Guidelines for Emergency Ventilation Smoke Control in Roadway Tunnels (2017)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

23 Tunnel Emergency Ventilation and Smoke Control Design Guidelines The purpose of controlling the spread of smoke and main- taining a tenable environment is to keep people in a smoke free or survivable environment as long as possible or for at least the duration of evacuation and rescue. This can mean one or both of the following: • Keep the smoke stratification intact, leaving more or less clean and breathable air suitable for evacuation under- neath the smoke layer to both sides of the fire (applicable to bi-directional or congested uni-directional tunnels and is typically achieved by zoned transverse ventilation or single point extraction as discussed in the following sections). • Completely push the smoke to one side of the fire (pref- erably applied to non-congested uni-directional tunnels where there are normally no people downstream of the fire and is typically achieved by longitudinal ventilation as discussed in the following sections). In all situations, people must be able to reach a safe place in a reasonably short amount of time (before tenability is in jeopardy) and cover a reasonably short distance. Emergency exits should be provided wherever necessary. Cross passages and egress stairs are common and an effec- tive method of providing means of egress. Spacing between cross passages and egress stairs shall be determined using egress and tenability analysis and depends on tunnel ventila- tion and FFFS activation and operation [16] and shall not exceed spacing limited by NFPA 502 Standard [1]. The stan- dard states that “Emergency ‘exits’ shall be pressurized in accordance with NFPA 92,” which means that they should be free from smoke in order to provide a safe evacuation path to evacuees. To achieve this, an airflow at air velocities greater than critical velocity that prevents smoke from get- ting into the cross passages/egress stairs shall be developed by the ventilation system with the direction of flow from the cross passage into an incident tunnel. This requires design analysis, but typically air velocities of not more than 2 m/s out of the cross passage/stair door(s) are sufficient when all desired egress doors are open. When the doors are closed the cross passages/egress stairs should be pressurized such that the force to open the doors fully, when applied to the latch side, should be as low as possible but in accordance with NFPA 502 [1], which states that the force to open shall not exceed 222N (50 lb.). Typically, such a requirement could be met with cross passage pressurization fans utilizing VFDs, barometric relief dampers, and sliding doors used in the path of egress. While multiple fire events at different tunnel locations at the same time are theoretically possible, the probability of such a catastrophic event is extremely low and has not hap- pened so far. It is a good engineering practice to design the sys- tems for a single fire event at a time, while emergency response plans may need to consider possible responses to more severe events. 3.1 Types of Road Tunnel Ventilation Systems Choosing the type of ventilation system is one of the most important decisions when designing a tunnel. While every situation has its unique features, some general conclusions can be drawn about the relative usefulness and efficiency of the various types of ventilation systems for smoke control. There are several types of natural and mechanical ventila- tion systems that have been used in road tunnels. Options for mechanical ventilation include longitudinal ventilation and transverse ventilation (fully transverse, semi-transverse, combination of, or modifications of the schemes discussed herein). A longitudinal ventilation system achieves its objec- tives through the flow of air along the tunnel roadway, while a full or semi-transverse ventilation system achieves its objec- tives by means of continuous uniform distribution, collection, or simultaneous distribution and collection of air throughout C h a p t e r 3

24 the length of the tunnel roadway. There are also many com- binations of different types of basic ventilation such as single point extraction, ventilation with intermediate shafts, etc. Each option is described in detail in this section. 3.1.1 Natural Ventilation Natural ventilation relies on natural phenomena and the traffic piston effect to ventilate the tunnel. Tunnel air is never still, as there is always some airflow caused by the piston effect and/or wind and other natural factors. During a fire emer- gency, it is expected that traffic will not be allowed to enter the tunnel and trapped vehicles will be stopped behind the fire location. The resulting piston effect generated by exiting residual vehicles leaving the tunnel and entering emergency vehicles will be minimal. However, adverse winds may have a significant impact on the tunnel airflow. Natural ventilated tunnels rely primarily on atmospheric conditions to maintain airflow and provide a satisfactory environment. The main factors affecting the environment are the pressure differential created by differences in elevation, the ambient air tempera- ture, and wind effects at the boundaries of the facility. The external effects that contribute to a difference in portal pressures are: • Portal entrance and exit pressure losses; • The stack (chimney) effect (in tunnels that have nonzero grades, particularly those with portals at different eleva- tions, a temperature dependent effect usually referred to as “the chimney effect” is often considered). Air will flow through a chimney, in a direction from the lower portal to the higher portal or vice versa, because of the difference between the weight of the column of air inside, the chim- ney (tunnel) and an equivalent column of air outside; • Atmospheric pressure differences (referred to the same elevation datum); and • Local wind effects at the portals. These effects combine to create a net external force acting on the tunnel air volume. The momentum equation is a balance of the following factors: Effect of wind; barometric pressure and temperature differences Piston Effect Friction losses in the tunnel plus exitandentrance losses { }       + =       Pressure losses at the entrance and exit portals are due to the sudden change in flow cross sectional area and can be expressed as the product of a loss coefficient and the local flow dynamic pressure at the portal. At the flow exit, the pressure loss (static or total) arises from the abrupt expansion into an essentially infinite cross section and is given by 2 Flow Exit 2 P U e∆ = ξ ρ The exit loss coefficient ξe is normally equal to 1.0 for the turbulent flows and cross section geometries found in most tunnels. The pressure loss is equivalent to the local dynamic pressure. Exit portals that have curved rather than square cor- ners (such as a bell-mouthed exit) will generally not reduce the exit loss coefficient. The pressure loss at the flow entrance is due to the abrupt contraction from an essentially infinite flow area to the tun- nel cross sectional area. This loss can also be expressed as the product of a loss coefficient and the local dynamic pressure: 2 Flow Entrance 2 P U i∆ = ξ ρ The entrance coefficient depends on the portal shape. For a flat square-edged entrance, the value of ξ i is usually taken as 0.5 with the resultant pressure loss equal to one-half of the entrance dynamic pressure. The entrance loss coefficient can be reduced significantly by rounding the portal corners (stream- lining) to achieve a smooth bell-mouthed entrance. Stream- lining reduces some of the flow separation to values as low as 0.1 for turbulent flows. For certain protruding entrance configurations, such as designs where the tunnel walls and roof extend significantly outward from a large sloped wall or hillside, entrance loss coefficients greater than 0.5 and up to 0.75 can be expected. If the portals are at different elevations, the hydrostatic difference in elevation head should be considered. For exam- ple, for long tunnels through mountain ranges, significant changes in barometric pressure (referred to the same eleva- tion datum) can result in a static pressure difference given by the equation: P P PAtm AtmO AtmL∆ = − In road tunnels that have nonzero grades, particularly those with portals at different elevations, a temperature depen- dent effect usually referred to as “the chimney effect” is often considered. P P P PAtmZo AtmZl elev atmO l− = ∆ + ∆ Typically, air will flow through a chimney in a direction from the lower entrance to the higher entrance usually because of the difference between the weight of the column of air inside the chimney (tunnel) and an equivalent column of air outside

25 the chimney (tunnel). If the interior temperature is higher (e.g., due to the heating effect of vehicles or soil conditions), then the tunnel air density will be lower, and the chimney effect will promote air flow in the direction of increasing elevation. Con- versely, if the tunnel air temperature is lower than that outside (e.g., due to heat transfer to cooler tunnel walls) the tunnel air density will be higher and the chimney effect will promote airflow in the direction of decreasing elevation. If it is assumed that the temperature variation inside the tunnel is approximately linear with elevation, then the force of the chimney effect can be calculated as: 8Chimney Linear F gA T Z Z ln T T T T ln T T T T r r l o Atm Atm Atm Atm L O L O L O L O [ ]( )= ρ − − − −       where TO and TL = the absolute temperatures of the tunnel air near the entrance and exit portals respectively; rrTr = the air density and absolute temperatures at a reference state (low portal); Zl; Zo = the elevations at the exit and entrance portals; TAtmL; TAtmO = the atmosphere absolute temperatures out- side the exit and entrance portals. External winds at tunnel portals can vary considerably in magnitude and direction as a function of time. Moreover, the net effect of wind action on the tunnel air flows depends considerably on the portal geometry and the surrounding topography. Portal winds can have significant favorable or unfavorable effects on road tunnel ventilation and must be considered. The difference of pressure caused by outside winds can be evaluated using the following simplified equation of Bernoulli: 1 2 sin 352P k W [ ]∆ = ρω φ where DP = the pressure induced by wind; r = the air density; w = the wind speed typically obtained from the wind rose for the area or from the local wind data; fW = the angle of wind with respect to the tunnel axis; k = a design parameter which depends on the configu- ration of the portals. This effect was studied by W. Blendermann [36] (Table 3.1 and Figures 3.1–3.2). The orientation of both tunnel portals with respect to the prevailing winds is a very significant parameter. The effective wind resistance (or thrust) is a function of the angle between the direction of the wind and the direction of the air flow entering/exiting the tunnel. ADDITIONAL FEATURE PORTAL ABOVE GROUND LEVEL PORTAL BELOW GROUND LEVEL WITH VERTICAL SIDE WALLS WITH SLOPING BOUNDS ---- Figure 3.2 (a) Figure 3.2 (d) * Dividing wall Figure 3.2 (b) Figure 3.2 (e) * Light adaptation section Figure 3.2 (c) Figure 3.2 (f ) Dam * * Figure 3.2 (g) Table 3.1. Configuration of tunnel portals tested by W. Blendermann [36]. Figure 3.1. Mean wind pressure coefficients [36].

26 The traffic condition may need to be evaluated. For exam- ple, in uni-directional tunnels, the assumption is made that in a fire emergency, the tunnel will be closed for on-coming traffic; some traffic will be trapped behind the fire while traffic downstream of the fire will leave the tunnel. The leaving traffic will cause a residual piston effect, driving smoke and airflow in the direction of travel. Residual air movement, caused by traf- fic approaching the fire location (piston effect) may also drive the airflow until the traffic movement in the tunnel is halted. The piston effect could be developed by emergency response vehicles as well while they move into the tunnel. The vehicle drag force FD and the vehicle drag coefficient CD are defined by the following expression involving the tun- nel air velocity U relative to the vehicle speed V and the vehi- cle frontal area AV: 2 2F C A V UD D V ( )= ρ − Figure 3.2. Configurations of tunnel portals tested by Blendermann [36].

27 The vehicle drag coefficients are measured using wind tun- nel tests. Heavy duty vehicles such as trucks and buses have larger drag coefficients than passenger vehicles typically in the range of 0.6 to around 1.0. This is due to the wide variety of heavy duty vehicle profiles including cab-over and conven- tional style tractors and a large number of trailer and body- work configurations. Average vehicle drag coefficient values of 0.4 for automobiles and 0.75 to 1.0 for heavy duty vehicles have been used in a number of highway tunnel studies. The equation for naturally ventilated tunnels with uni- directional traffic can be written as: 82 2C C A A NL V V U L D U D V i e  [ ]( )+ − = ξ + ξ + λ  where C ~ = wind, barometric pressure and temperature difference effects; N = number of vehicles traveling per unit length of tunnel; L = tunnel length; D = tunnel hydraulic diameter; l = The Darcy coefficient (a function of the Reynolds number and the ratio of the surface roughness to the hydraulic diameter and is determined from the Moody diagram or the Colebrook equation). Due to the fully developed turbulent flow and rough walls in tunnels, it can usually be assumed that the Darcy coefficient is a constant for a given tunnel with the variation from tunnel to tunnel primarily due to the type of wall lining used. Values commonly used for the Darcy coefficient in the design of road tunnels are 0.015 to 0.020 for very smooth lining, 0.025 for plain concrete, and 0.03 to 0.05 for bare rock with no lining. Several aspects must be taken into account for the effects of a tunnel fire on the air flow in a tunnel: • In the event of a large fire (where buoyancy forces are dominant), the high temperatures induce an increase of air volume (due to expansion) and therefore of air speed, as a result of which the air friction losses increase. • The density decreases, friction velocity increases, and the overall local losses increase. • The blockage effect of the fire on the longitudinal airflow produces a supplementary local head loss. • With steep grades of tunnels, the force of the chimney effect can rise to significant values. Natural ventilation systems can be very effective for the dilution of pollutants (especially for uni-directional tunnels), but it is difficult to rely upon natural ventilation for safety purposes. In fact, most of the natural ventilation factors are highly variable with time and therefore unreliable. Because of the number of different parameters that need to be considered in order to decide if mechanical ventilation is to be installed in a tunnel (length, location, traffic, type of vehicles using the tunnel, risks, and so forth), it is not possible at this moment to express universal recommendations about the limits of natural ventilation, particularly the allowable length without mechanical ventilation. NFPA 502 requires analysis of numerous factors to make a decision regardless of the length of the facility [1]. During a fire event with practically zero longitudinal air velocity and no significant grade, the smoke layer expands to both sides of the fire and smoke may spread in a stratified way for up to 10 minutes (to a distance of 400 to 600 m depend- ing upon the tunnel geometry and fire conditions). After this initial phase, smoke starts to mix over the whole cross section, unless by this time the mechanical ventilation, such as an extraction system, is in full operation. A tunnel that is long or experiences frequent adverse atmospheric conditions requires fan-based mechanical ventilation. 3.1.2 Longitudinal Ventilation Longitudinal ventilation introduces air into or removes smoke and gases from the tunnel at a limited number of points, primarily by creating longitudinal airflow through the length of the tunnel, from one portal to the other. Longitudi- nal ventilation can be accomplished either by injection; cen- tral fans; jet fans mounted within the tunnel; nozzles (often installed at the portals and called a Saccardo system); or through a combination of injection and extraction at inter- mediate points. The system must generate sufficient longi- tudinal air velocity to prevent backlayering of smoke. The air velocity necessary to prevent backlayering over stalled or blocked motor vehicles is the minimum velocity needed for smoke control in a longitudinal ventilation system. This veloc- ity is known as the critical velocity and was discussed previ- ously in Section 2.2. The limitations of longitudinal ventilation systems are often related to limitations of air velocity in road tunnels (see Sec- tion 3.2.1). The limitations on high air velocities are typically driven by normal ventilation requirements for the dilution of concentrations of vehicle emissions during free flowing and congested traffic and are seldom related to fire emergency conditions. Exceptions could be emergency response require- ments to limit smoke spread along the tunnel or turbulence that affects the smoke stratification downstream of the fire. This phenomenon is more evident at higher air velocities. Smoke stratification can also be disturbed by the longitudi- nal slope of the tunnel (especially when air flows downwards) and by vehicles. Longitudinal ventilation is particularly suited for tun- nels with free flowing uni-directional traffic because of the

28 assumption that drivers downstream of the fire are free to escape by continuing to drive towards the exit portal while drivers upstream remain behind the fire. Other types of traf- fic flow, including bi-directional traffic and uni-directional congested traffic, are better supported by other ventilation schemes. Figure 3.3 presents an example of a design process for the selection of the type of road tunnel ventilation system. Numbers presented in Figure 3.3 are for illustrative purposes only. In addition to the factors shown in the figure, the deci- sion process for ventilation scheme selection should include traffic volumes, tunnel location, environmental emission, and additional factors. One of the common methods of achieving longitudinal ventilation is the use of jet fans and this type of system has been installed in numerous tunnels worldwide (Figure 3.4). With this scheme, specially designed axial fans (jet fans) are typically mounted at the tunnel ceiling or side niches. This system eliminates the space needed to house large ventilation fans in a transverse ventilation system at a separate structure or ventilation building but may require greater tunnel height or width to accommodate the jet fans so that the fans are outside the tunnel’s dynamic clearance envelope. This enve- lope, formed by the vertical and horizontal planes surround- ing the roadway in a tunnel, defines the maximum limits of the predicted vertical and lateral movement of vehicles travel- ing on the roadway at the design speed. However, as tunnel length increases, disadvantages such as excessive air speed in the roadway and smoke being drawn the entire length of the roadway during an emergency fire event become apparent. The increase in pressure necessary to generate or main- tain the longitudinal flow in the tunnel is provided by the acceleration of the air flowing through the fan. Although jet fans deliver relatively small air quantities at high velocities (in the range of 25–45 m/s [4900–8900 fpm]), the momen- tum produced is transferred to the entire tunnel, inducing airflow in the desired direction. Jet fans are normally rated in terms of thrust rather than airflow and pressure and can be either uni-directional or reversible. It should be noted that 3. Will the tunnel NOT be used for bi-direc onal traffic during construc on, maintenance or other events?* 4. Are tunnel air veloci es less than (10 m/s [2000 fpm]) during normal, congested and fire emergency? 1. Is this a uni-direc onal tunnel? Consider * - In certain cases longitudinal ventilation could be justified for contra-flow traffic. longitudinal ven la on 2. Can the traffic management system manage traffic downstream of fire event? No No No No Yes Yes Yes 5. Is this a short tunnel less than 800 ‹. (240 m) long with constant uphill grade of 4% or more and no flammable and combus ble fuel vehicles? Yes Yes Consider transverse or semi-transverse exhaust, or single point extrac on ven la on scheme for smoke and hot gas management or modifica on of longitudinal ven la on Figure 3.3. Example of design process for consideration of longitudinal ventilation for the road tunnel ventilation system selection. Figure 3.4. Diagram of longitudinal ventilation controlling smoke and hot gases.

29 the decrease of air density during a fire event results in the lowering of the driving force of the jet fans that work in the hot air. Jet fan sizing is usually limited by the space available for installation in the tunnel (see Section 3.3). Calculations of jet fan capacity should take into account what air velocities should be sufficient for control of fire smoke. Usually the jet fans are designed to be reversible to allow for first responders to switch the direction of smoke movement after the evacuation is complete. For longitudinal ventilation using jet fans, the required number of fans is defined (once fan size and tunnel airflow requirements have been deter- mined) by the total thrust required to overcome the tunnel resistance (pressure loss), divided by the individual jet fan thrust, which is a function of the mean air velocity in the tun- nel. Fan thrust NS values provided by fan manufacturer can be defined by the equation: 1N C QUS = ρ where C1 = correction factor; r = air density; Q = air volume; U = air velocity The effective thrust imposed by the jet fan on the tunnel is based on the jet velocity relative to the tunnel air velocity. There is also an impact from the tunnel walls and from other parallel jet fans installed within the same group of fans. The effective fan thrust Ne can be written: 1 2 3N N U U C Ce s t j = −     where Ut = tunnel air velocity; Uj = jet velocity developed by jet fan; C2 = jet fan installation factor which depends on wall and/or ceiling proximity, jet fan inclination angle and tunnel niche effect; C3 = group of fans installation factor as a function of fan spacing, fan proximity to the entrance and exit portals. If the jet fan is placed in the center of the tunnel tube, ver- tically and horizontally, the installation factor will equal 1. However, such installation is impossible. The closer the jet fans are installed to the tunnel walls and ceiling, the stronger the wall roughness impacts the effective fan thrust. In order to create maximum space for the traffic in a road tunnel, the fans are often installed close to the ceiling. A considerable portion of the jet energy is lost to wall friction. The highest losses occur when the fan is installed in niches. Jet fan installation in niches should be accounted for by a reduction in the coefficient of efficiency of the impulse force of the fan to determine the longitudinal airflow in the tunnel. Moving the impulse fan further into this niche increases its eccentricity, until the fan is situated completely outside the tunnel tube profile. The additional losses due to wall fric- tion will grow correspondingly. As a result, jet fans can lose up to 30% of their efficiency due to niche installation [37]. Howden Standard 66-05.150 provides loss coefficients for niche installations. The total thrust developed by a number of fans in a tun- nel is the sum of the individual thrusts. The number of fans required is equal to the total tunnel thrust required divided by the effective fan thrust. Jet fans must not be installed too close to each other. Jet fans installed longitudinally should be installed at a sufficient dis- tance away from the next jet fan installation location to allow the tunnel air velocity profile to become approximately uni- form at the following location, which is at least 7 to 10 tun- nel hydraulic diameters (or sometimes identified as 100 fan diameters) apart so that the jet velocity does not affect the performance of the downstream fan. Otherwise, the jet stream of the preceding fan would not have fully decayed before being drawn into the suction side of the next fan. To ensure that the jet fans yield maximum efficiency, the jet stream from any fan must be fully decayed before it reaches the next jet fan. Jet fans installed side by side should be at least two fan diameters (centerline to centerline) apart. In this case they can be approximated as a single jet fan of equivalent cross sectional area. Reduction of spacing between jet fans should be accounted for in the jet fan installation factor. The axial installation distance between fans could be reduced or the installation factor improved if flow deflectors are installed. Jet fans installed close to the portal (at a distance less than 200 ft [60 m] or 40 fan diameters—see Figure 3.6) will lose their efficiency and will not produce the desired amount of air flow in the tunnel in a fire emergency. When jet fans are installed at a distance of less than 150 ft [46 m] or 30 fan diameters (see Figure 3.5 and Figure 3.6) from the portal and direct air out of the tunnel, air recirculation in the portal area will occur. Figure 3.6 shows an example of the tunnel airflow rate reduction in percentage as the result of the jet fan installation location proximity to the tunnel portal. The values depend on the tunnel geometry and Figure 3.6 is for illustration pur- poses only. Operating jet fans close to the fire is not recommended as air recirculation typically occurs locally at the jet fan loca- tions. The recirculation could destroy stratification and draw smoke and hot gases in the opposite direction of the intended air flow. Jet fans are often fitted with integral silencers and deflect- ing vanes to achieve an acceptable sound level and improve

30 air flow development in the tunnel. Silencers and vanes cause additional resistance to the airflow which results in a reduc- tion of fan thrust. Jet fans can have several speeds or be sup- plied with adjustable (variable) speed drives for reduced fan operation and power consumption during normal operation such as during night time when ventilation needs are lower compared to in emergency mode. 3.1.3 Longitudinal Ventilation Through Injection Using a High-Velocity Saccardo Nozzle The Saccardo nozzle functions on the principle that a high- velocity air jet injected at a small angle to the tunnel axis can induce a high-volume longitudinal airflow in the tunnel (Fig- ure 3.7). The phenomenon of air adhesion to the surface of the structure is used to induce secondary air movement in the tunnel in the same direction as the primary flow. The amount of induced flow depends primarily on the nozzle area, dis- charge velocity, angle of the nozzle, and the downstream air resistances. This type of ventilation is most effective with free flow uni-directional traffic flow. Saccardo injectors may operate in a flow induction mode (low tunnel air resistance) and in flow rejection mode (high tunnel air resistance). This means there may be flow reversal at the nozzle position with flow exiting the near portal, whereas jet fans always induce flow from one portal to the other. Saccardo nozzles only produce a limited pressure rise and there- Figure 3.5. Velocity vectors and pressure for jet fans installed close to the portal and blowing out of the portal [38]. Figure 3.6. Tunnel flow rate as a function of center of jet fan distance from the exit portal (example) [38]. 0 10 20 30 40 50 60 70 80 90 100 0 50 100 150 200 250 300 350 Tu n n e l F lo w R at e, P er ce nt o f D es ire d Distance from Center of Jet Fan to Portal, ft Tunnel Flow Rate as a Function of Jet Fan Distance From Portal Tunnel Grid Only Jet Fan Blowing Out of the Tunnel @ 106,000 CFM

31 fore, are suitable for relatively short tunnels unless supported by other ventilation schemes. The Saccardo nozzle ventila- tion scheme may not be as effective for fires near the injection point, possibly up to a few hundred feet downstream, as air recirculation may occur near the fire. Analysis should be per- formed to investigate the performance of the system for these fire locations (See Section 3.3 for Saccardo Nozzle fans). Another form of a longitudinal ventilation system is one with intermediate ventilation shafts: one for exhaust and one for supply producing a push-pull mode or a single shaft push- ing or pulling air from the tunnel. In this arrangement, part of the air flowing in the roadway is replaced by the interaction at the shafts, which reduces the concentration of contami- nants downstream of the shaft. This system is most effective in combination with a jet fan or Saccardo system. This type of longitudinal ventilation system can utilize the Coanda effect, the phenomenon which causes a jet stream to attach to and follow a smooth curved surface, to control the direction of air flow in the tunnel. By providing curved sur- faces into the tunnel, an air supply is blown onto the curved surface and the air in the tunnel will flow in the same direc- tion as shown in Figure 3.8. Reversible tunnel ventilation fans are often used in venti- lation shafts. For emergency ventilation, NFPA 502 requires that “reversible fans shall be capable of completing full rota- tion reversal within 90 seconds” [1]. A longitudinal ventilation system achieves its objectives through the longitudinal flow of air within the tunnel road- way pushing smoke out the portal. It typically does not require air ducts along the tunnel, while a transverse ventilation sys- tem typically does require air ducts. It achieves its objectives by means of the continuous uniform distribution and/or collection of air throughout the length of the tunnel road- way. It is very often that the longitudinal ventilation poses construction costs advantages compared to other ventilation schemes. Figure 3.7. Saccardo nozzle ventilation. Figure 3.8. Longitudinal ventilation system with shafts utilizing the Coanda effect.

32 3.1.4 Transverse Ventilation Transverse Ventilation is a system that is applied for smoke control when the smoke stratification must be kept intact, leaving more or less clean and breathable air underneath the smoke layer to both sides of the fire while extracting smoke from the fire site (applicable to bi-directional or congested uni- directional tunnels). The stratified smoke is taken out of the tunnel through exhaust openings located at the tunnel ceiling. With the classic full transverse ventilation system, air is sup- plied and exhausted all along the tunnel, which requires supply and exhaust ducts along the length of the tunnel and ventilation buildings to house supply and exhaust fans. This system can be found in many old long tunnels. This system can maintain smoke stratification while supplying fresh outside air needed for evacuees. It is recommended that the fresh air jets enter the tunnel near the road surface. The exit velocity and the distances between the individual jets should be small in order to obtain a uniform fresh air supply. A large tunnel fire creates strong longitudinal airflows to supply the oxygen to the fire. With a continuous transverse fresh air supply along the tunnel, this longitudinal air velocity is small, which minimizes the air mixing with the smoke layer. Fresh air jets entering from ceiling openings are unfavorable for smoke control. When air enters the tunnel ver- tically at the ceiling, it destroys the smoke layer, inducing smoke into the air jet and thus mixes smoke into the fresh air layer. It is recommended to position the fresh air outlets near the road surface. Survivors of the Holland Tunnel fire of 1949 reported that the supply air saved their lives providing breathable air to a tunnel environment filled with smoke. One recalls trapped firefighters breathing from the curb-level fresh air inlets during the Holland Tunnel fire. During a fire, exhaust fans in the full transverse system should operate at the highest available capacity and supply fans should operate at a reduced lower capacity (typically ½ to ¹⁄3 of the full capacity) with discharge air velocities not to exceed 3 m/s (591 fpm) when there is a fire. This allows the stratified smoke layer (at the tunnel ceiling) to remain at that higher ele- vation and be extracted by the exhaust system without mixing and allows fresh air to enter through the portals, which creates a breathable environment for both motorist emergency egress and firefighter ingress. Continuous extraction into a return air duct is needed to remove a stratified smoke layer out of the tunnel without disturbing the stratification. One way to control the longitudinal velocity is to provide several independent ventilation sections. When a tunnel has several ventilation sections, a certain longitudinal velocity in the fire section can be maintained by suitable operation of the individual air ducts. By reversing the fan operation in the exhaust air duct, this duct can be used to supply air and vice versa if the fans are reversible. Whatever the means of controlling the longitudinal air velocity are, their operation could be preprogrammed accord- ing to the location of the fire in the tunnel to assure opening of the required dampers and activation of the required fans, which would reduce the possibility of operator error. The longitudinal air velocity generated by a transversely ventilated tunnel is usually maintained below 2 m/s (394 fpm) in the vicinity of the fire incident zone. With higher veloci- ties, the vertical turbulence in the shear layer between smoke and fresh air quickly cools the upper layer and the smoke mixes over the whole cross section; most of the smoke from a medium size fire spreads to one side of the fire (little back- layering) and starts mixing over the whole cross section at a distance of 400 to 600 m (1312 to 1968 ft) downstream of the fire site. This mixing over the cross section can be prevented if the smoke extraction is activated early enough. Additional factors to consider are: • Vehicles standing in the longitudinal air flow strongly increase the vertical turbulence and encourage the vertical mixing of the smoke. • In a transverse ventilation system, the fresh air jets entering the tunnel at the floor level induces additional turbulence, which tends to bring the smoke layer down to the road. This is the reason for the recommendation to throttle the fresh air rate from ½ to ¹⁄3 of the full capacity, depending on the initial fresh air jet momentum. No fresh air is to be injected from the ceiling in a zone with smoke since this increases the amount of smoke spread and tends to sup- press the stratification. The disadvantage of the classic transverse system is that uniform exhaust along the length of the tunnel is not efficient for complete smoke extraction from the fire site and likely spreads smoke along the tunnel. The amount of exhaust air required for effective smoke extraction from the fire site is very large and impractical to achieve for long tunnels. Supply air could disturb and cool the smoke layer, destroying smoke stratification. 3.1.5 Semi-transverse Exhaust System A semi-transverse exhaust system is a modification of the full transverse system with a uniform exhaust air duct along the full length of the tunnel and no supply air. In a fire emergency, the system creates a longitudinal air veloc- ity in the tunnel roadway and extracts smoke and hot gases at uniform intervals. The disadvantage of this system is that uniform exhaust along the full length of the tunnel does not allow for complete smoke extraction from only the fire site, but spreads smoke along the tunnel. The amount of exhaust air required for effective smoke extraction from

33 the fire site is very large and impractical to achieve for long tunnels. There are many combinations of different types of basic transverse and semi-transverse ventilation such as single point extraction, ventilation with intermediate shafts, etc., which implement elements of longitudinal and transverse ventilation schemes. 3.1.6 Single Point Exhaust System Single Point Exhaust system is a modification of the semi- transverse exhaust system (Figure 3.9). The spreading of smoke over the whole tunnel length can be prevented by the large extraction of tunnel air directly above traffic with suitable extraction ports or large openings with remotely controlled dampers. This system works best in conjunction with jet fans, or Saccardo nozzles to localize smoke around openings and to prevent smoke driven by natural factors (such as wind and the tunnel grade) from spreading along the tunnel. In a fire event, single point extraction is achieved at the fire location by remote control of the dampers. To facilitate the maintenance of the equipment, there are systems in use where the large dampers are held by a magnet in closed posi- tion. In the fire zone, the magnets release the damper mecha- nism automatically by command from fire detectors, and the dampers then open by gravity force. However, this system does not allow the openings to close if a smoke plume moves to another place in the tunnel. Once a design fire and its smoke production rate have been chosen, a permissible length over which the smoke may spread has to be fixed. Depending on the type of exhaust openings (fixed or remote-controlled), the extraction capac- ity per unit tunnel length in the fire zone is derived. In gen- eral, an extraction system needs less total exhaust volume when remote single point extraction dampers are installed than with fixed openings. However, it also has to be consid- ered that in the first phase between the start of smoke spread and full operation of the exhaust system with large dampers, the smoke may have spread over a large distance (such as 1 km or more) from the fire site depending on the fire detection time and ventilation system operation design. Therefore, it may not be sufficient to only open a few exhaust openings near the fire but a minimum exhaust rate along the whole ventilation section could be considered. An extraction strat- egy has to be developed depending on the type of tunnel and its ventilation system. The extraction capacity over the length which is permis- sible for smoke to spread must exceed the smoke rate gener- ated by the fire because the openings will not only exhaust smoke but inevitably some fresh air as well and is discussed in Section 3.5. When fans are located close to or in the exhaust air open- ings of the single point extraction system, the exhaust fan temperatures must be properly evaluated in the design. To maintain smoke stratification (see also 3.5), a longitu- dinal air velocity is required to push smoke to one side of the fire. This can be achieved by jet fans or Saccardo (portal) nozzles. However, to activate the required number of jet fans within a few minutes after fire ignition is a complicated con- trol task due to the turbulent nature of tunnel airflow, a large cross sectional area, changing winds and other natural fac- tors. This requires air velocity measurements averaged over the cross section [40] as well as: • Required detection accuracy ± 0.3 m/s (60 fpm), • Short response time, and • Proper positioning of sensors. Figure 3.9. Tunnel with a single point extraction system [39].

34 Also, it is important that no jet fan is turned on in or near a place where there is smoke, as this would immediately destroy the smoke stratification. 3.1.7 Semi-transverse Supply System A semi-transverse supply system is not effective for smoke management as it is unable to maintain smoke stratification or provide smoke extraction at the fire site. It is sometimes used in combination with other systems to achieve longitudi- nal smoke movement along the tunnel. If a fire occurs in the tunnel, the supply air initially dilutes the smoke. If the system is equipped with reversible fans, supply semi-transverse ven- tilation should be operated in reverse mode in an emergency so that fresh air enters through the portals and creates a ten- able environment for both emergency egress and firefighter ingress. Therefore, a reversible supply semi-transverse venti- lation system should preferably have a ceiling supply (in spite of the disadvantages during normal operation) and revers- ible fans so that smoke can be drawn up to the ceiling dur- ing a tunnel fire. The conversion of the duct from supply to extraction must be done as quickly as possible to minimize the spread of smoke. 3.2 Tunnel Ventilation Systems Conditions for Application and Configurations Choosing a tunnel ventilation system is a complicated pro- cess that should consider both normal and fire emergency design strategies. In the past, tunnel ventilation systems were designed based on normal tunnel operation for vehicle emis- sions. With the constant reduction of vehicle emissions over the last 20 years, fire emergency conditions become the most important factors for determining the ventilation system in tunnels up to several miles long and this range expands every year. Many tunnels use the same ventilation system for both normal and fire emergency conditions. When designing and operating such a system, considerations should be given to requirements driven by surrounding infrastructure and pol- lution concentration controls including portal dispersion concentrations. The design objectives of the emergency ventilation system should be to control, to extract, or to control and extract, smoke and heated gases. Emergency ventilation system con- ditions for application are established by NFPA 502: In tunnels with bi-directional traffic where motorists can be on both sides of the fire, the following objectives shall be met: 1. Smoke stratification shall not be disturbed; 2. Longitudinal air velocity shall be kept at low magnitudes; 3. Smoke extraction through ceiling openings or high open- ings along the tunnel wall(s) is effective and shall be considered. In tunnels with uni-directional traffic where motorists are likely to be located upstream of the fire site, the following objectives shall be met: 1. Longitudinal systems a. Prevent backlayering by producing a longitudinal air velocity that is calculated on the basis of critical velocity in the direction of traffic flow. b. Avoid disruption of the smoke layer initially by not oper- ating jet fans that are located near the fire site. Operate fans that are farthest from the site first which are not de-rated by exposure to high air temperatures and do not cause de-stratification and recirculation of smoke in the immediate vicinity of the fire. 2. Transverse or reversible semi-transverse systems a. Maximize the exhaust rate in the ventilation zone that contains the fire and minimize the amount of outside air that is introduced by a transverse system. b. Create a longitudinal airflow in the direction of traf- fic flow by operating the upstream ventilation zone(s) in maximum supply and the downstream ventilation zone(s) in maximum exhaust. Based on these objectives the transverse ventilation or single point extraction system is more applicable for tunnels with bi- directional traffic or for uni-directional tunnels where motorists are likely to be located on both sides of the fire (unmanage- able congested traffic during fire event). Longitudinal ven- tilation is likely to be the choice for uni-directional tunnels with managed traffic downstream of the fire. See Figure 3.3 as an example of the design process for consideration of longitudinal ventilation for the road tunnel ventilation system selection. 3.2.1 Tunnel Length, Geometry, and Grades In the past it was considered that longitudinal ventilation was applicable to short road tunnels with uni-directional traffic only, while full transverse ventilation was commonly used for long tunnels or for tunnels with bi-directional traf- fic. The limitations of longitudinal ventilation systems are often related to the limitations of air velocity in road tunnels. However, these limitations are driven primarily by normal ventilation requirements for concentration of vehicle emis- sions during free flowing and congested traffic and are seldom related to fire emergency conditions. The exception could be emergency response requirements to limit smoke spread along the tunnel for first responder’s ingress.

35 Tunnel height and width and grades are factors to be con- sidered. Tunnel geometry affects the following two areas: • “critical velocity,” which has a direct impact on ventilation requirements, and • ventilation system design. 3.2.1.1 Impact of Tunnel Geometry on “Critical Velocity” The slope of the tunnel has an important influence on the dispersion of the flue gases. In general it can be said that due to the chimney effect, the dispersion velocity of the flue gases increases with the increase in tunnel slope. The longitudinal air velocity’s increase will lead to changes of FHRR and fire growth rate. Section 2.2 discusses critical velocity as the func- tion of fire heat release rate. Equations for “critical velocity” presented in Annex D of NFPA 502 show the relationship between the tunnel height, tunnel grade, and critical velocity. Roadway grade factor is shown in Figure 3.10 and is related to the buoyancy effects. Critical velocity is proportional to the grade factor. Smoke from a fire in a tunnel with only natural ventilation is driven primarily by the buoyant effects of hot gases and tends to flow upgrade. The steeper the grade, the faster the smoke moves and thus the higher the velocity needs to be developed by the ventilation system to overcome the buoyancy effect. One of the key parameters for the fire plume in a tunnel fire is the maximum ceiling gas temperature. The maximum gas temperature in a tunnel fire is mainly related to the effec- tive tunnel height, HRR, and ventilation velocity [1] unless buoyancy is impacted by the fire suppression system. Tunnel Width. The critical velocity in tunnels with aspect ratios of 1 to 3 is approximately independent of tunnel width. When the tunnel aspect ratio (tunnel width to tunnel height) is significantly lower than 1 or greater than 3, the effect of tunnel width may need to be considered. The critical velocity decreases when the width increases. For the aspect ratio lower than one and for high enough HRRs, the critical velocity sig- nificantly increases with tunnel width. The equation that best describes the relationship between FHRR and the tunnel width Wtunnel and fire width Wfire is: tunnel tunnel Runehamar RunehamarHRR B B HRR( )= where [29, 41, 42] 24 1 3 B W Wfire tunnel( )= + This equation allows one to estimate the design fire HRR against the values obtained in the Runehamar tests, consider- ing that QRunehamar = 203 MW, Wfire Runehamar = 2.9m, Wtunnel Runehamar = 7.3m, or in any other tests. Estimates show that for a 15 m (49.2 ft) wide tunnel, the design FHRR is about 100 MW (341 MBtu/hr). 3.2.2.2 Impact of Tunnel Geometry on Ventilation System Design Longitudinal Ventilation. The limitations of longitudi- nal ventilation systems are often related to limitations of air velocity in road tunnels. The maximum air velocity limita- tion is driven primarily by normal ventilation requirements for concentration of vehicle emissions during free flowing and congested traffic, and is seldom related to fire emer- gency conditions. Since the concentration of vehicle emissions increases linearly from the entrance to exit portals with longi- tudinal ventilation, there is a certain tunnel length that would require airflow in the tunnel of velocities exceeding 2200 fpm (11.0 m/s) in order to maintain acceptable air quality levels at the exit portal. Figure 3.10. Roadway grade factor [1] [11].

36 The maximum air velocity limitation for fire emergen- cies is set based on the ability of people to walk in a high air speed environment [1]. Intermediate ventilation shafts could be provided to maintain longitudinal tunnel air velocities below 2200 fpm (11.0 m/s). Intermediate vent shafts along the tunnel (if feasible) could extract vitiated air and supply outside air, which expands the limits of longitudinal ventila- tion length. The emergency response plan could be another limiting factor of longitudinal ventilation scheme. Longitudinal ventilation is impacted by natural factors, such as portal winds, tunnel grades, and portal elevations. Generated airflow should be able to overcome the tunnel chimney effect, sometimes called stack effect, to control tun- nel smoke. Jet fan longitudinal ventilation systems may require greater tunnel height or width to accommodate the jet fans so that they are outside of the tunnel’s dynamic clearance envelope. This envelope, formed by the vertical and horizontal planes surrounding the roadway in a tunnel, defines the maximum limits of the predicted vertical and lateral movement of vehi- cles traveling on the roadway at design speed. Jet fan sizing is usually limited by the space available for installation in the tunnel, which limits the system’s application. Saccardo nozzle ventilation systems require portal venti- lation buildings and significant space at the portal area for Saccardo nozzles installation. This type of system produces a limited pressure rise and therefore is only suitable for rel- atively short tunnels unless supported by other ventilation schemes, such as jet fans and/or exhaust air shafts. Longitudinal systems with intermediate ventilation shafts typically require excavations for the shafts, vent buildings and Saccardo nozzles, or spaces for jet fan installation. A brief comparison of the technical and economic fea- tures of the two longitudinal impulse ventilation systems reveals that: • Jet fans have little or no civil engineering costs for installa- tion, but have significant electrical cabling costs. • Saccardo nozzles (injectors) require expensive civil engi- neering work to construct ventilation buildings and install the fans at the tunnel portals with limited cabling distribu- tion costs. • Routine maintenance or emergency repair work on jet fans will usually mean disruption to the normal tunnel service and availability; this is not the case for Saccardo injectors that can be accessed externally. • Saccardo injectors eliminate electrical cabling within the tunnel; a cost advantage over jet fans. • Jet fans take up headroom in the tunnel ceiling which lim- its the effective kinematic envelope of the traffic, whereas Saccardo injectors are located outside the tunnel making them ideal in tightly spaced tunnels. • Saccardo injectors deliver their thrust at a single point, making them quite vulnerable to local tunnel fixtures. For example, a badly placed traffic sign, LED display, lighting equipment, or any significant blockage near the outlet of an injector will cause a dramatic drop in injector perfor- mance, whereas jet fans are less affected, as their thrust is distributed. • Jet fans are not only exposed to high temperatures while operating in the tunnel but also derated when operating at elevated temperatures in a fire environment (lower den- sity), whereas injectors are both safely outside the fire’s reach as well as immune to thrust reduction by virtue of using fresh air for primary intake. This offers some advan- tage in reliability of Saccardo injectors over jet fans. Transverse Ventilation. It could be perceived that trans- verse ventilation systems have no tunnel length limita- tions. However, they are limited by the air duct’s length to achieve uniform air distribution along the tunnel and by the feasible fan characteristics to produce the required air- flow and to develop sufficient pressure to overcome duct losses. The duct cross sectional area should change along the tunnel length to maintain a uniform ventilation rate. In addition calculating the friction factor of a duct in which there are irregular obstructions to the airflow is challenging. In a vehicular tunnel air duct the term “obstruction” can be used to describe an essential part of the tunnel system, such as water pipes, conduits, cables, etc. The duct length to duct hydraulic diameter ratio is sometimes limited to 300 to achieve the desired airflow distribution in the tunnel [8]. Multiple ducts each connected to independent fans could be constructed to overcome the technical limitations of one single long duct but costs of such designs would be signifi- cantly higher, and in many cases impractical. Tunnel height and width are other factors to be considered for the system selection relying on smoke stratification and protecting tunnel users to both sides of the fire. Tunnel grade is an important factor for the effect of buoyancy and stack effect on smoke and hot gases spread and could either assist or counteract tunnel ventilation. Tunnel Grade. Due to buoyancy effects, natural ventila- tion results in the extensive spread of smoke and heated gases upgrade of the fire, but relatively clear conditions downgrade of the fire. With a steep grade of the tunnel, the chimney effect can rise to significant values. With ducted transverse systems, depending on the num- ber of traffic lanes and tunnel width, airflow can be con- centrated on one side, or divided over two sides. Side walls ducted supply or exhaust may not be practical for over 3-lane tunnels. Air ducts at the ceiling may require deeper tunnel excavations.

37 These relative merits are crucial at the initial concept phase, when deciding on the type of ventilation system for any par- ticular tunnel. 3.2.2 Traffic Conditions Short rural tunnels with uni-directional light traffic with no flammable cargo and no HAZMAT could be justified for natural ventilation. NFPA 502 states that emergency ventila- tion shall not be required in tunnels less than 1000 m (3280 ft) in length where it can be shown by an engineering analysis using the design parameters of the particular tunnel (length, cross section, grade, prevailing wind, traffic conditions, types of cargos, design, fire size, etc.), that the level of safety pro- vided by a mechanical ventilation system can be equaled or exceeded by enhancing the means of egress, the use of natural ventilation, or the use of smoke storage, and shall be permit- ted only where approved by an AHJ. NFPA 502 provides the categories of road tunnels depend- ing on tunnel length and peak hourly traffic (Figure 3.11). Ventilation and other fire life safety systems requirements are based on the tunnel category. For example, a tunnel ventilation system is mandatory for all tunnels over 914 m (3000 ft) long with peak hourly traffic of 2,000 vehicles per hour per lane or less, which has the equivalent requirements to a 300 m (1000 ft) long tunnel with a peak hourly traffic of 6,000 vehicles per hour per lane. The tunnel ventilation system is a conditional mandatory for all tunnels over 245 m (800 ft) long with peak hourly traffic of 2,000 vehicles per hour per lane or less, which is equivalent to a 90 m (300 ft) long tunnel with peak hourly traffic of 5,500 vehicles per hour per lane. A tunnel that is long, has a heavy traffic flow, or experi- ences frequent adverse atmospheric conditions requires fan- based mechanical ventilation. Free flowing uni-directional traffic supports the choice of longitudinal ventilation, while bi-directional traffic or possible congested traffic supports other ventilation schemes (see Figure 3.3 in Section 3.1). In long tunnels with heavy traffic, the use of intermediate venti- lation shafts should be considered. Tunnels with heavy traffic volumes with Flammable Cargo and Heavy Goods Vehicles pose greater risk due to possibilities of a more severe fire event and require more complicated ven- tilation schemes. A single point extraction system supported by jet fans (or other longitudinal ventilation, such as Saccardo nozzles) is considered the most effective in smoke control for high risk tunnels and where vehicles are trapped on both sides of the fire. This system relies on smoke stratification for fires with significant heat dissipation and smoke capture for all fire types, produces low longitudinal air velocities, and does not impact the fire growth and heat release rate as much. However, this system is rather complicated and requires air velocity con- trols on both sides of the fire. It also needs coordination with sprinkler system activation. Additional means for providing protection of ventilation ducts, such as sprinkler protection of vent ducts, may be needed to avoid structural collapse. Full transverse ventilation is used in extremely long tunnels and in tunnels with heavy traffic volume. Short tunnels with light traffic and no flammable cargo and HAZMAT materials are less risky. There are some examples of relatively short rural tunnels with light traffic and bi-directional traffic when longitudinal ventilation system with jet fans was justified based on the risk analysis. French and some other international guidelines allow for longitudinal ventilation for short tunnels with bi-directional light traffic conditions. While most of the U.S. tunnels are uni-directional, many would consider using them as bi-directional during construction or maintenance in the parallel tube. Thus, bi-directional mode is often considered for fire design for uni-directional tunnels. Table 3.2 summarizes the requirements for longitudinal ven- tilation operation in case of fire. Table 2.3 presents a simplified example of tunnel fire safety risk and fire life safety systems needs based on the tunnel length and traffic conditions. Presence of a fire suppression system is also a factor in risk analysis. Other fire life safety means, such as parallel egress evacuation tunnels, other means of egress and etc., should also be considered for risky tunnels in addition to ventilation requirements. EU directive requires for tunnels longer than 3000 m (9842 ft), with bi-directional traffic, with a traffic volume higher than 2000 vehicles per lane and with a control cen- ter and transverse and/or semi-transverse ventilation, the Figure 3.11. Tunnel categories. Extracted from NFPA 502 (2014 edition) [1].

38 following minimum measures shall be taken with regards to ventilation: • Air and smoke extraction dampers shall be installed which can be operated separately or in groups. • The longitudinal air velocity shall be monitored constantly, and the steering process of the ventilation system (damp- ers, fans, etc.) adjusted accordingly. Table 3.3 lists the principles of smoke control with trans- verse ventilation systems recommended by PIARC [15]. For Case B, since the congested traffic would tend to move out of the tunnel, the longitudinal velocity at the extraction zone may be set higher than zero if the information is available to do so. Without remotely controlled dampers, the smoke is extracted through relatively small openings distributed along a long sec- tion. Therefore, the extraction rate near the fire is limited and much lower than using remotely controlled dampers. Smoke stratification, however, cannot be guaranteed but is more likely to occur when the longitudinal flow velocity is below about 1.5 m/s (300 fpm). Whether or not dampers can be controlled in the region of the fire, the longitudinal flow has to be controlled in order to ensure the desired flow velocities up- and downstream of the extraction zone. For this purpose, plausibility checks of the Evacuation phases Firefighting phase Longitudinal ventilation One tube with two-way traffic (not recommended in the U.S. and many other countries) The smoke stratification must not be disturbed: - longitudinal air velocity quite small - no jet fans working in fire/smoke zone Avoid backlayering of smoke: - higher longitudinal velocity - direction of airflow adaptable Longitudinal ventilation with one-way traffic Normal free traffic: Avoid backlayering of smoke: sufficient longitudinal air velocity in the same direction as traffic flow. Congested traffic, or fire at the end of the queue behind an accident, or one tube used bi-directionally: Same as one tube with bi-directional traffic for the two phases. Table 3.2. Longitudinal ventilation operation in tunnels with one-way and two-way traffic. Table 3.3. Transverse ventilation system smoke control strategies [15].

39 flow measurements have to be carried out in order to ensure that the flow measurements are reliable and representative for the actual flow situation in the tunnel. Interference of the airflow control devices with the smoke extraction should be avoided, e.g., by not using jet fans that are situated close to the smoke extraction zone. 3.3 Tunnel Ventilation Fans Utilization and Placement The ability of tunnels to function depends mostly on the effectiveness and reliability of its ventilation system, which is expected to operate effectively under the most adverse envi- ronmental, climatic, and vehicle traffic conditions. A tunnel ventilation system should be robust and designed with redun- dant fan(s) and more than one dependable power source to prevent interruption of service. The prime concerns in selecting the type, size, and number of fans include the total theoretical ventilation airflow capac- ity and pressure required. Fan selection is also influenced by how reserve ventilation capacity is provided either when a fan is inoperative, or during maintenance or repair of either the equipment or the power supply. Tunnel ventilation fans usually require a large volume of air at relatively low pressure. Some fans have low efficiencies under these conditions, so the choice of a suitable fan type is often limited to double width double inlet centrifugal fans, vane axial or jet fans. Factors affecting ventilation fan selec- tion include tunnel geometry, ventilation scheme, operation mode (reversible or not), pressure and air flow requirement. The number and size of fans should be selected by compar- ing several fan arrangements based on the feasibility, effi- ciency, and overall economy of the arrangement. Factors that should be studied include: (1) annual power cost for opera- tion, (2) annual capital cost of equipment (usually capitalized over an assumed equipment life of 50 years for road tunnel fans), and (3) annual capital cost of the structure required to house the equipment (usually capitalized over an arbitrary structure life of 50 years). The number and size of the fans should be selected to build sufficient redundancy and flexibil- ity into the system to meet the varying ventilation demands created by daily and seasonal traffic fluctuations and emer- gency conditions. Ventilation equipment can be a major source of noise in tunnels and therefore, noise limitation is an important factor for fan selection. Acoustic treatment by means of inlet and outlet silencers and/or casings with sound-absorbing lining may be required to reduce the amount of fan noise transmit- ted to the tunnel and to the outside environment. The sound power level of a fan increases very rapidly with increasing tip speed. For a given volume of air, a larger and slower rotat- ing fan will typically be quieter (but more susceptible to stall). Similarly reduction in mechanical noise can be achieved by efficient design of motor couplings, driving gear and adequate stiffening of the casing. Mounting fan equipment on insulated bases will reduce transmission of noise and vibration. 3.3.1 Jet Fans Although jet fans deliver relatively small air quantities at high-velocity, the momentum produced is transferred to the entire tunnel, inducing airflow in the desired direction. Jet fans are normally rated in terms of thrust rather than airflow and pressure, and can be either uni-directional or reversible. Jet fans are classified as impulse systems, since they impart a momentum to the tunnel flow, as the primary high-velocity jet diffuses out. At the startup, this thrust causes the air in the tunnel to accelerate until equilibrium is established between this force and the opposing drag forces due to viscous friction and the additional pressure losses due to tunnel portals, traf- fic, wind, and fire etc. In a jet fan system, this thrust is distrib- uted along the tunnel due to the installation of a series of jet fans along the tunnel. Jet fan sizing is usually limited by space available for installa- tion in the tunnel (Figure 3.13). Typically mounted outside of the tunnel’s dynamic clearance envelope on the tunnel ceiling (above the vehicle traffic lanes), or on the tunnel walls (out- side the vehicle traffic lanes), jet fans are sometimes placed in niches to minimize the height or width of the entire tun- nel boundary. However, niches must be adequately sized to avoid reducing the thrust of the fans. A typical jet fan niche arrangement is provided in Figure 3.12. Figure 3.12. Typical jet fan arrangement in niche (ASHRAE Applications Handbook, [11]).

40 Figure 3.13. Typical jet fan installations in road tunnels.

41 It is advisable to select the largest fan that can be fitted within the allocated space. Larger fans give a higher ratio of thrust to both capital and installation costs than smaller fans. For the lowest operating costs, choose a low speed and/or low pitch angle fan. The ratio of power to thrust is directly related to the fan outlet velocity, so for any given thrust requirements, the higher the velocity, the higher the power consumptions and higher noise levels. However, reducing the thrust for a given fan size increases the number of fans needed and hence the capital costs. For jet fans, additional sound-absorbent material in fan casing, and inlet and outlet silencers should be considered. Any increased head loss caused by a silencer, in some designs, can only be compensated for by increased fan energy con- sumption and hence higher potential noise levels. The design should ensure that during fire emergencies, noise levels in the tunnel do not exceed the levels defined in NFPA 502 and do not interfere with the use of emergency communication sys- tems and operations of first responders. As a note, reversible jet fans are marginally less efficient and slightly noisier than uni-directional fans. Jet fans being installed in the tunnel are subject to high temperatures during the fire events (Table 3.4). In case of a large fire (such as 300 MW [1020 MBtu/hr]) jet fans could be damaged over a distance of up to 300 to 500 m (984 to 1640 ft) downstream of the fire, which should be accounted for in the design. This distance depends on tunnel geometry, fire size and its spread, fan design, ventilation con- ditions should be analyzed using fire simulations tools. The fan damage can be significantly reduced if a FFSS is installed in the tunnel. Jet fans are typically specified to withstand 250°C (482°F) for 1 hour. Some designs require jet fans to be designed for 400°C (752°F) for up to 2 hours and not to fall down during the firefighting phase. British Standards provided data on distances over which jet fans were assumed to be destroyed by the fire [43]. Table 3.5 could be used as an example, but does not replace the calcula- tions required. Nominal FHRR, MW (MBtu/h) Temperature at Central Fans,a °C (°F) Temperature at Jet Fans,b °C (°F) 20 (68) 107 (225) 232 (450) 50 (170) 124 (255) 371 (700) 100 (340) 163 (325) 677 (1250) FHRR = Fire heat release rate aCentral fans located 700
. (213 m) from fire site. bJet fans located 170
. (52 m) downstream of fire site. Table 3.4. Maximum air temperatures experienced at ventilation fans during Memorial Tunnel Fire Ventilation Test Program [11]. Fire size, MW (MBtu/h) Distance upstream of fire, m () Distance downstream of fire, m () 5 (17 ) 20 (68) 10 (32.8) 40 (131.2) 50 (171) 20 (65.6) 80 (262.5) 100 (341) 30 (98.4) 120 (393.7) Table 3.5. Distances over which jet fans are assumed to be destroyed by a tunnel fire (BD 78/99) [43].

42 French guidance provides smoke temperatures at vari- ous distances (CETU, 2003). This is reproduced in Table 3.6, which could be used as an example, but does not replace the calculations needed. The following should be noted: • Jet Fans are not as efficient as axial fans operating in a ducted system. However, low capital cost and simplicity of instal- lation and maintenance may justify their use. • Fan performance is highly impacted by fan installation, spac- ing between fans, signs and other equipment located in the air stream. • Fans should be provided with anti-vibration mounting which should be fail safe. Safety chain installation is a good practice to secure fans. • Water entering the jet fans from any source needs to drain out. • Sealed-for-life bearings should be considered. • Designs should consider redundancy of jet fans due to loss of power or maintenance related issues. 3.3.2 Saccardo Nozzle Fans Injection longitudinal ventilation uses externally located supply fans to inject air into the tunnel through a high-velocity Saccardo nozzle. The Saccardo nozzle functions on the prin- ciple that a high-velocity air jet injected at a small angle to the tunnel axis can induce a high-volume longitudinal air- flow in the tunnel. The amount of induced flow depends pri- marily on the nozzle area, discharge velocity and angle of the nozzle, as well as downstream air resistances. Air veloc- ities discharged into the tunnel through Saccardo nozzles are similar to the jet fans discharge velocities. Saccardo fans (injectors) are located outside the tunnel in the ventilation buildings making them ideal in tightly spaced tunnels and easy for maintenance (Figure 3.14). Fans are not subject to high temperature exposure. Fans can be located horizontally or vertically depending on the vent building design. Fresh air intake for the supply fans should meet security require- ments. Special considerations should be given to the noise developed by the Saccardo fans due to high-velocity pressure requirements. Consideration should also be given to the geometry of the injector aerodynamic design to reduce pressure losses. Fans used in transverse ventilation schemes are located in the vent buildings and can be either centrifugal or axial flow and installed horizontally or vertically. The type of fan is determined by the required airflow and pressure for fire emergency and normal operating conditions and the available space in the tunnel configuration. Fans used in tunnel ventilation should be constructed to withstand the maximum pressure and temperature anticipated. Flow revers- ibility is frequently required in tunnel ventilation systems. Most tunnel ventilation fans are driven by electric motors. The fan motor selection is based on the full load horsepower requirements, fan speed and the starting characteristics. Downstream distance 10m (33 ) 100m (330 ) 200m (656 ) 400m (1,312 ) Light vehicle fire 250°C (482°F) 80°C 40°C 30°C HGV fire 700°C (1,292°F) 250°C (482°F) 120°C (248°F) 60°C Tanker fire >1000°C (>1,832°F) 400°C (752°F) 200°C (392°F) 100°C (212°F) Table 3.6. Smoke temperatures near the ceiling, with airflow close to critical velocity (CETU, 2003) [43]. Figure 3.14. Longitudinal ventilation with Saccardo nozzle.

43 Several fans installed in parallel typically serve a single ven- tilation duct due to high airflow requirements and relatively low pressure needed. Fans operating in parallel should be of equal size and have identical performance curves. Actual airflow capacities can be determined by plotting fan perfor- mance and system curves on the same pressure-volume dia- gram. If airflow is regulated by speed control, all fans should operate at the same speed. If airflow is regulated by dampers or by inlet vane controls, all dampers or inlet vanes should be set at the same angle. For axial flow fans, blades on all fans should be set at the same pitch or stagger angle. Fans selected for parallel operation may be required to operate in a particular region of their performance curves so that airflow capacity is not transferred back and forth between fans. This is done by selecting a fan size and speed such that the duty point total pressure, no matter how many fans are operating, falls below the minimum total pressure character- ized by the bottom of the stall dip or unstable performance range. Fans in the exhaust air duct are exposed to a mixture of very hot air from the immediate surroundings of the fire, which could be diluted by cooler air further away from the fire. The fresh air/smoke mixture is also subject to thermal exchange with the duct walls before reaching the fans. This mixture of hot and cooler air then travels in the duct and gets cooled down further (see Table 3.4). A fire resistance of the fans to 250°C (482°F) could be considered sufficient for most fire events, but needs to be checked for the design fire sce- nario. There can be residual thermal expansion of the gases passing through the fans. This phenomenon must be taken into account when establishing the thermal resistance criteria and capacities of the extraction fans. The thermal resistance of the fans must ensure that the extraction of the hot smoke is possible with any configuration. When the fire location is relatively close to the extraction point, the exhaust temper- ature may be significantly higher than 250°C. This was the reason for increasing the thermal resistance of exhaust fans in Austria and Germany to 400°C for 90 minutes, and in France to 400°C for 120 minutes depending on the location of the fans relative to the tunnel traffic [16]. 3.3.3 Centrifugal Fans Centrifugal fans can have either single or double inlet impel- lers with radial, forward curve, or backward curve blades. The most commonly used is the double width double inlet backward curve fan because of its relatively smaller space requirement, greater efficiency and non-overloading char- acteristics at selected speed. All centrifugal fans due to their designs require a larger amount of space than axial fans of the same duty. Some advantages over axial fans are that centrifu- gal fans are more efficient and less noisy and provide higher pressure capability. The performance of centrifugal fans is affected by: • variations in speed, • outlet damper control, and • variable inlet guide vanes. 3.3.4 Axial Fans These fans can be vane axial or tube axial with single stage or multistage construction. These fans have the ability to handle extremely large quantities of air and are frequently used in tunnel ventilation design. Axial fans can be mounted horizontally or vertically in ventilating shafts thereby reducing space requirements relative to centrifugal fans. Axial flow fans are often used in ventilation of major road tunnels. Capacities are often in excess of 100 m3/s. The large diameter fans are located in fan rooms with connecting shafts supplying and extracting air to or from the tunnel section or its full length depending on ventilation system design. An axial flow fan is one in which air passes between aero- dynamically shaped blades to enter and exit axially to the direction of rotation. Reverse flow may be achieved by revers- ing the direction of the rotation of the motor. System effects should be considered in the pressure loss calculations when designing the fan and duct configuration for all fan types and for all ventilation schemes, particularly the fan inlet and outlet conditions. Tunnel ventilation fans can have multiple speeds controlled by adjustable speed drives, or two- or three-speed motors or multiple motor drives. Axial fans can be produced with controllable blade pitch in motion (variable-pitch blades) for control of airflow and thrust. Fans can be aerodynami- cally stabilized by means of anti-stall ring which introduces on each side of the impeller, providing stable flow conditions and continuously rising fan characteristics in both flow direc- tions. When in the stall region, the separated and highly turbu- lent flow is removed from the main flow annulus and entered into the stabilized peripheral ring-shaped duct just upstream of the impeller blades. NFPA 502 requires that “tunnel ventilation fans that are to be used in a fire emergency shall be capable of achieving full rotational speed from a standstill within 60 seconds. Revers- ible fans shall be capable of completing full rotational rever- sal within 90 seconds” [1]. The emergency ventilation system should be capable of reaching full operational mode within a maximum of 180 seconds of activation. Fans could be acti- vated sequentially based on fire zones. Fan motors typically have adjustable speed drives, soft starters, or direct on line motor starters. The selection of the motor controller will affect the sequencing times, startup times, inrush current,

44 motor durability, generator sizing, flexibility of the system, efficiency, etc. 3.4 Effects of Ventilation on Tunnel Fires and Fire Sizes Ventilation has an influence on fire development, but it does not always conform to expectations; this influence depends on the location of the fire origin and the sufficiency of air [44]: 1. Due to increased ventilation, the development of a car fire can be slowed if the fire ignites at the front of the car. This is in contrast to the accepted view of supposed accelerated development due to ventilation. 2. The influence of increased ventilation on the observed fire behavior depends on the ignition location. Note that the majority of fires begin in the engine compartment (i.e., at the front). 3. Under the influence of a high ventilation velocity, fire devel- opment accelerates for a covered load at a rate 2–3 times faster than an uncovered load. The fire size is also 20–50% higher due to a high ventilation speed. Ventilation could cause flame deflection, which leads to the chance that the fire might spread to other vehicles and threaten the integrity of the tunnel structure on a larger sur- face, assuming the ventilation cooling effect and reduction in radiation at the source are insignificant. In most cases, mechanical ventilation will lead to complete combustion (the fire to burn fully). Thus, the total duration of the fire (if not timely extinguished) will be limited to the time to complete combustion. It is understood that there could be a negative effect of ventilation as forced ventilation may cause signifi- cant flame deflection and fire spread by convection. The increase in peak HRR and fire growth rate, due to increased velocity, is the result of more effective heat transfer from the flames to the fuel surface. In some cases, it results in a more effective transport of oxygen into the fuel bed, which enhances the mixing of oxygen and fuel. The oxygen unlimited larger pool fires (in wide tunnels) dominated by radiation heat flux from the flames are less affected by ventilation than small fires dominated by convective heat transfer from the flame volume. Starting a ventilation system when the fire has been ongo- ing for some time in a tunnel with high vehicle density and oxygen deficiency may offer some risk due to supplying oxy- gen to the fire and possible fire spread. However, as ventila- tion cannot reach its full operating mode immediately, the risk may be justified. Influence of ventilation rate on fire growth rate from the Benelux and Runehamar fire is presented in Table 3.7. Tests indicated that the fastest fire growth may occur at about 3 m/s airflow velocities. Both higher and lower ventila- tion rates may result in slower growth fires. These observa- tions were made on the basis of only a few experiments. More research is needed to confirm (or otherwise) the validity of these conclusions [45]. The flame spread rate in a tunnel fire is proportional to the ventilation velocity. However, there is some evidence of the decrease effect under highly ventilated conditions (the blow- off effect). The ventilation conditions are important for the combus- tion gases production. In an under-ventilated fire situation the yield of major toxicants is greater, compared to well-ventilated conditions. The main effects of longitudinal ventilation are an increase of the growth rate of the fire and an increased dilution of gases. Under high ventilation rate conditions only downstream flame exists, while under low ventilation condi- tions both upstream and downstream flames exist. Under low ventilation conditions, the total flame length increases with decreasing ventilation velocity despite that the downstream flame length is approximately invariant. The maximum total flame length is obtained when there is no ventilation in the tunnel, and it is approximately twice the downstream flame length in tunnel fires under high ventilation. 3.5 Fire Smoke Stratification and Length of Stratification and Its Impact on Emergency Ventilation This section applies to cases where stratification of smoke exists. During fires with a significant heat generation rate, the hot smoke flows upward due to the buoyancy force and impinges on the tunnel ceiling and then flows along the tun- nel ceiling longitudinally. A substantially smoke-free layer is formed below the smoke layer. As the smoke travels along the ceiling, the smoke temperature decreases rapidly with distance mainly due to heat loss to the tunnel structure. This indicates that the thermal pressure also decreases with dis- tance [30]. Therefore, smoke stratification becomes more Ventilation Rate Growth Rate Less than 1 m/s (200 fpm) About 5 MW/min (17 MBtu/hr/min) About 3 m/s (600 fpm) About 15 MW/min (51 MBtu/hr/min) About 6 m/s (1200 fpm) About 10 MW/min (34 MBtu/hr/min) Table 3.7. Influence of ventilation rate on fire growth rate [45].

45 and more difficult to maintain as the distance from the fire increases. Thermal pressure tends to maintain smoke stratifi- cation but the inertia force tends to destroy it. Due to a com- plex process of mass and heat exchange, the smoke is gradually cooled and mixed with the air. After a period of time, both upstream and downstream sections of tunnel can be completely filled with smoke. Therefore, stratification is a temporary phe- nomenon unless it is maintained by appropriate ventilation including extraction from the ceiling and control of longitu- dinal air velocity. Non-dimensional parameters that describe the balance between these two forces are the Richardson number (Ri) and local Froude number (Fr); both of them correlate the buoyancy force with the inertia force and indicate the sta- bility of the smoke layer, but in an inverse relationship to each other. As the Richardson number increases, the smoke stratification becomes more stable. As explained in Chapter 2, when Fr ≤ 0.9, there is severe stratification in which hot com- bustion products travel along the ceiling. The gas tempera- ture near the floor is essentially ambient. This region consists of buoyancy dominated temperature stratification and could be used for egress [30]. When the air temperature at the tunnel ceiling is signifi- cantly higher than at the level where the fire starts, the upward movement of the smoke plume may cease due to the lack of buoyancy and additional stratification may occur. The fire smoke layer and stratification may depend on the fire size (heat release rate), tunnel type, tunnel geometry, and longitudinal ventilation flow. When the longitudinal ventila- tion is gradually increased, the stratified layer may gradually dissolve. The need to maintain the smoke stratification upstream to the fire leads to the concept of critical velocity in emergency ventilation. In addition, recent research has provided the results on the length and duration of smoke stratification. In tunnels with natural ventilation and low air velocities (0–1 m/s [0–200 fpm]) the stratification of the smoke is usu- ally in the vicinity of the fire source. The backlayering length of the smoke is relatively long, and in some cases, the smoke travels nearly uniformly in both directions. When the veloc- ity increases and is close to about 1 m/s (200 fpm), the smoke upstream of the fire is inhibited by the ventilation and pre- vented from spreading further. The length of this backlayering smoke layer from the fire site could be in the order of 25 times the tunnel height. At longitudinal air velocities of 1 to 3 m/s (200 to 600 fpm), the stratification in the vicinity of the fire is strongly affected by the air velocity, especially at the higher velocities. The back- layering length could vary. For a ventilation velocity slightly lower than the critical velocity, smoke backlayering and good stratification exist upstream of the fire. However, the smoke stratification downstream becomes worse. At high air velocities, over 3 m/s (600 fpm), the stratification of the smoke downstream usually disappears and no back- layering exists upstream of the fire, which can be observed in tunnels with longitudinal ventilation. Cold air at high velocities could bypass the fire plume without mixing with smoke. The backlayering length, Lb (m), is defined in numerous publications [1, 3, 30, etc.] as the length of the smoke back- layering upstream of the fire when the ventilation velocity is lower than the critical velocity (Figure 3.15). In a longitudinally ventilated tunnel, a fresh air flow with a velocity not lower than the critical velocity at the designed HRR is created to prevent smoke backlayering, and therefore, the tunnel is free of smoke upstream of the fire site. Smoke stratification downstream of the fire will likely be destroyed as the longitudinal air velocity is too high. For this reason, H. Ingason with citation to O. Vauquelin and D. Telle proposed using a new term, “confinement velocity” [30] (see Glossary). For additional discussion on using such a velocity in the attempt to control back-layering and, at the same time, to preserve certain stratification refer to [30]. The backlayering length increases with the HRR for low HRRs and is nearly independent of HRR and dependent only on the ventilation velocity at higher HRRs. A small change in velocity will result in a greater change in backlayering length. Smoke extraction ventilation shown on Figure 3.16 [30] suggests that the spreading of smoke over the tunnel length can be prevented by a large extraction of tunnel air directly above the traffic with suitable extraction ports [3]. Used with permission of Springer. Figure 3.15. Schematic of backlayering of smoke in a tunnel fire [30].

46 The exhaust ventilation system should be sized to extract the smoke flow rate, which depends on the FHRR, gas tem- perature, tunnel geometry, and ventilation. When the smoke layer is relatively shallow, a high extract rate at any point may lead to “plug-holing”, where some air is extracted from below the smoke layer as opposed to the smoke itself. The traditional method always tries to avoid plug-holing, that is, the smoke flow is extracted slowly with multiple extracted points (Figure 3.16). The extract rate from one point should not exceed: 305 1 2 M g h z T T T Ts O O s [ ]( ) ( )= β − −  where M = extract rate, kg/s; g = acceleration due to gravity 9.81 m/s2; b = a numerical factor of 2 where the extract point is near the wall and 2.8 where the extract point is distant from the wall (limited experimental data); h = the tunnel height; z = the height of the layer interface (smoke free height); TO = the absolute ambient temperature, (K); TS = the average (absolute) smoke layer temperature, (K). If the two extract openings are close together, the flow around them will basically be the same as if they were one point. NFPA 92 provides guidance for the minimum separa- tion between extract points. Incoming airflows with a sufficiently large ventilation veloc- ity should be supplied from both sides of the smoke extraction systems to successfully prevent smoke from spreading further. In most practical applications, the main objective is to only partly remove the smoke flow. As the smoke flow cannot be completely controlled within a small region between vents and the fire source, it spreads to a much larger region. The fire location is determined by the fire detection sys- tem (see Chapter 4). The extraction ports or openings are typically equipped with control dampers. The exhaust near the fire site is achieved through the activation of the control dampers (opened or kept open) at the extraction opening or openings nearest the fire site upon detection and confirma- tion of the existence of a fire within the tunnel [1]. (See Fig- ure 3.9 in Section 3.1.) Meanwhile, the other control dampers of the remainder of the extraction openings should close or remain closed, therefore allowing the system to maximize the exhaust air flow adjacent to the fire site [1]. Critical or at least confinement velocities to both sides of extraction zone should be maintained. Some international guidelines recommend those velocities to be 1.5–2 m/s (300–400 fpm) [39]. For fires with HRR up to 500 MW H. Ingason et al. [30] recommends higher velocities of 2.9 m/s upstream of the fire source and 3.8 m/s downstream of the extraction open- ing with the critical extraction mass flow rate for confining the smoke to a small region between the fire and the extrac- tion vent, m˙ex, be estimated using the equation: 2m u Aex O C = ρ where rO = the outside air density; uC = the critical (or at least confinement) velocity; A = the tunnel cross section area. The number of and spacing between exhaust openings for a single point exhaust system depends on fire size, design vol- ume flow rate for each opening and on several other factors: • The accuracy of the fire detection system and the fire loca- tion relative to the exhaust openings. At least two, or prefer- ably three or more openings, should be opened to effectively extract smoke; • The activation time of the ventilation system. The smoke may have spread over 1 km or more from the fire site depending on the fire detection and ventilation system operation design; • The presence of a longitudinal ventilation system and con- trol system (including air velocity sensors in the tunnel) to control air velocities to both sides of the extraction openings; • Zone of tenability, tunnel geometry, egress features, etc. As the application of the tenability criteria at the perimeter of a fire is impractical and the zone of tenability should be Used with permission of Springer. Figure 3.16. Point extraction ventilation [30].

47 defined to apply outside a boundary away from the perimeter of the fire, it is practical to consider a smoke extraction zone of not less than 90 m (300 ft) long with at least 3 exhaust openings spaced at least 30 m (100 ft) apart. Note that smoke stratification may not happen when the fire heat generation rate is insignificant and the smoke pro- duction rate is significant. In such situations, tenability can be lost due to the ‘non-stratified smoke.’ This could be espe- cially dangerous before fire life safety systems are activated. As the heat generation rate is low, the smoke temperature would be close to ambient, while smoke could be very toxic and impair visibility. Such smoke could be managed by the tun- nel ventilation system with low critical velocity numbers, and smoke management could be achieved with airflow driven in the desired direction. The tunnel ventilation system shall be evaluated for this scenario.