Department of Transport Systems, Traffic Engineering and Logistics Faculty of Transport and Aviation Engineering

Fire ventilation tests in road tunnels

1. Measurements in the Laliki tunnel

On October 16th, 2016, the ventilation system efficiency was measured in the Emilia road tunnel. The Emilia tunnel is located on the S69 road and it links Żywiec and Zwardoń. The length of the tunnel is 678 m, the width is 11.9 m and the height is 6.55 m. The gradient (inclination) of the tunnel is 4%. The tunnel is not rectilinear. The northern portal has an elevation of 669 m a.s.l. and the southern portal has an elevation of 642 m a.s.l.

The tunnel is divided into 5 sections where the first is near the northern portal. Each section has two fans mounted. They produce 810 N of thrust each. All the fans are reversible, although the preliminary direction of their work is S→N.

Fourteen anemometers manufactured by the Sensor company were mounted on two pole stands. The height of the pole stands was 6.1 m and 4.5 m respectively. Eight anemometers were mounted on the taller pole stand and 6 anemometers on the shorter one. The distance between the anemometers was 0.77 m. The pole stands were moved during the measurements according to the previously determined measuring mesh.

This method allowed entire tunnel cross-section to be covered by the mesh 0.77×0.90 m. It comprised 76 measuring points in a particular cross-section of the tunnel. The fans had been turned on manually during the tests. Table 1 shows the details of the measurement series. W

Table 1. Measuring series - details

As the results the full air velocity profiles were obtained. The following conclusions may be drawn:

1. According to the required critical velocity, the work of one section of fans produces the required airflow toward the N portal (series 1 and 4). This is when the airflow direction is in the same direction as the natural stack effect. Average air velocities in the two series were 3.2 m/s and 3.3 m/s respectively
2. When the direction of the natural stack effect is opposite to the airflow produced by the fan section, it is necessary to activate another section of the fans (series 2 and 3). Otherwise, the critical velocity is not reached. Average air velocities in these two series were 2.5 m/s and 2.3 m/s respectively
3. Measurements in the tunnel under the Martwa Wisła in Gdańsk

Similar measurements were carried out on Jun 10^th 2017 in the Martwa Wisła tunnel. This is an urban tunnel which is located in the city of Gdańsk under one of the limbs of the Wisła river (Figure 1). The total length of the tunnel is 1378 m andit includes two tubes for both directions of the traffic. The height of each tube is 7.03 m and the width is 10.09 m. The minimal depth under theriver is approximately 8 m. The direction of the tunnel is from southeast (SE) to northwest (NW). The deepest part of the tunnel is roughly at its mid-section. The inclination towards the NW direction is 4% and towards the SEit is 3%. The crossfall of the roadway is 2.5%. The tubes are linked every 170 m by evacuation exits.

The tunnel is equipped with a longitudinal ventilation system. It is operated by 11 axial jet fans in each tube. They are mounted individually under the ceiling every 95m.

The axial fans are partly reversible. Their nominal thrust is 1200 N in the normal direction and 407 N in reversed mode. The diameter of the fan is 1.12 m and its length is 4.7 m. The normal direction of airflow is the same as the traffic direction. The basic airflow rate in the normal direction is 33.76 m3/s and in reversed mode is 20.25 m3/s. Exhaust velocity in the normal direction is 34.27 m/s.

The system of the stand poles was similar as previously, but due to different shape of the tunnel vault  it was a bit altered and covered 56 measuring points in a particular cross-section of the tunnel.

Five measuring series were carried out in the southern tube, where traffic is from northwest (NW) to southeast (SE). They are outlined in Table 2. Each series of airflow measurements was undertaken in the same cross section at a distance of 520 m from the NW portal (near the deepest point of the tunnel). The fans were turned off during the first measurement series. The first measurements were related to the natural draught effect.

Table2. Description of measuring series

The measurements confirmed the existence of a strong natural draught, which is the result of a combination of ambient conditions (mainly wind and air temperature at the tunnel portals) and tunnel geometry (mainly its inclination). Under typical operation of the tunnel, the piston effect arising from moving vehicles should also be taken into account.

The intensity of the natural draught in general is variable, thus it could unexpectedly influence the operation of the ventilation system. This phenomenon should be taken into account during the design phase of the tunnel ventilation system.

Fire development simulations in road tunnels

Sub-area: Safety in transport and logistics systems

Fire experiments in tunnels are very difficult to carry out as a real fire can damage the tunnel infrastructure. Experiments in the form of smoke tests with a fire of 1.5 MW (several times less than a car fire) and the production of artificial smoke are feasible. In such a situation, numerical simulations are preferable, but the problem of validating the obtained results arises.

3. Hot-smoke tests in Laliki tunnel

Two hot smoke tests have been conducted. They were based on the Australian Standard AS 4391-1999. Five smoke generators made by the Vulcan Company and two or four fire trays (containing ethanol) were in use. According to the above standard, two fire trays of A1 size filled with 16 dm3 of ethanol give a heat output (HRR) of 700 kW, whereas four fire trays the same size produce a heat output (HRR) of 1500 kW. Such trays configuration and the amount of fuel should assure the following burning sequence: 3 min of fire growth, 10 min of stable fire and 3 min of decay. The trays were placed in the middle of the road lane.

The second hot smoke test showed an increase of temperature on the windward side, evidently the upstream flow of hot gases was forced by the released heat. A reading of 46˚C was recorded at a height of 5.6 m, 5 meters from the last tray. During this test, the leeward part of the tunnel was completely filled with smoke

Due to relatively short time of the hot smoke test, the tunnel ceiling was not warmed up significantly – the temperature of the tunnel ceiling just above the fire, as measured by the thermo-vision camera, did not exceed 20˚C in both tests. This value is beneath the temperature threshold of fibro-laser sensor activation (about 40˚C), but in the first test this sensor should be also triggered by the rapid temperature rise (about 5˚C / 60 s).

4. Numerical investigation on fire accident in a urban tunnel

A fire outbreak in a road tunnel is the most dangerous threat to the people who are trapped inside. The main danger is the generation of large amounts of toxic smoke. Tunnel fires develop in an unpredictable way, but the conditions get worse very quickly. Therefore the self-evacuation involving suitable decisions and proficient actions is of crucial importance. The work presents a number of scenarios of the tunnel fires and the following evacuation. Different tunnel fire detection delays, types of ventilation system, traffic conditions, fire locations and fire powers were taken into account. The research was conducted by the combined numerical simulation: VISSIM software was used to simulate the traffic and the formation of congestion, FDS was used to reproduce the fire development and PATHFINDER was used to simulate the evacuation and to estimate injures. To embrace the great number of factors influencing the described issue the Taguchi method of experiment design was applied. It allowed also for determining the relative influence of these factors on the evacuation process.

The most important direct conclusions, which come from the described study are listed below:

1. The main factor threatening the people trapped inside a tunnel is smoke, which worsens the visibility and may contain toxic components.
2. The increased temperature is a threat of minor importance, just few people close to the fire source can be exposed to this danger.
3. The number of endangered people quickly grows up with the fire power and the traffic intensity.
4. Even in the case of shorter tunnels the installation of a semi-transverse ventilation system in a form of additional air supply vents located just above the street significantly increases the safety level. The observed gain is stronger for tunnels with for uni-directional traffic.
5. The quick and reliable fire detection is crucial. Therefore fire detection systems should be multiplied and be sensitive even to fires which develop in hiding. The fire detection rate is particularly important for bi-directional tunnels, where people could be trapped at both sides of the fire. Although, the delayed fire detection of a high power fire could result in a great number of exposed people, such scenario is not too likely.
6. Fire localization is of minor importance.

The results of the carried out numerical experiments allowed also to formulate several more suggestions of rather general nature:

7. The passengers of busses are especially threatened. Thus the bus drivers should be familiar with the principles of efficient evacuation, and the buses should be equipped with systems announcing unambiguous alarm signals.
8. Despite of the fact that the simulated evacuation ran without disturbances the number of threatened people was in some cases quite high. In the reality, with people being confused, indistinct messages and illegible marking, suboptimal decisions would be made, so the evacuation time could be significantly longer and the number of victims more severe. Thus it is desired to apply the moving light arrows, made in LED technology. Such signs mounted in the pavement or in walls just above it would be displayed in the alarm mode and would clearly guide people in the right direction. Additionally, one can imagine a dynamic control of such system which would take into account current conditions in the tunnel and choose relevant path for the evacuees.
9. A fire of a single passenger car is of low power, what results in the relatively low number of the threatened people. A fire of HGV could be dangerous for many more people, especially in tunnels with intensive traffic. Therefore, in a case of congested traffic in a tunnel the entry of large vehicles carrying flammable loads could be prohibited by variable message signs (VMS) and interim detours would be proposed as well. It would also be desirable to widespread the infrared monitoring of cars entering the tunnel in order to detect a possible fire ignition.
10. The road traffic is characterized by high variability, even for similar values of average traffic intensity. This is why in any tests of safety systems the least favorable conditions should be assumed.