Ninth Intermarional IBPSA Conference iBuildingSimulation Montreal Canada 2005 August 15-18 2005 NUMERICALSIMULATION OFFIRESPREADINTERMINAL 2OFBELGRADE AIRPORT Z. Stevanovic' T. Valentina' N. Kadic' Z. Markovic' M. Kadic and D. Mumovic² Institute of Nuclear Sciences Vinca University of Belgrade PO Box 522 11001 Belgrade Serbia and Montenegro 2The Bartlett School of Graduate Studies University College London 1-19 Torrington Place London WC1E 6BT England UK ABSTRACT new fire engineering approach (BSI 2003a) is more Increasingly the fire protection concerms at airports holistic performance based and provides more are being addressed using putational fluid fundamental and economical solution. Whilst the dynamics and most of these studies have been protection of life is the main objective of the fire based on the prescriptive remendations on safety regulations the financial impact of fire on a design of an airpont terminal building given by business as a result of direct property damage or appropriate standards. However this study is based lost productionmight alsobe important on the performance based principles of so called considerations. The recently adopted standard (BSI Qualitative Design Review (QDR) introduced by 2003b) introduces so called Qualitative Design recently adopted fire safety standard (BS 7974: Review (QDR) a structured technique that allows 2003-2004). Apart from the evaluation of the wider project engineers to address allfire safety issues by aspects of fire safety the main objectives of this identifying the worst case scenario. The new study were to provide sufficient information on the standard requires a parative study to be carried space and time distribution of temperature and out to assess the potential impact of system failures smoke concentration in the building and to assess i.e. it has been recognised that it is difficult to both the most critical fire scenarios and the establish the level of fire safety in absolute terms. performance of the ventilation system. Of particular interest was data monitored on the escaping route The international airport in Belgrade (Serbia and at the head height. This data was used to assess the Montenegro) was opened in 1927. As a constant safety of escaping passengers. traffic increase demanded significant airport Based on the simulation it is evident that a enlargement the new passenger terminal was built significant increase of temperature and smoke in 1960's. Apart of the so called ‘non-functional′ concentration at the monitoring point occurred after ponents such as concession areas rest rooms 4.5 minutes for given conditions. Time lag of the munication systems none of the elements temperature front to the smoke concentration front providing services directly related to a passenger is observed. This study was conducted as a part of have not been refurbished since. As expected from the refurbishment project of Terminal 2 of Belgrade the airport terminal built more than three decades Airport in Serbia and Montenegro. ago there are three clearly subdivided area within the building: (a) processing facilities (process INTRODUCTION passenger and their baggage) (b) holding facilities During the airport terminal fire in Dusseldorf (areas in which passenger wait for some events Germany in April of 1996 17 people died 70 such as passport control) and (c) flow facilities others were injured and hundreds were trapped in toxic smoke. With millions of pounds of damage and holding facilities).However today the this was one of the worst airport catastrophes Belgrade airport aims to maintain a high level of worldwide. service offering more space to catering and retail businesses enhancing thermal fort and Increasingly the fire protection concems at airports improving safety procedures in passenger are being adressed using putational fluid terminals. As the new standard imposes °objectives* dynamics (BRE 2004) and most of these studies rather than ‘'solutions' the QDR enables project have been basedontheprescriptive engineers to identify and incorporate necessary remendations on design of an airport terminal changes in architectural design of building while ）sppes udoddeq uiSuq maintaining the specific level of fire safety. 1996).Although the prescriptive approach This study conducted as a part of the refurbishment provides adequate technical solutions to most of the project of Terminal 2 of Belgrade Airport questions regarding the fire safety in buildings a pares the effect of introducing the new - 1171 -
mechanical ventilation system on fire spread in the terminal building of Belgrade airport with the situation prior to refurbishment by following the principles given in the new performance based’ standard. Of particular interest was data monitored on the escaping route at the head height. This data was used to assess the safety of escaping passengers. SIMULATION The number of possible fire scenarios in the case of Figure 1 The lcyout and geometry of the design fire an airport terminal is almost indefinite and it is not focation. feasible to assess the effect of all of them. The modelled fire scenarios include the following Design fire. Most of the fires can be characterised factors: (a) occupant characteristics (b) design fire by the following phases (BSI 2003c): incipient and (c) fire location. phase (slow initial growth phase characterised by smouldering).growth phase (fire propagation Occupant characteristics. Variations in evacuation period) fully developed phase (steady burning response time are related to occupant familiarity of rate) decay phase (the period of declining fire building. position of occupants in building severity) and the extinction (no release of energy). occupant alertness and occupant mobility. Detailed In order to model the worst case fire scenario this analysis of escape behaviour in fires and layout of qs Aian pue oseud undiou ou sounsse pns the terminal building lead to conclusion that the growth phase of 10 seconds in which heat release of international arrivals located at the underground fire linearly reaches 500 kW/m². As the objective of level would present the most critical area regarding this study was to evaluate the ability of the fire safety as most of passengers would not be occupants to escape from the fire the remaining familiar with the building. In addition passengers* decay phase was not considered. The modelling movement is limited before passing passport results presented in this study take into account the control and reclaimed baggage would block escape first 5 minutes of the growth and fully developed routes. fire only. Fire location. In a life safety analysis a fire located In a fire model with the standard k & turbulence adjacent to the widest exit route generally model the air flow induced by the fire has been represents the worst case scenario as it will mean oenbo uodsuen eous a Suapos Sq pasopad that the exit would not be available for escape (BSI for the mean flow variables in the well known 2003c). Furthermore as the location of the fire form: within the terminal building influences the time required by the fire services to begin to fight the fire once they have arived on site the special (1) attention was paid locating the design fire. The r international baggage claim area of the terminal building located at the underground level was There are eight equations in the form of equation selected for design fire location. Figure 1 shows the (1) with includes air velocity ponents a v w layout and geometry of the design fire location. of the vector U temperature T turbulent kinetic energy k and dissipation of turbulent kinetic The initial design fire area of 42.4 m² corresponds energy smoke mass concentration C and total with the actual size of the conveyer belts currently radiative heat flux R..P.
is the air density S is used at the airport. Mass release rate of fire of 0.6 the source term and I is the total diffusivity in the kg/s for bustible content of cotton leather equation for flow variable β. The equation (1) is wool cellulose and rubber was assumed. coupled with the mass continuity equation. The fire is taken as a volumetric heat source and appears in the source term of the conservation equation for temperature. Buoyancy force due to temperature variations was taken into account using Boussinesqueapproximation in equation for the vertical velocity ponent. - 1172 -
Non uniform Cartesian grid was applied with total were obtained in the case of smoke concentration number of cells in exceedance of 500.000. Number they have not been presented. of time steps was set to 60 for 300 seconds simulation. The simulation was performed using a Because of its buoyancy the smoke rises from the general CFD code PHOENICS 3.5. fire. Initially the buoyancy is low and the smoke DISCUSSION follows the ambient air curents present before the The numerical simulations of different fire fire occurred but as its heat release increases its buoyancy increases and begins to dominate air scenarios within the terminal building were performed assuming both‘not ventilated’case currents in the enclosure. Progression of these isosurfaces coincides with the similar experimental (prior to the refurbishment) and 'ventilated’ case (normal working conditions of the ventilation studies forming the plume zone just above the system). All doors along the evacuation routes were location where fire initially started and spreading towards the stairwell openings and the underground opened when fire started. Based on the simulation both temperature (100°C) and smoke concentration ceiling. (4000 ppm) isosurfaces were derived. Progression of these isosurfaces coincides with the similar experimental studies (FVLR 2002) forming the plume zone just above the location where fire initially started and spreading towards the stairwell openings and the underground ceiling. To evaluate the risk of smoke concentration on the escape route a numerical probe was located approximately 32 metres away from the design fire location at the height of 1.75 metres (Figure 2). Figure 3a Temperature distribution prior to refurbishment PROBE Figure 2 Location of α numerical probe within the intemational arrivals area. Figure 3 pares temperatures at the 1.75 m height assuming both (a) not ventilated case of the terminal prior to the refurbishment (b) ventilated case with the mechanical smoke protection ventilation system in place. Significant differences Figure 3b Temperature distribution with are apparent as the fire protection wall has been mechanical smoke protection venrilation system in added during the QDR stage of the project. It can place. be observed that the introduction of both measures provides a tenable environment for passengers and Figures 4 and 5 pare temperatures and smoke staff to enable their safe escape from the building. concentrations simulated in both cases (a) prior to Moreover by limiting the area of fire spread for a refurbishment (not ventilated case) and (b) after few hours it is believed that these measures have refurbishment (ventilated case) and recorded by assisted in property protection andmore numerical probe located in the exit area (see Figure importantly have enabled the entry and subsequent 2). Based on the *not ventilated’ simulation the operations of fire fighters. Note that the significant substantial increase in both temperature and smoke increase in temperature has not been predicted concentration occur approximately after 260 seconds. Assuming a passenger escaping speed of along the newly designed escape route (see Figure 3a numerical (red pen) probe. As similar patterns 0.6 m/s the passenger presentation time (interval - 1173 -
between the time at which a warning of a fire is time well exceeds 120 seconds. This is not given and the time at which a person reaches a acceptable as temperatures above 120 °C simulated place of safety assuming walking speed is after 100 seconds may cause skin pain and burns. unrestricted) from the international baggage arrival area would be an estimated 53 seconds. In the *ventilated’ case (which includes the newly suggested fire protection wall) the fire engineering tenability criteria are pletely satisfied meaning that a minimum clear layer height of 2.5 m above 38 the floor and a maximum upper layer temperature 36 not ventiated of 200 °℃. 0 1 4 32 ventilated Figure 6 and 7 show the critical 100 °C isosurfaces 2 in both cases *ventilated’ and °not ventilated’. With 26 mechanical ventilation in place both temperature and smoke concentration would reach the critical 22 values after 300 seconds only at the location where 20 fire initially started i.e. the intermational baggage 0 50 100 150 208 [sec ] 250 300 arrivals area (Figure 6). Although the smoke penetration through the stairwell openings in this Figure 4 Comparison of temperatures on the case would not be critical during the first five escape route (recorded by nuemerical probe) for minutesof the additionalstairwell both not ventilated and ventilated case pressurisation systems were remended in order to prevent smoke penetration to the ground level enabling all passengers and terminal staff to 25x10a evacuate the terminal promptly. The situation is significantly worse for not ventilated’ case. During 24xto the period of simulation the underground ceiling veintied zone is fully covered by the temperature front of 15xte 100 °C (Figure 7). At the same time the smoke 1axto* concentration front of 4000 ppm fully covers the underground ceiling area significantly penetrating 6ix10 the ground floor of the terminal building through the stairwell openings between the ground and the first floor. The stairwell openings between the 105 150 210 asc 300 ground and the first floor are affected as well. [set] Figure 5 Conparison of smoke concentrations on the escape route recorded by numerical probe)for both not ventilated and ventilated case Note that this ‘optimistic’time is calculated u wm snussed jo uonqsp upmes area ignoring disability of passengers to orient in the space that they are not familiar with and find exits disregarding passengers reluctance to enter heavily smoke-logged escape routes and finally assuming tolerable levels of all radiant convected and conducted heat (BSI 2003d). The total Figure 6 100 °C isosurfaces after 300 seconds evacuation time calculated (BSI 2004) assuming a simulated for the ‘ventilated’ case. case where the area was sparsely populated with a population density of less than one third of the design population is approximately 80 seconds. In a case where the fire affected area contains the maximum design population the total evacuation -1174 -
majority of passengers and terminal staff to evacuate the terminal promptly. Although the standard has not been enforced yet in most of European countries (as in the case presented in this study). it is believed that this standard recently adopted in the UK on application of fire safety engineering principles to the design of buildings (BSI 2003a) provides a flexible approach to design using performance Figure 7 100 °C isosurfaces after 300 seconds related objectives rather that prescriptive solutions. sinuxlated for the 'not ventilated’ case. Although this “freedom’ in the design stage of a project encourages innovation it also carries an CONCLUSIONS additional liability not just for project engineers The recently adopted standard on application of fire safety engineering principles to the design of working on the project if the assumption made during the project were not based on well Qualitative Design Review (QDR) a structured documented evidence. technique that allows project engineers to address NOMENCLATURE all fire safety issues by identifying the worst case generalised physical property scenario. As the new standard imposes *objectives* (=U，C k ..) rather than ‘solutions’ the QDR enables project u turbulence fluctuating velocity engineers to identify and incorporate necessary vector (m/s) changes in architectural design of building while k turbulent kinetic energy (m'/s) maintaining the specific level of fire safety. Using 6 dissipation rate of turbulent kinetic these principles apart of the mechanical smoke cnergy (m² / s) protection ventilation the new fire safety wall was pa fluid density (kg/m) introduced in the intermational baggage arrival area S6 source term for variable d (kg/s)-(Φ) early in the design process r generalised uodsuen coefficient variable Φ (m²/s) of Based on the simulation both temperature (100°C) and smoke concentration (4000 ppm) isosurfaces REFERENCE have been derived. Progression of these isosurfaces BRE (2004) Fire modelling. BRE Digest coincides with the similar experimental studies 367.Building Research Establishment. forming the plume zone just above the location Watford England. s ps BSI(1996) Fire precautions the design the stairwell openings and the underground ceiling. construction and use of buildings. Guide to fire safety codes ofpractice for particular In the case with no ventilation the temperature front applications. British Standards. has 2minutes time delaytothe smoke concentration front of 4000 ppm. Unlike of BSI (2003a) BS 7974: 2001 Application of fire safety engineering principles to the design of temperature which reaches 100°C during the first five minutes at the underground ceiling level only buildings: Code of Practice. British Standards. the smoke penetrates throughout the stairwell BSI (2003b) PD 7974-0: 2003 Application of fire openings towards both first floor and underground safety engineering principles to the design of level of the terminal. buildings: Part O: Guide to design framework and fire safety engineering procedures. British However with mechanical ventilation in place both Standards. temperature and smoke concentration would reach BSI (2003c) PD 7974-1: 2003 Application of fire the critical values only at the location where fire safety engineering principles to the design of initially started. Although the smoke penetration buildings: Part 1:Initiation and development of through the stairwell openings in this case would fire within the enclosure of origin (Sub-system not be critical during the first five minutes of fire 1) British Standards. the additional stairwell pressurisation systems were BSI (2003d) PD 7974-2: 2003 Application of fire remended inorder toprevent smoke safety engineering principles to the design of penetration to the ground level enabling vast buildings: Part 2:Spread of smoke and toxic - 1175 -
