Applied Engineering in Agriculture

Vol. 23(3): 347-355 � 2007 American Society of Agricultural and Biological Engineers ISSN 0883−8542 347

COLLECTION EFFICIENCY EVALUATION OF BAFFLE-TYPE PRE-SEPARATOR CONFIGURATIONS: EFFECTS OF

BAFFLE LOCATION AND INLET VELOCITIES

M. D. Buser, D. P. Whitelock, G. A. Holt, C. B. Armijo, L. Wang

ABSTRACT. Some cotton gins across the Cotton Belt use an abatement system consisting of a baffle-type pre-separator followedby cyclones to control the emissions from the cotton gin’s various exhausts. Very limited information exists in the literaturewhich discusses the effects of baffle location and critical velocity on the pre-separator’s collection efficiency. In this study,a range of critical air velocities and loading rates were evaluated to determine the effect of baffle location of the pre-separa-tor’s collection efficiency when using cotton gin waste. None of the treatments significantly affected the over-sized cycloneor over-all collection efficiency. The pre-separator collection efficiency was higher (81%) when the baffle placed at one-thirdthe overall width of the pre-separator from the inlet than when placed at one-half (78%) or two-thirds (75%). The pre-separa-tor collection efficiency was 79.4% at 18.3-m s-1 (3600-fpm) inlet velocity which was significantly higher than 78.2% at20.3 m s-1 (4000 fpm) and 78.5% at 22.4 m s-1 (4400 fpm). Loading rate did significantly affect the pre-separator efficiency,but not to the extent of inlet velocity. The sieve analysis indicated that the pre-separator removed the majority of materiallarger than 180 �m; however, the pre-separator did allow a substantial amount of lint to pass through to the cyclone. Thebaffle-type pre-separator performed well in reducing the course material loading rate entering the cyclone.

Keywords. Cotton gin, Cyclone, Particulate matter, 1D-3D, Emissions, Pre-separator, Trash, Waste.

he most common abatement devices used on cottongin exhausts are cyclones. These devices are typi-cally used as a single stage control technology andare therefore used to remove the larger gin waste

(e.g. sticks, burrs, lint, and leaf trash) and fine dust (particu-late matter less than 100 �m in diameter). Early cyclonesused in the cotton ginning industry were large-diameter, low-velocity devices designed primarily for the collection of largetrash. During the 1960s, the high-efficiency, small-diametercyclone, commonly referred to as the 2D-2D design, was de-veloped for the cotton ginning industry in an effort to reduce

Submitted for review in January 2006 as manuscript number SE 6311;approved for publication by the Structures & Environment Division ofASABE in March 2007.

Use of a trade name, propriety product or specific equipment does notconstitute a guarantee or warranty by the United States Department ofAgriculture and does not imply approval of a product to the exclusion ofothers that may be suitable.

The authors are Michael D. Buser, ASABE Member Engineer,Agricultural Engineer, USDA Agricultural Research Service, CottonProduction and Processing Research Unit, Lubbock, Texas; Derek P.Whitelock, ASABE Member Engineer, Agricultural Engineer, USDAAgricultural Research Service, Southwestern Cotton Ginning Laboratory,Mesilla Park, New Mexico; Greg A. Holt, ASABE Member Engineer,Agricultural Engineer, USDA Agricultural Research Service, CottonProduction and Processing Research Unit, Lubbock, Texas; Carlos B.Armijo, ASABE Member, Textile Technologist, USDA AgriculturalResearch Service, Southwestern Cotton Ginning Laboratory, Mesilla Park,New Mexico; and Lingjuan Wang, ASABE Member Engineer,Agricultural Engineer, Department of Biological and AgriculturalEngineering, North Carolina State University, Raleigh, North Carolina.Corresponding author: Michael D. Buser, USDA Agricultural ResearchService, Cotton Production and Processing Research Unit, 1604 E.FM1294, Lubbock, TX 79403; phone: 806-746-5353; fax: 806-744-4402;e-mail: mbuser@lbk.ars.usda.gov.

particulate matter (PM) emissions (Harrell and Moore, 1962;Baker and Stedronsky, 1967). In the late 1970’s, Parnell andDavis (1979) introduced the 1D-3D cyclone design whichwas reported to have a higher fine dust collection efficiencythan the 2D-2D cyclone design. Studies have shown thatsingle conventional cyclones, similar to the 2D-2D design,could remove 10-µm particles with 85% to 90% efficiency,5-µm particles with 75% to 85% efficiency, and 2.5-µm par-ticles with 60% to 75% efficiency (EC/R Incorporated, 1998)and could collect particles greater than 20 µm with 100% effi-ciency (Avant et al., 1976). The 2D2D and 1D-3D cycloneswere designed as fine dust collectors; however as previouslystated, most cotton gins use these abatement devices to re-move both the larger gin waste and the fine dust.

According to Baker et al. (1995), secondary abatementtechnologies such as baffle-type pre-separators have beenincorporated into some cotton gin abatement systems in orderto overcome problems encountered with using cyclones toremove both the large cotton gin waste and the fine dust.These problems have included: choking in the transition,narrow inlet of the original 1D-3D cyclone design, and trashoutlet; cyclone wear due to the presence of sand and otherabrasives found in the gin waste; and fluctuations in airflowrates entering the cyclones due to changes or modificationsto the processing system prior to the cyclones, which tend todecreased cyclone efficiencies. The typical pre-separatorused in the cotton ginning industry receives the materialladen air from all or the majority of the cotton gin processstreams and removes the large gin waste, thus reducing theloading rates, choking tendency, and wear while balancingthe airflow rates to the cyclones that follow the pre-separator.Although these devices have been used at commercial cottongins in Texas, California, and Australia since the mid-1970s(Baker et al., 1995), there is very little information in the

T

348 APPLIED ENGINEERING IN AGRICULTURE

literature discussing the design and collection efficiencies ofthe baffle-type pre-separator prior to the mid-1990s.

Mihalski et al. (1993) and Mihalski (1995) reported onstudies conducted using a small-scale baffle-type pre-separa-tor designed for a single 10.2-cm (4-in.) diameter 1D-3D or2D-2D cyclone. The test material, cotton gin waste with thefine dust removed and replaced with corn dust, providedknown quantities of both the gin waste and fine dust to beintroduced into the system. The results indicated that usinga pre-separator prior to a 2D-2D cyclone significantlyreduced the emission concentrations in 11 out of 12 compari-son tests and using a pre-separator prior to a 1D-3D cyclonesignificantly reduced the emission concentrations in onlyfour of nine comparison tests (Mihalski et al., 1993). Theoptimum pre-separator design included placing the baffle inthe center of the separator and operating at 5.1- to 6.1-m s-1

(1000- to 1200-fpm) critical velocity (Mihalski, 1995). Usingthe optimum design in series with a 1D-3D cyclone reducedemission concentrations from 170 to 16 mg m-3 whencompared to using only a 1D-3D cyclone to remove bothground cotton gin waste and fine dust at an inlet loading rateof 90 g m-3. Similarly, utilizing the optimum baffle-typepre-separator design in series with a 2D-2D cyclone reducedemission concentrations from 200 to 28 mg m-3 whencompared to using only a 2D-2D cyclone.

Columbus (1994) reported on using a 76.2-cm (30-in.)wide baffle-type pre-separator, with the baffle located at25.4-cm (10-in.) from the inlet and operated at a criticalvelocity of 4.6 m s-1 (900 fpm), in series with a 61.0-cm(24-in.) 1D-3D cyclone and in series with a 66.0-cm (26-in.)2D-2D cyclone to collect cotton gin waste from pickerharvested cotton. Results of the study indicated that thepre-separator reduced the total emissions by 15% to 17%.Baker et al. (1995) used a 76.2-cm (30-in.) wide baffle-typepre-separator, with the baffle located at 25.4-cm (10-in.)from the inlet and varied the critical velocity. The pre-separa-tor was about 80% to 90% efficient in removing stripperharvested cotton gin waste prior to a cyclone, and thepre-separator collection efficiency tended to increase as thecritical velocity decreased to about 8.1 m s-1 (1600 fpm).Results from these tests indicated that using a pre-separatorin series with a 1D-3D cyclone increased the cycloneemissions when compared to using only a 1D-3D cyclone,but this result was attributed to lint re-circulation near thecyclone’s trash exit. The pre-separator in series with a 2D-2Dcyclone tended to reduce emissions by about 8% whencompared to using only a 2D-2D cyclone.

A more recent study conducted by Wang et al. (2004) used10.2-cm (4-in.) 1D-3D and 1D-2D cyclones and a baffle-typepre-separator with the baffle located in the center of theseparator. This study was conducted using tub-ground WestTexas stripper-harvested cotton gin waste. Results from thestudy showed that the overall collection efficiency for bothsystems exceeded 97% and there were no significantdifferences between the emission concentrations for thepre-separator/1D-3D system and pre-separator/1D-2Dsystem.

In the current literature, only two studies have evaluatedbaffle-type pre-separators in terms of baffle location in termsof the distance from the inlet to the baffle and these twostudies were conducted on small scale pre-separators wherethe cotton gin waste was modified from its original state toprevent system choking. The purpose of this study was to

evaluate a 76.2-cm (30-in.) wide baffle-type pre-separator interms of the separator’s collection efficiency, the collectionefficiency of the over-sized cyclone that followed thepre-separator, and the overall collection efficiency withdifferent inlet air velocities, gin waste loading rates, andbaffle locations.

MATERIALS AND METHODSThe test system included: a variable speed conveyor to

regulate the amount of material entering the system (fig. 1);a variable speed fan to convey the material through thesystem; a baffle-type pre-separator (fig. 2); and an over-sizedcyclone. The 38.1-cm (15-in.) wide and 228.6-cm (90-in.)long conveyor belt was equipped with 22.9-cm (9-in.) siderails so that a known volume of material could be placed onthe belt prior to each test. The fan was capable of moving0.83 m3 s-1 (1750 cfm) of air through the entire system. Thelength of the baffle-type pre-separator, from the inlet to theoutlet, was 76.2 cm (30 in.), and the width, from side to side,was 48.3 cm (19 in.). The 3.2-mm (1/8-in.) thick baffle platespanned the width of the pre-separator and extended 55.9 cm(22 in.) into the main body of the pre-separator from the topof the pre-separator. The baffle plate could be positioned atdifferent points between the inlet and outlet. The main bodyof the pre-separator was 96.5 cm (38 in.) tall with a 106.7-cm(42-in.) tall transition down to a 20.3-cm (8-in.) dropper. Ahopper was attached to the bottom of the dropper to collectthe material removed from the air stream by the baffle-typepre-separator. The diameter of the piping entering and exitingthe pre-separator was 20.3 cm (8 in.). One side of thepre-separator was covered with Lexan so the material flowpatterns could be observed during the tests. The over-sizedcyclone that followed the pre-separator was an 86.4-cm(34-in.) diameter 1D-3D cyclone with a 2D-2D inlet and aD/3 trash outlet (Armijo et al., 1992). The cyclone’s trash exitwas equipped with a 208.2-L (55-gal) hopper to collect all thematerial the cyclone removed from the air stream. The

Figure 1. Material feeding system.

349Vol. 23(3): 347-355

Outlet

Adjustable Baffle{Height − 55.9−cm (22−in.)}

Airflow Inlet

Vacuum Dropper

183−cm (72−in.)

20.3−cm(8−in.)

20.3−cm(8−in.)

76.2−cm(30−in.)

76.2−cm(30−in.)

20.3−cm(8−in.)

Gin WasteRemoved

Figure 2. Baffle-type pre-separator.

cyclone exhaust was vented to the atmosphere. An over-sizedcyclone was used so that, while the pre-separator inletvelocity was varied, the cyclone inlet velocity would notfluctuate above and below standard inlet velocity. Cycloneinlet velocity was not critical as cyclone efficiency was notthe focus of the study, but rather was secondary informationcollected since a cyclone was in place to capture thepre-separator emissions.

The material fed through the system for all runs wastub-ground West Texas stripper-harvested cotton gin waste.The average bulk density of the gin waste was 160.2 kg m-3

(10 lb ft-3), based on 10 random samples collected from thewaste pile. Based on the bulk density of the gin waste and theconveyor belt dimensions, the size of the test lots were set at22.7 kg (50 lb). The gin waste was spread onto the conveyorbelt to a depth of 15.2 cm (6 in.) over the entire belt surface.

The study was set up and analyzed as a completelyrandomized block design. The main effects were inletvelocity and gin waste loading rate and the blocking variablewas baffle location. Baffle location was set as a blockingparameter because of the time and labor costs associated withmoving the baffle. Critical velocity is the velocity at thecross-sectional area between the inlet wall of the pre-separa-tor and the baffle and is a function of airflow rate and baffle

location. The three inlet air velocities were 18.3, 20.3,22.4 m s-1 (3600, 4000, and 4400 fpm) and corresponded torespective airflow rates of 36.1, 39.5, 43.5 m3 min-1 (1274,1396, and 1536 cfm). These airflow rates were combinedwith three baffle locations, set at one-third, one-half, andtwo-thirds the overall length of the pre-separator from theinlet [22.9, 38.1, 53.3 cm (9, 15, and 21 in.), respectively],resulted in nine critical velocities: 2.3, 2.6, 2.8, 3.3, 3.6, 3.9,5.5, 6.0, and 6.6 m s-1 (460, 504, 554, 644, 705, 776, 1073,1176, and 1293 fpm). The three gin waste loading rates were1.1, 1.8, and 2.6 kg (2.4, 4.0, and 5.7 lb) of gin waste per min.Forty-five runs, three inlet air velocities × three gin wasteloading rates × five replicates, were randomly assigned toeach of the three baffle locations. The overall designconsisted of 135 runs. Statistical analysis was performedusing SAS General Linear Models (SAS, 1999).

Prior to each run, velocity and static pressure readingswere made to ensure that the airflow rates were properly set.Barometric pressure, relative humidity, and ambient temper-ature readings were collected prior to each run and were usedto calculate air density. If significant differences weredetected in the air density values, then the airflow rate wasadjusted by adjusting the current supplied to the fan motor toaccount for the variation in air density. The parametersmeasured during and after the run included run time, mass ofmaterial captured by the baffle-type pre-separator, and massof material captured by the over-sized cyclone. Once a runwas completed, sub-samples of the material captured by thebaffle-type pre-separator and the over-sized cyclone were setaside for sieve analysis.

The baffle-type pre-separator collection efficiency wascalculated as the mass of material collected by the pre-sepa-rator divided by the mass of material introduced into thesystem [27.7 kg (50 lb)]. The over-sized cyclone collectionefficiency was calculated as the mass of material collected bythe cyclone divided by the mass of material introduced intothe system minus the mass of material collected by thepre-separator. The overall collection efficiency was calcu-lated as the sum of the mass of material collected by thepre-separator and the cyclone divided by the mass of materialintroduced into the system.

A sieve analysis was performed on the samples collectedfrom the pre-separator and the cyclones to determine the sizedistribution of the material. A 50-g sub-sample was placed inthe top sieve of a stack and shaken for 20 min in a Roto-Tap(W.S. Tyler, Mentor, Ohio) shaker. The following sizes ofsieves were used: 16.0 mm (5/8 in.), 9.5 mm (3/8 in.), 8.0 mm(5/16 in.), 4.75 mm (#4), 2.0 mm (#10), 1.4 mm (#14),710 µm (#25), 180 µm (#80), 75 µm (#200), and pan.

RESULTS AND DISCUSSIONThe pre-separator, over-sized cyclone, and overall collec-

tion efficiencies are reported in table 1. Pre-separatorcollection efficiency was significantly impacted by inlet airvelocity and loading rate; however, there were no effects dueto the interaction of inlet air velocity and loading rate.Baffle-type pre-separator collection efficiency was signifi-cantly higher at lower inlet air velocities [22.4 to 18.3 m s-1

(4400 to 3600 fpm)], ranging from 76.5% to 79.4%.Pre-separator efficiency for the lowest gin waste loading rateof 1.1 kg min-1 (2.4 lb min-1) was 77.6% and was significantly

350 APPLIED ENGINEERING IN AGRICULTURE

Table 1. Pre-separator, cyclone, and overall collection efficiencies[a] for the baffle-type pre-separator study.

TreatmentPre-separator

EfficiencyCyclone

EfficiencyOverall

Efficiency

Pre-separator Inlet Velocity,m s-1 (fpm) % % %

18.3 (3600) 79.4 a 78.6 a 95.6 a

20.3 (4000) 78.2 b 79.0 a 95.5 a

22.4 (4400) 76.5 c 78.9 a 95.1 a

Loading Rate, kg min-1 (lb min-1)

1.1 (2.4) 77.6 b 78.4 a 95.2 a

1.8 (4.0) 78.1 a 78.8 a 95.4 a

2.6 (5.7) 78.4 a 79.4 a 95.6 a

ANOVA - P value

Source of variation

Baffle location (block) < 0.01 < 0.01 < 0.01

Inlet velocity < 0.01 0.89 0.07

Loading rate < 0.01 0.58 0.19

Inlet velocity × loading rate 0.53 0.04 0.05

Model < 0.01 < 0.01 < 0.01[a] Means in the same column for a treatment followed by the same letter

are not significantly different as determined by Ryan-Einot-Gabriel-Welch test based on range, REGWQ, (P ≤ 0.05).

lower than that for the other two loading rates. Neithercyclone efficiency nor overall efficiency was significantlydifferent among inlet velocities or loading rates and averaged78.8% and 95.4%, respectively.

The baffle-type pre-separator collection efficiency washigher at baffle locations closer to the inlet. In Figure 3, thebaffle-type pre-separators collection efficiency is shown as afunction of inlet air velocity for the three baffle locations. Asseen in table 1, the pre-separator collection efficiencyincreased with decreasing inlet velocity, and the collectionefficiency increased for a particular inlet velocity as thebaffle location was moved closer to the inlet. The baffle-typepre-separator collection efficiency was also higher for closerbaffle locations for a particular gin waste loading rate (fig. 4).

When comparing pre-separator collection efficienciesamong critical velocities (fig. 5), it appeared that efficiencymay increase with increasing critical velocity. After furtherstudy, it was found that the increase was actually due todecreased distance of the baffle from the inlet. Also, for aparticular baffle location, the pre-separator efficiency de-creased with critical velocity. Based on the data shown infigure 5, it appeared that the airflow rate, and thus criticalvelocity, would have to have been substantially decreasedwhen the baffle was located further from the inlet [i.e.53.3 cm (21 in.) from the inlet] in order to obtain similarcollection efficiencies as those found when the baffle waslocated closer to the inlet [i.e. 22.9 cm (9 in.) from the inlet].

Figure 6 shows the over-sized cyclone collection efficien-cy as a function of pre-separator inlet air velocity and cycloneinlet velocity for the three baffle locations. It appears that theover-sized cyclone collection efficiency was highest whenthe baffle was located 53.3 cm (21 in.) from the inlet.However, when the baffle was located 53.3 cm (21 in.) fromthe inlet, the baffle-type pre-separator collection efficiencywas the lowest allowing more material to enter the cyclone.The cyclone inlet velocities ranged from 7.2 to 8.8 m s-1

(1414 to 1728 fpm), which were well below the recom-mended inlet velocity of 3200 fpm for 1D-3D cyclones.Reducing the inlet velocity from 8.8 to 7.2 m s-1 (1728 to1414 fpm) had little effect on the cyclones collectionefficiency.

There were no significant differences in overall (pre-sepa-rator followed by a 1D-3D cyclone) collection efficienciesdue to pre-separator inlet velocity or loading rate (table 1).The baffle location also had little effect on the overallcollection efficiency (fig. 7), although, the average overallcollection efficiency appeared to be slightly higher when thebaffle was located 53.3 cm (21 in.) from the inlet than whenthe baffle was located 22.9 or 38.1 cm (9 or 15 in.) from theinlet. In general, the pre-separator removed about 80% of thematerial entering the abatement system (fig. 8). Based onvisual inspection, the material removed by the pre-separatorlooked very similar to the material being fed through the

72%

74%

76%

78%

80%

82%

17.8 18.8 19.8 20.8 21.8 22.8

Inlet Air Velocity (m s−1)

Pre−

sepa

rato

r C

olle

ctio

n E

ffici

ency

3500 3750 4000 4250 4500Inlet Air Velocity (fpm)

22.9 (9)38.1 (15)53.3 (21)

Baffle Location cm (in.)

Figure 3. Pre-separator collection efficiencies at pre-separator inlet air velocities. Error bars represent two-sided 95% confidence limits for the means.

351Vol. 23(3): 347-355

73%

75%

77%

79%

81%

83%

1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6

Gin Waste Loading Rate (kg min−1)

Pre−

sepa

rato

r C

olle

ctio

n E

ffici

ency

2.2 2.7 3.2 3.7 4.2 4.7 5.2 5.7Gin Waste Loading Rate (lb min−1)

22.9 (9)38.1 (15)53.3 (21)

Baffle Location cm (in.)

Figure 4. Pre-separator collection efficiencies at gin waste loading rates. Error bars represent two-sided 95% confidence limits for the means.

system, while the material removed by the over-sizedcyclone appeared to consist mainly of lint and fine dust(fig. 9). Photographs of the sieved material: 16.0-mm(5/8 in.), 8.0-mm (5/16 in.), 2.0-mm (#10), 710-µm (#25),75-µm (#200), and smaller captured by the pre-separator areshown in figure 10. There were no significant differences inthe percentages of material captured by the pre-separatorcollected on any of the sieves among pre-separator inlet airvelocities, and only the percentage collected on the 2.0-mm

(#10) sieve was significantly different (Pvalue = 0.02) amonggin waste loading rates. Analysis of variance results showedthat material captured by the pre-separator had significantlyhigher levels of material collected on the 2.0-mm (#10) sieveat the lowest gin waste loading rate (30.2%) than at thehighest loading rate (28.6%). The reason for this differenceoccurring for only the 2.0-mm (#10) sieve was not apparent.

More significant differences were detected for the sieveanalysis of the samples captured by the over-sized cyclones

72%

74%

76%

78%

80%

82%

2.0 3.0 4.0 5.0 6.0 7.0

Critical Velocity (m s−1)

Pre−

sepa

rato

r C

olle

ctio

n E

ffici

ency

400 600 800 1000 1200 1400

Critical Velocity (fpm)

22.9 (9)

38.1 (15)

53.3 (21)

Baffle Location cm (in)

Figure 5. Pre-separator collection efficiency at pre-separator critical velocities. Error bars represent two-sided 95% confidence limits for the means.

352 APPLIED ENGINEERING IN AGRICULTURE

72%

74%

76%

78%

80%

82%

84%

86%

18.0 19.0 20.0 21.0 22.0Pre−separator Inlet Air Velocity (m s−1 )

Cyc

lone

Col

lect

ion

Effi

cien

cy

7.1 7.3 7.5 7.7 7.9 8.1 8.3 8.5 8.7Cyclone Inlet Air Velocity (m s−1 )

22.9 (9)38.1 (15)53.3(21)

Baffle Location cm (in.)

Figure 6. Over-sized cyclone collection efficiencies at pre-separator inlet air velocities and corresponding cyclone inlet air velocities. Error bars repre-sent two-sided 95% confidence limits for the means.

than by the pre-separator. The material captured by thecyclone at the lowest gin waste loading rate (0.3%) had asignificantly (Pvalue = 0.002) lower percentage of materialcollected on the 4.75-mm (#4) sieve than the medium andhighest loading rates (0.4 %). The 2.0-mm (#10), 1.4-mm(#14), and 710-µm (#25) sieves, and the < 75-µm pan showedsignificant differences in collected material among pre-sepa-rator inlet air velocities (table 2). For all three sieves, thelower inlet velocity resulted in significantly lower percent-age of material captured by the cyclone collected on thesieves; 2.1%, 3.4%, and 8.6% for the 2.0-mm (#10), 1.4-mm(#14), and 710-µm (#25) sieves, respectively. On the other

hand, there was a significantly larger percentage of thematerial captured by the cyclone collected on the < 75-µmpan for the lower pre-separator inlet air velocity (18.4%) thanfor the highest velocity (16.7%).

Figures 11 and 12 show the percentage of materialcollected on each sieve, based on the total amount of materialcaptured by the pre-separator and cyclone combined, for thepre-separator and cyclone, respectively. There appeared to bevery little difference in the sieved material captured by thepre-separator, due to baffle location, except for the 16.0-mm(5/8-in.) sieve, where the percentage of material decreased asthe baffle was moved away from the inlet (fig. 11). For

ÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖ

ÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖÖ

ÖÖÖ

ÖÖÖÖÖ

ÖÖÖÖÖ

ÖÖÖÖÖ

ÖÖÖÖ

ÖÖÖÖÖÖ

93%

94%

95%

96%

97%

17.8 18.8 19.8 20.8 21.8 22.8

Inlet Air Velocity (m s −1)

Ove

rall

Col

lect

ion

Effi

cien

cy

3500 3750 4000 4250 4500Inlet Air Velocity (fpm)

22.9 (9)

ÖÖ38.1 (15)53.3 (21)

Baffle Location cm (in.)

Figure 7. Over-all (pre-separator and cyclone) collection efficiencies at pre-separator inlet air velocities. Error bars represent two-sided 95% confi-dence limits for the means.

353Vol. 23(3): 347-355

0%

20%

40%

60%

80%

100%

2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5

Critical Velocity (m s−1)

Per

cent

of M

ater

ial R

emov

ed

400 500 600 700 800 900 1000 1100 1200 1300Critical Velocity (fpm)

BaffleCycloneNot Captured

Figure 8. Material captured by the pre-separator and by the cyclone, and the material escaping capture at pre-separator critical velocities.

Figure 9. Material captured by the baffle-type pre-separator and by thecyclone.

Figure 10. Material captured by the baffle-type pre-separator segregatedby sieve size.

material captured by the cyclone (fig. 12), the percentage ofmaterial on the 4.75-mm (#4), 2.0-mm (#10), 1.4-mm (#14),710-µm (#25), 180-µm (#80), and 75-µm (#200) sievesincreased as the baffle was moved away from the inlet. Itwould be expected that if a trend in the amount of material

Table 2. Results from sieve analysis with significant differences among pre-separator inlet air velocity for material

captured by the over-sized cyclone.

Inlet Velocitym s-1 (fpm)

Percentage Material per Sieve Size[a]

2.0 mm(#10)

1.4 mm(#14)

710 µm(#25)

< 75 µm(Pan)

18.3 (3600) 2.1 c 3.4 b 8.6 b 18.4 a

20.3 (4000) 2.5 b 3.9 a 9.7 a 17.7 ab

22.4 (4400) 2.8 a 4.0 a 10.2 a 16.7 b

Pvalue < 0.01 < 0.01 < 0.01 0.02[a] Means in the same column followed by the same letter are not

significantly different as determined by Ryan-Einot-Gabriel-Welch test based on range, REGWQ, (P ≤ 0.05).

collected by the pre-separator for a particular sieve withchanging baffle distance from the inlet was revealed, then theopposite trend in the amount of material collected by thecyclone for that sieve would follow, especially for largermaterial. Inexplicably, this was not seen in the analyses.Consistent with expectations of pre-separator and cycloneperformance, a comparison of figures 11 and 12 revealed thatthe pre-separator removed the majority of the materialranging from 180 µm #80 to 16.0-mm (3/8-in.) sieve, whilethe cyclone captured the majority of the material smaller than180 µm. One observation of note is that a large fraction of lintpassed through the pre-separator to the cyclone, which mayhave influenced performance.

CONCLUSIONA study was conducted to evaluate the effects of

pre-separator inlet air velocity, gin waste loading rate, andbaffle location on the collection efficiency of a baffle-typepre-separator followed by an over-sized 1D-3D cyclone.Results from the study indicated that the baffle-type pre-sep-arator’s collection efficiency ranged from 76.5% to 79.4%over the range of inlet air velocities [18.3 to 22.4 m s-1 (3600to 4400 fpm)] and was impacted to a greater extent by inlet

354 APPLIED ENGINEERING IN AGRICULTURE

ÒÒÒÒÒÒ

ÒÒ Ò

ÒÒÒÒ

ÒÒÒÒÒÒÒÒÒÒÒÒ

ÒÒÒÒÒÒÒÒÒÒÒÒÒÒ

ÒÒÒÒÒÒÒÒÒÒÒÒÒÒÒÒ

ÒÒÒÒÒÒÒÒÒÒÒÒ

ÒÒ Ò0%

5%

10%

15%

20%

25%

16.0 mm(5/8 in)

9.5 mm(3/8 in)

8.0 mm(5/16 in)

4.75 mm(#4)

2.0 mm(#10)

1.4 mm(#14)

710 µm

(#25)180 µm

(#80)75 µm

(#200)Pan

Sieve Size

Mat

eria

l Cap

ture

d

22.9 cm ( 9 in.)

Ò38.1 cm (15 in.)

53.3 cm (21 in.)

Baffle Location

Figure 11. Sieve analysis results for the material captured by the baffle-type pre-separator for each baffle location. Error bars represent two-sided 95%confidence limits for the means.

ÒÒÒÒ Ò Ò Ò Ò ÒÒ

ÒÒÒÒ

ÒÒÒÒÒÒ

ÒÒ

ÒÒ0%

5%

10%

15%

20%

25%

16.0 mm(5/8 in)

9.5 mm(3/8 in)

8.0 mm(5/16 in)

4.75 mm(#4)

2.0 mm(#10)

1.4 mm(#14)

710 µm(#25)

180 µm(#80)

75 µm(#200)

Pan

Sieve Size

Mat

eria

l Cap

ture

d

22.9 cm ( 9 in.)ÒÒ

38.1 cm (15 in.)

53.3 cm (21 in.)

Baffle Location

Figure 12. Sieve analysis results for the material captured by the over-sized cyclone, which followed the baffle-type pre-separator, for each baffle loca-tion. Error bars represent two-sided 95% confidence limits for the means.

velocity than loading rate. The pre-separator’s collectionefficiency increased as the baffle was moved closer to theinlet and decreased as the airflow rate was increased at agiven baffle location. The collection efficiency of theover-sized cyclone and over-all collection efficiency of thepre-separator followed by the cyclone were 78.8% and95.4%, respectively. There were no differences in either thecyclone collection efficiency or over-all efficiency for any ofthe treatments. Based on the sieve analysis, the pre-separatorremoved the majority of the material larger than 180 µm(#80 sieve), while the cyclone captured the majority of thefiner material. The pre-separator did allow a large portion oflint to pass through to the cyclone. The baffle-type pre-sepa-rator performed well at reducing the coarse material loadingrate entering the cyclone.

ACKNOWLEDGEMENTSThe partial support of this research by Cotton Incorpo-

rated is gratefully acknowledged.

REFERENCESArmijo, C. B., M. N. Gillum, S. E. Hughs, and P. J. Wakelyn. 1992.

Using cyclones effectively in cotton gins. Memphis, Tenn.: TheCotton Foundation.

Avant, R. V., Jr., C. B. Parnell, Jr., and J. W. Sorenson, Jr. 1976.Analysis of cyclone separator collection performance for grainsorghum dust. ASAE Paper No. 763543. St. Joseph, Mich.:ASAE.

355Vol. 23(3): 347-355

Baker, R. V., and V. L. Stedronsky. 1967. Gin trash collectionefficiency of small diameter cyclones. ARS No. 42-133.Washington, D.C.: USDA Agricultural Research Service.

Baker, R. V. M. N. Gillum, and S. E. Hughs. 1995. Pre-separatorsand cyclones for the collection of stripper cotton trash.Transactions of the ASAE 28(5): 1335-1342.

Columbus, E. P. 1994. A pre-separator for cyclones at cotton gins.In Proc. 1994 Beltwide Cotton Production Conferences,1741-1748. Memphis, Tenn.: National Cotton Council.

EC/R Incorporated. 1998. Stationary source control techniquesdocument for fine particulate matter. EPA Contract No.68-D-98-026. Research Triangle Park, N.C.: U.S.Environmental Protection Agency, Integrated Policy andStrategies Group.

Harrell, E. A., and V. P. Moore. 1962. Trash collecting systems atcotton gins. ARS No. 42-62. Washington, D.C.: USDAAgricultural Research Service.

Mihalski, K. D. 1995. The design of a pre-collector for cyclonecollectors. M.S. thesis. College Station, Tex.: Texas A&MUniversity, Department of Agricultural Engineering.

Mihalski, K., P. Kaspar, and C. B. Parnell, Jr. 1993. Design of pre-separators for cyclone collectors. In Proc. 1993 BeltwideCotton Production Conferences, 1561-1568. Memphis, Tenn.:National Cotton Council.

Parnell, Jr., C. B., and D. D. Davis. 1979. Predicted effects of theuse of new cyclone designs on agricultural processing particulateemissions. ASAE Paper No. 795102. St. Joseph, Mich.: ASAE.

SAS. 1999. SAS Online Doc. Ver. 8. Cary, N.C.: SAS Institute, Inc.Available at: v8doc.sas.com/sashtml/.

Wang, L., J. D. Wanjura, C. B. Parnell, Jr., B. W. Shaw, R. E. Lacey,S. C. Capareda, and M. D. Buser. 2004. Study of “baffle-typepre-separator plus cyclone” abatement systems for cotton gins.ASAE Paper No. 044017. St. Joseph, Mich.: ASAE.

356 APPLIED ENGINEERING IN AGRICULTURE