In Australia, a major expansion of the
rail network is underway, with new routes and upgrades to existing
facilities. Much of the investment has been in improving and expanding
city services, including increasing the capacity of existing lines. Most
rail lines are aboveground; however, there is increasing demand for
underground routes in dense, urban areas.
A recently completed underground rail
line, the Chatswood to Epping Rail Link in Sydney, is a 12 km-long twin
tunnel with three new stations connecting two existing stations. Further
expansion has been planned for Sydney, with new lines above and below
ground. Other cities, including Melbourne, Brisbane and Perth, are also
expanding their networks.
Australia uses heavy rail systems in urban areas,
with time tabled train operations and trains suited to suburban
commuting. Metro services have been proposed for the more densely
populated city centers either as completely new systems or by changing
the train car design and operations to increase passenger load and
reduce the transfer time at platforms.
SAFETY ASSURANCE IN THE AUSTRALIAN RAIL INDUSTRY
In Australia, the safety assurance
process is now a fundamental part of rail system design and operations.
This is a risk-based process that covers all design aspects that affect
safety, including signals, track design, occupational health and safety,
security (including terrorist risks) and fire protection engineering.
Rail regulators typically review risks as a whole, so they are keen to
keep fire safety risk slow. Fire protection engineers can expect their
fire strategies to be challenged in detail. Prescriptive codes may give
some direction for fire strategy, but reliance on these alone may not
demonstrate that the required level of safety has been achieved.
The ALARP (As Low as Reasonably
Practical) and related risk processes are generally used to assess
safety. Hazards are identified and risks assessed for design,
construction and operational safety. The risks are ranked using
consequence and probability descriptors to inform the design team of the
key risks. A generic example of a risk ranking matrix is shown in Table
1. Table 2 shows the required actions for the risk rankings.
FIRE SAFETY IN THE AUSTRALIAN RAIL INDUSTRY
There have been few fatal rail fires in
Australia, and none of the same severity as King's Cross, London, 1987.1 There
are several reasons for this, but an important factor is operational
staff acting quickly to prevent the escalation of fires. The challenge
is to maintain an acceptable level of fire safety with changing
operations, greater passenger numbers, and an expanding rail network.
In Australia, buildings must comply with the Building Code of Australia (BCA).2 The
prescriptive provisions do not address stations adequately, particularly
platforms with large travel distances and escalators used for egress,
nor do they address other issues such as electrical hazards and train
For tunnels, a new draft Australian Standard3 was
issued in 2009 for comment. This generally adopts a risk-based
approach, providing high level guidance, but avoiding prescriptive
NFPA 1304 is
often used as guidance for station and tunnel fire safety design in
Australia, although it must be adapted carefully to suit local rail
operations and fire safety industry practice.
Rolling stock fire safety design is
typically addressed through a range of fire safety standards and guides,
including in-house standards, ad hoc tests, and references to other
standards such as BS 6853.5 Some rail authorities are now
updating these standards to take into account developments in materials
testing and fire protection engineering methodologies generally.
The International Fire Engineering Guidelines (IFEG)6 set
out the fire protection engineering process, with some adaptation to
align with the safety assurance process. A risk-based approach is often
adopted for the fire safety issues in rail projects,with codes providing
supporting guidance for design solutions.
THE ROLE OF THE FIRE PROTECTION ENGINEER
Fire protection engineering, as a
discipline, is new to the rail industry. Prior to its involvement,
various disciplines would develop fire safety designs, often in
isolation of others, and not always with specialist fire safety
knowledge. This led to repeating previous designs, without understanding
the background. Code interpretation was often carried out
independently, with parties using different assumptions. Often, the
rolling stock, tunnel and station designs were developed in isolation of
each other. There was coordination of some key design elements, such as
the kinematic envelope for trains, but with less integration for fire
safety. With this approach, some fire hazards may have been left
uncontrolled for some elements or over designed in others.
There has been performance-based design
on infrastructure projects for nearly 15 years; an earlier example is
Sydney Olympic Park,which opened in 1998. The scope for fire protection
engineering has expanded, becoming the means to integrate fire safety
design solutions, as well as solving specific design problems. This has
occurred in parallel with new investment in the rail industry and the
introduction of the safety assurance process, as well as the
introduction of performance-based fire safety design. Many clients see
the value in fire protection engineering to develop a fire strategy
early, setting the parameters for the design of the tunnels, rolling
stock and stations.
DEFINING THE OBJECTIVES FOR FIRE PROTECTION ENGINEERING
An essential part of the process is
defining the fire protection engineering objectives. These include
safeguarding people from injury due to a fire, avoiding fire spread to
adjacent property and facilitating emergency services' activities. A
approach is adopted for all design elements to include consideration to
installation, operation and maintenance risk. Other objectives include
operational continuity and reliability, which can involve maintaining
operations after a relatively small fire, maintaining operations in
degraded mode after a significant fire, and implementing a process to
avoid escalation of a fire. These operational objectives require early
detection and response to incidents, including a robust communication
A risk-based approach has been used for
the selection of fire scenarios, categorising the scenarios as Base Case
Design Fires; High Challenge Design Fires; and Extreme Events.
The categories determine the safety
margin which is to be applied to the results, as shown in Figure 1.7 The fire scenarios can include a combination of fire sizes and occupant numbers.
The identified scenarios are assessed in terms of the Sub-Systems' defined in the IFEG, namely:
- Fire initiation and development and control.
- Smoke development and spread and control.
- Fire spread and impact and control.
- Fire detection, warning and suppression.
- Occupant evacuation and control.
- Fire brigade intervention.
ROLLING STOCK AND OPERATIONS
Rail fire safety is part of a dynamic
system. In a fire or other emergency, the number of passengers on a
train, number of trains in a tunnel and initial response of the rail
operator depend on the system operation. The best place for evacuation
is the station, where there is the capacity for the safe, rapid
evacuation of a train. Evacuation of passengers from tunnels can be
hazardous, even without a fire; therefore, it is better if a train with
afire can keep moving to a station. Until the train arrives,reasonable
conditions in the cars are required by limiting fire and smoke spread.
In Australia, specification of material fire performance varies,
although is typically by reference to BS 6853 with additional local
requirements. Newer international codes, such as EN 45545,8 are beginning to be adopted. The Duggan method9 is
often used as a simplified method to limit the train car interior heat
output. This can over predict the actual maximum fire size, but in the
absence of more detailed analysis, provides a fire size as a common
point of reference for train material specification and smoke
ventilation design. A more detailed analysis of fire growth in existing
rolling stock was conducted by Coles, Wolski and Taylor.10
The resilience of other train systems can improve fire safety, for example:
Fire rated floor to the passenger
compartment, to protect passengers from an under car traction or brake
Two cars to have independent power connection via fire rated (or
fire separated) cables to power transmission units;
Multiple motor units, so that a train
can continue to the next station, even with failure of one or two units.
These features are important in allowing a
train to continue safely to the next station, even with a severe fire
affecting one car.
New rolling stock generally has automatic fire detection
(in car and under car), CCTV to all areas of the car interior, and
various communications systems, such as public address, passenger
information displays, and passenger emergency information systems.
A difficulty often encountered is older
rolling stock that is run through tunnels that have fewer fire safety
features and less robust power and communications systems. Older trains,
or trains not specified for underground use, may not have the fire
safety features described above. In these cases, it is less certain that
a train could reach the next station, or that reasonable tenability
could be maintained in the passenger cars, while travelling to a
On some systems, reviews have been conducted on whether staffing
numbers can be reduced, including operating the trains without an
attendant or driver. In these cases, emergency egress response needs to
be assisted or directed by a remote operator in a central control room.
This requires robust communications, with levels of redundancy in the
train, tunnel and station. The fire protection engineer must understand
how these systems interact to be able to develop fire scenarios.
When a fire-affected train travels on to
the next station, a reasonable degree of safety must be provided. The
omission of inter-car doors has been examined to varying degrees by many
operators, who are keen to have an open train car design, in part for
better security. This can lead to greater smoke spread between cars.
This could impact tunnel egress design if there is a significant risk
due to smoke while a train continues travel or due to a train halting in
a tunnel. The rail tunnel systems in Australia are urban, with less
than 6 minutes travel time between stations; therefore, the risk is
ESCAPE VIA THE STATION
There are many factors that affect safe
egress from stations, including smoke control and occupant behavior. A
key design issue is predicting the occupant load numbers in an emergency
and acceptable evacuation times; typically, the methodologies in NFPA
1304 are used. Passenger numbers vary with different
scenarios, and this is addressed in NFPA 130. NFPA 130 requires
calculation of the occupant load with one missed headway; however,
beyond that, there are a variety of assumptions used - such as
crush-loaded trains arriving from both directions, rather than train
loads based on patronage estimates. This can lead to differences in
occupant load estimates. There is also much discussion about the
4-minute and 6-minute target evacuation times. There is often a dogmatic
attitude towards these targets, even with open air stations, despite
NFPA 130 allowing these time periods to be varied.
The occupant loads and fire scenarios can
be categorized into design basis, sensitivity and extreme. The
evacuation time can be better assessed using levels of service, with
maximum time limits for individuals to be held in a queue at restricted
points along the egress route. This is also compared against the
conditions achieved by the smoke control system. This provides a better
means to design and assess evacuation strategies than attempting to
interpret the exact meaning of four clauses in NFPA 130.
One of the challenges with evacuation
analysis has been to understand the effect of stations at depth. Some
stations have been planned up to 40 m deep. This can lead to challenges
on the movement speed of occupants if they need to climb fire stairs or
if using long escalator runs. A reduction in speed is likely, and stairs
are likely to become crowded, possibly crush-loaded, as the flow slows
or people stop to rest at upper levels. There is limited information for
movement speeds during continuous climbs at depth.
Solutions can involve factoring in
reduced flows and providing extended landings to provide resting spaces.
Proposals have been developed for the use of lifts as a large-capacity
system for general evacuation, as an alternative to stairs and
escalators. This is allowed in NFPA 130, and has been adopted for some
station designs internationally; however, in Australia this has only
been used for disabled passengers.
Egress analysis can use spreadsheets using
flow and walking speed data from various sources or computer modeling.
The computer modeling can be used to demonstrate overall flow through
the evacuation routes. It can also be used to visualise the effects of
some design options.
Smoke analysis at platform level can be done by computational fluid dynamics modeling.
The design and the modeling become more
difficult when platform screen doors (PSD) are introduced. In a fire,
there can be confusion, crowding on the platform, and smoke in the train
cars. It can be unclear if all passengers are off the train, and
whether the PSDs can be closed to contain smoke.
The smoke control system design must
allow for the PSDs to be open during the fire. The gaps between the PSDs
and the train are narrow, less than 100 mm typically; therefore, the
over track exhaust (OTE) system will extract little smoke via this gap.
Any smoke from a train is likely to enter the station platform area, and
a larger smoke extraction system is required at a high level over the
Design options are available to make better use of the OTE, such as
openable dampers above the PSDs to draw some smoke from the platform
area into the trackway OTE; however, these can conflict with other
design requirements, for example the damper opening conflicting with
deep beams over the PSD. This issue needs careful attention to provide a
practical design solution.
ESCAPE VIA THE TUNNEL
The probability of a train failing in a
tunnel due to fire is low, although the consequences are high if it were
to involve a major fire involving a car interior.
Walkways and equipment must be kept to a
safe minimum width to avoid increasing the tunnel diameter as part of an
economic design. Evacuation of a crowded train in a tunnel can be slow.
In rail tunnels, a longitudinal smoke
control system can be provided to protect passengers upstream of a fire.
Typically, the downstream condition is taken to be untenable and
ignored. This may be deemed acceptable if the most likely scenario is
considered to be an under car fire. With a fire-rated floor, passengers
would be able to walk over the fire in an upstream direction.
A major fire involving a car interior has
a lower probability; however, usually, there will be people downstream
of a fire. Consideration should be given to this scenario. If a serious
train fire involves a car interior, then occupants are unlikely to be
able to walk past the fire They would need to escape through the
downstream smoke. In Australia, this issue requires assessment as part
of the safety assurance process. This requires a more detailed
assessment than simply using 10 m or 7 m visibility as acceptance
criteria. The effects of visibility on egress need to be assessed, but
ten-ability must be assessed based on the effects of temperatures and
gases using a range of fire sizes. The risk may be determined to be low
through detailed analysis; however, if the risk is significant, then
additional measures are required.
A practical risk-reduction measure is
reducing the spacing of the cross passages, thus reducing the distance
required to travel through smoke. Cross passages provide an alternative
path for passengers to move away from the incident tunnel to the
non-incident tunnel and provide additional means for fire brigade
personnel to reach the incident. Australian Standards do not have
particular requirements for cross passages, although cross passages
every 240 m have been used as adapted from NFPA 130. This tunnel
standard requires a smoke control system for tenable conditions along
the egress route. Typically, this is with a longitudinal smoke control
system. to avoid back layering. A longitudinal smoke control system will
normally generate air flow at less than 2 m/s upstream of the train. It
is not intended to maintain tenable conditions downstream of the fire,
but may dilute smoke sufficiently to protect occupants from smaller fire
With increasing cross passage distances, it becomes more difficult
for the occupants downstream of afire to reach safety and for
firefighters to access, or retreat from, a fire. However, with well
designed rolling stock, there can be reduced probability of the failure
of a train leading to passengers needing to escape downstream of a fire.
In Australia, it is unlikely that a
distance between exits or cross passages greater than 240 to 300 m a
passenger rail tunnel would be accepted unless it can be conclusively
demonstrated that tenable conditions are possible downstream of a fire
up to the maximum size. Smoke temperatures would need to be limited to
enable occupants to travel the distances required without injury. Large
cross passage spacing could lead to a significant change in smoke
control, possibly even a smoke extraction system similar to that now
provided in road tunnels in Australia.
Joe Paveley is with Arup.
Fennel, D., "Investigation into the
King's Cross Underground Fire," Her Majesty's Stationary Office, London,
Building Code of Australia - Class 2 to Class 9 Buildings - Volume
1, Australian Building Codes Board, Fyshwick, ACT, Australia, 2008.
- DR AS 4825, "Tunnel Fire Safety," Standards Australia, Sydney, 2009.
NFPA 130, "Standard for Fixed Guideway
Transit and Passenger Rail Systems," National Fire Protection
Association, Quincy, MA, 2010.
BS 6853, "Code of practice for fire
precautions in the design and construction of passenger carrying
trains," British Standards Institution, London, 1999.
International Fire Engineering
Guidelines, Australian Building Codes Board, Canberra, Australia, 2005.
Johnson, P., Gildersleeve, C. &
Boverman, D., "Design Fires - How Do We Know We Have Them Right?" International Conference - Charting the Course, Society of Fire Safety, Melbourne, Australia, 2009.
EN 45545, "Railway Applications. Fire
Protection on Railway Vehicles," European Committee for Standardization,
- Duggan, G., "Usage of ISO 5660 Data in Railway Standards and Fire Safety Cases,"Fire Hazards, Testing, Materials and Products, Shrewsbury, Shropshire, UK, 1997.
Coles, A., Wolski, A. & Taylor, R.
"Rolling Stock Fire Performance: Heat Release Rates with Advanced
Computer Modeling," Fire Protection of Rolling Stock 2009, VIB Events, London, 2009.