Presentazione di PowerPoint

1
Numerical simulation for improving the design of
running gear Part 1 improvement of vehicle
dynamic behaviour Paolo BELFORTE, S. BRUNI
(Politecnico di Milano - Department of Mechanical
Engineering) Michael JÖCKEL (Fraunhofer Institute
for Structural Durability and System Reliability
- LBF)
2
MODTRAIN Project
MODTRAIN project
Innovative modular vehicle concepts for an
integrated European railway system
6th FRAMEWORK PROGRAMME PRIORITY 6.3
Transport 4 Years Project Started January 2004

• Modular approach to train design
• Interoperability new generation rolling stock
• Harmonised European criteria for rolling stock
  homologation

3
MODTRAIN Project
It consist of five different sub-projects

• MODBOGIE
• MODCONTROL
• MODPOWER
• MODLINK
• MODUSER

4
MODBOGIE SubProject

• ModBogie Subproject has 11 partners
• S.I. ANSALDOBREDA / ALSTOM / BOMBARDIER /
  SIEMENS
• Wheelset manufacturer LUCCHINI SIDERMECCANICA
• Operators DB / TRENITALIA / SNCF
• Research Institutes / Universities POLIMI /
  LBF-IWM / D2S.
• SubProject leader is ANSALDOBREDA

• ModBogie SubProject is dedicated to the
  optimization of the bogie, leading to
• improved performances in terms of energy
  efficiency
• enhanced bogie design for fulfill more demanding
  operational requirements
• wider dynamic performances with reduced
  environmental impact and maintenance costs.

5
INTRODUCTION NUMERICAL SIMULATIONS TOWARDS
VIRTUAL HOMOLOGATION

• In last years, the improved calculation
  technologies allowed the development of more
  detailed and accurate numerical models of rail
  vehicle dynamics, which can be used as a very
  useful tool for the design and development of a
  railway stock.

• With the development of new generations of HS
  trains, numerical simulations can give an
  important contribution in order to raise service
  speed and satisfy operators requirements which
  claims always for improved performance in terms
  of comfort and safety

6
INDEX
7
Vehicle model HS concentrated power locomotive
REFERENCE SYSTEMS
VEHICLE SCHEMATISATION
Loco of a concentrated power train
Carbody with two motor bogies Two motors
bogie-suspended by means of dedicated motor
hangers per each bogie
Only rigid modes also for the wheelsets ? problem
confined to low frequency
The equation of motion ? Lagrange equations
Vehicle inertia
W/R contact forces
8
Wheel rail contact forces model
9
COMPARISON A.D.Tre.S. SIMPACKEigenvalues and
time histories comparison
Natural frequencies comparison

• Straight track with concentrated track defect
• 5 mm lateral and 14 mrad roll
• 20 m wavelength
• speed 72 km/h.

Carbody natural frequencies
10
INDEX
11
Tuning procedure by sensitivity analysis
TYPE OF ANALYSIS parametric analysis on
primary suspension parameters and bogie
wheel-base straight track running behaviour -gt
critical speed curve negotiation -gt steady
state Q (vertical force values) steady
state Y (lateral force values) steady
state wear index
12
Tuning procedure by sensitivity analysis effect
of wheel-base
Reducing the wheelbase the vehicle has a better
steering behaviour
Reducing the wheelbase the critical speed
decreases
13
Tuning procedure by sensitivity analysis effect
of wheel-base
Radius curve m
Reducing the wheelbase the track shift force is
lightly increased
Wear index is lower in case of reduced wheelbase
14
INDEX
15
Analysis of technological options virtual
dynamic homologation simulation acc. to EN14363
Vehicle configurations taken into account for
EN14363 full analysis

• Three curve ranges are considered
•  Small radius curve (250 400 m)
• Medium-small radius curve (400 600 m)
• Large radius curve (600 2500 m) .

16
Analysis of technological options virtual
dynamic homologation simulation acc. to EN14363

• For each curve ranges a number of 30 sections, is
  considered. Per each section, a combination of
  the following parameters is chosen 
• Curve geometric parameters such as radius curve,
  cant and length of transition curve
• Wheel rail profiles
• Track irregularity (different small level one
  track irregularities)
• Speed, chosen randomly, imposing a cant
  deficiency of 110 of the admissible for at least
  20 of the complete simulation set.

17
Virtual dynamic homologation procedure main
curving indexes
Main parameters are obtained for all vehicle
configurations
TRACK SHIFT FORCE
Y/Q
VERTICAL FORCE
18
Virtual dynamic homologation procedure
critical speed and wear index.
Additional information is the wear index which
can be used for the evaluation of the
aggressiveness of the vehicle.
WEAR INDEX
CRITICAL SPEED
19
Parametrical analysis resultsSteady state
analysis
CRITICAL SPEED
GUIDING FORCE
TRACK SHIFT FORCE
WEAR INDEX
20
Sensitivity analysis and scatter prediction
Numerical simulation can be used even for the
evaluation of the impact of the scatter variation
of vehicles parameters on running behaviour.
21
Sensitivity analysis and scatter prediction
effect of damper parameters
Exemplary Simulation Results (12 Parameters
Varied Simultaneously) example of the
correlation of the damper parameters with
vertical wheel/rail contact forces.
Secondary suspension vertical damper (left)
Primary suspension vertical damper (left
front)
Each point Output for one sample-set (simulation)
Scatter of output
Max. normal force Fmax N
D11
? Strong correlation
? No correlation
Damper coefficient D1 Ns/m
Damper coefficient D2 Ns/m
22
INDEX
23
Methodology for the assessment of technological
optionsFULL FACTORIAL APPROACH

• Full factorial approach
• Dynamic performances analysis in straight track
  vehicle stability
• Dynamic performances analysis in curved track
  curving performance

24
Methodology for the assessment of technological
optionsFULL FACTORIAL APPROACH
Evaluate the influence of a simultaneous
variation of parameters

• Definition of factor and factor levels
• bogie wheelbase 3 m - 2.75m - 2.5 m
• lateral axlebox stiffness10-25-40 kN/mm
• longitudinal axlebox stiffness 10-30-50 kN/mm.
• ANOVA method distinction random and systematic
  variation ? polinomial equation of full
  factorial plan where coefficients a are
  determined applying the least square analysis

Reduced number of configurations
polynomial equation that describes the full
factorial plan
25
RESULTS IN STRAIGHT TRACK critical speed as a
function of bogie wheelbase and axle boxes
stiffness
265 km/h
245 km/h
24
BW 3 m
BW 2.75 m
BW 2.5 m
230 km/h
BW 2.5 m
26
RESULTS IN CURVEDTRACK wear rate as a function
of bogie wheelbase and axle boxes stiffness
Leading outer wheel frictional work small radius
curve
BW 3 m
18 kJ
BW 2.5 m
14 kJ
20
27
OPTIMIZATION results with different optimization
functions
Two different optimisation functions were used.
Wear index based optimisation
Reference vs. Opt.1 reduced wear 2
increased critical speed 5
Combined optimisation
Reference vs. Opt. 2 increased critical speed
of 16 increased wear of
4
28
CONCLUSIONS

• Numerical simulation can be used in order to
  complement physical testing for homologation
• Montecarlo approach coupled with multi-body
  simulations can account for the effect of scatter
  in component performances on ride safety
• Numerical simulations can also be used for
  optimising vehicle performances still meeting the
  constraints imposed by ride safety.

29
Thanks for your attention
BOGIE 07 Conference September 3rd - 6th, 2007
Budapest HUNGARY
Paolo BELFORTE paolo.belforte_at_polimi.it
Stefano BRUNI stefano.bruni_at_polimi.it
Michael JÖCKEL michael.joeckel_at_lbf.fraunhofer.de
30
COMPARISON A.D.Tre.S. SIMPACKEigenvalues
comparison
Natural frequencies with linearised contact
forces (Kalkers linear theory)
Natural frequencies computation
Fx -f33x Fy -f11h-f12f Mz f12h-f22f
Carbody natural frequencies
Carbody natural frequencies
31
COMPARISON A.D.Tre.S. SIMPACKTime domain
comparison

• Straight track with concentrated track defect
• 5 mm lateral and 14 mrad roll
• 20 m wavelength
• speed 72 km/h.

32
COMPARISON A.D.Tre.S. SIMPACKTime domain
comparison
Curved track without track defect R2000 m,
a.n.c. 1.32 m/s2, speed 185 km/h
33
Methodology for the assessment of technological
optionsSIMULATIONS PARAMETERS

• STRAIGHT TRACK
• Per each configuration
• MB simulations increasing speed (steps 5 km/h)
• Evaluation of rms values
• Evaluation of prescribed limits identification
  of critical speed
• Simulation parameters
• W/R profile theo. Rail / worn wheel cant 140
• Track irreg ERRI LOW
• The overall assessment of one vehicle
  configuration requires at least 50 simulations

• RMS calculation
• Fourier trasform of the last 10 s of the
  simulation
• Frequency f0 corrisponding to the maximum
  spectrum value identified
• Time history filtered with a band-pass filter
  f02 Hz

34
Methodology for the assessment of technological
optionsSIMULATIONS PARAMETERS

• CURVED TRACK
• Simulation parameters
• Steady state condition for different radius curve
  (300 2500 m) random combination of
• Track irregularity
• W/R profile
• Cant deficiency

Three tests zone small radius curves 250 -400
m small radius curves 400 600m radius
curves 600 2500m For each zone -gt 30
sections -gt data collected with simulations
35
Methodology for the assessment of technological
optionsOPTIMISATION PROCEDURE
Best vehicle w.r.t stability and wear ?
optimisation function
Ccs Cww ? critical speed and minimum frictional
work a b ? weighting coefficient All the
indexes prescribed in the standard were
considered as constrains
36
Results -- CURVED TRACKGuiding force as
function of bogie wheelbase and axle boxes
stiffness
Leading outer wheel guiding force small radius
curve
BW 2.5m
BW 3m
Low bogie wheelbase has positive effects on the
vehicle curving behaviour Longitudinal stiffness
reduces the bogie steering capability
37
Results -- OPTIMISATION
Best vehicle parameters optimisation procedure
result
Ref vs Opt.1 Increased critical speed of 16
Increased wear of 4 Ref vs Opt.2
Increased critical speed of 16
decreased wear of 2
high lateral stiffness and high boogie wheelbase
38
INDEX