The ten year old Huguenot Toll Tunnel is in the Western Cape on the main N1 route between Cape Town and Johannesburg and was excavated as an inverted horse shoe. At present, a single (South) bore with bi-directional traffic flow is operational. The North bore was excavated but not lined or equipped.

The ceiling and off centre vertical divider form two air ducts with cross sectional areas of 4m2 and 8m2. The smaller is the dedicated fresh air duct, which delivers air through secondary air ducts at 40m intervals along its length to a zone 750mm above the pavement (Fig 1). The larger duct carries fresh air, which is delivered to the traffic space through graded slots in the ceiling under normal operating conditions. In the event of fire, smoke and hot gases are drawn into this duct through the slots and discharged to the outside at one or both of the portals.

Each portal is equipped with two 3.2m diameter fans, which serve half the tunnel. They comprise a six blade fresh air fan capable of operating at 375 and 750 rev/min (45/350kW) and a nine blade exhaust fan operating at 1000 rev/min (1000kW). Six Visibility (Vis) and CO monitors sample tunnel air and initiate the fresh air fan, select the operating speed and hydraulically position the fan blades for optimum operation. Two fire detectors are required to initiate the operation of an exhaust fan. A major tunnel fire has been experienced only once2. The traffic mix in the tunnel is illustrated in Table 1. Note growth of heavy vehicles . The present typical daily average vehicle count is 7500/day (80% light/20% heavy). The year on year increase in monthly traffic volumes is 6%.

During commissioning of the fresh air fans, the duct slots were adjusted to give a reasonably equal flow of air along the tunnel at full fan capacity. This air then exited at each portal, sweeping exhaust gases with it. Unfortunately, this distribution only holds true for the one fan capacity and, as the speed and blade angles are variables, adequate distribution to the centre of the tunnel at lower throughputs is not easily achieved. The West exhaust fan was tested by lighting a 20l petrol fire into which 5l of oil had been poured. The fire detection and exhaust system functioned well throughout.

Manifestation of the problem

Shortly after the tunnel opened, the first complaints of bad air were received. A firm of pollution experts was commissioned to run a survey, which showed very low levels of smoke and oxides of nitrogen (NOx), and it was concluded that the odours were only detectable by persons sensitive to minute quantities of NOx. A brown smoky haze in the tunnel was attributed to poor HGV maintenance and lack of law enforcement in respect of heavy vehicles belching smoke.

Eventually it was decided to conduct tests. It was suspected that the strong winds from the east contributed in some way to the build-up of pollution but the exact mechanism of the problem was elusive. Trials proved that smoke blown out of the tunnel was swept up and a good portion drawn into the fresh air fan inlet and recirculated back into the tunnel. It was initially thought that this alone was responsible for the concentration of NOx, so it was decided to run the fresh air fans for much longer. This expensive exercise proved ineffective, so the decision was taken to gather meteorological data at each portal, together with wind speed and direction in the second bore, better to understand the phenomenon and arrive at a solution.

Analysis of the data collected concluded that:

  • The meteorological forces acting on the portals of the tunnel are relatively high for extended periods of time

  • There is a strong correlation between the measured air speeds in the North and South tunnels, and between these and the portal wind velocity E-W vectors

  • It appears that these meteorologically induced airflows in the tunnel do adversely affect the behaviour of the ventilation flows in the South tunnel. Under certain conditions, the meteorological forces are sufficient to ventilate the tunnel with natural portal to portal longitudinal ventilation.

Using a semi-transverse ventilation system, when the natural ventilation is not enough to cope with the particular traffic conditions, for every amount of fresh air introduced mechanically, the natural ventilation flow reduces by a similar amount. The amount of fresh air that must be introduced is more than that originally provided by the natural ventilation. The higher the meteorological forces, the higher the mechanical ventilation flows needed to cancel the natural ventilation flows and introduce sufficient mechanical ventilation flows to maintain the pollutant concentrations below the required levels. The mechanical ventilation operates best and should only be used when the airflows introduced by meteorological conditions are low.

As the meteorological conditions increase in force, part or all of the mechanical ventilation should be switched off and mechanical ventilation used only to assist the natural (longitudinal) ventilation rather than cancel it and create semi-transverse ventilation.

The Huguenot Tunnel has the luxury of being equipped with a very expensive instrument for measuring weather induced airflow in the South tunnel – the unused North bore.

New control philosophy

The existing fresh air and exhaust fans are controlled by dedicated control panels in the portal buildings. They are equipped with Siemens S5 115U (942) Programmable Logic Controllers (PLCs). These are to be replaced by new PLCs as part of the retrofit contract.

The existing CO and visibility monitors in the Western and Eastern half of the tunnel are hardwired to the Western and Eastern ventilation PLCs respectively. No communication exists between the Eastern and Western Control PLCs. Each PLC considers the signals from its respective CO and visibility monitors and automatically sets blade angle and fan speed to produce an air volume to reduce pollution levels. During a typical weekday afternoon traffic peak, readings of up to 55ppm CO and 5.5 K Vis are observed.

The primary objective of the new control system will be to achieve more effective control of the fans, taking into account the weather induced air flows. The fans will be treated as a system rather than as two independent fresh air fans, with all six CO and combined Vis measurements being used by the control system to control both fans. The new fan Control PLCs and cross connection PLCs will communicate via a dual redundant fibre optic network, with all CO and Vis measurements passed from the cross-connection where they are measured to both fan control PLCs.

Instruments will be installed at two positions in the South bore (towards the portals) to measure actual airflow. The existing airflow instrument in the North bore will be included in the new control system. All three airflow measurements will be passed via the control network to both fan control PLCs.

The strategy is then to use the North bore airflow measurement to control the fans. The North bore air speed measurement will be passed to both fan control PLCs. If the value is less than a pre-set minimum, say 1.5m/s in either direction, the system will operate as a semi-transverse ventilation system. As the North bore air speed measurement rises above the preset minimum in either direction, the controllers will make a decision to shut down one or other of the fresh air fans (i.e. the one which is producing a mechanical airflow in an opposite direction to the naturally induced airflow). The remaining fan will continue to operate, assisting the natural airflow to ventilate the tunnel longitudinally. The controllers will adjust the speed and blade angle of this fan to achieve the set CO and Vis levels.

Commissioning of the North bore will necessitate development of a system to determine the meteorological forces from prevailing weather conditions. It will also alter the traffic flow. Each bore will be used as a uni-directional tunnel under normal operating conditions. Future ventilation control strategies will have to take into account:

    &#8220The primary objective of the new control system will be to achieve more effective control of the fans, taking into account the weather induced airflows”
  • The extent of plug flow in a uni-directional bore, and whether this plug flow works against or with prevailing weather induced airflows

  • Both tunnels may revert to bi-directional operation for maintenance or repairs or if there is an accident in the other tunnel. This will re-introduce the problem of turbulence associated with a bi-directional bore.

Implementation of an expert system

The next step in the strategy is the implementation of an expert system to predict weather induced airflow from prevailing weather conditions, and consider other parameters in the control of the ventilation system. The most dominant meteorological force in the generation of induced air flow is the E -W wind vector at the Western portal, and one might be tempted to use this measurement as a direct substitute for the North bore air flow measurement in order to set the ventilation strategy as described above.

However, the authors’ plan is to implement a knowledge based expert system which has the facility to store the measured weather data, and, using current weather measurements, determine the induced airflow by inference. The principal objective of the expert system will be to improve CO and Vis levels in the tunnel by supplying expert advice to the fan controls and simultaneously reduce power consumption. The following data will have to be considered and rules or models established.

  • Historical weather data and determination of induced airflows. The actual portal windspeeds and, in particular, the E-W vector, will be used to determine the induced air flow expected under such conditions by comparison with historical data

  • Daily, weekly and annual traffic data and prediction of traffic patterns, using time of day, day of week, variable holiday calendars, traffic growth patterns and the future modes of operation of the twin bore system

  • Power consumption and electricity charges related to the generation of air volumes required. These requirements are illustrated in Fig 6. In generating the expert advice, the expert system will be required to perform two main functions:

  • Extracting more information from the data

  • Modelling the process or developing rules

    Expert advice will determine for both ventilation fans:

  • Start-up/shutdown times

  • Selection of which fan should run

  • Fan speed settings

  • Air volume settings (air volume control by PLCs.)

The application of this advice to the fan controls will maximise the benefit of weather induced airflow and traffic plug flow while keeping energy maximum demand requirements as low as possible. It is expected that the following will be observed once the system is operational:

  • Use by the system of only one fresh air fan under conditions of high portal winds

  • Start-up of fans before CO/Vis start points are reached in anticipation of traffic peaks

  • Limiting of maximum demand and hence energy costs

  • Smoother control of blade angle and air volume and elimination of peaks in power demand



Related Files
Ventilation System
Expert System Diagram