Download
this Paper – MS Word, Adobe Acrobat
Particle
Separation Efficiency of a Uniflow Deduster with Different Types of Dusts
Member ASHRAE
Based on a sensitivity analysis of design parameters, a prototype uniflow deduster - a cyclone type dust separator - was developed and evaluated. The prototype has a total pressure drop of 0.4 in water column (100 Pa) and a capacity of 288 cfm (135 L/s). The prototype was used to verify the uniflow particle separation theory and to evaluate the particle separation efficiency. Two types of dust were used in the evaluation; an artificial commercial dust and real swine building dust. It was confirmed that it is possible to separate particles smaller than 10 mm from an air stream for large volume of air flow. The cut-size of the deduster ranged from 1 to 4.5 mm, depending on whether a laser particle counter or an aerodynamic particle sizer was used. In terms of particle count concentration measured using particle counters, the particle separation efficiency of the deduster was 90% for particles larger than 10 mm and 77% for particles larger than 7 mm. In terms of mass concentration measured using mass samplers, the particle separation efficiency was 85%. Because most of the dust mass is attributed to the larger particles, the number separation and mass separation efficiency agreed very well. The deduster could be an effective method for air cleaning of dusty airspaces.
Keywords: Particle separation efficiency, air quality
Yuanhui
Zhang is an Associate Professor, Xinlei
Wang is a Graduate Research Assistant, Gerald
Riskowski and L. L. Christianson
are Professors and Steve Ford is a
Research Engineer in the Department of Agricultural Engineering, University of
Illinois at Urbana-Champaign, Urbana, IL, 61801.
This project was sponsored by ASHRAE Research Project 878-TRP.
INTRODUCTION
Possibilities for control of dust in an airspace have been extensively reviewed (Dawson, 1990, Zhang,
1999a). Common strategies for dust control include suppressing dust production, accelerating dust
sedimentation, separating dust from the air stream with air cleaning devices and ventilation. Some dust
suppression methods include spraying or sprinkling oil or oil-soap solution in the airspace (Takai et al.,
1993; Zhang et al., 1995, 1996). To accelerate the dust sedimentation onto the floor, air ionization
systems have been investigated (Veenhuizen and Bundy, 1990; Tanaka and Zhang, 1996).
A number of mechanical methods for dust control have been suggested (Owen, 1982; Carpenter,
1986). These methods include fiber filters, water or oil scrubbers, electrostatic precipitators andtraditional cyclone type aerodynamic dedusters. These methods may be associated with theventilation system in the barns, and involve large quantities of airflow through the equipment. It appears
to be too expensive to use mechanical air cleaners to clean large amounts of dusty air. Additionally,
existing mechanical filters were not applicable for separating respirable (smaller than 10 ?m) particles
(Carpenter, 1986). There was a need to explore potential techniques and thus to enhance our
knowledge or information of reducing airborne dust in dusty air spaces.
Conventional air cleaning devices employ a filtration media to scrub dust particles from an
airstream. The filtration media must make contact with dust particles to remove them from the air. This contact
process accumulates dust onto the media quickly and the media requires frequent maintenance or
replacement.
If dust particle removal from an airstream can be made through a non-contact process, an air cleaning
device may not need frequent maintenance or replacement. Traditional cyclones using similar principles
to the uniflow deduster (i.e., non-contact process) have been studied since the early 1930's, and a
number of commercial cyclones were developed (Perry and Green, 1984) since then. The application
of these cyclones are largely limited in industry primarily because of to two limitations: (1) high energy
consumption; and (2) low dust separation efficiency. These two limitations are in turn primarily due to
the high pressure drop across the cyclone. In traditional cyclone design, it was assumed that a high
pressure drop (which leads to high energy requirement) was required to create high swirl air velocities to
force the particle to separate from the airstream. These high air velocities were associated with high
turbulence intensity and particle reentrainment (low particle separation
efficiency).
Some important factors such as the traveling distance and turbulence intensity were also missing in the
existing particle separation theories (Heinsohn, 1991; Ogawa, 1984; Ogawa et al., 1993). There has
been little research on uniflow type cyclones in the past few decades. Zhang (1999b) developed a
particle separation equation for uniflow dedusters and identified key parameters affecting the particle
cut-size.
The purpose of this project was to (1) develop a uniflow deduster - a cyclone type particle separator -
with high particle separation efficiency and low pressure requirement; and (2) to evaluate the deduster
performance with different types of dust in different types of buildings.
PROTOTYPE
DEVELOPMENT
Based on a sensitivity analysis (Zhang, 1999b), key parameters affecting the particle cut-size and the particle separation efficiency were determined. In this prototype, the length of the separation chamber (L) was 42 inches (1.07 m), the average radius of the annular tunnel (Ra) was 3-1/3 in. (0.084 m), the gap of the annular tunnel was 1.5 in (0.038 m) and the vane angle was 60o. Tangential air velocity was maintained at or above 600 ft/min (3 m/s).
To minimize the electrostatic effect in the separation annular tunnel, several duct materials were tested:
galvanized sheet steel, plain PVC and acrylic sheet. Airflow carrying dust at a high pressure was forced
to pass along the surface of a duct. Then the dust collected on the duct was collected and weighed. It
was found that the surface roughness of duct materials such as galvanized sheet has a much higher effect
on dust particle removal than the electrostatic effect. Therefore, PVC duct was selected for the
deduster because it has a smooth surface. As shown in Figure 1, dusty air is drawn from the air inlet
passing through a set of vanes to establish a spiral flow pattern. The air then passes through the annular
tunnel and converges at the exit section above the dust bunker. Particles are collected in the dust
bunker and clean air is exhausted through the blower. This prototype has a total pressure drop of 0.4
in. water (100 Pa) and a capacity of 400 cfm (188 L/s).
Figure
1. The prototype of the uniflow deduster was fabricated based on the sensitivity analysis.
Air was drawn into the annular tunnel through a set of vanes with an angle of 60o.
To visualize the path of particles and flow patterns in the annular tunnel, a transparent acrylic outer tube
was used and a transparent dust deposit box was attached. A white inner tube was selected for better
visualization of airflow pattern studies. To reduce the turbulence intensity in the deduster chamber, inlet
and outlet endings were streamlined and fixed to the ducts (Figure 2). At the entrance of the
deduster, the bell-shaped fitting is connected to the outer tube and the hemisphere-shaped fitting is connected to
the inner tube, thus forming an annular nozzle shape to minimize the resistance and turbulence. Eight
vane blades were evenly placed around the annular tunnel and angled at 60? with respect to the center
axial line of the deduster (Figure 2a). A variable speed fan was attached to the outlet of the deduster to
pull air through the unit.
|
|
|
(a) (b)
Figure 2. The inlet and outlet of the deduster was
streamlined to reduce the turbulence intensity in the deduster chamber. (a) Eight vane blades were evenly placed around the annular
tunnel and angled at 60° with respect
to the duct central axis; and (b) the inner
tube of the annular tunnel was converged with a smooth streamline to guide the
air to exit the deduster.
PARTICLE SIZE DISTRIBUTION
Two different types of dust were used to evaluate the particle separation of the deduster, a
commercially available artificial dust, and real swine building dust. The commercial dust was used in a
full size test room in the Room Ventilation Simulator (RVS). The RVS is an air-conditioned single-zone
building capable of maintaining air temperature at any point between -25 to 40oC with an accuracy of
±1.5 oC yearly around. Thus the RVS can simulate the weather conditions yearly around regardless of
the seasonal changes. In this study, the temperature in the RVS was maintained at
22±1.5
oC. The test room used in this study measured 18 x 12 x 8 ft (L x W x H). The swine dust was measured in a swine
grower/finisher building. The size distributions of the two types of dusts are shown in Figure 3.
(a)
(b)
Figure 3.
Dust size distribution used in the evaluation of the deduster:
(a) a commercial artificial dust; and (b) dust in a swine grower/finisher
building.
The count mean diameter of the commercial dust was 1.61 ?m with geometric standard deviation of 1.58 µm. The count mean diameter of the swine building dust was 2.28 µm with geometric standard deviation of 2.1 µm. The mass mean diameter of the commercial dust was 9.1 µm with geometric standard deviation of 2.32 µm. The mass mean diameter of the swine building dust was 11.28 µm with geometric standard deviation of 1.75 µm. There were very few large particles (> 7 µm) in the commercial dust, while there are a large number of large particles in the swine building dust. Clearly, the commercial dust had a log-normal size distribution, while the swine building dust had a rather irregular size distribution.
PERFORMANCE EVALUATION
The deduster was evaluated in a laboratory setting and a typical swine building. All particle separation
efficiencies were calculated using the differences of dust concentrations in the upstream and downstream
air. Sampling conditions for the three samplers used, aerodynamic particle sizer, laser particle counter
and a multipoint dust mass sampler, were isokinetic.
Evaluation in the Room Ventilation Simulator (RVS)
The deduster was evaluated in a test room in the Room Ventilation Simulator (RVS) using the commercial dust. A turning table dust generation system was used to evenly supply dust in the room. Dust was distributed to the room airspace through 25 ports evenly placed on the entire floor area (Wang et al., 1999a). The room air was first filled with the commercial dust to a concentration of approximately 3 mg/m3 at the level of 2.67 ft (0.8 m) above the floor using a dust generation system. The dust concentration at this height from the floor varied between 1.2 –3.5 mg/m3 during the test. Because the RVS room is relatively small and has a slot air inlet on one side wall and a slot outlet on the other side wall, it was very difficult to achieve complete mixing conditions. Thus, it was very difficult to achieve a consistent spatial and temporal stable dust concentration when conducting the tests. Even in a well mixed flow pattern, the dust spatial distribution had a large gradient from floor to ceiling (Wang et al., 1999b) due to the gravity effect. Dust concentration was measured using an aerodynamic particle sizer that has a range of 0.5 to 20 µm in aerodynamic diameter.
To minimize the error due to the uneven spatial distribution and time variation, a total of 20 sets of data,
10 from upstream and 10 from downstream, all from 3 ft (0.9 m) above the floor, were collected in the
RVS for the deduster. The particle separation efficiency then was calculated using differences of dust
concentrations at the upstream and downstream. Each data point in Figure 4 represents 10 data points
with a standard deviation shown as the vertical bar.
Figure 4. Particle separation efficiency of the deduster using the commercial dust in the Room Ventilation Simulator
Because the count mean diameter of the commercial dust was only 1.61 µm, which is much smaller than the cut-size of the deduster, the dust separation efficiency is low for small particles (typically smaller than 3 µm). The dust separation efficiencies are plotted only for particle sizes of 0.5 to 7 µm because there were very few large particles in the commercial dust. From Figure 4, the separation efficiency is 76% for 7 µm and the cut size is 4.6 µm. It is reasonable to assume that the particle separation efficiency was higher than 76% for particles larger than 7 µm because the larger the particles, the larger the centrifugal force and thus the easier to separate from the air flow.
Evaluation
in a Swine Grower/Finisher Building
The deduster was evaluated in a swine grower/finisher building. The building measures 80 ft long and
40 ft wide. The capacity is 200 grower/finisher swine fed with mashed corn feed. The floor is partially
slotted. The total mean dust concentration in the room air was 1.7 mg/m3 at 0.8 m 2.67 ft (0.8 m)
above the floor.
The dust separation efficiency was measured using both an aerodynamic particle sizer (APS) and a laser
particle counter (LPC). Each point in Figure 5 is the mean value of 15 measurements. The standard
deviation is shown as the vertical bar across the data point. Because the swine building dust had much
larger particles, dust separation efficiency is plotted over the entire measurement range of the APS, from
0.5 to 20 µm. The dust separation efficiency achieved 90% for particles of 10
µm and larger (Figure 5). For particles smaller than 7 µm, the separation efficiency was 77% . The particle cut-size was 4.5
µm. For particles smaller than 7 µm, dust separation efficiencies were slightly higher than the
commercial dust.
Figure 5. Particle separation efficiency of the deduster in a swine grower/finisher building measured using an aerodynamic particle sizer.
The tendency of decline of the particle separation efficiency for particles larger than 10 µm is due to the sampling efficiency of the aerodynamic particle sizer (APS). The APS used in the study has specific sampling characteristics. The sampling efficiency of the APS decreases as the particle size increases. Typically, the sampling efficiency is only about 50% for particles larger than 10 µm. Although the particle separation efficiency is calculated using the difference between the upstream and downstream concentrations, which counteracts some error caused by low sampling efficiency , large variation and errors occurred in the particle separation efficiency.
To verify the particle separation efficiency, dust concentrations at upstream and downstream of the
deduster were also measured using a portable laser particle counter (LPC) and a dust mass sampler.
The LPC measures particle number concentration in six size ranges: 0.3 – 0.5
µm, 0.5 – 0.7 µm, 0.7 – 1.0 µm, 1.0 – 5.0 µm, 5 – 10 µm, > 10 µm. Since the small
particles (< 0.5 µm) are primarily from outside and not the major concern of this study (Zhang et al., 1994), only large particle separation
efficiency was calculated. The size distributions for the upstream and downstream of the deduster are
shown in Figure 6. Each data point in Figure 6 is the mean of five measurements, each was measured at
the same level as the APS. The total particle concentrations and separation efficiencies from the APS
and LPC data agree very well. However, the cut-size measured using the LPC was about 1
µm, which is much smaller than the APS measurement. The difference in particle cut-sizes measured using the
APS and LPC remain unclear. Both the APS and LPC were calibrated by the manufacturers prior to
the tests, and both were measured under isokinetic sampling conditions. The authors had no means to
perform a direct comparison of the APS and LPC.
The dust mass sampler consisted of six 37 mm diameter filters (0.8 µm porosity), each attached to a critical flow control venturi which maintained an accurate sampling rate at 20 liter per minute. Three filters were placed upstream of the deduster and three downstream. The filters were dried and weighed both before and after the air sampling in the swine building. The dust mass collected onto the filters was then used to calculate the dust concentration reduction at upstream and downstream. The dust mass reduction was 85%. Because most of the dust mass is attributed to the large particles, the number separation and mass separation efficiency agreed very well.
Figure 6. Dust separation
efficiency of the deduster in a swine grower/finisher room measured using a
laser particle counter.
There are two major differences between the dust in swine barns compared to the commercial dust. First, the dust spatial distribution in the swine barn is much more uniform and more stable over time than in the RVS. Second, there is a larger portion of larger particles (typically larger than 5
µm) in swine building dust than in the commercial dust in the RVS.
Comparison of the dust separation efficiency for commercial dust and swine building dust, the deduster showed a higher dust separation efficiency in swine buildings for large particles (> 7
µm). It was also observed that during the experiment, the deduster collected a noticeable amount of dust in the dust bunker in the swine building while much less was collected in the RVS with the commercial dust. There are at least three reasons for these variations. First, the commercial dust is primarily composed of small particles with a count mean diameter of 1.61
µm, while the swine building dust has a mean diameter of approximately 2.28
µm. Since the deduster is more efficient in separating large particles, there was a higher dust separation efficiency and more dust was collected for the swine building dust. Second, there is likely more oil in the swine building dust than the Arizona dust. It is a common practice to add 1-2%
of oil to swine feed, and the dust generated from the fecal materials and dander are high in protein content. Therefore, the swine dust is more adhesive and not easily re-entrained into the air flow once it is separated. The commercial dust is less adhesive and more easily re-entrained into the air stream. Third, the relative humidity was about 50% in the swine building and about 30% in the Room Ventilation Simulator. The relative humidity may also change the adhesion of the separated dust and hence the reentrainment of the separated particles.
CONCLUSIONS
A uniflow deduster was developed to verify that it is possible to separate particles smaller than 10
µm from an air stream with centrifugal action. This implies that it is applicable to use dedusters for large volume air cleaning in very dusty environments. Specifically, the following conclusions can be drawn from this study:
The cut-size ranged from 1 to 4.5
µm, depending on whether measurements were taken with a laser particle counter or an aerodynamic particle sizer.
In terms of particle count, the particle separation efficiency of the deduster was 90% for particles larger than 10
µm and 77% for particles larger than 7 µm.
In terms of dust mass, the particle separation efficiency of the deduster was 85%, which agrees well with the particle count separation efficiency.
ACKNOWLEDGEMENT
The authors wish to acknowledge ASHRAE for the financial support (project ASHRAE 878-TRP).
Thanks are extended to P. Stroot, J.R. Hubele, R. Brown for their assistance in collecting data.
REFERENCES
Carpenter, G. A. 1986. Dust in livestock buildings -- Review of some aspects. J. Agr. Engng. Res. 33: 227-241.
Dawson, J.R. 1990. Minimizing dust in livestock buildings - A review of possible alternatives to mechanical separation. Journal of Agricultural Engineering Research. 47: 235-248.Heinsohn, R.J. 1991. Industrial ventilation. John Wiley & Sins. New York. Pp. 475-476.Ogawa, A. 1984. Separation of particles from air and gases, Boca Raton, Florida: CRC Press Inc. Pp. 42-45.
Ogawa, A., K. Sugiyama, and K. Nagasaki. 1993. Separation mechanism for fine solid particles in the uniflow type of the cyclone dust collector. Paper presented at Filtech Conference,
Horsham, West Sussex, UK. Pp. 627-640.
Owen, J.E. 1982. Dust--filtration solutions and their cost. Farm Building Programs. 68: 19-23.
Perry, R.H. and D.W. Green. 1984. Perry's Chemical Engineer's Handbook: Chapter 20. McGraw-Hill, New York. Pp. 20-80 to 20-99.
Takai, H., F. Moller, M. Iverson, S.E. Jorsal, and V. Bille-Hansen. 1993. Dust control in swine buildings by spraying of rape seed oil. Proceedings of 4th Livestock Environment Symposium, ASAE, pp. 726-733.
Tanaka, A. and Y. Zhang. 1996. Evaluate the dust removal efficiency and static electrical effect of a negative ionization system in cold climate. J. Agr. Safety and Health 2(1): 39-47.
Veenhuizen, M.A. and D.S. Bundy. 1990. Electrostatic precipitation dust removal system for swine housing. ASAE Paper No. 904066. St. Joseph, MI. USA.
Wang, X., Y. Zhang, L.Y. Zhao, G.L. Riskowski. 1999a. Development of a multi-point aerosol sampler using critical flow control devices. Transactions of ASHRAE 105(2): SE-99-17-03.
Wang, X., Y. Zhang and G.L. Riskowski. 1999b. Dust spatial distribution in a typical swine building. In proceedings of International Symposium on "Dust Control for Animal Production Facilities", Aarhus, Denmark, May 30 – June 2, 1999.
Zhang, Y. 1999a. Engineering control of dust for animal facilities. In proceedings of International Symposium on "Dust Control for Animal Production Facilities", Aarhus, Denmark, May 30 – June 2, 1999. Pp. 6-11.
Zhang, Y. 1999b. Modeling and Sensitivity Analysis of Dust Particle Separation for Uniflow Dedusters. Transactions of ASHRAE: In review.
Zhang, Y., A. Tanaka, E. M. Barber and J. J. R. Feddes. 1996. Effect of frequency and quantity of sprinkling canola oil on dust reduction in swine buildings. Transactions of ASAE 39(3): 1077-1081.
Zhang, Y., E. M. Barber, J.J.R. Feddes and J.F. Patience. 1995. Identification of oils to be sprinkled in swine barns to reduce dust. Transactions of ASHRAE 101(2): 1179-1191.
Zhang, Y., L. Nijssen, E. M. Barber, J. J. R. Feddes and M. Sheridan. 1994. Sprinkling mineral oil to reduce dust in swine buildings. Transactions of ASHRAE 100(2): 1043-1052.
