Harmonic Vehicle/Track Interaction: Using Simulation Tools to - PowerPoint PPT Presentation

Harmonic Vehicle/Track Interaction: Using Simulation Tools to Increase Train Speeds and Safety Assurance Raphael Marotta Lucas Valente 1

MRS Logistic in Brazil MRS Logística is a concessionary that controls, operates and monitors the Southeastern Federal Railroad Network. The company has been in operation in cargo railway transportation since 1996 Across the rails, connect the three main economic centers of Brazil: Rio de Janeiro, Minas Gerais and São Paulo. MRS’ Shareholders 2

Why a presentation on harmonic excitations?  Some derailments are related with harmonic excitations. Methodology developed to analyze the  vehicle dynamic behaviour taking in Operation’s need to increase speed in many sections of MRS track safely. account the MRS’s features of vehicles, track and operation.  Case: Derailment with a GHS (Gondola‐Hopper) wagon with worn truck conditions. 3

What is a resonance phenomenon ?  A vibratory system will have energy dissipated when oscillating only under an initial disturbance.  An external force must be applied to maintain an oscillation motion, such as: harmonic, nonharmonic, transient or random.  Resonance is a harmonic excitation, when the frequency of excitation coincides with the natural frequency of the system. In a railway track irregularities are harmonic motion of the base 4

What is a resonance phenomenon ? 1  f Natural frequency  f 1 fr 2 Excitation frequency Resonance Single degree of freedom 5

Carbody modes The lower sway is a combination between rolling and lateral oscilation. 6

Carbody modes 7

Lower sway  The lower sway motion is related to the cross level irregularity.  The harmonic instability occurs when the excitation frequency is near to the carbody natural frequency of sway mode. 8

Vehicle modal analysis  Eingenvalue method • Although the freight car is suspension is non‐linear, this analysis allows to examine and understand about the carbody mode. The result is the free response of the system. Frequency Mode f n (Hz) ζ f d (Hz) f LS f LS 1 Lower sway LS f Y f Y 2 Yaw Y f n : Natural frequency f B f B 3 Bounce B ζ: Damping factor f P f P 4 Pitch f d : Natural frequency P f US f US 5 Upper sway US 9

Harmonic combination Connecting speed, frequency and wavelength, thus: Wavelength vs Speed Speed (km/h) Wavelength (m) Lower Sway Yaw Bounce Pitch Upper Sway 10

Lower sway frequency  Transient method • An external force is applied in the carbody center of gravity. • 2 simulations were performed, one with friction wedges(damping natural frequency) and the other without(natural frequency). Vehicle oscillation Natural frequency 11

Lower sway frequency 12

Geometry data processing  After modal analysis it is necessary  identify the λ from geometry data provided by the TrackSTAR  It is usual to apply Signal Processing Tools for this purpose Strip Chart (Cross level) v= λ f v= λ f Geometry Data Wagon Speed Defect amp 4 Track Wavelenght 2 0 v f  Excitation ‐2  frequency Km ‐4 13

Geometry data processing 2 1  The geometry data is defined as a sum of sines and cosines 0  λ is determinated using the Fast Fourier Transform method ‐2 Sin (0.02πx) 2 PSD 2 Geometry data sample 0 3 6 ‐2 1.5 Sin (0.04πx) 2 4 1 5 2 0 3 0 0 ‐1 ‐2 ‐2 ‐5 ‐4 2.5 Sin( 0.08πx +0.5) ‐3 ‐6 An alternative 2 4 f(x) = sin (0.02πx) + 1.5 sin (0.04 πx) way to represent 0 + 2.5sin (0.08 πx + 0.5) + sin (0.16 πx) track geometry ‐2 Sin (0.16πX) 14

Geometry data processing Expected behavior of a PSD P O W E R Frequency 15

Geometry data processing  With the modal analisys completed, the next step is to Modal Analisys determine the frequency bands of interest.  Upper and Low frequency limits must be defined. Track Inspetion What do we need to identify if the resonance Input data phenomenon occurs?  Determinate power within the frequency band. Geometry data  A higher power level correspond to most likely processing resonance.  Diagnosticate risk areas ( mile post) and confirm the diagnosis through simulation. Simulation Simulation 16

Geometry data processing FREQUENCY UPPER LIMIT POWER Frequency band LOW LIMIT Km Post POWER Power within the frequency band 17

Simulation – Vehicle Modeling Main vehicle modeling elements in VAMPIRE Mass Wheelset Stiffness elements Spring Friction elements Non linear stiffness

Simulation – Track Modeling Main track modeling elements in VAMPIRE Geometry – Track Geometry Car Rail and Rail and Fasntenings Fasntenings Tie Tie Vertical Stiffness Suballast Suballast Shoulder Shoulder  Track Modulus – Field Instrumentation, Ballast Track Inspection Vehicle. Lateral Stiffness Foundation Foundation  Rail to tie – TrackSTAR  Tie to ballast – STPT, Tamping Machines.

Case – GHS derailment 20

GHS – Vehicle Modeling  Vehicle GHS 100 tons with Ride Control 6” x 11” truck.  High conicity wheel profile.  Center of gravity height: 2.21 m.  Block side bearing.  Constant damping with worn wedges.  62 degree‐of‐freedom. Note: Friction wedges in worn conditions due to failure in the manufacturing process 21

Vehicle modal analysis  Eingenvalue method results: Frequency Mode f n (Hz) ζ f d (Hz) 1 Lower sway 0.94 0.04 0.94 2 Yaw 2.42 0.14 2.40 3 Bounce 2.46 0.07 2.45 4 Pitch 3.28 0.09 3.27 5 Upper sway 4.44 0.18 4.37 22

Lower sway frequency  Transient method results: Damping natural frequency Natural frequency 0.88 Hz 0.83 Hz 23

Harmonic combination Wavelength vs Speed 110 105 100 95 90 Lower Sway Speed (km/h) 85 80 Yaw 75 Bounce 70 65 Pitch 60 55 Upper Sway 50 45 40 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 Wavelength (m) 24

Harmonic combination  CWR length at MRS – 240 m Dominant Wavelength Parameter If the rail or welds Mode f n (Hz)  Bar length – 24 m Component (m) Affected joints were applied Lower sway 0.94 Yaw 2.42  The wavelength is more or less the same in a alternating 50 way the λ will be CWR length due the low stiffiness region 20 12 m Ballast Vehicle (Comfort) 50 km/h Low 10 2.42 stiffness 5 λ = 24 m λ yaw = 5.72 m 2 Dynamic forces 1 Rail (Track 0.5 deterioration) 0.05 Low stiffness 25

Track geometry at derailment site MAXIMUM DEFECTS Visually the cross level AMPLITUDES wavelengths seems to be Surface – 10 mm Alignment 12 mm about 20 m. Cross level – 17 mm Gauge – 20 mm It’s necessary to investigate more deeply No geometry exceptions 26

Track Geometry Analysis From the modal analysis Damping natural frequency ‐ 0.83 Hz Natural frequency ‐ 0.88 Hz FREQUENCY UPPER LIMIT = 0.05 Hz POWER Frequency band LOW LIMIT = 0.06 Hz Frequency band lower sway 0.83Hz to 0.88Hz Likely sites to find Likely site for Wagon Speed harmonic a harmonic Km Post λ = 15m ‐ 16 m 50 km/h behavior behavior   f = 1/ λ . v f f = 0.05Hz – 0.06Hz POWER Natural frequency 0.83Hz to 0.88Hz 27

Simulation results  Wheel unloading results: Vampire Pro TRANSIENT ANALYSIS Vampire Pro TRANSIENT ANALYSIS 13 abr 2016 13 abr 2016 16:47:28 16:48:25 percent percent 40 km/h 50 km/h 100 100 50 50 m m 0 0 100 200 100 200 -50 -50 -100 -100 Axle 1 left wheel Unloading Axle 1 left wheel Unloading Axle 1 right wheel Unloading Axle 1 right wheel Unloading Axle 3 left wheel Unloading Axle 3 left wheel Unloading Axle 3 right wheel Unloading Axle 3 right wheel Unloading VAMPIRE Plot VAMPIRE Plot 28

Results reached through simulation Speed Paraopeba Location Speed Gain Car type Gains Ferrovia do Aço Ferrovia do Aço 50 64 Gondola 2015 Fuel tank Paraopeba 30 40 car Cement Paraopeba 40 50 tank car Vale Paraíba 50 64 -  After the derailment at Vale do Paraiba, the STEPS TO INCREASE SPEED SAFELY harmonic behavior is taken into account to  L/V increase train speeds at MRS.  Hunting  Harmonic Behavior  No derailments related to harmonic behavior  Longitudinal Dynamic since 2013 ! 29

Conclusions  Harmonic excitation analysis is an important step to increase speed safely.  Multibody simulations can be applied to support new operation conditions.  The simulations can reduce cost and time in instrumentation field tests, if necessary.  The methodology presented is applicable for any vehicle, speed, and carbody modes. 30

Questions ? Raphael Marota raphael.marotta@mrs.com.br Lucas Valente lucas.valente@mrs.com.br 31