The Design of Tunnel Ventilation System for a Long Vehicular Tunnel

Guian Sin Kwa Land Transport Authority, Singapore

ABSTRACT Long vehicular tunnel coupled with long slip roads presents special challenges in tunnel ventilation system design. During normal traffic operation, the vehicular emissions have to be removed to ensure that the tunnel air quality is kept within acceptable limits. The pollutants discharged from the ventilation buildings to the atmosphere have an air quality impact on surrounding development which has to be carefully assessed. The ventilation system also has to be designed to prevent the building up of temperature beyond tolerable limit during traffic congestion where vehicles may pile up for several kilometres. When there is a fire emergency, smoke control to enable safe evacuation has to be addressed for main tunnel, long in-ramps and out-ramps. This paper describes the major elements of the design of the ventilation system for the 8.7 km tunnel of the Singapore Kallang / Paya Lebar Expressway. 1. INTRODUCTION The Land Transport Authority of Singapore has embarked on the construction of the Kallang / Paya Lebar Expressway to provide a link between the East Coast Parkway in the south and the Tampines Expressway in the north. Designed to carry a traffic volume of 6,600 vehicles per hour during peak hours, it is a complex project made up of tunnels, slip roads, viaducts, at-grade expressway and ventilation buildings. It will have a total length of 12 km of which approximately 8.7 km will be in tunnel. The Expressway tunnel will be linked by 8 entry and 9 exit slip roads and will be one of the longest dual 3-lane road tunnels in Southeast Asia when completed in 2007. 2. TRADITIONAL CONCEPT OF VEHICULAR TUNNEL VENTILATION Traditionally, three approaches have been used in the design of vehicular tunnel ventilation system, viz. longitudinal, semi-transverse and full transverse ventilation system. For short tunnels less than 250m in length, natural ventilation is a possible option subject to detailed engineering analysis. 2.1 Longitudinal ventilation system Longitudinal ventilation can be generated with a system of fans, which may be spaced intermittently at the ceiling above the roadway or may be housed in central ventilation buildings. Intermittent vent shafts, at points where air is either supplied or exhausted, can be introduced to increase the practical length of tunnels in which longitudinal ventilation can be effectively applied. The airflow direction follows the traffic direction and is usually more energy efficient because the system takes advantage of the vehicle piston effect. Under normal free flowing traffic, the tunnel is usually self-ventilated. The concentration of pollutant levels in the tunnel increases in the direction of airflow.

B04

In a fire emergency, smoke will be pushed downstream of the fire site towards the exit portal or captured at strategically located exhaust shafts. The Memorial Tunnel Fire Ventilation Test Program (MTFVTP) had concluded that a longitudinal ventilation system employing jet fans is highly effective in managing the direction of smoke spread for heat release rates up to 100MW (with 3.2% grade) [1]. Longitudinal ventilation can be achieved by using either jet fans mounted at the tunnel ceiling or by Saccardo nozzles [2] located at the tunnel ceiling. For jet fan system, the fans deliver airflow at a very high velocity (e.g. 20 to 35 m/s) and can create a momentum that will entrain the surrounding air to move in a desired airflow direction. For Saccardo nozzle system, air is injected at a high velocity (e.g. 30 m/s) from the nozzle. Similar to the jet fan system, the air exiting from the Saccardo nozzle can create a momentum that will entrain the surrounding air to move to a desired airflow direction. The jet fan system and the Saccardo nozzle system are illustrated in Figure 2.1 and Figure 2.2 respectively.
Jet fan

Figure 2.1 Jet Fan System with Air Exchange

Figure 2.2 Saccardo Nozzle System with Air Exchange

2.2 Semi-transverse ventilation system Semi-transverse ventilation system either supplies outdoor air or extracts exhaust air uniformly over the full length of the tunnel. Therefore, it always requires a supply air duct or an exhaust air duct running along the tunnel. With a semi-transverse supply or semi-transverse exhaust air system, the pollutant level in the tunnel are relatively uniform for uni-directional traffic flow. Oversized exhaust points at regular intervals along the tunnel are normally provided for emergency operation. The ventilation concept is illustrated in Figure 2.3.

Supply Air Duct

Supply Air Duct

Figure 2.3 Semi-Transverse Ventilation System with Enlarged Point Exhaust

In a fire emergency situation, the semi-transverse supply system will operate in reverse mode. Smoke will be captured over the length of the tunnel, with outdoor air entering the tunnel through the portals or adjacent air intake shafts. The aim is to maintain a smoke clear layer for passenger evacuation.

B04

2.3 Full-transverse ventilation system Full-transverse ventilation system supplies outdoor air and extracts exhaust air uniformly over the full length of a tunnel. Therefore, a supply air duct and an exhaust air duct running along the tunnel are always required. In this system, the amount of exhaust air equals the amount of supply air and the airflow is transverse to the longitudinal axis of the tunnel. With a full-transverse ventilation system, the pollutant level as well as the air velocity in the tunnel are relatively uniform. The ventilation concept is illustrated in Figure 2.4.

Exhaust Air Duct

Exhaust Air Duct

Supply Air Duct

Supply Air Duct

Figure 2.4 Transverse Ventilation System with Enlarged Openings for Smoke Extraction

In a fire emergency situation, smoke will migrate from its source along the underside of the ceiling. The full-transverse ventilation system attempts to control smoke and hot gases by extracting the smoke through the exhaust openings along the ceiling. The design aims at maintaining a smoke free layer for passenger evacuation. MTFVTP had shown that extracting smoke quickly from a location that is as close as possible to the fire can significantly reduce the migration of smoke and heat in undesirable directions, however, the extraction point needs to be oversized. 3. VENTILATION EXPRESSWAY DESIGN CRITERIA FOR THE KALLANG / PAYA LEBAR

The design criteria for the ventilation system have been developed after considering international practices and using guidelines from relevant infrastructure and technical bodies. 3.1 Air quality The tunnel air quality based on the recommendations of the Permanent International Association of Road Congress (PIARC) [3] for Year 2010 is shown in Table 3.1. As the predicted traffic mix contains mainly petrol engine driven vehicles (emit CO) and a small proportion of diesel engine driven vehicles (emit NOx), the airflow required to meet CO and visibility criteria will also meet NOx criteria. Table 3.1 Tunnel Air Quality Guideline (PIARC)
Traffic Situation Fluid peak traffic Daily congested traffic Exceptional congested traffic CO Concentration (Year 2010) 70 ppm 70 ppm 100 ppm Visibility 0.005 m-1 0.007 m-1 0.009 m-1

3.2 Air velocity The design air velocity in the tunnel during normal traffic operation is 8 m/s as any higher velocity will lead to stirring of debris and dust in the tunnel. Tests referenced in Subway Environmental

B04

Design Handbook [4] show that air velocity above 11.2 m/s will cause difficulty in walking and therefore could impede evacuation during emergency. Taking into consideration the blockage effect from vehicles, the maximum design bulk air velocity for emergency operation is 9 m/s. 3.3 Air temperature During congestion, the ventilation system is designed to maintain the tunnel environment at a temperature of not more than 10C above the ambient temperature. 3.4 Heat release rate Although Singapore has regulations to prohibit hazardous goods vehicles from entering the tunnel [5], in reality enforcement is difficult as evidenced by the occasional entry of petrol tanker into the Central Expressway. As such, the design heat release rate is 100MW which is based on a petrol tanker on fire. 4. MAIN TUNNEL VENTILATION For the main tunnel, the following 5 tunnel ventilation schemes have been considered: (a) (b) (c) (d) (e) Jet fan system with air exchange points Saccardo nozzle system with air exchange points Semi-transverse ventilation system with local extraction points Full-transverse ventilation system Electrostatic precipitator type with longitudinal ventilation system and air exchange points.

Calculations show that CO, instead of the visibility, is the dominant criterion for pollutant dilution. Therefore, electrostatic precipitator which cannot reduce the airflow requirement was not considered. 4.1 Comparison of different tunnel ventilation schemes 4.1.1 Operation Effectiveness (Normal Operation) - Jet fan and Saccardo nozzle schemes make use of the piston effect to ventilate the tunnel and are more energy efficient. The semi-transverse ventilation scheme uses less energy than the full-transverse one as local exhaust instead of ducted exhaust is used. Tunnel Air Quality - The pollutant concentration profile inside the tunnel increases in the direction of traffic flow for jet fan and Saccardo nozzle schemes. The concentration of pollutants inside the tunnel is almost constant and reaches the pollution limit for semi-transverse and full-transverse ventilation schemes. Based on the air quality design limit, the jet fan and the Saccardo nozzle schemes give better air quality as the average concentration of pollutant in the tunnel is lower for these two schemes. Tunnel Air Velocity - The air velocity inside the tunnel for jet fan and Saccardo nozzle scheme is almost constant. Their velocity is also higher than the transverse schemes. The air velocity for semitransverse scheme increases along the tunnel in the direction of traffic flow. 4.1.2 Maintenance of tunnel ventilation equipment For jet fan system, the fans are mounted at the tunnel ceiling and therefore special maintenance platform is required for the maintenance of the jet fans. Also, routine maintenance of jet fans may affect the traffic unless it is performed during the night when the tunnel is closed or partially closed. On the other hand, ventilation equipment for the other three schemes are housed inside ventilation buildings and impose fewer constraints on the operation of the tunnel. For semi-transverse and transverse ventilation schemes, it is necessary to provide enlarged openings with motorised fire and

B04

smoke dampers at the ceiling for smoke extraction. Maintenance of these dampers may affect traffic operation unless it is carried out at night when the tunnel is closed or partially closed. 4.1.3 Fire safety and smoke control For all the four schemes, the recommendations from MTFVTP have been adopted. Performance during fire emergency is good for jet fan and Saccardo nozzle schemes, and fair for semi-transverse and transverse ventilation schemes. 4.1.4 Initial cost Saccardo nozzle scheme has the lowest cost as it makes use of vehicle piston effect for ventilation during normal operation. The jet fan scheme is second lowest in cost due to extensive cabling in a long tunnel. Semi-transverse and full-transverse schemes have higher cost as they require concrete air ducts along the tunnel and dampers at intervals for smoke control. Full-transverse ventilation scheme has the highest cost as it requires two sets of ductwork and associated equipment for pollutant control. 4.1.5 Running cost Running cost is associated with the effective use of the piston effect. Therefore, both jet fan and Saccardo nozzle schemes have the lowest cost. The running cost for full-transverse ventilation scheme is the highest as it requires the operation of both supply and exhaust systems along the tunnel. 4.1.6 Spatial requirements in tunnel Full-transverse ventilation scheme requires two 15m2 concrete air ducts while semi-transverse ventilation scheme requires one 15m2 concrete air duct running along the entire tunnel length. Jet fan scheme requires local enlargement of the tunnel section so that jet fans can be mounted at the tunnel ceiling. This requires less space compared to the full-transverse and semi-transverse ventilation schemes. Saccardo nozzle requires local enlargement in the tunnel to house the Saccardo nozzle. Concrete air ducts are also required to connect the Saccardo nozzles to the ventilation buildings. 4.1.7 Construction method Full-transverse ventilation scheme and semi-transverse ventilation scheme require concrete air ducts running along the tunnel. As a result, pre-cast concrete slabs with openings are normally used for the construction of the air ducts. For jet fan scheme, tunnel lining with reinforcement is strong enough and can be used for the mounting of the jet fans. For Saccardo nozzle scheme, the concrete surface of the nozzle has to be smooth and the exit angle precise in order to ensure the performance of the nozzle. 4.1.8 Demand for land Full-transverse ventilation scheme and semi-transverse ventilation scheme require the largest ventilation buildings, as they cannot fully utilise the piston effect. Saccardo nozzle scheme requires larger ventilation buildings than that of jet fan scheme as additional space is required to house the Saccardo nozzle and its associated equipment. 4.2 Adopted ventilation scheme Based on the above analysis, Saccardo nozzle system with air exchange points is the most efficient and cost-effective ventilation system and was adopted. 4.3 Evacuation strategy during fire emergency The ventilation scheme for emergency operation as illustrated in Figure 4.1 is that axial fans upstream of fire run in supply mode and axial fans downstream run in exhaust mode to push the smoke in the direction of traffic, and a longitudinal air velocity greater than the critical velocity will be created to prevent smoke backlayering. The evacuation strategy is that people will evacuate towards the nearest escape staircase or cross passage upstream of the fire site where there is a smoke clear path.

B04

Stopped Exhaust Fan Operated Fresh Air Supply Fan Operated Stopped

Longitudinal Velocity Greater Than Critical Velocity to Prevent Smoke Backlayering

Figure 4.1 Main Tunnel Emergency Ventilation Concept

4.4 Validation of Saccardo nozzle design Subway Environmental Simulation (SES) program [6] is a one-dimensional computational tool, and as such, Computational Fluid Dynamics (CFD) tools were used to analyse the local effects of the design. 4.4.1 CFD analysis of air velocity downstream of Saccardo nozzle Analysis of air velocity downstream of Saccardo nozzle for one, two and three nozzles in operation was carried out based on congested traffic (no drag resistance) and the results are shown in Table 4.1. Table 4.1 Air Velocity Downstream of Saccardo Nozzle
Case 1 2 3 4 Description All three nozzles in operation Right and centre nozzle in operation Centre nozzle in operation Right nozzle in operation Results Maximum velocity at 1.5 m above road level is 8.7 m/s Maximum velocity at 1.5 m above road level is 7.5 m/s Maximum velocity at 1.5 m above road level is 6.4 m/s Maximum velocity at 1.5 m above road level is 7.0 m/s

Except for Case 1, all cases have air velocity less than the design criteria of 8 m/s. The chances of operating all three nozzles are remote as it is anticipated that only standstill traffic or fire emergency will require the use of three nozzles. The CFD result for Case 2 is illustrated in Figure 4.2. 4.4.2 CFD analysis of re-circulation effect of Saccardo nozzle CFD analysis was carried out to predict the amount of re-circulation and compared against the allowable re-circulated air flow based on maximum fresh air quantity of 495 m3/s calculated by SES program. Table 4.2 shows that the maximum separation of Saccardo Nozzle is about 2.5 km to cater for 3-lane standstill traffic. For longer separation, it will be necessary to operate the jet fans or operate intermediate ventilation buildings. The CFD result for Resistance Factor K=12 is shown in Figure 4.3. Table 4.2 Comparison of Allowable and CFD Predicted Re-Circulation
Tunnel Length (m) 1000 1500 2000 2500 3000 Maximum Fresh Maximum Supply Air Required Air from Nozzles (m3/s) (m3/s) 131 495 197 495 262 495 323 495 393 495 Allowable Re-Circulated Airflow 74% 60% 47% 35% 21% KFactor 5.27 7.94 19.54 13.21 15.88 CFD Predicted Re-Circulated Airflow Carried Over Less than 7% 15% 21 30% 30% Remarks

Acceptable with 2 nozzles Acceptable with 2 nozzles Acceptable with 3 nozzles Not acceptable.

B04

Air Velocity (m/s)

Percentage Re-circulation

Figure 4.2 CFD Analysis of Air Velocity Downstream of Saccardo Nozzles

Figure 4.3 CFD Analysis of Re-Circulation Effect of Saccardo Nozzle (K=12)

4.5 Ventilation buildings The 6 ventilation buildings are 25m in height with a 5m high exhaust stack above the roofline. Air quality impact assessment using SCREEN and ISC models developed by the US Environmental Protection Agency [7] was conducted. Due to pollution impact, air quality setback distance and height control of adjacent development have to be implemented for all the 6 ventilation buildings. 5. OUT-RAMP VENTILATION 5.1 Normal operation There are altogether nine exit slip roads ranging from 60m to 420m in length. For out-ramps that are shorter than 100m, the tunnel will be naturally ventilated using the piston effect of the vehicles to bring the pollutants out of the portal during normal operation. Where the out-ramp is longer than 100m, fresh air supply fans and jet fans will be provided to dilute the pollutants. 5.2 Emergency operation For out-ramps which are shorter than 100m, smoke will move uphill, stratify at ceiling and vent through the portal. The evacuation strategy is that people will evacuate towards the exit portal. For out-ramps that are longer than 100m, the ventilation system will operate to push the smoke, without backlayering, in the direction of traffic flow and towards the exit portal by combination actions of the main tunnels axial fans and the out-ramps jet fans as shown in Figure 5.1. The evacuation strategy is that people will evacuate towards the nearest escape staircase or cross passage upstream of the fire site.

Stopped

Longitudinal Velocity Greater Than Critical Velocity to Prevent Smoke Backlayering Stopped Fresh Air Supply Fan Operated

Jet Fan Operated in Reverse

Figure 5.1 Out-Ramp Emergency Ventilation Concept

B04

6. IN-RAMP VENTILATION There are altogether eight entry slip roads ranging from 150m to 730m in length. Longitudinal ventilation will be used for the in-ramps regardless of their length for normal and emergency operation. 6.1 Normal operation For free-flow traffic condition, the piston effect of vehicles will induce sufficient air into the tunnel. During congested traffic condition, jet fans at entrance portal will entrain outdoor air into the tunnel. 6.2 Emergency operation During fire emergency, the jet fans at the portal will operate to push the smoke in the direction of traffic. Exhaust point at the junction of the in-ramp and main tunnel connected to the main tunnel axial fan will operate to minimise smoke entering the main tunnel. The longitudinal air velocity will be greater than the critical velocity to prevent the backlayering of smoke. The evacuation strategy is that people will evacuate towards the entrance portal or nearest escape staircase or cross passage upstream of the fire site. The ventilation concept is illustrated in Figure 6.1.

Jet Fan at Entrance Portal Operated

Longitudinal Velocity Greater Than Critical Velocity to Prevent Smoke Backlayering Exhaust Fan Operated Stopped

Exhaust Point Operated

Stopped

Figure 6.1 In-Ramp Emergency Ventilation Concept

7. CONCLUSIONS A longitudinal ventilation system using Saccardo nozzles and air exchange points has been designed for the Kallang / Paya Lebar Expressway. It is an efficient and cost-effective design which has taken into account land use and air quality impacts, engineering constraints, construction cost, safety and efficiency of operation. The use of Computational Fluid Dynamics analytical tools has contributed to the verification of design concepts developed using Subway Environmental Simulation program. 8. REFERENCES [1] Memorial Tunnel Fire Ventilation Test Program (MTFVTP) Test Report, 1995 [2] 2003 ASHRAE Handbook - HVAC Applications, Chapter 13 Enclosed Vehicular Facilities [3] Technical Committee Report of the PIARC XIXth World Road Congress, Montreal, 1995 [4] Subway Environmental Design Handbook Volume I: Principles and Applications, 2nd Edition, 1976 [5] Road Traffic (Expressway Traffic) Rules, 1990 Edition of Road Traffic Act [6] Subway Environmental Design Handbook Volume II: Subway Environment Simulation (SES) Computer Program, Version 4.0, 1997 [7] SCREEN (Version 3) Model and ISC (Version 3) Model, US Environmental Protection Agency

B04