Saccardo nozzle ventilation system: cfd analysis to upgrade the performance of the ventilation system

This paper presents the results of a 3D numerical investigation carried out on a Saccardo ventilation system operating in road tunnel, with the aim to improve the Performance of the Ventilation System, Pvs. 1D mathematical model, here presented, has highlighted that the pitch angle and length of inflow opening are the main parameters that influence the Pvs. The results have shown that the pitch angle of the Saccardo nozzle, influences moderately the Pvs, instead the opening inflow lengths influence considerably the transferring of the momentum, increasing the mass flow rate. Pvs trend vs opening inflow length shows an improvement up to 47% with respect to the reference opening inflow length.


Introduction
In one-way road tunnel, the longitudinal ventilation systems are adopted to maintain acceptable air quality and thermal comfort.Tunnel ventilation system, typically, is designed to provide the optimal environmental condition for both normal and congested traffic scenarios, both in ordinary and emergency cases.The main systems adopted to provide impulse ventilation in tunnels include the Saccardo nozzles sometimes coupled with impulse exhaust dampers, traditional jet fan and its optimization such as the Banana Jet® and MoJet©.Tarada F. and Brandt R. [1] have presented a brief overview of impulse ventilation for tunnels, as guidance regarding their advantages and drawbacks.The verification of the goodness of the design choices or the activity of optimization for the ventilation systems is, usually, achieved by full scale or reduced scale experiments, both of which are expansive in terms of cost and time.Instead, the CFD analysis enables to carried out, regardless of the complexity of geometrical characteristics of boundary (spatial and time) conditions, a quite easy prediction of velocity and temperature fields both in ordinary situations and in emergency (in case of fire or traffic jam).Therefore, the CFD analysis, once validated on experimental data, is the cheapest way to optimize the performance of the fluid dynamic systems with respect to full or reduced scale experiments.Alston J. et al. [2] studied the retrofitting feasibility in a transverse ventilation system with a Saccardo nozzle placed at the entrance of a road tunnel.Sturm P. et al. [3] proposed in their research a simple calculation method for 1-D approach, derived from the conservation equation of momentum in order to calculate the Saccardo efficiency coefficient.Waymel, F. et al. [4] presented a methodology based on combination of 1D and 3D modelling used for the design and specific uses related to the Saccardo system design development.Hang, J.G. [5] et al. optimized nozzle design reaching a significant reduction in the overall maximum tunnel air temperature and proposed an alternative nozzle design.In particular, they present a detailed study of the performance of an optimized Saccardo Nozzle design using a CFD model combined with a mono-dimensional temperature prediction model to evaluate the carry-over effects in the tunnel.Ganjiazad, R. et al. [6] evaluated the influence of volumetric flow rate and inclination angle of air jet on the Saccardo ventilation system performance and inclination angle range to optimize the performance in fire scenario.Alston J. et al [7] have studied retrofitting of an existing transverse ventilation system to a longitudinal system using a Saccardo Nozzle using 3D CFD.Chammem T. et al [8] in their paper suggested a new longitudinal ventilation technique based on supplying external air into the tunnel by two inclined jets, combining the advantages of Banana Jet and Saccardo nozzles.Sturm P. [8] discussed the methods of fire ventilation with a focus on the requirements for sensors for Saccardo ventilation system.In this work the authors present a 3D CFD analysis of a longitudinal ventilation system equipped with Saccardo nozzle with the aim to improve the Performance of the Ventilation System, Pvs.A preliminary 1D mathematical model highlights in the pitch angle and length of inflow opening the main parameters that influence the Pvs.The CFD results show that the pitch angle of the Saccardo nozzle influences moderately the Pvs, instead the different opening inflow lengths influence considerably the mean air velocity in the tunnel and mass flow rate.In particular, a Pvs trend vs opening inflow length shows an improvement up to 47% with respect to the reference opening inflow length.In order to validate the CFD simulation, this paper refers to tunnel investigated to Hang et al. [5].

Details of the Tunnel and Ventilation System models
The physical domain under investigation is shown in figure 1 (not to scale); it represents a road tunnel of 1 km length, 14.6 m of width (with 3 lanes) and 6 m of height equipped with longitudinal ventilation system.The air flow is induced by Saccardo nozzles that provides fresh air, while the dumpers remove the exhausts.The three Saccardo nozzles and exhaust dumpers (as shown in figure 1

1D Mathematical model of the Ventilation System.
A 1D mathematical model to investigate the Saccardo system performance is presented.This model, even if simplified, can lead to recognize the main parameters to be investigated.Starting from the momentum balance (equation 1) referred to the volume control presented in figure 3, the equation (1) becomes the equation ( 4): where  is the total friction losses coefficient and L1, L2 tunnel are lengths upstream and downstream the control volume, respectively.This coefficient sums the friction due to the tunnel walls and the friction due to cars and the losses due to vorticity among cars [9,10].Considering, also, the continuity equation ( 5):     =     (5) the equation ( 4) can be rewritten as equation ( 6): The equation (6) shows that the velocity in the tunnel (  ) depends only on   and , while ̇, ̇ ℎ and   were design parameters.Through the CFD investigation, the influence of the pitch angle  and the jet velocity   are investigated in terms of Performance of the Ventilation System "Pvs", defined as ratio between the mass flow rate at the outlet tunnel portal (̇  ) and the mass inflow provided by Saccardo (̇), as reported in equation ( 7):

Boundary conditions
In this numerical CFD analysis, carried out by FDS simulator [11], the following boundary conditions were imposed: -the jet fans installed in Saccardo ventilation system are simulated as simple momentum source able to obtain: • total inflow mass flow rate ̇up to 202 kg/s; • total outflow mass flow rate of exhausts ̇ ℎ up to 202 kg/s; -the air is considered an ideal gas with properties valuated at the ambient temperature (20 °C); -the ceiling and floor roughness height of tunnel, modelled by "log wall model", are set equal to 0.005 m; -the pitch angle of the Saccardo system is set equal to 10° [5]; -the pressure at inlet and outlet portals is set equal to 101325 Pa.

Sensor positions
Three types of sensors were placed over of the simulation domain: velocity magnitude (v), axial velocity (vx), static pressure (p).The figure 4 shows how the sensors are arranged vertically and transversely in the tunnel.

Mesh choice and model validation
The domain is divided in two mesh zones, the mesh type A (0.5 m) covers the computational domain longitudinally (x-axis) from -500 m to -38.5 m and from 81.5 m to 500 m (see figure 5); the mesh type B (0.25 m) covers the zone where Saccardo nozzle and exhaust dampers have placed as shown in figure 5.The mesh types are the same (0.5 m and 2.5 m) in both congested and empty tunnel traffic.The mesh choice was carried out by comparing the CFD results and experimental ones (reference model) provided by Hang et al. [5], in terms of air flow velocity and total static pressure.The figure 6 shows the good agreement between the CFD simulation and CFD-JG Hang (validated by experimental results) in terms of air velocity and gauge pressure.This result allows the CFD model validation.Figure 7 shows the air average velocity and mass flow trends over time; one note that the stationary condition is reached after about 300 s.During the simulation, only one Saccardo fan was kept running, the one closest to the left wall of the tunnel, with the corresponding extraction module active [5].

CFD INVESTIGATION
In this paragraph, the Performance of the Ventilation System "Pvs" versus the magnitude and direction (pitch angle) of the inflow velocity vs is discussed.The Pvs is evaluated for different inflow air direction (pitch angle) in congested traffic condition and velocity magnitude in empty tunnel condition.About the first case, the cars are modelled by two overlapped parallelepipeds.The vehicle dimensions and their positioning inside the tunnel are reported in the table 2. In the second case, different velocity magnitude is achieved just reducing the inflow opening, with constant mass flow rate.

Pitch angle influence analysis
Three different pitch angles, 8°, 10° and 12°, were investigated numerically to evaluate the interaction between the air flux direction and cars with the consequently pressure losses.The pitch angle has not been further reduced (less than 8°) because the high jet velocity tends to deflect the jet towards the tunnel ceil due to the 'Coanda effect [8] increasing the friction losses and it has not been further increased (greater than 12°) because the pressure drops tends to increase due to the air flow of the Saccardo system directed on the cars.The figure 8 (left side) shows that the mass flow rate ̇  is not influenced significantly by the pitch angle and consequently the Pvs.The figure 8 (right side) shows that the pitch angle 8° degree seems improve slightly the ventilation in the road tunnel.Therefore, in the following a pitch angle of 8° is adopted.

Influence of inflow opening length of the Saccardo ventilation system
Having fixed the angle of inclination equal to 8°, we have investigated the influence of the inflow opening length reduction on Saccardo system performance.The different configurations investigated consist in reduction of the opening length of n x 1.8 m, for n=1 to 5, starting from the base configuration of 21 m.The results are provided in terms of Pvs vs opening length reduction (figure 9, left side) and Saccardo outflow velocity vs opening length reduction (figure 9, right side).We can note that as the opening length decreases the Pvs and Saccardo outflow velocity increase.The table 3 shows the influence of different inflow opening lengths on the mass flow rate at the outlet tunnel portal ̇out and consequently Pvs.

Conclusion
In this work the authors have presented the results of a 3D numerical investigation carried out on a Saccardo ventilation system operating in road tunnel.The boundary conditions and the geometrical dimensions of the ventilation system were the same of that investigated experimentally by Hang et al. [5] used to validate the CFD model.Through the CFD investigation, the influence of the pitch angle  and the jet velocity   were investigated in terms of Performance of the Ventilation System "Pvs", defined as ratio between the mass flow rate at the outlet tunnel portal ( ̇ ) and the mass inflow provided by Saccardo ( ̇).The results have shown that the pitch angle of the Saccardo nozzle influences moderately the Pvs, instead different opening inflow lengths influence considerably the mean air velocity in the tunnel and mass flow rate.A Pvs trend vs opening inflow length has shown an improvement up to 47% for an inflow length of 14.2 m (6.8 m inflow length reduction), this allows to obtain a significant energy saving.
on the right side) are arranged in side-by-side configuration, both placed in the middle of the tunnel at a distance of 500 m from inlet portal.The arrangement and geometrical dimensions of the ventilation system are shown in figure2and table 1; they are the same of that investigated experimentally by Hang et al.[5] and here used with the aim to validate CFD model.Both exhaust dampers and Saccardo nozzles engine are built by HVAC Vents connected to each other by ducts as shown in figure2.The ducts represent a continuous flow path, instead the nodes attached to vents surface represent the interface between the 1D-model and 3D-CFD regions.

Figure 3 .
Figure 3.Control volume.(  −   )  = ̇    − ̇    + ̇ ℎ  ℎ − ̇  cos()(1) where ̇  and ̇  are the mass flow rates at the outlet and inlet tunnel portal, respectively, and ̇̇ ℎ are the the mass inflow and outflow provided by Saccardo and exhaust respectively, and taking into account the pressure drops to the portals (equations 2 and 3):

) 3. CFD model 3 . 1 .
Mathematical model description The time Averaged mass Navier-Stokes combined with LES Standard model equations (8-10) describe the unsteady and turbulent states and three-dimensional flux, in the fluid region.These equations model the motion of viscous fluid substances:

Figure 5 .
Figure 5. Mesh size and connection of the two types of mesh adopted.

Figure 8 .
Figure 8. Left side: Mass flowrate vs time with pitch angle as parameter, velocity profile of air in the tunnel at right side: different 'x' distance from Saccardo inflow.
Parameters values wS Width of inflow duct of Saccardo nozzle 2.8 m Hduct High of inflow duct of Saccardo nozzle 2.2 m Lduct Length of duct of Saccardo nozzle 18 m Linflow Length of inflow area of Saccardo nozzle 21 m  Pitch angle 10° Aexhaust Exhaust dampers area 9.0 m 2 DS-d Distance between Saccardo Nozzle and Exhaust dampers 16 m  ̇ Saccardo mass flow rate kg/s  ̇ Exhaust mass flow rate kg/s ̇    − ̇    + ̇ ℎ  ℎ + −̇  () =

Table 2 .
Car dimensions and positioning inside the tunnel

Table 3 .
Inflow opening lengths vs mass flow rate and Pvs fixed the   ̇(=202 kg/s)