1、4681 (RP-1168) Exhaust Contamination of Hidden vs. Visible Air Intakes Ronald L. Petersen, Ph.D. Member ASHRAE John J. Carter Member ASHRAE John W. LeCompte Associate Member ASHRAE ABSTRACT A wind tunnel dispersion modeling study was conducted to investigate exhaust contamination of hidden versus vi
2、sible air intakes. A “hidden intake is typically on a building side- wall or on the sidewall of a roof obstruction opposite the exhaust source. A “visible” intake is at roof level or on top of an obstruction, directly above the hidden intake. Overall, the study has shown what designers suspected: pl
3、acing air intakes on building sidewalls is bene$cial when the stacks are on the roo$ Signijcant concentration reductions were found when air intakes are placed right below the building roof edge on the building sidewall. The farther down the building sidewall the air intake isplaced, the larger the
4、reduction. Howevel; the larg- est relative reduction between a visible and hidden intake is achieved by just moving the intake a few feet from the edge of the building roof to apointjust around the corner on the build- ing sidewall. INTRODUCTION This paper documents ASHRAE Research Project 1168- TW
5、on exhaust contamination of hidden versus visible air intakes. Throughout this paper, a hidden intake is typically on a building sidewall or roof obstruction sidewall, while a visible intake is at roof level or on top of an obstruction, directly above the hidden intake. Designers commonly place air
6、intakes on the walls of the building, just below the roof in the belief that these hidden intakes will have less contamination from roof- mounted exhaust sources than if the intakes are on the roof itself. This paper provides documentation supporting this design practice and also provides methods to
7、 quantify the level of concentration reduction that is achieved when intakes are hidden. The specified objective of this research was to compare the effects on exhaust-to-intake concentration reduction (dilu- tion) for hidden versus visible air intake locations and to produce a set of design guideli
8、nes and a concentration (dilu- tion) calculation procedure suitable for inclusion in the ASHRAE Handbook-HVAC Applications. To meet the project objectives, three major tasks were conducted: (1) a literature review of experimental data concerning the concen- tration reductions achieved at hidden inta
9、kes relative to the visible intakes; (2) an experimental program using scale models of three representative buildings in a boundary layer wind tunnel to study the effects of exhaust configurations, meteorological conditions, building configurations, and hidden versus visible intake configurations on
10、 concentration, velocity, and turbulence; and (3) analysis of the data to identifj key variables and to develop a simple method to predict concentration reductions at hidden intakes. General design guidelines regarding hidden intakes were also developed as part of task three. BACKGROUND INFORMATION
11、The literature review revealed a wealth of research regarding concentration predictions and observations at visi- ble intakes. Several studies of concentrations in building wakes(Huber l978,1988a, 1988b, 1989)aswellasafewstud- ies of clusters of buildings (Wilson et al. 1998; Hosker 1985) were ident
12、ified. Six studies that specifically tested hidden air intake locations on buildings of simple geometry are discussed below (Halitsky 1963; Wilson 1976, 1977a, 1977b; Li and Meroney 1983; Petersen et al. 1997). Each of the six studies added specific detail for under- standing the variables relevant
13、to concentration predictions at Ronald L. Petersen is vice president and John J. Carter is a senior project engineer at Cermak Peterka Petersen, Inc., Fort Collins, Colo. John W. LeCompte is a graduate student at Colorado State University, Fort Collins, Colo. 130 02004 ASHRAE. air intakes. The effec
14、ts of exhaust momentum, stack height, wind direction, stack location, and architectural screens were examined. Most of the studies based their predictions on stretched-sting distance, S, velocity ratio, V, /U, and exit area, A, but did not differentiate between hidden and visible intakes. Therefore,
15、 no beneficial effect was predicted when the intake was moved from the top of the roof edge to the top of the building sidewall. Only Li and Meroney (1 983) suggested a different distance dilution parameter, B, for hidden air intakes versus visible intakes (according to ASHRAE 19991, Chapter 43, but
16、 not stated outright in the published article). There are even conflicting conclusions between two of the studies. Wilson (1976) concluded that using a hidden air intake has no benefit in reducing concentrations, while Li and Meroney (1 983) concluded that there is a significant benefit. Three of th
17、e studies (Wilson 1976, 1977b; Li and Meroney 1983) were performed with very low exhaust velocities in order to simulate capped stacks or leaks and were not intended for use in the design of exhausts with significant exit veloci- ties. Halitsky (1 963) utilized a uniform mean approach veloc- ity wit
18、h negligible turbulence, which would not accurately represent the atmospheric boundary layer. Despite the value of this research, the results from all of these studies were disqual- ified for inclusion in this analysis for either of the following reasons: the velocity ratio was not varied through a
19、range in order to find the critical maximum concentration at the roof edge and at the hidden receptor, or only cases with insignifi- cant plume rise were investigated. Petersen et al. (1 997) investigated the influence of archi- tectural screens on exhaust dilution at visible and hidden air intakes.
20、 One building was studied with three exhaust veloci- ties and two wind speeds. The visible and hidden intake concentration data were considered for inclusion in this study but were disqualified since the overall maximum concentra- tions on the roof and sidewall were not determined. The results from
21、Petersen et al. (1997) were used to provide a rough check on the general equation that was developed as part of this study. Overall, the literature review provided information such that representative building geometries could be identified for this evaluation and provided further justification for
22、this research. WIND TUNNEL DATABASE Wind Tunnel Simulation of Airflow and Dispersion An accurate simulation of the boundary-layer winds and stack gas flow is an essential prerequisite to any wind tunnel study of difision around buildings. The similarity require- ments can be obtained from dimensiona
23、l arguments derived from the equations governing fluid motion. A detailed discus- sion of these requirements is given in Snyder (198 1). The crite- ria and experimental methods that were used for conducting the wind tunnel simulations are discussed in detail in Petersen and LeCompte (2002). Since th
24、is study was designed to be generic in nature, rectangular buildings were placed in one of two uniform roughness configurations. The roughness config- urations were designed to simulate either a rural environment with a surface roughness length of 0.30 m or an urban envi- ronment with a surface roug
25、hness length of 0.80 m. General Description of Test Plan Concentration and velocity measurements were obtained on 1 : 1 O0 scale models for three different building geometries with various exhaust stack configurations and meteorological conditions, The building geometries were determined by reviewin
26、g typical laboratory and commercial building shapes (Petersen and LeCompte 2002). Table 1 summarizes the dimensions of the buildings evaluated for this study and the building used by Petersen et al. (1997). Figure 1 is an isometric drawing of Building 1, a large flat-roofed, low-rise structure repre
27、sentative of a shopping center or commercial plant. For selected tests, a long or short barrier at one of two locations downwind from the stack was installed, as shown in Figure 1. Most testing was conducted with a flat, unobstructed roof. Concentrations were measured at hidden and visible intakes f
28、or a range of exhaust stack heights, volume flow rates, and wind speeds at each of three wind directions: normal to the short side, normal to the long side, and diagonal. Stack loca- tions were designated S l through S7, and receptor locations were designated R1 through R17, as shown in Figure 1. Us
29、ing the criteria described in Petersen and LeCompte (2002) and source characteristics specified in Table 2, the model test Table 1. Physical Characteristics of Buildings Evaluated Building Height Length Width Number Total Area Number ft (m) ft (m) ft (m) of Floors ft2 (m2) LH WH Comment 1 30 (9.1) 1
30、50 (45.7) 90 (27.4) 2 27,000 5.0 3.0 (2509.7) 2 75 (22.9) 300 (91.5) 150 (45.7) 5 225,000 (20,913.9) 4.0 2.0 3 150 (45.7) 375 (114.3) 75 (22.9) 10 281,250 (26,142.4) 2.5 0.5 4 50 (15.2) 100 (30.5) 50 (15.2) 3 15,000 2.0 1 .O Test Building For ASHRAE (1394.3) 805-TRI ASH RAE Transactions: Research 13
31、1 +- Source Exit Diameter, d Volume Flow Rate, Q Exit Velocity, Ve Description in. (m) cfm (m3/s) fpm (m/s) Low Flow 6.8 (0.17) 500 (0.24) 2000 (10.2) Medium Flow 21.4 (0.54) 5000 (2.36) 2000 (10.2) High Flow 67.7 (1.72) 50000 (23.6) 2000 (10.2) Building Type 1 Figure 1 Building dimensions with stac
32、k and receptor locations for Building 1-large, flat-roofed, low-rise shopping center or commercial plant. Table 2. Full-Scale Exhaust Parameters conditions were computed for three generic stack configura- tions designated low flow (500 cfm, 0.24 m3/s), medium flow (5000 cfm, 2.36 m3/s), and high flo
33、w (50,000 cfm, 23.6 m3/s). All stacks were tested in the rural environment for each build- ing. Only the medium flow stack was tested in the urban envi- ronment for each building. A full range of anemometer wind speeds (1 to 20 m/s, 2.24 to 44.7 mph) was tested in order to determine the highest conc
34、entrations achievable at the roof and sidewall receptors. Detailed information on the wind tunnel, instrumentation, and boundary layer documentation can be found in Petersen and LeCompte (2002). Mean velocity and longitudinal, lateral, and vertical turbulence intensity were measured in the rural rou
35、ghness configuration at heights corresponding to 5,1 O, 15,20,30, and 40 ft (1.52,3.05,4.57,6.10,9.15, and 12.2 m) aboveeachstack location for each wind direction. Velocity measurements using the urban approach roughness were only obtained for Building 2. The purpose of these measurements was to det
36、ermine how the building altered the approach flow and to develop inputs for the dispersion models. The results of these measurements are described in Petersen and LeCompte (2002). ANALYTICAL METHODS General Existing analytical methods were used as the starting point for developing and evaluating the
37、 method for estimating concentrations at visible and hidden air intakes. Chapter43 of the 1999 ASHRAE Handbook-HVAC Applications (ASHRAE 1999) describes a commonly used method for predicting the concentration at a visible intake. Another method is a basic Gaussian analysis (Turner 1994). Both of the
38、se methods, as well as an enhanced Gaussian analysis, are described briefly in this section. 132 ASHRAE Transactions: Research Concentration Prediction-ASHRAE Method In Chapter 43 of the 1999 ASHRAE Handbook-HVAC Applications (ASHRAE 1999), the critical dilution (concen- tration) is first calculated
39、 for a flush vent (a stack with h, = O), then adjusted for the actual stack height under consideration. This method uses the following parameters: emission mass flow rate; stack height, exit area, and exit velocity; stretched string distance; and simplified terms to describe the plume spread. Concen
40、tration Prediction- Standard Gaussian Method The Gaussian plume model can be simplified to the following equation for a centerline receptor at the surface of a reflecting roof or at ground level: where Cis the pollutant concentration (pg/m3); m is the pollut- ant mass emission rate (g/s); U, is the
41、mean wind speed at stack top (ds); oy and oz are the standard deviations of plume concentration distribution in the crosswind and vertical direc- tions at downwind distance x, respectively; h, is the stack height above the roof, receptor, or ground level as appropriate; and Ah, is the net effective
42、rise of the plume centerline above the stack top. For the standard method, the approach wind conditions at stack height are used rather than the local conditions at the stack top. This allows for straightforward calculations of U, oy and o,. The mean wind speed is characterized by a power law relati
43、onship: where U, is the mean wind speed at the anemometer height, z,; z, is the free stream height; nanem is the power law exponent at the anemometer; zs is the stack height above ground level; and nsite is the power law exponent at the build- ing site. The power law exponents, nsire and nane, can b
44、e calculated based on surface roughness classification or can be determined empirically by a curve fit to a model-scale approach velocity profile. The lateral and vertical plume dispersion coefficients, oy and oz, can be calculated using the empirical equations in the EPA AERMOD model (Cimorelli et
45、al. 1998) and can be represented as a function of two factors, and where ouo and o, are initial lateral and vertical plume disper- sion, and o and ozt are lateral and vertical plume dispersion due to ambient turbulence. According to Turner (1994), ? (4) - oyO - oz0 = 0.35d, where d is the stack diam
46、eter. According to Cimorelli et al. (1 998), the crosswind and vertical ambient turbulence disper- sion coefficients, oVr and ozt, are calculated as linear functions of distance from the exhaust source, based on the longitudinal turbulence intensity at stack height in the approach wind. - oyt - Iys,
47、apx = AiIxs,apX and (5) ozt = Izs,apx = A,Ixs,apx where and Izs,ap are the lateral and vertical turbulence intensities in the approach flow at stack height, Ix,ap is the longitudinal turbulence intensity in the approach flow at stack height, A, and A, are empirical constants, and x is the down- wind
48、 distance to the receptor location. For flat homogenous terrain, Snyder (1 98 1) suggests the following values: A, = 0.75 and A, = 0.50. (6) Ix,ap is determined by either an empirical relationship to height, heat transfer, mixing depth, temperature, gravity, and fluid properties (Cimorelli et al. 19
49、98) or by a curve fit to an observed approach longitudinal turbulence intensity profile. For this study, Ixap for the rural and urban approaches measured in the wind tunnel were found to agree with the following equations: 0.3631 0.3835 Ix,ap = 7.6h92(1) re - 7.5252(z) re (nirai) (7) 0.3849 (urban) (8) Ix,ap = 7.6798(L) 0.3591 re The second termwas addedafter the first term alone failed to produce an adequate fit. The best-fit constants were found by minimizing the normalized root-mean-square error (RMSE) over the range of z values used in the s
