Dynamic analysis of a new double deck passenger vehicle with bogie PW200

1
Dynamic analysis of a new double deck passenger vehicle with bogie PW200
Zhenggang Lu* Prof. Markus Hecht **
* Shanghai Tiedao University
** Technische Universitt Berlin
Abstract
This paper dealt with the dynamic behavior of a new double deck passenger vehicle with
bogie PW200, which was designed and produced by Puzhen Rolling Stock Works of China to
meet the demands of high speed railway passenger transportation in China. A full vehicle
model including all nonlinear factors was built in the mechanical software ADAMS/Rail. The
results of simulation were compared with the measurement data of field test and another
simulation results using random equivalent linearization method in frequency domain, to
verify the model and determine the track irregularity level. Finally the major bogie suspension
parameters were optimized, and the improvement of curve dynamic behavior was obtained
using optimum parameters.
Key words
Dynamic analysis, simulation, ADAMS/Rail, high speed, railway vehicle
1. Introduction
Since the first quasi-high speed rail line, from Guangzhou to Shenzhen in China, came to use
in 1994, the maximum travel speed of passenger train has reached to 160 km/h, and the travel
time has decreased to 75 minutes from Guangzhou to Shenzhen. All this made the railway
passenger transportation more compatible than ever in this region. In 1995 the passenger
trains speed in main rail lines of China was increased to 140 km/h, to meet the increasing
demands of railway passenger transportation. Three different new types of passenger vehicles
were designed and produced by different rolling stock works last year in order to increase the
passenger train speed to 200 km/h. A new double deck passenger vehicle with bogie PW200
was produced by Puzhen Rolling Stock Works.
The field dynamic test including these three passenger vehicles was finished last August, the
maximum test speed reached to 240 km/h. All safety and comfort index were good while this
vehicle run on the straight track, but the lateral ride comfort index in low floor and the wheel
set lateral force in curve were somehow to big, and lateral comfort index of carbody in curve
was greater than 2.5 while test speed exceeded 210 km/h
As all known, The good dynamic behavior of the vehicle on both straight and curve track
could be obtained by designing the suspension system carefully, but the bogie suspension
parameters have almost conflict effects on straight hunting stability and curve passing
behavior. For the high speed bogie, we usually put more attention to the hunting stability and
safety factors, this make the wheel set lateral force and carbody acceleration tendentiously big
in curve track. With the development of computer simulation technology, now we can
simulate the vehicle dynamic behavior and optimize suspension system early in design period
Dynamic analysis of a new double deck passenger vehicle with bogie PW200
2
, or make some improvement of existing vehicle through parameter optimization.
In this paper a full vehicle model of double deck passenger car with bogie PW200 including
all non-linear factors was built in the mechanical software ADAMS/Rail , Firstly the major
work of model setup was to parameterize suspension systems, and to set up special nonlinear
wheel-rail contact geometry of Chinese HLM profile wheel with R60 profile rail and other
track parameters.
Secondary straight comfort analysis was finished. Vertical, lateral acceleration and comfort
index of carbody were compared with the results of field test, so as to verify the model, and to
determine the track irregularity level in simulation. At the same time, these results also
compared with simulation results finished by author previously using random equivalent
linearization method in frequency domain.
Thirdly hunting stability was studied with parameter variation. The suitable value of primary
guide longitudinal stiffness and link bar stiffness between bogie and frame were chosen to
give the sufficient high critical hunting speed. The next step was to study straight response
and further optimize secondary suspension parameters include vertical and lateral damper
coefficient. The dynamic curve behavior including comfort and safety index on both ideal
track and real track with irregularities were simulated. Consequently the major bogie
suspension parameters affecting curve behavior were optimized.
The final dynamic simulation indicated that the improvement of curve passing behavior was
obtained using optimum parameters.
2. Vehicle and bogie structure and model setting up
2.1 Bogie structure
The bogie PW200, see Figure 1, has primary and secondary suspension systems. There are
two vertical helical springs and two elastic rubber guide cylinders inside of helical springs on
every axle box in primary suspension. Only helical springs endure static empty weight of
vehicle, both helical springs and rubber cylinders endure the dynamic force in loaded and
running case. The wheel set connects elastically to bogie frame through elastic guide
cylinders. The guide cylinders can supply different longitudinal and lateral wheel set guiding
stiffness. There is an vertical hydraulic damper in every axle box. The wheel set guide device
shows in figure 2(a) ,
1: Bogie frame 2: Wheel set and primary suspension 3: Bolster and secondary suspension 4: Braking system
Figure 1 Bogie PW200
The secondary suspension includes bolster, airspring, vertical and lateral hydraulic dampers
Dynamic analysis of a new double deck passenger vehicle with bogie PW200
3
and lateral rubber bumpstop. The bolster connects with bogie frame through longitudinal link
bar at its ends. The whole bolster swing link arrangement hangs to frame through bolster link.
At each end of bolster link, there is a elastic rubber joint ,see figure 2(b), it can decrease the
secondary lateral stiffness, the friction force and wear of relative revolve joins.
In vertical direction, the carbody is supported on the bolster through side bearing with low
friction coefficient. In longitudinal direction, it connects with bogie through center tracing
pivot.
An anti-roll bar is installed between bogie frame and carbody to decease carbody roll angle in
curve track.
The bogie is equipped with both conventional block braking system and disk braking system.
(a) Wheel set elastic guide device (b) Elastic linker rubber joint
Figure 2 Wheel set elastic guide device(a) and elastic link rubber joint (b)
2.2 Mathematics model setting up
2.2.1 Vehicle model
The full vehicle mathematics model included one carbody, two bogies frames, two bolsters,
four wheel sets and eight axle boxes, all these parts were assumed as rigid bodies. Because
the carbody was a double deck, it had a very strong structure, FEM analysis and mode test
analysis indicated previously that the first vertical, lateral bend and torsional mode
frequencies were more than 15 Hz, so this assumption could be acceptable for the comfort
and stability analysis. Table 1 showed the basic data of vehicle and bogie.
Vehicle basic data Table 1
Wheelset base 2.4 m
Wheel diameter (new) 915 mm
Bogie base 18 m
Carbody mass 39000 kg
Height of center of gravity 2.346 m
Bogie mass 6900 kg
The primary elastic rubber guide cylinder was modeled as bushing element, with linear three
direction stiffness and little structure damping. The vertical helical spring was modeled as
linear general force vector , to give very exact translational and rotational characteristics.
The secondary suspension model was little complex, the swing bolster arrangement was in
series with airspring. Because the whole lateral stiffness of bogie had been measured before
field test, so the secondary vertical ,lateral and yaw connections between bolster and bogie
frame were modeled as general force vector. In longitudinal direction, the link bar between
bolster and frame was modeled as force vector.
Dynamic analysis of a new double deck passenger vehicle with bogie PW200
4
The side bearing was modeled as nonlinear single force complement. The magnitude and
direction of output force depended on the relative velocity of carbody and bolster at its
location. Lateral bumpstop was modeled as nonlinear single force complement, its output
force was depend on the relative lateral displacement between carbody and bolster in its
location. Anti-roll bar was modeled as simple linear torsion spring.
The full vehicle model in ADAMS/Rail was shown in figure 3.
Figure 3 Full vehicle model in ADAMS/Rail
2.2.2 Wheel-rail contact model
As description previously, The wheel has the HLM profile, the normal rolling diameter is
915mm. The rail has the R60 profile. The rail inclination is 1:40 in China. The wheel-rail
contact model is shown in figure 4. The new wheel and rail profile were used to calculate the
contact geometry and contact table, the linear contact model level 2a in ADAMS/Rail was
used in stability parameter study, the non-linear contact model level 3 was used in dynamic
stability, straight and curve dynamic response studies.
Figure 4 Wheel rail contact profile: HLM/R60
2.2.3 Track model
To determine track irregularities is always difficult because there are many factors affecting
the track geometry and statues. The following formulas are power spectral densities(PSD) of
track irregularities suggested in China .
Alignment irregularity:
6 $
$ $
&
5 &
=
+ +

2
2 2 2 2
( ) ( )

(1)
Dynamic analysis of a new double deck passenger vehicle with bogie PW200
5
Cross level irregularity:
6 $
& 9
&
& 6 5
=

+ + +


2 2
2 2 2 2 2 2
( ) ( ) ( )
(2)
Profile irregularity:
6 $
9 9
&
5 &
=
+ +

2
2 2 2 2
( ) ( )
(3)
Where:
C
=0.8246,
R
=0.0206,
S
=0.4380 rad/m
With low level perturbation: A
A
=2.119x10
-7
, A
V
=4.032x10
-7

m.rad
With high level perturbation: A
A
=6.125x10
-7
, A
V
=10.8x10
-7
m.rad
Because we can not get the real measurement data of test track irregularities, all the
simulation are finished using IVPSD function in ADAMS/Rail, These PSD spectrum could be
transferred to track exciting input in time domain. The final track irregularity level used in
model would depend on the comparison and agreement between dynamic simulation and field
test.
3. Model verify and comparison with field test
Firstly the quasistatic curve model was used to verify the model, the simulation results
showed that the carbody accelerations corresponded correctly to the theoretic ones given by
speed, curve radius and cant.
Then the original bogie and vehicle parameters in field test were used to simulate the straight
response of vehicle. The calculated lateral and vertical acceleration and ride comfort index
on the floor of carbody above bogie were compared with test data correspondingly in figure 5
and figure 6.
Lateral acceleration
0
0,1
0,2
0,3
0,4
0,5
0,6
160 180 200 220
V [km/h]
m
/
s
^
2
Test data
A/Rail
REL
Lateral ride index
0
0,5
1
1,5
2
2,5
3
160 180 200 220
V [km/h]
W
z
Test data
A/Rail
REL
Figure 5 Lateral acceleration and ride comfort index comparison
Vertical acceleration
0
0,25
0,5
0,75
1
160 180 200 220
V [km/h]
m
/
s
^
2
Test data
A/Rail
REL
Vertical ride index
0
0,5
1
1,5
2
2,5
3
160 180 200 220
V [km/h]
W
z
Test data
A/Rail
REL
Dynamic analysis of a new double deck passenger vehicle with bogie PW200
6
Figure 6 Vertical acceleration and ride comfort index comparison
At the same time, these results were also compared with dynamic simulation finished by
author previously using random equivalent linearization method(REL) in frequency domain,
the REL values of lateral and vertical acceleration are the three times of standard deviation(3
) in figure 5 and 6. All simulations were finished under the low track irregularity level
from 160 to 220 km/h.
Based on these analysis and comparisons, the acceptable choose of track irregularity is low
level. Through this could affect the simulation accuracy, it is useful to optimize the
suspension parameters and foresee the improvement of dynamic behavior using different
parameters.
4. Dynamic simulation and parameter study
4.1 Stability analysis and parameter study
The first step is to study linear stability using linear wheel-rail contact model, to get the
general idea how suspension parameters affect on the vehicle hunting stability.
Figure 7 and 8 show how the critical hunting speed of the vehicle changes with wheel set
guide longitudinal stiffness K
PX
and lateral stiffness K
PY
.
Critical speed
0
20
40
60
80
100
120
140
4 6 8 10 12 14 16 30
Kpx [MN/m]
C
r
i
t
i
c
l
a

s
p
e
e
d
[
m
/
s
]
Eff. conicity :0.1
Eff. conicity:0.2
Critical speed
0
20
40
60
80
100
120
1 2 3 4 5 6 8 12
Kpy [MN/m]
c
r
i
t
i
c
a
l

s
p
e
e
d

[
m
/
s
]
Eff. conicity:0.1
Eff. conicity:0.2
Figure 7 Linear critical speed & K
PX
Figure 8 Linear critical speed & K
PY
In order to obtain sufficient high stability , we chose the wheel set guide longitudinal stiffness
K
PX
>12 MN/m, lateral stiffness K
PY
>4 MN/m.
To determine the effect of the nonlinear side bearing friction coefficient
SB
in series with
secondary link bar stiffness K
SX
, we use nonlinear dynamic stability model, and nonlinear
wheel-rail contact profile new HLM/R60, the vehicle excited by special track model. The
final critical hunting speed of full vehicle is about 350 km/h, this is sufficient for this vehicle
with the maximum running speed 200 km/h.
4.2 Straight response analysis and parameter study
The straight response analysis is finished to optimize secondary suspension parameters, such
as secondary suspension lateral stiffness K
SY
, lateral and vertical damper coefficient C
SY
and
C
SZ
, side bearing friction coefficient
SB
. Figure 9 to figure 12 show how these parameters
affect on the vehicle ride comfort. All the simulation speed is 60 m/s, the accelerations which
are used to calculate lateral or vertical ride index are measured on the middle floor of carbody
above lead and trail bogie.
Dynamic analysis of a new double deck passenger vehicle with bogie PW200
7
Lateral comfort index
1
1,2
1,4
1,6
1,8
2
2,2
2,4
2,6
2,8
3
0,1 0,14 0,18 0,2 0,22
Ksy [MN/m]
C
o
m
f
o
r
t

i
n
d
e
x

W
z
Lead point
Trail point
Lateral comfort index
1
1,5
2
2,5
3
0 10 20 40 60
Csy [KN.s/m]
C
o
m
f
o
r
t

i
n
d
e
x

W
z
Lead point
Trail point
Figure 9 Lateral comfort index & K
SY
Figure 10 Lateral comfort index & C
SY
Lateral comfort index
1
1,5
2
2,5
3
0,01 0,04 0,08 0,12
msb
C
o
m
f
o
r
t

i
n
d
e
x

W
z
Lead point
Trail point
Vertical comfort index
1
1,5
2
2,5
3
10 20 30 40 60 80
Csz [KN.s/m]
C
o
m
f
o
r
t

i
n
d
e
x

W
z
Lead point
Trail point
Figure 11 Lateral comfort index &
SB
Figure 12 Lateral comfort index & C
SZ
4.3 Dynamic curve analysis and parameter study
The dynamic curve analysis is finished to optimize bumper stop free clearance and contact
stiffness, to see how the side bearing yaw moment affects on the lateral force and yaw angle
of wheel set. All simulations is under the same track condition: curve radius is 1500 m,
transition curve length is 80 m, track initial cant is 80 mm, cant deficiency is 150 mm(curve
passing speed is about 170 km/h). Track irregularities are also considered to calculate the ride
comfort index in the circular part.
Figure 13 and 14 show the effect of bumpstop characteristic, figure 15 show how side bearing
friction coefficient
SB
affects on the wheel set lateral force( no track irregularities in this
case).
Lateral comfort index
1
1,5
2
2,5
3
0,25 0,5 1 1,5
Bumpstop contact stiffness [MN/m]
C
o
m
f
o
r
t

i
n
d
e
x

W
z
Lead point
Trail point
Secondary lateral stroke
40
50
60
70
80
0,25 0,5 1 1,5
Bumpstop contact stiffness [MN/m]
L
a
t
.

s
t
r
o
k
e

[
m
m
]
Lead point
Trail point
Figure 13 Bumpstop contact stiffness & comfort index in curve (Free clearance 40 mm)
Dynamic analysis of a new double deck passenger vehicle with bogie PW200
8
Lateral comfort index
1
1,5
2
2,5
3
30 40 50 60
Bumpstop free clearance [mm]
C
o
m
f
o
r
t

i
n
d
e
x

W
z
Lead bogie
Trail bogie
Secondary lateral stroke
40
60
80
100
30 40 50 60
Bumpstop free clearance [mm]
L
a
t
.

s
t
r
o
k
e

[
m
m
]
Lead bogie
Trail bogie
Figure 14 Bumpstop free clearance & comfort index in curve (Contact stiffness 0.4MN/m)
Leadig wst outside lateral force
0
10
20
30
40
50
0,03 0,07 0,14
musb
L
a
t
.

f
o
r
c
e

[
K
N
]
Leading wst Ver. force/Lat. force
0
0,1
0,2
0,3
0,4
0,5
0,03 0,07 0,14
musb
Q
/
P
Figure 15 Side bearing friction coefficient
SB
& wheel-rail lateral force
4.4 Dynamic behavior improvement through parameter optimum
Based on above dynamic analysis and parameter study, we find that original major suspension
parameters of bogie PW200 are suitable and very close to optimum values, such as wheel set
guide longitudinal and lateral stiffness, secondary link bar stiffness, vertical damper
coefficient, but secondary lateral damper coefficient, bumpstop free clearance and contact
stiffness could be optimized further to give improvement of curve passing behavior.
Here we simulate the dynamic curve behavior under the field test condition: curve radius is
3500 m, transition curve length is 40 m, track initial cant is 60 mm, curve passing speed is
220 km/h.
The simulation results show in figure 16 to figure 18.
Figure 16 Lat./ Ver. force of wst3 left side Figure 17 Lateral force of wst3 left side
Dynamic analysis of a new double deck passenger vehicle with bogie PW200
9
Fig 18 Lateral acceleration on the floor of cardody above leading bogie
5. Conclusion
With the help of ADAMS/Rail, the dynamic analysis of a new double deck passenger vehicle
with bogie PW200 has been finished, and the major suspension parameters are optimized. The
main conclusions can be made from this study :
This mathematics model of the full vehicle with bogie PW200 gives general acceptable
agreement with test data, and is suitable for dynamic simulation and parameter optimum.
The parameters study shows that major suspension parameters of bogie PW200 are
suitable and very closed to optimum values.
The track irregularity has important effect on the carbody acceleration and ride comfort
in both straight and curve track case. the further measurements of the track irregularity
are useful to simulate the behavior of vehicle accurately.
The effect of suspension positions on dynamic behavior could be studied using this
model, to decease roll motion of the carbody and improve lateral ride comfort of low
floor of the carbody.