07 25% 71,92 77,84 77,86 81,38 50% 70,21 75,96 78,00 80,53 Wheel and rail roughness Rail A Wheel B -5dB 65,15 71,30 71,23 74,78 Rail B Wheel B -5dB 66,92 72,84 72,86 76,38 Rail A Wheel B 69,41 75,68 75,55 79,12 Rail C Wheel B -5dB 70,04 75,72 75,85 79,34 Rail B Wheel B 70,15 76,30 76,23 79,78 Rail C Wheel B 71,92 77,84 77,86 81,38 Rail D Wheel B 75,04 80,72 80,85 84,34 Rail C Wheel B +5dB 75,15 81,30 81,23 84,78 Rail D Wheel B +5dB 76,92 82,84 82,86 86,38 Rail E Wheel B 79,23 84,75 84,95 88,42 Rail E Wheel B +5dB 80,04 85,72 85,85 89,34 Rail F Wheel B 83,94 89,40 89,62 93,09 Rail F Wheel B +5dB 84,23 89,75 89,95 93,42 Rail type EB50T 71,89 77,77 77,84 81,34 RATP52 71,92 77,84 77,86 81,38 QUIET-TRACK: Track optimisation and monitoring for further noise reduction 25
Results Parameter Parameter giving the lowest noise level Parameter giving the highest noise level Variation of rolling noise in db Vehicle speed 60 km/h 100 km/h 7,3 Axle load 10 T 14 T 0,7 Rail pad loss factor 50% 10% 2,5 Wheel or rail roughness Smooth Rough 15 Type of rail EB50T RATP 52 0,04 QUIET-TRACK: Track optimisation and monitoring for further noise reduction 26
Results Parameter Parameter giving the lowest noise level Parameter giving the highest noise level Variation of rolling noise in db Vehicle speed 80 km/h 160 km/h 9,4 Axle load 12 500kg 5 000kg 1,1 Track type Bi-block Holz 3,1 Rail pad loss factor 50% 10% 2,6 Wheel roughness Smooth Rough 8,5 Rail roughness Smooth Rough 3,9 Type of rail UIC 54 E UIC 60 0,7 QUIET-TRACK: Track optimisation and monitoring for further noise reduction 27