For buildings and industrial plants, detailed design guidelines and technical standards covering automatic fire-fighting systems have been available for many years. However, for road tunnels, this technology is only used in a limited number of projects. If the use of fixed firefighting systems is not required by national standards, PIARC recommends that the applicability be assessed on a project-by-project basis with quantitative risk and cost-benefit analyses.

The attractiveness of fixed fire-fighting systems will increase if the resilience of road connections becomes a more important economic factor. In summer 2016, there were two similar tunnel fires in Austria and Germany. In the Austrian Gleinalm Tunnel, a converted bus burnt out completely and the high temperatures involved in the fire destroyed a section of the tunnel ceiling. As a result, the tunnel was closed for six weeks for refurbishment.

In the German Jagdberg Tunnel, a truck fire occurred following a collision. The fire did not develop into a large fire, but was suppressed by a foam-based fire suppression system. The tunnel was closed for the removal of the truck, but was re-opened only seven hours later. These events did not lead to the installation of fixed fire-fighting systems in other tunnels in Austria and Germany, although they may serve as anecdotal evidence for an improved resilience.

MINIMISING FIRE SPREAD

Fixed fire-fighting systems are intended to contain a vehicle fire in a tunnel as early as possible to prevent the development of a larger fire. Overarching protection goals are for both personal and structural protection. The objectives can be qualitatively stated as

? Inhibiting the development of fire or reducing the heat release rate

? Reducing smoke production

? Preventing the fire from spreading to other vehicles

? Improving operating conditions for emergency Services

? Protecting the tunnel structure by cooling

? Maintaining the operating conditions of other safety systems, e.g. preventing the overheating of the ventilation system.

There are numerous studies on the application of water mist or deluge systems on tunnel fires. Their effect on tunnel safety facilities usually refers to reduced temperature exposure, water supply and drainage. The integration of fixed fire-fighting systems into the sophisticated and complex safety systems of road tunnels has been studied less than their effectiveness against fire development. The interaction of the fixed fire-fighting system with other safety provisions in the tunnel such as communication, egress and ventilation is not yet adequately investigated. The safety engineer cannot draw from established design practice.

Figure 1 shows the mutual influence of water mist and tunnel ventilation. In uncoordinated ventilation operation, e.g. one-sided smoke control, water mist will destroy smoke stratification. It acts as a resistance to the airflow. And the airflow may cause displacement of water mist away from the fire. In coordinated operation, e.g. minimum airflow in the tunnel, water mist supports airflow control. And the ventilation supports the dispersion of water mist at the fire. Still, the cooling effect may destroy smoke stratification. This has to be addressed in the egress concept. The schematic is simplified, as objectives of the safety concept may change during an event to support egress or firefighting.

Water mist has a significant impact on tunnel aero- and thermodynamics, e.g. on airflow resistance, flow rates, temperature and smoke stratification as well as smoke control. From our study1, we obtain a better understanding of the coordination required between water mist application and ventilation operation. The objective was to evaluate and, if possible, quantify the interactions between fixed fire-fighting systems and tunnel ventilation in design and operation. We isolated physical phenomena observed during water mist operation, cf. Figure 2. These can be described using classic fluid- and thermo-dynamics. While this represents a simplified model, it allows the identification of the most important parameters.

As a typical fixed fire-fighting system, we assumed water mist with a droplet diameter of 0.5mm and an area related flow rate of 6mm/min over a 75m length in a two-lane road tunnel. The design fire was defined as a truck fire with a nominal heat release rate of 30MW. The design ventilation airflow is 3m/s. The report1 includes the equations that allow a similar analysis for other parameters.

PRESSURE RESISTANCE

? The injection of water spray into a longitudinal airflow causes an additional pressure drop along the tunnel due to the acceleration of the droplets. This pressure drop shall be taken into account in the design of the tunnel ventilation system.

To estimate the momentum transfer from the airflow to the water droplets, a force balance is written for a single droplet. With some other simplifications, we obtain the equation of motion for the droplet. If the fire suppression agent is dispersed in small droplets, the initial momentum of the droplet is quickly lost to the airflow. The movement of the droplet is then described by the longitudinal airflow and by the droplet’s settling velocity. Displacement of droplets by the longitudinal airflow is not a significant problem, as long as the airflow velocity is not excessive and the droplets are not much smaller than 0.5mm.

The total flow rate is accelerated to the velocity of the airflow. In case of longitudinal ventilation, the additional jet-fan thrust required to meet the design airflow can be determined using the water-mass flow, air density, and airflow velocity. This is only an estimate since the interaction of fire and water mist can result in greater flow resistance. In our example, the thrust required to balance the pressure loss is 225N – about half the thrust of a 22kW jet fan.

THROTTLING EFFECT

? The flow resistance of the fire shall be taken into account in the design of the ventilation system, even if the pressure drop can only be estimated or determined numerically.

The flow resistance of the design fire in longitudinal flow (aka ‘throttling effect’) is not negligible compared to the other flow resistances in the tunnel. Currently, there is hardly any information available on the combined effect of water mist systems and the aerodynamic resistance of fires. Several studies have been published on fires in longitudinal ventilation without water mist. More detailed information on the throttling effect can be found in reference 22.

DESIGN AIRFLOW

? Longitudinal ventilation systems shall not be designed for reduced flow velocity when using fixed firefighting systems. The ventilation operation shall be defined in a safety concept that incorporates water mist, ventilation operation, egress and firefighting.

In ref1, several models are described for the heat transfer from the tunnel air to the droplet and for droplet evaporation. Evaporation and heat transfer models are then combined with the equation of motion. In the vicinity of the fire (at a temperature of 300°C), droplets up to 0.5mm in diameter evaporate completely before they hit the road surface. Even outside the hottest zone, the water droplets have a cooling effect on the fire fumes by absorbing heat through convection. With various model assumptions, it seems plausible that the use of a fixed fire-fighting system may reduce a nominal 30MW fire to an effective heat release rate of 15MW. This reduction in heat release rate, however, only has a rather small effect on the calculated critical velocity for smoke control.

In tunnels with considerable longitudinal gradient, when determining fire buoyancy, the application of fixed fire-fighting systems can be used to justify a reduction in the required jet-fan thrust. Still, without a significant change of the ventilation concept (e.g. smoke extraction vs. longitudinal ventilation), the cost saving potential is small. In this consideration, the failure modes of safety systems must be taken into account. Even in case the water mist system fails or is activated late, the ventilation system should be able to control the smoke. We conclude that there is little cost-saving potential from reduced flow losses by cooling the fire fumes.

The application of water mist will most likely destroy the smoke and temperature stratification in the tunnel. If temperature stratification is destroyed, a very small ventilation velocity is sufficient to control smoke propagation. Without temperature stratification, smoke back-layering is not expected. But the visibility downstream of the fire is greatly reduced. This shall be considered in the safety concept, which shall include water-mist operation, ventilation operation, egress and firefighting.

AIRFLOW VOLUME

? The volume of water vapour released in the tunnel does not have to be taken into account in the ventilation design, since this increase in volume is offset by the decrease in volume due to cooling.

The evaporation of water droplets leads to a density decrease and a volume increase of the airflow. On the other hand, cooling of hot fumes increases the density and reduces the volume.

The conversion of water to water vapour mainly takes place in the hot region close to the fire. The potential of water vapour to add to the longitudinal flow seems significant. At 100°C, 1kg of water expands into 1,672 litres of water vapour. However, water vapour at that temperature only occurs in the immediate vicinity of the fire. For the volume balance, we consider the flow rate downstream at 50°C, which is a design requirement of the Austrian RVS for water mist in road tunnels. The downstream flow rate with water mist at 50°C is slightly less than the flow rate without the water mist. In summary, evaporation and cooling reduce the airflow rates that have to be controlled by the ventilation system.

COOLING CAPACITY AND HUMIDITY

? The tunnel air downstream of the fire or water-spray section is almost completely saturated with humidity. Because of further cooling, water precipitates by condensation.

Downstream of the fire, the airflow temperature is significantly less than 100°C, e.g. 50°C according to the Austrian RVS. At this temperature, the amount of heat that can be absorbed by water evaporation is limited by saturation of the humid air. A large part of the extinguishing water remains in the vicinity of the fire or precipitates to the tunnel walls. It is disposed of by the tunnel drainage system.

In our example, a maximum of 15.5kg/s of water can be absorbed by the airflow. This includes water as a combustion product and the humidity of the ambient air. In the example, this is about 20% of the water mist flow rate. However, this is sufficient for heat absorption of 35MW. By reducing the longitudinal flow, the absorption capacity and thus the cooling effect by evaporation is reduced accordingly.

The capacity of the water mist to reduce the heat release rate is limited. Numerical simulations of water mist systems on concealed fires do not show a continuous reduction of the heat release rate with increased water flow rate. Water mist is very effective for low water-flow rates. At some point, a further increase of the water supply will have little effect on the heat release rate. A quick activation of a small water flow-rate may be very effective for tunnels where water is scarce. And a precise detection of the fire location is the key.

With increasing distance from the fire, the saturated air cools down. This leads to condensation on the road surface and the tunnel wall. The evaporation enthalpy is released as heat into the concrete. The condensed water is disposed of by the tunnel drainage. Any equipment installed in the airflow (in the tunnel, ventilation duct, ventilation outlet etc.) has to be selected for 100% humidity including condensation.

WATER MIST ACTIVATION

? Fixed fire-fighting systems shall be activated as soon as possible after fire detection to inhibit fire development.

The cooling of the fire by fixed fire-fighting systems limits the pyrolysis of solid fuel or the evaporation of liquid fuel. This inhibits fire development. In a developed fire, only a small proportion of the heat release rate is required for pyrolysis or vaporisation. Water mist systems have a smaller effect on the heat release rate in a developed fire.

When determining the activation timing, a tradeoff arises between the effects of the fixed fire-fighting system on the fire (requiring activation as early as possible) versus the chances of escape for tunnel users (requiring delayed activation). If people are in the tunnel section with water mist and smoke, their chances of escape are significantly reduced. In addition to the ventilation concept and fixed fire-fighting system, the egress concept and communication to tunnel users must be considered when determining the activation time. There is need for further research on his topic.

VALIDATION IN FDS

The study1 included a series of numerical simulations, using Fire Dynamic Simulator FDS. In the simulations, different flow configurations were tested in order to separate and quantitatively assess individual influencing variables. This goal was only partially achieved. Without fire, the pressure losses obtained by simulation agree very well with the analytical models, e.g., for the pressure drop due to water-mist injection. However, the simulations with fire result in pressure losses that deviate significantly from the results of the analytic approach.

In 2018 to 2019, FDS has been used to further investigate the pressure distribution around tunnel fires2. It was not possible to define a model with a plausible grid-independent pressure profile over the tunnel length. We discussed this in the FDS User Forum with other users and developers of FDS. Since December 2018, the problem of pressure calculations in fire simulations for tunnels has been treated as Issue #7040 by the FDS developers. Our simulations were carried out with FDS versions 6.7.0 and 6.7.1. Since no plausible results could be obtained and as a consequence of the discussions, we decided not to use FDS for this purpose until the issue was resolved. At the time of writing this article, the most recent version of FDS is version 6.7.6. While Issue #7040 remains open, the release notes describe a new feature to improve simulations in long tunnels.

CONCLUSIONS

Fixed firefighting systems have an impact on tunnel ventilation design, specification and operation. The design of longitudinal ventilation shall include the additional pressure drop from water mist injected into the airflow. Smoke extraction systems shall not be designed for an increased extraction flowrate. The specification of any ventilation equipment in the airflow downstream of the water mist section must allow for 100% relative humidity and for water condensation.

More importantly, the operation of the emergency ventilation and the activation of the water mist system must form part of a coordinated safety concept with the objectives of:

? Optimum conditions for the egress of tunnel users.

? Support of emergency services access, and

? Protection of the tunnel structure.