Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Download
Standard view
Full view
of .
Save to My Library
Look up keyword
Like this
0Activity
0 of .
Results for:
No results containing your search query
P. 1
Pilot and Industrial-Scale Experimental Investigation of Numerically Optimized Cyclones

Pilot and Industrial-Scale Experimental Investigation of Numerically Optimized Cyclones

Ratings: (0)|Views: 8 |Likes:
Published by Veronica Anescu
kjhvkkhcgkhx
kjhvkkhcgkhx

More info:

Published by: Veronica Anescu on Jun 29, 2013
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

06/29/2013

pdf

text

original

 
Pilot-and Industrial-ScaleExperimental Investigation of Numerically Optimized Cyclones
RomualdoL.Salcedo*
,†
and Ma
´
rioJ .Pinho
Departamento de Engenharia Quı 
´
mica, Faculdade de Engenharia da Universidade do Porto,Rua Dr. Roberto Frias s/ n, 4200-465 Porto Codex, Portugal, and QUIMlGAL, Quı 
´
mica de Portugal S.A.,Quinta da Industria 3860 Estarreja, Portugal 
A new geometry of reverse-flow gas cyclones obtained by numerical optimization was shown atthe laboratory scale to be significantly more efficient than other high-efficiency designs. However,it is usually recognized that experimental results obtained with laboratory-scale or samplingcyclones cannot be extrapolated to pilot or industrial scales. The present paper confirms, atthese larger scales, the significantly larger collection efficiencies obtained with the numericallyoptimized design compared to a competing high-efficiency design available on the marketplacefor the capture of fine sulfanilic acid (median volume diameter of 17
µ
m)at a Portuguese chemicalmanufacturer. A partial recirculation system within a collector-first arrangement further reducesemissions without an appreciable increase in pressure drop. The experimentally verifiedefficiencies at the industrial scale varied between 99.58 and 99.64% for sulfanilic acid withpressure drops around 2.5 kPa. The numerically optimized cyclones, when coupled with a partialrecirculation system, extend the applicability of these simple devices to the fine particle collectionthat is typical of more expensive devices, such as venturis and online pulse jet bag filters.
Introduction
Cyclones are gas
-
solid separation devices character-ized by low investment and operating costs that can beused on heavily loaded gases or at high temperature andpressure. They are employed in many industries for twocomplementary purposes: air pollution control by re-duction of atmospheric emissions to maximum admis-sible levels and raw material or product recovery. Thedevelopment of cyclones with collection efficienciessignificantly above those obtainable with currentlyavailable geometries, especially for particle diametersbelow 5
µ
m, could have a strong impact on the chemicalprocess industries.
1
Many industries could then usethese low investment and operating cost dedusters withthe required efficiency to avoid having to resort toequipment or with much higher capital, operating, ormaintenance costs, such as venturis and bag filters.Also, in some processes at high temperature (up to 1300K) and pressure (up to 100 bar), cyclones are currentlythe only dedusters available.In a typical reverse-flow cyclone (Figure 1), which isthe most common type used, the dirty gas enters thecyclone tangentially at the top, describes a descendingouter vortex, inverts the direction of motion, andascends by an inner vortex exiting at the cyclone topthrough the vortex finder. The heavier (larger or denser)particles are swept to the cyclone wall by centrifugalforces that oppose a radial drag and are carried towardthe cyclone bottom by the descending outer vortex. Thefiner particles exit at the top with the gas, together withcoarser particles that might have been reentrained andswept by the ascending inner vortex. Reentrainment canoccur as a result of an entry velocity that is excessivefor the cyclone geometry and particle density.
2
-
4
The present work shows, at the pilot scale, thebehavior of the numerically optimized RS
VHE cyclonedesign, extending previous work done at the laboratoryscale.
5,6
It also shows, at both the pilot and industrialscales, the performance of the RS
VHE cyclone coupledwith a partial recirculation system. A comparison is alsomade with the performance of a high-efficiency cycloneavailable on the marketplace and with an online pulsejet bag filter.The results show that the proposed RS
VHE designtypically reduces the emissions by 50%in comparisonwith the tested HE cyclone at comparable pressuredrops, in agreement with the laboratory-scale data
* To whom all correspondence should be sent. Fax:
+
351225081440. E-mail: rsalcedo@fe.up.pt.
Faculdade de Engenharia da Universidade do Porto
QUIMlGAL.
Figure1.
Reverse-flow cyclone.
145
Ind. Eng. Chem. Res.
2003,
42,
145
-
154
10.1021/ie020195e CCC: $25.00 © 2003 American Chemical SocietyPublished on Web 11/19/2002
 
comparing the RS
VHE and Stairmand HE designs.
6
The recirculation further reduces emissions by another50%, putting this system at a level that is characteristicof high-performance venturis and online pulse jet bagfilters.
Numerical Optimization of Reverse-flowCyclones
The scientific literature is scarce on the subject ofcyclone optimization most probably for two reasons: upto recently, no single theory could predict with reason-able accuracy the behavior of an arbitrary-geometrycyclone under different operating conditions,
5,7
-
10
andthe possibility exists for commercial spin-offs derivedfrom innovative geometries.
11
Also, it is highly unlikelythat the optimum design can be found by empiricaltesting, as too many design parameters are involved.Thus, the problem of the optimization of reverse-flow cyclones has mostly been tackled on a trial-and-errorbasis.
1,12
-
14
To our knowledge, only a few works are directlyrelated to cyclone optimization based on a simulationmodel.
6,15
-
19
Dirgo and Leith
15
have used the Leith andLicht
20
collection model to predict cyclone performance.Pilot-scale tests with the optimized geometries, however,have failed to produce any significant improvements.Iozia and Leith
16
and Ramachandran and Leith
17
haveused the logistic empirical collection efficiency model ofIozia and Leith
8
for simulation purposes. However,these authors have not performed a full optimization,as only five degrees of freedom (
e
,
,
,
, and
) wereemployed, whereas the other three cyclone designvariables (
,
, and
b
) were held fixed.In previous papers,
6,18
it was demonstrated that it ispossible to design significantly improved reverse-flow cyclones by solving adequate numerical optimizationproblems. These equations concern the maximization ofsome profit function, viz., global efficiency, subject tothe model equations (equality constraints) and inequal-ity constraints (geometry, pressure drop, and saltationvelocity). The Mothes and Lo¨ffler
7
theory was used asa modeling environment, coupled with an empiricalestimate of the particles’ turbulent dispersion coef-ficient.
5
This estimate predicts grade efficiency curvesthat agree reasonably well with the experimental datafrom various authors and is given bywhere
Pe 
p
)
r
p
/
r
is the radial particle Peclet number,
r
is the radial dispersion coefficient, and
R
p
)F
p
r
/
 µ
is the radial particle Reynolds number. The optimizationwas performed using a stochastic adaptive randomsearch global optimizer.
21
Ravi et al.
19
also followed a similar approach, usingmultiobjective optimization and genetic algorithms butconsidering a constant value for the particles’ turbulentdispersion coefficient.The main result of the numerical optimization wasthe geometry RS
VHE, which is significantly differentfrom other high-efficiency geometries available in theliterature,
6
including the optimum cyclone designsobtained by Iozia and Leith
16
and by Ravi et al.
19
Thisgeometry is detailed in European patent applicationEP099670006, and its superior behavior at the labora-tory scale, in comparison with a Stairmand HE designfor the capture of fine Ca(OH)
2
can be found elsewhere.
6
For example, a reduction in emissions of 53% wasachieved on average, and the Moore and MacFar-land
22,23
dimensionless performance index of theRS
VHE design, as compared with that from fivedifferent designs,
14
showed lower (better) values.
Recirculation Systems
One of the methods of increasing the collectionefficiency of cyclone dedusters is to promote somerecirculation of the exhaust gases. Recirculation systemscomposed of a dust concentrator, which can be a uniflow or straight-through cyclone (concentrator), coupled toa reverse-flow cyclone (collector) are described in theliterature.
2,24,25
These systems, similar to that shownin Figure 2, can, under some circunstances, shocollection efficiencies well above those of reverse-flow cyclones. Their collection efficiency is given bywhere
η
c
and
η
s
are, respectively, the cyclone (collector)and concentrator efficiencies, both computed within theloop. Thus, these systems are more efficient than singlecyclones (
η
>
η
c
) of comparable geometry and sizewhenever
η
s
>
1/(2
-
η
c
), viz., the concentrator musthave a collection efficiency larger than that of thecollector, both measured within the loop. An efficientsystem with a concentrator-first arrangement is de-scribed in European patent EP0430647.However, it is possible to alter the system configura-tion to a collector-first sequence (Figure 3), where thesystem will always have a collection efficiency abovethat of the single reverse-flow cyclone operating withinthe loop.
3
A system composed of a reverse-flow collectorfollowed by a straight-through concentrator is describedin detail in International patent application WO0141934,with an efficiency given byBecause both eqs 2 and 3 share the same denominatorand the numerator in eq 3 is larger, the collectionefficiency of this system is always better. Also, it willalways be better than that of the single reverse-flo
Pe 
p
)
0.0342
R
p1.263
(1)
Figure2.
Concentrator-first recirculation system.
η
)
η
c
η
s
1
-
η
s
+
η
c
η
s
(2)
η
)
η
c
1
-
η
s
+
η
c
η
s
(3)
146
Ind. Eng. Chem. Res., Vol. 42, No. 1, 2003
 
cyclone operating within the loop, because the necessarycondition
η
s
(
η
c
-
1)
<
0 is always verified. To be fair,comparing the systems in Figures 2 and 3 makes senseonly if they have similar geometries and sizes, whichmight not be the case because, in the proposed system(collector-first arrangement, Figure 3), the entire flow must pass through both the collector and concentrator,whereas in the system depicted in Figure 2 (concentra-tor-first arrangement), only the concentrator receivesthe entire flow, which might allow the collector, forcomparable pressure losses, to be somewhat smaller andthus more efficient.It was recently verified
26,39
that reverse-flow cyclonesfitted with a post-cyclone (PoC) at the exit of the vortexfinder can increase the collection of fines, but a signifi-cant reduction in emissions (about 33%)only occurs witha conical finned PoC at the expense of a large increasein pressure drop (about 1.2 kPa). However, when thebleed flow from the PoC is recycled back to the cyclone,significant collection improvements for submicrometerparticle diameters occur with an increase of only 34%in the pressure drop above that of the cyclone. In thiscase, a reduction in emissions of about 26%has beenreported.In this work, the performance of a recirculationsystem with a collector-first arrangement is reported,with the collector being a single RS
VHE of 0.46-minside diameter for the pilot-scale experiments and aset of 12 RS
VHEs of 0.50-m inside diameter in aparallel arrangement for the full-scale experiments. Thepilot-scale concentrator has the same diameter as thepilot-scale RS
VHE cyclone (purely for convenience),whereas the two industrial-scale concentrators werelarger to accommodate flow from six cyclones each, fora maximum increase in pressure drop of about 30%overthat of the reverse-flow cyclones.
Experimental Section ResultsPilot-ScaleCyclones,RecirculationSystem,andOnlinePulseJ etBagFilter.
At the laboratory scale,the 0.02-m-inside-diameter RS
VHE design showed, onaverage, a 53%reduction in emissions of fine Ca(OH)
2
(log-normal distribution, mean mass diameter of 1.37
 µ
m, geometric standard deviation of 2.23) as comparedto a similar-sized Stairmand HE design at comparablepressure losses.
6
When fitted with a collector-firstrecirculation system using a small venturi to promoterecirculation (Figure 3), the emissions were furtherreduced by 57%, on average.
27
Thus, the recirculationsystem with the RS
VHE cyclone reduced emissionsby almost 80%as compared to a single Stairmand HEdesign. A 50% reduction in emissions due to therecirculation has also been obtained at the laboratoryscale with CaCO
3
(median mass diameter of 1.96
µ
m),again using an RS
VHE collector. These results weresufficiently encouraging to extend the RS
VHE andrecirculation systems to pilot- and eventually to full-scale testing.The pilot-scale test rig was installed at the sulfanilicacid production facility of a Portuguese manufacturerof benzene-derived organic chemicals. Basically, a fluid-ized-bed dryer conveys sulfanilic acid dust to a 0.80-m-diameter high-loadings process cyclone that returns afraction to the dryer and conveys the finer fraction to apulse jet bag filter at a flow rate of about 10 000 m
3
/h,a moisture content of 4%(molar basis), a temperatureof 350 K, and concentrations up to 20 g/m
3
. By installingthe pilot rig in parallel with the bag filter, a smallfraction of the flow (about 10%) could be diverted by aninduced fan to a single RS
VHE cyclone, fitted withan optional recirculation straight-through cyclone andcorresponding induced recirculation fan. Both fans wereequipped with variable-speed motors to allow the re-circulation and main flow rates to be varied. Figure 4shows the pilot-scale system, which was designed tohandle up to 1000 m
3
/h of process gas and 500 m
3
/h ofrecirculation gas.The sampling locations (inlet, outlet, and recirculationpipes) for the pilot-scale cyclones and for the filter wereestablished, and the velocity was measured by completepitot traverses using a Testo 400 unit. The measure-ments revealed fairly flat distributions at the inlet andoutlet and a much more irregular distribution at therecirculation pipe, as can be seen in Figure 5a,b, becauseof constraints on the sampling location (directly up-stream from the recirculation fan). Resampling with apitot tube at the inlet and outlet for the same operatingconditions showed consistent results (all within
(
5%),but resampling at the recirculation pipe showed flow rates for the same operating conditions varying by asmuch as
(
22%. Thus, data derived from these measure-ments (recirculation flow rate, concentrator collectionefficiency, and particle size distributions of recirculateddust) should be viewed with caution.
Figure3.
Proposed collector-first recirculation system.
Figure4.
Pilot-scale system (RS
VHE cyclone and concentrator).
Ind. Eng. Chem. Res., Vol. 42, No. 1, 2003
147

You're Reading a Free Preview

Download
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->