AIHAJ
62:313–321 (2001)
Klaus Willekea Saulius Trakumasa,c Sergey A. Grinshpuna Tiina Reponena Mikhaylo Trunova,d Warren Friedmanb a Aerosol Research and Exposure Assessment Laboratory, Department of Environmental Health, University of Cincinnati, P.O. Box 670056, Cincinnati, OH 45267-0056; E-mail:
[email protected]; b Office of Lead Hazard Control, U.S. Department of Housing and Urban Development, 451 7th St. SW (P 3206), Washington, DC 20410; c On leave from Institute of Physics, Vilnius, Lithuania; d On leave from Odessa University, Odessa, Ukraine
Test Methods for Evaluating the Filtration and Particulate Emission Characteristics of Vacuum Cleaners The overall filtration efficiency of a vacuum cleaner traditionally has been tested by placing the vacuum cleaner in a test chamber and measuring aerosol concentrations at the chamber inlet and outlet. The chamber test method was refined and validated in this study. However, this chamber test method shows an overall filtration efficiency of close to 100% for most of the industrial vacuum cleaners and for most of the newly developed household vacuum cleaners of midprice range or higher because all these vacuum cleaners have a high-efficiency particulate air (HEPA) or other highly efficient filter installed at the exhaust. A new test method was therefore developed through which the vacuum cleaner was probed in various internal locations so that the collection efficiency of the individual components could be determined. For example, the aerosol concentration upstream of the final HEPA filter can thus be measured, which permits one to estimate the life expectancy of this expensive component. The probed testing method is particularly suitable for field evaluations of vacuum cleaners because it uses compact, battery-operated optical particle size spectrometers with internal data storage. Both chamber and probed tests gave the same results for the aerosol filtration efficiency. The probed testing method, however, also gives information on the performance of the individual components in a vacuum cleaner. It also can be used to determine the dust pickup efficiency and the degree of reaerosolization of particles collected in the vacuum cleaner. Keywords: efficiency, lead paint abatement, motor emission, vacuum cleaner
he U.S. Department of Housing and Urban Development Guidelines for the Evaluation and Control of Lead-Based Paint Hazards in Housing indicates that final cleanup after lead-based paint abatement should be performed using a special vacuum cleaner equipped with a high-efficiency particulate air (HEPA) exhaust filter.(1) Lioy et al.(2) showed that a year-long repetitive cleaning protocol, which included an industrial HEPA vacuum cleaner, could significantly decrease lead levels in rugs and other exposed surfaces in homes, and thus reduce the potential for exposure to lead among the occupants. Rhoads et al.(3) reported that the blood lead level decreased by 17% for children living in houses where HEPA vacuuming and wet-mopping were regularly performed during the year. Such reduction is possible only if the dust is picked up and removed efficiently
T This research was supported by the U.S. Department of Housing and Urban Development, Office of Lead Hazard Control, grant no. OHLHR0026-97.
Copyright 2001, AIHA
by the vacuum cleaner. Ewers et al.(4) showed through a laboratory-based study that different models of commercially available HEPA vacuum cleaners may remove different amounts of dust when tested on new carpets embedded with a test dust. Furthermore, vacuuming may increase the level of airborne particles in the surrounding air due to leakage in the vacuum cleaner or the dispersion of dust when loosened by a beater bar or the movement of the nozzle over dusty surfaces.(5,6) These findings indicate that a highly efficient vacuum cleaner must be used for cleaning areas that are contaminated with hazardous particulate materials. Recently, several manufacturers have introduced new household vacuum cleaners with two or more filtration stages and a final HEPA filter. The price of such a new household vacuum cleaner (purchased in 1999) can be as low as about $150, AIHAJ (62) May/June 2001
313
THEORETICAL & EXPERIMENTAL STUDIES
AUTHORS
Ms. #204
THEORETICAL & EXPERIMENTAL STUDIES
FIGURE 1. Chamber facility for vacuum cleaner testing
whereas the price of an industrial HEPA-equipped vacuum cleaner ranges from about $1000 to several thousand dollars. The manufacturers of vacuum cleaners usually test their products in an enclosure using special test particles.(7) The initial filtration efficiency of a vacuum cleaner is determined by measuring the aerosol concentrations upstream and downstream of the enclosure in which the vacuum cleaner is operated in a stationary mode. Unfortunately, such tests do not provide information on the various components in the vacuum cleaner and their changes in performance due to dust loading, air leakage, or material degradation. For example, in a vacuum cleaner with a low efficiency primary dust collector the final HEPA filter gets loaded very quickly by the uncollected dust, resulting in frequent and expensive replacements of the final HEPA filter. Knowing the collection efficiency of the primary dust collector, therefore, allows one to assess the dust load on the final HEPA filter. The goal of this study was therefore twofold: to evaluate the procedures currently used in chamber tests and to develop a new method through which a vacuum cleaner and its various components can be tested outside a chamber, such as in the field.
EXPERIMENTAL MATERIALS AND METHODS Vacuum Cleaner Testing in Chamber Experimental Setup The chamber facility for vacuum cleaner testing is shown schematically in Figure 1. The basic components of this test facility are similar to the ones described by Lioy et al.(7) However, this new facility includes recently developed portable particle size spectrometers that also are being used for vacuum cleaner testing in the field. Thus, field testing methods can be evaluated first in the laboratory facility. The test chamber is made of stainless steel and contains an air 314
AIHAJ (62) May/June 2001
volume of about 0.6 m3 (;21 ft3). The upper part of the chamber is cylindrical, 0.8 m in diameter, and 1.2 m high. The lower part of the chamber funnels into an effluent tube, about 89 mm (53½ inches) in diameter. The vacuum cleaner being tested is placed on top of a metal grid that is installed at the base of the cylindrical chamber section, approximately 30 mm above the bottom of the chamber. During normal testing the vacuum cleaner in the chamber draws air from the laboratory through open Valve 3 and a HEPA filter. This filtered airflow, labeled QFILTERED in Figure 1, entrains the test particles prior to their entering the test chamber. The air emitted by the vacuum cleaner is pushed out of the bottom of the chamber. The airflow through the vacuum cleaner can be increased by Fan 1 when Valve 1 is open. In practice, however, Fan 1 was used primarily to evaluate the particle injection system and possible aerosol losses in the test chamber. In the vacuum cleaner test facility described by Lioy et al.,(7) the vacuum cleaner being tested was surrounded with a ‘‘sheath’’ flow of air. To evaluate this mode of operation, Fan 2 provided additional HEPA-filtered airflow, labeled QSHEATH. A small flow deflector at the upper inlet to the chamber and a panel with 4 mm diameter holes (spaced 25 mm from each other) ensured that the sheath flow was evenly distributed at the top of the chamber. The test aerosol particles were dispersed by a three-jet Collison nebulizer (BGI Inc., Waltham, Mass.). According to ASHRAE, standard filtration tests usually are performed with potassium chloride (KCl) particles.(8) Such particles were generated from a 0.5% KCl solution. The pressure applied to the nebulizer did not exceed 15 psi (;1.02 atm). The test aerosol was diluted with dry HEPAfiltered air, QCLEAN (typically 25 L/min, ;0.9 ft3/min), and then acquired Boltzmann charge equilibrium passing through a 10-mCi 85 Kr electrostatic charge neutralizer (model 3012, TSI Inc., St. Paul, Minn.) before being mixed into QFILTERED. Narrowing of the airflow tube from 4 inches (;102 mm) to 3 inches (;76 mm) diameter and introduction of the aerosol flow along the centerline ensured proper mixing of the test particles into the clean air stream. The particle concentration in the flow entering the test vacuum cleaner (or the test chamber when there is no vacuum cleaner inside), CIN, and in the flow leaving the test chamber, COUT, were measured with optical particle size spectrometers (model 1.108, Grimm Technologies Inc., Douglasville, Ga.). These devices count particles from 0.3 to 20 mm and register them in 15 particle size ranges. The upper particle size limit of these particle size spectrometers is higher than needed for evaluating vacuum cleaners, because high-performance vacuum cleaners essentially collect all particles larger than 1.0 mm, even if no final HEPA filter is installed. High-performance vacuum cleaners with a final HEPA filter installed remove close to 100% of 0.3 mm and larger particles. Lower-performance vacuum cleaners usually do not include a final HEPA filter. They are approximately 100% efficient for 3.5 mm and larger particles. The particle size spectrometers were particle-size calibrated with polystyrene latex (PSL) test particles. These devices report the particle concentrations as a function of particle diameter, dp. Their sampling flow rate, QS, is 1.2 L/min (;0.04 ft3/min). These devices can operate under pressure drops up to DPS ;40 inches H2O. The particle concentrations were measured in 6 sec intervals and were recorded to memory cards that are removable from these particle counters. After or during testing, the data were downloaded to an IBM-compatible personal computer. The sampling lines for measuring the aerosol concentrations at the vacuum cleaner inlet, CIN, and at the test chamber outlet, COUT, were of the same inside diameter and length, and in both
Test Procedure The vacuum cleaner to be tested was placed on the metal grid in the chamber, which was completely sealed against the room environment. The nozzle of the vacuum cleaner was connected to the aerosol delivery system through an adapter. The chamber was flushed with HEPA-filtered sheath air before each experiment until Optical Particle Counter 2 (Figure 1) recorded less than 1023 particles/cm3 of air at the outlet of the chamber. After operating the vacuum cleaner for 30 min, the motor emission, CMOTOR OUT, was recorded by Optical Particle Counter 2 sampling from the outlet of the test chamber. The aerosol generation system was then activated and aerosol concentrations CIN and COUT were recorded three times during a 4-min interval. The test aerosol generator was then deactivated and CMOTOR OUT was recorded one more time after 10 min. The filtration efficiency, E, of the vacuum cleaner was calculated by Equation 1:
1
2
C 2 CMOTOR OUT E 5 1 2 OUT 100% CIN
(1)
where CMOTOR OUT is the average of the motor emissions measured at the beginning and the end of the experiment. The average collection efficiency and standard deviation were calculated from three measurements of CIN and COUT.
Probed Vacuum Cleaner Testing Without Use of a Chamber To be able to test the performance of vacuum cleaners in the field, a test method was developed that measures the particle concentrations near the inlet and outlet of the vacuum cleaner by compact, portable devices mounted on the vacuum cleaner. This is possible now through the availability of the abovementioned optical particle counter, which is quite small (240 mm 3 120 mm 3 70 mm), can be operated by an internal battery, and contains a data logging memory card. Following paragraphs show that this
THEORETICAL & EXPERIMENTAL STUDIES
cases sampled from the center of the 3 inches (;76 mm) diameter tubes. To ensure that such aerosol sampling is representative, the aerosol concentrations also were measured at four other points, equally spaced from each other and located 8 mm from the inner tube wall. The average concentration in the chamber inlet measured near the tube wall differed from the one measured at the tube center by less than 10% over the particle size range of 0.3– 3.5 mm. At the chamber outlet the variability was less than 8%. The lower particle size limit of 0.3 mm was set by the lower threshold of the optical particle size spectrometer. The upper particle size limit was set to 3.5 mm, above the size at which the fractional collection efficiency of non-HEPA-filtered vacuum cleaners reaches 100%. The 3.5-mm size is also the approximate upper limit for counting a statistically sufficient number of particles when the indicated aerosol generation method is used. The flow rates QFILTERED, QSHEATH, and QEFFLUENT were calculated from the flow velocity values measured by air velocity meters (VelociCalc Plus model 8384, TSI Inc., St. Paul, Minn., and model 430–3, Kurz Instruments Inc., Carmel Valley, Calif.). The air temperature and relative humidity were monitored by sensors in the effluent airstream (models TRH-100–20FT and P300–5PSID, Pace Scientific, Inc., Charlotte, N.C.). The pressure drop in the chamber and at the vacuum cleaner inlet was measured with pressure sensors (P300–5PSID, Pace Scientific, Inc.). All of these sensors were calibrated by the manufacturers. The values of flow velocity, temperature, relative humidity, and pressure drop were recorded by a Pocket Logger (model XR440, Pace Scientific, Inc.) connected to a personal computer.
FIGURE 2. Testing of a vacuum cleaner and its components in ambient air (CAMBIENT) without use of a test chamber. Dashed lines represent potential leakage flows. The aerosol generation system upstream of QIN is the same as in Figure 1.
method gives the same results as the chamber test. As seen in Figure 2, the aerosol sampling probes, similar to those used in the chamber test, were installed at the nozzle inlet, CNOZZLE IN, and outlet CNOZZLE OUT, at the bag compartment inlet, CBAG IN, and outlet, CBAG OUT, and at the motor inlet, CMOTOR IN, and outlet, CMOTOR OUT. The sampling probes were inserted into the vacuum cleaner by drilling 2.5 mm diameter holes and positioning the probes with epoxy glue. If desirable, the probes could be removed and the holes closed with either tape or glue. The optical particle counter also measures the ambient aerosol concentration, CAMBIENT, and the aerosol concentration at the outlet of the vacuum cleaner, such as CHEPA OUT after the HEPA filter shown in Figure 2. The overall filtration efficiency of a vacuum cleaner, tested by the probing method, is determined by use of Equation 2:
1
E5 12
2
CHEPA OUT 100% CNOZZLE IN
(2)
Additional probing lines may be installed depending on the number of filtration stages and components of interest in the vacuum cleaner being tested. The collection efficiency of each individual component can be found through Equation 2 by replacing CHEPA OUT/CNOZZLE IN with the ratio of the aerosol concentrations measured at the outlet and inlet of this component. The test particle generation system is the same as the one shown in Figure 1 when the probing method is used in the laboratory.
RESULTS AND DISCUSSION Chamber Versus Probed Testing Figure 3 shows the measured collection efficiency when a vacuum cleaner was tested by the chamber method (Figure 3A) and by probing outside the chamber (Figure 3B). The tested device was AIHAJ (62) May/June 2001
315
THEORETICAL & EXPERIMENTAL STUDIES
FIGURE 3. ‘‘Collection’’ efficiency of a test vacuum cleaner as a function of particle size, as measured by an optical particle size spectrometer. The total aerosol concentration input for both tests was about 3 3 102 particles/cm3 (dp $ 0.3 mm). The calculated collection efficiency is given in quotation marks because it depends on the aerosol concentration of motor emissions relative to the input concentration of test aerosols when the probe samples after the motor before the HEPA filter. (A) Tested in chamber; (B) tested by sampling probes without use of a chamber.
an upright vacuum cleaner of the kind shown in Figure 2. It contained three filtration stages: filter bag, motor prefilter, and final HEPA filter. All results presented in this article were obtained by testing this vacuum cleaner with new filters installed before each test in order to avoid possible changes in performance of the vacuum cleaner due to test dust loading on the filters. During each run with test aerosol (lasting about 25 min) the pressure drop and flow rate through the vacuum cleaner remained constant (DPNOZZLE OUT 5 27 inches H2O, DPBAG OUT 5 226 inches H2O, QIN 5 40 ft3/min); that is, there were no changes in the performance of the vacuum cleaner. Each data point in Figure 3 represents the average value calculated from three measurements. Where no error bars are shown, the standard deviation is less than the size of the symbol. As can be seen from Figures 3A and 3B, the overall collection efficiency of the vacuum cleaner with a HEPA filter installed at the exhaust was higher than 99.98 60.01% over the entire size range from 0.3 to 3.5 mm during both tests (square symbols). All of the HEPA vacuum cleaners from different manufacturers tested so far (one industrial and five household) had initial collection efficiencies higher than 99.97% for particles larger than 0.3 mm. The lines with triangular symbols in Figures 3A and 3B show the collection efficiencies of the vacuum cleaner when the final HEPA filter is removed. In this case, particles of average size 0.35 and 0.45 mm were removed less efficiently: (97.2 60.4)% for dp 5 0.35 mm and (98.9 60.3)% for dp 5 0.45 mm, in the chamber test, Figure 3A; and correspondingly (96.3 60.1)% and (98.8 60.1)% in the probed test outside the chamber, Figure 3B. This shows that at a 5% confident level both test methods give equivalent results for a given particle size. Thus, either method can be used. During the chamber test with the HEPA filter removed, the collection efficiency was calculated by subtracting the motor emissions through Equation 1. During testing with the sampling probes without use of the chamber (Figure 3B, triangles) the 316
AIHAJ (62) May/June 2001
HEPA filter also was removed to ensure the same flow rate through the vacuum cleaner as when tested in the chamber. The collection efficiency was calculated by Equation 2 using CMOTOR IN and CNOZZLE IN. In the probed test the collection efficiency was calculated from the particle concentration measured upstream of the motor. In the chamber test the average motor emission rate has to be included in the calculation. Thus, the probed test (with one variable removed) is more accurate. The lines with circular symbols in Figures 3A and 3B show the ‘‘collection’’ efficiency of the vacuum cleaner tested without a HEPA filter installed and calculated without subtracting the motor emissions (see Equation 1). This is not the actual collection efficiency, because the particle concentration measured downstream of the vacuum cleaner depends on the magnitude of the motor emissions relative to the magnitude of the upstream test aerosol concentration. The data presented in Figures 3A and 3B show that the results obtained during well-defined and controlled vacuum cleaner testing in a chamber facility (Figure 3A) also can be obtained without the use of a chamber by directly probing the vacuum cleaner (Figure 3B). Figure 3 also shows that in the tested vacuum cleaner only a small amount of test aerosol entering the nozzle of the vacuum cleaner loads the final HEPA filter (curves with squares minus curves with triangles). However, the total load on the final HEPA filter is much higher due to motor emissions (curves with triangles minus curves with circles).
Performance of the Test Chamber Facility As shown in Figure 1, the test aerosol flows through several tube contractions and bends, which may result in particle losses in the test chamber facility. Therefore, tests were performed on the influence of potential particle losses on the measured filtration efficiency of the vacuum cleaners tested. Since the sampling probes
THEORETICAL & EXPERIMENTAL STUDIES
FIGURE 4. Proof of minimal particle losses in the vacuum cleaner test chamber. The concentration of KCl test aerosol was measured at the inlet and outlet of the test chamber, with no vacuum cleaner present. Each symbol represents the average of three measurements. The standard deviation for each data point is less than the size of the data symbol. The total measured background particle concentration in the chamber before generating aerosol concentration CIN was less than 1023 particles/cm3. (A) QIN 5 21 ft3/min; (B) QIN 5 42 ft3/min; (C) QIN 5 84 ft3/min.
extract particles from the test aerosol stream at constant velocity, but the different vacuum cleaners to be tested pull airflows of different magnitude through the test facility, the sampling probes were evaluated for nonisokinetic sampling. Proof of Minimal Particle Losses in Vacuum Cleaner Test Chamber Figure 4 shows the KCl particle concentrations (dp 5 0.35–5.0 mm) measured at the inlet and outlet of the test chamber when there was no vacuum cleaner inside the chamber. The airflow rate through the chamber was set by the rotational speed of Fan 1 (Valve 1 open, Valve 3 closed, QSHEATH 5 0, see Figure 1). Figure 4A shows data for QIN 5 21 ft3/min (;600 L/min); Figure 4B shows data for twice that flow rate, QIN 5 42 ft3/min (;1200 L/min); in Figure 4C the flow rate is doubled again to QIN 5 84 ft3/min (;2400 L/min). These flows cover a wide range of commercially available vacuum cleaners. Each data point in Figure 4 represents the average of three measurements. The standard deviation for each data point is less than the size of the data symbol. As seen in Figure 4, the aerosol concentration at the chamber outlet matches the aerosol concentration at the chamber inlet at all three flow rates. Thus, it is concluded that there are minimal particle losses in the test chamber at flow rates through the chamber ranging from 21 to 84 ft3/min (600 to 2400 L/min). (The manner in which the aerosol concentration data are presented in Figure 4 and in subsequent figures deserves comment. The particle size intervals in the output of the optical particle counter are not quite equal on a logarithmic particle size scale, i.e., the logarithm of the upper particle size minus the logarithm of the lower particle size is not quite the same for each size interval [0.3, 0.4, 0.5, 0.65, 0.8, 1.0, 1.6, 2.0, 3.0, and 4.0 mm]. Presentation of the measured aerosol concentrations per logarithmic particle size interval—typical for aerosol physics journals—may result in somewhat smoother curves, but does not directly display the measured aerosol concentrations, nor would it affect the conclusions derived from Figure 4 and subsequent figures. Therefore, the authors chose to present the actual particle concentrations measured in each size interval, instead of adjusting them.)
FIGURE 5. Proof that the aerosol flows in the test chamber may be sampled nonisokinetically over the indicated particle size range at flow rates typical for vacuum cleaners. Each symbol represents the average of three measurements. The standard deviation for each data point is less than the size of the data symbol. (A) Aerosol concentration; (B) aerosol concentration multiplied by the velocity ratio V0/Vs.
Collection Efficiency of Sampling Probes Under Nonisokinetic Conditions It is well known that the representativeness of an aerosol sample depends on how the sample is extracted from its environment.(9) During these experiments all aerosol sampling lines were short (about 130 mm ;5 inches long) and of identical diameter (2.42 mm ;0.095 inch). They were operated at a constant flow rate to the particle size spectrometer, QS 5 1.2 L/min, which results in an aerosol sampling velocity through the sampling probe of VS 5 4.35 m/sec (;856 ft/min). For isokinetic sampling the aerosol velocity through the surrounding duct thus has to be V0 5 856 ft/min. This corresponds to a flow rate through the duct of Q0 5 42 ft3/min. A typical flow rate through a household vacuum cleaner varies from 40 to about 80 ft3/min. By varying the rotational speed of Fan 1, tests were performed at half (Q0 5 21 ft3/ min) and twice that flow rate (Q0 5 84 ft3/min). For Q0 5 21 ft3/min, the duct velocity was V0 5 428 ft/min, whereas the sampling velocity was VS 5 856 ft/min. Thus, the sampling probe samples superisokinetically at VS 5 2V0. Conversely, for Q0 5 84 ft3/min, the sampling probe samples subisokinetically at VS 5 0.5V0. Figure 5A shows aerosol concentrations measured by the optical counters at the three different flow rates while the aerosol input remained the same. Thus, the aerosol concentration was AIHAJ (62) May/June 2001
317
THEORETICAL & EXPERIMENTAL STUDIES
FIGURE 6. Particle emissions by a vacuum cleaner motor measured in the test chamber. Filtered air input (no test aerosol). Each data point represents the measurement performed during 1 min.
highest at the lowest airflow rate of Q0 5 21 ft3/min, and is lowest (four times) at flow rate Q0 5 84 ft3/min. When these measured aerosol concentrations are multiplied by velocity ratio V0/VS, as shown in Figure 5B, all these curves coincide. If one considers that the original test aerosol stream is diluted by airflow Q0, then multiplication of the measured aerosol concentrations C by velocity ratio V0/VS takes the dilution factor back out. The finding that the three curves in Figure 5B are superimposed on each other proves that there is no sampling bias at these low sampling velocities over the particle size range of 0.3 to 3.5 mm. Thus, nonisokinetic sampling does not affect the determination of the filtration efficiency in this test chamber facility. If the test facility had been designed for higher air velocities and the particle sizes had been measured for particles larger than 5.0 mm (not needed for these evaluations), correction factors would have been considered for nonisokinetic sampling.
Motor Emissions Although vacuum cleaners are designed to remove dust particles from surfaces and to capture them, these devices also generate high concentrations of particles themselves. This is primarily due to the wearing of the rotating parts in the motor. Lioy et al.(7) reported that the motor emissions are composed of carbon, chemical binders, and metal impurities created by abrasion of the carbon brushes and commutators. Figure 6 shows time traces of particle emissions from the tested vacuum cleaner motor during a 60-min time interval. Each point in Figure 6 represents a 1-min emission measurement. The motor emissions were measured with the vacuum cleaner operated in the test chamber without test aerosol input (CIN 5 0, QSHEATH 5 0). The results presented in Figure 6 show that almost all of the emitted particles (;99%) are quite small, with dp ,0.65 mm. These motor emission data also show that there is no significant change in emitted particle concentration during 1 hour of operation. Similar particle size distributions of motor emissions were recorded 318
AIHAJ (62) May/June 2001
FIGURE 7. Effect of sheath air on vacuum cleaner performance data. (A) Aerosol particle concentration measured at the outlet of the test chamber, COUT, and at the outlet of the vacuum cleaner nozzle, CNOZZLE OUT, with and without sheath air through the chamber; no test aerosol input. (B) Collection efficiency determined for the vacuum cleaner with different sheath airflows applied; tested with KCl aerosol input.
for several other vacuum cleaners. The particle emission rate from this vacuum cleaner motor, ERMOTOR 5 CMOTORQOUT, was about 7 3 107 particles/min in the size range dp 5 0.3–4.0 mm (Figure 6). The motor emissions may affect a vacuum cleaner’s filtration characteristics during use. In some vacuum cleaner models particles emitted by the motor are captured by the same final filter as the particles that have penetrated through the primary filter of the vacuum cleaner. For instance, in the vacuum cleaner shown in Figure 2 the motor emissions are mixed into the airflow and thus add to the loading of the HEPA filter. In this case, high motor emissions will shorten the lifetime of the final HEPA filter, which is usually somewhat costly. In some vacuum cleaner models the motor emissions are removed by a separate filter or are discarded to the surrounding room air.
Influence of Sheath Air During Chamber Testing As discussed earlier, vacuum cleaners may emit particles generated by the motor and particles that have penetrated through the filters. Unless removed by a well-sealed HEPA filter, these particles are emitted into the air surrounding the vacuum cleaner. Some of these may be drawn back into the vacuum cleaner through leak sites such as those indicated by dashed lines in Figure 2. This
THEORETICAL & EXPERIMENTAL STUDIES
FIGURE 8. Particle concentration measured in different stages of the vacuum cleaner: (A) the nozzle compartment; (B) the bag compartment; (C) the motor compartment. The concentration increases are due to aerosol leakage and/or aerosol generation. The vacuum cleaner was tested outside the test chamber with filtered air input (no test aerosol). The standard deviation is less than the size of the data symbol, when not shown.
reentrainment is illustrated in Figure 7A. With the upright vacuum cleaner placed on the grid in the test chamber (Figure 1) and operated without test aerosol input (CIN ,1023 particles/cm3), the optical particle counter at the chamber outlet registered a high concentration of particle emissions from the motor (open circles, COUT). Some of these emitted particles were drawn back into the vacuum cleaner through leakage in the nozzle, resulting in CIN ,CNOZZLE OUT ,COUT (solid circles). The concentration at the nozzle outlet, CNOZZLE OUT, was measured through a probe by an optical particle counter placed in the chamber. When sheath air was added (65 ft3/min), the particle concentrations measured in both locations were reduced (triangles) due to the dilution effect of the sheath flow. However, when the collection efficiency was determined from data for either no sheath flow or sheath flow up to 65 ft3/min (Figure 7B), the effect of particle leakage back into the vacuum cleaner was found not to affect the collection efficiency determination. The need for sheath flow in vacuum cleaner testing is therefore questionable. Also, separate experiments have shown that the amount of emitted particles reentering the vacuum cleaner depends on the design of the vacuum cleaner, for example, the amount of motor emissions and the location of the leakage sites with respect to the motor emissions. All of these factors determine the proportion of clean sheath air versus emitted particle-laden airflow reaching the leak sites of the vacuum cleaner. (In some vacuum cleaner models the user can add a flow of ambient air to the suction flow to regulate the flow and pressure at the nozzle inlet.) It is difficult to assess how much sheath flow might best simulate an actual field situation. If the sheath flow is eliminated, when testing a HEPA filter-equipped vacuum cleaner in the test
chamber, the aerosol concentrations measured at the chamber outlet are highest, which helps the measurement accuracy. Therefore, elimination of all sheath flows is the best strategy in the opinion of the present investigators.
Probed Test Outside the Test Chamber Without Test Aerosol Input Vacuum cleaners typically are not sealed against ambient air leaks. Thus, ambient aerosol particles may be pulled into the system through components other than the nozzle inlet. This is demonstrated in Figure 8. The diagram of the vacuum cleaner components above the data relates the data to their measurement locations. The vacuum cleaner was tested outside the test chamber exposed to laboratory air with clean air input to the nozzle inlet. As seen in Figure 8A, the particle concentration measured at the nozzle outlet, CNOZZLE OUT, was significantly higher than that measured at the nozzle inlet, CNOZZLE IN, for all particles in the size range dp 5 0.3–3.5 mm. The particle concentration increased due to ambient air leakage and wear of the rotating parts in the nozzle. The aerosol concentration measured at the bag compartment inlet, CBAG IN (open triangles in Figure 8B), was somewhat lower than that measured at the nozzle outlet, CNOZZLE OUT (solid circles in the Figure 8A). Some decrease of the particle concentration may have occurred due to particle settling onto the inner hose wall. The collection efficiency of the new filter bag used in this particular vacuum cleaner was measured to be close to 100% for particles larger than 1.0 mm and about 75% for 0.35-mm particles. In spite of the bag’s high collection efficiency, the aerosol concentration measured at the bag outlet, CBAG OUT, was higher than AIHAJ (62) May/June 2001
319
THEORETICAL & EXPERIMENTAL STUDIES
FIGURE 10. Aerosol particle transmission through different stages of the test vacuum cleaner
FIGURE 9. Aerosol concentration at different stages of the test vacuum cleaner, evaluated outside the test chamber in ambient air with aerosol concentration CAMBIENT and with KCl as test aerosol input
that measured at the bag inlet, CBAG IN (Figure 8B). This was due to the leakage into the bag compartment through insufficient sealing between the bag housing and its lid, which was large with a long perimeter. Ambient aerosol entering the bag compartment through these leakage sites entrained powder particles on the gasket, which apparently had been added so that the gasket would not bond to the surfaces of the lid and housing. As a result, CBAG OUT was higher than CAMBIENT for particles larger than 0.5 mm. The recorded CBAG OUT was almost equal to CBAG IN when all possible leak sites were sealed (dashed line in Figure 8B). The curves in Figure 8C show that the motor prefilter captured all particles larger than 0.8 mm (open squares) and that the rotating parts of the motor emit a large number of particles (solid squares).
Probed Test Outside the Test Chamber With Test Aerosol Input A probed test similar to the one already discussed was performed using KCl particles as test aerosol. The particle size distributions measured at different points of the vacuum cleaner are shown in Figure 9. It is seen that the particle concentrations decreased in each successive stage. Only at the motor stage was the outlet particle concentration, CMOTOR OUT, higher than the inlet concentration, CMOTOR IN; this was due to motor emissions. Figure 9 demonstrates that leakage of particles from the surrounding air into the tested vacuum cleaner does not have any significant effect on the efficiency determination, when the concentration of the test aerosol is much higher than the concentration of ambient air particles. In the example shown in Figure 9, the test aerosol concentration at the nozzle inlet is about 20 times higher than the ambient aerosol concentration, and the aerosol 320
AIHAJ (62) May/June 2001
concentration measured at the bag outlet is higher than the ambient air concentration. The curves presented in Figure 10 show the percentages of particles in the size range from 0.3 to 3.5 mm transmitted through the different stages of the vacuum cleaner. For example, at the nozzle exit, the particle concentration was only 60–80% of the particle concentration entering the nozzle (solid circles). Although some of the test particles may have deposited onto the inner parts of the nozzle and onto the rotating beater bar, the majority of the particle concentration decrease in the nozzle was probably due to air dilution through leak sites in the nozzle. As seen in Figure 10, the aerosol transmission curve for the nozzle is fairly flat at 60– 80% over the entire particle size range, whereas for the hose it decreases from about 90% for smaller particles to about 35% for the larger test particles. The latter is typical for particle losses in a transmission line: the larger particles gravitationally settle to the inner wall or are removed inertially in bends of the hose.(9) The flatness of the transmission curve for the nozzle suggests that the primary cause for the shape of this curve is different. The researchers postulate that air leakage into the nozzle diluted the aerosol concentration over the entire particle size range. (Room air was relatively clean compared with the test aerosol concentration.) Potential leak sites at the nozzle are shown in Figure 2. As discussed before, the magnitude of the CNOZZLE OUT curves in Figure 7 prove that there was leakage into the nozzle because there was no test aerosol input for this test. The curve with solid triangles shows that the aerosol concentration of 1.0 mm and larger particles downstream of the bag was about 5% of the particles upstream of the bag; that is, only about 95% of these particles were collected, although earlier it was found that the initial collection efficiency of the same bag was close to 100% for the larger particles. This discrepancy is due to leakage of ambient aerosols into the bag compartment, as previously proven in Figure 8B. Only a small percentage of particles smaller than 0.5 mm penetrated the combination of bag and motor prefilter. Transmission through the HEPA filter at the vacuum cleaner exit was about zero for all particles larger than 0.3 mm.
Chamber Versus Probed Testing The data presented in Figures 3A and 3B show that testing a vacuum cleaner in a test chamber versus testing it in the open environment through sampling probes gives similar results for the
CONCLUSIONS he chamber test method was evaluated and shown to be effective in determining the overall filtration efficiency of a vacuum cleaner. A new vacuum cleaner testing method by probing was introduced. This new method utilizes a compact, battery-operated optical particle size spectrometer with internal data storage. With this device the aerosol concentration can be measured at any location in the vacuum cleaner that is accessible to probing. The probed testing method measures the same overall initial filtration efficiency as the chamber test when the vacuum cleaner is probed at its inlet and outlet. Testing of vacuum cleaners equipped with a final HEPA filter has shown that all test particles larger than 0.3 mm are collected with higher than 99.97% efficiency. This indicates that parameters other than the initial overall filtration efficiency may need to be considered when choosing a vacuum cleaner with final HEPA filter. For example, measurement of the aerosol concentration upstream of the final HEPA filter permits one to estimate the life of the HEPA filter (which is expensive to replace). If several vacuum cleaners equipped with a final HEPA filter are tested, and all show an initial collection efficiency of about 100%, measurement of the initial collection efficiency of the primary filter provides valuable information related to the frequency and cost of replacing the final HEPA filter. The data presented in this study show that probed vacuum cleaner testing provides more information than overall testing in
T
a chamber. With the probed testing method one can track changes in the particle concentration as the aerosol flow passes through individual stages of the vacuum cleaner. These data allow determination of the particle transmission through and collection by different elements of the vacuum cleaner. The leakage sites into or out of the vacuum cleaner also can be located by the probed testing method. The probed testing method is well suited for field evaluations of vacuum cleaners. In addition to the filtration efficiency, the dust pickup efficiency by the nozzle and the degree of reaerosolisation of dust particles collected in the vacuum cleaner also may be assessed when the vacuum cleaner is probed in field evaluations.
REFERENCES 1. U.S. Department of Housing and Urban Development (HUD): Guidelines for the Evaluation and Control of Lead-Based Paint Hazards in Housing (HUD publication 1539-LBR). Washington, D.C.: HUD/Office of Lead Hazard Control, 1995. 2. Lioy, P.J., L.M. Yiin, J. Adgate, C. Weisel, and G.G. Rhoads: The effectiveness of home cleaning intervention strategy in reducing potential dust and lead exposures. J. Expos. Anal. Environ. Epidemiol. 8: 17–36 (1998). 3. Rhoads, G., A.S. Ettinger, C.P. Weisel, et al.: The effect of dust lead control on blood lead in toddlers: a randomized trial. Pediatrics 103:551–555 (1999). 4. Ewers, L., S. Clark, W. Menrath, P. Succop, and R. Bornschein: Clean-up of lead in household carpet and floor dust. Am. Ind. Hyg. Assoc. J. 55:650–657 (1994). 5. Ronborg, S.M., L.K. Poulsen, P.S. Skov, and H. Mosbech: Effect of two different types of vacuum cleaners on airborne Fel d I levels. Ann. Allergy Asthma Immunol. 52:307–310 (1999). 6. Woodfolk, J.A., C.M. Luczynska, F. de Blay, M.D. Chapman, and T.A. Platts-Mills: The effect of vacuum cleaners on the concentration and particle size distribution of airborne cat allergen. J. Allergy Clin. Immunol. 91:829–837 (1994). 7. Lioy, P.J., T. Wainman, J. Zhang, and S. Goldsmith: Typical household vacuum cleaners: The collection efficiency and emissions characteristics for fine particles. J. Air Waste Manage. Assoc. 49:200– 206 (1999). 8. American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc.: Method for Testing General Ventilation AirCleaning Devices Used for Removal Efficiency by Particle Size [ASHRAE Standard 52.2–99]. Atlanta: ASHRAE, 2000. 9. Brockmann, J.E.: Sampling and transport of aerosols. In K.Willeke and P.A. Baron, editors, Aerosol Measurement: Principles, Techniques and Applications, pp. 77–111. New York: Van Nostrand Reinhold, 1993.
AIHAJ (62) May/June 2001
321
THEORETICAL & EXPERIMENTAL STUDIES
overall filtration efficiency. When the vacuum cleaner was tested with the final HEPA filter installed, the collection efficiency was close to 100% in both cases over the particle size range 0.3–3.5 mm (square symbols). Similar good agreement between the two test methods also was found when the vacuum cleaner was tested without a HEPA filter installed (triangles). The data presented in Figure 7A show how leakage to the nozzle was detected. In this case a combination of the two test methods was used. The same nozzle leakage also was detected when only the probed testing method was used (Figure 8A). Furthermore, leakage to the bag compartment also was located during this test (Figure 8B). These data and the data presented in Figures 9 and 10 lead to the conclusion that the probed testing method yields more information than the chamber testing method. Also, only probed testing can be used for vacuum cleaner evaluations in the field.
