IMO SMOKE CONTROL AND VENTILATION. Research project on smoke control systems on passenger ships. Submitted by Sweden

Transcription

1 INTERNATIONAL MARITIME ORGANIZATION E IMO SUB-COMMITTEE ON FIRE PROTECTION 46th session Agenda item 4 FP 46/INF.6 November ENGLISH ONLY SMOKE CONTROL AND VENTILATION Research project on smoke control systems on passenger ships Submitted by Sweden SUMMARY Executive summary: This paper presents a summary of a research project on smoke control systems on passenger ships performed at Lund University Action to be taken: Paragraph 3 Related documents: FP 4/, FP 4//, FP 4//, FP 44/5, FP 44/5/ and FP 46/4/ The benefits and drawbacks of smoke control have been investigated in a number of different research projects. In Sweden, a research project on smoke control systems on passenger ships has been carried out during at Lund University. A short summary of this project is presented in the attached annex. The complete report will be available on the website of the Institute of Fire Safety Engineering at Lund University in January at The project has studied a number of different accommodation spaces on passenger ships. Simulation of the smoke spread with and without smoke extraction has been done. The input to these simulations has been taken from existing ships. The main conclusions from the project were that:. efficient smoke control can be achieved with modest costs;. within large public spaces and atriums the capacity of the normal comfort ventilation are in most cases sufficient; and.3 in cabin areas sufficient smoke control can be achieved by extracting smoke from the corridors, the capacity of the normal ventilation system is however in this case not adequate. Action requested of the Sub-Committee 3 The Sub-Committee is invited to note the information provided in the annex and take action as appropriate. *** For reasons of economy, this document is printed in a limited number. Delegates are kindly asked to bring their copies to meetings and not to request additional copies.

2

3 EXTRACT FROM THE REPORT: SMOKE CONTROL SYSTEMS ABOARD A RISK ANALYSIS OF SMOKE CONTROL SYSTEMS IN ACCOMMODATION SPACES ON PASSENGER SHIPS, ANDERSSON C., SÄTERBORN D., LUND, SWEDEN Executive summary: This document contains results and conclusions from computer simulations concerning smoke control systems for accommodation spaces aboard large passenger ships. The conclusions are set out as a whole in Chapter 6. The main conclusions are: The use of smoke control systems works out well and is recommended for the spaces studied. Generally, for larger public spaces the capacity of the HVAC-system is sufficient. For the cabin areas studied, a separate system for smoke extraction or a higher capacity HVAC-system is needed. The new regulation in SOLAS II- entering into force of July concerning atriums may in many cases not give a sufficient capacity of the smoke control systems. Other criteria, such as height to the smoke layer, should be used. To find the required extraction capacity for a space, calculation models based on a design fire and the geometry of the space should be used.. BACKGROUND Chapter II- in SOLAS, concerning the fire safety aboard ships, has been altered a number of times over the years. The many changes and adaptations made the regulations more difficult to apply. In 99 it was decided to make a comprehensive review of chapter II-. During this review it was discussed to include new regulations concerning smoke control on passenger ships. It was however decided that these requirements should be issued as a MSC circular instead, except a generalised requirement for smoke extraction in atriums. The purpose of this study is to evaluate the risk level combined with usage of smoke control systems aboard ships in order to provide further information to be used during the development of the circular. The studies are presented in detail in the main report: Smoke Control Systems Aboard a risk analysis of smoke control systems in accommodation spaces on passenger ships by Andersson, Caroline and Säterborn, Daniel. The report was conducted to meet the requirements for a Master of Science degree in Risk Management and a Bachelor of Science degree in Fire Protection Engineering at the Department of Fire Safety Engineering, Lund University, Sweden. The work has been supervised by Techn. lic. Johan Wikman at the Swedish Maritime Administration and prof. Sven Erik Magnusson at the Department of Fire Safety Engineering in Lund. The main report is to be completed by the end of and will be available on the Internet: LIMITATIONS The three major limitations made in this study are: The study only comprises smoke control systems of the accommodation spaces in the object, meaning that high-risk areas like engine rooms and cargo decks have been left out. A separate analysis will have to be made for these areas.

4 Page The study is focused on how smoke control systems affect the height of the smoke layer. The effect of smoke control systems on smoke temperature and radiation from the smoke to evacuating persons has not been evaluated. Sprinkler systems have been neglected. To be able to analyse the worst probable cases for a fire on a typical passenger ship, the fire suppressing / extinguishing effect of a water sprinkler was omitted. The main report specifies more detailed information about the limitations and assumptions made. 3. PERFORMANCE CRITERION The evacuation time always has to be shorter than the time it takes for critical conditions to occur inside a building, or a ship. Parameters used to evaluate the level of critical conditions are for example visibility, radiation from the smoke layer, temperature, toxic gases and a combination of these. The performance criterions used in this report follow the recommendations from Swedish building authorities stated in Boverkets Byggregler, BBR 998:38 : Visibility: Temperature: The smoke free height, i.e. the height from the floor to the smoke layer, has to be higher than.6 + (. H) meters, where H is the ceiling height. A maximum temperature of 8 C in the air below the smoke layer. Radiation to a target (person): A short radiation intensity of maximum kw/m or a short heat release rate of maximum 6 kj/m added to the energy from a radiation of kw/m. The main criterion in this report is visibility. It should be observed that it s only the height of the smoke layer, and not the level of obscuration in the smoke that is used as a criterion. The criterion is set for escape routes and communicating spaces leading to escape routes. The other criteria have been taken into consideration during every simulation, and they have not been exceeded. 4. METHODOLOGY 4. Analysed objects The base for this study is the M/S Skåne, a Ro/Ro-passenger vessel operating in the Baltic Sea. The vessel has been used as an input source for different conditions concerning geometry, fire load and configurations of the existing ventilation system. Although M/S Skåne is a large ship, its public areas do not fully represent the range of possible variations in geometry that are becoming more and more common in modern passenger ships. The wish to be able to use geometries more representative for passenger vessels in general resulted in creating alternative geometry s based on the geometries of M/S Skåne. These alternative geometry s are called M/S Alternative. Detailed information on input data such as the design fire, geometry and ventilation is presented in Chapter 5.

5 Page 3 4. Identifying scenarios and ventilation configurations The semi-quantitative risk analysis method used to determine the worst probable scenarios is a method called PHA, or Preliminary Hazard Analysis. This type of analysis is mainly used as a first step to identify and estimate the possible hazards on a low-detail level in an existing structure. The purpose is to decide which hazards are in need of a more extensive analysis. This decision is based partly on the evaluated level of risk, i.e. probability and consequence. But it is also based on the location of the hazard since it is of interest to study how the location of a fire affects the final conditions like temperature, smoke spread etc. In the latter the purpose is to give preference to scenarios that affect different parts of the ships accommodation spaces. With consideration of location, three different scenarios were chosen to be objects for further analysis on M/S Skåne. The chosen fire scenarios are originating in a cabin, in a playroom for children by an arcade and in a cafeteria. These design fires have also been applied to M/S Alternative s geometries. All of the chosen scenarios have been further analysed in computer simulations, as shown in figure. Cabin, Scenario Cabin, Scenario 4 M/S Skåne Cafeteria, Scenario Children s playroom, Scenario 3 M/S Alternative Cafeteria, Scenario 5 Assembly hall, Scenario 6 Atrium, Scenario 7 Figure : Event tree of fire scenarios on M/S Skåne and M/S Alternative. Three different ventilation configurations were added to each of the scenarios above, i.e. the ventilation system was: not activated, corresponding to an emergency shut down situation; activated with the capacity of the existing HVAC-system installed on M/S Skåne. (excluding Scenario 4-7); and activated with the capacity required to keep escape routes free from smoke to such an extent that the requirements of BBR are fulfilled. The required capacities has for each scenario been compared to either the existing HVAC-system (Scenario -4) or to a fictive HVAC-system with a capacity of replacements/hour (Scenario 5-7). The repl./h capacity is a requirement for the air conditioning system to achieve comfortable conditions in public spaces. It means that the air in the entire space should be replaced times per hour. 4.3 Computer simulations All simulations have been carried out on the computer program CFAST, a two-zone model for prediction of smoke spread. The program uses two control volumes to describe a room, the upper volume contains hot combustion products and the lower contains fresh air. Each layer is

6 Page 4 characterised by one temperature, one gas concentration and one smoke density. CFAST contains a number of sub-models, i.e. flame height, smoke velocity, the position of the interface, smoke temperature etc. These models are very complex and output from one sub-model becomes input to another. All sub-models contain approximations and assumptions. The total error of a prediction is a combination of all the assumptions made as input and the errors in sub-models. 5. ANALYSED OBJECTS 5. Cabin area The simplified geometry for the computer simulations of the cabin fire on M/S Skåne, Scenario, is presented in Figure. The ceiling height is. m for all compartments. The compartments are numbered from to 7 representing:. Cabin fire compartment (3m 4 m).. Corridor, part I (m.m) 3. Corridor, part II (m.m) 4. Corridor, part III (7.5m.m) 5. Reception Hallway (8m 6.5m) 6. Hallway (8m.5m) 7. Hallway on deck above. Connected to compartment 5. (8m.5m) Figure : Geometry for simulations in CFAST The capacities used in simulations of the different scenarios are listed in the table below. Compartment # Existing ventilation capacity (m 3 /s) Scenario Required ventilation capacity (m 3 /s) Scenario Required ventilation capacity (m 3 /s) Scenario The design fire is one starting in one of the cabin beds, reasonably as a consequence of smoking or arson. The fire then spreads to adjacent beds reaching a total heat release rate of approximately 7 kw after 5 minutes. The fire then descends, equivalent to full-scale experimental data 3. Since the design fire includes a decay phase the smoke layer rises towards the end of the simulation. This is due to the reduction of smoke production. The lowest level of the smoke layer during the simulation has been considered as the critical smoke height. Lundin J., Uncertainty in Smoke Transport Models, Department of Fire Safety Engineering, Lund University, Särdqvist, S., Initial Fires, Dept. of Fire Safety Engineering, Lund University, Lund, 993

7 Page 5 The results are evaluated on basis of the following acceptance criterion: Smoke free height of at least.8 m above deck within the time for evacuation, 9 s.,5,5,5 5,5,5,5 5 Comp Comp 3 Comp 4 Comp Comp 3 Comp 4 No ventilation: With the HVAC-system completely shut down the smoke layer in the corridor descends to a critical level within 3 minutes. This is the case for adjoining areas as well. Existing ventilation: With the existing HVAC-system in the corridor running, tenable conditions are almost achieved. Compartment, 3 and 4 has a capacity of. m 3 /s each. The worst conditions concerning temperature and smoke will occur in compartment since this is the compartment closest to the fire. Here the smoke free height reaches a minimum level of.6 m.,5,5,5 5 Comp Comp 3 Comp 4 Required ventilation: This is a solution with the entire smoke extraction system concentrated to the corridor. With one exhaust in every corridor compartment (comp., 3 and 4) with a capacity of.7 m 3 /s each, the criterion is met for all escape routes, i.e. the corridor and the hallways. The same design fire as for Scenario has been used for the cabin scenario on M/S Alternative, Scenario 4. The only difference is that the cabin size is twice as large (6.m 8.m). The results given from computer simulation on M/S Alternative are presented below. The results are evaluated on basis of the acceptance criterion: Smoke free height of at least.8 m above deck within the time for evacuation, 9 s.,5,5, Comp Comp 3 Comp 4 No ventilation: With the ventilation shut down the smoke rapidly descends below the acceptance criterion. Since the design fire has a decay phase the smoke production decreases with time and the smoke layer therefore rises some at the end of the simulation. The smoke will then spread to adjoining compartments, but will be more and more diluted. At 3 minutes the acceptance criterion in the corridor is exceeded.

8 Page 6,5,5, Comp Comp 3 Comp 4 Required ventilation: The capacity required for a smoke control system with exhausts in the corridor alone is at least 3. m 3 /s. With this capacity tenable conditions can be reached for the corridor as well as the adjoining escape ways. The smoke free height reaches its minimum after 5 minutes (comp. ), stabilises for a few minutes and then rises. For the cabin area (Scenario ) a total capacity of. m 3 /s is required in the corridor to keep the smoke layer above the critical height. The existing HVAC system has a capacity of.66 m 3 /s, which corresponds to 3% of the required. For Scenario 4 with 3. m 3 /s required this decreases to %. 5. Large public spaces The simplified geometry for the computer simulations of the public spaces (Scenarios, 5 and 6) are presented in Figure 3. The compartments are numbered from to representing:. Cafeteria, part I, fire compartment (9.m 5.5m). Cafeteria, part II (9.m 5.5m) 3. Driver s dining room (5.5m 7.m) 4. Stairway (6.m 8.m). Hallway, on deck above connected to compartments 4 and (8.m.m). Stairway (4.m 6.m) Figure 3: Geometry for simulations in CFAST The capacities used in simulations of the different scenarios are listed in the table below. Comp. # Existing ventilation capacity (m 3 /s) Scenario Required ventilation capacity (m 3 /s) Scenario Required ventilation capacity (m 3 /s) Scenario 5 Required ventilation capacity (m 3 /s) Scenario 6 (3MW) Required ventilation capacity (m 3 /s) Scenario 6 (6MW) For Scenario the ceiling height was set to. m. The design fire starts in the middle of the cafeteria, originating in a television set and spreading to nearby wooden structures. The fire peaks at 3 kw after approximately 6 / minutes and the heat release rate is then kept at this level through the whole simulation.

9 Page 7 The results are evaluated on basis of the acceptance criterion: Smoke free height of at least.8 m above deck within the time for evacuation, 9 s.,5,5,5 5 Comp No ventilation: This is the solution used today on M/S Skåne. It will lead to a complete smoke filling of the cafeteria. The acceptance criterion is exceeded within four minutes,5,5,5 5 Comp Existing ventilation The existing exhausts of the cafeteria have a total capacity of 4. m 3 /s, in the simulations divided into two equal exhausts in the two cafeteria compartments. Running this system results in almost acceptable evacuation conditions but the smoke free height cannot be kept at.8 m.,5,5,5 5 Comp Required ventilation: To achieve equilibrium between smoke produced by the fire and the smoke extracted, the capacity of the smoke control system has to be at least 5. m 3 /s. The exhausts are divided equally in the two cafeteria compartments. With this capacity the hot smoke layer stops descending at a level of.8 m after about 7 minutes. The compartment area in the cafeteria on M/S Alternative is the same as on M/S Skåne. In Scenario 5 the ceiling height has been risen to 4.4 m. The same design fire as for Scenario has been used. The results are evaluated on basis of the following acceptance criterion: Smoke free height of at least. m above deck within the time for evacuation, 9 s Comp No ventilation: Without a smoke control system the conditions in the cafeteria become critical within 6 minutes. At this time the smoke can also spread to adjoining compartments or fire zones if any door has been left open during the evacuation procedure.

10 Page Comp Required ventilation: The capacity of the smoke control system in the cafeteria has to be at least 3.5 m 3 /s to keep the smoke layer above the critical height. Since the critical smoke layer height,. m, is lower than the door heights,. m, the smoke can continue to spread to other compartments; for example the stairways. Due to this the capacity of 3,5 m 3 /s is an absolute minimum for this geometry. The geometries of the cafeteria also have been altered to represent an assembly hall on M/S Alternative. Thus, in Scenario 6 the ceiling height has been risen to m. The design fire has a growth rate corresponding to NFPA s Fast with a peak heat release rate of 3. MW after approximately 4 minutes. This can be compared to a fire in two upholstered chairs 4. Since the amount of interior varies from ship to ship, as well as the materials used, the magnitude of the fire has been sensitivity analysed. As a result of this the maximum heat release rate is doubled to 6. MW, keeping the fast development. The results are evaluated on basis of the acceptance criterion: Smoke free height of at least.6 m above deck within the time for evacuation, 9 s Comp No ventilation: Without running the ventilation system, the smoke will reach a critical height after about 4 minutes. The smoke filling process will be similar up to this time for both a 3MW fire and a 6MW fire. As the 6MW fire continues to develop the smoke will become much thicker and warmer than for the 3MW fire. Smoke spread to adjoining compartments due to open doors will possibly occur after 4 ½ minutes Comp Required ventilation: The 6 MW fire required an exhaust capacity of 8.5 m 3 /s to meet the criterion. A system of this size also prevents smoke from spreading to adjoining compartments. For the 3. MW fire the total required capacity of the smoke control system was 5 m 3 /s. As for the 6 MW fire the spread of smoke to the surrounding compartments can be avoided. In the cafeteria (Scenario ) a capacity of 5. m 3 /s is required to create tenable conditions during the entire evacuation procedure. The existing HVAC-system has a capacity of about 4 m 3 /s according to drawings. This capacity corresponds to 8% of the required capacity. Measured capacities of the existing HVAC-system on M/S Skåne shows that the capacity in reality is even higher, in fact closer to 5 m 3 /s, which would meet the criterion. 4 Särdqvist, S., Initial Fires, Dept. of Fire Safety Engineering, Lund University, Lund, 993

11 Page 9 The meeting of the criterion for the smoke layer height and the comfort requirement of replacements/h seems to have an analogous relation for larger public spaces, atriums with some exceptions though. A comparison with the geometries on M/S Alternative (Scenario 5 and 6) shows that this requirement ( repl./h) on the HVAC system even results in capacities higher than the ones necessary to keep the smoke layer at the critical height (see table below). 5.3 Arcade Type of system Cafeteria Scenario Capacity (m 3 /s) Cafeteria Assembly Scenario Hall 5 Scenario 6 HVAC-system according to drawings or with repl./h Smoke control system with required capacity (6MW) (3MW) The simplified geometry for the computer simulations of the arcade, Scenario 3, is presented in Figure 4. The compartments are numbered from to 3 representing:. Children s playroom, fire compartment (4.6m 6.m). Arcade, part I (3.5m.m) 7. Air-seat lounge (5.5m 6.6m) 8. Arcade, part II (3.m 6.5m) 9. Reception-hallway (6.5m 8.m). Hallway, on deck above connected to compartment 9 (.5m 8.m) 3. Tax-free shop (7.m 8.m) Figure 4: Geometry for simulations in CFAST The capacities used in simulations of the different scenarios are listed in the table below. Compartment # Existing ventilation capacity (m 3 /s) Required ventilation capacity (m 3 /s) Arcade only Arcade only Whole fire zone,6,6 -,, 4, 7 -, - 8,, 4,

12 Page Since children playing with fire was said to be the cause of this scenario, the fire originates in the children s playroom. The children s playroom consists mainly of a pool of small plastic balls. To decide a probable HRR the fire was approximated as a polymer pool fire. The exact polymer, which these balls are made of, is not known but was approximated with the fire specific properties of PVC. Hand calculations resulted in a HRR close to. MW. The growth rate corresponds to NFPA s Fast. The results are evaluated on basis of the acceptance criterion: Smoke free height of at least.8 m above deck within the time for evacuation, 9 s.,5,5,5 5 Comp Comp 8 No ventilation: This solution, used today on M/S Skåne, will lead to a complete smoke filling of all escape routes, as well as all other compartments, within the analysed fire zone. The acceptance criterion for the arcade is exceeded within two minutes.,5,5,5 5 Comp Comp 8 Existing ventilation: The capacities of the existing exhaust in the escape ways are.6 m 3 /s in Comp. (negligible) and. m 3 /s in Comp. and 8 respectively. Still, this capacity is too small to keep the smoke layer above.8 m.,5,5,5 5 Comp Comp 8 Running all existing exhausts in the fire zone affects the level of the smoke layer in a more distinct way. Even though the acceptance criterion isn t fulfilled the conditions in the escape ways are practically tenable with a smoke free height of.5 m for almost 6 minutes. In this case the total capacity of the smoke control system is.63 m 3 /s.,5,5,5 5 Comp Comp 8 Required ventilation: The required exhaust capacity in arcade (comp. and 8) is 4. m 3 /s. The recommended solution for the smoke control system in Scenario 3, is to extract smoke through the exhausts in the escape ways alone. With a total exhaust capacity of 8. m 3 /s equally distributed in the arcade the conditions meet the criterion of.8 m.

13 Page The HVAC system in the arcade has a rather low capacity,.4 m 3 /s, compared to other public areas. If the exhausts are placed as recommended in the main report i.e. in the escape ways the total capacity of the smoke control system would have to be about 8 m 3 /s for the set design fire. The HVAC-capacity in the arcade corresponds to / of the required. 5.4 Atrium The simplified geometry for the computer simulations of Scenario 7 is presented in Figure 5. Compartment represents the atrium, and is also the fire compartment. The atrium measures 9. m 5.5 m and is. m high. Figure 5: Geometry for simulations in CFAST The capacities used in simulations of the different scenarios are listed in the table below. Compartment # HVAC system with Required ventilation capacity (m 3 /s) repl./h (m 3 /s) 6.4 MW, BBR 6.4MW, US proposal 3. MW, BBR 3. MW, US proposal 3,,5, The fire originates in a television set and spreads to nearby wooden structures. The development of the heat release rates for TV s and particle boards are taken from separate experimental data 5 and put together into a design fire peaking at 3. MW after approximately 6 / minutes. Due to the large volume and different possible uses of an atrium the maximum heat release have also been doubled, continuing from 3. MW to 6.4 MW with a growth rate corresponding to NFPA s Ultra Fast. The results are evaluated on basis of two different criterions:. BBR: Smoke free height of at least 3.8 m above the atrium floor within the time for evacuation, 9 s.. US proposal: Smoke free height of at least.8 m above the upper most deck within the time for evacuation, 9 s. This corresponds to a smoke free height of.6 m. 5 Särdqvist, S., Initial Fires, Dept. of Fire Safety Engineering, Lund University, Lund, 993

14 Page No ventilation: For the two design fires in the atrium the BBR acceptance criterion will be exceeded in about 4 minutes when not running any system for smoke extraction. After 7 minutes the atrium will be completely filled with smoke, with higher temperature and obscurity for the 6.4 MW fire than for the 3. MW fire. In the aspect of smoke filling the graph to the left is representing both fires. Required ventilation: Six different scenarios have been analysed to find the different required capacities of the smoke control system in the atrium: MW fire - BBR To meet the acceptance criterion of 3.8 m, the exhaust capacity required for the atrium is. m 3 /s. When activated immediately this will keep the atrium completely free from smoke during the first eight minutes, but due to changes in temperature the smoke layer will descend, leading to a state of equilibrium at around 3.8 m above deck MW fire US proposal To be able to keep the upper most deck free from smoke for this design fire, a capacity of at least 3. m 3 /s is needed for the smoke control system. By meeting this criterion the prevention of any smoke from spreading to adjoining compartments is made MW fire - BBR For the smaller design fire a smoke control system with an exhaust capacity of at least.5 m 3 /s will be needed to meet the acceptance criterion of 3.8 m.

15 Page MW fire US proposal Keeping the upper most deck free from smoke for this design fire requires a smoke extraction system with the capacity of at least. m 3 /s. As for the 6.4 MW fire, further smoke spread will be prevented. The results above shows that an HVAC-system with the capacity m 3 /s (or repl./h) do not guarantee a safe egress environment depending on the structure of the atrium. An atrium with high placed openings to adjacent decks, would for the large design fire need a smoke control system with about 4 % higher capacity than that given from repl./h. This in order to keep the smoke layer on these upper most decks at a tenable level (.8m). However, if one allows the smoke to descend further in the atrium, or when looking at the smaller design fire, the system can be said to be sufficient. The exhaust capacity requirement in SOLAS chapter II-, stating that the entire volume of the atrium should be extracted within minutes is also compared to the BBR and US proposal criteria. For the atrium analysed, with a volume of 6479 m 3, the exhaust capacity would be.8 m 3 /s. This capacity is simulated both for the 3. MW fire and the 6.4 MW fire MW fire The simulation of a system designed to exhaust the entire volume of the atrium in minutes results in a smoke free height of 3. m. This is not in accordance with any of the criteria for the atrium used in this report MW fire For the larger fire, the smoke layer descends as low as. m above deck, clearly showing the insufficiency of the system as both the criteria set are exceeded. Studying the graphs above one can see that designing the smoke control system on basis of the existing SOLAS regulations for atriums cannot be considered as recommendable. Following these regulations of exhausting the entire volume in minutes will result in large differences in smoke extraction capacities for different volumes. Depending on the fire load in the atrium one cannot ensure that the smoke will be kept at a tenable height above deck with this method. The insufficiency can clearly be seen for the simulated fires were the smoke descends as low as. m above the lowest deck.

16 Page 4 6. CONCLUSIONS After consulting ventilation experts, the following conclusion was drawn: 6. General conclusions These conclusions are general for all scenarios: When designing a smoke control system one should use an acceptance criterion based on the smoke free height in a space, and not a criterion fixed to the volume of the space. The latter do not account for different fire load or the geometry of the specific space. An emergency shut down of the HVAC system, in the purpose to hinder smoke spread does not necessarily improve the evacuation conditions. If one continues to run the system in the affected fire zone the conditions will become better, provided that re-circulation of the extracted air is prevented by dampers. In other research this has been verified by full scale tests 6. Over-pressurisation of adjoining fire zones hinders or hardens the smoke spread to these. This is especially important when evacuation has to take place through another fire zone causing doors to be held open. The over-pressurisation can be managed with a Smoke Control Strategy. A smoke control system affects the pressure inside the area it is operating. One therefore has to pay special attention to the doors placed in the escape ways. To open a door, one should only need a maximum force of 3 N, according to the Swedish building regulations. This value could be considered to be required also for ships. Further research has to be made though. Another alternative, which also has to be evaluated further, is the possibility to change the opening direction of the doors in the escape routes. Exhausts should be placed at ceiling height to have the best effect. This since the hot gases are concentrated at ceiling level. A high placement decreases the amount of fresh air extracted. It is important to make sure that the system is resistant to hot smoke during a certain time (e.g. the fire resistance time for the divisions). The fans have to be fully operable during this time. The risk for sparks igniting uncombusted gases also has to be considered. It is also very important that the smoke control system installed hinders smoke spread to other parts of the ship trough the system. As an example one has to have at least one separate system for each fire zone. 6 Bengtsson S. et al, Smoke Distribution on Large Passenger Ships and the Effect of the Ventilation System, FOA, 99.

17 Page 5 6. Cabin area The following conclusions can be drawn for the cabin area: It is necessary to install a separate smoke control system in the cabin area or increase the capacity of the HVAC-system. The smoke control system has to be designed for a design fire where, among others, parameters like fire load and cabin size will affect the smoke production. As a comparison, scenario 4, with twice as large cabin as in scenario, requires a total smoke exhaust capacity of 3. m 3 /s compared to. m 3 /s. The exhaust should be placed in the corridor, preferably equally distributed. A separate smoke control system in each cabin is not recommendable. According to Scenario, extraction in each cabin requires a capacity of 4 m 3 /s. A system with this capacity would take up a lot of space and generate unnecessary installation costs. Supply air should as a rule be taken from the cabins existing supply air system. If this capacity is insufficient one can complement the amount of supply air with air from adjoining stairways if these are over-pressurised. One example of this is if the doors between the cabin area and the stairways are kept open when the system is operating. This situation is very likely to appear during an evacuation procedure. 6.3 Large public spaces The following conclusions can be drawn for larger public spaces: 6.4 Arcade If the HVAC-system in public spaces is designed to provide repl./h, the system most probably has enough capacity to be used as a smoke control system. In this case it is not necessary to provide more powerful fans or larger dimensions on the ducts since these are designed for the required capacities. This strongly encourages to design HVAC-systems for these areas to be able to operate as smoke control systems as well. This should be considered both when installing new systems and when upgrading existing. The following conclusions can be drawn for smaller public spaces: The HVAC system installed in a public space cannot straight off be considered to have the required capacity of a smoke control system. In contrast to the larger public spaces this typical example has a capacity that is far below the required. Hence, every system has to be designed according to specific design fires. The purpose of the smoke control system is to meet the criterion for critical conditions in the escape routes, in this case the arcade, during the time for the evacuation. The recommended solution is to place the smoke control extractions in the arcade, in the same manner as for the corridor in the cabin area (Scenario and 4). This solution could be used since the tax-free shop and the lounges are rather small compartments. It is in this case not reasonable to place a separate smoke control system in each of these rooms. Obviously this depends on the specific geometry.

18 Page 6 Another solution for these compartments is to limit the amount of fire load. This would lead to a smaller design fire and consequently a lower extraction capacity is required to meet the acceptance criterion. 6.5 Atrium The following conclusions can be drawn for atriums: The existing regulations in SOLAS does not require sufficient capacity of the smoke control systems in order to fulfil the acceptance criteria used in this report. The applicability of the SOLAS requirement is limited since no consideration is taken to different fire load (design fire size) and the desirable requirement that the smoke layer should be kept at a safe level above deck in all evacuation routes, independent of their level in the atrium. Regulations for smoke control should use the height to the smoke layer as acceptance criterion and not a criterion fixed to the volume of the atrium. The normal HVAC-system is in most of the cases studied barely sufficient and consequently it is relatively easy to upgrade it to a smoke control system. Achieving a smoke free height of.8 m at the upper most deck in the atrium is the most demanding requirement. According to the simulations rather high extraction capacities was required compared to those provided by a repl./h HVAC-system. This leads to higher installation costs and some loss of space due to larger ducts. It might instead be more appropriate to allow the smoke to descend to a lower level in the atrium and to prevent the smoke from spreading to adjacent spaces by other means, for example by smoke tight barriers.