article

Evaluating safety aspects for underground metro networks

Posted: 6 May 2011 | Prof. Dr.-Ing. Alfred Haack, former Executive Board Member at STUVA (Research Association for Underground Transportation Facilities) and Past President of ITA (International Tunnelling and Underground Space Association) | No comments yet

A functioning mass transit system is most important to back up mobility in large cities and conglomerations. This has been recognised for nearly 150 years: the first metro worldwide was inaugurated in London in January 1863. It was a steam driven system. The first electrified trains started in November 1890, again in London. The first metro on the continent followed in 1896 and was installed in Budapest. Paris started its metro in 1900, Berlin in 1902 and Hamburg in 1912.

The size of the route network depends on the population of the city or the area served by the subway system. London operates its underground network with a length of 410km and 270 stations – by far the largest network in Europe (and worldwide), followed by Madrid with 325km, Moscow with 300km and Paris with 215km.

A functioning mass transit system is most important to back up mobility in large cities and conglomerations. This has been recognised for nearly 150 years: the first metro worldwide was inaugurated in London in January 1863. It was a steam driven system. The first electrified trains started in November 1890, again in London. The first metro on the continent followed in 1896 and was installed in Budapest. Paris started its metro in 1900, Berlin in 1902 and Hamburg in 1912. The size of the route network depends on the population of the city or the area served by the subway system. London operates its underground network with a length of 410km and 270 stations – by far the largest network in Europe (and worldwide), followed by Madrid with 325km, Moscow with 300km and Paris with 215km.

A functioning mass transit system is most important to back up mobility in large cities and conglomerations. This has been recognised for nearly 150 years: the first metro worldwide was inaugurated in London in January 1863. It was a steam driven system. The first electrified trains started in November 1890, again in London. The first metro on the continent followed in 1896 and was installed in Budapest. Paris started its metro in 1900, Berlin in 1902 and Hamburg in 1912.

The size of the route network depends on the population of the city or the area served by the subway system. London operates its underground network with a length of 410km and 270 stations – by far the largest network in Europe (and worldwide), followed by Madrid with 325km, Moscow with 300km and Paris with 215km.

In Germany, the largest networks are operated in Berlin with 150km and 173 stations, Munich with 105km and Hamburg with 100km. Significant parts of all these networks are running on surface or viaducts, especially in the outskirts of the cities. In Germany, there are a total of 17 cities operating metros or light-rail rapid transit systems with significant underground sections. Their remarkable contribution to inner urban mobility is enlightened by the average transport capacity taken from the statics of VDV (the Association of German Public Transport Companies1). According to these ascertainments, approximately 21 million people are using the systems on every normal work day. It is obvious treating such high numbers of users asks for highly sophisticated safety concepts, especially considering fire events on a train.

Emergency situations in tunnels and the definition of a standard emergency scenario

Essentially a distinction has to be drawn between an operational breakdown and an emergency. When breakdowns occur there is no danger for passengers, but normal services have been interrupted. They can be triggered by various factors. In such a case, passengers are only affected by e.g. delays in travelling times.

Emergency situations always involve potential danger for passengers, for example during derailments, collisions or fires. Defective operating equipment can also lead to emergency situations should this result in accidents resulting in people being injured or worse or fires occurring. An emergency situation always involves people being endangered. Examples for the causes of emergency situations are:

  • Vehicles: a derailment; a collision with an obstruction; a crash with another train/accident resulting from collision and a vehicle fire
  • Stations: defective operating equipment; cable fires; casualties/suicides; escalator fires and fires in sales outlets and/or service premises.

It is essentially extremely difficult to estimate the effects and possibilities to provide protection in conjunction with attacks carried out by terrorists within public transit tunnels. As attacks of this nature do not represent a specific problem for tunnels, they normally are disregarded.

If one compares a case of emergency involving a fire and a large number of passengers who are no longer capable of saving themselves with an emergency where there is no fire, then the scenario with the fire always has to be assessed as more unfavourable and more critical. During a fire incident, injured people must be able to escape or be rescued in a very short period of time in order to prevent smoke poisoning and burns as much as possible.

Design fire of rolling stock

Experience shows that fires in underground, suburban and urban rolling stock occur extremely rarely. During the last 50 years in Germany for instance, no resultant deaths have been recorded and the number of injured is extremely small2. Fires can be caused, for example, by defects in the vehicle electronic system or by arson. Fires in modern rolling stock are self-extinguishing – except when exposed to an outside source of energy for some length of time or if criminals plant substances designed to accelerate combustion. Once ignition occurs, the further course of the fire depends on the amount of available oxygen, the quantity of combustible materials in the vehicle and their flammability (see Figures 1 and 2).

Public commuter transportation coaches are not all equipped in the same way to resist fire on account of their long service life (in some cases up to around 30 or 40 years). DIN 55106 has applied for preventive fire protection in trackbound vehicles since 1988. All rolling stock built since then must comply with this norm. Older vehicles are refitted during regular maintenance checks (e.g. seats, cables) so that these vehicles, by and large, now comply with DIN 5510.

The spreading of smoke in an underground stop

The hot smoke gases, which rise from the fire seat, largely spread themselves in layers along the tunnel roof or the ceiling of an underground stop, providing that the air flow in the tunnel’s longitudinal direction is low. Existing eddies and back-flows of the hot smoke gases lead to an undefined border area between the upper hot gas layer and the significantly cooler cold gas layer located below it. This cold gas layer is also known as the low smoke layer.

People cannot survive in the hot gas layer without additional protective measures. The low smoke layer (cold gas layer) must possess a sufficient thickness so that people can survive within it. Furthermore, the temperature prevailing there must be acceptable for people (T<50°C), there must be sufficient oxygen (>14 Vol.-%) and the toxic substance concentrations in this layer must not exceed the permissible limit values (e.g. CO < 500 ppm). Furthermore, visibility within this layer amounts to at least approximately 10m so that people trying to escape can orientate themselves and do not panic. This visibility is possible providing the surrounding lighting amounts to around 40 lux and the optical density is approximately 0.13 m-1 at the most. When the fire starts the visibility is far greater and diminishes correspondingly first as the fire progresses.

It is no longer possible for passengers in underground stations to save themselves if the facility contains an excessive amount of smoke. This is the case when the thickness of the low smoke layer (cold gas layer) above platform level is less than 2.5m – or, however, the optical density in the low smoke layer exceeds 0.13 m-1.

The fire services demand that during the self-rescue phase or rescue by a third party an average approximate 2.5m (see Figure 3 on page 50) to approximately 1.5m thick low smoke layer is retained above platform level with suff – icient visibility for a duration of at least 15 minutes or 30 minutes from the fire starting. Through these demands, it is intended to ensure that:

  • Passengers are able to escape during the self-rescue phase without outside help unhampered providing there is sufficient visibility and
  • Those requiring attention from the emergency services during the third-party rescue phase such as e.g. disabled people or those under shock are able to receive sufficient air to breathe and can be rescued by the fire brigade.

Various constructional measures can be under taken to effectively restrict smoke spreading in underground stops (e.g. smoke curtains, smoke removal shafts). In this way the period of time available for the evacuation of the passengers can be substantially extended. Smoke curtains are applied to ensure that, at least for the duration of the self-rescue phase, stairways and distribution levels are protected against smoke. Smoke curtains must allow a clearance height of around 2m for passengers escaping at all entrance/exit areas (see Figures 4 on page 50 and Figure 5).

Hot smoke gases can be transferred systematically via smoke removal shafts in the underground station ceiling to the ground surface (see Figure 4). The shafts can act either naturally through the thermal effect of the hot smoke gases or can be equipped with a mechanical ventilation unit.

Flow technical calculations have to be undertaken to ensure the shafts are designed to best suit the purpose and to establish the most favourable positions for the smoke access openings at any given stop. In addition, it must be observed that the access openings for the smoke removal shafts are kept away from streets as far as possible.

Zone and field computational models are made use of to calculate the smoke numerically. Given the same general conditions, both models for geometrically homogenous and less irregular spaces provide approximately the same results for important parameters such as e.g. the thickness of the low smoke layer and visibility. However, a field model is essential if flow conditions have to be investigated chronologically and locally in complex facilities such as the platform level or distribution level taken together or in the route tunnels.

Evacuating a station

In the selected scenario ‘Vehicle on fire but still reaches the next stop’, it is presumed that the fire breaks out when the vehicle is travelling through the route tunnel and is discovered after 0.5 minutes by a passenger and reported to the driver 1 minute after it started. However, the vehicle reaches the next stop 2 minutes after the fire started thanks to the emergency brake bridge-over. In the assumed scenario, the driver requires a further minute to investigate the fire incident and sends an additional report to the control centre 3 minutes after the fire began. He calls on the passengers to leave the train immediately at the stop.

Generally, a reaction time for the affected passengers must be taken into consideration for evacuation analyses after the alarm is sounded. This can be very different depending on the circumstances. A reaction time of 1 minute was accepted as realistic in conjunction with the assumed vehicle fire. Thus the evacuation of the train and the station starts some 4 minutes after the fire started – including travelling time and investigation time.

Certain people are bound to react more quickly to the alarm being sounded than others. These people can more or less select their walking speed at will as the throng of people en route to the stairways is still relatively low.

Individual pedestrians, who are capable of moving freely, attain walking speeds of between 1 and 1.6 m/s depending on their age and health. Normally, an average walking speed of 1 m/s corresponding with NFPA 1304 is assumed for calculating the evacuation.

The values that apply to walking speeds of individuals on stairs in salient literature are related to the vertical, inclined or horizontal length of the stairway. This must be considered when comparing various walking speeds on stairs. A walking speed related to the vertical height components of the stairs of 0.25 m/s was selected in accordance with NFPA 130 for fixed stairs leading upwards.

The escape-way width (walking track width) on stairs amounts to 60cm. In literature depending on the general conditions (e.g. normal/dangerous situation, persons with/without luggage) very different stair capacities of 19 to 46 persons per minute and per escape-way are given.

A stair capacity for fixed stairs leading upwards of 37 people per minute and escapeway after NFPA 130 seems appropriate for calculating the evacuation times for underground stops.

The following definitions were agreed on for escalators:

  • Generally speaking, escalators in Germany are 1m wide. Nonetheless, it is recomm – ended to select an escape-way corresponding to a width of 60cm for an escape situation per escalator. In other words, the evacuation calculations preclude people overtaking one another on escalators. As a result, safety reserves are available through this definition.
  • It should be assumed that the escalators are switched off in the event of a fire although in practice escalators running upwards should be kept operating as long as possible – as this consequently speeds up evacuation in general and alleviates the procedure for older or disabled persons.
  • For the evacuation calculations, it is to assume that in each case one escalator is out of commission as it can be the case that escalators are not usable because e.g. elements have been removed on account of maintenance work being carried out.

Fixed stairs must possess a width of at least 2m in accordance with BOStrab5. Taking the escapeway width of 60cm into account, such a stairway possesses three parallel escape-ways.

It is possible to make use of recognised methods to determine the evacuation times, as for example:

  • The methods applied in the American Standard NFPA 130
  • Calculation methods according to Russian tests2
  • Special programmes to calculate evacua – tion times in underground transportation facilities such as e.g. SIMULEX, STEPS or ASERI.

On the basis of the established definitions relating to the number of people, walking speeds and stairways, and taking the dimensions of the station into consideration, the walking time on the platform, the overall walking time till the ground surface is reached and all waiting times (e.g. in front of the stairways) are determined for the person, who is the last to leave the platform. Towards this end, it is assumed that the escaping passengers distribute themselves more or less uniformly over the stairways in keeping with the available stair capacities (so-called hydraulic flow-off principle). The total walking time and all waiting times are added together for the most unfavourable escape route to establish the decisive evacuation time.

Comparison of evacuation and smoke times

The smoke times are compared with the evacuation times above the various stairways for each particular station under consideration. Generally speaking, the evacuation times must always be shorter than the smoke times. The previously cited fire service requirements must also be observed. In the end, the result can be:

  • The time span allotted for self-rescue is too short unless constructional fire protection measures are undertaken. If so, additional measures have to be carried out.
  • In many cases, smoke curtains at the stairways as the sole additional measure do not suffice in order to safeguard the conditions for self-rescue of passengers. The smoke time, however, is substantially extended through the smoke curtains.
  • Non-endangered self-rescue of passengers is possible if an adequately dimensioned smoke removal shaft for natural extraction or for mechanical extraction in complicated cases is provided in addition to the smoke curtains.

Summary

The objective must be to attain short evacuation times and as lengthy smoke times as possible through the constructional measures. Towards this end, the two following methods are available with regard to planning underground stations:

1. The smoke time can be extended by increasing the clear height of the stop (larger smoke storage volume), setting up smoke curtains and smoke removal shafts as well as by the application of mechanical smoke extraction units, should space for the stairways be restricted so that it is not possible to shorten the evacuation time.

2. The evacuation time can be shortened by increasing stair capacities (e.g. more or wider stairs) if for example there is not sufficient space for large smoke removal shafts (e.g. for shaft openings on the surface) to extend the smoke time.

Planners and safety experts must work together closely from the very beginning of the planning stage. It is their responsibility to determine the number of persons for instance, for a particular stop, which should be taken into account for the evacuation calculation. In addition, the fire behaviour of the vehicle (e.g. smoke release rate), walking speeds, stair capacities, escapeway lengths etc. should be established on the basis of the prevailing local conditions.

References

1. Annual Statistics of the Association of German Public Transport Companies, Cologne; edition 2009

2. Haack, A./Schreyer, J.: Emergency Scenarios for Tunnels and Underground Stations in Public Transport; 4th International Symposium “Tunnel Safety & Security”, 17.-19. March 2010, Frankfurt/Main, Germany

3. Bemessungsbrände für S-Bahnen und den gemischten Reisezugverkehr – Anwenderhandbuch; Bericht der STUVA, Juni 2010 (Design Fires for S-Bahnen and the mixed passenger transport – manual; report prepared by STUVA, June 2010)

4. NFPA 130: Standard for Fixed Guideway Transit and Passenger Rail Systems, edition 2010, National Fire Protection Association, Quincy, USA

5. Bundesministerium für Verkehr: Richtlinien für den Bau von Tunneln nach der Verordnung über den Bau und Betrieb der Straßenbahnen (BOStrab) (BOStrab- Tunnelbaurichtlinien); edition April 1991; edited by Verkehrsblatt 45 (1991), issue 10, pp 464-469 (Ministry of Transport: Guidelines for the construction of tunnels according to the standards of construction and operation of trams)

6. DIN 5510. Vorbeugender Brandschutz in Schienen – fahrzeugen; 1988 (Preventive fire protection for railway coaches; 1988)

About the Author

Prof. Dr.-Ing. Alfred Haack is a civil engineer and studied at the Technical Universities of Hannover and Berlin. He obtained his PHD degree for research on fire protecting during tunnelling in 1971. He has been working with STUVA for more than 40 years in widely varying fields of tunnelling and micro tunnelling such as drainage methods, waterproofing techniques, various lining systems, occupational health and safety, fire prevention and fire fighting. Since 1995 until 2007 he has been an Executive Board Member of STUVA and, since 1996, an honorary professor in the field of civil engineering at the Technical University of Braunschweig. Between 1995 and 2004, Prof. Haack was member of the Executive Council of ITA – International Tunnelling and Underground Space Association and between 1998 and 2001 the ITA’s President. Since 2008 he is retired and works as a free consultant for STUVA.