United States
Environmental Protection
Agency
Industrial Environmental Research^
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-81-148 May 1982
Project Summary
• ^v
Flux Force/Condensation
Scrubbing for Collecting Fine
Jtarticulate from Iron
Cupolas
qhard. D- Chrhielewski and Seymour Calvert
A flux force/condensation (F/C)
scrubbing system for controlling
paniculate emissions from an iron and
steel melting cupola demonstrated its
industrial feasibility during a 6-month
test period.
A particle collection efficiency of 84
percent was required to meet the
emission limit for a melting rate which
ranged around 12,500 kg/hr. The
demonstration plant was designed to
enable experimental determination of
operating characteristics when
attaining efficiencies around the
emission limit. Particle size
distributions, flow rates, pressure
drops, temperatures, and other
parameters were measured and
utilized for performance analysis and
comparison with theoretical
predictions.
After an initial period during which
equipment and operating methods
were refined, the system performed
well and without significant difficulty.
An optimized F/C scrubbing system
for this application would require only
about 65 percent of the predicted
power requirement for a conventional
high energy system.
This Project Summary was devel-
oped by EPA's Industrial Environ-
mental Research Laboratory,
Research Triangle Park, NC. to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
Flux force/condensation (F/C) scrub-
bing has been developed by Air Pollu-
tion Technology, Inc. under
EPA-sponsorship for the past several
years (see Table 1). The object of F/C
scrubbing is to reduce the power
requirement for collection of fine parti-
cles as compared to conventional high
energy scrubbers. The improvement in
scrubber performance due to condensa-
tion effects is most apparent in the sub-
micron size range where the
mechanism of inertial impaction is diffi-
cult to apply economically. Water vapor
from a hot saturated gas can be
condensed by contacting the gas with
cold liquid.
Three condensation effects are util-
ized in F/C scrubbing. The suspended
particles in the gas act as condensation
nuclei, resulting in growth of particle
mass due to the condensation of water
vapor. The transfer of water vapor
toward the cold liquid results in diffusio-
phoresis, causing particles to collect by
movement of the condensing vapor
toward the cold surface. Simultane-
ously, tr\e temperature gradient estab-
lished 'in the condenser creates a
thermal force which leads to collection
-------
by thermophoresis. The three effects
enhance the particle collection effi-
ciency of the F/C scrubber system.
Background
Flux force effects on particles have
been known for many years, and the
background has been reviewed and dis-
cussed in depth by many authors. The
research and development work sup-
ported by the EPA was aimed at first
assessing the potential of F/C scrub-
bing, then developing rational means
for exploiting it, and then demonstrating
the technology.
The feasibility of F/C scrubbing and a
rational design basis were established
in the first four studies listed in Table 1.
The next step was a pilot-scale demon-
stration, carried out on a secondary
metal recovery furnace. Control of the
entire source effluent, a maximum flow
rate of 200 AmVmin (7,000 acfm) at
temperatures up to 800°C, required
both particle collection and acid gas
absorption. Demonstration plant perfor-
mance was consistent with the preced-
ing laboratory, bench scale, and
pilot-plant studies.
Earlier studies (see Table 1) included
theoretical development of design
equations for F/C scrubbing and the
development of a computerized
mathematical model to predict particle
collection in a sieve plate F/C scrubber.
The predictions compared well with
experimental results, but the computer-
ized model was cumbersome. Inthefirst
demonstration study, a revised
mathematical model was developed
from the same basic relationships, and
the design procedures were greatly sim-
plified for prediction of F/C scrubber
performance.
This report gives results of a demon-
stration project to test further the design
methods and economics of F/C scrub-
bing on a large scale in an industrial
environment. An F/C system for the
total exhaust gas stream form a 12,500
kg/hr iron melting cupola was
designed, built, and tested. The operat-
ing experience included summer, fall,
and winter in a demanding environment
and resulted in a good test of equipment
design and materials selection.
System Design
As a replacement system in an oper-
ating plant, the F/C scrubber had to fit in
the space around the existing scrubber
system. Once the new apparatus was in
place, it had to be connected and the old
Table 1. F/C Scrubbing Studies by A. P. T. for EPA
Title
Subject
Wet Scrubber System Study — "Scrubber
Handbook" (Vol. I). Report EPA-R2-72-
118a (NTIS PB 213-016). August 1972
Feasibility of Flux Force/Condensation
Scrubbing for Fine Paniculate
Collection, Report EPA-650/2-73-036
(NTIS PB 227-307), October 1973
Study of Flux Force/Condensation
Scrubbing of Fine Particles, Report
EPA-600/2-75-018 (NTIS PB
249-297), August 1975
Study of Horizontal-Spray Flux
Force/Condensation Scrubber,
Report EPA - GOO/2-76-200 (NTIS
PB 262-669). July 1976
Fine Particle Collection by Flux
Force/Condensation Scrubber:
Pilot Demonstration. Report
EPA-600/2-77-238 (NTIS
PB 277-075), December 1977
F/C scrubbing principles and potential.
Theoretical analysis, mathematical
modeling, computer prediction,
experimental data from bench-
scale F/C scrubbers.
Laboratory pilot-plant study of
F/C scrubbing in sieve plate and
spray apparatus.
Laboratory pilot-plant study of
F/C srubbing in horizontal-
spray apparatus.
Pilot demonstration of F/C
scrubbing with a sieve plate
column on a secondary metal
recovery furnace, analysis of
results, and further development
of mathematical modeling.
system disconnected with the minimum
loss of productive operation of the cup-
ola. This and the space constraints
imposed by the foundry operations had
the major influence on F/C system
layout.
Conservation of project cost required
that the maximum amount of existing
auxiliary equipment be used. Thus, the
existing afterburner, scrubber pump
and sump, fan, and stack were used.
The F/C system was built to operate
in two basic configurations, induced
draft (I.D.) and forced draft (F.D.). The
F.D. mode has the potential advantage
of reducing the fan power requirement
by about 6 percent, with the possible
disadvantage of entrained liquid enter-
ing the fan.
The system flow diagram in Figure 1
is for the I.D. configuration. Only the
ducting was modified to switch from
F.D. to I.D.; the remaining components
remained unchanged.
Process Description
The cupola melts scrap iron and steel
using heat generated by the combustion
of coke. Raw materials (when the pro-
gram started) were scrap structural
steel, engine blocks, metal stampings
and turnings, cast iron piping and fit-
tings, and other miscellaneous metal
parts. During the construction period,
the cupola was converted to water
cooled operation and the metal charge
was changed to small steel scrap.
The scrap and coke are charged
together in 1,800 kg batches with an
iron-to-coke mass ratio of 6. Small
amounts of limestone, 20-30 kg, are
also charged to form slag. The nominal
rate is 12,500 kg/hr and the normal
tuyere air flow rate is 3.5 DNmVs.
The exhaust gas from the bed of mate-
rial in the cupola is composed typically
of 14 percent CO, 13 percent COa, and
73 percent Nz. Additional combustion
air, entering the charging door 2.3
DNmVs, allows for complete conver-
sion of CO to CO2. An ignition burner
ensures that a flame source is available
for combustion of the CO to COa. (It was
not needed or used during initial tests.)
Combustion is completed in a brick-
lined afterburner. The exhaust gases
should leave the afterburner tank at a
temperature of 800-1,000°c; The
exhaust gas flow rate at this point is
about 6.6 DNmVs. After the cupola
modification and the change to steel
scrap charge, the cupola gas would not
burn consistently, so the gas tempera-
ture was lower and more variable than
in preliminary tests.
-------
Water
Figure 1. F/C demonstration plant flow diagram.
The hot gas leaves the afterburner
and is cooled in the saturator by water
sprays to a temperature approaching
adiabatic saturation. Typically, the satu-
ration temperature was 71 — 77°C. The
maximum water spray rate in the satu-
rator is 4.5 l/s (75 gpm).
The condenser and cooling tower sys-
tem has a design heat rejection rate of
9.6 x 106 J/s (33 x 106 Btu/hr). The
water flow rate can be varied up to 69
l/s (1,100 gpm). The cold gas tempera-
ture leaving the counter-current
packed-bed condenser is about 55° C at
design flow rates.
The exhaust gases leave the con-
denser and enter either the exhaust fan
or the scrubber, depending on the duct-
ing. The primary particle collection
device in the system is a gas-atomized
scrubber with a variable-area throat, an
integral water sump, and an entrain-
ment separator.
The
plant
blade
at -96
The
(1.10C
The
induced draft.
fan used in the demonstration
was rebuilt from that used in the
previous system and is a straight radial
design. The flow rate of the fan is
approximately 14 AmVs (30,000 cfm)
cm W.C. (-38 in. W.C.).
condenser water comes from the
coolinj tower, designed to cool 69 l/s
gpm) of water from 66 to 38°C
with ambient air. This is accomplished
by eveporative cooling.
settler was designed to remove
particles collected by the scrubbing liq-
uid. A 1.3-l/s (20-gpm) stream of sump
water is clarified by the settler. The
thickened sludge (10-30 percent by wt)
is disposed of as landfill. The clarified
water (100 ppm solids) is returned to the
sump.
Performance Testing
In the course of system start-up, a
significant problem with the original
exhaust fan caused excessive power
consumption and tripping of the electri-
cal overload device. This condition con-
tinued until a new A.P.T.-designed fan
wheel was installed. Some problems
due to foaming and entrainment carry-
over had to be resolved, but system
operation was otherwise quite good
during the testing. No production time
was lost by the foundry due to problems
associated with the operation of the F/C
scrubber system.
The major test program variables
were the condenser liquid flow rate and
scrubber pressure drop. The condenser
liquid flow rate affected the amount of
cooling in the condenser and, therefore,
the condensation ratio. The scrubber
pressure drop could be varied by chang-
ing the scrubber-throat flow area and
the scrubber liquid-to-gas ratio.
Samples were usually taken at the
condenser inlet and the scrubber outlet.
-------
For some runs, samples were also taken
at the scrubber inlet. Sampling at three
locations allowed the condenser and
scrubber performance to be evaluated
separately.
The particle size distribution and
mass concentration were measured
with University of Washington cascade
impactors equipped with precutters.
Several modified EPA Method 5 total
filter runs were made to determine total
mass concentration for compliance
tests and to confirm impactor mass
loadings.
From measurements of cumulative
mass concentration as a function of par-
ticle size, particle penetration was com-
puted as a function of particle size. The
data was analyzed using a computer
program which calculated impactor cut
points and cumulative mass concentra-
tions based on weight gain per stage,
impactor flow rate, and impactor cali-
bration data.
Performance Model
The system performance model can
be used to predict the emission from a
F/C scrubber system installed on a pol-
lutant source with a known particle con-
centration and size distribution. The
model allows independent variation of
condensation ratio and scrubber pres-
sure drop.
Without going into details of the
mathematical model, the basic concepts
and outline of the approach are dis-
cussed. The gas leaving the source is
hot and has a water vapor content
which depends on the source process.
The first step is to saturate the gas by
quenching it with water.
In the condenser there is diffusiopho-
retic and thermophoretic deposition,
some inertia! impaction, and (perhaps)
Brownian diffusion. The particles in the
gas leaving the condenser will have
grown in mass due to the layer of water
they carry.
Subsequent scrubbing of the gas will
result in more particle collection by iner-
tia I impaction. There may be additional
condensation, depending on water and
gas temperatures, and its effects can be
accounted for.
The mathematical model for the F/C
demonstration plant accounts for the
mechanisms and assumptions outlined
below.
Saturator
1. Gas is humidified and cooled to
adiabatic saturation temperature.
2. No condensation occurs.
3, Particle collection in the saturator
is negligible.
Condenser
1. Particles are collected by impac-
tion in packed column.
2. Condensation occurs, causing
growth of particles.
3. Collection occurs in condenser
due to diffusiophoresis.
Scrubber
1. Grown particles are collected by
impaction in scrubber.
2. Negligible condensation occurs.
Experimental Results
Experimental performance measure-
ments, made during the last 2 months of
the testing, represent the best level
attained after system adjustment was
nearly completed. The "36 series" of
runs were in the I.D. mode.
Runs 36/12 and following were
made after a foaming problem was
resolved and the scrubber liquid piping
revised to better purge the internal
sump. The "37 series", of runs were
made in the F.D. mode, except for 37/21
and following, which were in the I.D.
mode.
The hot gas temperature leaving the
afterburner was, for most of the runs,
significantly lowerthan anticipated. The
design-basis hot gas temperature was
1,000°C, as measured in preliminary
tests. The maximum temperature mea-
sured was about 850°C. The low gas
temperature reduced the attainable
condensation ratio: The maximum was
about 0.2 g H2O/g D.G. (dry gas). It had
been anticipated that a condensation
ratio of 0.3 g H2O/g D.G. could be
achieved with the higher gas
temperature.
Particle Mass and Size
The average particle mass concentra-
tion at the condenser inlet was 2.4
g/DNm3 (33 percent higher than the
design basis mass concentration of 1.8
g/DNm3).
Condenser inlet mass concentration
measurements varied significantly from
run to run, reflecting the problem of
having to sample during various periods
in the charging cycle. Inlet sampling
times had to be much shorter than
outlets because the inlet concentrations
were so high. Consequently, one inlet
sample would not be representative of
the period over which the outlet sample
was taken. Computations of scrubber
penetrations, on the basis of inlet and
outlet samples for each run, varied
greatly and randomly. To cope with this
variation, an average condenser inlet
size distribution was used in determin-
ing grade penetration curves.
The average cumulative mass inlet
concentration distribution was used in
combination with the cumulative mass
outlet concentration distribution for
each run to compute the grade penetra-
tion relationship for that run. Figures 2
50
40
30
S? 20
1 »
q' = 0.08
\
--— Predicted
— Measured
0.3 0.4 0.5
1.0
2.0 3.0 4.0 5.0
10.O
Figure 2. Particle penetration. Run 36/10.
-------
through 5 present experimental and
predicted grade penetration curves for
runs 36/10 through 36/13.
The grade penetration curves show
that predictions and experiment com-
pare well in some cases but not in oth-
ers. Experimental penetrations at
around 1/umA tend to be higher than
predictions, at both F/C and non-F/C
conditions.
The performance in the F.D. mode
was erratic and not improved as much
by condensation as it was in the I.D.
mode. Some of the condensed water
may have evaporated in the fan so that
the particles entering the scrubber were
smaller than those entering the fan.
This question could not be resolved dur-
ing the test.
The variation of computed penetra-
tions appears to be caused by the uncer-
tainty of the inlet size distribution. The
agreement between predictions and
experiment for many of the series-36
runs is considered good for field experi-
mental conditions, where the equip-
ment is being modified and the source
process is cyclic.
Economic Analysis
An optimum F/C scrubber system
design for this application would not be
significantly different from the demon-
stration plant, and the process design
would remain substantially unchanged.
Major modifications recommended for
an optimum system involve equipment
redesign and specifications to reduce
capital expense and to ensure that the
gas temperature would be high.
Significant cost savings can be
achieved by using a single combined
condenser/scrubber vessel. Experi-
ments with the demonstration plant
showed that the vessel diameter and
the packing depth can be smaller than
used.
Because of space requirements, the
cooling tower was on a platform, above
the saturator and sump. Consequently,
the most compact tower was chosen,
not the most economical or energy effi-
cient. A woodfilled cooling tower would
have been the best choice if space had
not been at a premium. Such a tower
would require only 22 kW(30 hp), com-
pared to 44 kW for the tower used on the
demonstration plant. The wood packing
may have some advantage over epoxy-
coated steel for corrosion resistance. A
second circulating water pump would
be required for the* system, but the total
cost would be less because an expen-
sive structural steel platform would not
be needed.
50
4O
30
S? 20
10
Q>
q' = 0.1
"-—-Predicted
Measured
I I
I I
0.3 0.4 0.5
1.0
2.0
3.0 4.0 5.0
10.0
1 I
Figure 3. Particle penetration. Run 36/10.
50
40
30
^ 20
q' = 0.1
--- Predicted
Measured
0.3 0.4 0.5 1.0 2.0 3.O 4.0 5.0
dpa,umA
Figure 4. Paricle penetration. Run 36/12.
10.0
The saturator, designed to fit into
space available at the foundry, would be
smaller in an optimum system. The
settler used on the demonstration sys-
tem provided acceptable performance;
however, the condenser, scrubber, and
sump all were, subject to deposition of
sandy material which required frequent
removal. Sloped bottoms should be
used in all of these vessels.
It is also recommended that a moving-
belt filter be used instead of the gravity
settler to remove this sandy material.
This would cost less than the settler
used on the demonstration plant.
To evaluate the potential benefits of
an F/C scrubber system (compared to a
conventional one), the costs for a ven-
turi scrubber system were estimated.
The process conditions and emission
criteria were the same for both systems.
Capital Cost Estimates
The capital cost of an optimum F/C
system would consist of direct costs for
delivered major equipment and other
-------
items (e.g., installation, piping, and
electrical), as well as indirect cost (e.g.,
engineering, construction overhead,
and contingencies). Table 2 shows the
estimated direct and indirect costs for
the F/C system and the high energy
scrubber. All costs were adjusted to
December 1979, using the Marshall
and Stevens cost index.
Operating Costs
The operating cost of the air pollution
control system consists of the annual
cost of the utilities (power and water),
raw material, and maintenance. Table 3
shows the power requirement for both
the F/C scrubber system and a conven-
tional scrubber. The F/C scrubber sys-
tem would require only 65 percent of
the powerof the conventional scrubber.
The exhaust fan for the conventional
scrubber would require more than twice
the power of that for the F/C system.
The estimated total operating costs
are summarized in Table 4. The total
operating cost for the F/C scrubber sys-
tem was $131,230, compared to
$150,300 for the conventional
scrubber. The annual operating cost of
the F/C scrubber system would be
about $19,100 less than the cost for a
conventional scrubber for the condi-
tions stated. If the operating time were
reduced below 4,000 hr/yr, the F/C
system would be less favorable; how-
ever, longer operating time would make
it more favorable. Power cost will also
have an obvious effect on economics.
Conclusions
The following conclusions can be
drawn from this study.
1. The performance data were in
general accord, with predictions
based on design methods deve-
loped by APT. on previous EPA
contracts.
2. The F/C scrubber system was
capable of complying with all
applicable air pollution regula-
tions for the site.
3. The F/C scrubber system needs
only about 65 percent of the pre-
dicted power requirement of a
conventional high energy
scrubber to achieve the same
performance.
4. The particle mass concentration
and size distribution emitted dur-
ing the cupola charging cycle var-
ied greatly. Continuous or
semi-continuous cupola charging
with a conveyor system would
50
40
30
* 20
^
1 10
§
Q.
5
2
/
I
q' = 0.16
Predicted
^—— Measured
I I I I I
0.3 0.4 0.5
1.O
2.0 3.0 4.0 5.0
10.0
Figure 5. Particle penetration. Run 36/13.
Table 2. Direct and Indirect Costs
F/C System
Conventional
Scrubber
Item
Ratio
Cost, $
Ratio
Cost. $
Direct
Equipment
Installation
Instruments
Piping and Ducting
Electrical
Site Preparation
Total Direct Costs
Indirect
Engineering
Construction Overhead
Contractor's Fee
Contingency
Total Indirect Costs
Fixed Capital Investment
1.00
0.40
0.10
0.40
O.JO
0.05
2.05
0.40
O.45
0.10
0.40
1.35
3.40
$102.380
40,952
10,238
40.952
10.238
5,119
$209.879
40.952
46.071
10,238
40,952
$138,213
$348,092
1.00
0.40
0.10
0.40
0.10
0.05
2.05
0.40
0.45
0.10
0.40
7.35
3.40
$73,765
29,506
7,377
29,506
7,377
3,687
$151,218
29.506
33,193
7,377
29,506
$99.582
$250,800
Table 3. Summary of Power Requirements
Power, kW (hp>
Item
Exhaust Fan
Saturator Pump
Scrubber Pump
Cooling Water Pumps
Cooling Tower Fan
Total Power Required
F/C Scrubber
156(210)
10 (13)
2(2)
33 (44)
22 (30)
223 (299 )
Conventional Scrubber
332 (445)
10 (13)
3(4)
345 (462)
-------
probably reduce the variation in
emissions and gas temperature.
5. The solid/liquid separation pre-
sented some maintenance prob-
lems. Additional attention should
be given to equipment design,
both to prevent settling in the
s-crubber equipment and to
improve solids separation in the
water treatment system.
6. Corrosion of the cooling tower
packing was severe. Acids must
be neutralized to prevent signifi-
cant deterioration of steel compo-
nents in the system. Maintenance
of the proper circulating water pH
is also important if stainless steel
materials are used, because of the
chloride content present.
7. Operation over 6 months showed
that the F/C system can perform
well and without significant diffi-
culty under the demanding condi-
tions of a ferrous foundry. During
the entire test period, the F/C
scrubbing system did not cause
any production stoppages.
8. F/C scrubbing becomes more
economically advantageous than
conventional scrubbing as either
power cost or plant operating time
increases.
Recommendations
Demonstration plant experience has
led to several recommendations on
equipment changes; most of them are
discussed above under "Economic
Analysis." Essentially, the recom-
mendations:
1. Minimize costs of equipment re-
quired to obtain F/C effects.
2. Maximize cupola flue gas
temperature.
3. Provide adequate solids removal
capacity for the scrubbing-liquid
recycle stream.
Table 4. Summary of Estimated Annual Operating Costs
F/C Scrubber Conventional
Item Unit Cost System Scrubber
Capital Cost
Maintenance
Operating Labor
Power Costb
Water Use
Raw Materials
@ 10%ofF.C.I.a
Materials @ 3% of F. C.I.
Labor @ 8 hr/wk shift.
$9/hr
@ 12 hr/wk shift. $9/hr
$0.065/kWh
$0.035/ 1.000 liters
Soda ash $0.265 /kg;
flocculant $ 1. IS/ liter
$ 34,810
10.440
7,200
10.800
57.980
2.000
8,000
$ 25.080
7,520
7.200
10.800
89,700
2,000
8.000
Total Operating Cost
$131,230
$150.300
a Fixed capital investment.
*Estimated for 4,000 hr/yr.
R. D. Chmielewski and S. Calvert are with Air Pollution Technology, Inc., San
Diego. CA92117.
Dale L. Harmon is the EPA Project Officer (see below).
The complete report, entitled "Flux Force/Condensation Scrubbing for Collect-
ing Fine Paniculate from Iron Melting Cupolas," (Order No. PB 82-196 866;
Cost: $13.50, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
Postage and
United States Center for Environmental Research Fees Paid
Environmental Protection Information Environmental
Agency Cincinnati OH 45268 Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300 ,,
PS 0000329
U S ENVIR PROTtCTlON AGENCY
REGION 5 LIBRAHY
230 S DEARBORN STREET
CHICAGO IL 60604
------- |