- 110 - AIR BARRIERS USED FOR SEPARATING SMOKE FREE ZONES IN CASE OF FIRE IN TUNNEL Gregory Krajewski 1 1 Building Research Institute Fire Research Department, Poland ABSTRACT The aim of this paper is to take the advantage of CFD application in calculating, optimising, and designing air curtains used to separate smoke free zones in case of fire in tunnel. Air curtains can be a good solution in case when the usage of solid obstructions is not feasible (for example in a big tunnel).a properly designed air curtain produces a pressure drop which prevents transversal flow through the opening. An accurate CFD calculation of an air curtain is challenging because of the high air velocity and relatively thin nozzle. Most air curtains are tested on scaled down models which are difficult to extrapolate. Tests in a real scale model are performed and the tests results are used to verify the chosen turbulence model. The intention of this paper is to present the comparison between the CFD calculations and tests results. Key words: air barrier, CFD calculations, fire, tunnels 1. INTRODUCTION To avoid standstill and facilitate the flow of vehicles and people through doorways of buildings and other enclosures, solid doors are often replaced or supplemented by air curtains (air screens, air planes). Simultaneously, the air screen eliminates or reduces the transfer of heat and mass through the opening. Air curtains have become popular in the 60 s of the 20th century; nevertheless, the principles of the air planes dates back to 1904. The concept of air screens was founded by Theophilus van Kennel [3][4][5] and his idea has become a forerunner of modern air curtains. The flow of air across the doors is caused by the difference of pressure between two volumes of fluid, the dissimilar temperature values, and the presence of a ventilation system. Air curtain devices are often used in the entrances to public buildings, cooling rooms and refrigerators, as well as in chemical and electronic industry. The knowledge that a direct exposure to fire is not the most immediate threat to people s lives, has been displayed by previous experience and research. A vast majority of fatalities connected with fire are triggered by the smoke-inhalation. Therefore, to decrease the number of fatalities air curtains can be used as virtual screens to stop smoke spreading in a building. Safe evacuation of people and secure intervention of fire fighters are of significant importance and need to be taken into consideration [11]. Air screens can be used to obstruct or reduce the movement of toxic smoke while enabling full access to emergency exits. Properly designed air curtains produce a pressure drop which preventstransversal flow through the opening (Fig.1). Figure 1: Illustration of zone separation
- 111 - There are many applications of such kind of solution in example. tunnel junctions, elevators connected to the stations in underground tunnels, separation between tunnel and station.(fig. 2) Figure 2: Application of air curtain The main criterion for the air curtain is its efficiency which is the rate of mass and heat transfer crossing it, in comparison with the same opening without the curtain. Air curtain design is considered relatively difficult. It is essential to know the conditions on both sides of the air curtain to choose its parameters correctly. If the outlet velocity is too high and the blowing angle is not optimal, the air curtain can increase the heat and mass transfer through itself. On the other hand, if the jet velocity is too low, the air curtain will not be tight enough. Currently, most of the installations are experimentally set up on scaled down or full scale physical models. Nonetheless, it is hard to extrapolate results from the scaled down model to the different geometrical dimensions because the Euler number similitude is unavailable. As it is written in literature that kind of extrapolation provides an wrong results by creating to high or to low speed [5]. To study the performance of air curtains the application of computational fluid dynamics (CFD) is useful. It is critical to properly define initial and boundary conditions and compare gained results from simulations with analytical equations. 2. AIR CURTAIN DESCRIPTION 2.1. Velocity distribution There are numerous publications involving experimental data and mathematical analysis presenting the theory of a free stream jet as velocity profile and deflection of the centreline axis. Significantly in this subject, publications of Abramovich (1963) and Rajaratnam (1976) offer the most fundamental piece of information [1]. Particularly, depending on the height and the stream of air, a jet shows two, three or four regions. It is possible to distinguish between the potential core zone, the transition zone, and the developed zone or the impinging zone (Fig.3):
- 112 - Potential core zone: characteristic for this region is that the centreline velocity is almost constant and equal to the outlet velocity U 0. Transition zone; this region starts with the velocity decay and the amplification of the jet expansion. It generally starts after approximately 5 e from the nozzle. Analytical solution of the velocity can be described by: e y + U ( x, y) 1 = 1 + erf 2 1 U 0 2 x σ Eqs. 1 e - characteristic dimmension of nozzle [m] U o - outlet velocity on the nozzle [m/s] Developed zone in this region velocity decay remains constant. Velocity decay expressed with non-dimensional quantities. It generally starts after approximately 20 e from the nozzle. Analytical solution of the velocity can be described by [10]: U x, y) = U 3 2 7.67e 1 + tanh x ( 2 0 0 ( x) x = C1 C2 U e 1 2 y 7.67 x Eqs. 2 U C Eqs. 3 C 1 and C 2 depends on the nozzle shape and on the boundary conditions. They are in the range 1.9 < C 1 < 3.0 and -8 < C 2 < 10 Impinging zone: this region is in the vicinity of the floor. The flow in that zone is very complex and still not well known. The thickness of that zone is approximately 15% of the total height. core zone transition zone developed zone impinging zone Figure 3: Zones of a free air jet 2.2. Deflection of jet axis According to F. C. Hayes and W. F. Stoecker [7][8], the problem of a plane jet, subjected to a lateral side pressure and the blowing angle directed to the side of higher value of the pressure is treated in the literature with the conclusion that the axis of the jet represents a circular curvature Looking at picture below (Fig. 4) there are 8 variables describing the problem. Air curtain can be set up 5 non-dimensional numbers which are representative of the phenomena. There are Euler number, geometric aspect ratio, the Reynolds number, turbulence intensity at exit and blowing angle.
- 113 - Figure 4: Air curtain ΔP H U 0e = f,, I 0, α 1 2 ρu e ν 0 2 g c - gravity [m/s²] ΔP - pressure difference [Pa] ρ - density {kg/m³] e - characteristic dimension of nozzle [m] U o - outlet velocity on the nozzle [m/s] α - nozzle angel [ ] ν kinematic viscosity [m²/s] Eqs. 4 3. EXPERIMENT Before studying the air curtain used to prevent smoke movement, the numerical model isverified.the verification was done in a tunnel of 8 m length, 1 m width, and a 2 m height. The test equipment enables to carry out tests with a blowing angle of the jet varying from 0-45 and velocities up to 30 m/s. The nozzle width is 20 mm (changeable). The pressure difference on both sides of the air curtain could be set in a range from 0 Pa to 200 Pa. A schematic drawing of the test facility is presented below in Figure 4. The real scale model was build in Fire research Department in Building Research Institute (Fig. 5). Figure 5: Illustration of test facility
- 114-3.1. The free air jet The initial test, which was done to validate the numerical model, was a free jet test. Testes were carried out for three different air velocities at the nozzle outlet (10 m/s, 20 m/s and 30 m/s). Referring to free jet centreline velocity decay and transversal distribution of the jet velocity were compared with the CFD results. The velocity measurements in the centreline were done using hot wire thermo anemometry with three wires to measure velocity in two directions (Fig.6) and viualisation.. Figure 6: Velocity at the nozzle outlet 4. SIMULATION In order to evaluate the parameters and the effectiveness of the air curtain, it is necessary to perform series of calculations using the CFD. To confirm the correctness of the boundary conditions and CFD models, numerical calculations of free jet were performed. Defined initial conditions were the same as in the experiment. Blowing angle of the jet is equal to 0, the outlet velocity from the nozzle is 10 m/s, 20 m/s and 30 m/s. Turbulence intensity is the same as in experiment (about 5%). The three-dimensional model of the analyzed domain was build according to the experimental setup. In the middle of the ceiling an air curtain outlet was created. The domain has been divided into a finite number of control volumes using an unstructured hexahedral grid. The total quantity of control volumes was approximately 2 500 000 with dimensions ranging from 2 mm in the area of the air curtain outlet to 20 mm on the peripheries of the domain. The three-dimensional model is presented in Figure 7. Figure 7: View of the CFD model
- 115 - The numerical simulations were conducted using ANSYS Fluent 14.5. software. RANS k-ε Realizable turbulence model was applied in the calculations. Boundary conditions are defined as walls for the side, bottom and top of the analysed domain. Additionally, the air curtain slot is defined as velocity inlet; the upstream and downstream part of a model are pressure inlets. Various configurations to check the behaviour of the air stream have been investigated. For instance, diverse pressure differences and outlet velocities are simulated.the results from the CFD calculations were compared with those from the experiment. 5. RESULTS The effects of the numerical calculations of the free jet are in good agreement with the examined results from the experiments with exception of the first part of the stream for almost all turbulence models (k-ε. k-ω, LES). As illustrated in Figure 8, the outcomes of the centreline velocity reveal a significant similarity to those obtained in the experiment. In order to visualize the velocity profile, the velocity counters in cross-section of analytical domain are presented in figure 9. Comparing detailed analyzes of measurements it seems that time averaged RANS k-ε model describes air curtain phenomenon very well. Figure 8: Centreline Velocity (α=0, U=10 m/s) Figure 9: Velocity counters in cross-section of numerical domain The experimental result of using the air curtain as a separation of two zones (one full of smoke and one free of smoke) is shown in Figure 10.
- 116 - free of smoke zone full of smoke zone Figure 10: Division of two zones using air curtain Visualisations of the air curtain shapeunder the influence of lateral pressure are given in Figures 11 and 13. Figure 11: Velocity profile in cross section e=5 cm, U= 10 m/s Figure 12: Temperature range in cross section e=5 cm, U=10 m/s (K) Figure 13: Velocity profile in cross section e=10, U=10 m/s Figure 14: Temperature range in cross section e=10 cm, U=10 m/s (K)
- 117 - In Figure 12 and 14 the simulation results of a division of two separate zones of fluid is presented. One of them has higher temperature then second one, on the opposite side of the air curtain, is not higher then 5 K that it was in the initialisation moment. On the left side of air curtain (normal temperature) an underpressure was defined, that why the jet has an arc shape. The amount of heat (smoke) which is transferred through the air curtain is significantly low. More detailed analysis of air curtain leakages will be done in further research using CFD calculations [12] and scale model tests. 6. CONCLUSION This paper demonstrates the possibility of using CFD methods to properly design and analyse air barrier. According to conducted simulations, it is crucial to declare that air curtains can be used as a division of smoke free zones in case of fire. In addition, properly defined boundary conditions and chosen turbulent model have fundamental influence on received effects. Research has been done in the framework of the project "Innovative measures and effective methods to improve the safety and durability of buildings and transport infrastructure in the strategy of sustainable development" co-financed by the European Union from the European Regional Development Fund under the Innovative Economy Operational Programme. 7. REFERENCES [1] Abramovitch G.N.: The theory of turbulent jets. Massachusetts M.I.T. press, 1963 [2] ANSYS Fluent 13.0.0. Technical Documentation [3] Goyonnaud L., Solliec C.: Mass transfer analysis of an air curtain system, Transactions on Engineering Science vol. 18, 1998 WTT Press [4] Gugliermetti F., Santrapia L., Zori G.: Air curtain applied to fire smoke pollution control, Transactions on Ecology and the Environment vol. 66, 2003 WTT Press [5] Guyonnaund L., Solliec C., Dufresene de Virel M., Rey C.: Design of air curtains used for area confinement in tunnels, Experiments in Fluids 28 (2000) 377-384 Springer- Verlag 2000 [6] Gupta S., Pavageau M., Elicer-Cortes J. C.: Cellular confinement of tunnel sections between two air curtains, Building and Environmant 42 (2007) 3352-3365 [7] Hayes F. C., Stoecker W. F.: Heat Transfer Characteristics of the Air Curtain, ASHRAE Transactions No. 2120, 1969 [8] Hayes F. C., Stoecker W. F.: Desig Data For Air Curtains, ASHRAE Transactions No. 2121, 196 [9] Rajaratnam N.: Turbulent jets. Elsevier, Amsterdam, 1976 [10] Schlichting: Boundary layer theory.mc Graw-Hill Book, New York, 1968 [11] Sztarbala G., Krajewski G.: Application of CAE in designing process of fire ventilation system based on jet fan system, EBECC London, UK, 2009 [12] Krajewski G. Verification of CFD turbulense model used to simulations of air curtain, Building Physics in Theory and Practise 2013