1、4676 Comparison of Diffusion Characteristics of Aerosol Particles in Different Ventilated Rooms by Numerical Method Bin Zhao, Ph.D. Xianting Li, Ph.D. Member ASHRAE ABSTRACT Particle difusion with gravitational sedimentation in displacement and mixing ventilated rooms is investigated numerically. Th
2、e driftflux model, which considers the settling ofparticles under the efect of gravitational sedimentation, is adopted to simulate particle dtfusion, while the simplijied model for solving the continuous jluidjlow is combined. Since the PM 2.5 and PM 10 particles are mostly concerned in indoor envir
3、onment, passive contaminant and 2.5-20 micron particles are investigated in this paper. The numerical results show that in a mixing ventilated room, the distribution of nonpassive particles does not difer much fiom that ofpassive contaminant when the particle diameter is less than 20 microns. Meanwh
4、ile, in the displacement ventilated room, the gravitational settling should be taken into account when particle diameter is larger than 10 microns. Comparison also demonstrates that displacement venti- lation will bring particle contaminants generated by manne- quins to the working zone and upper pa
5、rt of the room; in contrast, mixing ventilation will carry up the pollutant in a relatively small amount. INTRODUCTION Aerosol particles are a ubiquitous pollutant indoors and outdoors around the world and are regarded as significant pollutant sources in the indoor environment. The aerosol parti- cl
6、e concentration in a room greatly influences the indoor air quality (IAQ). It is known that aerosol particles may be suspended in an occupied area so that they may be inhaled by the occupants and deposited on the nasal passage with poten- tial harmful effects. Thus, predicting and controlling diffus
7、ion Zhao Zhang Dongtao Huang, Ph.D. characteristics of indoor aerosol particles is important for ventilation design and indoor environment evaluation. Aerosol particles are often assumed to be passive contam- inants; that is, they are assumed to move in the same manner as the airflow (Hu et al. 2002
8、). This assumption may be adequate when predicting the diffusion characteristics of small-diameter particles in some indoor environments. However, the movement of particles in ventilated areas is influenced by many factors, such as airflow pattern, particle properties, geometry configurations, etc.
9、Murakami et al. (1 992) numerically investigated the diffusion characteristics of various sizes of particles with gravitational settling in clean rooms. Their study may be used as a guide for engineers to decide how the effects of gravitational sedimentation of parti- cles are considered when design
10、ing the indoor environment of cleanrooms. But few comparative studies of diffusion charac- teristics of aerosol particles with different sizes in different ventilated rooms have been found to date. As it is difficult to examine the effect of gravitational sedimentation in rooms by model experiments,
11、 the purpose of this paper is to compare the diffusion characteristics of various sizes of particles inside displacement and mixing ventilated rooms by a three-dimen- sional numerical method. MATHEMATICAL MODEL To calculate the three-dimensional and nonisothermal airflow inside ventilated rooms quic
12、kly and correctly, a well- validated simplified methodology combined with the N-point air supply opening model (Zhao et al. 2003) and a zero equa- tion turbulence model (Chen and Xu 1998) are applied. For indoor aerosol particle diffusion, it is not adequate to assume the particles as passive contam
13、inants when comparing the Bin Zhao is apostdoctoral fellow and Dongtao Huang is a professor in the Department of Engineering Mechanics and Zhao Zhang is a student and Xianting Li is an associate professor in the Department of Building Science, Tsinghua University, Beijing, China. 88 02004 ASHRAE. di
14、ffusion characteristics of various sizes of particles since the settling under the effect of gravitational sedimentation may play an important role on the particle diffusion. To take the gravitational sedimentation of particles into account, the drift flux model, which has been applied for indoor pa
15、rticle diffu- sion simulation successfully (Murakami et al. 1992; Holm- berg and Li 1998), is adopted. The model considers the gravitational sedimentation of particles and, thus, may be applied to investigate the difision characteristics of various sizes of particles. Governing Equations sion in vec
16、tor form are as follows: The governing equations for airflow and particle diffu- at +v. (pV v) = v. (peflVv)-VP+ f (2) m+V(pv) = V.(QV)+S, at 30 (3) ao + v. (p( v+ VS)C) = v. (QVC) + s, (4) at 3C where p, I: and P are the air density, velocity vector, and pres- sure, respectively. V, is the settling
17、 velocity of particles. C is the mass concentration of particle. The effective viscosity is the sum of molecular and turbulent viscosity. Turbulence is modeled by the zero equation turbulence model (Chen and Xu 1998). is the scalar quantity of air. For this study, it stands for enthalpy (or temperat
18、ure). The nondimensional numbers cjO and represent the turbulent diffusivity of Q, and C, respectively. Here the values are set as 1 .O. fis the body force due to air density (temperature) difference, which is modeled by using the Boussinesq approximation. Sc is the particle generation rate indoors.
19、 Key assumptions used for the simulation of particle diffu- sion in this study include: The effect of particles on turbulence is not considered, as it is believed that the low particle loadings and, compara- tively, small particle settling velocities have a negligible effect compared to the high inf
20、low turbulence levels (Elghobashi 1994). The particle size distribution will not be altered by coagu- lation due to low particle loadings. The body force due to particle/fluid density difference is neglected, as it is much smaller compared to the force caused by temperature difference for the partic
21、le sizes in the present study. Thus, the particle diffusion may be simulated based on the convergent velocity field. The settling velocity of a particle derived by equaling the fluid drag forceon the particle with the gravitational force can be expressed as (5) where C, is the drag coefficient, dp i
22、s the diameter of the parti- cle, pp and pa are the density of the particle and the ambient air, respectively, and g is the gravitational acceleration. The settling velocity always has the same direction as gravitation, namely, perpendicularly downward. The drag coefficient is either derived by the
23、Stokes equa- tion (Re 1, the settling velocity is derived by iterating the drag balance equation based on Equation 6. Boundary Conditions The supply inlet is described by the N-point supply open- ing model to consider the complicated geometry of actual diffusers (Zhao et al. 2003), and all variables
24、 are defined at the supply inlet. Outlet boundary conditions are set as the Neumann boundary condition; that is, mass flow boundaries are specified to ensure the mass flow rate out ofthe domain just as the mass flow rate into the flow domain. For the zero equa- tion turbulence model, wall functions
25、are not needed for the region near the walls, where the algebraic equations of turbu- lent viscosity may be applied directly (Chen and Xu 1998). The wall boundary condition of particle is simplified as zero normal gradients. That is, X/dn = O, where n is the direction normal to the wall. As the amou
26、nt of deposited parti- cles is usually much smaller than that of particles exhausted by ventilation (Nazaroff and Cass 1989), the deposition process cannot exert a strong influence on the flow field in an indoor environment. Numerical Methods The equations above are discretized into algebraic equa-
27、tions by finite volume method (FVM), and the coupling between velocity and pressure is solved by SIMPLE algorithm (Patankar 1980). In the course of discretization, the power law scheme and the second-order central difference are, respec- tively, implemented for the convection and diffusion terms. Th
28、e resulting set of discretized equations for each variable is solved by a line-by-line procedure, combining the tridiagonal matrix algorithm (TDMA) and the linear under-relaxation iteration. Nonuniform staggered grid system is employed in ASH RAE Transactions: Research 89 Figure I Items Human simula
29、tors X Quantities (W) 75* x 2 (a) Displacement Schematic of the two ventilation cases (I-cabinet, difuser, 7-exhaust; (6) 6-exhaust, 7-grille. Table 1. Total Heat Transfer for the Two Cases I cornouters 1 108 + 173 (close to window) 1 External heat transfer Gross 710*, 728f * One can find different
30、values in practice. These were chosen for validation pur- oses among the detailed experimental conditions. Displacement ventilation t Grille ventilation the present study, with denser grids clustering near the bounded walls so as to resolve the boundary layer properly. A well-validated CFD program d
31、eveloped by the authors, STACH-3 (Zhao et al. 2003), which fulfilled the simplified methodology mentioned above, is used as the simulation tool. CASES STUDIED To compare the diffusion characteristics of various sizes of particles under different ventilation conditions, a full-scale room with two dif
32、ferent ventilation methods (one displace- ment and one mixing) is selected. The room geometry is L(X) x H(r) x W(Z) = 5.16 m x 2.43 m x 3.65 m. The mixing case is ventilated by one grille diffuser on a side wall, and the grille size is 0.28 m (2) x 0.18 m (Y). The displacement diffuser size is 0.53
33、m (2) x 1.1 m (Y). Both cases have a 0.43 m X 0.43 m exhaust. Both the inlets and outlets are symmetrical about the middle line in the XY plane (Z = 1.825 m). Figure 1 gives the schematic of the two cases. The two cases also have the same air supply volume and simi- lar heat sources. Thus, the condi
34、tions are similar so that the results are comparable. Tables 1 and 2 present the values of the internal heat sources and air supply parameters, respectively. To compare the difision characteristics of various sizes of particles, five different sizes of particles-2.5, 5, 7.5, 10, and 20 pm-are consid
35、ered in the present study. The particles are assumed as spherical, and the density is 1050 kg/m3. Two X (b) Mixing 2-table, 3-compute 4-person, 5-lamp). (a) 6-displacement Table 2. Ventilation Displacement Mixing (Grille) Air Supply Parameters of the Two Cases 22.2 5.0* I 1.26 I 15.1 I 24.5 One can
36、find different values in practice. These were chosen for validation purposes among the detailed experimental conditions. persons are assumed as particle generation sources at the rate of 0.0916 pg/ (s.person) for each size of particle. RESULTS AND DISCUSSION Figures 2 and 3 show the velocity and tem
37、perature distri- bution of the two cases at the symmetrical plane (Z = 1.825 m). The displacement ventilation produces a ther- mal plume in the room, and the temperature is stratified. For the mixing ventilation, the jet attaches to the ceiling for a distance of about 4.0 m and then drops down and r
38、eturns by the exhaust. The air in the mixing ventilation room is better mixed than in the displacement one, and the temperature is more uniform. The detailed validation of the velocity and tempera- ture between numerical results and measured data has been performed by the authors (Zhao et al. 2003).
39、 The agreements are satisfied for engineering accuracy, The following will focus on the analysis of particle difision characteristics. Particle Diffusion in the Displacement Ventilated (DV) Room Figure 4 shows the five positions of monitoring points investigated in this study. Figures 5a to 5f indic
40、ate the simu- lated particle distribution in the room. There is only some slight difference between the results of particles smaller than 10 microns and that of passive contaminant. For the larger particles of 20 microns, due to relatively high settling velocity (about 1.4 cms), the concentration is
41、 much lower above the floor, and the settling on the floor increases much more. In 90 ASHRAE Transactions: Research 2.0 3.89 2.58 1.13 O.O(mL 1 .o 0.9 CI - - - . 0.0 2.0 4.0 (a) Displacement (b) Mixing Figure 2 The velocity distribution for the two cases (at symmetrical plane Z = 1.825 m). Tf%npclSO
42、15171 19.1 21 1232290 Tanpcil501691720Z24?43260 2 g, O wm wm (a) Displacement (b) Mixing Figure 3 The temperature distribution for the two cases (at symmetrical plane Z = 1.825 m). general, the particle concentration is higher at the low right comer and the zones above it. It may be because the pred
43、om- inant rightward near-floor airflow brings particles to the low right comer and then the strong upward near-wall flow carries part of them to the upper part (see Figure 2a). The most inter- esting thing is that for 1 O-micron particles, the concentration is much higher in this case, which is unli
44、ke that of the small particles as well as the large particles. The unidirectional upward flow can make smaller particles escape and middle- sized particles suspended, but it is not strong enough to hold particles as large as 20 microns in diameter, which leads to larger concentration of 1 O-micron p
45、articles indoors (see also Figure 6). Figure 6 shows the concentration of different particles at five points shown in Figure 4. Points 1 and Points 2 are in the region near both the particle source and the air supply open- ings. The concentration is lower near the ceiling and the floor and higher in
46、 the height about 0.8 m to 1.8 m where the manne- quin generates particles carried up by the upflow. Point 3 represents the central part of the room that is away from heat 2.41 1.825 1.31 * Oh= 5 4 *3 *2 1 t” I Z Figure 4 Positions of samplingpoints noted in this paper: and particle sources and is j
47、ust under the exhaust area. Points 4 and 5 are in the region far from the air supply opening and near the right wall. It shows the particle concentration is higher for the two points, which agrees with the analysis of Figure 5. Although the concentration value of larger particles differs greatly fro
48、m smaller ones, the distribution tendency along alti- tude is more or less similar. We can see that except for the near- floor area, the concentration summit always appears at the alti- tude of 1.2-1.4 m. Getting down to the flow field shown in Figure 2a, we can explore the reason: below this altitu
49、de there is air circulation, and the air can be entrained to the upper part only by upward flow near mannequins and the right wall. Once the particles fall into this circulation, they will either recircu- ASHRAE Transactions: Research 91 -2 E -1 ; -0 C:Qi: O O 1 2 3 4 5 X Im X Im (d) Non-passiveparticbe, Dp-7.5p (e) Nowpassive particbe, Dp=lO/rm Figure 5 Comparison of dispersion of diferent sizes of particles in the displacement ventilated room (ut symmetrical plane Z = 1.825 m). 3 2.5 2 - E, E, 1.5 a E 1 0.5 0 (a) Pomt 1 0) pont 2 (c) Point 3 (d) Point 4 Figure 6 Concentration
