Safety in railway tunnels and selection of tunnel concept

ESReDA 23rd Seminar, November 18-19, 2002, Delft University, Netherlands

Mr. Terje Andersen
Det Norske Veritas (DNV), Veritasv. 1, N-1322 Hvik, Norway
Mr. Brre J. Paaske
Det Norske Veritas (DNV), Ingvald Ystgaardsvei 15, 7496 Trondheim, Norway
Abstract:
A number of serious tunnel accidents have put tunnel safety on the public agenda. Concerns have
been directed towards the safety of both road and rail tunnels. The choice of tunnel concept for double
tracked rail lines has been given much attention.
Two alternative tunnel concepts are discussed in a safety perspective in this paper:
1. Single bored tunnel, i.e two parallel tracks in the same tube with escape ways to open air
through the tunnel portals or through intervening cross cuts.
2. Parallel twin bored tunnels, i.e two parallel tubes with one track in each with intervening
connections between the two tubes equipped with fire doors or smoke traps.
The risk and safety arguments for various concepts are examined and pros and cons for each of the
concepts are discussed.
An investigation of known tunnel and metro fires is used to assess how the choice of tunnel concept
may have influenced the outcome of the accidents. Further, typical results of CFD calculations of
tunnel fires and risk analysis results are presented.
1. Introduction and scope
Several serious accidents in tunnels have recently years put safety in tunnels on the public agenda,
both in Norway and in other countries. Concern has been directed towards both road and rail tunnels,
and the scope is here limited to safety in rail tunnels.
The choice of tunnel concept for double tracked rail lines has been given much attention. Two
alternative tunnel concepts are discussed in a safety perspective:
1. Single bored tunnel, i.e. two parallel tracks in the same tube with escape ways to open air through
the tunnel portals or through intervening cross cuts or specially designed escape ways.
2. Twin bored tunnels, i.e. two parallel tubes with one track in each with intervening connections
between the two tubes equipped with fire doors or smoke traps.
From a safety point of view it has in general been argued that twin tube tunnels are safer than single
tube tunnels, and that new tunnels therefore should be built according to a parallel twin tube concept.
Twin tube tunnels with frequent intervening connections have some obvious safety advantages, but
also some disadvantages, which are not equally obvious. Each bore of two parallel tunnel tubes has in
general a much smaller cross sectional area than a double tracked tunnel. The larger cross-sectional
area is an obvious advantage of double track tunnels.
If each parallel tube in a twin bored concept is built equally large as a double tracked tunnel, two
parallel tubes will be advantageous in any aspects, except economy. Such concepts would be
prohibitively expensive and may cause further environmental impacts due to depositing of removed
material. Such concepts are not discussed in this paper.

The presentation deals with following issues:

-

Discussion of advantages and disadvantages for different concepts

Double track single bore tunnel: Large cross sectional area and long escape way
Twin bore tunnels: Reduced cross sectional area and shorter escape way
Today's status for a selection of countries with examples of different solutions and risk analysis
results
Description of tunnel accidents and how the choice of tunnel concept may have influenced the
outcome of these accidents.

2. Advantages and disadvantages of different concepts

2.1 Single bored tunnels without meeting track or block stations
This is the traditional railway tunnel for less used lines with a single track. The tunnel is normally
without sophisticated technical installations, except necessary technical installations for operating the
trains. During normal traffic there will be one train only in the tunnel, and in general there will be no
signals controlling traffic in the tunnel which may prevent the train from driving out of the tunnel.
The ventilation direction is determined by the traffic direction, as long there is a train in motion in the
tunnel. If a train, of any reason stops inside the tunnel, the air flow will quickly decrease because of
the speed of the train and the ventilation direction may be turned, depending on the topography and
temperature conditions inside and outside the tunnel. A fire may contribute to a change of ventilation
direction.
In Figure 4 is shown the smoke concentration results of a CFD calculation for a stopped train on fire in
a single track tunnel after 4 min exposure of a 10 MW fire. No mechanical ventilation was applied
only the general draft produced by the train movement prior to stopping.
2.2 Single bored tunnels with meeting track and/or block posts
These tunnels are in principle the same as the tunnels described in 2.1, but with traffic controlling
installations, like block posts and/or meeting track with signal installation making it possible to handle
several trains simultaneously in the tunnel. Signals controlling the traffic will be present in such
tunnels, and they may request trains to stop in the tunnel. A meeting track in the tunnel will also
include switches which may be elements of increased risk to the traffic. In case of an accident, several
trains will make the rescue work more complicated, but only one train will be present in each tunnel
section (signal section).
The ventilation direction in the tunnel may be unpredictable if there is possibility for several trains in
different directions in the tunnel. When a train drives into the tunnel, the flow pattern may be changed
rapidly.
2.3 Double track in single bored tunnel
Both tracks are located in the same tube. For tunnels blasted in rock this normally gives the lowest
investment costs. Many factors are contributing and among these are:
- Less work faces during construction
- Less total mass to remove
- Less rock/ground surfaces to seal and secure.

Figure 1: Sketch of double track in single bored tunnel concept

This is the traditional tunnel concept for double tracked rail tunnels blasted in rock in most countries.
In general the cross sectional area for common railways is large (80 115 m2 for new tunnels) with
big air volumes under the tunnel roof. The cross sectional area is normally substantial less for metro
tunnels. The tracks are normally equipped with one or more block stations to handle subsequent trains
and in longer tunnels it is possible to build one or more connections between the tracks within the
tunnel.
Normally there will be more than one train in the tunnel and if an immediate evacuation from a train in
the tunnel is necessary, it is important to regard any possible traffic on the other tracks. The natural
ventilation direction in the tunnel is relatively unpredictable with several trains in different directions
in the tunnel, but the tunnel has large air volumes which normally will give good smoke stratification
during the first phases nearby the fire.
In Figure 4 is shown the smoke concentration results of a CFD calculation for a stopped train on fire in
a double track tunnel after 4 min exposure of a 10 MW fire. No mechanical ventilation was applied
only the general draft produced by the train movement prior to stopping.
2.4 Twin bored tunnels
In this concept there are two parallel tubes, one for each track, with possibility for intervening
connections and escape ways between the tunnel tubes. This tunnel concept is especially suitable for
very long (15 20 km) tunnels, without any possibility for cost effective escape ways to open air. In
full profile bored tunnels, it may also be used in shorter tunnels because the cost picture is different for
such tunnels, and there might be stability problems during construction while boring a full double
track profile in soft soils.
The twin bore tunnel concept may also be used in other tunnels were two separate tracks are required,
but will normally not be a cost effective solution for tunnels which are made by traditional rock
blasting techniques. The ventilation direction in the tunnel tube will be predictable and are given by
the traffic direction, but may change when train stops or be influence from a fire. In longer tunnels one
or more connections between the tracks may often be necessary. Such connections will break the
separation between the tunnel bores unless expensive gate constructions are established.
The cross sectional area in each bore in a double tube tunnel is substantially less than for a double
track tunnel. Smoke and heat accumulation in the individual bores will be as for s single track tunnel
as presented in Figure 4, and smoke accumulation will occur much more rapidly than in a double track
tunnel, especially in head height for walking persons. This may be an element of great importance to
evacuating passengers.

Figure 2: Sketch of twin bored tunnel

2.5 Double bored tunnel with service tunnel
For long high traffic tunnels under mountain massifs/plateaus or sub-sea tunnels it may be difficult to
ensure access to the tunnel for servicing, maintenance or emergency assistance. In such tunnels it can
be relevant to evaluate a separate service tunnel along the whole distance or part of it. The Eurotunnel
from France to Britain is of this type.
A concept with service tunnel is often combined with separate traffic tunnels for each direction, but
there are also examples where service tunnels are combined with a single bore double track traffic
tunnel. The subsea part of the Seikan tunnel in Japan is of this type. A concept with service tunnel has
obviously many advantages regarding safety and rescue and access for inspection, tests and
maintenance of technical equipment. However, such service tunnels also represent a considerable cost
element and are not risk free.
3. Existing status
3.1 Summary of regulations and practice in some countries
In Table 1 below is made a brief summary of regulation and practice in some countries as well as
recommendations from UIC (International Railway Union).
Table 1: Regulations and practice with regard to tunnelling concepts in various countries
Country:

Regulations and practice:

UIC

UIC the International Railway Organisation is promoting standardisation within rail traffic. In
2001 UIC issued a draft for the document Safety in Railway Tunnels Recommendations for
Safety Measures. /8/.
The document identifies and evaluates various safety measures for railway tunnels. Twin bore
tunnels is not generally recommended, but a measure to be evaluated for new tunnels under
certain conditions.

Germany

New tunnels on high speed lines have generally been constructed as single bore double track
tunnel. A new regulation calls for escape and rescue access for every km where the overburden
is less than 60 m.

France

Tunnels on the new LGV Mediterranne, including the Marseille tunnel (7.8 km) are built as
single bore double track tunnels. The LGV network in France in general has few tunnels. Plans
for a new line from Lyon to Turin is based on parallel twin bore single track tunnels.

The Channel Tunnel Rail Link from London to the Channel tunnel involves several full profile
bored tunnels. The tunnel under the eastern part of London and the river Thames is constructed
as parallel twin bore single track tunnels with cross connection every 750 m.

UK

For the shorter North Downs tunnel (3.2 km) a single bore double track concept was chosen.
Italy

Italy is the European country with the largest length of railway tunnels and has an extensive
program for construction of new tunnels. In general they are built as single bore double track
tunnels. Regulations

Switzerland

Selection of tunnel concept is based level of traffic, length, soil condition and economy. Single
bore double track concepts are in general preferred on main lines for economic reasons.
Exception is the cross-Alpine baseline Gotthard and Ltschberg tunnels which are built as
parallel twin bore single track tunnels with frequent cross connections (i.e. each 375m) between
the bores.

Sweden

Regulation of tunnel safety matters in Sweden belongs to the Building Authority and rail
tunnels are dealt with as an ordinary building. The building (tunnel) should be equipped with
exits or cross connections such that the maximum distance to an exit does not exceed 150 200
m. New tunnels under Hallandssen and the city of Malm are built as parallel twin bore single
track tunnels

Denmark

Historically, does not have many rail or metro tunnels. The new subsea tunnels under the Great
Belt and the resund are built as twin bore single track tunnels. The same is the case with the
new Copenhagen Metro. For the bored tunnels (Storeblt and Copenhagen metro), the concept
is not chosen entirely from safety reason as it has other benefits.

Todays practice can be summed up as follows:

-

Different concepts may be used for new tunnels to ensure desired safety level

Double track tunnels is far the most used concept in short and medium long tunnels, sometimes in
addition to requirements for maximum distance between escape ways to open air.

Twin bored tunnels (also with service tunnels) are mainly used in extremely long tunnels where it
is difficult top make escape ways to open air with reasonable distances because of topographic
conditions. Soil conditions and construction method may also benefit twin bored tunnels under
certain conditions, especially for full profile bored tunnels in soft soil or submerged underwater
tunnels.

3.2 Risk analysis results; example and general conclusion

In Figure 3 is shown some of the results of a risk analysis of various tunnel concepts for different
length tunnels for a new railway line in Scandinavia. The absolute risk values should only be taken as
indicative, but the following general conclusion can be drawn:

Twin bore tunnel concept has very marginal safety benefits if any for tunnels up to 10 km
For tunnels above 10 km the safety advantage increases with increasing tunnel length.

4. Accident experience at rail and metro tunnels

A study of known tunnel and metro accidents may be a contribution to evaluate the conditions
connected to the accident and may also gives an indication on which initiatives that considerably could
have reduced the consequences. Data from the accidents are mainly collected from references /1/-/7/.
During the period from 1940 to 2001 we have totally identified 26 serious accidents in rail and metro
tunnels and stations from different references. 14 of the accidents involved fires in passenger trains or

passenger facilities at stations. A list of those fires with information on the event is presented in Table
2. The overview is not complete and there are no uniform criteria for selection of these accidents,
except that they occurred in tunnels or in subsea spaces of metro systems. It is reasonable to believe
that the most serious fires during this period are included. The author has not detailed knowledge
about the actual metro system and tunnel concept for several of the oldest accidents.
4.1 Example of rail passenger vehicle fire incidents in tunnels
The most serious accident occurred in the Armi tunnel, Italy, in 1944 were 400-500 people were killed
because of carbon monoxide poisoning caused by the smoke from the 2 steam locomotives hauling the
train and did not manage to drive through the tunnel. Finally, the train had to reverse but by this time,
most of the passengers had died. From information on the event that happened during WW2 it seems
that the passengers may have transported on open flat-cars. This incident is not a traditionally fire
accident and is not likely to be relevant for tunnels of today, but the combustion intensity and smoke
production in the two steam engines (6-15 MW thermal effect) may have been comparable with what
can be expected in a relatively severe fire in a single passenger car in todays train or metro system.
Among other serious accidents, it is worth to mention the fire in the Metro of Baku in 1995 (289
people killed) and the fire on the cableway to Kitzsteinhorn, Austria, in 2000 (155 people killed). Both
tunnels had relatively small cross sectional area (Kitzsteinhorn 10 m2 and Baku Metro 28 m2). It is
reasons to believe that the narrow cross sectional area of these tunnels have contributed significantly
to the severity of the accidents because most of the people that died did not manage to get out of the
train, or got out very late. In the Baku fire approximately 245 casualties were found in the train and
only approximately 40 were found in the tunnel. In both incidents, there were problems with opening
the doors, but rapid development of the fire and smoke accumulation also made a considerable
contribution to the casualties. Larger cross sectional area may have given better time for evacuation
before the heat and smoke concentration became unbearable. Improved escape ways from the tunnels
would not have reduced the consequences significantly, but may have saved a few persons. Regarding
the accident in the Baku Metro it should be mentioned that most of the people that were killed as a
result from mechanical damages (i.e. they were trampled because of panic onboard the train).
Another serious accident occurred in 1972 in the double tracked Hokuriku tunnel (13,9 km) when a
fire started in a restaurant wagon in a night train. The train stopped halfway in the tunnel to disconnect
the actual wagon, but was not able to drive further from this place. The train carried more than 700
passengers and 30 of these were killed. The tunnel was not sufficiently equipped regarding ventilation
and lightning and this was heavily criticised after this accident.
There are also examples of serious train fires in trains that have stopped inside a tunnel and were the
passengers by themselves have rescued themselves out of the tunnel both in double tracked and twin
bored tunnels.
The accident on the BART-metro in San Francisco in 1979 shows that a twin bored tunnel concept
with frequent intervening cross cuts is no warranty for safety in a case of fire and do not necessarily
lead to sufficient working conditions for the rescue team. This tunnel had a service tunnel in addition
to two single tracked tubes, but still one person from the rescue team was killed and several were
injured in this fire.
5. Brief analysis of accidents
Totally about 1400 people have been killed in the 26 identified accidents of which 1000 in the 14
listed fires and toxification accident (Armi). The remaining accidents have been collisions, tunnel
collapses and station overcrowding. The large majority of the casualties (90%) were found onboard
the train or within station areas. Just a small proportion of the victims were suffocated in the tunnel
outside the train. Also in fire accident most of the people were killed inside the train. Therefore it
seems equally important to ensure the possibility to evacuate the train as to ensure safe escape from
the tunnel.

In 4 of the 14 fire accidents the train stopped in the tunnel due to a technical failure that also started
the fire, or the fire resulted in a technical failure that caused the train to stop. The Baku and the
Kitzstein-horn accidents as well as other severe accidents were of this type. In 2 other fire accidents
(Hirschen-graben & BART) the train were forced to stop in the tunnel due to application of the
emergency brake. After application of the emergency brake the trains were not able to continue out of
the tunnel and the train had to be evacuated inside the tunnel. These two stops in tunnel could
potentially have been avoided if the emergency brake could have been overridden by the driver, or if
passengers were instructed not to use emergency brakes in tunnels.
For these 6 fire accidents with un-wanted stop, the train must have stopped quite arbitrarily along the
tunnel. Hence, to ensure safe evacuation in these cases there must be possible to carry out rapid
evacuation of the train at all locations in the tunnel, and the interval between two cross-connections to
a second tube/escape way must be quite short or the tunnel have a large cross-sectional area.
In 4 scenarios (Eurotunnel, Hamburg, Hokuriko and Simplon) the train was deliberately stopped in the
tunnel or at an underground station to evacuate people, decouple carriages on fire, and/or fight the fire.
Apart from the Hamburg U-bahn incident where the train remained at a station platform, the other
events happened in relatively long tunnels (14 km or longer). In the Eurotunnel fire the train stopped
next to an emergency exit but the concentration of smoke and fire gases following the train made it
difficult to use the exit until a bubble of fresh air at overpressure was injected into the tunnel at fire
through the emergency exit. In Simplon the carriage at fire was decoupled and the diesel powered train
moved out of the tunnel with most of the passengers. In the Hokuriko accident the train were not able
to move after the effort to decouple the carriage at fire, and 30 out of the 700 passengers lost their
lives before the remaining passengers were saved by trains on the neighbouring track.
The remaining 4 accidents include events with train on fire driving through the tunnel without stop
(Salerno 1999), CO-toxification from steam locomotive not able to pass the tunnel (Armi 1944), metro
station fire (Kings Cross 1987) and collision between trains with subsequent fire (Batignolles, 1921).
In two of the events (Eurotunnel 1996, BART 1979) the passengers escaped into parallel tunnels. For
the other accidents it is doubtful whether the tunnel concept allowed such actions.
The tunnel fire accidents with the highest number of fatalities have all occurred in tunnels with narrow
profiles with a single track bore, either on a single track line or as part of a double tube tunnel concept.
6. Conclusion
To conclude the paper, the following statements can be made:

A large cross sectional tunnel area is a substantial safety benefit that should not be neglected in
choice of tunnel concept.
To promote fire safety in tunnels, it is equally important to ensure the possibility to evacuate the
train as it is to ensure safe escape from the tunnel.
To promote fire safety in tunnels, the evacuation concept must also cover for trains which are not
able to stop at a defined location (i.e at a crosscut or escape way) for evacuation.
For long tunnels twin tubes with intervening cross connections should be thoroughly evaluated
against single tube with large cross sectional area in the context of the above conclusions.
A risk based cost benefit evaluation of these concepts, including assessments of mechanical
ventilation systems should typically form the basis for selection of tunnel concept.

7.

References

/1/: Semmens, Peter,

Railway disasters of the World
/2/: Kichenside, Geoffrey
Great Train Disasters The Worlds worst railway accidents
/3/: Hall, Stanley
Hidden Dangers Railway Safety in the Era of Privatisation
/4/: First International Conference on Safety in Road and Rail Tunnels, Basel, Switzerland, 23rd 25th
November 1992
/5/: Second International Conference on Safety in Road and Rail Tunnels, Granada, Spain, 3rd 6th
April 1995.
/6/: Third International Conference on Safety in Road and Rail Tunnels, Nice, France, 9th 11th
March 1998.
/7/: Department of Transport; Investigation into the Kings Cross Underground Fire, HMSO 1988.
/8/: UIC (International Union of Railways). Safety in Railway Tunnels; Recommendations for Safety
Measures. October 2001. Ernst Basler & Partners.

Fatalities per billion passengerkm

6,00

Fatality
risk per
billion
passenger
km

5,00
4,00

Total risik
for car
drivers per
billion
personkm

3,00
2,00
1,00
0,00
Open
line

Single Double Twin

track
track
tube
7 km
7 km
7 km

Single Double Twin

track
track tube
15 km 15 km 15 km

Single Double Twin

track
track tube
24 km 24 km 24 km

Figure 3: Risk results for various tunnel concepts for a new railway line in Scandinavia

Figure 4: Smoke concentration after 4 min around a stopped train calculated by a CFD tool
for a 10 MW fire in a single track tunnel. No mechanical ventilation.

Figure 5: Smoke concentration after 7 min fire around a stopped train calculated by a CFD
tool for a 10 MW fire in a double track tunnel. No mechanical ventilation.

Table 2: Passenger railroad and Metro Tunnel Fires - excerpt from extensive list of 26 accidents (international)
No
Location:
Date:
Accident
Tunnel
Concept:
Fatalities/ Passeng Comments/descriptions:
category:
length:
Injured:
ers
(total):
155
167
The cable car caught fire at the bottom of the tunnel immediately after departure, and came to a
1
Kitzsteinhor 2000,
Fire
3,4 km / Single
bore
fatalities
stop 0.6 km inside the 3.4 km long tunnel. The car light went out, and at first the doors were
n, Austria
Nov.
43o
inclinati tunnel for
impossible to open. After a while they nevertheless managed to open some of the doors. The
11th
on
cable cars
cross section of the tunnel is very narrow (diameter 3.6 m) and the space at the side of the cable
car is very narrow. The steep inclination (43o) made the tunnel a chimney in it self. People
coming out of the train and trying to escape upwards the tunnel, were caught by the smoke and
warm flue gases and died inside the tunnel. Only 12 of the passengers survived the cable
conveyor fire. They managed to escape the train at an early stage by a broken window and
evacuated downwards in the tunnel. 3 of the fatalities had been staying in the station building at
top of the tunnel.
2
Salerno,
1999,
Deliberate About
?
4 fatalities 1100
A fire was started in one car of a 13-car train carrying 1100 football supporters on their way back
Italy
May
ly started
10 km
/ 9 injured
home from a match. The fire was most likely arson or the result of a smoke bomb. The incident
23rd
fire in
took place in a long tunnel, but there are no indications that the train stopped in the tunnel. Thus
train
the tunnel may not be of essential significance to the outcome of the incident. This was the
second football supporter train fire in Italy during a 2 week period.
3
Eurotunnel,
1996,
Fire
50 km
Twin bore 0 / ?
28?
One of the trucks on the train caught fire before the train entered the tunnel. The train was one of
tunnel
UK/France
Nov
the Euroshuttle train for trucks with truck drivers in a separate car next to the locomotive. The
18th
with
train continued at its speed (~120 km/h) for about 10 minutes before it stopped beside an exit to
service
the adjoining service tunnel. I was impossible to disconnect the burning part of the train, and the
tunnel.
power from the voltage conducting wire disappeared relatively quickly. The fire was at that
Cross
moment very strong (max 100 MW) and it rapidly spread to cars nearby. Thick smoke forwarded
connectio
quickly in the train due to other trains running in the current tube. This made the evacuation
ns every
more difficult. Persons in the staff car and the truck drivers managed to evacuate through the
375 m
neighbouring door leading to the parallel service tunnel. Overpressure from the service tunnel
door created a "fresh-air bubble" when opened. Locomotive staff was at a later stage rescued by
a rescue party from the adjoining service tunnel. The fire caused huge structural damages to the
tunnel.
4
Baku Metro, 1995,
Fire due
Metro,
Twin bore 289
About
Each tunnel bore has a relatively narrow cross-section (H=5,6m W=5m). The tunnel is equipped
Aserbadjan; Oct.
to
about
tunnel.
fatalities
1100
with a controllable ventilation system.
between
28th
electrical
2,2 km
Possibly
and 265
A fully loaded 5 car metro train stopped about 200 m after Uldus station due to sparkover/
Uldus and
fault on
between no coninjured.
electric arc in electrical equipment in the rear end of the fourth car. The fire rapidly spread to car
train
stations
nections
Narimanov
About 245
5. Because of problems with the opening of doors in car 4, the passengers were forced to
station.
between
killed in
evacuate through car 3. Many travellers, narrow tunnel cross-section and the door problems lead
the tubes
train, 40
to slow evacuation, and the people panicked. The ventilation direction of the tunnel was changed
in tunnel
during the event and much of the smoke was drawn in the same direction as many of the
passengers escaped. 245 of the casualties were afterwards found inside the remains of the train,
most of them either squeezed or stamped to death. 40 of the casualties were found in the tunnel.
95 % of the persons who managed to evacuate the train survived.

10

Double
tracked
(A=70
m2)

0 / 0?

Metrosystem

Station

31/?

Fire in
train

Metrosystem

0/3

1979,
Jan.
17th

Fire

5,9 km
BARTmetro

Twin bore
tunnel
with
service
tube.

1/58

40
passeng
ers +
firecrew +
+

1972,
Nov

Fire

13,9
km,

Double
tracked

30
fatalities /

> 700

Hirschengraben
tunnel,
Zurich,
Switzerland

1991,
April
16th

Fire

Kings
Cross Tbane,
London, UK

1987,
Nov.
18th

Fire in
escalator

Hamburg
U-Bahn,
Germany

1980,
Apr
8th

BART; San
Francisco,
US

Hokuriku
tunnel,

1,3 km

50 in
train S9
90 in
train S5

A fire was observed in the second last car of a 6-car local train, S9, as it left Zurich Hbf and
drove into the Hirschengrabentunnel. The train stopped almost halfway through the tunnel, as
someone activated the emergency brake. The tunnel has double tracks, and was rapidly filled
with smoke. The escape routes leading to the tunnel gateways were about 500-800m long. It took
2-3 minutes from the train stopped until evacuation was initiated. Before evacuation started the
passengers were encouraged to stay in the train. Nearly at the same time train S9 stopped inside
the tunnel, train S5 entered the tunnel in the opposite direction from Stadelhofen station. Train
S5 was stopped before it reached the accident area. The train driver takes the other driver's room
in use, and after 4 minutes starts to drive slowly back towards Stadelhofen. After 100m he stops
the train to pick up evacuating passengers from train S9. After a while the voltage disappeared
from the overhead contact line leaving the trains stranded in the tunnel. All of the passengers
were requested to evacuate the train and to leave the tunnel on foot. The fire brigade and rescue
personnel were present outside the tunnel gateway 10 minutes after the burning train had come to
a stop. Their duty was first of all to advise and help the last of the evacuating persons. They were
all able to walk, and nobody needed to be carried out. The last passenger came out of the tunnel
about 20 minutes after the burning train had stopped.
Kingss Cross is one of the most frequently used stations in London Transportation's tube
network. The station is operated by 5 different lines in 4 different levels. I addition there is a
distribution hall with ticket offices, placed below street level. From the distribution hall several
escalators descend to the different platforms. The fire started in escalator 4, which serves
Piccadilly Line's platforms. After about 15 minutes moderate fire, suddenly an flashover of the
whole upper part of the escalator shaft and the distribution hall occurred. The length and great
inclination of the escalator shaft contributed to intensify the fire. Trains that continued running
for some time after the fire outbreak, rescued the passengers on the lower platforms.
A fire broke out in a train car in the Hamburg U-bahn. A passenger at the Altona station gave
alarm, and the train driver immediately tried to put out the fire with a fire extinguisher. The
attempt failed, and alarm was given to the control room, which called the fire department. The
fire department was quickly in place, but the platform was totally covered by smoke, so all work
had to be performed with smoke proof breathing equipment. 3 firemen suffered from smoke
injuries from the incident. No passengers were injured. The fire started in a seat, and the
prevailing car was totally over ignited after about 9 minutes.
A fire broke out in a circuit breaker in the 5th and 6th of 7 cars in all. The train was stopped by the
emergency break and could not be restarted. An unsuccessful attempt to disconnect the burning
cars delayed the evacuation and escape of the passengers with about 30 minutes. The tunnel was
filled with smoke despite of induced draught outlet every 300m and activating of the ventilation
system. The passengers were taken to the service tunnel and out in the open air through the other
main tunnel. Smoke was drawn into the service tunnel and the other main tunnel. The fire caused
1 fatality and 58 serious or minor injuries, most of them caused by flue gases or poisonous gases
from combustion of plastics. The death victim was a fireman who died due to flue gas poisoning.
The tunnel system has intervening connections from the main tunnel to the service tunnel every
100 m.
A dining car in the late night train "Kitaguni" caught fire while the train was in the tunnel. The
train driver tried to disconnect the burning car, but a power failure brought the train to a stop in

11

Japan

6th

10

Vranduk
tunnel,
Yugoslavia

1971,
Feb.
14th

11

Simplontunnel,
Switzerland
/Italy

12

13

14

built
1962.

tunnel

many
injured

Fire

1,6 km

Single
track?

34
fatalities

Ca 200

1969,
Nov
8th

Fire

19,8 km

Twin bore

0/0

London, UK
Shepards
Bush/
Holland
Park
Armi
tunnel, Italy

1958,
July
28th

Sparkover
electricarc - fire

Metrosystem

1 fatality/
51 injured

1944,
Mar.
2nd

Carbon
monoxide
poisoning

Ca 1-1,5
km

Single
track?

426
fatalities/
60 injures

Batignollestunnel,
Paris,
France

1921,
Oct.
5th

Collision
Fire

Ca 1 km

Double
tracked

>28
fatalities

2 trains

the tunnel in darkness. The passengers evacuated the train and tried to get away from the fire and
to escape the tunnel. Trains on the other track picked up some of the passengers trying to escape
by foot. Rescue trains were driven in from both sides and most of the passengers were rescued
(30 were killed), but there were many injuries. After the accident, criticism was raised at
inadequate illumination and ventilation.
A Diesel-electric locomotive in a train carrying 200 passengers caught fire in the Vranduk tunnel
between Zeneca and Doboi in Bosnia. The train stopped ca. 300 m from the northern end of the
tunnel. The locomotive staff attempted to put out the fire, but it spread to the cars. Some of the
passengers that jumped from the train and tried to escape from the smoke filled and dark tunnel
were unable to get out.
A fire started in a diesel-powered train immediately after the train had entered the tunnel, most
likely in the machinery in the aft end of the train. The passengers were directed forward in the
train. The train stopped at an emergency station with an emergency telephone. The burning car
was disconnected from the rest of the train and the rest of the train with the passengers could
drive out of the tunnel with only a 14-minute delay. 10 of the passengers however did not follow
the instructions from the conductor, and under guidance from to railroad employees they
proceeded to the tunnel gateway. The light conditions were poor and the passengers were also
obstructed by smoke. A train driving carefully towards them from the opposite direction rescued
these passengers. There were no injuries. The fire extinguishing started after one hour and lasted
for about three hours. The tunnel was re-opened the following day.
A sparkover in the electro-technical equipment caused a powerful electric arc that resulted in a
subsequent fire. Delayed evacuation of the train caused intense smoke exposure of the
passengers during evacuation. 1 person died in hospital after the accident.
Two steam locomotives were hauling a train from Balvano to Potenza and the train experienced
problems in a steep tunnel due to low quality coal. The train remained in the tunnel for a long
time without getting through, and in the end had to back out of the tunnel. Most of the
passengers in the train had suffocated of carbon monoxide poisoning by then. The incident took
place during the allied invasion of Italy in World War II. One source indicates that the travellers
were stowaways on a freight train, while others have indicated that it was a matter of evacuation
of civilians, transportation of troops etc. From the description of the course of events, it seems
possible or probable that open cars were used.
A local train heading for Versailles stopped in the Batignolles tunnel, which is situated right
outside Gare St Lazaire in Paris. Due to a mistake the subsequent train received permission to
drive into the tunnel. The train drove into the rear end of the first train. Some people died in the
collision, but even more died as a result of a subsequent fire caused by discharging gases from
the train's illumination system which caught fire. After this accident all gas illumination on trains
in France was replaced with electrical illumination. There are reasons to believe that the number
of fatalities could be higher than the official number of 28 fatalities.

12