NUMERICAL SIMULATIONS OF FLOWS IN PUSH-PULL FUME CUPBOARD

Transcription

1 NUMERICAL SIMULATIONS OF FLOWS IN PUSH-PULL FUME CUPBOARD Ming-Jyh CHERN, and Wei-Ying CHENG Department of Mechanical Engineering, National Taiwan University of Science and Technology, Taipei 10607, Taiwan ABSTRACT A push-pull fume cupboard shown is an innovative device to remove pollutants from a fume cupboard to the outdoor environment. A push flow coming from the bottom of a sash and a pull flow behind the doorsill are adopted to form an air curtain in the fume cupboard. Meanwhile, the roof of a push-pull fume cupboard is open to the environment and ambient fluids are drawn from the roof. To investigate the performance of a push-pull fume cupboard, a numerical model based on the finite volume method is utilized. Various ratios of push flows to pull flows are considered. Four characteristic flow fields are found at a variety of speed ratios. They are concave curtain, straight curtain, under suction and over blow modes. According to numerical results, a concave curtain mode with a fast pull flow and a weak push flow is recommended for the operation of a push-pull fume hood. Furthermore, the manikin effect on the flow filed of a push-pull fume hood is also examined in this study. A manikin is located in front of the fume cupboard according to the ANSI/ASHRAE 110/1995 standard. Numerical results reveal that fluid flows in a push-pull fume cupboard are not affected by a manikin. As a result, an operator standing in front of a push-pull fume cupboard is protected from pollutants inside the cupboard. KEYWORDS push-pull fume cupboard, manikin effect, and turbulent diffusion INTRODUCTION A fume cupboard is a common local ventilation device at a laboratory. The main purpose of a fume cupboard is to remove toxic pollutants from the laboratory using a capture flow. A common way to generate a capture flow is to use a top fan at the roof of a fume cupboard. A suction flow is generated at the roof and ambient fluids are drawn from the opening of the fume cupboard. Due to the capture flow, pollutants inside the fume cupboard can be removed to the outdoor environment. A design example for an exhaust fume cupboard can be found in ACGIH (2004). Researches regarding flow variations inside an exhaust fume cupboard and performance are available in several references such as Fletcher and Johnson (1992 a and b), Ivary et al. (1989), Durst & Pereira (1991), Ekberg & Melin (1991), Ozdemir et al. (1993), and Tseng et al. (2006). When the capture flow enters an exhaust cupboard, the direction of the flow is changed and the angle between the inlet and outlet directions is almost 90 o. Due to the steep change of the flow direction, several vortices are formed inside an exhaust cupboard. The structure of vortices has been reported in Hu et al.(1998), Kirkpatrick (1998), and Graham et al. (1998). As a result, they not only remain some pollutants inside an exhaust cupboard but also cause the leakage of pollutant from edges of the opening. Moreover, the leakage of pollutants becomes more serious when an operator stands in front of the opening of an exhaust cupboard. An operator in front of the cupboard is an obstacle in the captured flow. A pair of vortices is formed between the operator and the cupboard. Due to the pair of vortices, pollutants are Corresponding Author: Tel: , Fax: address: mchern@mail.ntust.edu.tw

2 accumulated in front of the operator. Consequently, an exhaust cupboard can not prevent the operator from breathing pollutants. Since the formation of inner vortices and the pair of vortices in front of an operator is due to the capture flow, rearrangement of the capture flow may improve a fume cupboard. The main purpose of this study is to introduce the push-pull air curtain technique in a fume cupboard. The push-pull air curtain technique has been adopted in capture of pollutants coming from an open surface tank with a large area. Its relevant studies can be found in several references such as Malin (1945), Battista (1947), Hama (1957), Huebener & Hughes (1985), Robinson & Ingham (1995a, b), Rota et al. (2001), Huang et al. (2005), and Chern & Ma (2006). The basic principle of the push-pull air curtain is to use a weak push flow and a strong pull flow at two sides of the open surface to form an air curtain. The air curtain prevents escape of pollutants and transports them to the pull sides. Huang et al. (2006) proposed the push-pull air curtain technique to a fume cupboard. A push flow is arranged at bottom of the sash and a pull flow is located behind the doorsill of the fume cupboard. Furthermore, the roof of the fume cupboard should be open to the indoor environment. Therefore, a push-pull air curtain is formed and parallel to the sash and the operator. Clean air is drawn from the opening and the roof. Subsequently, pollutants inside the cupboard are removed by the pull flow. To verify the capture performance and details of the flow field in a push-pull fume cupboard, a numerical model based on computational fluid dynamics is established. The verifying procedure is according to the ANSI/ASHRAE 110/1995 standard. A variety of push and pull flow speeds are considered to investigate flow features and the distribution of pollutant concentration. Moreover, the manikin effect is considered to examine if the performance of the fume cupboard is reduced. The final goal of this study is to find an optimal mode to operate a push-pull fume cupboard and to prevent an operator from toxic pollutants. MATHEMATICAL FORMULAE AND NUMERICAL APPROACHES To study flows in a fume hood numerically, governing equations for fluid motions have to be established first. An incompressible fluid is considered in this study. All fluids simulated in the numerical model should obey the laws of conservation of mass and momentum. Under normal circumstances, the air flow inside a fume cupboard should be turbulent, so Reynolds decomposition is used to obtain a Reynolds averaged equation for the mass conservation, which is denoted as U = 0, = 1, 2, 3 (1) x where U is a mean velocity component. Furthermore, the Unsteady Reynolds Averaged Navier-Stokes (URANS) equations are presented as U ( U U ) i i P U i ' ' ρ + = + µ ρu u, i, = 1, 2, 3 i (2) t x xi x x where P is the mean pressure, ρ is the density of the fluid, and µ is the dynamic viscosity of the fluid. It ' ' is found that new variables, ρu u i, called the Reynolds stresses appear in Eq. (2). We do not have sufficient equations to solve these unknowns. To close this problem, Boussinesq's approximation is adopted whereby

3 W 1 = 59 cm W 2 = 374 cm W 3 W 3 = 26.5cm H 1 = 60 cm H 2 = 132 cm H 3 sash W 1 H 3 = 445 cm H 1 H 2 doorsill slit W 2 Figure 1 Schematic of a push-pull fume cupboard. ' ' U U i 2 uiu ν = t + kδ i, (3) x x i 3 where ν t and k are the eddy viscosity and the turbulence kinetic energy, respectively. Moreover, the k-ε two-equation model proposed by Jones and Launder (1972) is used to determine the kinematic eddy viscosity. The turbulent mass transfer equation based on Reynolds' decomposition approach is used to explore concentration variations and is denoted as C U C C ' ' + = D u c, = 1 ~ 3 m t x x x (4) where C and c' are the temporal mean and fluctuating value of concentration, respectively. The concentration fluctuation term ' ' C u c = Dt x ' c ' u is proportional to the concentration gradient, i.e.,, (5) where D t is the turbulent diffusion coefficient. Since turbulent flows are complex and highly changeable, regarding D t as a constant is usually unreliable for most applications. Therefore, we adopt the turbulent Schmidt number (Sc) defined as the ratio of the turbulent momentum diffusion rate to the turbulent mass diffusion rate, i.e. ν t /D t. The turbulent Schmidt number relates to the ratio of the viscous diffusion to mass diffusion, and can be assumed as a constant. Depending on the situation under consideration, this value will vary from 0.8 to 1.2 (see Durst and Pereira, 1991). We use Sc = 1 for the following studies. The continuity equation, URANS equations, and the turbulent mass transfer equation are solved numerically for the present study. Air is the working fluid, with density and dynamic viscosity as kgm -3 and 1.855x10-5 kgm -1 s -1 at 25 o C, respectively. The physical domain is 3.73x3.89x3.16 m 3. Dimensions of the push-pull fume cupboard can be found in Figs. 1. A rectangular Cartesian coordinate system is used for the solution and its origin is located at the center of the doorsill.

4 Y 1.5 (m) X (m) Figure 2 The computational mesh at the vertical central plane of the push-pull fume cupboard. Computational meshes consisting of 545,228 cells for the push-pull fume cupboard are used. Figure 2 shows the 2-D central plane of the meshes for the push-pull fume cupboard. Cells are clustered densely inside the cupboard. Constant pressure 1 atm is imposed at the far field boundaries. Various speeds of push and pull flows are used for the boundary conditions of a push-pull fume cupboard. The push flow condition is imposed at the bottom of the sash. A pull flow channel is considered in the model and the pull flow condition is imposed at the exit of the channel instead of the slot behind the doorsill. The wall functions are used to determine boundary conditions at solid walls. Variations in velocity and concentration are predominantly normal to solid walls. In addition, SF 6 is used as a tracer gas in the numerical model. Its density and molecular diffusivity in air are 6.04 kgm -3 and 3x10-5 m 2 s -1. The gravitational effect is not considered in this study. The boundary condition of concentration at the exit of the eector is 100 %. A quiescent flow field is used as the initial condition in the model. Simulations do not terminate until steady state solutions are reached. For each steady solution, the ratio of the maximum temporal variation of velocity to its magnitude is less than The software STAR-CD TM based on the finite volume method is employed to solve Eqs. (1), (2), and (4). The 2nd order Crank-Nicolson scheme and the QUICK scheme are used for the temporal and advective terms, respectively. The PISO scheme is utilized for the pressure-velocity iteration. The maximum mass residual from all computational cells must be less than RESULTS AND DISCUSSION Flow Modes and Concentration Distribution of Push-pull Fume Cupboard A variety of parameters including push and pull flow speeds denoted as V b and V s and sash opening are considered in this study to investigate the performance of a push-pull fume cupboard. Essentially, all flow variations can be categorized into four modes, i.e. concave curtain, straight curtain, under suction, and over-blow modes. Figure 3 presents vertical streamline plots of a concave curtain example in a fully open fume cupboard. The pull flow is much faster than the push flow (V b = 3 ms -1 and V s = 14 ms -1 ). The resultant air curtain starts straight at the bottom of the sash and becomes concave in the vicinity of the pull side, i.e. the front bottom of the cupboard. Figures 3 (a)-(d) show the streamline plots of various vertical planes from the central section to the side wall successively. Only a small vortex denoted as E a occurs at the rear bottom of the cabinet and consists of clean air inhaled from the

5 Figure 3 Streamline plots the concave mode at various vertical planes. Figure 4 Streamline plots of the under suction mode at various vertical planes. environment. The concave curtain mode is recommended for operating a push-pull fume cupboard. When the speed ratio of push to pull flows becomes large, the flow in the fume cupboard becomes a straight curtain mode as shown in Figures 4(a) - (d)(v b = 5 ms -1 and V s = 12 ms -1 ). It is found that a vortex occurs above the doorsill and in front of the air curtain. Due to the air curtain, the vortex consists of clean ambient fluids. This is the only difference compared with a concave curtain mode. All inhaled air from the top can be pulled to the bottom along smooth streamlines. As a result, pollutants coming from the eector are captured by the flow and pulled into the slot behind the doorsill. However, a straight curtain mode is not appropriate for the operation of a push-pull fume cupboard due to the manikin effect discussed later. Furthermore, an under-suction example appears as shown in Figures 4 (a) - (d) as the speed ratio increases (V b = 2 ms -1 and V s = 4 ms -1 ). The air curtain still exists and becomes concave near the bottom of the cupboard. Nonetheless, a primary vortex denoted as E d is found inside the fume cupboard. Consequently, inhaled fluids from the top cannot be transported to the pull slit and brought out. In other words, no mechanism can remove pollutants from the fume cupboard in a under-suction mode. Therefore, it is recommended that a fume cupboard not be operated in an under-suction mode. If the push flow is faster than the pull flow, then an over-blow mode will be observed. Due the page limit, it is not presented in the manuscript. Not only does a huge primary vortex occur inside the fume cupboard, but a vortex also develops in front of the air curtain. Due to the primary vortex, pollutants cannot be brought out from the fume cupboard. In fact, pollutants may be transported to the environment by the primary vortex from the top of the fume cupboard. This mode is certainly not recommended for operating a push-pull fume cupboard. To observe whether pollutants can be brought to the pull slit, streamlines starting from the exit of the eector are drawn to visualize possible pollutant paths. Figures 5 (a) - (d) present 3-D streamlines in various modes. Obviously, streamlines are pulled to the slit at the bottom soon after exit from the

6 Figure 5 3-D streamline plots of four modes. Figure 6 Characteristic figure for a fully open push-pull fume cupboard. eector in the concave curtain and straight curtain modes. Nonetheless, streamlines in a under-suction or over-blow mode are strongly affected by vortices inside the fume cupboard in Figures 5 (c) and (d). Pollutants coming from the eector remain inside the fume cupboard. Characteristic Figures of Push-pull Fume Cupboards Flow structures inside a push-pull fume cupboard are affected by the speed ratio of push-pull flows and the opening. Various ratios and openings are considered in this study in order to observe their effects. Figure 6 indicates the characteristics of flow patterns obtained at various speed ratios in a fully open fume cupboard. Numerically predicted results are compared with Huang et al.'s (2006) experimental results. The agreement between numerical and experimental results is acceptable as can be seen in Figure 6. An operator may refer to those figures when they use a push-pull fume cupboard. Characteristic figures for other openings from 25% to 75% are available upon request. As

7 Figure 7 Comparisons of vertical concentration contours between a conventional (a) - (d) and a push-pull fume cupboard (e) - (h) at the same flow rate 0.5 m 3 s -1. The push and pull speeds are 2 and 12 ms -1. This speed ratio is in a concave curtain mode. mentioned in the previous section, one should adust the speed ratio of push-pull flows to a concave curtain mode according to various openings. Comparisons between Exhaust and Push-pull Fume Cupboards The manikin effect on the capturing capability of an exhaust fume cupboard is extremely negative. This is because a manikin plays the role of an obstacle in a uniform fluid flow. As a result, a pair of vortices occurs between the manikin and an exhaust fume cupboard and pollutants accumulate in front of the manikin. To investigate the manikin effect on a push-pull fume cupboard, a manikin is located 7.5 cm in front of the sash according to ANSI/ASHRAE Standard As under suction and over-blow modes are not appropriate for operation, only a concave curtain mode is examined and compared with an exhaust fume cupboard including the manikin effect. The push-pull fume cupboard is designed to overcome the disadvantages of an exhaust fume cupboard. We now examine whether this aim has been achieved. These cupboards are in operation at the same volumetric flow rate of suction, 0.5 m 3 s -1. An upward flow appears in front of the manikin in an exhaust fume cupboard, so an operator may breathe pollutants transported by the upward flow. There is no such upward flow in a push-pull fume cupboard. The flow in front of the manikin is smoothly pulled into the slit at the bottom. Moreover, Figures 7 (a) - (h) reveal the vertical concentration contours of those fume cupboards. The local concentration values in front of the manikin in an exhaust fume cupboard are extremely high as shown in Figures 7 (a) - (d). In contrast, the local concentration values in front of a manikin in a push-pull fume cupboard are less than 1 ppb. Since pollutants may also leak out of the cupboard from gateposts and the doorsill, the ANSI-ASHRAE Standard does not define the required minimum concentration values at the rest areas on the front face of the sash. In order to investigate variations of concentration at the rest areas, twelve points on the vertical face of the sash are examined. Figure 10 shows an example in a concave curtain mode. Hollow and solid symbols refer to the local concentration values in a push-pull fume cupboard and an exhaust fume cupboard, respectively. Consequently, all local concentration values in a push-pull fume cupboard are less than 0.1 ppm. The values at P2, P3, P6, and P7 are even less than 1 ppb, so no hollow symbols appear in Figure 8. In addition, an exhaust fume cupboard obviously does not satisfy the requirement of 0.1 ppm except for P6 and P7. In general, an exhaust

8 Figure 8 The local concentration values at 12 selected points in a concave curtain mode. The push and pull flow speeds of the push-pull fume cupboards are 2 and 12 ms -1, respectively. The pull flow rate of the push-pull fume cupboards is as same as the exhaust fume cupboard. fume cupboard does not perform better than the corresponding push-pull fume cupboard in terms of the local concentration values at those selected points. All those results show that a push-pull air curtain fulfills its purpose successfully. CONCLUSIONS A push-pull fume cupboard for removing toxic pollutants is introduced and examined. A numerical model based on the finite volume method and the standard k-ε turbulence model is utilized to investigate the flow and concentration fields inside the fume cupboard. The flow variations with respect to a variety of push-pull flow speed combinations are classified into four characteristic modes, i.e. concave, straight, under-suction, and over blow modes. The concave mode with a weak push flow and a strong pull flow is recommended to use for the operation of a push-pull fume cupboard. The manikin effect is examined in the push-pull fume cupboard. In terms of comparisons with an exhaust fume cupboard, the push-pull fume cupboard is not affected under the manikin effect. The concentration measurements at twelve points on the opening face of the push-pull fume cupboard are all less than the standard 0.1 ppm even under the manikin effect. ACKNOWLEDGEMENTS We are grateful to National Science Council Taiwan for the financial support (NSC E ). Computational resources provided by National Center for High-Performance Computing Taiwan are appreciated. REFERENCES 1. American Conference of Governmental Industrial Hygienists (ACGIH) (2004) Industrial ventilation - a manual of recommended practice, published by American Conference of Governmental Industrial Hygienists 25th ed., Chapter 3, pp B. Fletcher and A.E. Johnson (1992a) Containment test of fume cupboards - I methods, Annals of Occupational Hygiene, Vol 36, B. Fletcher and A.E. Johnson (1992b) Containment test of fume cupboards - II Test room

9 measurements, Annals of Occupational Hygiene, Vol 36, R.E. Ivary, M.W. First, and L.J. Diberardinis (1989) A new method for quantitative in-use testing of laboratory fume cupboards, American Industrial Hygiene Association Journal, Vol 50, F. Durst and J.C.F. Pereira (1991) Experimental and numerical investigation of the performance of fume cupboard, Building and Environment, Vol 26, L.E. Ekberg and J. Melin (1991) Required response time for variable air volume fume board controllers, Annals of Occupational Hygiene, Vol 44, P. Hu, D.B. Ingham, and X. Wen (1998) Effect of baffles and a louvered bypass on the airflow and the convective patterns of contaminant inside a fume cupboard, American Industrial Hygiene Association Journal, Vol 59, A.T. Kirkpatrick (1998) Numerical simulation of laboratory fume cupboard airflow performance, ASHRAE Transactions, Vol 104, P.N. Graham, P.C. Raymond, and L.D. Mervyn (1998) Computational fluid dynamics as a method for assessing fume cupboard performance, Annals of Occupational Hygiene, Vol. 44, B.S. Malin (1945) Practical pointers on industrial exhaust systems, Heat and Ventilation, Vol. 42, W.P. Battista (1947) Semi-lateral tank ventilation cupboard controls contamination, cut costs, Heat Piping Air Condition, Vol. 19, G.M. Hama (1957) Supply and exhaust ventilation for metal picking operations, Air Condition, Heat Ventilation, Vol. 54, M. Robinson and D.B. Ingham (1995a) Numerical modeling of the flow patterns induced by a push-pull ventilation system, Annals of Occupational Hygiene, Vol. 40, M. Robinson and D.B. Ingham (1995b) Recommendations for the design of push-pull ventilation systems for open surface tanks, Annals of Occupational Hygiene, Vol. 40, R. Rota, G. Nano, and L. Canossa (2001) Design guidelines for push-pull ventilation systems thourgh computational fluid dynamics modelling, AIHA Journal, Vol. 62, R.F. Huang et al. (2005) Aerodynamic characteristics and design guidelines of push-pull ventilation systems, Annals of Occupational Hygiene, Vol. 49, M.J. Chern and C.H. Ma (2007) Numerical investigation and recommendations for push-pull ventilation systems, Journal of Occupational and Environmental Hygiene, Vol. 31, R.F. Huang et al. (2006) Development and evaluation of air curtain fume cabinet with its consideration of aerodynamics, accepted by Annals of Occupational Hygiene. 19. W.P. Jones and B.E. Launder (1972) The calculation of low-reynolds number phenomena with a two-equation model of turbulence, International Journal of Heat and Mass Transfer, Vol F. Durst and J.C.F. Pereira (1991) Experimental and numerical investigation of the performance of fume cupboard, Building and Environment, Vol. 26,