PAPER

Analyses of Contact Force Fluctuation between Catenary and Pantograph

Mitsuo ABOSHI Katsushi MANABE
Senior Researcher, Current Collection, Chief Researcher,
Railway Dynamics Div. Research & Development Promotion Div.

One of the most important subjects of overhead contact line and pantograph system is
to reduce the contact loss of pantograph in high-speed operation. In this research, the
relation between the wave motion of contact wire and the contact force fluctuation of pan-
tograph is studied; the contact force fluctuation in a hanger span cycle is analyzed; and
methods to reduce the contact force fluctuation are proposed. It is shown that the contact
force fluctuation of pantograph in high-speed operation is mainly caused by the incident
wave of contact wire that is generated by the unevenness of contact wire and is reflected at
hangers. This influence becomes extremely large when the train speed is close to the wave
velocity of contact wire. Increasing the wave velocity of contact wire or decreasing the
reflection factor of wave motion at hangers is expected to effectively reduce the contact
force fluctuation.

Keywords : Current collection, Contact wire, Wave motion, Pantograph, Contact force fluc-
tuation

1. Introduction enary. Because this equipment consists of three wires

(so-called "compound type"), the elasticity at the support
For high-speed operation in electric railways, reduc- point is comparatively small. And the total tension of
ing the contact loss of pantograph is one of the most im- this catenary amounts to 53.9 kN. Therefore, it is consid-
portant subjects. It has been confirmed in past research1) ered that the contact force fluctuation in a support span
that when the train speed is close to the wave velocity of cycle is comparatively small, and that in a hanger span
contact wire, the contact loss ratio becomes extremely cycle is influential in high-speed operation.
large and contact loss occurs in a hanger span cycle. It is
necessary to clarify the causes of the contact force fluc-
tuation between catenary and pantograph, and to pro-
pose an improvement method to reduce the contact force
fluctuation.
In this paper, the relation between the wave motion
of contact wire and the contact force fluctuation of panto-
graph is studied; the contact force fluctuation in a hanger
span cycle is analyzed; and methods to reduce the con-
tact force fluctuation are proposed.

2. Contact force fluctuation

The best contact condition is that the contact force
between contact wire and pantograph is always constant
and equal to the stationary contact force. But the con-
tact force changes in high-speed operation due to several
causes. The main causes are considered as (1) fluctua-
tion mechanism in a support span cycle, (2) fluctuation
mechanism in a hanger span cycle, (3) unevenness of con-
tact wire (undulating wear, for example), (4) aerodynam-
ics disturbance and others. The magnitude of the con-
tact force fluctuation in a support span cycle and that in
a hanger span cycle are indices to evaluate the dynamic
performance of overhead equipment.
Concerning the contact force fluctuation in a support
span cycle, the following improvement methods have been
shown in past research2), (1) reducing the variability of
catenary elasticity, and (2) increasing the whole tension
of catenary. A heavy-compound catenary is normally used
in Shinkansen lines. Fig.1 shows an outline of this cat-
Fig. 1 Standard overhead equipment for Shinkansen
(Heavy-compound catenary)
3. Analysis of contact wire wave motion
3.1 Method of measuring wave motions
The wave propagating velocity (phase velocity) of the
contact wire is one of the most important indices for the
current collecting performance. The wave velocity is ob-
tained by estimating the frequency at which the addition
or the subtraction of accelerations of 2points is equal to
0 3). A measurement example is shown in Fig.2. It is found
that the wave velocity increases at high frequency be-
cause of its flexural rigidity, and that the measured val-
ues agree well with the calculation.
In order to estimate the wave motion of the contact
wire, it is necessary to separate the measured vibration

182 QR of RTRI, Vol. 41, No. 4, Dec. 2000

into the forward wave and the backward wave. If the wave frequency of the incident wave, respectively; β is the ratio
motion y(x,t) is expressed by Eq.(1), two wave motions of the moving speed of pantograph to the wave propagat-
propagating in opposite directions are obtained by Eq.(2) ing velocity of contact wire; Z t, Z p are the mechanical im-
by using the gradient and the wave velocity of the wire c. pedance of contact wire and pantograph; and i is the
imaginary unit. The sign "+" is used when the wave mo-
y( x,t ) = f ( x − ct ) + g( x + ct ) (1) tion comes incident from the front of the pantograph, while
1 ∂y( x,t ) "-" is used for the wave motion from behind. Eq.(4) means
f ( x − ct ) = ｛y( x,t ) − c ∫ dt｝
2 ∂x that the magnitude of the contact force fluctuation is pro-
(2) portional to the vertical velocity amplitude of the wave
1 ∂y( x,t )
g( x + ct ) = ｛y( x,t ) + c ∫ dt｝ incident to the pantograph.
2 ∂x
A measurement example of the wave separation is
shown in Fig.3 (vertical acceleration limited under 60 Hz).
It is seen that the wave motions propagate in different
directions on the contact wire.

Fig. 4 Analysis model for contact force fluctuation caused

by wave motion of contact wire

In order to verify the above-mentioned theoretical

Fig. 2 Measurement example of wave propagation velocity
analysis, we measured the wave motion of the contact
wire incident to the pantograph and the contact force
fluctuation at the same time. An actual catenary and
pantograph are available for the experiment with cur-
rent collecting test equipment. In order to measure the
contact force, load-cells and accelerometers are set un-
der the contact strips of PS200A type pantograph. A con-
tact wire is only provided for the running experiment,
in order to simplify observation of the wave propagation.
The result of the experiment in rigid support is shown
in Fig.5 (speed:100 km/h, β =0.28). "V12_F""V34_G" are
the vertical velocities of the wave motion passing through
the measuring points (limited under 60 Hz). It is found

Fig. 3 Measurement example of wave separation

3.2 Contact force fluctuation caused by wave

motion of contact wire

In the analysis model as shown in Fig.4, if the dis-

placement of the wave motion incident to the pantograph
y is expressed by Eq.(3), the amplitude of contact force
fluctuation Fp caused by the incident wave is obtained by
Eq.(4),

y = A exp{iω( t - x/ct )} (3)

1
F p = (1 # β )ω A Z , Z= (4)
1/ Zt +1/ Z p
where A and ω are the amplitude and the angular Fig. 5 Result of fundamental experiment

QR of RTRI, Vol. 41, No. 4, Dec. 2000 183

that the wave motion propagates forward  , reflects which two infinite strings are combined with a spring-
at the next support point  , comes incident to the pan-
tograph  , and repeats the reflection between the sup-
port point and the pantograph  . A large contact
force fluctuation is caused when the wave motion comes
incident to the pantograph. The measuring results agree
well with the theoretical one.
damper element. The complex reflection factor R (the
ratio of amplitude of reflected wave A1r to that of incident
wave A1i) is obtained by Eq.(8), where the spring constant,
the damping factor and the mass of a hanger is k , D, mt,
mm, and the line density and the tension of contact wire
and messenger wire are ρt , ρ m , T t , T m.

3.3 Generation of wave motion by unevenness of contact A1r Zh

wire γR = =−
A1i Zt + Zh
As shown in Fig.6, we suppose that the contact wire 1
Z h = iω mt + (8)
is an infinite string which has the unevenness of a wave 1 1
+
length
and one-side amplitude B. When a pantograph Z m + iω mm D + k / iω
runs under the contact wire (speed;v), the amplitude of
vertical velocity of the generated backward wave and for- Z t, Zm are the mechanical impedance of contact wire
ward wave |V1|, |V2| and the angular frequency ω , ω2 and messenger wire, respectively, shown in Eq.(9).
are obtained by Eq.(5) and Eq.(6). The contact force fluc-
tuation caused by the unevenness of contact wire Fue is Z t = 2 ρ t Tt , Z m = 2 ρ mTm (9)
obtained by Eq.(7).

Zp 1 1
V1 = ω B, ω1 = ω0 (5)
Zt + Z p 1+ β 0 1+ β
Zp 1 1
V2 = ω B, ω2 = ω0 (6)
Zt + Z p 1− β 0 1− β
ZtZ p
Fue = −iω 0 B exp iω 0t (7)
Zt + Z p
2π v v
ω0 = , β=
λ ct
From these equations, it is indicated that the fre-
quency and the vertical velocity amplitude of the forward
wave increases when the speed becomes high, and the
forward wave is especially influential at high-speed. Con- Fig. 7 Analysis model for wave reflection at hanger
cerning the relation between the unevenness of contact
wire and the amplitude of generated wave motion, we have
confirmed that the measuring results of fundamental ex- 4. Mechanism of contact force fluctuation in
periments with low spring constant hangers agree well hanger span cycle
with the theoretical analysis 4).
As shown in Fig.8, we suppose that wave motion ① is
3.4 Reflection of wave motion at hanger point generated by the dip of contact wire between hangers, and
② reflects at the nearest hanger, ③ becomes incident to
As shown in Fig.7, we suppose an analysis model in the pantograph and causes the contact force fluctuation.

Fig. 6 Analysis model for wave generation by unevenness Fig. 8 Analysis model for contact force fluctuation in
of contact wire hanger span cycle

184 QR of RTRI, Vol. 41, No. 4, Dec. 2000

 Fig.10 shows a measurement example of Shinkansen
that the forward wave is generated by the dip of contact
wire between hangers and propagates forward. Under the
conditions of
=5 m, c t= 114 m/s, v=61.1 m/s (220 km/h),
the frequency of the generated wave motion is calculated
at 26 Hz from Eq.(12). The frequency of the forward wave
observed by this measurement is mostly in agreement with
this calculated value. It is confirmed that the contact wire
wave motion is also generated by the dip between hangers
in actual equipment.

Fig. 9 Measurement example of unevenness of contact

wire in Shinkansen

Fig.9 shows a measurement example of unevenness of

contact wire of Shinkansen. The unevenness of wavelength
λ = 5 m (wave number =0.2) is equivalent to the dip be-
tween hangers. The shorter wavelength is also seen in this
measurement result.
If the dip between hangers is assumed to be a second-
ary curve, the k-th coefficient of Fourier series a k is ex-
pressed by Eq.(10), where L h is the span length between
hangers.
2
ρ t g ⎛ Lh ⎞ Fig. 10 Example of wave generation by dip between
ak = ⎜ ⎟ cos (π k ) (k ≠ 0) (10) hangers
2Tt ⎝ π k ⎠
The amplitude of the k-th unevenness is 1/k 2 times the Next, if it is assumed that the forward wave motion
primary unevenness, and its wavelength is 1/k times. reflects and returns γ R times at a hanger point, the verti-
Therefore, it is shown from Eq.(6) that the vertical veloc- cal velocity amplitude and the angular frequency of the
ity amplitude of the contact wire wave motion generated incident wave to a pantograph |V3|, ω 3, and the amplitude
by the k-th unevenness becomes 1/k times in the case of and the angular frequency of contact force fluctuation
primary unevenness. In other words, the influence of the |Fwv|,ω wv are obtained as follows.
higher order unevenness of the dip between hangers to the
generation of contact wire wave motion is comparatively g Zp β Lh
small. V3 = γ R・ ・
If the contact wire unevenness is assumed to be only π Zt + Z p ω =ω
1 − β ct
0

the primary one of the dip between hangers, the amplitude β ct (14)
of contact wire unevenness B is expressed by Eq.(11). ω 3 = 2π
2
1 − β Lh
g ⎛ Lh ⎞
B = a1 = 2 ⎜ ⎟ (11)
2c t ⎝ π ⎠ 2 ρ t gLh Zp Zp (1 + β )β
Fwv = γ R・
π Zt + Z p ω =ω 0
Zt + Z p ω =ω wv
1− β
The vertical velocity amplitude and the angular fre-
(1 + β )β ct
quency of the generated wave motion of contact wire (for- ω wv = 2π
ward wave) |V2|, ω2 are obtained by Eq.(12), and the am- 1 − β Lh
plitude of the contact force fluctuation caused only by the (15)
unevenness of contact wire |Fue| is obtained by Eq.(13).
Fig. 11 shows the contact force fluctuation caused by
g Zp β Lh the contact wire wave motion Fwv presumed from the mea-
V2 = ・
π Zt + Z p ω =ω
1 − β ct surement result of the contact wire unevenness and re-
0
sidual diameter in the case of Shinkansen. The presumed
β ct (12)
ω 2 = 2π contact force fluctuation by the contact wire wave motion
1 − β Lh Fwv is obtained by deducting the contact force fluctuation
by contact wire unevenness Fue from the contact force fluc-
2 ρ t gLh Zp tuation Fue+Fwv equivalent to residual diameter under sup-
Fue = ・β
π Zt + Z p ω =ω 0
position that contact wire wear is proportional to the con-
tact force. This example is for the PS202 pantograph type
ct (13)
ω 0 = 2πβ at 200 km/h. The tendency of the contact force caused by
Lh the contact wire wave motion Fwv becoming higher is regu-

QR of RTRI, Vol. 41, No. 4, Dec. 2000 185

larly seen at the 0.7~1.0 part of hanger span. The phenom- ning speed approaches the wave propagation velocity of
enon in which the generated contact wire wave motion re- contact wire, is mainly based on the contact force fluc-
peats reflection between a front reflecting point and a pan- tuation in a hanger span cycle.
tograph is observed by the above-mentioned fundamental The amplitude of contact force fluctuation caused by
experiment. In this case, the first and second positions x 1, the reflective wave motion |Fwv| is compared and shown
x2 at which the contact wire wave motion becomes incident in Fig.13 for the case of heavy-compound (HC) catenary
to a pantograph are 0.72 and 0.92 respectively in a hanger and the case of high-tension heavy-compound (HTHC) cat-
span. These points agree well with the point that the con- enary for which the tension of contact wire is increased to
tact force fluctuation caused by wave motion Fwv is higher 19.6 kN. In this case, it is calculated with the parameter
in this figure. This means that the wave reflection at the of PS202. In the case of high-tension heavy-compound
nearest hanger is influential to the wave incident to the catenary, the wave propagation velocity of contact wire is
pantograph. about 15 % larger, and the reduction effect of contact force
It has been confirmed that the contact loss in a hanger
span cycle occurs near the center between hangers and
that the residual diameter near the center between hang-
ers is larger. We think that these phenomena result from
this mechanism. That is, the contact force fluctuation in a
hanger span cycle is considered to be the phenomenon in
which are compounded with the contact force fluctuation
by the dip of contact wire between hangers Fue and that by
the contact wire wave motion reflected at a hanger Fwv.

Fig. 12 Calculated example of contact force fluctuation

Fig. 11 Contact force fluctuation by wave motion presumed

11
from measured unevenness and residual diameter

5. Contact force fluctuation in high-speed

operation

5.1 Speed characteristic of contact force

fluctuation in a hanger span cycle
Fig. 13 Comparison of contact force fluctuation
Fig.12 shows an example of the amplitude of contact
force fluctuation |F ue|, |Fwv |which are calculated by
Eq.(13) and Eq.(15). These curves show the characteris-
tics to non-dimensional speed in the condition of heavy-
compound catenary and Zp>>Zt. The contact force fluctua-
tion by contact wire unevenness |Fue| increases linearly
with speed. On the other hand, when exceeds about 0.7,
the contact force fluctuation by the reflective wave mo-
tion |Fwv| increases remarkably. This characteristic is
based on the so-called Doppler effect as shown in Eq.(15).
Consequently, it is shown that the influence of wave re-
flection becomes larger especially in high-speed opera-
tion.
This characteristic of contact force fluctuation agrees
well with the radically increasing tendency of contact loss
ratio. Therefore, we think that the phenomenon, which Fig. 14 Comparison of contact force fluctuation
the contact loss ratio increases remarkably as the run- (Non-dimensional speed)

186 QR of RTRI, Vol. 41, No. 4, Dec. 2000

fluctuation by decreasing β is seen. Fig.14 shows the com-
parison of these cases at non-dimensional speed β, these
characteristics have almost the same curves. As shown
in Eq.(15), this is because the amplitude of contact force
fluctuation |Fwv| has almost the same non-dimensional
speed characteristic when the contact wire is the same
type and the reflective coefficient is the same.
Here, by using Fig.13, we will estimate the maximum
speed that a pantograph can run without contact loss.
We suppose that the stationary uplift force of the mov-
ing pantograph is 74 N (static uplift force 54 N + aero-
dynamic uplift force 20 N) and when the amplitude of
contact force fluctuation caused by wave motion reflec-
tion of contact wire exceeds the stationary uplift force,
contact loss will occur. On the speed characteristic of
this figure, it is indicated that the maximum speed with-
out contact loss is about 270 km/h for the heavy-com-
pound catenary, and is 320 km/h for the high-tension
heavy-compound catenary. In an actual case, it is thought
that contact loss occurs below such speed because of other
unevenness except for the dip between hangers. However,
the speed shown here becomes the standard of the maxi-
mum operation speed as an index of contact performance
of overhead equipment.
5.2 Improvement methods of catenary for high-speed
operation
From Eq.(15), improvement methods for high-speed
operation are shown as follows.
(1) Increasing wave propagation velocity of contact wire.
(2) Decreasing density of contact wire.
(3) Reducing wave reflection of contact wire.
(4) Shortening hanger span length.
The present overhead equipment of Shinkansen for
300km/h operation has a high wave propagation velocity
of contact wire by heightening the tension of contact wire,
and adoption of lightweight contact wire. This is desir-
able from the viewpoint of this research. Moreover, if it
is the same wave propagation velocity, the smaller the
line density of contact wire is, the better the performance
is.
In order to reduce the reflection of contact wire wave
motion at a hanger point, use of hangers which have a
spring and a damper mechanism is effective. Although
the use of these damping type hangers cannot remark-
ably change the increasing tendency of contact loss de-
pending on the relation of moving speed and the wave
propagation velocity of contact wire, it is expected to re-
duce the contact force fluctuation below the critical speed.
A reflection factor of wave motion with a damping hanger
will be calculated as shown in Fig.15. A coil spring is used
in the friction damping hanger, and a piece of rubber is
used in the rubber damping hanger. From the results of
the field test carried out on Sanyo Shinkansen line, it is
found that there is a close correlation between the con-
tact loss ratio and the vertical velocity of backward wave
motion. The rubber damping hangers can reduce the
magnitude of the wave motion incident to the pantograph
and the contact loss ratio.
In Shinkansen lines, the measurement results shows
QR of RTRI, Vol. 41, No. 4, Dec. 2000
Fig. 15 Reflection factor of wave motion at hanger
that the contact loss ratio in the section of hanger span
length 3.5 m is smaller than that in the section of 5 m.
We think that the result obtained by this research ex-
plains such measurement result.
6. Conclusion
In the existing catenary equipment for Shinkansen
lines, the contact force fluctuation in a support span cycle
is comparatively small because it is a compound type and
the whole tension is higher. Therefore, the contact force
fluctuation in a hanger span cycle is comparatively large.
From the above study, we get the conclusions summarized
below about the mechanism of contact force fluctuation
and the improvement methods for high-speed operation.
(1) It is shown that the contact force fluctuation of panto-
graph in high-speed operation is mainly caused by the
incident wave of contact wire that is generated by the
unevenness of contact wire and is reflected at hang-
ers.
(2) This influence becomes extremely large when the train
speed is close to the wave velocity of contact wire. The
amplitude of contact force fluctuation exceeds the sta-
tionary uplift force of a pantograph at about 0.7 or
more non-dimensional speed.
(3) Increasing the wave velocity of contact wire or decreas-
ing the reflection factor of wave motion at hangers is
effective to reduce the contact force fluctuation.
References
1) Manabe,K.:High-speed contact performance of cat-
enary -pantograph system, JSME, Vol. 54, No. 504,
pp. 1843-1851, 1988
2) G.Gilbert: Pantograph motion on a nearly uniform
railway overhead line, PROC. IEE, Vol.113, No3,1966
3) Aboshi,M.: Research for contact loss reduction
method by damping catenary vibration, WCRR'97,
Vol. C, pp151-159, 1997
4) Aboshi,M.: Relation between wave motion and un-
evenness of contact wire, JSME, D&D'98, 1998
187