EFFECT OF FEED CHARACTERISTICS ON THE PERFORMANCE
OF EPA'S MOBILE INCINERATION SYSTEM
by: James P. Stumbar
Robert H. Sawyer
Gopal D. Gupta
Foster Wheeler Enviresponse, Inc.
Edison, NO 08837
Joyce M. Perdek
Frank J. Freestone
Releases Control Branch, USEPA
Edison, NJ 08837
ABSTRACT
During the past four years, the EPA Mobile Incineration System (MIS),
has processed a wide variety of feeds. Besides the incinerating the
hazardous materials for which the MIS was designed, the unit has also
incinerated contaminated debris including wood pallets, steel and fiber
drums, and plastics. This paper identifies significant physical and
chemical characteristics of various feed materials and their relationship
to MIS performance. The paper also correlates the effect of these feed
characteristics on specific MIS components. Corrective actions taken to
mitigate several problem characteristics are presented. The operating
experience with the MIS has provided valuable data on the limits of
incineration capacity as well as reliability of the unit in relation to
various feed stocks. This information is also discussed. The information
contained in this paper is directly applicable to field use of mobile and
transportable incinerators at Superfund and other industrial cleanup sites.
-------
INTRODUCTION
Under the sponsorship of the Office of Research and Development of the
U.S. Environmental Protection Agency (EPA), the Mobile Incineration System
(MIS) was designed and constructed to demonstrate high-temperature
incineration of hazardous wastes (1). The system essentially consisted of
a refractory-lined rotary kiln, a secondary combustion chamber (SCC), and
an air pollution control system. These three components are mounted on
three separate heavy-duty semi-trailers. Monitoring equipment is carried
by a fourth trailer.
The MIS was rigorously tested in Edison, New Jersey, during 1982 and
1983 with PCB-contaminated and other chlorinated organic liquids (2). The
system was transported to the Denney Farm site in McDowell, Missouri, in
December 1984 for a dioxin trial burn and field demonstration (3,4). A
total of 900,000 kg of solid and 81,600 kg of liquid dioxin-contaminated
materials was incinerated between July 1985 and February 1986. During
1987, the MIS was modified to double its capacity and to improve its
reliability. A second trial burn was conducted on both solids and liquids
contaminated with chlorinated organic compounds and PCBs during August and
September of 1987 (5). Since 1987, an additional 3,200,000 kg of solids
and 31,000 kg of liquids have been decontaminated or destroyed.
Over the lifetime of the MIS, a wide variety of feed materials have
been processed. These materials exhibited differences in characteristics
that affected the MIS in various ways. Often a particular characteristic
or a combination of characteristics would affect the MIS performance
adversely. The experiences gained from field operations of the MIS during
the past four years have increased the understanding of the interplay of
feed characteristics with hardware.
This paper describes the effects of feed characteristics on the MIS
performance; correlates various feed characteristics with affected parts of
the system; describes actions taken to mitigate the resulting problems; and
discusses the limits imposed on capacity and reliability by the various
feed characteristics.
FEED CHARACTERISTICS
Both the physical and the chemical properties of the feed determine
incineration system performance (6). Important physical properties
include: heating value, morphology, density, rheology, ash particle-size
distribution and fusion characteristics. Important chemical properties
include the composition of the feed as shown by: organic content, organic
hazardous constituents, acid forming elements such as sulfur and the
halogens, moisture content, and inorganic ash components. These properties
can affect the operating parameters, the capacity, and the reliability of
the incineration system. Many of these properties are interdependent as
far as their effect on the incinerator performance. The manner in which
these properties affect the performance of incineration systems, based on
the experience with the MIS, is summarized in Table 1.
-------
TABLE 1. PROPERTIES AFFECTING INCINERATION SYSTEM PERFORMANCE
Property
Hardware affected
Operating parameter affected
Effect on performance
Feeds of concern
Heating value
Morphology
Density
Rheology
Halogen and sulfur content
Moisture
Particle size distribution
Ash fusion characteristics
(Determined by chemical
characteristics, e.g. alkalis)
Rotary kiln
Rotary kiln temperature. Flue gas Feed capacity. Fuel usage
residence time in SCC
Feed system, Ash gates
Rotary kiln. Corbel Weight of material held by kiln
Feed interruptions
Feed capacity
Feed system,
Rotary kiln
Ouench system, Air
pollution control
equipment design and
operation
Feed system, SCC
Cyclone, SCC, Ducts,
Wet Electrostatic
Precipitator (UEP),
Instrumentation
Rotary kiln. Cyclone,
Ducts, Ouench elbow.
Instrumentation
Frequency of feeding, Ash purity. Feed capacity
Kiln rotation speed
Pump cavitation, pH control. Feed capacity. Caustic usage
Slowdown rate. Particulars emissions
Plastics. Tresh,
Wooden pallets,
Brominated sludge
Steel barrels, Steel
rings, Plastics,
Wooden pallets,
Vermiculite
Brominated sludge
Flue gas residence time in SCC,
Rotary kiln temperature
Feed capacity. Fuel usage
Kiln draft. Paniculate emissions Fouling of ducts, cyclone.
Excess oxygen control. Temperature SCC, process Mater system.
Control & Instruments
Kiln draft. Temperature, Excess
O control
Slagging of kiln. Plugging
of instruments & downstream
equipment
Trial burn mixture,
Brominated sludge
Muddy soil.
Brominated sludge
Erwin Farm soil,
Broninated sludge.
Vermiculite
Plastics, Trash
Brominated sludge
-------
EFFECT OF HEATING VALUE
Heating value of the feed material affects both feed capacity and fuel
usage of the Incinerator. As the heating value of feed material increases,
kiln temperature can increase and sometimes become uncontrollable. The
kiln also requires greater amounts of oxidant to complete combustion and
greater quantities of inert material to control kiln temperature. This
temperature increase can limit feed capacity. The MIS reached its capacity
limit at 1.33 to 1.61 megawatts (MM) heat input to the kiln. Feed
materials, such as plastics, trash, wooden pallets, and brominated sludge,
had capacity constraints caused by high calorific values. Maximum feed
capacities for these materials ranged from 90 kg/hr for pure plastics to
859 kg/hr for brominated sludge as shown in Table 2.
Solid materials with high calorific values cause transient behaviors
that sometimes further limit feed capacity. Plastics, trash, and wood
ignite almost immediately after they are fed to the kiln. Gases evolved
from these materials burn rapidly producing a sharp increase in kiln
temperature and a sharp decrease in excess oxygen. Prior to the 1987
modifications, the MIS was extremely sensitive to these transients, which
caused many feed stoppages.
After the addition of the LINDER Oxygen Combustion System (OCS), the
MIS response to the above transient behavior was improved, and feed
stoppages due to low oxygen were virtually eliminated (7). However, there
were still many feed cut-offs caused by excessive kiln temperatures. These
were minimized by operating the kiln at the lower end (790°C) of the
temperature range, allowed by the RCRA permit, and by using water injection
to control kiln temperatures.
For brominated sludge, the behavior of the MIS was somewhat different.
Large oscillations of the kiln temperature and excess oxygen level occurred
even when the kiln was operated at 790°C. The resulting
over-temperatures (greater than 1040 °C) caused feed cut-offs and loss of
the kiln burners. Loss of the burners increased the length of the feed
cut-off period. The operating changes required to alleviate this
phenomenon are described below.
To reduce the amplitude of the temperature excursions, an automatic
feed cut-off was introduced into the kiln control system. This stopped
feed whenever the kiln temperature exceeded 945°C or the oxygen level
dropped below 4% (wet). This action minimized overtemperature incidents,
but the oscillation frequency was still large. As shown in Figure 1, about
four oscillations occurred per hour. Feed was cut off for approximately
eight minutes during each oscillation. Feed rate was limited to 450 kg/hr
under these operating conditions.
Observations of the kiln during the oscillations showed that the sludge
was not igniting rapidly. Several batches of sludge would be fed by the
ram before ignition occurred. After ignition, flame would fill the kiln
and oxygen flow and temperature would increase rapidly. After the sludge
-------
TABLE 2. FEED CHARACTERISTICS AND EFFECT OF VARIOUS FEEDS ON MIS PERFORMANCE
Feed material
X of Particles Heating X Moisture Organic Halogen Kiln Maximum Limitation on feed Operating
<100 microns value content content temp, feed rate concerns
(Ash)
(cal/g)
(X)
(X)
( C) (kg/hr)
Dermey Farm soil-clay
(1985-1986 before mods)
1.24 g/cc density
(1987-1988 after mods)
Mud » Soil
Mud * Plastics
Eruin Farm soil-silt
(1985-1986 before mods)
(1987-1988 after mods)
Mud «• Soil
Mud + Plastics
Plastics & trash
0.9-2.2 g/cc
4.8
S3.4
1500
1500
20
>70
20
>70
3600-10500 0-10
2.5
50-100
0.13 871-927 900
SCC res. time, Ram Jamming of Ran,
capacity 3-day SCC cleanout
req. every 20 days
788-815 2275 dry Overflow of kiln
corbel
11 1140 mud Doctor blade j arm ing
680
Doctor blade. Kiln 3-day kiln cleanout
Slagging, Kiln temp, cleanout every 10
BTU input days
871-927 900
SCC res. time, Ram Jamming of ram,
capacity 3-day SCC cleanout
every 5 days
•• 788-815 2275 dry Overflow of kiln
corbel
11 " 1140 mud Doctor blade jamming
» "680 Doctor blade, Kiln 3-day kiln cleanout
Slagging, Kiln temp, every 10 days
BTU input
0-57 788-815 90-320 Doctor blade jamming 3-day kiln clean out
Kiln slagging, Kiln cleanout every 10
temp., BTU input days
(continued)
-------
TABLE Z (continued)
Feed material
Wooden pallets
0.5-0.8 g/cc
Steel barrels
Verrnicul ite
Brominated sludge
Trial burn mixture
Solids
Liquids
X of Particles Heating X Moisture Organic Halogen Kiln Maximum Limitation on feed
<100 microns value content content temp, feed rate
(Ash)
-------
had burned several minutes, the flames would extinguish and the oxygen flow
and temperature would decrease rapidly to complete the cycle of
oscillation. It became apparent that steady ignition of the material was
required to prevent the oscillations.
. The kiln temperature was Increased from 790°C to 900°C to provide
the necessary energy to evaporate the water and volatile organics so that
ignition could be sustained. This operating change was successful in
reducing the oscillations to a minimum as shown in Figure 2. Maximum feed
rate was Increased to 900 kg/hr by the above changes.
For feeds with high heat content such as brominated sludge, the
capacity of the MIS is increased by water injection. Due to its high heat
capacity, water provides a very effective heat sink. Consequently, when it
is used to control kiln temperature, moisture increases the SCC residence
time as compared to the use of excess air. Figure 3 shows that the use of
water injection can increase capacity by about 20% over the use of excess
air at a given SCC residence time. The use of oxygen in the kiln enhances
the effect of water injection by allowing further capacity increases. At
an enrichment to 40% 0? in the combustion air, capacity can increase by
60%.
EFFECT OF MORPHOLOGY
The morphology of the feed material affects the feed system by causing
periodic jams. Most of these problems are caused by materials that are
poorly prepared by the shredder due to their morphological
characteristics. Problem materials consist of wooden pallets, metal drum
closure rings, plastics, trash, clothing, and mud (8,9).
The feeding of these materials restricts the MIS capacity. As shown in
Table 2, relatively dry soils can be fed at rates up to 2275 kg/hr but the
presence of plastics and mud reduces the feed rate to 680 kg/hr.
The shredder is used to prepare the solid feed materials for
incineration. For most materials, the shredder works extremely well.
However, wooden pallets and metal drum closure rings often cause feed
blockages when shredded in the present equipment. While the shredder
breaks most of the wooden pallet into 5-cm wood chips, an occasional board
will position itself to go through the shredder as a 5-cm wide by 1.3-m
long sliver. The same is true for the drum lid rings. The shredder
sometimes drags a ring through, straightening it but not cutting it. In
each case, plugging of the conveyor, weigh scale, or ram follows. The best
solution has been to manually separate and prepare these feed materials by
cutting them Into small pieces (about 24 cm in length) prior to shredding.
The shredder is unable to handle pipe or thick metal pieces. These
must be manually classified, cut to proper size, and placed on the main
feed conveyor downstream of the shredder.
The shredder also performs poorly on materials such as plastic,
clothing, trash, and mud. These poorly shredded materials often jammed at
-------
Temp.
°C
1095
955
815
675
NOTES:
- Brominatod Sludge
- September 2, 1988
- Before Operating Changes
Figure 1. Oscillations in temperature and
excess oxygen in rotary kiln.
NOTES:
- Bromlntt*d Sludg*
- September 8, 19&8
- Att*f Operating Changes
Figure Z. Kfln temperj;jre ana excess oxygen
after opera! "q change.
-------
the doctor blade or restricting dam that was originally used to level the
granular material on the conveyor belt as the belt exits the shredder
hopper. The doctor blade worked quite well for granular material but it
created large blockages when materials such as shredded plastic, clothing,
trash, metal, or mud were fed. A roller has been installed to replace the
doctor blade, and this has reduced the number of jamming Incidents.
In one Instance, the combination of metal and mud damaged the main
conveyor belt. The belt was slit for 21 of its 26-m length by a metal
shard which was embedded in mud. The clump of mud containing the shard had
stuck to the belt, passed under the top-end belt wiper, and lodged in the
underside rollers. The top-end wiper was repositioned to provide a more
positive wiping action.
Finely granulated material affected the operation of the ram feeder.
It would bypass the ram head and collect on the backside. The material on
the backside would periodically prevent the ram from fully retracting. A
small chain-plug conveyor was installed and timed to convey the bypassed
material to the front side of the ram. This solution has worked quite
well.
EFFECT OF DENSITY
The density of the feed determines capacity for many feed materials.
For feeds of typical densities (1.5 g/cc), such as soils, the maximum feed
rates were 2275 kg/hr. The maximum feed rate was obtained at a kiln
revolution rate of 1.6 rpm, which gives a typical solids residence time of
30 minutes. For low density materials such as vermiculite (0.096 g/cc),
feed rates up to 364 kg/hr were feasible.
The density of the feed determines capacity for many feed materials,
because the density of a material is inversely proportional to the volume
it occupies, and the ultimate feed rate for a given material is limited by
the volume capacity of the kiln. The volume capacity of the kiln is
determined by the amount of material that can be held by the kiln without
overflowing the corbel. The corbel is an annular lip on the front end of
the kiln, which rotates with the kiln. There is a space between the corbel
and the front plate of the kiln. When material gets into this space, the
seals of the kiln are damaged. The maximum volume that the kiln can hold
without spilling over the corbel is about 15% of its total volume.
EFFECT OF RHEOLOGY
The rheology of the material affects either the feed system or the
decontamination behavior. Muddy soils, fed to the MIS, formed clumps of
material, which were caught by the doctor blade and also would stick to the
conveyor belt, weigh scale and ram trough. The resulting buildup would
periodically plug various parts of the feed system thus reducing overall
feed rates to about 1140 kg/hr. This is approximately 50% of the maximum
rates achievable with dry soils. Addition of vermiculite has eliminated
the sticking of the muddy material while adding only a small amount to the
throughput weight.
-------
The brominated sludge that was fed had a tendency to form balls up to 8
cm in diameter. Since the time required for burnout of a sphere is
proportional to the square of its diameter, the large balls require a much
longer residence time in the kiln for decontamination. At the normal 1.6
rpm revolution speed of the kiln for soil, small smoking particles would
exit the kiln with the kiln ash. This required limiting feed rate to about
450 kg/hr. However, when the residence time of the sludge was increased by
reducing the revolution speed to 0.8 rpm, feed rates up to 900 kg/hr were
achievable without smoking.
EFFECT OF HALOGEN AND SULFUR CONTENT
Incineration of brominated and chlorinated wastes generate the acid
gases hydrogen bromide (HBr) and hydrogen chloride (HC1). These acid gases
affect the capacity, the blowdown rates that control total dissolved solids
(TDS) in the process water system and the particulate emissions.
A capacity limit of 115 kg/hr of acid-forming organic chlorides was
encountered during the 1987 trial burn (5). The capacity limit was caused
by pump cavitation in the quench system, which cools the gases exiting from
the SCC to about 95°C. The cavitation reduced the quench water flow rate
which activates the protective instrumentation resulting in cut-off of the
feed and shutdown of the burners. This cavitation was produced by
excessive chlorinated waste feed rate as follows: The quench water is
treated with caustic solution to neutralize acid gases. Reaction between
HC1 in the combustion gases and sodium hydroxide (NaOH) in the quench sump
produces effervescence. The amount of effervescence increases to violent
levels as HC1 flow rate increases. The violent effervescence reduces the
available net positive suction head (NPSH) of the pump, which causes
cavitation at very high HC1 loads.
High organic chloride loads also affect particulate emissions through
the phenomenon of mist carry-over into the stack. The amount of carry-over
is determined by both the HC1 loading of the flue gas and the TDS of the
process water (5,10).
In tests performed prior to the trial burn, particulate emissions were
found to exceed the allowable emissions (180 mg/dscm) by as much as a
factor of three. The relationship between particulate emissions and
organic chloride loading is shown in Figure 4. Table 3 presents the
analysis of the Method 5 particulate filter cakes, which shows that the
major portion of the particulate was sodium chloride (NaCl) and sodium
hydroxide (NaOH). The emissions were brought into compliance after a mist
eliminator was installed.
However, data taken during the 1987 trial burn showed that TDS of the
process water also affected particulate emissions. As shown in Figure 5,
particulate emissions were proportional to TDS during the trial burn
tests. The data shows that operation with TDS at 20,000 ppm adequately
controls particulate levels at high chlorine loadings.
The TDS is controlled by adjusting the blowdown rates as follows: For
-------
TABLE 3. RESULTS OF ANALYSES OF METHOD 5
PARTICIPATE FILTER CAKES FROM HIGH
ORGANIC CHLORIDE LOADING TESTS OF HIS
Test Description
Date (1987)
High Ch1or1d« *
without
nlst eliminator
6/18-19
High Chloride **
with
mist eliminator
7/20
Total participates
Iron (Fe)
Chromium (Cr)
Sodium (Na)
Chloride (Cl)
Aluminum (A1)
As Sodium Chloride (Had)
Remaining Na as Sodium Hydroxide
(NaOH) (Excess caustic)
(weight In grus)
0.3412
0.0202
0.0022
0.1560
0.1367
(weight 1n grams)
O.OS24
O.OOS7
0.22S3
0.1174
0.0175
0.0200
0.0022
0.0329
0.0079
Organic chloride loading • 71.5 kg/hr
Organic chloride loading - 83.4 kg/hr
TABLE 4. RESULTS OF ASH DEPOSIT ANALYSIS
FROM HIS CYCLONE RISER DUCT
Analyte
Silicon
Aluminum
Titanium
Iron
Calcium
Magnesium
Sodium
Potassium
Sulfur
Phosphorous
(SI)
(Al)
T1
(Fe)
(Ci)
(Kg)
(Na)
(K)
(S)
(P)
Reported as
Silicon Dioxide
Aluminum Dioxide
Titanium Dioxide
Ferric Oxide
Calcium Oxtde
Magnesium Oxide
Sodium Oxide
Potassium Oxide
Sulfur Trioxide
Phos. Pentoxide
Amount
(wt. %)
25.6
8.5
0.3
5.0
36.7
1.5
0.4
0.4
21.0
0.5
Analytical
technique
X-ray
X-ray
X-ray
X-ray
X-ray
AA
AA
X-ray
X-ray
X-ray
AA • Atomic Absorption Spectroscopy
-------
0 0.2 0.4 O.t 0.1 1 1.2 1.4 1.( 1.1 2
NOTES:
Solid Feed Heating Value t .50 Keal/g
Kiln Temperature 925*C
SCC Temperature 1200*C
A - Water Infection using 40% oxygen-enriched air tor
combustion
B - Water injection using air for combustion
C - Air cooled using either air or 40% oxygen-enriched
air for combustion
Figure 3. Effect of cooling media on SCC residence time.
TOO
(00 -
£- wo -
o
55
*oo .
t
JOO .
(ii
§
o ""^
(00 .
20 40 M K
ORGANIC CHLORIDE LOADING (kg/hr)
100
NOTES:
i- Before Addition of Mist Eliminator
After Addition of Mist Eliminator
r»gur« 4. Effect of organic chloride loading
on participate emissions.
-------
a chlorinated waste 1.65 kg NaCl 1s formed per kg Cl 1n the waste. To
maintain the IDS at 20,000 ppm, 82.4 kg of process water must be drawn from
the system for every kilogram of Cl processed. This illustrates that the
acid content determines the required process water blowdown rate.
EFFECT OF MOISTURE
Moisture content affects incinerator performance and can adversely
affect rheological behavior as described above. Depending upon the heat
content of the waste, moisture can either improve or impede incinerator
performance.
When using feeds with high heat content moisture acts as a heat sink to
control kiln temperatures.
Conversely, when using feeds with low heat content, moisture increases
auxiliary fuel requirements and decreases SCC residence time causing a
capacity debit. The effect of moisture on SCC residence time is shown in
Figure 6.
EFFECT OF PARTICLE SIZE DISTRIBUTION
The particle size distribution of the ash generated from the waste
determines the amount of particulate carry-over from the rotary kiln to the
rest of the system. The importance of this characteristic can be
demonstrated by the MIS experience with Oenney Farm soil and Erwin Farm
soil. As shown in Table 2, Denney Farm soil is much coarser than the Erwin
Farm soil. Up to 25% of the Erwin Farm soil would carry over from the kiln
to the SCC. This caused a rapid buildup of solids in the SCC. The solids
buildup necessitated a 70-hr shutdown for removal of the slag after each 96
to 120 hours of operation (45,000 kg of soil processed). The behavior of
the silt caused the unit to be unavailable for operation an average of 40%
of the time due to the need to clean out the SCC. On the other hand, the
unit could process the coarser Denney Farm soil for about 600 hours
(270,000 kg of soil processed) before a shutdown for slag removal from the
SCC was required. The unit was unavailable for 10% of the time due to SCC
clean-outs with the coarser Denney Farm soil. In both cases, the buildup
of solids in the SCC significantly reduced the availability of the
incinerator.
The problem was mitigated in 1987 when a cyclone was added, between the
kiln and the SCC, to remove the fines carried over from the kiln. The
system operated over a three-month period and processed over 500,000 kg of
solid material without requiring a shutdown for slag removal from the SCC.
Although the cyclone has alleviated the solids buildup in the SCC, fine
particulates still have caused problems with the operating instruments.
The large number of fine particulates associated with brominated sludge
fouled the kiln oxygen meter and the SCC thermocouple about once every
eight hours. This increased the number of over-temperature incidents in
the rotary kiln, caused the incinerator feed cutoff to actuate due to a
false SCC low temperature measurement, and increased the fuel flow to SCC
-------
u
240
220 -
200 -
1H
1(0 .
140 .
120 .
100 .
K .
to 4
40
20
0
M 40
TOTAL DISSOLVED SOLIDS - TOS (ppoi)
NOTES:
Solids Feed Rat* 1800 kg/hr
Liquids Feed Rat* 60-160 kg/hr
Organic Chlorids F*ed Rats 50-74 kg/hr
figure 5. Effect of IDS on participate emissions.
I
u
ut
cc
u
30
MOISTURE {%)
so
NOTES:
Solid Fssd Rats 1820 kg/hr
Kiln Tsmpsraturs 925°C
SCC Trnnpsraturs 1200*C
30% Oxygsn-Enrlchsd Air
r'9ure 6. Effect of feed moisture content
on SCC residence time.
-------
Sample 1
QUENCH ELBOW
1.5x
Sample 2
CYCLONE RISER
DARK SURFACE
Figure 7. Macrophotograph of samples from Quench Elbow (1)
and Cyclone Riser (2) .
-------
Figure 8. SEM photomicrographs of the cross section
of the deposit from the Cyclone Riser.
-------
burners. The thermocouple problems have been eased by changing the
thermocouple location and using a thermowell rather than an aspirating
thermocouple. No satisfactory solution has been found for the kiln oxygen
measurement.
EFFECT OF ASH FUSION CHARACTERISTICS
The ash fusion characteristics of some feed materials caused the
formation of hard slag deposits in the kiln; consolidated deposits in the
ductwork between the kiln and cyclone, between the cyclone and SCC, and in
the cyclone exit tube; and consolidated deposits in the SCC exit venturi.
The ash fusion characteristics are determined by the chemical composition
of the ash.
Ashes containing elements such as sodium (Na) and potassium (K) have
low slagging temperatures. Ashes from plastics, glasses, wood, and other
components of trash are rich in these compounds. The increased slagging
tendency of trash was experienced in the rotary kiln, which required a
system shutdown about every ten days to remove the slag build up caused by
incineration of trash.
Ashes containing significant quantities of calcium (Ca), iron (FeJ,
sulfur (S), or phosphorous (P) have moderate slagging temperatures.
Although brominated sludge, containing both Ca and S did not slag the kiln,
the ash produced a consolidated deposit, shown in Figure 7, which fouled
the ductwork between the kiln and the SCC. A consolidated deposit also
occasionally formed in the quench elbow upstream of the quench nozzles.
This fouling necessitated a system shutdown about every twelve days to
remove the deposits.
Samples of the deposits were analyzed to determine the mechanism of
deposition. More details on this topic are provided in reference 11. The
deposit mechanism was found to be similar to those operative in boilers
fired with subbituminous coal. The deposits were formed by sintering of
calcium sulfate (CaSO*) in the temperature range of 870 and 980°C in an
ash containing 14% CaS04, 23% calcium oxide (CaO), and about 2.5% sodium
oxide (NagO). The formation of fused calcium silicates as a result of
the decomposition of CaSC^ in the presence of quartz and aluminosilicates
between 870-980°C was also an important factor in the mechanism of
deposition. A photomicrograph showing the sintering is presented in Figure
8. Table 4 gives the bulk composition of the deposit from the cyclone
riser.
Most ashes consist mainly of aluminum (Al) and silicon (Si), which
generally have good fusion characteristics (fusion temperatures above
1650°C). Both the Denney Farm and Erwin soils were composed mainly of
SiOg and Al20o. The lack of slagging and of troublesome deposits
experienced with the MIS when processing these materials demonstrates these
good fusion characteristics of Al and Si.
-------
CONCLUSIONS
The paper has shown how various feed characteristics affected the MIS
performance. Concerns stemming from these effects have been discussed and
are summarized below:
o Increased heating value of the feed often reduces the fuel
requirements of the unit. However, as heating value increases
control of kiln temperature becomes Important and feed capacity
can be restricted. Water injection can increase kiln capacity
under these conditions. The experience with the MIS also shows
that operating conditions must be adjusted to insure rapid
ignition of solids that have high heating value.
o Proper feed preparation is very important for maximizing
throughput. Problem materials should be sorted out and specially
prepared prior to feeding.
o Rheology can affect either the feed system or the decontamination
behavior.
o Increased halogen content increases mist formation and can
increase particulate emissions. TDS of process water can also be
important in controlling particulate emissions.
o Moisture content increases fuel requirements and can create
feeding problems for muddy feeds.
o Materials containing large quantities of micron-sized particles
can foul critical instruments and decrease reliability of the
system.
o Ash fusion characteristics can cause formation of slag or other
deposits in various parts of the system. This also decreases
reliability. The processing of pure trash produced slagging
problems in the rotary kiln.
Most of this experience is directly applicable to other mobile or
transportable incinerator systems.
-------
REFERENCES
1. Yezzi, J.J., Jr. et al. Results of the Initial Trial Burn of the
EPA-ORD Mobile Incineration Systems. In: Proceedings of the 1984
National Waste Processing Conference, ASME, pp. 514-534.
2. Yezzi, J.J., Jr. et al. The EPA Mobile Incineration Systems In:
Proceedings of the 1982 National Waste Processing Conference, ASME, pp.
199-212.
3. Lovell, R.O., et al. Trial Burn Testing of the EPA-ORD Mobile
Incineration System. EPA-600/0-84-054, Municipal Environmental
Research Laboratory, Cincinnati, Ohio, 1984.
4. Mortensen, H. et al. Destruction of Dioxin-Contaminated Solids and
Liquids by Mobile Incineration. EPA Contract 68-03-3255, Hazardous
Waste Engineering Research Laboratory, Cincinnati, Ohio, 1987.
5. King, G., and Stumbar, J. Demonstration Test Report for Rotary Kiln
Mobile Incinerator System at the James Denney Farm Site, McDowell,
Missouri. EPA Hazardous Waste Engineering Research Laboratory,
Cincinnati, Ohio, 1988.
6. Brunner, C, Incineration Systems Selection and Design. Van Nostrand
Reinhold Company, New York, 1984.
7. Ho, M., and Ding, M. G. Field Testing and Computer Modeling of an
Oxygen Combustion System at the EPA Mobile Incinerator. JAPCA, Vol.
38, No. 9, September 1988.
8. Gupta, G.D., et al. Operating Experiences with EPA's Mobile
Incineration System, In: Proceedings of the International Symposium
on Incineration of Hazardous, Municipal, and Other Wastes. American
Flame Research Committee, Palm Springs, CA, 1987.
9. Freestone, F.J., et al. Evaluation of On-site Incineration for Cleanup
of Dioxin-contaminated Materials. Nuclear and Chemical Waste
Management, Vol 7, pp 3-20, 1987.
10. Gupta, G.D., et al. MIS Modifications Trial Burn and Operations
February 1986 to September 1987 Draft Report EPA Contract 69-03-3255,
Risk Reduction Engineering Laboratory, Edison, New Jersey, 1988.
11. Bryers, R.W. Deposit Analysis: Cyclone Riser/Quench Elbow - Denney
Farm Site, Foster Wheeler Development Corp., Livingston, New Jersey,
1988.
------- |