Tests show that automatic sprinkler systems in tunnel fires are only suitable for tunnels with low velocities. It is recommended that automatic sprinkler systems be used in tunnels with transverse ventilation or in bi-directional tunnels, rather than in longitudinally ventilated tunnels with high velocities.
Interest in tunnel fire safety has increased dramatically in recent years owing to numerous catastrophic tunnel fires and extensive media coverage. New technologies, such as water sprinkler systems, have been developed to improve the fire safety in tunnels.
Yet it is generally accepted that automatic water sprinkler systems (AWSs) can be adversely affected by ventilation during tunnel fires, but this viewpoint has not been fully investigated. This was why, before adopting a position on the matter, we decided to investigate how AWS systems function in various longitudinal ventilation flows. The system widely accepted today is the deluge system, where one or two zones of sprinklers are activated in the event of a fire.
The heat from a fire plume rises vertically, if there is no wind in the tunnel. However, when there are longitudinal air flows, the heat and flames will be deflected and accumulate downstream of the fire. For example, it is possible that sprinklers are activated too remote from the fire to deliver water effectively.
We also wanted to address whether too many sprinklers might activate, thereby exceeding the water delivery capacity of the system. A deluge system activates one or two zones and is not sensitive to the longitudinal air flows. However, a fully AWS system in a tunnel is assumed here to cover the entire length, without being split into different zones. When individual sprinkler bulbs activate over a large area, the AWS system will fail as it cannot deliver the amount of water needed to control the fire.
This study aims to find out why such systems fail. A scale model was used, as it is cost-effective and can, if well designed, provide vital, reliable information.
Test Series
In all, 28 tests were carried out in a 1:15 scale tunnel model, which was 10m long, 0.6m wide and 0.4m high, see Figure 1. The scale model corresponds to a tunnel that is 150m long, 9m wide and 6m high. The fire spread between wood cribs spaced 1.05m apart (15.75m in full scale) was tested, as were the effect of different ventilation velocities and water flow rates on the activation of nozzles. The heat release rate, fire growth rate and fire spread were measured. See SP Report 2011:31 for further information about the tests.
Activating auto sprinklers
The tests show that the heat release rate at activation of the first nozzle (sprinkler head) increases linearly with ventilation velocity. The activation temperatures are about 206oC and 109oC for a link temperature of 141oC and 68oC, respectively.
The tests also show that the location of the first activated nozzle depends mainly on the ventilation velocity. The other nozzles in the measured region will be activated shortly after the activation of the first nozzle.
Fire suppression
During a tunnel fire, the suppression resulting from an AWS system can be divided into two modes: the fire is controlled by suppression through surface and gas cooling near the fire, while downstream it is controlled by gas cooling.
At the preliminary stage, few nozzles close to the fire source are generally activated shortly after ignition to sufficiently suppress the fire development and cool the hot gases above the fire. The nozzles activated at this stage play the most important role in the fire development. At the second stage, many nozzles downstream of the fire may be activated to cool down the hot gases.
Collapse of the system
The activation range is directly related to the longitudinal ventilation velocity and the water flow-rate of the automatic sprinklers, as shown in Figure 2.
Failure of an AWS system during a fire is defined as it having an activation range equal to or greater than 100m at full scale, i.e. a dimensionless activation range of about 15 (15 times the tunnel height). High ventilation and low water flow-rates can result in the failure of an AWS system during a tunnel fire. Note that the water flow-rates tested corresponds to 16.5mm/min, 20mm/min and 25mm/min at full scale, respectively. Note that nozzles having such flow-rates can suppress the fire efficiently as part of a deluge system.
Under the specified water flow-rate tests, longitudinal ventilation plays the most important role in the failure of a system by stimulating development of the fire (by increasing the rates of maximum heat release and spread). It can be concluded that, on the basis of the water flow-rates tested, the most important parameter for an AWS system in a tunnel fire is ventilation velocity, rather than the water flow-rate; under such conditions, fire development is closely related to ventilation velocity, and almost independent of the water flow-rate.
Special strategies
To improve the performance of an AWS system in a tunnel fire, special strategies of variant ventilation and special control were also tested.
The variant ventilation strategy changed the longitudinal velocity from 2m/s (full-scale 8m/s) to 0.5m/s (full-scale 2m/s), following an estimated fire detection time. Using the variant ventilation strategy effectively suppresses the fire development and reduces the maximum heat release rate. However, the maximum ceiling temperature is slightly higher.
The special control strategy activated the sprinkler located 0.6m upstream. Using the special control strategy for nozzles also can efficiently suppress the fire development, reduce the gas temperature and prevent the failure of an AWS. Although the tested ventilation velocity was 2m/s (8m/s in full scale), it should still work under low ventilation.
Use AWS system under low ventilation
High ventilation promotes fire development, resulting in the failure of an AWS system. It should be used in tunnels with low ventilation velocities or natural ventilation.
Therefore, it is recommended that AWS systems be used in tunnels having transverse ventilation or in bi-directional tunnels, since in these tunnels longitudinal ventilation velocities will be relatively low.
AWS systems are not recommended for use in longitudinally-ventilated tunnels with high ventilation velocities, except when variant ventilation strategies or special control strategies are applied.
Automatic Sprinkler Systems in Tunnel Fires by Ying Zhen Li and Haukur Ingason from SP Fire Technology.
Main Image from Oliver Fire Protection