Chapter

AMERICAN RAILWAY ENGINEERING AND

MAINTENANCE OF WAY ASSOCIATION
Practical Guide To Railway Engineering

EuropeanCurve
and Turnout
Mechanics
12-
12-1
©2003 AREMA®
AREMA COMMITTEE 24 - EDUCATION & TRAINING

European Curve and Turnout Practices

Miodrag Budisa
(P.E Railroad Engineer – Germany)
(P.E Railroad Engineer – Yugoslavia)

Consoer Townsend Envirodyne Engineers

Chicago, IL
[Link]@[Link]

12-
12-2
©2003 AREMA®
 Chapter
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

European Curve and

Turnout Mechanics
Railroading is railroading, although the methods that we use to get
to the end product may vary significantly. An in-depth
understanding of the geometrical relationships that are common to
all railway configurations is essential, whether designing, building
or maintaining a high-speed passenger line or a 40 mph drag coal
line.

12.1 Introduction

T he purpose of this chapter is to examine mathematically some key components

of the track structure, curves and turnouts, but from a European perspective.
The European railway, in many ways, is significantly different than the typical
railway of North America. High-speed rail plays a very significant role in not only
Western Europe, but also in Eastern Europe as well. In Europe, curves are built broad
and long, favoring high-speed operations. The approach in North America is entirely
different. Privatized North American railways were built to keep construction costs
down and to bridge great distances as quickly as possible. This translates into sharp
curvature and heavy grades, even for heavy haul activities. The need to move bulk
commodities over great distances favored the loading and design approach used by
North American railways. The excellent road system developed over the vast majority
of the populated segments of North America has relegated the majority of passenger
rail travel to a limited few heavily populated corridors. In Europe, the exorbitant cost
of fuel, a very high population density within countries the size of states or provinces,
along with a nationalized system of railways, has necessitated and enabled the
development of an extensive passenger rail based system.

Naturally, the European and North American rail networks evolved in two very
different directions. The one system is based solely on the reliable movement of heavy

12-
12-3
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

tonnage at the lowest cost, the other on pure speed. Vehicle and truck/suspension
development also took diverging paths between the two continents. European trucks
are set much further in than their North American counterparts. Truck (bogey)
suspensions handle track anomalies much differently in Europe. Europeans do not
stagger joints as is done in North America. Rock-off is unheard of in jointed territory.
On the other hand, vehicle bounce can be accentuated.

Today though, the North American engineer may need to take a new look at how our
counterparts "across the pond" have surmounted the problems of dealing with
operating at high speeds within existing alignments. North American railways are
looking for ways to operate faster at lower costs. European tangential turnouts have
been successfully installed in a number of heavy-haul territories. There is renewed
interest in high-speed inter-city passenger trains with a number of feasibility studies
underway. A new 110-mph Amtrak/IDOT service between Chicago and Springfield,
Illinois will soon initiate service.

This chapter does not pretend to cover all the significant design approaches used by
European railways in the handling of curvature and turnouts, particularly at high speed.
Nor does it present itself as being totally inclusive of European practices. There is
wide variance between systems in Western and Eastern Europe. But the reader
hopefully will get an appreciation of why Europeans have taken the approach they
have to these two topics. In addition, the commonality will also be apparent and
hopefully, the reader will secure a better appreciation of why we in North America
have developed the standards that we have.

12.2 Curves
12.2.1 Curve Definition
Prior to discussing curve engineering, one must have a common method of defining a
curve. There are two ways of describing curvature in common practice. In North
America, a railway curve is described by the angle in degrees subtended by two radii,
whose end points on the curve form a chord of 100 feet in length. In other parts of
the railway world, the length of the radius described above, measured in meters,
describes the curve.

The circumference of a complete circle is 2 ⋅ R ⋅ Π . Since the full body of a curve is

theoretically circular in construction, each degree of curvature will describe 100 feet of
chord (at relatively small degrees of curvature, the arc distance is approximately equal
to the chord distance). Thus, the circumference for such a circle made up of 100 foot
chords, each describing 1° of curvature, would for practical purposes, be 3600 feet and
the radius would be:

12-
12-4
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

36000 36000
R= = = 5730 feet
2Π 2 ⋅ 3.14

Thus, a 1-degree curve has radius of 5,730 feet. (Divide 5,730 by the degree of the
curve to get the radius in feet.) Converting 5,730 feet into meters, we find the same 1°
curve has a radius of 1746m.

In North America, one says, e.g., 10-degree curve. Outside of North America, one
would say, e.g., 174.6m radius to describe the same curve.

12.2.2 Gage
Obviously, excellent gage is a requirement for high-speed track, especially in curves. In
Europe, the distance between the two rails is also called the flangeway.

As in North America, the gage (flangeway) between the rails is measured on the gage
face at right angle to the axis of the rail. Whereas in North America, gage is measured
5/8" below the top of the rail regardless of rail section; in Europe, the distance
measured down is dependent on the particular rail section in use, a distance of 10 – 15
mm below the top of rail. The specific distance down is associated with the point on
the gage face where the radii begin. This is the controlling point, which emphasizes the
conicity of the
wheel/rail interface
as the wheel
progresses around
the curve.

Allowable gage
variance in Europe is
determined not on
the basis of the
speed operated (as is
done in the United
States and Canada -
under the FRA
213.53 Gage
requirements), but
simply on an Figure 12-1 Wheel and Rail Interface – Courtesy of Miodrag Budisa
allowable deviation
in tolerance for all
speeds.

Determination of the tolerance limits is derived directly from the wheel dimensions
and permissible wear rates that have been established. The standard gage width is

12-
12-5
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

1435mm. For curves where the radius R ≥ 250m, the Office for Research and
Experiments – International Union of Railways (ORE UIC) has determined that the
flangeway for a speed of V=100km/h (62 mph) should be 1432mm because it causes a
decrease in the truck hunting of the vehicle and wheel angle of the attack.

In Europe, wheels are condemned when the flange angle is worn down to 14° versus
the new wheel flange angle of 25°. (See Figure 12-1) Per Figure 12-1, when the flange
and the rail are new, the overall gage width is 1426mm. Thirteen hundred and sixty
millimeters is the inside to inside wheel mounting dimension and the flange thickness
for each wheel is 32.5mm; thus E, the gage, is:
E = 1360 mm + 2[32.5 mm] = 1425mm (56.102”)
On the French SNCF railway, the gage is within this value:

E = 1432 +−52 mm

and if the maximum tolerance is to be considered, this will permit a total gage width of:

E = 1435 + 3 = 1438mm (56.614”),

which allows the possibility of a lateral movement of the wheel of:
1438 – 1425/2 = 6.5mm (0.255”)
This translates to a permissible lateral wheel movement from center of slightly more
than 1/4".
Obviously, as wheel flange wear increases, so does truck hunting. On the SCNF
system, wheels are replaced when the flange wear exceeds 9.5mm. Rail is replaced
when the gage wear exceeds 9mm. On the German DB system, wheels are replaced
when the flange wear exceeds 8mm (approximately 5/16”).
With these tolerances, it is possible to get a maximum lateral wheel movement of
16.25mm (approximately 5/8”) under the most extreme conditions. For both wheels,
total lateral movement is:
6.5 + 16 + 10 = 32.5mm (approximately 1-1/4”) total
Note: The FRA 213 Regulations permit a maximum deviation in gage of 1” for
passenger equipment operating at Class 5 (up to 90 mph). This does not consider
additional movement due to wheel wear.

12-
12-6
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

12.2.3 Elevation in Curves

In Chapters 6 and 8, we learned that the centrifugal force experienced by a body
operating around a curve is defined by:

C = V2/R,

Where V = Speed of the body and

R = Radius of the curve

V = Speed km/Hr.

v = the speed traversed in 1 second

(meters/sec)

or 1000 meters/60min x 60 sec/min

C = v2/3.62 R
Sin α = h/s

Where h = amount of cant

(superelevation) and

s = gage
Figure 12-2 Centrifugal Force – JZ (Yugoslavian) Superstructure
Regulations
a = (C –g tan α) cos α

a = (v2/13R –g tan α) cos α

Where a = Uncompensated Car Body Acceleration, g = gravitational constant = 9.81

m/sec2 and α = angle between the tilting car and the plane across the top of the rails.
See Figure 12-2.

For very small angles, tan α is approximately sin α and cos α = 1. Thus:

éæ v 2 ö h ù v2 h
a = êçç ÷÷ − g ú ⋅ 1 = − 9.81
ëè 13R ø s û 13R s

m feet
a max = 0.65 2
≈ 2.13 2
s s

12-
12-7
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

v2
a=
equilibrium appears as a= 0. The resultant: R and

v 2 s 11.8 ⋅ V 2
hequilibrium = ⋅ =
R g R

At hequilibrium, the effects of centrifugal force are nullified when the vehicle is traveling at
equilibrium speed. The formula above, of course, is derived for metric units, but is
obviously determined in the same manner with the English units used in North
America.

12.2.4 Elevation Transition

The transition from a segment of track with no crosslevel to another segment with full
cant (superelevation) is performed with the help of the transition ramp. The
superelevation at the beginning of the ramp is h = 0, but at the ending point it has the
full value of the full body of the curve. Transition ramps for the increase in elevation,
for high-speed purposes, are more important than the transition horizontal alignment
curve (spiral) running from tangent track to the full body of the curve. They directly
affect the safety and comfort of the ride. Most European railway companies (like
North American railways) make the superelevation possible by raising the elevation of
the outer rail in the curve, while the inner rail stays at its original position and height, as
shown in Figure 12-3. In some rare cases, the superelevation is attained in the manner
as shown in Figure 12-4. In these situations, the outer track is raised half of the desired
superelevation amount while the inner rail is lowered. This method was used in the
construction of the Japanese Tokaido Railway and also in the Swiss railway network
where the extensive costs associated with increasing overhead clearance in tunnels
justifies the cost of doing such.

Figure 12-3 – Courtesy of Miodrag Budisa Figure 12-4 – Courtesy of Miodrag Budisa

The same procedure is applied for superelevation in reverse curves without the tangent
segment between the curves. The type of coupling system used in Europe will permit
the degree of rotation that coupled cars would undergo in this situation. The advantage
of this type of ramp is the fact that the gravity of the vehicle stays in the central axis of
the track; thus allowing the passengers a smooth transition into the following curve.

12-
12-8
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

The reason why this type of ramp is not used often is because it is very hard to
maintain.

The vertical ramp can be linear (as in North

America), curved or sinusoidal.

In the case of linear ramps, the superelevation

increase is directly proportional to its location
along the ramp as shown in Figure 12.5. Figure 12-5 – Courtesy of Miodrag Budisa

The slope of the linear ramp equals:

h
1: n =
1000 ⋅ l (length)

Therefore, the length of the transition curve must be:

n
l= ⋅h
1000

The length and slope of the transition ramp for the superelevation must provide for
the required stability of the train on the ramp. The ramp characteristics must:

• Permit adequate stability so that the train does not derail.

• Remain within an acceptable range for rate of elevation increase.

• Remain within acceptable limits for applied lateral shock (jerk).

The slope of the ramp is dependent on the projected speed to be operated. Steep
ramps are not allowed because of the danger of the wheel flange climbing the rail and
causing the vehicle to de-rail. Within Europe, the limit for the rate of inclination in
slope is around 1:400 (read 1-meter rise in 400-meter run).

The correct value of the ramp slope “n” is calculated by considering the minimum
wheel flange height being 25mm (1”) and the truck spacing being 10m. (Note: In Europe,
the bogies (trucks) are set much further in from the end of the car than in North America.) For the
wheel not to climb the rail, the maximum permissible slope of the ramp would be:

25 1
n= =
10000 400

In order to prevent the vehicle from rolling over when traversing the ramp, the rate of
elevation increase should be within the given formula:

12-
12-9
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

dh
v= = n ⋅V
dt

The maximum permissible rate of elevation increase will vary with different railway
companies. The SNCF network permits up to 70mm/sec, on the DB (Deutche Bahn)
network up to 35mm/sec and on the JNR (Japanese National Railroad) up to
42mm/sec.

According to Prof. A. Prud’homme, a rate of elevation increase of 100mm/sec. is

theoretically possible for high-speed trains, but is not the practice of the European
railways.1

If the rate of elevation increase on the ramp is 28mm/sec, we can get the horizontal
component length of the ramp by:

l = 10 ⋅ V ⋅ h

The value determined is generally rounded to the nearest 5m or so.

For heavy curvature or gradients, the permissible rate of elevation of increase is

increased to 35mm/sec. or:

l = 8 ⋅V ⋅ h

For high-speed rides, Deutche-Bahn AG recommends:

l = 12 ⋅ V ⋅ h

Shock (Jerk)

In Europe, a parameter called shock (jerk) is utilized to determine the configuration of

the ramp. The magnitude of the vertical jerk is defined as a change in the vertical
acceleration within a given time span:

da æ m ö
Ψ= ç ÷
dt è s 3 ø

The maximum jerk value is set at by experience at:

Ψ = 1 .0 m
s3

and the common range of values of the jerk is:

1 Professor A. Prud’homme, “General Revue for French Railroad,” November 1976, Paris, France.

12-
12-10
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

Ψ = 0 .3 − 0 .5 m
s3

The curved ramp is the second type of ramp configuration used in Europe to
transition in superelevation and is shown in Figure 12.7.

The slope of this ramp varies along the

curve and the greatest slope is located in the
middle of the ramp, at point S in Figure
12.6. It follows:

2h
1 : ns = Figure 12-6 JZ (Yugoslavian) Superstructure Regulations
1000 ⋅ l

The curved ramp (spiral) is longer than the linear ramp and thus, it is more suitable for
higher speeds. The curved ramp is used for speeds up to V=180 km/h (112 mph) and
its length is defined by:

Lcurve = 1.41 ⋅ Lstraight

where Lstraight is the length of the associated linear ramp (i.e., about 41% longer)

The greater the rate in acceleration change, the longer the required ramp must be. In
the sinusoidal ramps (utilized in the Tokaido Railroad), the length of ramp is calculated
by:

Lsin = 1.60 ⋅ Lstraight

In the case of the curved ramps, the resultant curve and the slope increase in the form
of two squared parabolas touching, but whose slope constantly changes.

In the case of sinusoidal ramp, the largest curve is at the beginning and at the end of
the ramp, while the vertical velocity is continual.

Sinusoidal Ramp

Sinusoidal ramps are longer in length than the other forms of transition ramps and
allow higher speeds. For

V ≤ 180km / h

max .n s = 4V and n s = 400

12-
12-11
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

2 ⋅ ns h
l=
according to the equation: 1000

we can get the shortest length of the curved ramp.

8 ⋅V ⋅ h
min .l =
1000 and min .l = 0.8 ⋅ h

V2
l =8
If we insert the value of superelevation R

V2
8V ⋅ 8
l= R
we get: 1000

0.064V 3
l=
R for V > 100km / h (62 mph)

V 2 6.4 ⋅ V 2
l = 0.8 ⋅ 8 =
R R for V ≤ 100km / h

If we use the minimum value for the superelevation in the upper equation

V2
hmin = 11.8 − 100
R

we get the shortest length of the curved ramp.

0.0944 ⋅ V 3
− 0.8 ⋅ V
R for V > 100km / h

l min =

9.44 ⋅ V 2
− 80
R for V < 100km / h

The higher value is the value chosen.

In Europe, the minimum length of curved track on which superelevation is imposed is:

12-
12-12
©2003 AREMA®
 C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

G = V/2 ≥ 30m (98.4 feet)

which is again based on the wider truck spacing utilized.

In a compound curve, each individual curve making up the complete compound curve,
has a different radius and different superelevations h1 and h2. The transition ramp will
be made on the length of the transition (spiral) curve between the circular curve
components, where:

l = n(h2 − h1 )

When there is no transitional curve between the two circular curves, the transitional
ramp will be made in the curve with a longer radius, as shown below in Figure 12.7.

Figure 12-7 JZ (Yugoslavian) Superstructure Regulations Figure 12-8 JZ (Yugoslavian) Superstructure Regulations

The crosslevel of the two intersecting ramps in a sinusoidal ramp must be equal for the
both ramps. See Figure 12-8.

12.2.5 Track Warp

The change of crosslevel over the given length
of the transitional ramp or the transitional
curve is defined as track warp. The formula
is:

dh æ mm ö
warp = ç ÷
∆l è m ø

When the car truck (bogie) moves over the Figure 12-9 Track Warp – JZ (Yugoslavian)
Superstructure Regulations

12-
12-13
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

transition ramp, all points of contact are not positioned in the same plane, so the
applied loads are not equally loaded on all four wheels (See Figure 12-9).

Therefore, the slope of the ramp (mm/m) should be as low as possible and within the
allowed limits. This becomes more critical with a speed increase.

Excessive warp, combined with speed, has a significant influence on the dynamic track
stability of the railway and reduces the level of safety provided.

On the railway Paris-Lion the warp is limited to a 0.6 mm/m.

12.2.6 Horizontal Transition Curves

Where a circular curve joins directly onto a piece of straight track, there is at the
tangent point, the condition that at one moment no superelevation is wanted and that
the next the full superelevation appropriate to the curve is required.

If the alignment is to remain unaltered, one of the following conditions must exist:
a. Run the superelevation gradually off
onto the tangent so as to ensure the full
superelevation through the body of the
curve;
b. Begin increasing the elevation at the PC
and attain full superelevation out in the
body of the curve; or
c. Attain and run out the superelevation
partly on the tangent and partly in the
curve body.

These alternatives are all bad both in theory

and practice. (See Figure 12-10)

Figure 12-10 - Courtesy of Miodrag Budisa

In case “a,” the weight of the train gradually becomes shifted to the inside rail while still
on the tangent track, and then is suddenly thrown back to where the superelevation is
at the maximum (the beginning of the curve). The sudden shock produced translates
directly as a rolling action to the locomotive car body and cars, thereby creating track
anomalies.

12-
12-14
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

In case “b” the centrifugal forces are unbalanced at the entrance to the curve and the
leading outside wheels of the vehicles have to bear the brunt of the impact. Even if the
wheel flange did not climb up on the ball of the rail, there would be excessive gage
wear, and not to mention uncomfortable traveling until the proper amount of
superelevation was reached.

In case “c,” there is a combination of the above faults with the same shock at the
tangent point and conditions conducive to rolling, excessive wear and discomfort.

The one method of getting over these serious difficulties is to insert a horizontal
transition between the tangent and the circular cure. The radius of the transition curve
at the full body of the curve is the radius of the full body curve. At the other end of
the circular unit (point of tangency), the radius becomes infinitely great. This also has
the effect of gradually increasing the curvature to the full amount of the curve, instead
of the sudden application of full curvature.

In this type of horizontal curve, the offsets from the tangent at any point on the
tangent, increase in proportion to the cube of their distance from the point of origin.

Such a parabola plotted with

ten offsets is shown in Figure
12-11.

For convenience, the offset at

the end has been divided into
1000, and where this offset is
fixed in value, the method of
obtaining the offset at any point
is by proportion:

Offset at 5 =
4 ⋅ 125 1
= = 0 .5 m
1000 2 Figure 12-11 German Superstructure Regulations

Similarly, the radius of the curve at Sta. 4 is half the radius of the curve at Sta. 2. At Sta.
6, it is one-third of the radius at Sta. 2, and so on. A transition curve, such as this,
cannot be inserted between an existing tangent track and an existing circular curve
without shifting inward the circular
curve to a position parallel to its
original alignment.

In Europe, this arrangement is

commonly known as the “platelayer
transition” (North American cubic
parabola spiral) and it will be seen
from Figure 12-12 that the easing of
Figure 12-12 German Superstructure Regulations

12-
12-15
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

the curve inward has led to an overall increase of curvature in the circular curve.

Using basic spiral theory, the proper

arrangement of a transition curve is
shown in Figure 12-13. The
original circular curve TC is
tangential to TZ. The curve is
shifted to ZY, and TZ is the
amount of the “shift.” The
transition curve ANY bisects the
Figure 12-13 JZ (Yugoslavian) Superstructure Regulations
“shift” TZ at N.

The total length of the transition curve (measured along the tangent) is AB, and AT is
equal to TB.

It is important to remember that any transition is better than none.

12.2.7 Theory of the Transitional Curves

The sudden change of movement occurs when the vehicle enters the curve. That is, a
lateral velocity change occurs. The flange of the outside wheel strikes the outside gage
corner of the rail because of the centrifugal force applied. This causes a tilting of the
vehicle to the high side. In order to eliminate these undesirable movements, the
transitional curves are implemented between the location where R = ∞ and the
centrifugal force is c = 0 and the location in the curve for which those parameters
have certain (given) values. Transitional curves can then provide a smooth change of
movement and direction.
The transitional curve has to fulfill certain requirements:

a) The increase in superelevation should be linear. That is, the radius, which changes
from ρ = ∞ to ρ = R at the point of full body of the circular curve, must match
the required superelevation at every point of the transitional curve. The car body
tilt progresses linearly with movement through the transition curve.

b) The superelevation at each point should match its calculated amount.

l l
c) The curve should be continual from = 0 to = Z , respectively transitioning
R R
m ⋅V 2
from tangent with ρ = ∞ and c=0 to circular curve with ρ = R and c = .
R

12-
12-16
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

The cubic parabola, the fixed cubic parabola and “clothoidal” are suitable for these
conditions.

The simple cubic parabola has been

discussed in this Manual in Chapters 3 and
6.

The fixed parabola has been used in

Europe for transitional curves in lengths
up to:

L = 4 0.64 ⋅ R 3

The cubic parabola is shown in Figure 12-

14.

For transitions that are longer in size, the

Hefer Cubic Parabola is used in Europe.
The equation is:

3
é æ l ö2 ù Figure 12-14 – Courtesy of Miodrag Budisa
ê1 + ç ÷ ú
êë è 2 R ø úû
y= ⋅ x3
6⋅ R⋅l

This formula gives better values, but its accuracy is not 100%.

The Clothoidal transition curve has been used for high speeds because all necessary
requirements are met.

In the Clothoidal, the transition and circular curves have a joint tangent at the end of
the transitional curve/point of the circular curve.

The Clothoidal equation is:

R ⋅ L = const. and

R ⋅ L = A 2 where A is a Clothoidal parameter.

The Quadric parabola is used for a higher speed with a curved transitional ramp. See
Figure 12-15.

The Sinusoidal transitional curve is shown in Figure 12-16.

12-
12-17
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

Figure 12-15 - Courtesy of University of Sarajevo Figure 12-16 – Courtesy of University of Sarajevo

Figure 12-17 shows lateral shock (jerk)

æmö
change Ψ ç 3 ÷ in transitional curves,
ès ø
such as:

a) Cubic Parabola

b) Quadric Parabola

c) Sinusoidal Curve

Figure 12- 17 – Courtesy of University of Sarajevo

12-
12-18
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

The lateral shock (jerk) is defined as a change of velocity in the given time:

da da
Ψ= =V ⋅
dt dl

12.3 Gradient Change

Consideration is given to variations in design level for run-off, vertical curves and
change in gradient, specifically the deviation from these parameters.

Vertical curve is provided when the change in two neighboring gradients is larger than
2‰ (promils). The radius ρ of such a curve is calculated by the equation:

V2
ρ=
2

and in extreme situations, it is permissible to use:

V2
ρ=
4

The minimum acceptable radius of curvature for a vertical curve in Europe is provided
by ρ = 2000m (6,567 feet).

In Europe, vertical grade changes cannot be located within turnouts and crossings,
transition run-offs, spiral curves and on open deck bridges. The beginning and ending
of the vertical curve must not be located closer than 5m from the above objects.

The tangent length of the vertical curve is given by:

ρ i1 ± i2
T = ⋅
2 1000

where ι1 and ι2 are measured in "promil," which is equivalent to a rise of 1m/1000m of

run.

A + sign is used whenever the track grade is ascending and a – sign is used when the
grade is descending.

The tangent distance T for the vertical curve produced from a descending grade
running into level track is given by the equation:

12-
12-19
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

ρ i
T = ⋅
2 1000

Ordinates of the vertical curve are given by:

x2
y= as shown on Figure 12-18.
2ρ

Figure 12-18 – Courtesy of Miodrag Budisa

12-
12-20
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

12.4 Turnouts and Turnout Design

Turnouts commonly used throughout Europe include the following:

§ Lateral - Provide connection of two tracks (See Figure 12-19)

§ Single

§ Complex

§ Three-way

§ Three-way Curved

§ Double turnouts – Provide connection for three tracks (See Figure 12-20)

§ Slip turnouts – Single and double slip turnouts for track crossings under
different angles. (See Figure 12-21)

Lateral Turnouts

Figure 12-19 JZ (Yugoslavian) Superstructure Regulations

12-
12-21
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

Double Turnouts

Figure 12-20 JZ (Yugoslavian) Superstructure Regulations

Slip Turnouts

Figure 12-21 JZ (Yugoslavian) Superstructure Regulations

Lateral single turnouts provide divergence from the main straight track to a sidetrack.
Lateral turnouts can be right or left, depending on the diverging direction. Frogs can
be the typical straight flangeway for both straight and diverging moves or may have the
diverging flangeway curved. Some frogs actually carry the closure rail curvature all the
way through the frog. These curved frogs are typically used in light rail situations.

12-
12-22
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

Some single turnouts with curved points have a curve between P.S. and P.F. with a
reverse direction for the diverging track. In this case, a movable point frog may be
used in order to reduce required ROW for connection to multiple tracks, such as
where parallel tracks are located on a ladder. The advantage of this is a bigger angle
between the ladder and the main track.

For situations where the turnout is located within the curve, the P.S. and frog have a
special design. The flangeway of the frog in the main track direction is in a curve. The
diverging track curves can be concave or convex to the main track.

Double turnouts are designed to connect two tracks laying on one or two sides. The
advantage of using double turnouts in stations and yards is that it significantly reduces
the length of the interlocking.

12.4.1 Measuring the Frog Angle

If at x feet from the gage line intersection, a spread of y feet is found, the unit angle is
x
.
y

x
If = N , - the angle is 1-in-N.
y

There are 3 methods of measuring the frog

angle, which are known as:

1. The centerline measure.

2. The right angle measure.

3. The isosceles measure.

Figure 12-22 shows the measurement

principles.

The unit angle of any frog measured by Figure 12-22 - Courtesy of Miodrag Budisa
the three methods is not the same. Thus,
a frog with a unit angle, which by the central line method is 1 in 8, is a unit angle of 1 in
7.969 by the right angle method, and 1 in 8.016 by the isosceles method. In Europe
(like North America), the centerline measure is the most common method used to
measure the frog angle.

12-
12-23
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

12.4.2 Turnout Calculations

Conventional European turnout calculations are based on a geometric projection of
the length and the relative direction at the points where the connections are made.
This method assumes that kinematic forces rather than dynamic forces determine
vehicle response.2 To determine the essential equations, one uses one of the projection
methods.

The open polygon method option is shown on Figure 12-23 and the closed polygon
method is shown on Figure 12-24. Elements of the turnout are projected on a
convenient coordinate system. From the geometric length relations, the equation can
be written as:

The equation representing the open polygon of Figure 12-23 is:

A ′E ′ = a ⋅ cos α + b ⋅ cos β + c ⋅ cos γ − d ⋅ cos δ

Figure 12-23 – Courtesy of University of Zagreb Figure 12-24 – Courtesy of University of Zagreb

The equations representing the closed polygon in Figure 12-24 are:

a ⋅ cos α − b ⋅ cos β − c ⋅ cos γ = 0

a ⋅ sin α + b ⋅ sin β − c ⋅ sin γ = 0

On the European Railway network, the tangent of an angle rounded off to the nearest
degree expresses the diverging angle.

2 Mejgyeri, J., Geometric Movement in Turnout Development, Austrian Railway Archive Volume 40, Pages

59–65, 1985.

12-
12-24
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

For example:

(
1 : n = 1 : 10 ∴ α = 50 42'38" or)
(
1 : n = 1 : 14 ∴ α = 4005'08" )
The maximum permissible lateral acceleration is 0.8m/sec2. Thus, one could develop a
curve in the closure rail with a radius that, for a desired diverging speed, generated a
lateral acceleration of a = V2/R, not exceeding 0.8m/sec2. This curve would be
independent of the frog number utilized. In effect, Europeans will utilize a variety of
turnout closure rail curves with the same movable point frog in order to secure
different diverging speeds. The resultant turnout, obviously lengthens significantly as
the radius of the curve grows. This is in direct opposition to the turnout practices of
North American railways, that utilize one given closure rail curve for a given turnout
number.

The effective radius is performed by calculating the offset at the switch heel, based on a
chord (12.2m or 40’) centered about the switch heel. The calculated offset will provide
the effective radius ( R ) at the switch heel by using:

C2
R=
8 ⋅V

On the JZ (Jugoslavian Railroad), the chord length of 12.2m is the shortest distance
between wheel set centers (truck centers currently in service).

Thus, for example, in high-speed operations, the diverging angle can be:

α = 2°29'22"∴ (tan α = 1 : 23.5)

and the permissible diverging radius is R=1390m (4,560 ft.). This permits a diverging
speed of 80 km/h in 120 km/H territory.

For 100km/h in a diverging route, we need a curve with a radius of 1500m.

As determined by the following equation:

Vmax = 2.91 R1/2

This radius is coupled with a tan α = 1:18.5 or diverging angle α = 3º05’38.4”.

For Turnout Model EW (1:40.154) (DB Railroad), α=1º12’ 7.5” (tan α=1:40.154)
with a radius of 6100m (20,013 ft), which allows a diverging speed of 200 km/h.

12-
12-25
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

Recognize that although the lateral acceleration component will safely permit diverging
speeds of this magnitude, European railways operate at slower diverging speeds
because of maintenance issues.

Switch Blade (Point) Calculation

For a straight switch blade, the

diverging switch angle is
composed of two angles η and
δ . For a standard straight switch
point angle η = angle δ ,
δ is the angle at the heel of the
switch blade. The clearance at the
heel of switch between the gage Figure 12-25 JZ (Yugoslavian) Superstructure Regulations
side of stock rail and the field side
of the point is approximately 58
mm, and e = e0 + g where
e0 = width of the heel spread,
g = width of the rail head.

e
Equation: sin η = sin δ = ,
p
where p= blade length. See
Figure 12-26.

For a curved switch blade, the

amount of theoretical overlapping
of the blade, where if it were
extended past the point of switch,
would be in the range of u=7-
20 mm, where u equals the
amount of theoretical switch
blade overlapping. (See Figure 12-27) Figure 12-26 JZ (Yugoslavian) Superstructure Regulations

γ
p = SK = Rπ ⋅
1800

where SK is the arc formed.

p 1800
γ = ⋅
R π

12-
12-26
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

æ γö
e = t sin çη + ÷
è 2ø

γ
t = SK = 2 R sin
2

u = R(1 − cosη )

Angle η can be a maximum of: η = 1 0 30 '

Angle δ = η + γ

And if angle η is known, we can calculate “e”.

Figure 12-27 Wheel Movement Through a Conventional Turnout – Courtesy of Weichenwork Brandenburg GmbH

12.4.3 Clothoidal Turnout

For high-speed passenger operations, Europeans utilize the Clothoidal turnout with an
allowable diverging speed up to 230 km/h (143mph). This type of turnout minimizes
lateral acceleration and jerk at key turnout locations. This compliments car body
features such as automatic tilting to maximize speeds through the curve. The following
figures illustrate the use of this type of turnout.

12-
12-27
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

Figure 12-28 Berlin – Halle (Leipzig Line), 300km/H (200km/H diverging) - Courtesy of Weichenwork Brandenburg GmbH

The frog number for this turnout is a #40.154, which corresponds to an angle of
1° 12’ 7.5”. Note the magnitude of the lateral acceleration as one moves through the
turnout. Similarly, note the locations of greatest jerk. The sinusoidal curvature of the
blade (points) minimizes the impact felt at the point of switch as compared to the very
large impact felt in conventional North American turnouts. Figure 12-29 illustrates the
chlotoidal switch point. Figures 12-30, 12-31 and 12-32 show the blade (point) shape
along the switch point.

Kinematic Gauge Optimization is an advanced method to further reduce lateral forces

when a rail car enters a turnout. This design takes advantage of the use of the rolling
radius difference between the two wheels mounted on an axle when the truck enters a
turnout by bending a specially designed stock rail outward. This increases the rolling
radius difference and guides the axle movement in the desired direction.

12-
12-28
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

Figure 12- 29 – Courtesy of JEZ Sistemas Ferroviorios, Renfe, Spain

Courtesy of JEZ Sistemas Ferroviorios, Renfe

Figure 12-30 Section AA Figure 12-31 Section BB Figure 12-32 Section CC

Figure 12-33 illustrates the action of the wheel as it moves through the clothoidal
switch.

Figure 12-33 Movement Through the Clothoidal Switch - Courtesy of Weichenwork Brandenburg GmbH

12-
12-29
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

12.5 Speed Raising Improvements

There are several alternatives if one desires to increase the speed on an existing railway,
for example, with the improvement of transition ramps and transition (spiral) curves.

In Method 1, one might increase superelevation through the curve and lengthen the
transition ramp (spiral). The minimum superelevation required is derived from the
following formula:

V2
min. h = 11.8 ⋅ − 100
R
Using the value for the highest allowed superelevation

m inches
h = 150mm and a lateral acceleration of: a = 0.65 2
∴ 25.59
s s2

Solving for V yields the highest speed permissible for the given curve with a radius R:

V = 4.6 R

For a curve of the radius R = 800m,

the maximum permissible speed is:

V = 4.6 800 = 130 km (81mph)

h

See. Figure 12-34. Note the

similarity with the Vmax equation Figure 12-34 JZ (Yugoslavian) Superstructure Regulations
used in North America for
maximum permissible speed operated over a given curve.

The length of the linear ramp (spiral) is:

0.08V 3 0.08 ⋅ 130 3

l= = = 220m (715 feet)
R 800

and the curve is shifted for:

l2 220 2
f = = = 2.51m (8’ 1-3/4”)
24 R 24 ⋅ 800

In Method 2, if we used the curved transitional ramp (described earlier):

12-
12-30
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

The length of the curving ramp would be:

0.064V 3 0.064 ⋅ 130 3

l= = = 176m (572 feet)
R 800

and the curve is shifted for:

l2 176 2
f = = = 0.81m (2’ 7-1/2”),
48R 48 ⋅ 800

which is 1.70 m (5’ 7”) less than with the straight ramp. If the transition curve is 30m
(98.4 feet), the required offset shift of the curve from the tangent would be:

= 302/24 x 800) = 0.046m (1-7/8”)

So it is necessary to move the curve with the curved ramp by:

81 – 4.6 = 76.4 cm (30”)

This is quite often possible to

do. Thus, we can see the
advantage of applying the
curved ramp and re-aligning the
alignment for high speeds.

Method 2 is very convenient

for long curves and is based on
the following criteria:

The end of the transition ramp

will be constructed in such a Figure 12-35 JZ (Yugoslavian) Superstructure Regulations
way that the elevation run-off
will initiate in the tangent and will not reach the full elevation until it is some distance
into the body of the curve as shown in Figure 12-35.

In this case, the transition curve will have a minimal length, by equation:

ì 0.0944V 3
min .l = í − 0.8V
î R

12-
12-31
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

4.72V 2
− 40
R

for n=8V and l=0.4h, where n = rate of change.

The maximum superelevation deficiency is when the ∆hmax = 100mm or (4”) at the
point of curve, and the ramp height is defined. In this case, the maximum lateral
m
a = 0.65
acceleration sec 2 (a = 2.13 ft/sec2) is only at the point of curve.

In Method 3, we can apply a longer transition curve. In this case, the beginning of the
ramp can be moved into the tangent segment in front of the beginning of the
transition curve.
Only the transition ramp is extended, but the curvature stays the same. On the points
AB, we have the negative lateral acceleration and it is the largest at the beginning of the
transition ramp. At the point B, the acceleration is a=0 and it still rises till the
beginning of the curve is in a positive direction. At that point, there is the maximum
unbalanced superelevation (permissible for European equipment) and it is
∆h = 100mm (4”).

12.5.1 Curve Improvements

A very important parameter of track geometry is the direction of mid-ordinate throw at
each station when lining curves. Direction becomes more important with high-speed
tracks.

It is necessary that the mid-ordinate throws be balanced (i.e., the net result of the
throws be 0), which will lessen the twist of a vehicle. In a perfectly lined curve, the
mid-ordinate readings would be the same at any point along the full body of the curve.
They are measured on a base (string length) of 10m and 20m and in spirals, the mid-
ordinate measured will vary according to its location along the transition curve.

The value of a mid-ordinate is measured as follows:

s2
f =
8⋅ R

where s is the length of the chord used to measure the midordinate.

Permissible deviation in the mid-ordinates measured is as follows:

12-
12-32
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

56 ⋅ s
∆f = for V > 60 km/h and
V

∆f = s for V < 60 km/h

One is allowed to have a difference between adjacent mid-ordinates after curve lining
of:

28 ⋅ s
∆f = for V >60 km/h and
V

s
∆f = for V< 60 km/h
2

At the Japanese Railway Network (JNR), the maximum mid-ordinate is determined by:

400
f max =
V

The curve is in compliance if the mid-ordinate difference is within permitted limits.

Otherwise, the curve is re-lined or a speed reduction is applied, which will permit a
larger ∆f .

12.5.2 Surfacing and Lining

The following methods of surfacing and lining are commonly utilized in Europe:

a) Plus-minus method (Comparable to the North American Bartlett or Bracket

Method)

b) Mechanized 3 points method for Matisa machines

c) Three point’s leveling method and 4 point’s method for Plasser machines.

Figure 12-36 shows the effect of throwing a curve at one point at which the mid-
ordinate is significantly different from those produced by the adjoining chords.

The effect of a throw outward at one point is to increase the mid-ordinate at that point
(by the amount of the throw), and to decrease the mid-ordinate on each side at the
next point by the amount equal to half the throw.

12-
12-33
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

Similarly, a throw inwards is to decrease the mid-ordinate at that point by the amount
of the throw, and to increase the mid-ordinate on each side at the next points by an
amount equal to half the throw.

The Matisa machine

applies the same
methods as above.
“Matisa” automatically
measures the mid-
ordinate and calculates
the necessary correction.
It simultaneously
determines the mid-
ordinate direction (i.e.,
either sharpens or
flattens the curve).

The leveling principle for

the vertical direction is
shown schematically in
Figure 12-37. In this Figure 12-36 JZ (Yugoslavian) Superstructure Regulations
three point measuring
system, point A is located on the already corrected track. The track is to be given an
overall lift equal to the lift height BC. The track at working point D is raised, and the
tie is tamped so that it comes to rest on the straight line AB at E. Leveling can be
expressed mathematically as:

b a
Ζ n (n ) = Ζ n (x − a ) + Ζ o (x + b )
l l

Figure 12-37 – Courtesy of Plasser & Theuer, Theuer, Germany

The machine works in two modes: Normal mode and Automatic mode.

Use of the Four Point Measurement System

12-
12-34
©2003 AREMA®
C H A P T E R 1 2 – E U R O P E A N C U R V E A N D T U R N O U T
M E C H A N I C S

The vertical lining system is based on a 4-point measurement. A computer measures

three ordinates Y1, Y2 and Y4, as seen in Figure 12-38, which determines a square
parabola. At C, where there is a work point, the track is thrown so that it lays on this
curve.

The correct position is

verified by mid-ordinates h
and H and their quotient is a
fixed value. The leading point
of the long chord D follows
the old track geometry
whereas A and B follow the
new geometry which has
been already corrected.

Figure 12-38 – Courtesy of Plasser & Theuer, Theuer, Germany

Figure 12-39 High Speed Turnout, 8 Machines in the

Point, 3 in the Frog, 300km/H (160km/H diverging)
- Courtesy of Weichenwork Brandenburg, GmbH

12-
12-35
©2003 AREMA®