PLEASE SCROLL DOWN FOR ARTICLE
This article was downloaded by: [National Taiwan University] On: 14 April 2009
Access details: Access Details: [subscription number 908033515] Publisher Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Journal of Occupational and Environmental Hygiene
Publication details, including instructions for authors and subscription information: http://oeh.informaworld.com/soeh/title~content=t713657996
Correlation Between Airflow Patterns and Performance of a Laboratory Fume
Hood
Li-Ching Tseng a; Rong Fung Huang b; Chih-Chieh Chen a; C. -P. Chang c
a Institute of Occupational Medicine and Industrial Hygiene, National Taiwan University, Taipei, Taiwan b
Department of Mechanical Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan c Institute of Occupational Safety and Health, Council of Labor Affairs, Sijhih City, Taiwan
First Published on: 01 December 2006
To cite this Article Tseng, Li-Ching, Huang, Rong Fung, Chen, Chih-Chieh and Chang, C. -P.(2006)'Correlation Between Airflow Patterns and Performance of a Laboratory Fume Hood',Journal of Occupational and Environmental Hygiene,3:12,694 — 706 To link to this Article: DOI: 10.1080/15459620601015695
URL: http://dx.doi.org/10.1080/15459620601015695
The American Conference of Governmental Industrial Hygienists (http://www.acgih.org/) and the American Industrial Hygiene Association (http://www.aiha.org/) have licensed the Taylor & Francis Group to publish this article and other materials. To join the American Conference of Governmental Industrial Hygienists visit http://www.acgih.org/Members/. To join the American Industrial Hygiene Association visit http://www.aiha.org/Content/BecomeMember/becomemember-splash.htm.
Full terms and conditions of use: http://oeh.informaworld.com/terms-and-conditions-of-access.pdf
This article maybe used for research, teaching and private study purposes. Any substantial or systematic reproduction, re-distribution, re-selling, loan or sub-licensing, systematic supply or distribution in any form to anyone is expressly forbidden.
The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.
Journal of Occupational and Environmental Hygiene, 3: 694–706 ISSN: 1545-9624 print / 1545-9632 online
Copyrightc 2006 JOEH, LLC DOI: 10.1080/15459620601015695
Correlation Between Airflow Patterns and Performance
of a Laboratory Fume Hood
Li-Ching Tseng,
1Rong Fung Huang,
2Chih-Chieh Chen,
1and C.-P. Chang
3 1Institute of Occupational Medicine and Industrial Hygiene, National Taiwan University, Taipei, Taiwan 2Department of Mechanical Engineering, National Taiwan University of Science and Technology,Taipei, Taiwan
3Institute of Occupational Safety and Health, Council of Labor Affairs, Sijhih City, Taiwan
To understand the physical mechanisms of the contaminant dispersion and containment leakage during the ventilation pro-cess through a laboratory fume hood, the complicated three-dimensional flow patterns and the real-time tracer gas (SF6)
leakage were studied via the laser-assisted flow visualization method and the standard/special gas sampling technique, re-spectively. Through flow visualization, the large-scale vortex structures and boundary layer separations were found around the side poles and doorsill of the hood. In the near-wake re-gion of the manikin, large recirculation zones and wavy flow structures were also identified. When tracer gas concentration measurements were conducted point-by-point across the sash opening, the areas near the doorsill, the lower parts of the side poles, and the sides of the manikin showed significant contaminant leaks. These areas with high contaminant leaks exactly corresponded to where the flow recirculated or sepa-rated. However, when the ANSI/ASHRAE 110–1995 protocol was used to measure the concentration of SF6at the breathing
zone of the manikin, no appreciable leakage was detected. It is suggested that a method based on the aerodynamic features and multipoint leakage detections would reflect a more realistic evaluation of overall performance of laboratory fume hood than a single-point sampling method at the manikin’s breathing zone.
Keywords laboratory fume hood, performance, smoke visualiza-tion, tracer gas, turbulence
Address correspondence to: Rong Fung Huang, Department of Mechanical Engineering, National Taiwan University of Science and Technology, 43 Keelung Road, Section 4, Taipei, Taiwan, R.O.C.; e-mail: rfhuang@mail.ntust.edu.tw.
INTRODUCTION
A
laboratory fume hood is a box-like structure with an ad-justable front opening. It is designed to capture, contain, and remove harmful fumes generated inside the enclosure. The performance and characteristics of laboratory fume hoods have been widely studied during the past few decades. Previous studies(1−7)have suggested that operating the hood at a highface velocity does not ensure a high hood containment effi-ciency because no correlation between the face velocity and the hood performance was observed. The primary physical mechanism leading to contaminant leakage, therefore, may not be the global parameter of the suction flow rate or the flow distribution. Instead, the detailed flow structure may play a significant role and require systematic investigation.
Earlier studies(8−14) on the local, detailed flow structure concentrated mainly on the use of numerical analysis to pre-dict flow patterns through laboratory fume hoods. Turbulent flow has been recognized as a significant factor contributing to contaminant leakage. Therefore, a better understanding of the underlying dynamics of the basic structures in turbulent flow to describe the transport mechanism of contaminant is required, but detailed experimental data on this subject and the extent of understanding the physical mechanisms governing contaminant leakage during the ventilation process are limited and warrant further investigations.
The ANSI/ASHRAE 110-1995 Method of Testing Perfor-mance of Laboratory Fume Hoods provides a test mthod that is both quantitative and qualitative for evaluating the containment of laboratory fume hoods.(15) This widely accepted protocol consists of a three-part evaluation process that included a face velocity test, flow visualization, and a tracer gas test. This test method was developed as an attempt to measure the breath-ing zone exposure of a fume hood operator. Accordbreath-ingly, the concentration at the breathing zone is representative of the concentration inhaled by a person in front of a laboratory fume hood. However, the test protocol does not account for important aerodynamic features of fume hood operating exposure.
In this study, the time-averaged face velocity, the velocity fluctuation, and the real-time tracer gas concentration along the face of the hood were measured. The laser light sheet-assisted flow visualization method was first incorporated to provide details of the flow patterns inside and around the sash opening of the fume hood. The data was used for guiding the subsequent quantitative containment experiments. The effects of the fume hood geometry and the presence of the manikin on
NOMENCLATURE
CSF6 concentration of SF6, in ppm
H operating sash height of fume hood
H0 designed maximum sash height of fume hood
Q volumetric suction flow rate of fume hood V0 face velocity of fume hood (= Q/ area of sash
opening)
W width of hood opening in sash plane X horizontal coordinate originated at left side
boundary of hood opening in sash plane Y vertical coordinate originated at left side
boundary of hood opening in sash plane Z normal coordinate originated at left side boundary of hood opening in sash plane
the contaminant leakages evaluated by tracer gas concentra-tion measurement methods were investigated. By comparing the results of the flow structures and the tracer-gas tests, the physical mechanism leading to the contaminant leakage is discussed and a strategy for testing the laboratory fume hood with considerations of the local leakage instead of the single-point test around the worker’s breathing zone is suggested. EXPERIMENTAL METHODS
T
he experimental study included two parts: the laboratory test and the on-site measurements.Laboratory Test Apparatus
The experimental setup included a laboratory fume hood model, an extract fan, and instruments, as shown in Figure 1. The prototype fume hood used for the test was made of transparent acrylic plates so laser beams could pass through. The height and width of the working aperture of the test fume hood were 850 mm and 1200 mm, respectively. The suction flow rate was measured by a homemade Venturi flow meter and a calibrated pressure gauge. The accuracy of the suction flow rate measurement was less than 1% of the reading. Experiments were carried out in a well-controlled test room that was 19 m3×
17 m3× 5 m3. During the experiment, the interference induced
FIGURE 1. Experimental setup
FIGURE 2. Experimental setup; (a) apparatus for flow visualization and tracer gas test, (b) test grid arrangement by the external sources such as the air supply diffusers, doors,
and traffic in the room was restricted. Flow Visualization
An experiment of flow visualization was conducted at a sash height of 600 mm. The face velocity V0was set at 0.5 m/sec.
Paraffin oil mist was continuously delivered to the smoke ejec-tor of a small-diameter, stainless steel tube from a home-made smoke generator. The smoke ejector (or smoke tube) was placed in the hood at the proper position to release the smoke and scatter the laser light, as shown in Figure 2a. The Stokes number of the smoke particles was in the order of magnitude of 10−6and could, therefore, properly follow the flow fluctuations up to at least 10 kHz.
The streams of the smoke were discharged by two dif-ferent means. (1) For simulating the contaminant transport
process, the smoke was released through a smoke ejector that was made following the ANSI/ASHRAE 110-1995 method for tracer gas measurement, as shown in Figure 2a. Both the positioning of the smoke ejector and the release rate of the smoke flow all followed the ANSI/ASHRAE 110-1995 pro-tocol for tracer gas measurement. (2) For better visualiza-tion of local flow structure in some target regions, the smoke was released at a velocity lower than 0.05 m/sec through a small-diameter (4 mm) stainless steel tube positioned around the target regions. The laser beam emitted from a Nd:YAG laser (Dpss Green Laser AMGC-100, Onset Electro-Optics Corp., Taipei, Taiwan) was transmitted through an optical fiber and connected to a 20◦ laser light sheet expander. The par-ticle images were recorded by a CCD camera (DCR-TRV 900, Sony Corp., Japan) that was equipped with an asyn-chronous variable electronic shutter so the exposure time could
be adjusted from 1/12000 sec to 1/30 sec at a framing rate of 30 ft/sec.
Experiments were performed for both “occupied hood” (with a manikin standing in front of the hood) and “unoccupied hood” (without the manikin) to examine the effect induced by the presence of the manikin.
Tracer Gas Test
The experiment was conducted at sash heights of 600 mm and 850 mm. The face velocity V0was set at 0.5 m/sec. The
tracer gas ejector position, flow rate, and manikin placement were the same as those used in the flow visualization, as shown in Figure 2a. The tracer gas ejector was fabricated following the ANSI/ASHRAE 110-1995 test standard. The mean and peak values of SF6concentration were measured and recorded with
a high precision infrared detector (MIRAN SapphIRe Ambi-ent Air Analyzer, Thermo Electron Corp., Franklin, Mass.). The detection range from 0 to 4 ppm was calibrated by the manufacturer and checked in the laboratory before use. The results of the laboratory check showed good agreement with the factory calibration in this range. The detection ranges from 4 to 100 ppm were calibrated in the laboratory.
Separated calibration curves and regression formulas were obtained for the ranges 4–37 ppm and 37–100 ppm. The max-imum deviations of the calibration curves from the supplied standard concentration values were 10% of the reading for the range 0–1 ppm, 20% of the reading for the range 1–4 ppm, and 12% of the reading for the range 4–100 ppm. The output data could be set on the control panel to ppm with a resolution of 0.01 ppm or to ppb with a resolution of 1 ppb. In this study, the resolution was set to ppm. The sampling flow rate was 14 L/min, the internal sample rate of the detector was 20 Hz, and the average value over a 10-sec period was recorded as one reading.
The experimental procedures were conducted in accordance with Test Mode A and Test Mode B of ANSI/ASHRAE 110-1995.
Test Mode A. The tracer gas was 10% by volume fraction
of sulfur hexafluoride (SF6) in nitrogen. Tracer gas samples
were taken point-by-point over the 12 locations in a rectangular area in the sash opening plane through a 13 mm i.d. stainless steel tube, as shown in Figure 2b. Both the occupied hood and the unoccupied hood were tested. Sampling was taken for 480 sec, and the data for the initial period of 180 sec was discarded.
Test Mode B. Measurements were conducted following
the ANSI/ASHRAE 110–1995 test protocol by using 100% SF6 as the tracer gas. The manikin with the same shape and
size as described in the flow visualization method was placed in position. Tracer gas samples were taken through a 13 mm i.d. stainless steel tube. The detector probe was arranged to penetrate through the head of the manikin so that the suction tip was fixed below the nose. The center of the probe was 660 mm above the work surface and 75 mm in front of the sash. The detector readings were recorded for 300 sec while the tracer gas was released into the hood.
On-Site Measurement
The relationship between face velocity and flow patterns to the tracer gas leakage was investigated on nine laboratory fume hoods randomly selected from nine universities in Taiwan. The mean face velocity measurement, flow visualization, and tracer gas test following the same procedure of the laboratory test were carried out.
RESULTS AND DISCUSSION Effect of Hood Geometry
Figure 3 shows the side-viewed smoke flow pattern across the symmetry plane of the hood when the sash height H was set at 600 mm. A large, counterclockwise-rotating vortex was ob-served behind the sash when smoke was released from a smoke tube placed behind the sash. This large-scale flow structure was induced by the separation of boundary layer occurring at the bottom edge of the sash when the airflow was drawn through the sash opening into the cabinet. Unsteady, complex, three-dimensional vortices were formed there. If the smoke was released from the smoke ejector placed on the work surface (which meant smoke was not released from the smoke tube behind the sash), the large recirculation bubble would become invisible. Instead, smoke released from the smoke ejector was drawn to the back of the cabinet and expelled through the bottom slot of the baffle. The flow pattern of Figure 3 implies that although a large recirculation bubble may have existed in the back of the sash, the contaminants released from the lower level of the cabinet may not have had the opportunity to disperse drastically up to the recirculation bubble behind the sash.
Figures 4a and 4b show the flow patterns around the right side wall and side pole. The laser light sheets were arranged across the horizontal planes on the middle and lower levels. Smoke was released from a smoke tube placed behind the doorsill near the right lower corner of the cabinet. The smoke was not drawn directly to the back of the cabinet. Instead, it was carried by an internal lateral stream to the area near the side wall, where some characteristic flow structures were observed. The flow in that region was very turbulent; the recirculation motion of the smoked flow was apparent.
Therefore, three-dimensional flow structure must exist in the cabinet, otherwise the smoke would not appear there. When the flow passed over the front edge of the side pole, the bound-ary layer separated. The separated boundbound-ary layer evolved from the edge of the side pole and entrained smoke from the region near the side wall so that the region near the side pole, which was outside the sash opening, was contaminated by the smoke. If contaminants in the cabinet recirculated to this area, dispersion through the shear layer out to the environment via the turbulent diffusion mechanism would be possible.(16)
Figure 4c shows the flow pattern around the doorsill. The laser light sheet was arranged vertically. When the environmen-tal air was drawn into the cabinet, the boundary layer separated from the front edge of the doorsill and formed a recirculation zone behind the doorsill. A series of coherent flow structures
FIGURE 3. Side view of smoke patterns in interior of hood cabinet, H= 600 mm, V0= 0.5 m/sec appeared in the shear layer. The region above the doorsill
out of the sash opening was contaminated by the smoke. The flows in the recirculation zone were highly turbulent. It is very probable that the contaminants were carried to this region and diffused through the turbulent shear layer to the atmosphere.
Table I shows the results of the tracer gas experiments using Test Mode A. When sash height H is set to either 600 mm or 850 mm, the levels of SF6 leakages measured at the upper
grids, i.e., the grids U1, U2, U3, and U4 shown in Figure 2b, were generally low for the unoccupied hood. On the middle level, substantially high SF6 concentrations were detected at
grids M1 and M4, which were located near the side poles. SF6 concentrations detected at M2 and M3 were still low.
On the lower level above the doorsill, all the detected SF6
concentrations were high. Particularly, at locations near the side poles (L1 and L4), containment leaks were much more severe than those detected at other positions. Obviously, under the condition of static test, the regions around the bottom edge of the sash and the middle region of the sash opening were not the most probable areas for containment spillage; the most highly probable areas existed in the regions around the side poles and the doorsill.
FIGURE 4. Smoke patterns near (a) middle level of side wall, (b) lower level around side right side pole and side wall, and (c) doorsill. H= 600 mm, V0= 0.5 m/sec.
The results of the flow visualization experiment establish that the recirculation zones formed behind the sash and around the side poles and the doorsill were the most probable sources for containment leakage, as stated earlier. The aforementioned results of SF6 concentration tests reveal that severe
contam-inant leakage occured around the areas near the side poles and the doorsill. However, the SF6 concentrations detected
directly beneath the bottom edge of the sash were relatively low. The three-dimensional flow motion in the cabinet, in
conjunction with the separated boundary layers induced by the inappropriate design of hood geometry, induced flow recircu-lation. Unsteady turbulent dispersion was, therefore, one of the important causes for hood leakage. The separated shear layers around the side poles and the doorsill shown in Figure 4 (which consist of unsteady coherent flow structures and recirculated containments) were located outside the hood cabinet, whereas the large recirculating vortex behind the sash, as shown in Figure 2, was totally enclosed in the cabinet.
TABLE I. SF6Concentrations Measured by Test Mode A (Point-by-Point Method)
H = 600 mm H = 850 mm
CSF6 CSF6 CSF6 CSF6
Unoccupied Hood Occupied Hood Unoccupied Hood Occupied Hood
Grid Position Mean (peak), ppm Mean (peak), ppm Mean (peak), ppm Mean (peak), ppm
U1 0.04 (0.05) 0.26 (0.66) 0.08 (0.14) 0.14 (0.54) U2 0.02 (0.02) 0.02 (0.02) 0.03 (0.05) 0.04 (0.08) U3 0.02 (0.02) 0.04 (0.08) 0.03 (0.07) 0.03 (0.08) U4 0.03 (0.06) 1.63 (2.43) 0.08 (0.28) 2.85 (6.84) M1 1.30 (4.58) 16.70 (25.78) 3.78 (5.79) 27.68 (39.42) M2 0.01 (0.01) 0.01 (0.01) 0.02 (0.03) 0.05 (0.07) M3 0.02 (0.02) 0.01 (0.02) 0.03 (0.05) 0.07 (0.15) M4 1.14 (3.24) 22.42 (31.76) 3.05 (8.59) 38.73 (50.41) L1 1.14 (4.78) 24.65 (39.29) 2.30 (5.02) 37.96 (51.20) L2 0.26 (0.68) 5.76 (38.81) 1.46 (5.27) 12.14 (57.97) L3 0.39 (3.90) 22.14 (57.97) 2.07 (7.36) 37.65 (76.52) L4 1.88 (6.33) 77.65 (96.52) 2.49 (4.50) 66.48 (86.29)
Note: Averages were taken over the 3rd to 8th min.
Therefore, the possibility of dispersion of the contaminants out into the atmosphere from the areas around the doorsill and the side poles would inevitably be higher than dispersion around the bottom edge of the sash. The airfoil structure or streamlined doorsill design commonly seen in commercial hoods may have positive effects on alleviating the formation of the doorsill boundary layer separation and, as a result, could reduce some containment spillage from that area if the arrange-ment is aerodynamically appropriate.
Effect of Manikin Presence
When smoke was released from the ejector, as shown in Figure 5a, none was found in the breathing zone of the manikin. As smoke was introduced to the area around the manikin’s head, as shown in Figure 5b, the suction force developed by the fan drew the smoke directly into the hood chamber. No recir-culation was found. The downwash current flushed downward from the head of the manikin was the primary flow physics responsible for the disappearance of the wake recirculation expected to appear there. The tracer gas measurements on the upper level grids for the occupied hood in Table I show that the values of SF6concentration around the breathing zone of
the manikin (grids U2 and U3) were quite low; concentrations were similar to those of the unoccupied hood. This result correlates well with the observation of flow patterns shown in Figure 5.
To examine flow around the chest of the manikin, smoke was released from the smoke ejector in the hood at the posi-tions indicated in Figure 6a. No smoke was found in front of the manikin’s chest. However, as the smoke was introduced into the area around the manikin’s chest, as indicated in Fig-ure 6b, the flow in front of the manikin’s chest exhibited a large, recirculating flow zone. This large recirculation
bub-ble was actually a “bluff-body wake” that was induced when air in the environment was drawn into the hood and across the manikin (the manikin is a bluff body from the viewpoint of aerodynamics). The recirculation bubble extended into the hood and may have transported the contaminant to the area near the manikin’s chest. The characteristic flow structure in-duced by the flow across a circular cylinder has been studied experimentally.(17,18)
These studies reveal that the presence of a bluff body (e.g., the manikin) can create a wake zone with recirculation and vortex shedding. The wake zone could draw contaminants from the regions in the hood where they might build up. This phenomenon is inevitable in fume hoods of conventional de-sign because the flow drawn into the sash opening always has a horizontal component that would make the Reynolds number based on the shoulder width of the manikin generally larger than 10000. In Table I, the results of the tracer gas measurements for the occupied hood on the middle level show that the values of SF6 concentration around the chest sides
(grids M2 and M3) were not significantly larger than their counterparts for the unoccupied hood.
The existence of chest recirculation seems not to have en-trained the SF6 out of the hood, which is in contrast to the
expectation obtained from the flow visualization result of Fig-ure 6b. This may be because the SF6released from the ejector
was drawn through the slots of the baffle and did not have op-portunity to disperse up to the recirculation zone, as illustrated in Figure 6a. However, this may not apply to all hood designs. The chest recirculation is still a dangerous flow structure that could cause contaminant leakage.
When the manikin was placed in front of the hood, highly unstable recirculation zones and wavy structures were observed in the horizontal and vertical planes around the side
FIGURE 5. Smoke patterns around breathing-zone of manikin. Smoke released from (a) smoke ejector placed in cabinet, (b) head side of manikin. H= 600 mm, V0= 0.5 m/sec.
corner and doorsill area, as shown in Figures 7a and 7b, respectively. These flow patterns were similar to those ob-served in the case of the unoccupied hood shown in Figures 4b and 4c. By comparing Figures 7a and 7b with Figures 4b and 4c, the recirculation zones induced by the flow separation when the manikin was present are apparently larger than those of the unoccupied hood. The presence of the manikin seems to have had the effect of enhancing the intensity and size of the separation and recirculation zones around the middle and the lower levels of the hood.
The tracer gas test results of Table I support the evidence of the visual findings. In grids M1 and M4 (near the side poles) on the middle level, the occupied hood exhibited much larger containment spillage than the unoccupied one. For instance, for H = 600 mm, CSF6 = 16.70 ppm was detected at M1
and 22.42 ppm at M4, whereas for the unoccupied hood, SF6
concentrations were only 1.30 ppm at M1 and 1.14 ppm at M4. On the lower level, the containment spillage was sub-stantially much more severe when the manikin was present, e.g., at the right corner (L4), the sampled SF6 concentration
FIGURE 6. Smoke patterns in front of chest of manikin. Smoke released from (a) smoke ejector placed in cabinet, (b) chest side of manikin. H= 600 mm, V0= 0.5 m/sec.
at the presence of the manikin attained 77.65 ppm, which was drastically larger than 1.88 ppm of the unoccupied hood. The interaction between the recirculation bubble in the manikin wake and the separated shear layer above the doorsill also caused large contaminant leakages. For example, at L2 and L3 (Table I) the detected SF6concentrations in the occupied
hood were substantially higher than those in the unoccupied hood. Figure 8 shows the time histograms of the recorded SF6
concentrations at L1–L4 for both the unoccupied and occupied hoods. The presence of the manikin caused drastic increase of
the time-averaged leakage levels when compared with levels in the unoccupied hood. Also, the fluctuations of the measured SF6concentrations of the occupied hood were critically larger
than those of the unoccupied one, which shows the unsteady characteristics of the containment dispersion induced by the increased turbulence levels.
Discussion of the ANSI/ASHRAE 110-1995 Method As shown in Table II, the measured time-averaged mean concentrations of SF6at the breathing zone of the manikin
FIGURE 7. Smoke patterns near (a) lower level around side right pole and side wall, (b) doorsill. Manikin placed in front of hood. H= 600 mm, V0= 0.5 m/sec
using the ANSI/ASHRAE 110-1995 method were less than 0.01 ppm for H = 600 mm and lower than 0.02 ppm for H = 850 mm. These levels are almost undetectable and far below the containment leakage threshold (0.05 ppm) recommended by ANSI Z 9.5-2003(19) for the class of “as manufactured
(AM).” The hood employed for the current investigation was a basic-type fume hood, which means no bypass design inside the hood above the sash, no auxiliary air near the face of the hood above the worker, no doorsill airfoil installation, no doorsill compensation air, no variable air volume (VAV)
operation, no adaptive back baffle for sash vortex manipulation (e.g., the Bi-stable Vortex Fume Hood), and no vortex-isolation technique (e.g., the Berkeley Hood), etc. The hood passed the ANSI/ASHRAE 110-1995 test, although it presented severe containment leakages from the regions around the side poles and doorsill by using the point-by-point diagnostic method described in previous sections.
This result is not unexpected according to the arguments and discussions made previously. In the ANSI/ASHRAE 110-1995 test, the tip of the sampling probe was positioned under
FIGURE 8. Time histograms of SF6leakages of unoccupied and occupied hood at lower lever grids. (a) L1, (b) L2, (c) L3, (d) L4. H = 600 mm, V0= 0.5 m/sec.
the nose of the manikin and was at a distance 75 mm away from the sash plane; whereas the data listed in Table I were obtained by positioning the tip of the sampling probe on the sash plane, using Test Mode A. Although the concentration of SF6released
from the ejector by following the ANSI/ASHRAE 110-1995 protocol was 100%, the containment leakage was still difficult to detect around the breathing zone of the manikin. Never-theless, as discussed previously, severe contaminant leakages occurred in the regions near the side poles and the doorsill regardless of whether the manikin was present. The repre-sentative samples measured in the breathing zone by using the ANSI/ASHRAE 110-1995 method seemed not to reflect accurately the overall behavior of the contaminant leakage through the hood.
The sampling strategy of the method focuses on detecting the concentration of SF6in the breathing zone of the manikin
by releasing the tracer gas ejector from a height of about 350 mm from the work surface. It does not ensure the detection
of the most probable areas of containment leakage around a fume cupboard. Based on results of the present investiga-tion, when satisfactory containment results are obtained by following the ANSI/ASHRAE 110-1995 protocol, it should be conservatively claimed that the tested hood has “good” performance because the point-by-point measurements across the whole sash opening may reveal results in contrast to the conclusion.
On-Site Measurement Results
Table III shows the in situ test results of nine fume hoods randomly selected from nine universities in Taiwan. Fume hood A has a face velocity (V0) lower than 0.30 m/sec. Fume
hoods B–G have a face velocity ranging from 0.30–0.50 m/sec. Fume hoods H and I have a face velocity greater than 0.5 m/sec. The results of tracer gas tests using Test Mode A and B were compared. Test Mode A (point-by-point method) results show
TABLE II. SF6 Concentrations Measured by Test Mode B, ANSI/ASHRAE 110–1995 Method
H= 600 mm H= 850 mm
CSF6 CSF6
Test Time Interval Mean (peak), Mean (peak),
(min) ppm ppm
0–5 <0.01 (0.01) 0.01 (0.02)
3–8 <0.01 (0.01) 0.02 (0.02)
extremely high containment leakage. However, when Test Mode B (ANSI/ASHRAE 110-1995 method) was employed, the measured SF6 concentrations were generally much lower
than those obtained by Test Mode A. Two of the nine fume hoods (C and I) passed the containment criteria for “as installed (AI)” and “as used (AU)” of 0.1 ppm recommended by ANSI Z 9.5-2003(19)for hoods set at fully open sash height. Test re-sults also confirm that lower face velocity may not necessarily provide poor containment(1−7)and may be capable of passing the criteria. Local flow structures are more important than the global parameter of the face velocity when considering the overall containment leakage problem.
TABLE III. On-Site Measurement Results
CSF6Test Mode BB (ppm) CSF6
Fume V0 Test Mode AA H/H0= H/H0=
Hood (m/sec) (ppm) 50% 100% A 0.29 >100 48.49 62.30 Bottom-middle edge B 0.32 >100 0.06 0.39 Bottom-middle edge C 0.33 31.14 0.02 0.03 Bottom-middle edge D 0.44 >100 2.60 1.94 Bottom-middle edge E 0.44 7.36 0.46 8.35 Bottom-right corner F 0.45 22.20 0.07 2.55 Bottom-middle edge G 0.46 20.89 0.41 0.79 Bottom-middle edge H 0.56 1.74 0.01 0.67 Bottom-left corner I 0.76 > 100 0.20 0.06 Bottom-middle edge
AMeasurements taken at most severe area; point-by-point method used. BANSI/ASHRAE 110-1995 method used.
CONCLUSIONS AND RECOMMENDATIONS
T
his study clearly demonstrates the influences of geometric features and the presence of an operator on laboratory fume hood performance by correlating the results obtained from the aerodynamic flow patterns and tracer gas concen-tration measurements. Complicated turbulent flow patterns, for example, were induced around the bottom edge of the sash, side poles, and doorsill when environmental air drawn by the suction fan of the hood passed over the hood structure and hood operator. The induced flow structures involved the separation of turbulent boundary layers, unsteady motions of shear layers, vortical flows, wake instabilities, and interactions among these characteristic flow structures. These complex flow behaviors can easily contribute to the spread of contaminant leakage and increase the probability of laboratory workers’ exposure. The most probable sources for containment leakage of a conventional fume hood are around the doorsill and side poles. Results suggest that a comprehensive strategy based on the entire hood face exposure assessment would reflect a more realistic evaluation of the overall effectiveness of laboratory fume hoods than a methodology based on the single-point, breathing zone sampling.ACKNOWLEDGMENT
T
his research was supported by the Ministry of Education of Taiwan, Republic of China.REFERENCES
1. Ivany, R.E., M.W. First, and L.J. Diberardinis: A new method for quantitative, in-use testing of laboratory fume hoods. Am. Ind. Hyg. Assoc. J. 50:275–280 (1989).
2. Fletcher, B., and A.E. Johnson: Containment testing of fume cupboards. I. Methods. Ann. Occup. Hyg. 36:239–252 (1992).
3. Fletcher, B., and A.E. Johnson: Containment testing of fume cup-boards. II. Test room measurements. Ann. Occup. Hyg. 36:395–405 (1992).
4. Saunders, T.: Laboratory Fume Hoods, A User’s Manual. New York: John Wiley & Sons, 1993. p. 56.
5. Volin, C.E., R.V. Joao, J.S. Reiman, E. Party, and E.L. Gershey: Fume hood performance: Face velocity variability inconsistent air volume systems. Appl. Occup. Environ. Hyg. 13:656–662 (1998).
6. Maupins, K., and D.T. Hitchings: Reducing employee exposure poten-tial using the ANSI/ASHRAE 110 Method of Testing Performance of Laboratory Fume Hoods as a diagnostic tool. Am. Ind. Hyg. Assoc. J. 59:133–138 (1998).
7. Ekberg, L.E., and J. Melin: Required response time for variable air volume fume hood controller. Ann. Occup. Hyg. 44:143–150 (2000).
8. Durst, F., and J.C. Pereira: Experimental and numerical investigations of the performance of fume cupboards. Build. Environ. 59:153–164 (1991).
9. Durst, F., and J.C. Pereira: Numerical simulation and optimization of the aerodynamics flow field inside fume cupboards. J. Wind Eng. Ind. Aerodynam. 40:127–145 (1992).
10. Hu, P., D.B. Ingham, and X. Wen: Effect of the location of the exhaust duct, an exterior obstruction and handle on the air flow inside and around a fume cupboard. Ann. Occup. Hyg. 40:127–144 (1996).
11. Hu, P., D.B. Ingham, and X. Wen: Effect of baffles and a louvered bypass on the airflow and the convective patterns of contaminant inside a fume hood. Am. Ind. Hyg. Assoc. J. 59:303–312 (1998).
12. Nicholson, G.P., R.P. Clark, and M.L. De Calcina-Goff: Computational fluid dynamics as a method for assessing fume cupboard performance. Ann. Occup. Hyg. 44:203–217 (2000).
13. Lan, N.S., and S. Viswanathan: Numerical simulation of airflow around a variable volume/constant face velocity fume cupboard. Am. Ind. Hyg. Assoc. J. 62:303–312 (2001).
14. Kirkpatrick, A.T., and R. Reither: Numerical simulation of laboratory fume hood airflow performance. Trans. ASHRAE Career and Technical Education 104:999–1011 (1998).
15. American Society of Heating, Refrigeration and Air Conditioning
Engineers, Inc. (ASHRAE): Method of Testing Performance of
Labora-tory Fume Hoods. (ANSI/ASHRAE Standard 110–1995). Atlanta, Ga.: ASHRAE, 1995.
16. Tennekes, H., and J.L. Lumley: A First Course in
Turbulence. Cambridge, Mass.: The MIT Press, 1972. pp. 223– 247.
17. Kim, T., and M.R. Flynn: The effect of contaminant source momentum on a worker’s breathing zone concentration in a uniform freestream. Am. Ind. Hyg. Assoc. J. 53:757–766 (1992).
18. Flynn, M.R., and B. Ljungqvist: A review of wake effects on worker exposure. Ann. Occup. Hyg. 39:211–221 (1995).
19. American National Standards Institute, Inc. (ANSI): Laboratory Ventilation (ANSI/AIHA Z9.5–2003). [Standard] New York: ANSI, 2003.
