United States EPA-340/1-81-002
Environmental Protection Office of General Enforcement February 1981
Agency Washington, DC 20460 T>/?$ /- I I'')
Stationary Source Enforcement Series
An Investigation of
Corrosion in Particulate
Control Equipment
-------
EPA-340/1-81-002
An Investigation
of Corrosion in Particulate
Control Equipment
by
Thomas E. Mappes, Project Manager,
and Robin D. Terns, PH.D.
PEDCo Environmental, Inc.
Cincinnati, Ohio 45246
Contract No. 68-01-4147
Task No. 107
Project Officers:
Kirk E. Foster, Technical Support Branch
Division of Stationary Source Enforcement
Henry Onsgard, Air Programs Branch
Air and Hazardous Materials Division
Region V
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of General Enforcement
Division of Stationary Source Enforcement
Washington, DC 20460
February 1981 i,.S. " wl roimcntal Protection
-
bPL-j6)
Gtvcbt, Room 1670
0604
-------
DISCLAIMER
This report was furnished to the U.S. Environmental Protec-
tion Agency (EPA) by PEDCo Environmental, Inc., Cincinnati,
Ohio, in partial fulfillment of Contract No. 68-01-4147, Task
No. 107. The opinions, findings, and conclusions expressed in
the report are those of the authors and not necessarily those of
the EPA. The mention of trade names or commercial products does
not constitute an endorsement or recommendation for use by the
authors or the EPA.
11
-------
CONTENTS
FIGURES V
TABLES vi
ACKNOWLEDGEMENT vii
CONVERSION FACTORS viii
1. INTRODUCTION 1
Background Information 1
Description of the Study 2
2. SUMMARY OF FINDINGS 4
3. MODES OF CORROSION IN CONTROL EQUIPMENT 7
Acid Dewpoint Corrosion 7
Corrosion by Scrubbing Waters 12
Uniform attack 13
Pitting 13
Crevice corrosion 14
Stress corrosion 14
Weld decay 14
Erosion-Corrosion 16
Other Modes of Corrosion 18
High-temperature corrosion 18
Galvanic corrosion 18
4. FACTORS THAT CONTRIBUTE TO CORROSION IN CONTROL
EQUIPMENT 20
Process Characteristics 20
Chemical constituents 20
Flue gas temperatures and moisture content 22
Process operating cycles 23
Operating and Maintenance Practices 25
111
-------
CONTENTS (continued)
Page
5. AVOIDING CORROSION IN CONTROL EQUIPMENT 27
Controlling Acid Dewpoint Corrosion 27
Reduction of sulfur trioxide and moisture
levels 27
Control of flue gas temperatures 28
Thermal insulation 29
Air inleakage 31
Special alloys and protective coatings 31
Controlling Scrubber Corrosion 34
Selection of materials for scrubbers 35
Control of scrubbing liquor pH 38
Abrasion in scrubbers 39
REFERENCES 41
APPENDIX A CASE HISTORIES A-l
APPENDIX B REVIEW OF LITERATURE ON CORROSION IN AIR
POLLUTION CONTROL EQUIPMENT B-l
IV
-------
FIGURES
Number Page
1 Wireburning Incinerator Scrubber that was
Totally Destroyed by Corrosion 2
2 Uninsulated Fabric Filter Dust Hopper with a
Patch Over Area of Corrosion 9
3 Cast Iron Cupola Fabric Filter that was
Scrapped after 5 Years of Service Because
of Corrosion 10
4 Corner of a Fabric Filter (from inside) Showing
Corrosion of Mild Steel Structure 10
5 Corrosion Damage to the Frame of an Uninsulated
Entry Hatch in a Fabric Filter 11
6 Pitting of Stainless Steel in a Municipal
Incinerator Scrubber Vessel 13
7 Stress Corrosion Cracking of Stainless Steel
Ductwork of a Sewage Sludge Incinerator
Scrubber 15
8 Weld Decay in Stainless Steel Ductwork of a
Sewage Sludge Incinerator Scrubber 15
9 Erosion-Corrosion at the Inlet of a Cyclonic
Mist Eliminator of a Venturi Scrubber Serving
a Hox-Mix Asphalt Plant 16
10 Scouring and Erosion of a Fan Blade of an
Induced-Draft Fan Serving an Industrial Boiler
Fabric Filter 17
11 Electrostatic Precipitator with a Complete
Blanket of Insulation Covering the Chambers,
Hoppers, and Ductwork 30
12 Insulated Fabric Filter with Structural Steel
Protruding through the Insulation 32
v
-------
FIGURES (continued)
Page
13 Gap in the Sheathing Covering an Insulated
Duct 33
14 Cracked Vibration Sleeve in Ductwork 33
TABLES
Number Page
1 Galvanic Series of Some Common Metals in
Seawater 19
2 Properties of Materials Used in the Construction
of Wet Scrubbers and Auxilliary Components 36
VI
-------
ACKNOWLEDGEMENT
This report was prepared for the Environmental Protection
Agency, Division of Stationary Source Enforcement (DSSE), by
PEDCo Environmental, Inc., Cincinnati, Ohio. PEDCo sincerely
appreciates the assistance provided by Mr. Thomas Rigo, Ohio
Environmental Protection Agency; Messrs. Ray Goetz and Gary
Stoneburner, Virginia State Air Pollution Control Board; and Mr.
E. Walter Linna, Chicago Department of Energy and Environmental
Protection in arranging the plant visits; and by Mr. Sidney R.
Orem, Industrial Gas Cleaning Institute in arranging the in-
terviews with control equipment manufacturers.
The EPA Project Officers were Mr. Kirk Foster, DSSE, and
Mr. Henry Onsgard, EPA Region V. Mr. Thomas E. Mappes served as
PEDCo"s Project Manager and principal author. Dr. Robin D. Terns
authored portions of the report and served as photographer.
VII
-------
CONVERSION FACTORS
1 meter (m) = 3.281 ft
1 meter (m) = 39.37 in.
1 meter2 (m2) = 10.76 ft2
1 meter3 (m3) = 35.31 ft3
1 meter3 (m3) = 1000 liter (£)
1 metre3/second (m3/s) = 2.119 x 103 ftVmin (cfm)
1 liter (£) = 0.264 gal (U.S. liquid)
1 kilogram (kg) = 2.205 Ib
1 kilogram (kg) = 1.102 x lo"3 short tons (2000 Ib)
1 megagram (Mg) = 1.102 short tons (2000 Ib)
1 joule (J) = 9.471 x lo"4 Btu (mean)
1 joule (J) = 2.778 x io~7 kWh
1 watt (W) = 1.341 x io"3 hp
1 pascal (Pa) = 1.450 x 10~4 lbf/in.2 (psi)
1 pascal (Pa) = 4.019 x io"3 in. of H20 (60°F)
1 kilopascal (kPa) = 0.01036 atmospheres (Atm)
1 kilopascal (kPa) = 4.019 in. of H20 (60°F)
Degrees Celsius = (degrees Fahrenheit - 32) -r 1.8
Degrees Kelvin = degrees Celsius + 273.15
Vlll
-------
SECTION 1
INTRODUCTION
The Division of Stationary Source Enforcement (DSSE) of the
U.S. Environmental Protection Agency is concerned about malfunc-
tions in particulate control equipment at stationary sources
that result in violations of emissions regulations. Previous
studies have led DSSE to believe that corrosion is one of the
leading causes of these malfunctions. DSSE conducted this study
to provide to agency personnel and control equipment users back-
ground information on corrosion and guidelines for avoiding such
corrosion.
BACKGROUND INFORMATION
In 1978 the President's Council on Environmental Quality
conducted a survey to identify the causes of excess emissions
from controlled stationary sources that had initially been in
compliance with emissions regulations. Inspection of 20 such
sources showed that corrosion-related malfunctions in the con-
trol equipment were the primary cause of excess emissions at 3
sources and contributing causes at 6 sources.1 Subsequent
source inspections conducted for DSSE confirmed that corrosion
is a common cause of malfunction in particulate control equip-
ment.2 These inspections also revealed that many sources do not
devote sufficient attention to corrosion control when they
select particulate control systems. In many cases proper corro-
sion control measures have been determined by trial and error at
great expense (Figure 1). In addition, the transfer of knowl-
edge gained in this manner is often slow, allowing mistakes to
be repeated in subsequent applications.
-------
Figure 1. Wire-burning incinerator scrubber
that was totally destroyed by corrosion.
Based on these results, DSSE saw a need to assist particu-
late control equipment users and State Enforcement Agency per-
sonnel in coping with corrosion problems. They commissioned
this study to characterize the effects that corrosion can have
on the performance and operating life of particulate control
equipment and to assemble guidelines to help reduce the fre-
quency and severity of excess emissions incidents due to corro-
sion. They hoped such guidelines would also help reduce costs
attributable to corrosion by encouraging the use of more relia-
bly designed equipment.
DESCRIPTION OF THE STUDY
The study included four phases. The initial phase was a
review of technical literature to determine the extent of knowl-
edge on corrosion in particulate control equipment. Information
obtained by a computerized search of engineering and environ-
mental indexes at the Air Pollution Technical Information Center
-------
of the Office of Administration, Information Services Division,
U.S. EPA, Research Triangle Park, North Carolina, was supple-
mented with information from other sources and summarized in an
indexed bibliography.
The second phase of the study was the identification of
particulate emissions sources where control equipment corrosion
was known to be a problem. Most of these candidate facilities
were identified through the assistance of State Agency field
inspectors.
The third phase of the study was the inspection of a samp-
ling of the identified sources to gain first hand information
about specific corrosion problems. This phase involved the
inspection of 38 control devices at 18 facilities, all of which
participated voluntarily. To help elicit honest and complete
responses from personnel at these facilities it was agreed
before-hand that neither facility names nor identifying process
information would be listed in this report. Several facility
managers insisted that no photographs be taken within their
facilities before agreeing to participate.
The final phase of the study consisted of a series of
interviews with design engineers at seven manufacturers of
particulate control equipment. As in the case of the emissions
sources, these manufacturers are not identified in this report.
Section 2 of this report lists the principal findings of
the various phases of this study. Section 3 describes common
modes of corrosion reported in the literature, observed during
the study, or reported by the equipment manufacturers. Sec-
tion 4 discusses the possible effects of process characteristics
and plant operation and maintenance practices on control equip-
ment corrosion. Section 5 presents guidelines for avoiding
corrosion in particulate control equipment. The appendices
include inspection reports for each of the 18 facilities in-
spected during the study and the indexed bibliography about
corrosion in particulate control equipment. The inspection
reports are referenced as examples in the following sections of
the text.
3
-------
SECTION 2
SUMMARY OF FINDINGS
This section lists eight major findings of the study. Some
of the findings are illustrated by the case histories described
in the Appendix A inspection reports. References are made,
therefore, to these inspection reports where appropriate.
1. The study confirmed that corrosion in air pollu-
tion control equipment is a common underlying
cause of equipment malfunction. In some cases,
corrosion has led to surprisingly rapid destruc-
tion of control equipment. (See Appendix A,
Sources 12 and 17). In other cases, less severe
corrosion has resulted in lower particulate
removal efficiency or decreased system reliabil-
ity.
2. Literature addressing corrosion in particulate
control equipment is beginning to appear in
technical journals (see Appendix B). Numerous
papers and reports have been published about
operation and maintenance problems, including
corrosion, in particulate control equipment at
municipal incinerators. Several papers have also
been published about similar problems in equip-
ment serving various iron and steel industry
processes, industrial and utility boilers, pulp
and paper mills, and cement plants. In addition,
two seminars have been held about corrosion in
air pollution control equipment. These were
sponsored jointly by the National Association of
Corrosion Engineers, the Industrial Gas Cleaning
Institute, and the Air Pollution Control Associa-
tion.
Certain problems observed during the plant visits
of this study or cited by the control equipment
manufacturers have not been addressed adequately
in the literature. These include corrosion in
equipment serving the primary and secondary
nonferrous metals industries, lime kilns, crushed
-------
aggregate dryers, and the phosphate fertilizer
industry; pH control in scrubbers; and the
relationship of wet quenching to dewpoint corro-
sion.
Corrosion can increase particulate emissions at
controlled sources. In some cases corrosion has
completely destroyed the control equipment, re-
sulting in extended periods of uncontrolled emis-
sions (Source 17). In other cases corrosion has
resulted in frequent short-term use of bypass
stacks during periods of malfunction (Sources 6
and 13). Some control equipment remains opera-
tional after suffering corrosion damage, but the
particulate removal efficiency is often reduced
by the corrosion (Sources 11, 12, 14, and 16).
Corrosion-related system failures often add sig-
nificant, unexpected costs to the operation of
the controlled process. These failures sometimes
neccessitate replacement of expensive capital
equipment long before the end of its designed
life (Sources 1, 2, 6, 10, 12, 14, 17, and 18),
cause substantial increases in maintenance costs
(Sources 4, 5, 7, 15, and 16), or cause substan-
tial loss-of-production costs at facilities where
the regulations do not permit uncontrolled emis-
sions during repair of control equipment (Source
13).
Several recurring causes are responsible for most
of the corrosion problems in control equipment:
poor control of flue gas temperature, which leads
to acid dewpoint problems (Sources 5, 6, 7, 9,
10, 12, and 16); inappropriate materials of
construction and/or poor pH control in scrubbers
(Sources 1, 8, 13, 17, and 18); and the combined
effects of corrosion and abrasion (Sources 15 and
16). Acids derived from sulfur oxides and nitro-
gen oxides are the most common agents responsible
for corrosion. Chlorides are also a common cause
of corrosion.
Particulate control devices that seem to be par-
ticularly susceptible to corrosion are those
serving cast iron cupolas, municipal incinera-
tors, coal-fired boilers, lime kilns, cement
kilns, rotary aggregate dryers, nonferrous-metals
industries, and the phosphate fertilizer indus-
try.
-------
7. Most of the corrosion failures observed during
the study could have been prevented or reduced in
severity by application of existing technologies.
In many cases the sources selected inappropriate
equipment designs because of inexperience or
because of a perceived cost savings. Poor oper-
ating practices have also contributed to corro-
sion in some cases. Scrubber manufacturers
indicated that a major problem in scrubber design
is the frequent inability of sources to provide
complete process information when ordering par-
ticulate scrubbers. The manufacturers claim that
scrubbers can generally be designed to withstand
corrosive environments if all corrosive constitu-
ents affecting the scrubbers are identified prior
to design. Many scrubber users do not provide
chemical analyses of the raw materials, flue
gases, or scrubbing waters to the manufacturers.
8. The findings in 7 indicate that the need is not
as great for further research as it is for a more
effective application of existing technologies.
Research could be beneficial in a few areas, how-
ever. One such area is the development of fuel
cost models to compare the costs of increasing
process temperatures with potential benefits
derived from reducing dewpoint corrosion. The
cement industry is a good candidate for this type
of study. Another area is the analysis of pH
control systems to identify features that could
improve system reliability. The emphasis of such
research should be to develop pH control systems
that can be properly maintained at small sources
with small, poorly trained staffs. Finally, it
may be beneficial to run a series of in situ
tests of various alloys and protective coatings
in a variety of corrosive control equipment
applications. Tests such as these would increase
knowledge about the behavior of materials used in
various control equipment environments.
-------
SECTION 3
MODES OF CORROSION IN CONTROL EQUIPMENT
Despite the wide variety of particulate sources and the
numerous designs and configurations of particulate control de-
vices, only a few modes of corrosion are responsible for most
failures reported in this study. These modes include acid dew-
point corrosion, corrosion by scrubbing liquors, and erosion-
corrosion (or abrasion). Each of these modes is described in
detail below.
ACID DEWPOINT CORROSION
Acid dewpoint corrosion is the predominant mode of corro-
sion in fabric filters and electrostatic precipitators (ESP's)
serving hot sources. The acid dewpoint generally refers to the
temperature at which sulfur trioxide vapor (S03) will combine
with water vapor to form sulfuric acid (H2SO4). Sulfur trioxide
is found most frequently in flue gases produced by processes
using sulfur-bearing fuels such as coal, coke, and oil. Sulfur
trioxide can also be formed in primary metals-refining proces-
ses, such as copper smelting, because the ores typically contain
large quantities of sulfur.
Combustion of sulfur-bearing fuel first results in the for-
mation of sulfur dioxide (SO2); some of this SO2 is further oxi-
dized to SO.,. The exact amount of SO9 converted to SO_ is de-
*^ ^ *3
pendent on many variables including gas temperatures in the com-
bustion zone, the configuration of the combustion chamber, the
availability of oxygen in the flame, and the concentrations of
impurities such as vanadium which can act as catalysts.3
-------
Sulfur trioxide and water have a strong affinity for each
other; when temperatures are lowered to the dewpoint the two
combine rapidly to form sulfuric acid molecules. The sulfuric
acid molecules in turn have a high affinity for water. As they
condense they draw additional water molecules from the gas
stream forming a concentrated acid solution. Therefore, when
flue gases containing relatively small concentrations of SO,,
reach the dewpoint temperature, droplets of concentrated sulfur-
ic acid can condense on the cooler surfaces. For example, an
82.5 percent sulfuric acid solution (by weight) will condense at
148 °C from flue gases containing as little as 40 ppm (by volume)
SO3 and 10 percent (by volume) water vapor.4 Most fabric filter
and ESP enclosures and much of their internal hardware are
constructed of mild steel, which is very susceptible to sulfuric
acid attack.
After review of available data, Verhoff and Banchero devel-
oped an empirical relationship from which the dewpoint tempera-
ture can be calculated when the percentages of water vapor and
sulfur trioxide are known (Equation I).5
T _ _ 1,000
DP 1.7842+0. 2691 ogP,, n-0.10291ogPcn +0. 03291 ogPu
M2 3
where
Tnp = dewpoint temperature in degrees kelvin
?„ _ = vapor pressure of water in atmospheres
H2°
P = vapor pressure of sulfur trioxide in atmospheres.
so3
This equation has agreed with most experimental results to
within about 7 degrees kelvin. In cases where measurement or
estimation of S03 and water vapor concentrations is not practi-
cal, gross estimates of the dewpoint can be based on sulfur
content of the fuels.6 In either case it is advisable to add a
factor of safety to the estimated dewpoint temperature to allow
8
-------
for variations in flue gas composition and to compensate for
nonuniform flue gas temperatures throughout the system.
Acid dewpoint corrosion is most likely to occur in loca-
tions where flue gas temperatures are the coolest or where steel
surfaces are the coolest. Temperatures in dust hoppers are
often cooler because gas detention is longer than at other
locations (which allows more time for cooling) and because the
hoppers have a large surface-to-volume ratio (which increases
the rate of radiant heat loss to ambient air). Thus, hoppers
without properly designed heaters or insulation are frequently
affected by acid dewpoint corrosion (Figure 2). Outer walls
(especially corners) of a filter or an ESP are other locations
that are often cooler and subject to acid dewpoint corrosion
(Figures 3 and 4). Entry hatches and hatch frames are also more
likely to corrode because of cool air inleakage and because the
hatches are often not as well insulated as the rest of the
structure (Figure 5).
Figure 2. Uninsulated fabric filter dust hopper with
with a patch over area of corrosion penetration.
-------
Figure 3. Cast iron cupola fabric filter that was
scrapped after 5 years of service because of corrosion.
Figure 4. Corner of a fabric filter (from inside) showing corrosion
of mild steel structure. Note that new horizontal beams have
been welded to existing beams to reinforce corroded beams.
10
-------
Figure 5. Corrosion damage to the frame of an uninsulated
entry hatch in a fabric filter.
11
-------
Components that suffered from acid dewpoint corrosion in
the 11 fabric filters studied include dust hoppers (Sources 5-7,
16), sidewalls and corners (Sources 5-7, 12, 14, 16), entry
hatches (Sources 9, 12, 14), bag hangers and other bag hardware
(Sources 9, 12, 16), filter compartment isolation dampers
(Sources 11, 16), and "top end" components such as bag support
structures and bag cleaning mechanisms (Sources 9, 11). Five
fabric filter manufacturers reported that dust hoppers, top end
components, and bag hardware are the components most frequently
affected by corrosion.
ESP enclosures are subject to cooling effects similar to
those affecting fabric filter enclosures; therefore, many corro-
sion problems reported in fabric filters also appear in ESP's.
As in fabric filters, acid dewpoint corrosion often attacks dust
hoppers, corners of the enclosures, access hatches, hatch
frames, and the undersides of roof plates. It can also occur in
penthouses that contain support insulators. At installations
where flue gas temperatures are very low, acid dewpoint corro-
sion can attack structures within the ESP, such as discharge
plates and discharge wires. Examples of acid dewpoint corrosion
in ESP components are described in the Source 10 report.
CORROSION BY SCRUBBING WATERS
The interaction of process gases and scrubbing liquids in
particulate scrubbers very often creates liquors that are ex-
tremely corrosive. Unlike acid dewpoint corrosion, corrosion by
scrubbing waters may involve a wide variety of corrosive agents,
including sulfuric acid, hydrochloric acid, hydrofluoric acid,
nitric acid, organic acids, metallic salts, and others. The
presence of these agents causes many scrubbing liquors to have a
very low pH level and/or a high level of dissolved solids, both
of which tend to cause corrosion in metals. Stainless steels
and high nickel alloys, which are resistant to corrosion in many
acid solutions, can suffer corrosion in scrubber service because
12
-------
of the synergistic effects of chlorides or fluorides with acids
in the liquors.
Corrosion by scrubbing waters can manifest itself in a
variety of forms. The most common of these are listed below.
Uniform Attack
Uniform attack commonly occurs when mild steel is exposed
to an acid, such as sulfuric acid. Uniform attack is not limi-
ted to mild steels, however, and it can be caused by a variety
of corrosive agents.
Pitting
Pitting is a localized form of corrosion that results in
holes in the surface of the exposed metal. After pit initia-
tion, rapid perforation of the metal can occur although sur-
rounding surfaces remain relatively unaffected by corrosion
(Figure 6). Most pitting of scrubber components results because
of high levels of chlorides in the scrubbing liquors.7 The
Source 1 report describes a case of severe pitting in the stain-
less steel walls of a scrubber vessel which resulted because of
chlorides in the scrubbing liquor.
Figure 6. Pitting of stainless steel in a
municipal incinerator scrubber vessel.
13
-------
Crevice Corrosion
Crevice corrosion is the localized corrosion of a metal
surface within crevices or other shielded areas exposed to cor-
rosive liquids. The corrosion rate is often accelerated in a
crevice because the concentrations of oxygen and metallic ions
in the relatively stagnant fluid within.the crevice can differ
greatly from concentrations in the more turbulent fluid outside
the crevice. Potential sites for crevice corrosion in scrubbers
include bolts and other mechanical fasteners, incompletely
welded seams (e.g., skip welds and one-sided butt welds), and
holes or tears in coatings and linings. The Source 1 and
Source 2 reports describe corrosion in scrubbing liquor recircu-
lation pumps which occurred in crevices under tears in their
linings.
Stress Corrosion
Stress corrosion is a particularly serious form of deterio-
ration caused by the simultaneous effects of tensile stress and
corrosion (Figure 7). Various metals can suffer the effects of
stress corrosion cracking dependent on the corrosive agents pre-
sent. Austenitic stainless steels are most likely to suffer
stress corrosion cracking in chloride environments. The
Source 1 and Source 2 reports describe stress corrosion cracking
of high nickel alloy fans serving municipal incinerator scrub-
bers.
Weld Decay
Weld decay is the corrosion of a normally resistant metal
in the vicinity of a weld (Figure 8). Weld decay occurs in
stainless steels because the heat involved in the welding pro-
cess can cause chromium in the alloy to precipitate with carbon
to form insoluble carbide particles within the affected areas of
the metal. The lowered chromium content of the metal remaining
in these sensitized areas reduces the metal's resistance to
corrosion. Examples of weld decay are described in the Source 1
and Source 3 reports.
14
-------
Figure 7. Stress corrosion cracking of stainless steel
ductwork of a sewage sludge incinerator scrubber.
Figure 8. Weld decay in stainless steel ductwork of
a sewage sludge incinerator scrubber; outleakage
has caused additional corrosion on the exterior
surfaces.
15
-------
EROSION-CORROSION
Erosion-corrosion is the third mode of corrosion commonly
found in participate control equipment. Erosion-corrosion is an
acceleration of corrosive attack due to the removal of otherwise
protective oxide films by suspended abrasive particulate. By
this removal, fresh metal is continually exposed to the cor-
rosive medium. Erosion-corrosion is most likely to occur at
points where flue gases, sprays, or scrubbing liquids must make
sudden changes in direction, e.g., at elbows, at the inlet to
cyclonic mist eliminators (Figure 9), and at blast plates in
fabric filters. Erosion-corrosion also can occur in the turbu-
lent zones of venturi scrubbers and in fans (Figure 10). Source
reports 2, 15, and 17 describe erosion-corrosion failures in
scrubber vessels and Source reports 5, 7, and 16 describe abra-
sive failures of fans.
Figure 9. Erosion-corrosion at the inlet of a cyclonic
mist eliminator of a venturi scrubber serving a
hot-mix asphalt plant.
16
-------
Figure 10. Scouring and erosion of a fan blade of an
induced-draft fan serving an industrial boiler fabric filter.
17
-------
OTHER MODES OF CORROSION
Other modes of corrosion include high-temperature corrosion
and galvanic corrosion. Although these modes do not affect par-
ticulate control equipment as frequently as the three just dis-
cussed, they can become significant problems in some situations.
High-Temperature Corrosion
At temperatures over 325°C the importance of water in
corrosion is overshadowed by other agents, such as molten salts,
lead, vanadium, sodium, oxygen, sulfur dioxide, and hydrogen
sulfide. At elevated temperatures corrosion can be extremely
rapid and sometimes catastrophic. Although high-temperature
corrosion does not often affect particulate control equipment,
it can affect hot ductwork, dampers, quench chambers, and heat
exchangers upstream from the particulate control equipment. The
Source 6 report describes a case of high-temperature corrosion
in the ductwork and dampers leading to an ESP.
Galvanic Corrosion
Galvanic corrosion is caused by the coupling of electro-
chemically dissimilar metals in an electrolyte. In a galvanic
couple, the more anodic metal will corrode, discharging positive
metal ions into the surrounding electrolytic solution. These
ions typically combine with dissolved oxygen or ionic species in
solution, such as hydroxyl ions, to form corrosion products.
Electrons from the anodic metal are transmitted through the
couple to the cathode where they participate in the various
cathodic reactions. The cathodic metal in a galvanic couple
generally does not corrode while the anodic metal tends to
corrode at a faster rate than it would if not part of the cou-
ple. Table 1 contains a list of several common metals arranged
according to their galvanic activity in seawater. The farther
apart two metals lie in the galvanic series, the more severe
will be the corrosion of the anodic metal., Common galvanic
couples which result in corrosion include (the anodic or corrod-
ing metal is listed first) mild steel/stainless steel, mild
18
-------
steel/copper, aluminum/copper, aluminum/mild steel, and alumi-
nun/stainless steel. These couples can sometimes be inadver-
tently constructed in piping systems with their many valves and
fittings, and in mechanically fastened metal parts such as
bolted joints, riveted joints, and access hatch hardware.
TABLE 1. GALVANIC SERIES OF SOME COMMON
METALS IN SEAWATER8
Active or anodic
Noble or cathodic
Magnesium alloys
Zinc
Aluminum
Mild steel
Cast iron
Lead
Tin
Muntz metal
Manganese bronze
Naval brass
Admiralty brass
Copper
Nickel
76 Ni-16Cr-7 Fe
Type 410 stainless steel (passive)
Titanium
Type 304 stainless steel (passive)
Silver
Gold
Platinum
Note: The order of various metals in the galvanic series may
be different in media other than seawater.
19
-------
SECTION 4
FACTORS THAT CONTRIBUTE TO CORROSION IN CONTROL EQUIPMENT
Many factors contribute to the formation of corrosive
conditions in participate control equipment. If corrosion is to
be prevented, it is important to recognize these factors and to
understand how they are interrelated. The first part of this
section is a discussion of factors which are inherent in the
process being controlled, and the second part is a discussion of
factors which vary according to operation and maintenance prac-
tices at a plant.
PROCESS CHARACTERISTICS
Many of the factors that determine the corrosivity of the
environment affecting a particulate control device are a func-
tion of the process being controlled. Of prime importance is
the chemical nature of the service environment. Also important
are the temperature and moisture content of the flue gases, the
operating schedule of the controlled process, and any process
changes that are made after particulate control equipment is
installed.
Chemical Constituents
Corrosive chemicals can enter the particulate control
system from several sources including the process raw materials,
the process fuels, combustion and dilution air, quench waters,
and scrubbing waters. Although some of these constituents may
be noncorrosive in their original forms or concentrations, they
can be altered chemically or concentrated by various industrial
processes. Some chemicals that are noncorrosive by themselves
may have synergistic corrosion effects with other chemicals.
20
-------
The two most common corrosives that affect particulate con-
trol equipment are sulfuric acid and chlorides. Sulfuric acid
is a contributing factor in most corrosion problems affecting
fabric filters and electrostatic precipitators. Observations
made during the field inspections of this study and verified by
the scrubber manufacturers interviewed, indicate that chlorides
are responsible for the majority of corrosion problems in scrub-
ber applications. Chlorides in combination with acidic scrub-
bing waters are especially troublesome.
Raw Materials-
Process raw materials are a primary source of corrosive
contaminants in particulate-laden gases entering particulate
control devices. Ores, clays, sands, aggregates, recycled scrap
metals, plastics, slags, and other raw materials may contain a
variety of corrosive secondary components in addition to their
primary, economically valued components. For example, many ores
(e.g., copper and iron) contain high percentages of sulfur, some
of which can be oxidized to sulfur trioxide during refining pro-
cesses. Crushed aggregates used in asphalt production, es-
pecially those mined in coastal areas, sometimes contain chlo-
rides .
Recycled scrap metals used in secondary metals industries
can contain a variety of corrosive contaminants. Feedstocks for
secondary lead furnaces consist predominantly of lead-acid bat-
teries, which contain sulfuric acid and chloride-bearing plast-
ics. Feedstocks for the secondary aluminum and secondary copper
industries can contain chlorinated rubbers and plastics. Feed-
stocks for cast iron cupolas typically contain a large percent-
age of internal combustion engine parts, which contain oil,
lead, and paint residues. These generate a variety of flue gas
contaminants.
Steel mill slags, which are sometimes used as raw materials
for rockwool insulation cupolas, crushed aggregate dryers, and
other processes, can contain sulfur. Finally, clays used in
21
-------
brick production, phosphate used in fertilizer production, and
feldspar used in glass production often contain fluorides.
Fuels—
Fossil fuels, especially coal, metallurgical coke, and
residual fuel oils, contain significant quantities of sulfur
which can oxidize to sulfur trioxide during combustion. When
such fuels are used in a process controlled by a dry particulate
collection device, acid dewpoint corrosion may occur. In in-
stallations where wet scrubbers are used to control processes
using fossil fuels, it is common for acidic conditions to devel-
op in the scrubbing liquors.
Some fuels can introduce corrosive contaminants other than
sulfur into the particulate control device. Polyvinyl chloride
plastics and chlorinated rubbers, commonly burned in municipal
incinerators and in wire burning incinerators, can generate
hydrochloric acid. Some coals are an overlooked source of
chlorides, in addition to being a source of sulfur.9
Scrubbing Liquors—
Scrubbing liquors can be a source of corrosive contaminants
as well as a medium in which contaminants from flue gases can
collect. Recycled scrubbing waters can be especially trouble-
some because recycling can increase the concentrations of corro-
sive contaminants by several orders of magnitude. Chlorides,
acids, and other electrolytic species found in benign concentra-
tions in the feed water or the flue gases can become quite cor-
rosive as a result of the concentrating effects of recycling.
Flue Gas Temperatures and Moisture Content
Flue gas temperatures can have a great effect on corrosion
and other forms of materials degradation in particulate control
devices. Extremely high temperatures in combination with cor-
rosive contaminants can result in high-temperature corrosion in
ductwork and other components. High temperatures can also
damage fabric filter bags, protective coatings and linings, and
fiberglass-reinforced plastic components.
22
-------
Gas temperatures can be reduced by dilution air dampers,
quench chambers, radiant cooling loops, or heat exchangers.
Reducing gas temperatures, however, can set the stage for acid
condensation. Therefore, when hot flue gases are to be cooled,
it is important to determine their moisture and sulfur trioxide
contents so that the sulfuric acid dewpoint can be estimated.
All possible sources of flue gas water vapor should be consid-
ered, including the water vapor added with dilution air or
evaporative cooling and that which forms as a byproduct of
fossil fuel combustion. If analysis indicates that temperatures
will fall below the acid dewpoint it is advisable to reduce the
amounts of water vapor and/or sulfur in the flue gases or to
select an alternative particulate control strategy.
Process Operating Cycles
The operating cycles of a controlled process can sometimes
influence the corrosion activity in a particulate control de-
vice. Operating cycles can affect dewpoint corrosion, because
of the close relationship between process variations to varia-
tions in flue gas temperatures. In general, the more frequently
a hot process starts up and shuts down or varies between full
production and partial production, the more frequently the flue
gases in the control device will pass through the acid dewpoint.
A number of operating cycles are possible in industrial
processes, and they produce a variety of flue gas temperature
cycles. At one extreme is the characteristic cycle of a cast
iron cupola, which typically operates for melts lasting less
than 24 hours between startup and shutdown. While a cupola melt
is in progress, there are frequent charges of fuel, scrap iron,
and fluxing; periodic blasts of combustion air through the
tuyeres; and frequent iron and slag taps. Each of these opera-
tions effects the temperature of the cupola flue gases. Wide
fluctuations in cupola flue gas temperatures can be tempered by
gas cooling devices such as dilution air dampers, quench cham-
bers, and heat exchangers. It is usually difficult with these
23
-------
devices, however, to maintain sufficient control of flue gas
temperature to simultaneously protect the particulate control
equipment from high temperature stresses and acid dewpoint
corrosion. At most cupola installations, high temperature
protection has priority over dewpoint control because a single
high-temperature excursion can do immediate damage whereas
dewpoint corrosion is a cumulative problem.
The other extreme in operating cycles is represented by the
base-loaded utility boiler or a fully operational cement kiln
which run for months at a time near their maximum firing rates.
Industrial boilers fall somewhere between the two extremes in
operation. These usually operate for periods lasting 5 to 6
days in length and can vary considerably in their firing rates
at various times during a day.
Operating cycles can affect more than flue gas temperatures
in some processes. For example, in iron and steel processes
such as an oil-fired open hearth furnace (Source 8), the addi-
tion of lime fluxing can add a temporary burst of alkaline
particulate to the flue gases that can raise the pH of scrubbing
liquors. During periods between lime additions, sulfur oxides
from the fuel combustion can lower scrubber pH drastically.
These fluctuations in flue gas composition must be considered
when designing a scrubber liquor pH control system. Similarly,
the variability of refuse entering a municipal incinerator can
affect the corrosivity of the flue gases. Residential refuse
charged to an incinerator after a rainy period is likely to
produce flue gases with a higher moisture content. Charging an
incinerator with refuse that contains high percentages of poly-
vinyl chloride plastic is likely to produce flue gases with
higher levels of hydrochloric acid.
The significance of the operating cycle of a process must
be considered when developing a corrosion control strategy for
particulate control equipment. Variability in flue gas tempera-
tures and chemical constituents may be a determining factor in
the type of cooling device chosen, the sophistication required
24
-------
in a scrubber pH control system, or the type of particulate
control device selected.
OPERATION AND MAINTENANCE PRACTICES
Operation and maintenance practices often affect corrosion
in particulate control equipment. Operating parameters that in-
fluence corrosion are the control of flue gas temperatures, the
variability of raw materials and fuels, the frequency of pH ad-
justments to scrubbing liquors, and scrubbing liquor blowdown
and makeup rates.
In attempting to control the sulfuric acid dewpoint, it is
often advisable to preheat a cold fabric filter or electrostatic
precipitator before introducing moist, sulfur-bearing flue
gases.10 For processes having wide fluctuations in tempera-
tures, it is advisable to vary the amount of dilution air or
quench spray to match the changing temperatures, rather than
presetting the amount of cooling for the highest expected temp-
erature. Automatic cooling controls are preferable to manual
controls. In processes such as cement kilns or hot-mix asphalt
dryers, operators must avoid the temptation to lower process
exhaust temperatures to save fuel unless they have also consid-
ered the effects that lower temperatures may have on dewpoint
corrosion.
Changes in raw materials sometimes alter corrosion activity
in a particulate control device. For example, a change in fuel
from distillate oil to residual oil or coal can increase the
sulfur levels in the gases. A change in the source of coal can
change the amounts of sulfur and chlorides introduced into the
system. Some clays used in brick and refractory production
contain significant quantities of fluorides, whereas others do
not. Likewise, some aggregates used in hot mix asphalt produc-
tion contain high levels of chlorides while most do not. Final-
ly, refuse containing high levels of plastics and chlorinated
rubber will produce hydrochloric acid when burned in a municipal
25
-------
incinerator, whereas refuse mixtures without these materials
will not.
Management of scrubbing waters is also important for con-
trol of corrosion activity. In some scrubber systems pH control
is critical. Continuous pH control is usually superior to
periodic adjustments, but it requires good pH instrumentation
and frequent operator monitoring. One scrubber manufacturer
interviewed during this study reported that as many as 50 per-
cent of the pH control systems for particulate scrubbers are not
used at all because of the attention they require from the
operators. The potential impact of an increase in the liquor
recycle rate is also an important consideration. Increasing
recycle rates increases the concentrations of acids, and other
corrosive constituents.
Maintenance practices that can affect corrosion activity in
particulate control equipment include the frequency at which
various components are inspected and the speed at which malfunc-
tioning components are repaired. Among the items which should
be inspected frequently are pH control instrumentation (es-
pecially the pH probes), temperature control instrumentation,
hatches, hoppers, fans and fan vibration sleeves, insulation,
and mist eliminators.
26
-------
SECTION 5
AVOIDING CORROSION IN CONTROL EQUIPMENT
Corrosion in particulate control equipment can be addressed
in most applications through careful design and through the use
of proper operation and maintenance procedures in both the con-
trol equipment and the controlled process. This section de-
scribes the most important of these tools for avoiding control
equipment corrosion.
CONTROLLING ACID DEWPOINT CORROSION
Most of the manufacturers interviewed during this study
agreed that acid dewpoint corrosion can be adequately controlled
in nearly all fabric filter and electrostatic precipitator
applications. The principal tools available for controlling
acid dewpoint corrosion are reduction of sulfur trioxide and
moisture levels in the flue gases, control of exhaust gas temp-
eratures, use of proper thermal insulation, and prevention of
cool air inleakage into the control device. In some cases the
use of protective coatings and special alloys can be helpful;
however, the cost of the latter is usually prohibitive for the
major structural components of fabric filters and electrostatic
precipitators.
Reduction of Sulfur Trioxide and Moisture Levels
Assuming that there is no moisture inleakage into a control
device, the levels of sulfur trioxide and moisture in the flue
gases are dependent on process characteristics. The best way to
keep the sulfuric acid dewpoint temperature low is to select
fuels with as low a sulfur content as possible and to select
dry-type flue gas cooling systems whenever possible.
27
-------
It should be noted that the use of low sulfur fuels is no
panacea for acid dewpoint corrosion, however. Low sulfur coals
and residual oils can generate sufficient sulfur oxides when
burned to cause acid dewpoint corrosion; the use of these fuels
merely lowers the dewpoint temperature which makes acid conden-
sation easier to prevent. Economics usually governs the ulti-
mate decision on the type of fuel to be used. Any cost analysis
of available fuel options should consider the possible costs of
corrosion damage to the particulate control equipment and lost
production following corrosion failures as well as the relative
costs of the fuels.
The types of systems available for cooling flue gases are
evaporative cooling (i.e., quench chambers), dilution air, and
heat exchangers. Evaporative cooling injects considerable mois-
ture into the flue gases. Dilution air cooling also can intro-
duce moisture into flue gas streams, depending on the relative
humidity of the ambient air. From the standpoint of corrosion,
heat exchangers are the preferred cooling method because they
add no moisture to the flue gases. As with fuel, economics
plays a major role in the selection of a flue gas cooling sys-
tem. Dilution air systems usually entail the lowest capital
investment, whereas heat exchangers usually entail the largest
capital investment. Part of the costs of heat exchanger systems
often can be recovered by reuse of the waste heat removed.
Recuperative heat exchangers can be used to preheat combustion
air or for other energy conservation purposes. Installed costs
of evaporative cooling systems usually fall somewhere between
those of dilution air systems and heat exchanger systems; how-
ever, they also require a supply of water.
Control of Flue Gas Temperatures
Flue gas temperatures are a function of the process char-
acteristics and of the flue gas cooling systems. Processes such
as cast iron production are not ammenable to flue gas tempera-
ture control at the process. Sources such as these require the
28
-------
use of flue gas cooling devices. Primitive systems such as a
manually operated dilution air damper are usually not sufficient
to insure that flue gases remain above the dewpoint. Control
devices with temperature sensors to automatically control dilu-
tion air are preferable. Evaporative cooling is not recommended
for cast iron cupolas, but recuperative heat exchangers have
been used successfully.
In processes such as cement kilns, where operating condi-
tions are relatively stable, it is sometimes possible to control
exhaust temperatures by adjusting the firing rates and the
amounts of excess combustion air. In such processes, it is
helpful to have temperature sensors in the control device which
can automatically control the process firing rates or sound
alarms in the plant's control room whenever temperatures in the
control device fall below a safe level. The increase in fuel
consumption required to maintain high kiln exhaust temperatures
must be weighed against the savings attributable to reduced
corrosion in the particulate control system.
Another important means of controlling acid dewpoint corro-
sion is to preheat the control device prior to the introduction
of corrosive flue gases.10 For example, a rotary dryer can be
preheated by firing the burner at a reduced rate for a period of
time before introducing the wet aggregate. A cupola can be pre-
heated by firing natural gas afterburners for a period of time
prior to lighting the sulfur-bearing coke charge. The advantage
of preheating is that the hot corrosive flue gases will not con-
tact cold steel surfaces in the control device during startup as
would occur without preheating. As in the case of other flue
gas temperature control measures, the advantage of preheating
must be weighed against the additional fuel expenses.
Thermal Insulation
Thermal insulation is another means of maintaining flue gas
temperatures above the acid dewpoint. The ultimate purpose of
insulation is to reduce the amount of flue gas cooling that
29
-------
takes place between the time the gases enter the ducting to the
control device and the time they exit the stack. At some in-
stallations, insulation alone will solve acid dewpoint corrosion
problems; in others it is only a supplementary measure.
Insulation will be most effective if a few simple guide-
lines are followed.11 Any insulation must be thick enough to
sufficiently impede the flow of heat through the external sur-
faces of the control device. The insulation must also cover all
exposed surfaces including the dust hoppers, access doors, and
the ductwork. The insulation must be installed in such a manner
that it will remain intact over the life of the equipment. Most
types of insulation should have a protective sheath (usually
sheet metal) on the outside to repell rainwater and to prevent
mechanical damage. An insulation blanket should be sealed at
the top and bottom to prevent a "chimney effect" circulation of
outside air through or underneath the blanket. An example of a
properly insulated electrostatic precipitator is shown in
Figure 11.
Figure 11. Electrostatic precipitator with a complete blanket of
insulation covering the chambers, hoppers, and ductwork.
30
-------
Among the common errors made in insulating control equip-
ment are the failure to insulate all exposed ductwork (including
reverse air ductwork for the cleaning of fabric filter bags),
the failure to insulate hatches, and the failure to attach the
insulation securely. Another common error (Figure 12) is to
allow structural members to protrude through the blanket of
insulation; these members radiate heat away from interior sur-
faces of the control device and promote condensation at these
locations. Figure 13 shows improper sheathing of the insulation
surrounding a duct. The gap in the sheathing can channel rain-
water into the insulation and cause the exterior surface of the
duct to remain moist.
Air Inleakage
Unexpected air inleakage can spoil a well-designed acid
dewpoint control program in a fabric filter or an electrostatic
precipitator. Inleakage allows not only moisture from the
ambient air to enter the control system, but also causes severe
localized cooling at the point of entry. Both phenomena can
contribute to acid dewpoint corrosion. Inleakage is a self-
feeding process—the metal adjacent to the initial penetration
tends to corrode at an accelerated rate, which promotes en-
largement of the penetration. Inleakage is common in poorly
fitting hatches (Figure 5), hatches with worn seal gaskets,
rotary air locks with worn seals, cracked vibration sleeves
(Figure 14), and incomplete welds. It can also occur in duct-
work or at the process. Inleakage should be repaired promptly
to avoid further damage.
Special Alloys and Protective Coatings
The relatively large size of most fabric filters and elec-
trostatic precipitators makes it uneconomical to use higher
priced stainless steels or nickel alloys as the principal mater-
ials of construction. Most structural components of these
devices are constructed of carbon steel or, in some cases, Cor-
ten or cement block. Special alloys are used, however, for some
31
-------
it-
12. Insulated fabric filter with structural steel
protruding through the insulation.
32
-------
Figure 13. Gap in the sheathing covering an
insulated duct.
Figure 14. Cracked vibration sleeve in ductwork, which
allows air inleakage into the system.
33
-------
of the smaller components within these ducts, such as filter bag
hardware and electrostatic precipitator discharge wires.
Nearly all exterior surfaces of fabric filter and electro-
static precipitator enclosures receive a protective coating.
However, the interior surfaces of these devices are not always
coated. One fabric filter manufacturer reported that only 10
percent of its filters receive internal coatings. The reason
that coatings are not often used is that few seem to be very
effective on hot sources. Despite claims of some coatings
manufacturers, long-term performance of most available coatings
systems is not good in hot corrosive flue gases. All control
device manufacturers interviewed reported that most coatings
fail when exposed for prolonged periods to acid bearing flue
gases above 300 to 400°F. In cases where a coating is appropri-
ate, it is important that the coating be applied in the proper
manner to a properly prepared surface. In most cases, this
requires that the coating be applied in the shop rather than in
the field.
CONTROLLING SCRUBBER CORROSION
Corrosion problems arise frequently in scrubbers because
exhausts from many industrial processes contain gases and par-
ticulates that are corrosive when dissolved in water. The
combustion of fossil fuels such as coal, metallurgical coke, and
fuel oils can produce sulfuric acid in scrubbing liquors. Coals
can also contain sufficient chlorides to cause scrubbing liquors
to become corrosive to many alloy steels.9 Combustion of poly-
vinyl chloride plastics, commonly found in incinerator feed-
stock, also yields chlorides. Rotary aggregate dryers can
contribute chloride- and fluoride-bearing particulate to ex-
hausts. Fluoride-bearing particulate is especially troublesome
in the phosphate fertilizer and the feldspar industries.
Prevention of corrosion in scrubbers is best handled
through proper choice of materials and through scrubbing liquor
34
-------
pH control. Recirculation of scrubbing liquors greatly in-
creases the concentrations of any corrosive agents they contain.
Reduction or elimination of liquor blowdown can increase the
concentrations of corrosive constituents in the liquor that
otherwise acceptable materials of construction may become sus-
ceptible to corrosion.
Selection of Materials for Scrubbers
Materials of construction for scrubber applications must be
carefully selected to withstand corrosive and abrasive agents in
the gases and liquors and any high temperatures that are likely
to occur. Most manufacturers interviewed contended that if the
process conditions are properly defined before design begins,
the experienced scrubber manufacturer can design a scrubber that
will withstand its service environment. Many scrubbers fail be-
cause inappropriate materials are selected after a superficial
investigation of process conditions, or because less resistant
materials are substituted to reduce costs.
Investigation of each scrubber application should include
chemical analysis of the raw materials, combustion products, and
scrubbing liquids and a review of the operating histories of
scrubber installations in similar applications. Finally, review
of literature about materials performance is recommended, and
when materials performance data are not available, in situ
coupon tests may be required. After all relevant information
has been compiled, the designer should prepare a list of materi-
als suitable for the expected service. Selection of materials
of construction from this list will be based in part on their
relative costs.
Although the above-mentioned procedures should always be
followed in the selection of materials, some general aspects of
materials applications can be summarized here. Table 2 lists
the major categories of materials available for scrubbers and
ancilliary components, together with their major properties,
general corrosion behavior, and relative costs.
35
-------
TABLE 2. PROPERTIES OF MATERIALS USED IN THE CONSTRUCTION OF WET SCRUBBERS AND AUXILIARY.COMPONENTS7••
8.12
Material
Properties
Corrosion resistance
Uses
Cast iron
Carbon steel
Martensitic
stainless
steels
(410, 416,
420, 440)
Ferritic
stainless
steels
(405, 430,
442, 446)
Austenitic
stainless
steels
(201, 202,
301, 302,
304, 310,
316, 317,
alloy 20)
High strength; low ductility;
brittle; hard; low cost
Good strength, ductility, and
workability; low cost
Iron-chromium alloy, hardenable by
heat treatment; costs 2 to 5 times
more than carbon steel
Iron-chromium alloy, not hardenable
by heat treatment; costs 2 to 4
times more than carbon steel
Iron-chromium-nickel alloy; not
hardenable by heat; hardenable
by cold working; nonmagnetic; cost
3 to 10 times more than carbon
steel; alloys with "L" designation
(e.g., 316L) have lower carbon con-
tent for improved weldability
Gray or white cast irons exhibit
fair resistance to mildly corro-
sive environments; high-silicon
cast irons exhibit excellent re-
sistance in a variety of environ-
ments (hydroflouric acid is an
important exception); cast irons
are susceptible to galvanic cor-
rosion when coupled to copper al-
loys, stainless steels
Fair to poor in many environments;
low pH and/or high dissolved
solids in moist or immersion
service leads to corrosion;
properly applied protective
coatings give appropriate pro-
tection in many applications;
susceptible to galvanic corro-
sion when coupled to copper
alloys, stainless steels
Good
iood; better than martensitic
stainless steels; resists stress
corrosion caused by chlorides;
better chloride resistance than
austenitic stainless steels; sus-
ceptible to crevice corrosion and
pitting
iood in oxidizing environments;
fair in non-oxidizing environ-
ments; susceptible to pitting
and stress corrosion cracking in
chloride solutions. Type 310 re-
sists high temperature corrosion;
types 316 and 317 contain molyb-
denum for better chloride and pit-
ting resistance; alloy 20 has ex-
cellent resistance in many envir-
onments
Pump casings, valve casings, piping;
often used with linings in corrosive
service
General purpose in noncorrosive en-
vironments
Typically used for machine parts; rarely
used in scrubber applications
General purpose; certain alloys good for
high-temperature service; not as commonly
used in scrubber applications as austen-
itic stainless steels
Scrubber vessels; fans; stacks and
ductwork; mist eliminators; quench
chambers
(continued)
-------
Table 2 (continued)
Material
Nickel alloys
(Inconel ,
Incoloy,
Monel,* .
Hastalloys -
Chlorimet,
and others)
Titanium
Alumi nun
al1oys
Fiberglass-
reinforced
plastics
(FRP)
Wood
Properties
Good strength; costs more than 10
times as much as carbon steel;
also expensive to fabricate; com-
monly alloyed with chromium, iron,
or copper
High strength; light weight (60%
that of steel); costs at least 10
times more than carbon steel
Good strength-to-weight ratio; good
electrical and thermal conductiv-
ity
Good chemical resistance; poor
abrasion resistance; cannot be
used in high-temperature service;
low hardness
High tensile and shear strength
perpendicular to grain; low ten-
sile and shear strength parallel
to grain; low hardness; poor abra-
sion resistance; cannot be used
for high-temperature service
Corrosion resistance
Excellent resistance in most en-
vironments; not resistant i"n
strong oxidizing solutions such
as ammonium and HN02; most have
good resistance to stress corro-
sion; some nickel-copper alloys
have good resistance to hydro-
fluoric acid
Exceptional resistance at ambient
temperatures; excellent resist-
ance at other temperatures, ex-
cept that crevice corrosion is
possible in chloride solutions
above 250°F
Good atmospheric corrosion re-
sistance (except in salt mist
exposure for certain alloys);
some alloys suitable for immer-
sion in certain acids; suscept-
ible to galvanic corrosion when
coupled to most other metals
Excellent in many corrosive en-
vironments; actual results de-
pend on type of plastic resin
used
Good resistance in dilute acids
(including hydrofluoric acid);
susceptible to biological at-
tack under certain conditions;
deteriorates in alkaline solu-
tions
Uses
Scrubber vessels, fans; stacks and
ductwork; mist eliminators
Not widely used in scrubber service
because of costs; sometimes used in
severe environments in fans and other
components
Not widely used in scrubber service
Scrubber vessels, piping, mist
eliminators, ductwork and stacks
Scrubber vessels, tanks,
especially in fluoride expo-
sure; fir and cypress are popular
species
a Registered trademark of Huntington Alloys, Inc.
Registered trademark of the Stellite Division of the Cabot Corporation
Registered trademark of the Duriron Company, Inc.
-------
The success of protective coatings and linings in scrubber
applications has been variable. Several manufacturers reported
successful use of linings in certain applications, including the
use of rubber linings in coal-fired utility boiler scrubbers and
flaked-glass linings in hot ductwork. Rubber linings are also
common in scrubber liquor pumps. Other protective coatings used
in some scrubber applications include various epoxies and fiber-
glass reinforced plastics (FRP's).
Difficulty of application and repair are inherent problems
in the use of coatings and linings. Proper application of coat-
ings and liners requires a well prepared metal substrate and
good application techniques. A poor anchor pattern after sand-
blasting or the presence of oil, dirt, or condensed water on the
surface can cause a coating to disbond. Improper formulation or
mixing of a coating in the liquid state, the inclusion of air
bubbles during application, or improper curing after application
can also cause coating failure. Finally, coatings are suscepti-
ble to mechanical damage and repair of improperly applied or
damaged coatings is usually difficult. One vendor interviewed
prefers to select resistant materials of construction rather
than to resort to coatings and linings. The same manufacturer
indicated that solid FRP presents fewer problems than FRP-lined
steel.
Control of Scrubber Liquor pH
The neutralization of extremely acidic scrubbing liquors
can often reduce the need for expensive materials of construc-
tion in scrubbers. For example, mild steels which fail when
exposed to low pH solutions are less likely to fail in neutral
solutions. Likewise, many forms of chloride attack in stainless
steels occur only under low pH conditions. Control of pH,
therefore, can substantially reduce the costs of materials in
scrubbers and lengthen the useful life of many components.
Systems for the control of pH have not always proved to be
reliable, however. Most scrubber manufacturers admitted that
38
-------
the majority of scrubber pH control systems do not perform
adequately. In most cases the problems are a lack of technical
expertise on the part of the operators and a lack of regular
maintenance. In some installations delivered pH control systems
are not used because the operators are unable or unwilling to
perform the required maintenance. One scrubber manufacturer
insisted that the pH control systems can work if properly de-
signed and reported that some systems have operated for as long
as six months without calibration. In contrast, other pH con-
trol systems (see Sources 1 and 8, Appendix A) require calibra-
tion as often as once per shift and still do not provide entire-
ly satisfactory service.
A pH control system that is to be the principal defense
against corrosion will usually require maintenance at frequent
intervals, especially at the pH electrodes. Maintaining clean,
calibrated probes is often difficult, but calibrated probes are
vital to the success of a pH control system. Ultrasonically
cleaned probes usually provide greater reliability than ordinary
probes. To assure a representative measurement, the pH probes
must be located at carefully selected points in the system.
Probes located in stagnant areas may provide unrepresentative
data, whereas probes subjected to high turbulence or fouling
conditions may require frequent replacement.13 The required
frequency of chemical addition to scrubbing liquors will depend
on the variability of the flue gas acidity, the detention time
of the liquor in the treatment and recirculation systems, and
the type of chemicals used for pH adjustment. In highly varia-
ble processes such as open-hearth furnaces and cupolas, a large
settling basin can help dampen sudden changes in liquor pH, as
will the addition of chemicals with a high buffering capacity,
such as lime.
Abrasion in Scrubbers
Abrasion can occur when gases or scrubbing liquors contain-
ing high concentrations of abrasive particulate are in the
39
-------
turbulent mode or are subjected to a sudden change in flow
direction. Typical wear areas in scrubbing systems include
venturi throats, walls of centrifugal mist collectors near the
inlet duct (Figure 9), and elbows in the ductwork.14 One manu-
facturer cited the wet-dry line in scrubbers serving hot sources
as particularly susceptible to corrosion and abrasion. Solu-
tions to abrasion wear include the use of precleaning devices
and the use of large-radius turns in ductwork.
Rotating equipment such as fans, pumps, and clarifiers must
receive special attention in scrubber service because of poten-
tial abrasion, plugging, and corrosion. Key wear areas in these
components are the bearings and any components rotating in the
fluid stream.15
Fan wear is a commom problem. Forced-draft fans often
suffer abrasion because of exposure to particulate-laden gases.
Wear in forced-draft fans can be reduced by use of special
wear-resistant alloys, by reduction of fan rotation speeds
(i.e., installing a larger fan), or by moving the fan to an
induced-draft location on the clean-air side of the scrubber
system. Blades of induced-draft fans can suffer corrosion or
solids buildup if mist is carried over from the liquid entrain-
ment separator. Induced-draft fan problems can be addressed by
use of corrosion-resistant materials and by improving liquid
entrainment separation.
Pump wear is also a common problem in scrubber systems.
Pump housings, impellers, and seals are subject to abrasion and
corrosion by scrubber slurries. The use of rubber linings and
special alloy pump materials will reduce abrasion and corrosion
of the housings or impellers, and installation of a water flush
in the seals can help reduce wear of the seals.15
40
-------
REFERENCES
1. Booz, Allen, and Hamilton, Inc. Final Evaluation of North
Carolina's Program to Regulate Air Pollution From Station-
ary Sources. Prepared under President's Council on Envir-
onmental Quality, Contract No. EQ8AC015, Bethesda, Mary-
land, July 1, 1979.
2. PEDCo Environmental, Inc. Unpublished data obtained during
SIP audit inspections under contract to U.S. EPA, Division
of Stationary Source Enforcement, Contract No. 68-01-4147
Tasks 110, 131, and 137, 1979-80.
3. Balasic, P. J. Electrostatic Precipitator Corrosion Prob-
lems on Recovery Boiler Applications. Paper No. 185,
presented at the National Association of Corrosion Engi-
neers, Corrosion/79, Atlanta, Georgia, March 12-16, 1979.
4. Pierce, R. P. Estimating Acid Dewpoints in Stack Gases.
Chemical Engineering 84(4): 125-28, 1977.
5. Verhoff, F. H., and Banchero, J. T. Chemical Engineering
Progress, Vol. 70, p. 71, 1974.
6. Katz, J. The Art of Electrostatic Precipitation, Chap-
ter 7. Precipitator Technology, Inc., Pittsburgh, Pennsyl-
vania, 1979.
7. Fontana, M. G. and N. D. Greene. Corrosion Engineering,
2nd ed., McGraw-Hill, New York, New York, 1978.
8. Brasunas, A. des., and N. E. Hamner (eds.). NACE Basic
Corrosion Course, National Association of Corrosion Engi-
neers, Houston, Texas, 1977.
9. lapalucci, T. L., R. J. Demski, and D. Bienstock. Chlorine
in Coal Combustion, Report 7260, U.S. Department of Interi-
or, Bureau of Mines, 1969.
10. Beggs, T. W., and U. M. Patankar. Accelerated Baghouse
Corrosion in a Waste Oil Burning Asphalt Concrete Plant
Presented at the 72nd annual meeting of the Air Pollution
Control Association, Cincinnati, Ohio, June 24-29, 1979.
11. Landrum, R. J. Equipment. Chemical Engineering 77(22)-
75-82, October 12, 1970.
12. Benzer, W. C. Steels. Chemical Engineering 77(22), Octo-
ber 12, 1970.
41
-------
13. PEDCo Environmental, Inc. Lime FGD Systems Data Book.
Electric Power Research Institute, EPRI-FP 1030, Palo Alto,
California, May 1979.
14. Schifftner, K. C. Venturi Scrubber Operation and Mainte-
nance. Presented at the U.S. EPA Environmental Research
Information Center Seminar on Operation and Maintenance of
Air Pollution Equipment for Particulate Control, Atlanta,
Georgia, April 1979.
15. Czuchra, P. A. Operation and Maintenance of a Particulate
Scrubber System's Ancillary Components. Presented at the
U.S. EPA Environmental Research Information Center Seminar
on Operation and Maintenance of Air Pollution Equipment for
Particulate Control, Atlanta, Georgia, April 1979.
42
-------
APPENDIX A: CASE HISTORIES
INSPECTION REPORTS FOR EIGHTEEN PARTICULATE
CONTROL SYSTEMS THAT HAVE EXPERIENCED CORROSION
A-l
-------
LIST OF INSPECTIONS
Source 1:
Source 2:
Source 3:
Source 4:
Source 5:
Source 6:
Source 7:
Source 8:
Source 9:
Source 10:
Source 11:
Source 12;
Source 13:
Source 14:
Source 15:
Source 16:
Source 17:
Source 18:
Page
Wet Scrubbers Serving Municipal Incinerator A-3
Venturi Scrubbers Serving Municipal
Incinerator A-15
Wet Scrubbers Serving Sewage Sludge
Incinerators A-22
Electrostatic Precipitator Serving Glass
Manufacturing Plant A-28
Fabric Filter Serving Cast Iron Foundry
Cupolas , A-33
Fabric Filter Serving Cast Iron Cupolas A-38
Fabric Filters Serving Cast Iron Foundry
Cupolas A-42
Wet Scrubbers Serving Open Hearth Furnaces A-50
Fabric Filter Serving Portland Cement Plant A-55
Electrostatic Precipitator Serving Portland
Cement Plant A-62
Fabric Filters Serving Slag Cupolas A-70
Fabric Filter Serving Gas-Fired Aggregate
Rotary Dryer A-75
Wet Scrubber Serving Rotary Lime Kiln A-79
Fabric Filters Serving Lime Kilns A-86
Wet Scrubbers Serving Rotary Lime Kilns A-91
Fabric Filter Serving Industrial Boiler A-97
Venturi Scrubber Serving Coal-Fired
Industrial Boilers A-104
Wire-Burning Incinerator Controlled by Wet
Scrubber A-lll
A-2
-------
SOURCE 1
WET SCRUBBERS SERVING MUNICIPAL INCINERATOR
SUMMARY
This facility houses two identical incinerators for burning
household and light industrial wastes. Each incinerator has a
wet scrubber for controlling particulate matter emissions.
From startup, the air pollution control system has been
subject to severe corrosion. All major equipment in the control
system has now been replaced except for the scrubber vessels
which are in a poor state of repair. Corrosion-related expenses
are well in excess of $1 million.
One of the scrubbers has never achieved compliance with
particulate matter emission standards, and the other only re-
cently achieved compliance. Experiments are continuing in an
attempt to find a modification in the first scrubber that will
achieve compliance.
SOURCE DESCRIPTION
The source is a municipal incinerator that began operating
in 1969. The plant operates 24 hours per day, 7 days per week,
and 52 weeks per year.
The facility consists of two identical incinerators that
each burn 270 Mg of refuse per day. The quality of the refuse
is quite variable and dependent on seasonal effects (e.g., grass
cuttings) and on the unpredictable nature of industrial dispos-
als. In each incinerator, the waste is burned in a moving grate
supplied with underfire air and followed by a rotary kiln.
Trash such as paper and cardboard is used for starting the fire.
Combustion efficiency is dependent on the quality of waste
A-3
-------
available because there are no facilities for additional firing
of natural gas or oil.
The gases pass from the kiln to either a bypass stack or to
the air pollution control system. Each air pollution control
system consists of a quench chamber, followed by a spray scrub-
ber, mist eliminator, wet fan, and stack. The gases exit from
the stack at about 77°C.
The quench section is constructed from refractory-lined
mild steel. The scrubber shell, spray nozzles, and 6.35-mm
diameter mist eliminator support rods are constructed from 316L
stainless steel. Experiments are continuing with different
demister media; currently, Telerets and B-Gon polypropylene
moldings are being evaluated. The fan impeller and housing are
fabricated from Incoloy 825. The stack and the ductwork above
the fan are constructed from fiberglass reinforced plastic
(FRP).
Each scrubber treats 47.2 m3/s of flue gases. This re-
quires the use of 1140 £/min of water in the quench section and
1920 £/min of water in the spray vessel. Of this, about 380
£/min are emitted through the stack as water vapor, and 190
2/min are bled off to the lagoon. The rest of the water is
recycled. Makeup water is monitored on only one scrubber.
Nine months after startup, pH control of the recycled water
was instituted. The plant currently uses once a day grab sample
analysis as the basis for estimating the amount of ammonia that
must be added to maintain a pH of approximately seven.
Ash from the incinerators is landfilled, as is any material
deemed to be noncombustible. A company is planning to recover
metals from the ash.
MALFUNCTIONS DUE TO CORROSION
Problems occurred soon after startup of the system, and
they have continued throughout the operating history of the
plant.
A-4
-------
Scrubber Pumps and Pipework
Within 6 months of startup, the black iron pipework failed
due to general corrosion. The pipework was replaced with cera-
mic-lined mild steel. No further problems have been reported.
The housing and impellers of the recirculation pumps, which
were constucted from Nihard, failed after a short time in ser-
vice. The replacement pumps, which had rubber-lined impellers
and housings, also failed after only a short service life.
Duriron pumps were then installed, and these have worked well
for 8 years.
Quench Chambers
The original quench chambers were constructed from Inconel
600. These failed after 6 years and were replaced with refrac-
tory-lined steel.
Scrubber Shells
The combined scrubber and mist eliminator shells are fabri-
cated from 316L stainless steel. After 3 years, the shells
began to corrode. After about 6 years, the inside of each shell
was sandblasted. The upper portion was lined with FRP, and the
lower portion was lined with castable concrete. The FRP was not
successful and soon began to fail. Reportedly, chemical attack
gave the FRP a burned appearance. The concrete also fell apart.
At the time of the inspection, the scrubber shells were
suffering from severe pitting (Figure Al-1). Many pits had
penetrated the scrubber shells allowing scrubbing liquor to flow
down the outside of the shells and evaporate and leaving concen-
trated corrosive solutions and corrosive salt deposits on the
external surfaces of the shells (Figures Al-2 and Al-3).
Because the corrosion appeared to be associated with the
welds, improper welding techniques may be responsible for some
of the failures. It was also apparent that some plates of the
shell were attacked more severely than others (Figure Al-1),
possibly because of the internal configuration of the scrubber
(alignment of sprays or demisters, delamination of lining, etc.)
A-5
-------
Figure Al-1. Severe pitting apparent on external surface
of one of the scrubber vessels.
or because some of the plates were of a different grade of
stainless steel than that specified.
Scrubber Internal Components
Continual problems have been experienced with the mist
eliminators. Different materials and configurations have been
used in attempts to meet emission standards. The original mist
eliminators were constructed from stainless steel and lasted
1 month. The plant is currently using a 0.3-m layer of
Telerets, followed by five layers of B-Gon polypropylene mold-
ings supported by 6.35-mm diameter, 316L stainless steel rods.
Spray nozzles are constructed from 316L stainless steel and
have a life of 6 to 7 years. Tests have been carried out with
other materials, but none has proved to be more durable (e.g.,
brass spray nozzles lasted only 6 months).
A-6
-------
Figure Al-2. Salt deposits at corrosion leaks in the scrubber vessel
The corrosion appears to be structurally dependent, possibly
affected by the internal configuration of the scrubber.
A-7
-------
Figure Al-3.
Salt deposits on the external surface
of the scrubber vessel.
Fans
The original induced draft fans were constructed from mild
steel that lasted about 3 years. They were replaced by fans
constructed from Incoloy 825, which serve moderately well but
are prone to stress corrosion cracking. The Incoloy fans are
removed from service periodically and welded in an attempt to
repair developing cracks. Occasionally, fans fail catastroph-
ically and must be replaced. Once when this happened, a fan
constructed from Carpenter 20 was installed because of the
limited availability of fans constructed from Incoloy 825. The
Carpenter 20 fan corroded and failed after 2 years.
Stacks
The original system was built with one stack for each
incinerator. Consequently, each stack operated continuously,
A-8
-------
sometimes in the hot dry (bypass) mode and sometimes in the warm
wet (scrubber) mode. The original stacks were fabricated from
Inconel 600; these required patching during their 3-year life.
In 1973, the original single stack on each incinerator was re-
placed with a dual stack system. A hot dry (bypass) stack was
constructed from mild steel lined with refractory brick, and a
warm wet (scrubber) stack was constructed from FRP. The stacks
are currently holding up well, though there is some leakage at
joints in the FRP stacks (Figure Al-4).
Figure Al-4. Leaks in fiberglass-reinforced plastic ductwork
leading to the stack.
DESIGN EVALUATION
Materials of Construction
Wet scrubbers that control emissions from incinerators are
subject to extremely corrosive gases and dusts. Therefore,
rapid attack and failure of a scrubber vessel would be expected
unless careful consideration had been given to the materials of
construction. Recirculation of the scrubbing water also tends
A-9
-------
to accentuate corrosion because of the buildup of aggressive
ions in the water. Incineration of household waste releases
significant quantities of HCl from plastics (e.g., PVC) and
other constituents. In acidic chloride environments, austenitic
stainless steels (e.g., 304, 316) are prone to pitting and would
not usually be expected to give satisfactory service. It is not
surprising, therefore, that the scrubber vessel has suffered
from severe localized corrosion. Additionally, the corrosion of
the original pipework and the pumps should have been expected
because of the highly acidic pH of the recirculating water.
This plant is particularly interesting because it was the
subject of a scrubber materials evaluation program during the
second year after startup.1 Various alloys were evaluated.
Specimens immersed in a channel carrying effluent scrubber water
rapidly became coated with adherent deposits consisting of PbS04
and CaS04-2H2O combined with barium, silicon, and aluminum com-
pounds. Reportedly, these deposits afforded some protection to
the samples. The behaviors of the specimens are described in
Table Al-1. Generally, Ti-6Al-4V, Hastelloy C, and Inconel 625
were found to be superior. The 316L stainless steel was found
to be susceptible to stress corrosion cracking. The composition
of the scrubber effluent is given in Table Al-2. It is inter-
esting to note the effects upon the salt concentrations of
recirculating the water. The flue gases were believed to con-
tain HCl, SO2, SO3, HF, and organic acids. As a result of the
materials investigation, the plant replaced the failed mild
steel fans with ones constructed from Incoloy 825; these have
performed fairly well.
A second study was performed at a sister plant of similar
design. The plant differed in that the scrubber was operated
with zero blowdown, there was no pH control system, and the
incinerators were fueled by a higher percentage industrial of
wastes. These differences resulted in chloride levels of 30,000
ppm in the scrubbing liquor and pH values of about 2.5. This
second study, however, came to conclusions similar to those of
A-10
-------
TABLE Al-1. EVALUATION OF ALLOYS UNDER INCINERATOR SCRUBBER CONDITIONS
Alloy
Ti-6Al-4V
Haste! loy C
Inconel 625
Hastell oy F
Hasten oy C-276
Hastell oy G
T175A
S-816
Carpenter 20
Incoloy 825
316L
310
446
Inconel 600
Inconel 601
Armco 22-13-5
USS 18-18-2
304
Performance
Good
Good
Good
Good
Good
Good
Good
Good
Pitted
Pitted
Pitted, Stress corrosion
Pitted
Pitted
Trenches
Trenches
Pitted
Pitted
Pitted, Stress corrosion
cracking
cracking
TABLE Al-2. SCRUBBER WATER COMPOSITON DURING EXPOSURE PERIOD1
Total hardness as
CaC03, ppm
Calcium as Ca, ppm
Magnesium as Mg, ppm
Sulfate as S04, ppm
Chloride as Cl , ppm
Specific conductance
mmho/cm
Total solids, ppm
Suspended solids, ppm
Total acidity as CaC03,
ppm
PH
Date
8/31/71
2330
568
262
1261
2355
7499
5916
114
402
4.4
10/8/71
2760
784
230
1308
2355
7639
5900
14
388
4.3
11/2/71
2760
1704
573
1933
1610
11,160
11,386
10
352
5.1
1 2/8/7 la
272
71
27
231
183
833
732
114
68
4.0
On this date the scrubber water was not being recirculated.
A-ll
-------
the first study. The most satisfactory materials were found to
be Hastelloy C-276 and Inconel 625, although it was reported
that even these materials were likely to suffer localized corro-
sion in the low-pH, high-chloride environment (titanium products
were not studied). Types 316 and 304 stainless steels were
found to be unsatisfactory.2
Plant Design
Elements of the original design have helped to accentuate
corrosion and consequent air pollution problems.
The use of a single stack for both bypass and scrubber
modes gives particularly arduous service conditions for stack
materials. No materials have been proven adequate for such
service.
The lack of pH control for the recirculation water led to
problems in piping and pumps, and probably initiated problems in
the scrubber itself. The necessary pH control was instituted
9 months after startup.
The lack of provision for additional firing of gas or oil
is responsible for wide fluctuations in the quality and quantity
of the flue gas because burning cannot be controlled. This
makes both effective scrubbing and pH control difficult.
Operating and Maintenance Practices
Maintenance at the plant is minimal. There are no sched-
uled turnarounds for the scrubbers or incinerators. The fans
are repaired periodically as necessary.
The pH control system was initially installed as a fully
instrumented automatic system that used ammonia to control acid-
ity. The plant, however, had continual problems with probes
scaling and breaking. Thus, although the actual electronics
worked well, the plant went to a once a day grab sampling method
to avoid the probe problems. This method has led to some fluc-
tuations in the pH of the recirculating water (Figure Al-5)
because of the slow response time and because of the fact that
A-12
-------
M
U>
10.0
9.0
8.0
7.0
6.0
: 5.0
L
4.0
3.0
2.0
1.0
0
— UNIT 2 INLET
- UNIT 1 INLET
— UNIT 2 OUTLET
••• UNIT 1 OUTLET
Sa Su M
Sa Su M
Sa Su M
Figure Al-5. Variations in scrubber liquor pH over a 22-day
period as measured in lab from daily grab samples (Mondays
through Fridays).
-------
pH is measured only 5 days a week while the plant operates
7 days a week. Fluctuations in the pH of the recirculating
water are also indicated in Table Al-2. Another problem with pH
control has been insufficient allocation of funds. In the last
fiscal year, only about 30 percent of the funds needed to pur-
chase sufficient ammonia has been allocated.
The scrubber is protected from temperature excursions by a
high-temperature monitor that operates the bypass at 121°C.
CORROSION-RELATED EMISSIONS AND COSTS
Emissions resulting from corrosion failures have been
significant. One scrubber is reported to have never met emis-
sions standards; the other scrubber only recently achieved
compliance.
The costs of the corrosion failures have been enormous.
The cost of installing the new stack system was in excess of
$1 million. The installation of the new stacks took 9 months,
during which time the solid waste was directly landfilled at an
additional unreported cost. No cost data were supplied for fan
replacement, but a reasonable estimate would be $150,000. It
was not possible to accurately estimate the cost of repairs and
modifications to the scrubbers and mist eliminators.
REFERENCES
1. Krause, H. H., D. A. Vaughan, and P. D. Miller. Corrosion
and Deposits From Combustion of Solid Waste (Part 2).
J. Engineering for Power, July 1974, pp. 216-222.
2. The Corrosion Resistance of Nickel-Containing Alloys in
Flue Gas Desulfurization and Other Scrubbing Processes,
International Nickel Company, publication CEB-7, 1980.
A-14
-------
SOURCE 2
VENTURI SCRUBBERS SERVING MUNICIPAL INCINERATOR
SUMMARY
The facility consists of six furnaces that were each de-
signed to burn 195 Mg per day of domestic refuse. Only five of
the six furnaces are in use. Each of these five is now con-
trolled by a venturi scrubber to reduce particulate matter
emissions.
The venturi scrubber system has experienced both minor and
major corrosion problems. The major problems have included
premature failure of Incoloy ID fans and deterioration of the
refractory brick stack lining in one of the three stacks. To
date, it is estimated that these problems have cost in excess of
$400,000.
The plant did not release any data on possible violations
of particulate emissions standards that have been caused by
corrosion-induced equipment failures.
SOURCE DESCRIPTION
Source 2 is a municipal incinerator that began operation in
1959. At the time of construction, it was the largest municipal
incinerator facility in the United States. The plant burns
approximately 970 Mg per day of domestic refuse; its operating
schedule is 10 days on and 2 days off. The plant is located in
a region that experiences severe winters.
Each of the six furnaces is designed to consume 195 Mg per
day of domestic refuse. Each furnace has a primary and a sec-
ondary burning chamber. The grates in the primary burning
chamber are of the inclined type and may be controlled either
A-15
-------
manually or automatically. Six fans with variable speed drives
(16 m3/s) supply air to the furnaces for combustion. One auxil-
iary oil burner (3 x lo9 J) is located in each of the primary
and secondary burning chambers of each furnace. Auxiliary
dampers controlled by an electric eye are located on both sides
of the secondary chamber to allow air to enter when needed for
further combustion of incompletely incinerated material. Ash
from the incinerators is landfilled. Flue gases from the incin-
erators pass into the air pollution control system for particu-
late matter removal.
The original air pollution control system had six spray
chambers. In each chamber, four banks of sprays had a capacity
of 410 £/min. Each chamber consisted of three passes; the final
pass was designed to dry the flue gases.
The present air pollution control system was installed
during 1969-1976. In 1969, an Overtron scrubber was installed
on Furnace No. 2. In 1972, 1973, 1974, and 1976, a scrubber
manufactured by Combustion Equipment Associates (CEA) was in-
stalled on each of the furnaces Numbered 3 through 6. In an
attempt to reduce plant costs, Furnace No. 1 was abandoned;
therefore, no pollution control devices were installed on the
furnace. It was believed that the capacity of the five con-
trolled furnaces would be adequate.
The venturi scrubbers are constructed with brick venturi
throats and 316L stainless steel venturi cones. On each scrub-
ber, the venturi is followed by a 12-ft diameter, stainless
steel cyclonic separator, where tangential collection of the
scrubbing liquor is achieved. The flue gases enter the venturi
throat at a temperature of about 650°-760°C, and they exit the
separator at a temperature of about 77°C. From the separator,
the flue gases pass to a fan (Incoloy 800 impeller, carbon steel
shaft, and stainless steel housing) and through ductwork to the
stack. The ductwork is constructed of mild steel lined with a
Ceilcote liner. The stacks are constructed of brick with an
airspace annulus and a refractory-brick lining; the stacks are
A-16
-------
76 m in height. Furnace No. 2 has its own stack; Furnace Nos. 3
and 4 share a stack; and Furnace Nos. 5 and 6 also share a
stack.
Each of the venturi scrubbers uses 3400 £/min of water. Of
this, 3060 £/min is recirculated, 380 £/min is purged to the
clarification system, and the other 230 £/min is vented through
the stack as water vapor. In the event of a temperature excur-
sion, additional emergency water is available. The pH of the
recirculating water is controlled by an automated system that
uses sodium hydroxide to control acidity. The system is checked
hourly by an engineer.
MALFUNCTIONS DUE TO CORROSION
The air pollution control system that was installed when
the plant was constructed was a primitive system consisting of a
water spray that impinged on the flue gases as they left the
incinerator. Reportedly, the system had various materials and
corrosion problems and was not very effective in removing par-
ticulate matter. Several experimental scrubbing systems were
investigated before the Overtron and CEA scrubbers were chosen.
Negligible data are available on these original and experimental
systems.
The present air pollution control system has experienced a
few costly corrosion problems which are detailed in the follow-
ing subsections.
Scrubber Recirculation Pumps
The recycle pumps are constructed of rubber-lined mild
steel. The linings were prone to damage caused by pieces of
refractory that had broken from the venturi lining. The situa-
tion has been remedied by the installation of a stainless steel
screen between each scrubber and its recycle pumps. Before this
modification, the pump impellers had to be replaced frequently—
sometimes as often as once a month.
A-17
-------
Scrubber Internal Components
An inspection of the interior of one of the CEA scrubbers
found it to be very clean, with only slight evidence of corro-
sion. A structural member just below the venturi was suffering
from high velocity erosion. The venturi cones, which are con-
structed from 316 stainless steel, have a life of only 3 years
because of erosion-corrosion. The scrubber shell, which was
constructed from either 304 or 316L stainless steel, showed no
evidence of corrosion.
Fans
Stress corrosion cracking of the ID fans is one of the more
costly aspects of corrosion experienced by the scrubber system.
Each fan wheel is fabricated from Incoloy 800 and is welded
to an Incoloy 800 hub on a carbon steel shaft. Because the
Incoloy is susceptible to stress corrosion cracking in the weld
region, fan life is limited to approximately 4 years. Figure
A2-1 shows a stress corrosion crack running around the weld; the
Figure A2-1. Incoloy hub removed from the carbon steel fan shaft. A section
of the hub has been cut away for metal!ographic analysis. A fracture
resulting from stress corrosion cracking can be seen running in a
semi-circle around the weld.
A-18
-------
crack is most visible in the region of the cutaway (cutaway was
removed for metallographic analysis after the development of the
crack).
Each fan housing was originally constructed of rubber-lined
mild steel. Wear and temperature excursions resulted in perfo-
rations of the lining. Particularly susceptible sections of the
lining have been repaired by welding on stainless steel plates,
which has sometimes resulted in further damage to the remaining
rubber lining.
Ductwork and Stacks
The ductwork was constructed of carbon steel with a
Ceilcote lining that failed because of poor application. The
ductwork was recoated and has lasted well.
The stacks were designed for high-temperature service
(315°-370°C). The present venturi scrubbers, however, use
10 times the quantity of water that the original spray chambers
used; this results in lower gas temperatures (38°-95°C), high
moisture content, and stack lining problems. One stack has been
relined by placing chicken wire and gunite over the existing
acid refractory brick; this nessitated constructing temporary
ductwork to reroute the flue gases to the other stacks. The
temporary ductwork, constructed from unlined mild steel, suf-
fered severe corrosion, but it did serve the purpose for which
it was constructed.
DESIGN EVALUATION
Materials of Construction
Overall, the materials of construction used in the air
pollution control system have been adequate for the service
conditions.
The stack problems have been typical of those experienced
in scrubber retrofit situations, where the quantity of moisture
in the flue gases is greatly increased over original design con-
ditions and where the temperature of the flue gases is lowered.
A-19
-------
These problems, however, were due to design rather than to
materials of construction.
The failures of the fan impellers and hubs because of
stress corrosion cracking may have been avoided-if a more re-
sistant alloy had been selected. Ti-6Al-4V, Hastelloy C, and
Inconel 625 have all been found to have superior corrosion
resistance in scrubber environments; Titanium 75A may also be
satisfactory.1 Both Hastelloy C and Inconel 625 would entail a
higher capital investment than Incoloy 800, but the cost of fans
fabricated from titanium would be comparable with the cost of
Incoloy 800.2 The higher cost of the Hastelloy C or Inconel 625
fans may be justified by reduced maintenance and replacement
costs.
Plant Design
The plant experienced both minor and major design problems
that have now been remedied.
The rubber linings of the recirculation pumps were damaged
continually by pieces of refractory that had broken away from
the interiors of the scrubbers. This costly problem was reme-
died by the installation of stainless steel screens between the
scrubbers and the pumps; the screens remove the pieces of broken
refractory.
The lining failed on one stack because the stacks were
designed for a high-temperature, dry environment (315°-370°C),
and the venturi scrubbers provided a low-temperature, high-
humidity environment that was deleterious to the refractory
brickwork. To date, only one stack has been repaired, but it
seems likely that repairs will be required in the other two
stacks in the near future. The problem could have been avoided
if hot electrostatic precipitators had been installed instead of
venturi scrubbers to prevent particulate pollution.
Some aspects of the plant design have helped to minimize
the effects of corrosion. The use of a continuous automatic pH
control system that performs to design specifications (pH con-
trol systems tend to be unreliable) is one of the primary
A-20
-------
factors in reducing scrubber corrosion. Materials such as
stainless steels that would otherwise have failed in scrubber
service have been used successfully with only minor corrosion
problems. The incinerator design, which allows for auxiliary
fuel firing and for the injection of additional combustion air,
has also been beneficial because the f•• reman can more accurately
control the furnace conditions, achieve more complete combustion
and a more uniform flow of flue gases, and attain more effective
scrubbing and better pH control.
Operation and Maintenance Practices
The plant operates on a planned maintenance schedule that
helps to reduce equipment failures. The pH control system is
serviced weekly and is checked hourly by an engineer. The fans
are checked with a vibration meter every 2 hours.
CORROSION-RELATED EMISSIONS AND COSTS
The plant released only limited data on costs related to
corrosion failures. They estimated that each fan failure costs
$30,000 for materials, giving a total cost for fan failures of
at least $200,000 over the life of the new scrubber system.
Man-power costs for fan installation and the landfilling costs
for disposing of the refuse that would have been incinerated
were not disclosed. No data were supplied on the cost of relin-
ing the stack, but an estimate of at least $200,000 would seem
reasonable.
The plant did not report any emissions problems related to
corrosion failures.
REFERENCES
1. Miller, P.D., et al. Corrosion Studies in Municipal Incin-
erators, EPA SW-72-2-3-3 and NTIS PB-213 378, 1972, p. 110.
2. Feige, N. G. Corrosion Service Experience and Economics of
Titanium's Usage in Gas Scrubbing Equipment for Refuse
Incinerators. Presented at Corrosion/74, paper No. 138.
National Association of Corrosion Engineers, Houston,
Texas.
A-21
-------
SOURCE 3
WET SCRUBBERS SERVING SEWAGE SLUDGE INCINERATORS
SUMMARY
This municipal sewage treatment plant utilizes four multi-
ple hearth incinerators to dispose of sludges accumulated within
the plant. Exhausts from each incinerator are treated for
particulate removal by three-stage impingement plate scrubbers.
Corrosion has been a chronic problem in the fans and stacks
and in the ducting connecting them to the scrubbers. The corro-
sive agents seem to be sulfuric acid and chlorides originating
in the sludges and the scrubbing liquors. Corrosion is exacer-
bated by poor welding practices. No significant particulate
emissions have resulted because of the corrosion, as corrosion
has occurred downstream of the scrubbers. Costs of the corro-
sion damage appear to have been significant, but estimates for
these costs are not available.
SOURCE DESCRIPTION
Source 3 consists of a set of four sewage sludge incinera-
tors serving a large metropolitan sewage treatment plant. Each
of the four incinerators is a multiple hearth incinerator with a
quench chamber for cooling the flue gases and an impingement
type tray scrubber for particulate control. The sewage treat-
ment plant is more than 20 years old, and the incinerators are
approximately 15 years old.
Approximate]y 18,000 Mg of sludge are burned in the incin-
erators each year. Sludge is drawn from both the primary set-
tling chamber and the secondary treatment clarifiers. Sludge
first passes through degritters to remove large abrasives that
A-22
-------
are difficult to pump and then to anaerobic sludge digesters.
Digested sludge is thermally conditioned to facilitate dewater-
ing; it is then thickened by sedimentation and dewatered by
eight vacuum filters. Dewatered sludge fed into the incinera-
tors from the vacuum filters contains only 50 percent water.
The sludge is self-burning, but the incinerators have the
capability of using No. 2 oil or digester gas to aid in combus-
tion. Each incinerator contains nine hearths, has a maximum
capacity of 775 kg of sludge per hour, and can operate 24 h/da,
7 da/wk. Sludge enters the top of an incinerator where it is
first dried on the upper hearths by hot, rising combustion gases
and then burned as it moves slowly down through the lower
hearths. Residual ash is removed at the bottom hearth. Temper-
atures in the furnace are about 320°C in the lower ash cooling
hearth, 750° to 1100°C in the central combustion hearths, and
540° to 650°C in the upper drying hearths. Ash resulting from
the incineration process is discharged into a water-filled tank
and pumped as a slurry to lagoons located at the treatment plant
sites.
Exhaust gases from each incinerator pass through a quench
chamber, a scrubber, and an induced draft fan, and then out
through a stack. The incinerators, quench chambers, and scrub-
bers are located inside a building while the fans and stacks are
on the building roof.
Each scrubber contains three impingement plate trays with
scrubbing water flowing across the top of each tray and water
sprays beneath each tray. Scrubbing water (as well as quench
water) is once through, and is obtained from the plant's second-
ary clarifiers. Each scrubber also has a chevron-type liquid
entrainment separator. Scrubber exit temperatures are approxi-
mately 95°C. Temperature monitors at the scrubber outlet are
interlocked to emergency bypass dampers. The monitors will
actuate the dampers when high-temperature excursions are detec-
ted such as during a loss of scrubber water.
A-23
-------
The fan wheels and chevron mist eliminators are constructed
of type 316 stainless steel. Sections of the ductwork leading
from the scrubbers to the fans and from the fans to the stacks
are type 302 stainless steel, and other sections are type 304
stainless steel. Much of this ductwork is thermally insulated.
The scrubber vessels, stacks, and most other scrubber components
are also stainless steel, but records listing the specific
alloys used are not available.
MALFUNCTIONS DUE TO CORROSION
Corrosion in the Source 3 scrubber systems seems to be
confined to components located outside the incinerator building
(i.e., on the incinerator roof). Most of the fan components and
the ductwork were replaced after 5 years of service because of
corrosion. Some of these components have been repaired again
since then, and most will require replacement a third time in
the near future (Figure A3-1). All four stacks have also been
replaced, and may likewise require replacement again (Figure
A3-2). Corrosion in the stacks and ductwork is most severe at
the welds. Plant personnel have reported little corrosion in
the quench chambers or in the scrubbers.
DESIGN EVALUATION
The probable agents of corrosion at Source 3 are sulfuric
acid and chlorides. Dried municipal sewage sludges generally
contain between I and 2 percent sulfur that can oxidize during
incineration to form sulfur dioxide and sulfur trioxide (temper-
atures within the incinerator are not sufficiently high to form
nitrogen oxides).1 Sulfur trioxide formed in this manner can
combine with the scrubber water to form sulfuric acid. Chlo-
rides can come from two sources in the scrubbing system. Sec-
ondary clarification water, such as that used for the Source 3
scrubbing liquor, typically contains 90 ppm chlorides, and
sludge incinerator exhaust gases typically contain 750 ppm
hydrochloric acid.1'2
A-24
-------
I
N)
cn
Figure A3-1. Stainless steel ducting (alloy
unknown) leading from a scrubber to the fan
inlet. Severe corrosion is visible at each
weld. Less severe corrosion is visible
elsewhere.
Figure A3-2. Base of a scrubber stack showing
severe corrosion at welds (light is visible
along corroded weld seams).
-------
Source 3 maintenance personnel explained that until very
recently, the liquid entrainment separators were not totally
effective (recent modifications have solved liquid entrainment
problems). Poor liquid entrainment separation resulted in mist
carryover with the flue gases, which transported sulfuric acid
and chlorides to the ductwork, stacks, and fans.
The optimum alloy choices for each component have not been
established. Several stainless steel alloys including types
302, 304, and 316 have been used in the system, but plant per-
sonnel have not kept track of which alloys were used in par-
ticular components. Most corrosion seems to be concentrated
around welds, which suggests that poor welding practices are as
great a problem as alloy selection (Figure A3-3). Corrosion in
the welds is probably due to the weld decay mechanism, while
corrosion immediately adjacent to the welds could be the result
of stress corrosion cracking (the stresses being induced in the
metal by the heat of the welding process).
Figure A3-3.
Closeup of a weld in stainless steel ducting showing
corrosion at weld.
A-26
-------
CORROSION-RELATED EMISSIONS AND COSTS
There have been virtually no corrosion-related emissions
from the Source 3 scrubbers, because corrosion generally has
occurred in components downstream from the scrubber vessel. The
main effect of the corrosion has been reduced availability of
the scrubber systems during repair operations. Corrosion of the
stacks and ducts may have also had a moderate impact on plume
rise and dispersion.
The costs of scrubber system corrosion at Source 3 have
probably been significant, but estimates of the costs are not
available.
REFERENCES
1. Compilation of Air Pollution Emission Factors, Third Edi-
tion. EPA-AP-42, Part A, August 1977.
2. Lamb, J. C., III. Advanced Treatment of Municipal Waste-
waters-A Survey of Current Practices. Department of En-
vironmental Sciences and Engineering, University of North
Carolina. Publication ESE-361, November 1973.
A-27
-------
SOURCE 4
ELECTROSTATIC PRECIPITATOR SERVING GLASS-MANUFACTURING PLANT
SUMMARY
Source 4 produces glass in two natural gas-fired melt
furnaces. Exhausts from the two furnaces pass through a single
quench chamber for cooling and through an electrostatic precipi-
tator (ESP) for particulate removal. During the first 5 years
of ESP operation, there have been chronic corrosion problems in
ductwork and dampers between the furnace exhausts and the quench
chamber. The ESP has been bypassed approximately 1 percent of
the furnace's production time to facilitate repair of corroded
ductwork.
The plant has tried various steels, including stainless
steels, but none has lasted longer than 1 year in the hot fur-
nace exhausts. The ESP manufacturer has assisted the plant in a
program to test various alloys. Type 310 stainless steel ap-
pears to be the most durable alloy tested.
The plant has spent $50,000 to $100,000 per year in mainte-
nance of the ESP system. Most of these expenses have been for
replacement of corroded ductwork and dampers.
SOURCE DESCRIPTION
The facility consists of two melt furnaces that exhaust to
one ESP system for particulate removal. The furnaces, which are
designed to operate continuously, use about 51,000 m3 of natural
gas per day. The furnaces have been in operation about 7 years.
Source 4 was initially required by the State agency to
install a particulate control system on the furnaces prior to
plant startup; because this compliance schedule caused design
A-28
-------
problems, however, the agency gave a 1-year extension to Source
4 to finalize the designs and to install particulate controls.
This time was used to choose between ESP, scrubber, and fabric
filter systems; to select the manufacturer; and to design the
system with the manufacturer's assistance. The ESP was placed
in service about 1 year after plant startup, and it passed the
compliance tests 3 months later.
Exhausts from each furnace are normally ducted into a
single quench chamber preceding the ESP. Gases enter the quench
chamber at 600° to 650°C and exit at a temperature of about
270°C. The quench chamber uses air-atomized water to cool the
furnace gases. Water use is about 240 £/min—all of which is
evaporated. The exhaust gases contain silicate particulate as
well as traces of fluorides and lead oxides. A draft is induced
through the quench/ESP system by a fan located between the ESP
and the main plant stack. Airflow through the system is about
11.3 m3/s.
During ESP or quench chamber upset conditions, the furnace
exhausts can bypass the particulate control system and exit
through emergency stacks. Each furnace has its own emergency
bypass stack equipped with a top cap and two shutoff dampers to
divert gas flow where the emergency stack and quench chamber
ducts join. The furnaces, a portion of the ductwork, and the
bypass dampers are located indoors. The quench chamber, ESP,
fan, and remaining portions of the ductwork are located out-
doors. Because of severe winters at the plant site, the ESP is
well insulated, and its hoppers are completely enclosed so that
condensation will not occur.
The original materials of construction in the ESP system
included Corten for the shell, quench chamber, ductwork, and
electrodes; 316L stainless steel for the dampers; 316 stainless
steel for the quench chamber spray nozzles; and mild steel for
the fan.
A-29
-------
MALFUNCTIONS DUE TO CORROSION
High-Temperature Ductwork
The high-temperature furnace gases have caused extensive
corrosion in the ductwork between the furnaces and the quench
chamber. Corten failed after a few months service in this envi-
ronment. The plant replaced sections of corroded Corten ducting
several times and then switched to 316L stainless steel ducting;
this material failed in as little as 3 months. The plant has
begun a test program with the assistance of the ESP manufacturer
to determine the most effective alloy for this environment. It
appears that 310 stainless steel is the most effective material;
several sections of 310 ductwork are still in service after 6
months.
Dampers and Bypass Stack Caps
The dampers have had corrosion problems similar to those in
the high-temperature ductwork, but the corrosion is more severe.
The dampers are exposed to the highest temperatures in the
system without the cooling effects of ambient air that benefit
the ductwork. The best service thus far has been provided by
310 stainless steel.
The Corten bypass stack caps have also experienced high-
temperature corrosion problems, but the problems are less severe
because exposure to furnace exhausts is intermittent. The caps
have been replaced only once in 7 years.
Other ESP System Components
No significant corrosion problems have been experienced in
the lower temperature portions of the ESP system. Corten steel
has served well in the quench chamber, ESP, and connecting duct-
work. The 316 stainless steel quench nozzles have lasted 5
years; they failed due to erosion rather than corrosion. Final-
ly, mild steel has proven to be appropriate for the fan. The
fan is cleaned and balanced annually and has experienced no
problems.
A-30
-------
DESIGN EVALUATION
Corrosion problems in metal ductwork containing high-tem-
perature gases are difficult to solve. Present technology in
refractory metals does not provide alloys that are resistant to
corrosion in all high-temperature applications. Fluorides and
lead oxides in the furnace gases have a synergistic effect with
the elevated temperatures in causing corrosion.1 Fluorides
alone have the potential to cause significant corrosion in 310
stainless steel at temperatures above 400°C.2 The use of 310
stainless steel for ductwork has mitigated the corrosion but it
is doubtful that this alloy will eliminate corrosion entirely.
The best that can be expected with 310 stainless steel is a
longer service life between corrosion failures. Source 4 plans
to use refractory-lined mild steel ductwork on any new ESP
systems it might install.
One design option that might reduce the costs of corrosion
failures is to relocate the quenching chamber further upstream
to reduce the linear footage of ducting exposed to the highest
temperatures.
CORROSION-RELATED EMISSIONS AND COSTS
Corrosion has increased particulate emissions at the plant
in two ways. The most significant increase occurs when a duct
failure necessitates bypassing of the ESP system; the plant
estimates that this has occurred approximately 1 percent of the
operating time of the furnaces. The second most significant
increases occur because noncatastrophic duct failures allow
in-leakage into the system. Because the ESP system was designed
with minimum spare capacity, the increased flow requirements
resulting from in-leakage exceed the design capacity and reduce
the ESP's particulate removal efficiency. Source 4 plans to
purchase a second ESP to provide an increased factor of safety
in the system.
A-31
-------
Annual operation and maintenance costs for the ESP system,
not including power consumption and depreciation, are estimated
at $50,000. These costs are primarily due to replacement of
corroded ductwork. The particulate control system cost approxi-
mately $1.2 million when installed.
REFERENCES
1. Brasunas, A. deS. Alloy Behavior at High Temperatures,
Chapter 13, NACE Basic Corrosion Course. National Associa-
tion of Corrosion Engineers, Houston, Texas, 1977.
2. Moran, J. J., Jr. Corrosion at High Temperatures, Chapter
12, NACE Basic Corrosion Course. National Association of
Corrosion Engineers, Houston, Texas, 1977.
A-32
-------
SOURCE 5
FABRIC FILTER SERVING CAST IRON
FOUNDRY CUPOLAS
SUMMARY
This cast iron foundry operates a cupola battery that pro-
duces iron for a variety of cast iron products. A shaker-type
fabric filter is used to control particulate emissions from the
cupolas. The filter provides adequate control of particulate
emissions when operating properly, but corrosion failures and
fan abrasion have resulted in several fabric filter bypass
episodes since the foundry began production. It is estimated
that particulate emissions during bypass of the filter are
approximately 270 kg/h. It is also estimated that the corrosion
in the fabric filter system has cost the source more than
$250,000 since plant startup.
Filter corrosion is the result of sulfuric acid and other
impurities condensing from the hot flue gases on the cooler fil-
ter surfaces. Elimination of this acid corrosion is difficult
to accomplish without costly modifications. Source 5 is inves-
tigating the possibility of replacing the quench chamber with a
recuperative heat exchanger. This modification may prevent acid
condensation by eliminating the source of most of the moisture
in the flue gases, but it would cost up to $900,000.
SOURCE DESCRIPTION
Source 5 operates a pair of 31-Mg/h cast iron cupolas with
afterburners and a fabric filter to control carbon monoxide and
particulate emissions. The cupolas operate on alternating days
for melts lasting 16 to 20 h. The foundry and associated air
pollution control equipment began operation in 1974.
A-33
-------
Each cupola is a carbon steel cylinder lined with refrac-
tory brick. Prior to ignition, a cupola is charged with scrap
steel (predominately automobile scrap), metallurgical grade
coke, and limestone flux. Additional raw materials are added as
needed during a melt to maintain production. Combustion gases
exit at the top of the cupola through an afterburner section
where the carbonaceous particulates and carbon monoxide in the
gases are oxidized. The afterburners for each cupola can use
natural gas or oil for fuel and have a maximum heat input of 6.8
x 109 J/h. Flue gases exit the afterburners at 700°C.
Cupola exhausts pass from the afterburner through brick-
lined carbon steel duct work to a quench chamber. The 4-m
diameter quench chamber is constructed of carbon steel, lined
with luminite gunite. A series of water sprays with a capacity
of 560 £/min cool cupola exhausts to between 230° and 260°C.
Temperature sensors at the quench chamber outlet control the
water flow through the sprays. These sensors will also trigger
the opening of the cupola bypass caps should the quench sprays
fail to cool the flue gases below the temperature limits of the
filter bags. The relatively large quench chamber reduces gas
velocity so that the quencher also serves as a settling chamber
to remove the largest sized particulates.
Unlined carbon steel ducting leads from the quench chamber
to the system fan. The carbon steel radial fan is designed to
draw 75 m3/s of cupola exhausts through the quench system and
force it through the fabric filter.
The fabric filter is a 14-compartment, shaker-type design
with a roof monitor exhaust. The filter contains 1200 silicone-
treated glass bags having a combined surface area of 7080 m2.
At the designed airflow, the filter has an air-to-cloth ratio of
2.1 to 1. Collected dust falls from the bags into collection
hoppers that empty into screw conveyors. The screw conveyors
discharge into covered bins that are emptied as needed. The
fabric filter enclosure is constructed .of a carbon steel frame-
work and corrugated transite walls.
A-34
-------
MALFUNCTIONS DUE TO CORROSION
The quench chamber and the ductwork have experienced few
corrosion problems since installation of the filter. The fan
has suffered from abrasion, however, and the carbon steel por-
tions of the filter have suffered from corrosion.
Fan
The fan is located on the dirty side of the fabric filter
where particulate is present in the gas stream. This particu-
late abrades the fan blades so that they eventually become
unbalanced or their clearance with the fan casing becomes too
great. The fan has required major maintenance or replacement at
least once a year since startup. During unexpected fan outages,
the cupola emissions must be exhausted untreated through the
cupola bypass caps.
Source 5 is experimenting with surface treatment of the fan
blades as a method of reducing wear. Fan components subject to
high wear are sprayed with a chromium-nickel treatment. High-
magnesium wear plates have also been used. These measures seem
to reduce wear although experiences with these measures have
been too short to determine their overall effectiveness.
Fabric Filter
The fabric filter has suffered corrosion of all carbon
steel parts, including the tube sheet, bag hangers, dust hop-
pers, and structural members of the filter enclosure. Corrosion
is caused by the condensation of sulfur oxides from coke combus-
tion with the water vapor from the quenching, forming acids on
filter surfaces. Acid condensation is most frequent during
winter startup sequences. The plant has had to replace or
repair many filter components and it must sandblast and paint
steel parts as frequently as once a year.
A-35
-------
DESIGN EVALUATION
The design of the fabric filter system at Source 5 provides
adequate control of cupola emissions. Several changes could be
made, however, to reduce fan abrasion and filter corrosion,
thereby increasing system reliability.
Fan abrasion could be avoided if the fan was relocated to
the clean air side of the filter. To alter the fan to this
configuration, however, would require a major rebuilding of the
filter roof monitor exhaust and the installation of a stack.
The filter has two design features for minimizing corro-
sion. These features are the construction of the filter enclo-
sure so that structural steel is outside the corrugated transite
walls, away from the hot sulfur-bearing flue gases; and the use
of protective coatings to protect these and other steel parts
from corrosion. These design features have helped reduce the
severity of corrosion, but they have not proven to be a total
solution. The structural steel is still subject to some corro-
sion in its outside location, and steel components inside the
filter are subject to the full effects of acid condensation.
The plant has experimented with several coatings, but most have
failed in less than a year. The present high-temperature epoxy
coating has provided more than 15 months service.
Source 5 is considering the replacement of the quench cham-
ber with a recuperative heat exchanger system. This system
would cool cupola exhausts without increasing the moisture
content of the flue gases, which would reduce the potential for
condensation in the filter. The heat removed by the heat
exchanger might be usable for plant heating and for preheating
the combustion air injected into the cupola. This system is
expected to cost between $650,000 and $900,000.
CORROSION-RELATED EMISSIONS AND COSTS
Control system corrosion failures have increased particu-
late emissions at Source 5. Corrosion-related emissions are
A-36
-------
predominantly in the form of control equipment bypass episodes
because of unexpected corrosion failures. The filter system has
had several periods of uncontrolled emissions lasting 2 to 3
days since plant startup. Most of these periods were due to fan
outages or because the sandblasting and painting of the filter
required more than the annual 10-day period reserved for major
plant maintenance. During normal operation, the filter removes
up to 99 percent of cupola particulate emissions and emits no
visible particulates. Decreased performance can occur, however,
because of bag failures or corrosion of the tube sheets, bag
hangers, and dust hoppers. Uncontrolled emissions from a cupola
battery such as at Source 5 have been estimated to be in the
range of 270 kg/h.1
The original cost of the filter system, including the fan
and the quench chamber, was about $175,000. Since installation
of the filter, the plant has spent about $25,000 annually to
sandblast and paint the filter and up to $20,000 annually to
repair abraded fan parts. Additional costs have been incurred
because of corrosion damage to the hoppers and other filter
components. In six years of filter operation, corrosion-related
maintenance costs have probably totaled more than $250,000.
REFERENCE
1. Controlled and Uncontrolled Emission Rates and Applicable
Limitations for Eighty Processes, EPA-340/1-78-004, 1978.
A-37
-------
SOURCE 6
FABRIC FILTER SERVING CAST IRON CUPOLA
SUMMARY
This facility produces iron-shot abrasives in a cast iron
cupola. A fabric filter was added after the facility operated
uncontrolled for 1 year. The uncontrolled cupola can liberate
approximately 8.5 kg of particulate per Mg of iron produced.1
At the time of installation, the filter provided adequate
control of emissions, but its efficiency and reliability deteri-
orated significantly over time as a result of corrosion. The
causes of filter corrosion were poor control of the acid dew-
point as a result of poor filter insulation, and mist carryover
from the quench chamber. After only 5 years of service, the
filter system, representing a capital investment of approxi-
mately a quarter million dollars, was replaced with a venturi
scrubber system.
SOURCE DESCRIPTION
The typical cupola production schedule is one 10-h shift
per day, 250 days per year. Approximately 900 kg of scrap steel
(predominately automobile scrap), 135 kg of metallurgical grade
coke, and limestone flux are added to the cupola each hour. The
cupola was placed into service more than 6 years ago, and it
operated for 1 year before the fabric filter was added to con-
trol particulate emissions.
A quench chamber was used for cooling the cupola emissions
to a temperature compatible with the fabric filter. Cupola
exhausts entering the quench chamber at 980°C were cooled to
A-38
-------
170° to 190°C before they entered the fabric filter for partic-
ulate removal. The pulse-jet fabric filter was equipped with a
thermocouple for recording inlet gas temperatures, a bypass
damper activated by the thermocouple, a differential manometer
for indicating pressure drop across the filter medium, and a
magnehelic gauge indicating pulse-air pressure. Under normal
conditions, fabric filter differential pressure varied from 0.25
kPa immediately after a cleaning cycle to 2.2 to 2.5 kPa just
prior to a cleaning cycle. Pulse-air pressure was maintained
between 550 and 690 kPa.
The filter was constructed of epoxy-lined steel, filter
bags were made of felted Nomex, and the quench chamber was
constructed of uncoated mild steel. Dust was removed from the
filter hoppers through a rotary airlock and a screw conveyor.
The filter system was designed to treat 18.9 m3/s of cupola
gases at 980°C.
After 5 years of operation, the fabric filter was replaced
with an adjustable throat venturi scrubber. The 304 stainless
steel venturi scrubber is designed to treat 21.2 m3/s of cupola
gases at 980°C. Scrubbing water is recirculated and treated for
pH control. The new scrubber system is equipped with tempera-
ture sensors and fan ammeters; water flow meters and pH monitors
are presently being installed.
MALFUNCTIONS DUE TO CORROSION
Corrosion became a problem in the fabric filter system
after 2 years of operation. By the end of 5 years, the system
was unusable; that is, corrosion had penetrated the filter walls
and hoppers, weakened the structural supports, and caused fail-
ure of 80 percent of the ducting. During the life of the unit,
bag cages and bag hangers suffered repeated failures. The fan
was removed from service because of corrosion and wear about 1
year after the filter was removed from service. The quench
chamber, however, provided good service throughout the life of
A-39
-------
the fabric filter system, although moderate corrosion had become
evident by the end of 5 years use.
The cause of corrosion was acid attack of mild steel struc-
tures within the filter. The acids were formed by combustion
products, such as nitrogen oxides and sulfur oxides, that had
condensed with water on the cold outer surfaces of the filter
and ductwork.
DESIGN EVALUATION
Design requirements for a fabric filter system restrict
temperatures to a specific range. Ideally, the quench chamber
should cool the gases to a temperature below the limitations of
the bag fabric yet above the sulfuric acid dewpoint. Tempera-
ture control in a cupola is complicated by startup sequences,
shutdown sequences, and numerous charging operations during each
shift, which cause the exhaust temperatures to vary.
Source 6 indicated that moisture carryover from the quench
chamber was a problem throughout the life of the filter. Even
when quench water was completely evaporated, condensation within
the filter was likely because it was not insulated, and it was
located outside where it was exposed to the severe winters at
the plant site. The system was especially susceptible to con-
densation as the filter cooled after shutdown of the cupola.
Source 6 listed several filter problems in addition to
corrosion. First, the actual airflow was about half that of the
designed airflow; this reduced draft in the cupola .limited the
plant production rate. Second, sparks penetrated the quench
chamber and on rare occasions caused fires in the filter; these
fires may have damaged the epoxy coating on the structural
steel. Third, the rotary airlock and the screw conveyor were
subject to plugging, which the plant management attributed to
tramp materials entering the filter. Because of problems with
the fabric filter, especially with quench chamber moisture
carryover, Source 6 replaced the filter with the scrubber
system.
A-40
-------
CORROSION-RELATED EMISSIONS AND COSTS
Corrosion of the fabric filter greatly increased the par-
ticulate emissions of Source 6. The manufacturer reported the
efficiency of the filter to be 99.9 percent, but the plant mana-
ger estimated the efficiency to be only 90 percent throughout
most of the 5 years of filter operatxon. This low efficiency
was due in part to corrosion. The plant was frequently out of
compliance during this period; during the final 1% years of
filter operation, the plant was out of compliance most of the
time. The enforcement agency allowed the plant to operate in
this manner while attempts were made to improve the system.
Because compliance problems were obvious, neither Source 6 nor
the State performed a stack test on the filter.
Neither the filter manufacturer nor the distributor ac-
cepted responsibility for any of the corrosion problems or other
design problems. Source 6 was unable to invoke a guarantee,
written or implied, to recover any of the financial losses of
filter failures. The costs of filter system malfunction included
total depreciation of the $270,000 filter system in less than
half of its planned 15- to 20-year life, an estimated $30,000 to
$40,000 in annual maintenance costs, and an undetermined amount
of lost production. The capital costs of the filter failure can
be attributed almost entirely to corrosion. More than 60 per-
cent of the annual maintenance costs were due to bag failures
rather than corrosion, however, and nearly all the loss of
production costs were due to the inadequate flow rate rather
than corrosion.
REFERENCES
1. Formica, P. N. Controlled and Uncontrolled Emission Rates
and Applicable Limitations for Eightly Processes. EPA-340/
1-78-004. April 1978.
A-41
-------
SOURCE 7
FABRIC FILTERS SERVING CAST IRON
FOUNDRY CUPOLAS
SUMMARY
This foundry operates two cupola lines that produce a vari-
ety of cast iron products. Cupola emissions were originally
uncontrolled, but the plant became subject to State Implementa-
tion Plan (SIP) emission regulations in the early 1970's.
Fabric filters were then retrofitted to each cupola. These
filters performed satisfactorily at first, but fan abrasion and
corrosion problems appeared in each filter system within a year.
Fan and filter failures have resulted in a number of control
equipment bypass episodes and have cost a considerable amount in
replacement or repair of equipment. It is estimated that par-
ticulate emissions during bypass of the filter are approximately
150 kg/h for each cupola line. It is also estimated that filter
system corrosion has cost the source in excess of $300,000 in 8
years of operation.
Filter corrosion is due to the difficulty in maintaining
filter temperatures above the sulfuric-acid dewpoint. Acid
condensation is especially severe during cold startup of a
fabric filter. Elimination of dewpoint troubles may require
costly design changes or changes in operating procedures.
Changes that may be required include the improving of filter
insulation, the preheating of the filters prior to cupola and
quench startup, or the replacement of the water quench with
recuperative heat exchanger.
A-42
-------
SOURCE DESCRIPTION
Source 7 operates two cast iron cupola batteries with a
fabric filter on each for control of particulate emissions.
Each battery includes two cupolas that operate on alternating
days to facilitate maintenance. One battery has an 18 Mg/h melt
rate and operates 24 h/day while the other battery has a 16 Mg/h
melt rate and operates 16 h/day. The fabric filters were retro-
fitted to the cupolas in 1972 and 1973.
Each cupola is a carbon steel cylinder lined with refrac-
tory brick. Prior to ignition, a cupola is charged with scrap
steel (predominately automobile scrap), metallurgical grade
coke, and limestone flux. During the melt, additional raw
materials are added as needed to maintain production. Combus-
tion gases exit at the top of the cupola through an afterburner
where any carbonaceous particulates or carbon monoxide remaining
in the gases are oxidized. Each afterburner contains two burn-
ers that can use natural gas or oil for fuel and that have a
maximum heat input of 4.6 x io9 J/h. Flue gases exit the after-
burner at 700° to 1000°C.
Cupola emissions are ducted to a quench chamber through
water-cooled carbon steel. The quench chamber is a 3-m diameter
carbon steel cylinder containing a series of sprays for cooling
the cupola exhausts below the temperature limitations for the
filter material. Each spray system has a maximum capacity of
260 £/min. The quench chambers are lined with a layer of lumi-
nite gunite. The relatively large quench chambers reduce gas
velocity so that the chambers also serve as settling chambers
for removal of the largest sized particulates. Temperature
sensors at the quench chamber outlets control the water flow
rates in the spray sections. In addition, these sensors will
trigger the opening of the cupola bypass caps whenever a failure
occurs in the quench chamber. During normal operations, quench
chamber exhausts vary between 230° and 260°C.
A-43
-------
Cooled cupola exhausts exit the quench chambers through un-
lined carbon steel ducting to the cupola fans. Each cupola bat-
tery has an identical carbon steel centrifugal fan designed to
draw 34 m3/s through the quench chamber. The quenched cupola
exhausts pass through carbon steel ducting to the fabric filter
inlets.
The two fabric filters are similar in design, each having
10 compartments, shaker-type cleaning, and roof monitor ex-
hausts. Each filter contains 600 silicone-treated glass bags
with a combined surface area of 3540 m2. During typical opera-
tion, a filter's air-to-cloth ratio is 1.9. Collected dust
falls from the bags into collection hoppers that empty into
screw conveyors. The screw conveyors discharge into covered
bins that are removed for disposal as needed. The fabric filter
enclosures are constructed of corrugated transite sheeting
attached to a carbon steel framework.
MALFUNCTIONS DUE TO CORROSION
The quench chambers and the ductwork have experienced few
corrosion problems since installation of the filters. The fans
and the filters, however, have not provided satisfactory ser-
vice. The fans have suffered from abrasion, and the filters
have suffered from acid dewpoint corrosion.
Fans
The fans are located on the dirty side of the fabric fil-
ters where parti culate is present in the gas stream. This
particulate abrades the fan blades so that they eventually
become unbalanced or their clearance with the fan casings be-
comes too great. The fans have required major overhaul or
replacement about once a year because of this abrasion. Unex-
pected fan outages require about 3 days to repair, during which
time the cupolas must exhaust uncontrolled through the bypass
caps.
A-44
-------
The plant is presently experimenting with surface treatment
of the fan blades. The plant has used high-magnesium wear
plates and has sprayed .fan parts with a chromium-nickel treat-
ment. These measures seem to reduce wear, although experience
with these measures has been too short to make a final determi-
nation of their effectiveness.
Fabric Filters
The fabric filters have suffered corrosion of all carbon
steel parts. Corrosion appeared within the first year of filter
operation and has at times threatened the structural integrity
of the filter enclosure. In 7 years of operation, each filter
has required replacement of the tube sheet, the perimeter
H-beams (Figure A7-1), the bottom hoppers, and the bag hangers.
During this same period, the catwalks and stairways have been
partially replaced. The plant now sandblasts and paints the
steel structures about once a year.
The filter corrosion is caused by acid condensation result-
ing from the combustion products of sulfur from the coke combin-
ing with moisture added during quenching. Corrosion is most
severe on cooler surfaces in the filter where condensation is
most frequent (Figure A7-2). The plant reports that condensa-
tion is more likely during winter startup sequences than during
summer startup sequences or during shutdown sequences. The
intermittently operated cupola battery suffers less severe
filter corrosion than the continuously operated battery, in
spite of a presumably larger number of cold startups.
DESIGN EVALUATION
The design of the fabric filter systems at Source 7 seems
to provide adequate control of particulate emissions from the
cupolas. From a corrosion standpoint, however, several design
changes could have increased system life and reliability. In
fairness to the plant management, it should be noted that the
A-45
-------
Figure A7-1. Inside Source 7 fabric filter showing structural steel
corrosion and new cross-members welded alongside
weakened original cross-members.
A-46
-------
•m- '
Figure A7-2. Closeup of Source 7 fabric filter cross-member
showing that dewpoint corrosion is more severe along
surfaces nearest the cool outer wall.
severity of filter corrosion proved to be greater than antici-
pated during the design phase, and that many design options that
could have reduced corrosion problems would have been considera-
bly more expensive.
The fan abrasion problems might have been avoided during
filter system design by locating the fans on the clean side of
the filters. To alter the filters to this configuration after
they have been constructed would probably be prohibitively ex-
pensive because the roof monitor exhausts would have to be re-
worked, the fans would have to be moved, and stacks would have
to be built. The present experimentation by Source 7 with wear
plates and surface treatments may increase fan life sufficiently
so that the present configuration will be acceptable.
Filter corrosion could be reduced in at least three ways.
First, corroding filter parts could be reconstructed of corro-
sion-resistant alloys. Such alloys would be excessively
A-47
-------
expensive, however, and the coupling of these alloys to existing
carbon steel parts could cause galvanic corrosion of the latter.
Second, Source 7 could continue to rely on protective coatings
to reduce corrosion, but experience thus far indicates that
coatings are not a total solution. A third way to mitigate
filter corrosion is to achieve better control of the acid dew-
point within the filter.
Improving dewpoint control seems to be the most practical
way of reducing filter corrosion. Dewpoint control could be
improved by a combination of design changes and changes in op-
erating procedures. One design change that could help is the
rebuilding of the filter enclosures and the hoppers with a thick
layer of thermal insulation on the outside. Concurrent improve-
ments in quench control instruments might allow their use for
preventing temperature excursions below the dewpoint during
cupola melts as well as for preventing high-temperature excur-
sions. Condensation during cold startup would not be addressed
by these measures, however. Startup condensation could be re-
duced by preheating the filters before the cupolas and quench
sprays are started. The afterburners might serve as preheaters
if they could be operated at a rate low enough to prevent damage
to the filter bags.
CORROSION-RELATED EMISSIONS AND COSTS
Control system corrosion failures increase particulate
emissions substantially at Source 7. Corrosion-related emis-
sions are predominately in the form of bypassing of the control
system because of unexpected corrosion failures. Uncontrolled
emissions from cupolas such as at Source 7 have been estimated
to be in the range of 150 kg/h.1
Each of the two filter systems has had several periods of
uncontrolled emissions lasting 2 to 3 days since their installa-
tion. Most of these periods were due to fan outages. During
normal operation, the filters remove up to 99 percent of the
A-48
-------
cupola particulate emissions and emit no visible participates.
Decreased performance does occur, however, because of bag fail-
ures or corrosion in the tube sheets and bag hangers. During
the inspection of Source 7, a light to moderate haze was visible
at the roof monitor exhaust for one filter.
The original cost of the pollution control system, includ-
ing fans and quench chambers, was about $1 million. The major
effect of the corrosion has been a reduction in the useful life
of the equipment. It is expected that the filter enclosures
will have to be rebuilt in the next several years. In addition
to reducing the life of this capital equipment, corrosion-re-
lated maintenance costs have been substantial. Costs of fan
repairs and the painting of the filters have probably exceeded
$30,000 per year. In addition, approximately $25,000 was re-
quired to rebuild the filter hoppers. Corrosion-related costs
in the filter systems have probably totalled more than $300,000
since installation of the filters.
REFERENCE
1 Controlled and Uncontrolled Emission Rates and Applicable
Limitations for Eighty Processes, EPA-340/1-78-004, 1978.
A-49
-------
SOURCE 8
WET SCRUBBERS SERVING OPEN HEARTH FURNACES
SUMMARY
This source is a steel mill using retrofitted wet scrubbers
to control particulate emissions from three open hearth fur-
naces. The scrubbers reduce emissions to well below the appli-
cable regulatory standards, but corrosion in scrubbing water
treatment system components has caused the mill to bypass the
scrubbers on numerous occasions. Corrosion has been caused
primarily by poor reliability in the pH probes that control lime
additions. Inappropriate materials choices have also contrib-
uted to the corrosion.
Uncontrolled particulate emissions during scrubber bypass-
ing episodes have been about 23 Mg/day. Based on the number of
malfunctions requiring bypassing of the scrubbers, Source 8 may
have emitted as much as 450 Mg of particulate during the first
year of scrubber operations. The costs of the corrosion fail-
ures have totalled more than $70,000 in replacement parts and
labor.
SOURCE DESCRIPTION
Source 8 produces steel in three open hearth furnaces that
use No. 2 and No. 6 oils as fuel. The mill operates only two
furnaces at a time to facilitate periodic furnace maintenance
without interrupting continuous production.
Furnace exhausts have been treated for particulate removal
by a battery of six parallel scrubbers since the fall of 1978.
These scrubbers are designed so that only three are in use while
the other three are being cleaned and/or repaired. The old
A-50
-------
furnace stacks are available to permit bypassing of the scrub-
bers when more than three scrubbers require servicing.
The scrubbers are the hydrosonic aspirator type in which
draft is induced by scrubbing water that flashes into steam as
it mixes with hot flue gases at the scrubber orifice. Flue
gases enter the scrubber at temperatures between 480° and 650°C.
The scrubbers use water at the rate of 4540 £/min; of this, 1140
2/min are evaporated and 3400 £/min are recycled. Used scrub-
bing water passes from the scrubber into a mixing tank where
lime slurry is added for pH control. The lime treatment system
has duplicate lime pumps and piping because of their propensity
for plugging. Scrubbing liquor passes from the lime-mixing tank
into clarifiers, where particulate matter settles as sludge.
Sludge is dewatered by vacuum filtration.
Lime additions to the scrubbing water are controlled auto-
matically by pH instrumentation, because of frequent variations
in the composition of furnace exhaust gases. Without proper pH
control, sulfur oxides from combustion of the fossil fuels can
combine with the scrubbing liquor to form acids that can reduce
liquor pH to as low as 2.0. During periods immediately follow-
ing addition of limestone flux to the furnaces, scrubbing liquor
pH remains high without lime additions.
Materials of construction for the scrubber system include
carbon steel for the scrubber vessel, 316L stainless steel for
the scrubbing water atomizers, and fiberglass-reinforced plastic
(FRP) for troughs used to collect the scrubbing liquor. The FRP
resin is a phenolic type that can withstand temperatures up to
200°C, although the temperature of the liquor is usually about
70°C. Lime slurry vessels and piping are also constructed from
FRP, except for the piping near the slurry pumps, which is
stainless steel. The pump casings are rubber-lined cast iron.
A-51
-------
MALFUNCTIONS DUE TO CORROSION
Corrosion failures have plagued various components in the
scrubbing water treatment system since 4 months after the scrub-
ber system startup. These failures not only have required
numerous short-term shutdowns of the scrubbers to facilitate
repairs, but also have caused one or both of the operating
furnaces to emit uncontrolled exhausts through the bypass
stacks.
Corrosion first appeared in the sludge dewatering vacuum
blowers at the water treatment plant. The mild steel casings
for these blowers deteriorated so quickly that temporary jack-
eting and patching had to be welded to the blowers several times
before replacement casings could be obtained. The replacement
casings were constructed of rubber-lined steel.
The next corrosion failures were the lime slurry pumps.
These pumps alternately corroded during periods of low pH and
plugged during periods of high pH. Problems were so severe on
several occasions that duplicate standby pumps failed before the
primary pumps could be serviced. These pumps have been replaced
with pumps that have rubber-lined casings, but the linings have
been only moderately successful because of abrasion damage.
Other corrosion failures have occurred in the piping in the
scrubbing water treatment system. Mild steel pipes and fittings
were installed by the contractor at some locations instead of
specified stainless steel fittings. These components have
failed and have been replaced by components made of.the specif-
ied corrosion-resistant materials.
There have been no serious corrosion problems in the carbon
steel scrubber vessel or in any of the FRP components.
DESIGN EVALUATION
Corrosion problems occurred in the Source 8 scrubber system
primarily because of difficulty in controlling the pH of the
scrubbing liquor. A contributing cause of corrosion was the use
A-52
-------
of inappropriate materials in locations subjected to intermit-
tent (albeit unexpected) acid conditions. Corrosion problems
are likely to be less frequent in the future because maintenance
personnel devote more effort to monitoring the pH control system
than when the scrubbers were first installed and because unsuc-
cessful materials have been replaced with more expensive corro-
sion-resistant materials.
Although the pH control circuitry is quite reliable, the pH
probes do not remain calibrated for very long. Plant engineers
describe the pH probes as the "weakest link" in the scrubbing
system. Since installation of the scrubbers, plant engineers
have experimented with several probe types, and have tested
ultrasonic probe-cleaning methods. These efforts have produced
only minor improvements in pH system reliability, so it is still
necessary to analyze grab samples of scrubbing liquor in the
mill laboratory at least once per shift to verify the calibra-
tion of the pH probes.
Compounding the pH control problem is the 5-min detention
time in the lime-mixing tank. Variations in scrubbing liquor pH
would be dampened if a larger tank could be used, but there is
insufficient room near the scrubbers to enlarge the tank by the
required amount.
CORROSION-RELATED EMISSIONS AND COSTS
Corrosion failures in scrubbing system components have
required that the scrubbers be bypassed on numerous occasions
during the first year of scrubber operation. Bypassing has
occurred at least 20 times for periods up to 24 h in length
because of failures in the sludge dewatering vacuum blowers,
lime slurry pumps, and piping. During each emissions episode,
Source 8 reported the nature and magnitude of the problems to
the State enforcement agency.
Uncontrolled particulate emissions at the mill are esti-
mated to be 23 Mg/day. When operating, the scrubbers remove
A-53
-------
more than 99 percent of the particulates from the furnace ex-
hausts, reducing mill emissions to 20 percent of the allowable
rate. Excess emissions caused by bypassing the scrubbing system
were, therefore, in the range of 450 Mg during the first year of
scrubber operation.
Corrosion problems in the scrubbers increased plant operat-
ing costs by about 10 percent during the first year of scrubber
operation. The costs of replacing corroded pumps, vacuum blow-
ers, and piping were more than $50,000 in materials and $20,000
in labor. The original capital investment in the scrubber sys-
tems was more than $3 million, but these expenditures will be
partially recovered by reductions in refractory maintenance and
by increases in furnace capacities made possible by the scrub-
bers .
A-54
-------
SOURCE 9
FABRIC FILTER SERVING PORTLAND CEMENT PLANT
SUMMARY
This facility produces portland cement in a rotary kiln
fired with high-sulfur coal. Exhausts from the kiln pass
through a set of multiclones, several radiant cooling loops, and
a reverse-air fabric filter. Within the first year of filter
operation, the rate of filter bag failures due to corrosion of
support rings and bag hangers became excessive. Changing the
materials in the support rings and changing the design of the
hangers increased the life of these components to acceptable
levels. Corrosion has also attacked portions of the filter
enclosure. Corrosion of the bag hardware and the filter enclo-
sure has resulted because of acid condensation within the filter
and rainwater inleakage into the filter.
During the first year of filter operation, the corrosion
failures undoubtedly increased the particulate emissions above
regulatory limits on several occasions. The cost of repairing
the corroded parts was in excess of $24,000, but it was shared
by Source 9 and the filter manufacturer.
SOURCE DESCRIPTION
Source 9 produces about 250,000 Mg per year of portland
cement in a single, dry process kiln. The kiln is fired with 2*5
percent sulfur coal and runs continuously for campaigns lasting
up to 4 months. The kiln is shut down three to four times a
year for repair of the refractory lining and five to six times a
year (briefly, for less than 8 hours) for unexpected mainte-
nance .
A-55
-------
Exhaust gases from the kiln pass first through a set of
multiclones for removal of large particulates, then through a
set of four radiant cooling loops, and finally through the
fabric filter (Figure A9-1). The old stack (available for
emergency bypass) and the system fan are between the cooling
loops and the filter. The cooling loops reduce filter inlet gas
temperatures to between 230° and 260°C.
Figure A9-1. View of Source 9 with rotary kiln on the left, fabric filter
on the right, radiant cooling in the center, and emergency bypass stack
(partially obscured) behind the cooling loops.
The reverse-air fabric filter has 10 compartments and a
roof monitor exhaust. The filter is about 4 years old; it is
constructed of mild steel, and the interior steel surfaces were
originally coated with paint (type unknown). All outside walls
of the filter except the entry hatches are thermally insulated.
The fiberglass filter bags have wire rings sewn into the fabric
for preventing bag collapse; these rings, originally constructed
of mild steel, are now cadmium-coated steel. The filter system
A-56
-------
instrumentation includes a manometer in each compartment, a
manometer across the entire filter with an indicator in the
control room, a filter thermocouple, fan ammeters, a fan tachom-
eter, and reverse-air fan malfunction warning lights. The ther-
mocouple is linked to a warning light and to an alarm to indi-
cate high-temperature excursions that could damage the filter
bags.
MALFUNCTIONS DUE TO CORROSION
Bag Support Rings and Hangers
Within 1 year of fabric filter startup, bag failures had
become frequent because of rapid corrosion of the mild steel bag
support rings (Figure A9-2). Broken rings not only allowed the
bags to flex but also provided sharp points that induced punc-
tures and tears in the fiberglass fabric. A second cause con-
tributing to bag failure was corrosion of the bag support
springs. Support spring failures reduced the bag tension,
interfered with the cleaning of the fabric, and occasionally
allowed the bags to fall into the filter hoppers.
Both of the above corrosion problems were alleviated by
changing the materials of construction. All replacement filter
bags have cadmium-coated steel rings rather than the plain steel
rings. The bag support mechanisms were redesigned so that a
heavy chain-type hanger replaced the light spring-type hanger.
The manufacturer was helpful in correcting these materials prob-
lems and shared some of the expenses attributed to ring and
hanger failures.
Structural Steel
The structural steel showed signs of corrosion during the
inspection, but there have been no structural failures to date.
The corrosion was most apparent near the roof exhaust (Figure
A9-3) and near the access doors where heat losses and condensa-
tion are most likely (Figure A9-4). In general, the structural
A-57
-------
Figure A9-2. Filter bag showing tear caused by
abrasion with corroded, mild steel support ring.
Figure A9-3. Structured steel near filter showing
corrosion of the roof support beam.
A-58
-------
corrosion did not appear to pose a serious problem; however,
repair of corroded members, especially near the access doors,
will be required during the life of the filter. An undetermined
type of paint visible on the structural steel within the filter
had protected the steel during the first 4 years of filter
operation (Figure A9-5). Much of this coating, however, had
failed by the time of inspection, exposing the steel to the kiln
exhaust gases.
Other Filter System Components
The plant did not report any serious corrosion problems in
the fan, cooling loops, ductwork, or any other components.
DESIGN EVALUATION
Combustion of coal that contains 2% percent sulfur creates
the potential for corrosion in any downstream equipment for
handling flue gas. To minimize corrosion caused by acid conden-
sation in a dry particulate collection system, it is highly
desirable to maintain temperatures above the sulfuric acid dew-
point and to prevent moisture inleakage. It is apparent that
some amount of acid condensation does occur within the Source 9
fabric filter. It is also possible that rainwater occasionally
enters the filter at the roof monitor exhaust and at the entry
hatches.
Evidence of corrosion is most apparent in the relatively
cooler areas near the filter access doors and roof and less
apparent on the interior structural steel. Corrosion in both
locations is limited somewhat by the formation of a cement-like
cake on exposed steel structures. Portland cement is known to
inhibit corrosion of mild steel in most environments.1 The
severity of the corrosion in the mild steel bag rings and hang-
ers may have resulted because no cement cake formed on these
components. Contact with the fabric probably abraded the ring
surfaces clean, and the hangers probably remained clean because
they were on the clean side of the filter.
A-59
-------
>
o
Figure A9-4. Mild steel fabric filter
enclosure showing corrosion near the
uninsulated hatch. The corrosion pattern
in the upper left corner of the hatch
frame is due to condensation of acid from
warm flue gases leaking through the hatch
seal.
Figure A9-5. Mild steel structural support
showing the remnants of the original
protective coating.
-------
Replacement of the plain steel rings with cadmium-coated
ones seems to be an appropriate design change, as the coated
ones are providing much longer service than the plain rings.
The use of heavier chain-type hangers also seems appropriate;
their heavier gauge construction will provide longer service
than the lighter springs before failing from corrosion. It
might also be helpful to check the filter enclosure for leaks;
several places were observed in the roof and near the entry
hatches during the filter inspection that might allow entry of
rainwater.
CORROSION-RELATED EMISSIONS AND COST
During the first year of filter operation, there were un-
doubtedly excursions above the emission limits because of numer-
ous bag failures. Plant personnel, however, were not able to
quantify the frequency, duration, or severity of these excur-
sions. Plant personnel described deviations above regulatory
emission limits as infrequent during the first year of filter
operation and as especially rare since the filter bag rings and
hangers have been redesigned.
The costs of the corrosion failures have been limited to
those for replacing all filter bags after only 1 year of service
plus those for changing the hanger designs. The filter manu-
facturer assumed some of the financial burden associated with
these changes. The plant reported that a set of new bags costs
about $24,000. The total capital investment for the filter was
about $1 million.
REFERENCE
1. Hamner, N. E. Coatings for Corrosion Protection, Chap-
ter 14, NACE Basic Corrosion Course. National Association
of Corrosion Engineers, Houston, Texas, 1977.
A-61
-------
SOURCE 10
ELECTROSTATIC PRECIPITATOR SERVING PORTLAND CEMENT PLANT
SUMMARY
This wet process portland cement plant has a four-chamber
electrostatic precipitator (ESP) for control of particulate
emissions from the rotary kiln. During the first 8 years of
operation, several ESP components have failed as a result of
acid dewpoint corrosion. These failures have been repairable,
but it is expected that a major reconstruction of the ESP will
soon be required. The designed life of the control system is 30
years.
The potential for corrosion is inherent in an ESP serving a
wet process kiln, especially when coal is used as fuel. Corro-
sion could probably be prevented by maintaining higher kiln ex-
haust temperatures or by switching to sulfur-free fuels, but a
substantial penalty in fuel costs would be paid.
The source reported no major incidents of excess emissions
attributable to corrosion in the ESP. The costs of the corro-
sion have been substantial, however, and will undoubtedly in-
crease as additional ESP components require replacement.
SOURCE DESCRIPTION
Source 10 is a wet process portland cement plant with a
single, coal-fired kiln. Kiln exhausts pass through a four
compartment ESP for particulate removal. Gas flow through the
ESP is approximately 190 m3/s. Draft in the system is induced
by a fan located between the ESP exhaust and the 90-m plant
stack. Most plant equipment is 8 years old and represents state
of the art technology. The plant stack is the major remnant of
A-62
-------
a previous cement plant that was scrapped in order to build the
present more modern plant.
The kiln operates continuously for campaigns averaging 3
months in length. Chains have been attached to the inside of
the kiln to improve transfer of heat to the product. These
chains enable the kiln to operate using less fuel and, conse-
quently, the kiln exhaust temperature is lower than it would be
without the chains. The backend product temperature averages
about 540°C, and the exhaust gas at the ESP inlet is 230°C. The
outlet temperature of the ESP (in the breeching at the stack)
averages 145°C. Because the ESP is well insulated, the ESP
exhaust gas temperatures vary no more than 30°C between the
extremes of summer and winter.
Space constraints imposed by the location of the existing
stack required that the four ESP chambers be stacked in pairs
(Figure A10-1). Gas flow from the kiln divides into four paral-
lel streams, each passing through an individual chamber. Ad-
joining chambers share dust hoppers and screw conveyors. The
plant performs detailed maintenance on all ESP components twice
a year while the kiln is out of service for rebuilding of the
refractory.
The ESP chambers are constructed of 5-mm thick mild steel,
and the inside surfaces are uncoated. Structural support is
provided by box type steel beams (Figure A10-2). Because the
plant uses the wet process and is located in a cold climate, the
ESP chambers and hoppers are well insulated. The mild steel
ductwork is not insulated and there are no hopper heaters. The
collection plates are also mild steel and are hung in the
chambers as indicated in Figure A10-2. The discharge wires are
plow steel and the wire weights are cast iron. The kiln stack
is concrete with a steel ring at the top.
MALFUNCTIONS DUE TO CORROSION
The first corrosion-related malfunctions were in the sec-
ondary circuit insulator compartments. The casings of the
A-63
-------
CHAMBER 1
GAS1
INLET
CHAMBER 2
v~ — ' — "
FIELD
2A
~i
N
1 '
FIELD
2B
• — 7
Y-
FIELD
2C
7
\_5\ GAS
V ^OUTLET
^
,
DISTRIBUTION
PLATE
GAS
INLET
DUST HOPPER AND
SCREW CONVEYOR
(SHARED BY TWO
CHAMBERS)
1 ^
FIELD
4A
— 7
p ^
FIELD
4B
7
f — •
FIELD
4C
r^7
,GAS
I OUTLET
SCREW
CONVEYOR
Figure A10-1. General arrangement of Source 10 ESP.
A-64
-------
SHEET METAL SHEATH
OVER THERMAL INSULATION
THERMAL INSULATION,
7
ESP WALL,
0.5 cm
MILD STEEL
PLATE
COLD SPOT
"BOX" BEAM CONFIGURATION
INSULATION
GAP
MILD STEEL
"BOX" BEAM
GAS FLOW
COLD SPOT
(especially in the outer
plates and in lower
fields)
oooooooooooo
-PLATE HANGER
-COLLECTION PLATE
-PLATE STABILIZER
COLLECTION PLATE HANGING METHOD
Figure A10-2. Locations of cold spots within
the Source 10 ESP.
A-65
-------
insulator compartments corroded severely after only 1 year of
service. Source 10 retrofitted a heated, positive pressure
ventilation system to the insulator compartments to prevent
condensation. This system has been effective in preventing
corrosion within the compartments.
During semiannual ESP maintenance inspections, several
areas of corrosion have been identified within the ESP chambers.
These areas of corrosion seem to be within sections of the cham-
bers that are likely to be cooler and therefore more likely to
fall below the sulfuric acid dewpoint. One cold spot is at the
"box" beam structural supports for the chamber walls. As indi-
cated in Figure A10-2, these beams represent a gap in the ESP
chamber thermal insulation. Beam surfaces exposed to the sul-
fur-bearing flue gases have consequently suffered severe dew-
point corrosion. Corrosion proceeded from the exposed surfaces
and penetrated many of the beams. Condensation then occurred
within these beams and corrosion of the beams continued from
within. This corrosion could have caused a major structural
collapse of the ESP chambers if it had not been detected early.
The remedy for this problem was to repair the most severely
corroded portions of the damaged beams and to fill the centers
of all the beams with concrete. The concrete provides addi-
tional structural support and fills the void within the beams,
thereby limiting the possibilities of condensation. Concrete
can also inhibit corrosion where it contacts the surfaces of the
steel.*
Other cold spots occur within the ESP chamber near the cor-
ners formed by the chamber walls and the dust hoppers, and in
the corners of the hoppers themselves. These cold spots appear
to be most severe in the two lower chambers and near the cooler
outlet ends of the chambers. The hoppers of the upper chambers
are adjacent to the warm roofs of the lower chambers and are
therefore less subject to heat loss (Figure A10-1). Dewpoint
corrosion has occurred at several points near the cold spots.
Corrosion has occurred in both the ESP walls and in the hopper
A-66
-------
walls. Damage to these walls, although troublesome, can be
repaired by patching during regular maintenance. Cold spots,
however, extend into the lower reaches of the ESP chambers
causing corrosion at the bottoms of the ESP plates (Figure
A10-2). This has resulted in significant corrosion of the plate
structures, especially at the crevices formed where the stabi-
lizer bar is bolted to the plates. Several plates have corroded
completely away from the stabilizers. These unfastened plates
can contact discharge electrodes, taking a field out of service.
Replacement of the plates is a much more difficult and expensive
operation than the patching of the chamber walls described
above.
In addition to corrosion in the ESP chambers, there has
been corrosion at three other locations in the kiln flue gas
treatment system. Corrosion has occurred in the stack sampling
ports, the steel ring at the top of the stack, and in the duct-
work from the fan to the stack. Each of these components has
required repair or replacement during the first 8 years of plant
operation.
DESIGN EVALUATION
The use of coal to fire the Source 10 kiln adds significant
amounts of sulfur oxides to the exhaust gases, and the use of
the wet process in the blending of the raw materials adds sig-
nificant amounts of water vapor to the exhaust gases. Under
these circumstances, acid condensation and subsequent corrosion
can be avoided only by consistently maintaining exhaust gas
temperatures above the acid dewpoint.
Source 10 has attempted to maintain high temperatures in
the ESP through the use of extensive thermal insulation. This
strategy has been successful at most locations within the ESP,
but it has not worked at the more exposed extremities of the
chambers and hoppers. In addition, the tail-end components such
as ducting near the stack, the stack sampling ports, and the
stack ring also seem to be subject to acid condensation.
A-67
-------
Several options are available for dealing with the acid
condensation problem at Source 10. One option would be to
convert the plant from the wet process to the dry process. This
option would reduce the water vapor in kiln exhausts, thereby
lowering the acid dewpoint temperature. Disadvantages of this
option are the considerable capital expenditures involved in
converting raw materials blending facilities, the possible
detrimental effects the conversion might have on production
rates or
product quality, and the adverse effects the change may have on
the particle resistivity and collectability in the ESP.2
A second option is to switch from coal to natural gas or
low-sulfur fuel oil to decrease the amount of sulfur entering
the flue gases. This would also lower the acid dewpoint temper-
ature. Disadvantages of this option include higher fuel prices
and possible supply interruptions.
A third option is to fire the kiln at a higher temperature
so that exhaust gases enter the ESP at higher temperatures. The
obvious disadvantage of this option is the increase in fuel use.
A study of temperature variations within the ESP would be help-
ful prior to implementation of this option so that the minimum
required temperature increased can be defined.
A final option is to simply continue present operating
practices and to replace the ESP components as they fail.
Source 10 anticipates that a major rebuilding of the ESP will be
required after about 10 years of service because of corrosion
(design life for plant equipment is 30 years). Where feasible,
rebuilt ESP components should be constructed of materials more
resistant to sulfuric acid.
The ultimate choice among the above options will require
economic analysis as well as consideration of the technical
merits for each option.
A-68
-------
CORROSION-RELATED EMISSIONS AND COSTS
No specific incidents of excess emissions are attributable
to corrosion in the ESP. Corrosion damage has not yet caused
catastrophic failure of the ESP system. Damage to the ESP has
been repaired during periods when the kiln is out of service.
The costs of the corrosion have been substantial. The
total installed cost of the ESP was approximately $800,000 (1970
dollars) and the auxiliary equipment such as pilings, noise bar-
riers, rotary air locks, screw conveyors, insulator compartment
blowers, etc. cost more than $1 million (1970 dollars). It is
expected that most of the ESP components and some of the auxil-
iary components will require major repair or replacement after
only 10 years of the 30-year design life.
REFERENCES
1. Hamner, N. E. Coatings for Corrosion Protection, Chapter
14, NACE Basic Corrosion Course. National Association of
Corrosion Engineers, Houston, Texas, 1977.
2. Kulujian, N. J. Inspection Manual for the Enforcement of
New Source Performance Standards: Portland Cement Plants.
EPA 340/1-75-001, September 1975.
A-69
-------
SOURCE 11
FABRIC FILTERS SERVING SLAG CUPOLAS
SUMMARY
This plant includes two slag cupolas that are controlled
for particulate emissions by fabric filters. During the first 3
years of operation, severe degradation has occurred in the
filtering medium and also moderate corrosion in the mild steel
filter enclosure. The causes of these material failures have
been a lack of adequate flue gas temperature control within the
filter and the presence of sulfur oxides, water vapor, and
traces of fluorides in the raw materials.
A manually controlled dilution air cooling system has
allowed cupola flue gas temperatures to vary widely. During the
initial years of operation, the filters apparently operated
below the sulfuric acid dewpoint a significant fraction of the
time. These moist conditions led to the destruction of the
fiberglass bags and to corrosion of the mild steel filter enclo-
sures. The presence of unexpected traces of fluorides in the
flue gas probably accelerated the degradation of the fiberglass.
Source 11 replaced the fiberglass bags with polyester bags that
were subsequently destroyed by a high-temperature excursion.
Additional high-temperature excursions have been prevented by
installing high-temperature alarms in each filter. The acid
condensation problem has not been completely solved, but the
severity of acid condensation has been decreased by operating
the cupolas with slightly less dilution air added to the flue
gases. Fluorides have been virtually eliminated by careful
selection of slag feedstocks for the plant.
A-70
-------
SOURCE DESCRIPTION
Source 11 consists of two mineral wool cupola lines, each
containing a slag cupola, a mineral wool spinner, a packaging
line, and a fabric filter for treating the cupola exhausts.
Both cupola lines are designed to operate 24 h/day, but during
the first 3 years of operation market conditions have not per-
mitted full operation. In addition, production has been inter-
rupted periodically by a variety of production line or fabric
filter malfunctions. The actual operating schedule during the
first years of operation has been characterized by numerous cold
startups. At the present time, market conditions are improved
and most production problems have been solved; thus the cupola
lines are approaching their designed continuous operation.
Each cupola is a water-cooled steel cylinder with a charg-
ing door near the top, a trap door type bottom, and a slag pit
beneath the cylinder bottom. Combustion air is provided by
tuyeres around the lower portion of the cupola, and dilution air
for flue gas cooling enters through the charging door and
through a damper on top of the cupola. Flue gases exit the
cupola through a duct that leads to the fabric filter.
After startup, each cupola is charged at approximately
20-minute intervals. Raw materials for the cupolas include
steel-mill and copper-smelter slags as well as metallurgical
coke fuel. Cupola exhausts contain a mixture of fly ash, soot,
and sulfur compounds, which can vary in chemical composition as
raw materials vary. Sulfur compounds originate from the combus-
tion of the coke as well as from the melting of some of the
slags. The steel slags typically contain 1.3 percent sulfur,
and the copper slags contain negligible amounts of sulfur.
Copper slags, however, can contain as much as 1 percent lead as
well as traces of fluorides.
During the first years of operation, both copper slag and
steel slag were used. Determining factors in their use were
relative costs, availability, and the quality of mineral wool
A-71
-------
they produced. More recently the copper slag has not been used
in an attempt to reduce the exposure of workers to lead and to
reduce the amount of fluorides entering the fabric filter.
Exhaust gas temperatures typically ranged between 55°C
after a charge to 105°C just prior to a charge, during the first
years of operation. Later, the source improved exhaust gas
temperature regulation (through more careful operation of the
cupola and the dilution air damper) so that exhaust gas tempera-
tures typically ranged between 70° and 165°C. During most of
this early period of operation, however, the cupolas were sub-
ject to an unknown number of temperature runaways. This problem
has only recently been corrected by the use of high-temperature
alarms that give the operators sufficient time to control unusu-
ally high exhaust gas temperatures.
Each fabric filter contains four compartments and is equip-
ped with both reverse-air and shaking modes of cleaning. The
filters are also designed so that individual compartments may be
isolated with dampers to allow for on-line repair. The filter
shells are made of carbon steel with no coating on the inside.
The filter walls and roof are thermally insulated. The original
filter bags were fiberglass, but Source 11 has replaced these
with polyester bags. Gas flow through the filters is induced by
fans located between the filters and the stacks.
MALFUNCTIONS DUE TO CORROSION
During the first 3 years of operation, several components
of the filters have failed because of materials breakdown. The
most troublesome of these failures have been repeated failures
of the bag fabric. Fiberglass fabric was originally chosen
because of the expected high temperatures. Fiberglass bags,
however, failed within a few months of service. The failure of
the fiberglass bags appeared to be the result of self-abrasion
of the fibers during the shaker cleaning mode. Subsequent
chemical analysis of the cupola raw materials revealed traces of
A-72
-------
fluorides in the slags, which probably played a significant role
in the degradation of the fiberglass material. The replacement
polyester bags also failed once, because of a high temperature
excursion.
Other materials problems in the filter have included corro-
sion of copper tubing serving the filter magnehelic gauges and
general corrosion of the carbon steel shells. This latter prob-
lem has not yet resulted in any component failures, but if per-
mitted to continue, it may threaten the structural integrity of
the filters.
DESIGN EVALUATION
The principal design deficiency in the filter systems is
the inability to maintain cupola exhaust gas temperatures below
the maximum safe level for the filter fabric, but above the
sulfuric acid dewpoint. Exhaust gases are cooled by dilution
air, but the dilution air dampers do not automatically respond
to process changes. Typical practice has been to set the damper
positions so that exhaust gases are cooled safely below the
upper temperature limit of the fabric. This can allow exhaust
gas temperatures to fall below the acid dewpoint during cupola
startup or immediately after cupola chargings.
Failure of the fiberglass bags probably resulted from the
effects of condensed moisture and from the presence of fluorides
in the exhaust gases. Fiberglass fabric has a tendency to
self-abrade when flexed under moist conditions and its fibers
are weakened when both fluorides and moisture are present.1
After the first set of polyester filter bags were destroyed
by fire, Source 11 installed temperature alarms in the filters
to warn of high-temperature excursions. Since installation of
these alarms, there has been no further high-temperature damage
to the bags. Source 11 has not altered the filter systems in
any way to assure that flue gas temperatures remain above the
dewpoint. Automatic thermocouple-controlled dampers would
A-73
-------
probably improve control of low temperatures. This could reduce
the condensation of water and sulfuric acid and thereby reduce
degradation of the mild steel filter enclosures, as well as the
bag fabric and the copper magnehelic gauge tubing.
An additional factor contributing to internal moisture in
one filter enclosure was poor welding. Inspection of the inside
of this filter revealed a gap in one of the welded seams at the
roof of one compartment. This gap allowed cool ambient air and
rainwater to be drawn into the filter by the induced draft.
CORROSION-RELATED EMISSIONS AND COSTS
Source 11 maintains no records that would allow a quantifi-
cation of either excess emissions or costs resulting from mate-
rials failures. Excess emissions caused by materials failures
have consisted of substantially decreased particulate collection
efficiency because of to bag failures. The number and duration
of these episodes cannot be determined, nor can the emissions
rate during these episodes. Costs accrued because of filter
system material failures have included labor costs and materials
costs associated with replacing the bag sets several times, the
costs of lost production because of shutdowns required during
repair of the filters, and the costs of a decrease in the useful
life of the filter enclosures.
REFERENCE
1. Strauss, W. Industrial Gas Cleaning, second edition.
Pergamon Press, New York, 1975, p. 298.
A-74
-------
SOURCE 12
FABRIC FILTER SERVING GAS-FIRED AGGREGATE ROTARY DRYER
SUMMARY
This facility operates a gas-fired rotary dryer as part of
a process to prepare steel-mill slag for use in the glass indus-
try. Exhausts from the dryer were originally treated for par-
ticulate removal by a reverse-air, shaker-type fabric filter.
After approximately 1 year of service, the filter was retired
because of severe corrosion and chronic plugging. The filter
was replaced by a venturi scrubber, which has given good service
for about 1 year. The apparent cause of the filter corrosion
and plugging was water and acid condensation. Condensation
occurred because ductwork from the dryer to the filter was quite
long and exposed to ambient air. The average January tempera-
ture at the plant site is approximately 0°C, and temperatures
of -15°C are not uncommon during the winter.
SOURCE DESCRIPTION
Source 12 purchases blast furnace slag from a neighboring
steel mill and crushes it to specifications for use as a raw ma-
terial in manufacturing glass. Wet slag is received in rail
cars directly from the steel mill slag pit. Wet chunks of slag
are fed into the rotary dryer countercurrent to the direction of
gas flow. A natural gas-fired burner supplies heat to the
dryer. The dryer has the capability to use No. 2 fuel oil in
lieu of natural gas, but natural gas has generally been availa-
ble more than 99 percent of the time. The drying process gener-
ates a large quantity of dust that is separated from the dryer
exhaust gases by the particulate control system. Dried slag is
A-75
-------
segregated according to size and crushed in a series of crush-
ers, using a Stedman mill. The slag is crushed to a 16-mesh
size and is sold as a feed stock for glass production. The
plant operates one or two shifts a day, depending on the market
for its product.
The first particulate control system at the plant was a
fabric filter that used both reverse-air and shaking mechanisms
for cleaning the fabric. Draft for the system was provided by a
fan between the dryer exhaust and the filter. The filter was
installed in an unused area of a building adjacent to the dryer,
a location that required a very long outdoor duct between the
fan and the filter. The filter was designed for an inlet tem-
perature of 70°C. After a short period of operation, it became
apparent that dryer exhausts were being cooled to below the acid
dewpoint during transit through the duct; thus, a modest layer
of insulation was placed around the duct.
The fabric filter was eventually rendered inoperative by
corrosion and plugging, and was replaced by the present scrubber
system. The scrubber system consists of a venturi and a cy-
clonic mist eliminator. Because the scrubber is located adja-
cent to the dryer and the fan, considerably less ductwork is
required than for the fabric filter system. Scrubbing water for
the system is recirculated, and system blowdown is clarified in
settling ponds. There is presently no pH control or other
chemical treatment for scrubbing water. Both the venturi and
the mist eliminator are constructed of mild steel, but there
have been no corrosion failures during their first year of
operation.
MALFUNCTIONS DUE TO CORROSION
Corrosion first became evident in the fabric filter system
less than 3 months after startup. At this time, the interior
structures of the filter were hand-brushed and coated with
zinc-chromate paint. The paint was not effective because of the
A-76
-------
difficulty of properly preparing the filter surfaces prior to
application of the paint. Problems continued throughout the
first year of filter operation as a result of moisture condensa-
tion within the filter. Corrosion attacked all steel structures
within the filter and also portions of the ductwork between the
dryer fan and the filter. Insulation was placed around the
ductwork in an attempt to control temperatures within the fil-
ter, but gas temperatures apparently continued to fall below the
dewpoint. Within 1 year of startup, the filter had failed
completely because of corrosion and bag-blinding problems.
DESIGN EVALUATION
Process Characteristics
The corrosion in the fabric filter resulted because the
flue gas temperatures fell below the sulfuric acid dewpoint.
Coke used in the steel mill added sulfur to the slag, and cool-
ing water at the mill slag pit added moisture. The sulfur and
moisture were driven off during the drying process and condensed
as sulfuric acid within the filter. Attempts to prevent this
condensation by insulating the ductwork failed because the
modest layer of insulation was not sufficient to maintain high
gas temperatures in such a long duct. In addition, some of the
insulation had broken away from the duct after a year of ser-
vice.
Materials of Construction
The fabric filter structural parts and the associated duct-
work were made of mild steel, which is a poor choice for acid
mist service. Materials resistant to acid conditions, however,
would have been prohibitively expensive compared with other
design changes that would have prevented condensation.
The zinc-chromate coating applied to the fiter was not
totally inappropriate for the conditions in the system. The
field application over a hand-brushed surface, however, did not
A-77
-------
produce a high-quality coating. Had the coating been applied in
the shop over a sandblasted surface, it may have provided
partial protection from corrosion.
Equipment Arrangement
The configuration of the equipment was the primary reason
for acid dewpoint troubles in the fabric filter. If the filter
had been constructed nearer the dryer and if the filter enclos-
ure and the ductwork had been well-insluated, acid dewpoint
problems may have been prevented. The equipment manufacturer
was unable to foresee the acid dewpoint problems and was unable
to find a solution to the problems when they appeared.
CORROSION-RELATED EMISSIONS AND COSTS
The fabric filter system was so unreliable as a result of
corrosion and plugging that emissions were well above regulatory
limits during much of the year it was in operation. The local
enforcement agency permitted the source to operate without par-
ticulate controls whenever the filter system was malfunctioning.
The cost of the abandoned filter system was in excess of
$30,000. The source spent undetermined amounts for unscheduled
maintenance and repair work to the filter and for the scrubber
system that replaced the fabric filter.
A-78
-------
SOURCE 13
WET SCRUBBER SERVING ROTARY LIME KILN
SUMMARY
This plant is an 11-year old rotary lime kiln that has had
a succession of exhaust gas particulate control systems. The
first particulate control system was a relatively ineffective
settling chamber. Under pressure from the State Agency, mul-
tiple cyclones were added in series with the settling chamber.
These multiple cyclones improved particulate control considera-
bly, but emissions still remained above regulatory limits. The
plant then installed a carbon steel wet scrubber which met the
regulatory limits, but that failed because of corrosion during
the first year of operation.
Corrosion of the carbon steel scrubber was caused by the
presence of sulfur oxides and nitrogen oxides in the kiln ex-
hausts which formed sulfuric and nitric acids in the scrubbing
liquors. Corrosion affected the scrubber but not the quench
chamber because the bulk of the alkaline components in the flue
gases are removed in the quench chamber. Source 13 eventually
replaced most carbon steel scrubber system components with 316L
stainless steel components. The 316L alloy is withstanding the
corrosive scrubber environment.
Scrubber corrosion increased the particulate emissions con-
siderably during the first year of scrubber operation and was
very costly. Corrosion-related excess emissions probably ex-
ceeded 20 Mg during this period. Repair and replacement of
corrosion-damaged scrubber components cost approximately
$300,000 (1978 dollars). In addition, the plant lost between
$300,000 and $400,000 in corrosion-related lost production
during the first year of scrubber operation.
A-79
-------
SOURCE DESCRIPTION
The Source 13 lime facility began production approximately
11 years ago at an existing limestone quarry. The principal
component of the facility is the coal-fired rotary kiln, which
was moved to its present location from another plant. Other
components in the original facility, including a settling cham-
ber for particulate control, were installed new.
Because the settling chamber did not reduce particulate
emissions to regulatory limits, Source 13 added a set of mul-
tiple cyclones to improve collection efficiency. The multiple
cyclones reduced emissions substantially, but emissions con-
tinued to exceed the standards. Source 13 then installed a
carbon steel wet scrubber system. This scrubber system has
recently been replaced with a 316L stainless steel scrubber
system of nearly identical design.
The kiln is designed to run continuously for campaigns
lasting up to 60 days. Actual campaigns have lasted an average
of 20 days, with occasional shutdowns occurring in as few as 7
days. Coal sulfur content typically ranges between 1.0 and 1.25
percent. The kiln, as presently fired, uses significant excess
air, because of excessive air inleakage into the kiln.
Kiln exhaust gases exit through refractory brick-lined
carbon steel ducting. This ducting leads to the scrubber system
quench chamber that is designed to quench exhaust gases having a
maximum temperature of 1090°C. The quench chamber water pumps
can deliver up to 2270 £/min of water. Actual kiln exhaust
temperatures vary between 760° and 870°C, and actual water use
is not measured. Typical quench chamber outlet temperatures
vary between 90° and 120°C. Because quench water sprayed into
the vessel is not completely evaporated, the quench chamber also
serves as a spray type precleaner for the venturi scrubber. Ap-
proximately 70 percent of the particulate is collected by the
unevaporated water droplets in the quench chamber.
A-80
-------
Cooled and saturated gases pass from the quench chamber to
the venturi scrubber for final particulate removal. The venturi
scrubber is designed to operate with a pressure drop of up to
7.5 kPa, and its pumps can deliver up to 760 £/min of water.
Actual pressure drop across the venturi averages about 4.5 kPa
because of the air inleakage at the kiln. Pressure drop across
the venturi is measured by a permanently mounted differential
manometer. As in the quench chamber, water use in the venturi
scrubber is not measured. Venturi outlet gas temperatures vary
between 40° and 65°C.
The venturi scrubber is followed by a cyclonic dropout
chamber for collection of entrained water droplets, a 200 hp
induced draft fan, and the kiln stack. Each component from the
venturi to the kiln stack (including the ductwork) has been
replaced with 316L stainless steel. Water in both the quench
chamber and the venturi scrubber is once through. Spent slurry
from both vessels is combined into one stream and discharged
into an unused limestone pit. Combined pump capacity for both
vessels is 120 hp.
Gas temperatures in the scrubber system are measured at
three locations using thermocouples. Thermocouples at the
quench inlet and the quench outlet are connected to high temper-
ature alarms, but the thermocouple at the fan inlet has no
alarm. The scrubber system has a bypass damper for emergency
use, such as during a loss of water, but it is manually actuated
rather than actuated by the temperature alarms.
MALFUNCTIONS DUE TO CORROSION
During the first year of service, there were numerous
corrosion failures in the scrubber system. The first failure
was in the carbon steel induced draft fan, which failed after
only 4 weeks of service. This fan was rebuilt four times during
its first 8 months of operation. During one of the earlier
rebuilds, the carbon steel fan blades were replaced with 316L
A-81
-------
stainless steel because the blades seemed to suffer the most
severe corrosion. During the final rebuild, all remaining
carbon steel fan components were replaced with 316L stainless
steel. This new stainless steel fan has provided good service.
Many other scrubber components also failed during the first
year of operation. Failures included the venturi, dropout
chamber, ductwork, and stack, all of which were originally
carbon steel. Failures in these components required frequent
patching and occupied the attention of two of the maintenance
crew nearly full time for the first year of service. The plant
initially patched each failure by welding a carbon steel plate
over the affected area, but it was soon apparent that carbon
steel was unsuitable anywhere downstream from the quench cham-
ber. The remedy for corrosion in these carbon steel components
was eventual replacement with 316L stainless steel.
DESIGN EVALUATION
The present scrubber system at Source 13 represents a sig-
nificant improvement in particulate collection efficiency com-
pared with the multiple cyclones. When the scrubber system is
operating properly, the lime kiln meets applicable particulate
emissions regulations. Problems in the scrubber system have
been in reliability and have been caused almost exclusively by
corrosion.
When the State required Source 13 to upgrade the lime kiln
particulate control system, the plant considered a fabric filter
as well as wet scrubber systems proposed by several vendors.
The fabric filter alternative was rejected because the expected
maintenance requirements were too great a burden on the small
maintenance staff, because temperature control was expected to
be a problem, and because the filter was expected to require
more energy to operate than a scrubber. The existing multiple
cyclones had required a 250-hp fan, and a fabric filter required
A-82
-------
a 450-hp fan to handle kiln exhausts plus dilution air for pro-
tecting the filter fabric from high temperatures. The wet
scrubber system that was eventually installed required a 200-hp
fan.
The wet scrubber designs that were considered varied
greatly in costs, delivery times, and materials of construction.
One vendor recommended an expensive all-stainless-steel design
(alloy 316L). Another vendor recommended that only part of the
system be constructed of stainless steel; however, this vendor
had such a backlog of orders, that it could not promise delivery
for 2 years. Source 13 selected a third vendor who contended
that a less expensive scrubber constructed of carbon steel would
withstand any corrosive constituents in the exhaust gases be-
cause of the acid-neutralizing effects of the lime in the col-
lected particulate. This vendor was also able to construct and
deliver a scrubber without delay. This vendor had never de-
signed a scrubber for a lime kiln, but offered an 18-month
warranty with the system.
After the first series of corrosion failures in the carbon
steel scrubber system, Source 13 could not get satisfactory re-
sponse from the vendor concerning the warranty. At this point,
the plant management investigated, for the first time, scrubber
systems in other lime kiln applications. These investigations
revealed that carbon steel is generally not acceptable for
scrubbers serving coal-fired lime kilns.
Source 13 also had the kiln exhausts analyzed in two loca-
tions and the scrubbing liquors tested in three locations.
These chemical analyses revealed that significant quantities of
acid-forming gases were present in the exhausts. As would be
expected in a coal-fired process, quantities of sulfur trioxide
in the exhaust gases formed dilute sulfuric acid in the scrub-
bing liquors. Because the level of chlorides was found to be
low, it seemed that stainless steel would be a suitable mate-
rial. Because of the high level of excess combustion air used
A-83
-------
in the kiln, however, there were unusually high levels of nitro-
gen oxides in the gases. These formed high concentrations of
nitric acid in the scrubbing liquors, eliminating the possibil-
ity of using less expensive 302 or 304 stainless steel alloys.
Because of the nitric acid, the more expensive 316L alloys were
required.
Chemical analysis of the scrubber gases and liquors also
indicated that most of the alkaline constituents in the kiln
exhaust gases were scrubbed out in the quench chamber. Approxi-
mately 70 percent of the particulate collected by the scrubber
system appeared in the quench chamber outlet slurry and only 30
percent in the scrubber outlet slurry. Corrosion affected the
scrubber but not the quench chamber because of this uneven
-k.
distribution of lime-bearing materials in the slurries and be-
cause of a less violent mixing in the quench chamber. The water
sprays in the quench chamber were able to scrub much of the
alkaline particulate out of the flue gases, but allowed most
acid-forming gases to pass into the scrubber. In the scrubber,
a more violent mixing caused a greater transfer of the acids
into the scrubbing liquor that contained the smaller quantity of
alkalis.
The high water flow is necessary in the quench chamber to
prevent solids deposition and plugging within the vessel. Even
with this high water flow, the vessel must be cleaned fre-
quently. Attempts to recycle the scrubbing liquors have not
been successful because of plugging in the piping and pumps.
Source 13 regrets the purchase of the original scrubber
system, because of the vendor's lack of experience in lime kiln
scrubber applications and his recalcitrant attitude towards
honoring the warranty. Management still feels, however, that a
scrubber constructed of the proper alloys is the best alterna-
tive, because of the high particulate collection efficiency and
the low maintenance costs. Plant management also feels that the
State Agency complicated the decisionmaking process by requiring
action in too brief a time period.
A-84
-------
CORROSION-RELATED EMISSIONS AND COSTS
Corrosion-related excess emissions during the first year of
scrubber operation occurred during approximately 50 bypass
events. These bypass events were between 3 hours and 3 days in
length. Based on the estimated uncontrolled emissions rate
(Table A13-1), corrosion-related excess emissions during the
first year of scrubber operation may have exceeded 20 Mg. Stack
tests have shown that kiln emissions are below the regulatory
limits of 13.6 kg/h when the scrubber is operating properly.
TABLE A13-1. PARTICULATE EMISSION RATES FOR SOURCE 13 LIME KILN
Uncontrolled emissions
Emissions from original settling
chamber
Emissions from settling chamber
followed by multiple cyclones
Mass emissions regulations
Emissions from scrubber system
>136 kg/h
>91 kg/h
>27 kg/h
13.6 kg/h
<13.6 kg/h
The costs of the scrubber corrosion have been substantial.
The original carbon steel scrubber system cost $135,000 (1978
dollars). Source 13 has spent an additional $400,000 to repair
and replace corroded scrubber components. The scrubber manu-
facturer has been reluctant to honor the equipment warranty and
is still negotiating the warranty refund.
A greater cost to the plant has been the lost production
during corrosion-related shutdowns. These have cost the plant
an estimated $300,000 to $400,000. During the first 6 months of
scrubber operation, the kiln was shut down approximately 30 per-
cent of the time because of scrubber corrosion. In addition,
the shutdown during the rebuilding of the scrubber lasted 6
weeks. The State Agency did not permit the plant to continue
production using the scrubber bypass for a period of this
length. An additional cost of the scrubber corrosion is a de-
crease in the life of the kiln refractory because of frequent
startups and shutdowns. These costs cannot be precisely deter-
mined.
A-85
-------
SOURCE 14
FABRIC FILTERS SERVING LIME KILNS
SUMMARY
This facility contains one rotary lime kiln and a battery
of five shaft lime kilns. A fabric filter is used to remove
particulate from the rotary-kiln exhaust gases, and a similar
filter serves the five shaft kilns. Corrosion has not become a
problem in the rotary kiln filter even though the kiln is fired
with oil. The shaft-kiln filter is suffering severe corrosion,
however, despite the use of natural gas as fuel. The determin-
ing factor in filter corrosion is not the fuel used, but the
reliability of kiln exhaust gas temperature control. The batch-
type shaft kiln process allows frequent periods of low exhaust
gas temperatures, which promote moisture condensation on steel
surfaces within the shaft kiln filter.
Corrosion in the shaft kiln filter has not yet resulted in
the use of the filter emergency bypass. Filter efficiency, how-
ever, may have been reduced on occasion because of corrosion.
The excess emissions attributable to reduced filter efficiency
because of corrosion cannot be quantified. The costs of filter
corrosion have included a modest amount in repair work at the
filter access doors and a substantial decrease in the useful
life of the filter.
SOURCE DESCRIPTION
Source 14 consists of one rotary lime kiln and five shaft
lime kilns located at a limestone quarry. The rotary kiln is
fired with No. 2 fuel oil and operates continuously. The five
shaft kilns are fired with natural gas and operate intermittent-
ly.
A-86
-------
Exhaust gases from the rotary kiln are treated for particu-
late removal by a 5-year old shaker-type fabric filter. The
kiln is equipped with a heat exchanger to recover heat for
preheating the limestone aggregate and to cool the exhaust gases
from approximately 375°C to a temperature below the maximum safe
temperature for the filter bag fabric.
Exhaust gases from the five shaft kilns are treated for
particulate removal by a 4%-year old fabric filter similar to
the one serving the rotary kiln. The shaft kiln exhausts are
ducted to a common flue and then through a cooling loop that
cools the gases from a maximum temperature of approximately
350°C to a temperature compatible with the filter bag fabric.
Both fabric filters are constructed of corrugated transite
mounted to the outside of a carbon steel support structure.
Filter hoppers are carbon steel with thermal insulation on the
outside. Each filter is equipped with a temperature alarm that
actuates when inlet gas temperatures reach 290°C. At the sound-
ing of an alarm, plant personnel can divert gas flow through
bypass dampers to protect the bag fabric from heat stress.
MALFUNCTIONS DUE TO CORROSION
Corrosion has varied markedly in the two fabric filters at
Source 14. The filter serving the rotary kiln does not seem to
be greatly affected by corrosion, but the filter serving the
shaft kilns has suffered severe corrosion at the access doors
(Figure A14-1). The structural supports are also suffering from
corrosion, which will eventually threaten the structural integ-
rity of the shaft kiln filter (Figure A14-2).
An additional materials problem in both filters is the in-
ability of the filter compartment isolation dampers to seal
individual compartments from the kiln exhaust gases. The isola-
tion dampers are designed to allow maintenance activities within
an individual compartment while the kilns are operating. Be-
cause of heat warpage, the dampers do not seal sufficiently to
A-87
-------
c»
oo
*fe»-**&•*%;: vtf:**
*^S5L~ f^>^m^ ~W ,•*
;^;^%^^i- -A
:V*\:'^SSC?ii ; ^:i
Figure A14-1. Corrosion at shaft kiln filter
access door and door frame. Opening permits
inleakage of cool ambient air, accelerating
moisture condensation and further corrosion.
Figure A14-2. Corrosion of carbon steel beams
within the shaft kiln filter. Layers of lime
bearing deposits are visible on many surfaces.
The deposits have disbonded from the surface
of the vertical beam.
-------
maintain a safe working environment in an isolated compartment.
Corrosion of the dampers may also be contributing to this seal-
ing problem.
DESIGN EVALUATION
The differences in corrosion rates in the two filters can
be attributed to differences in inlet gas temperatures. Both
filter systems are designed to handle exhaust gases with temper-
atures of approximately 230°C so that neither condensation nor
heat stress of the fabric will occur. The continuously operated
rotary kiln seems to produce exhaust gases that are consistently
near the designed temperature. The temperature of the combined
exhausts from the shaft kilns, however, varies considerably.
The shaft kilns operate in a batch fashion with frequent
startups and shutdowns. The number of shaft kilns operating at
a time can range from none to five. The cooling provided by the
cooling loops can also vary, depending on the temperature of the
ambient air. These factors combine to allow gas temperatures in
the filter to fall so that moisture condensation on steel sur-
faces is frequent. Water vapor in the exhaust gases originates
from the limestone feedstock to the kilns, from combustion air,
and from the combustion of natural gas. Condensed water can
absorb traces of sulfur trioxide that may be present in the
exhaust gases and form dilute sulfuric acid (it is reported1
that natural gas contains an average of 4.6 kg of sulfur per 106
m3, and the limestone feedstock at Source 14 is reported to
contain 0.02 percent sulfur).
Corrosion is most severe in the corners of the filter and
at the uninsulated access doors, where exhaust gases are most
likely to be cooler. Significant corrosion has not occurred in
the insulated hoppers or in the warmer center of the filter.
The high concentrations of lime in the collected dusts have not
provided corrosion protection to the steel components. The
lime-bearing dusts form a solid cake of material that disbonds
A-89
-------
from the steel surfaces (Figure A14-2). Disbonded lime cake may
contribute to corrosion rather than mitigate it because moisture
can condense more readily under the cake. The cake layer appar-
ently insulates steel surfaces from the heat of the exhaust
gases within the filter, and it prevents ventilation of the
surfaces, which retards the drying of any condensed moisture.
CORROSION-RELATED EMISSIONS AND COSTS
Source 14 reported that corrosion in the fabric filters has
not resulted in any filter bypass events. All corrosion-related
maintenance has been accomplished during kiln shutdowns. It is
possible that corrosion in the filter may have reduced filter
efficiency on occasion. Air inleakage at corroded access doors
may affect filter performance and isolation damper leakage may
make speedy repair of broken bags impossible.
The costs of filter corrosion thus far have been limited to
the costs of replacing corroded access doors and door frames.
These costs have not been determined, but they are negligible in
comparison to the total value of the filters ($500,000 each in
1975 dollars). The principal expense caused by filter corrosion
is a decrease in filter life, which will necessitate early re-
placement of the shaft kiln filter. Source 14 had hoped to get
at least 10 years of service from the filters, but the shaft
kiln filter is not likely to last more than 7 years. The rotary
kiln filter, however, will probably provide more than 10 years
service.
REFERENCES
1. Compilation of Air Pollutant Emission Factors, Third Edi-
tion, EPA-AP-42, August 1977.
A-90
-------
SOURCE 15
WET SCRUBBERS SERVING ROTARY LIME KILNS
SUMMARY
This facility operates three coal-fired rotary lime kilns
that have a total of six wet-fan scrubbers for particulate
control. During the first 5 years of service the carbon steel
scrubbers have experienced corrosion, abrasion, and plugging
problems. Sulfuric-acid corrosion has occurred in the demister
sections of the scrubber vessels and in the scrubber stacks.
Abrasion has occurred in the precleaner chambers of the scrubber
vessels, where particulate loadings are the heaviest. Plugging
has occurred in both the precleaner sections and in the scrub-
bing liquor piping.
No excess particulate emissions from any of the scrubbers
can be attributed to corrosion. Scrubber performance, however,
appears to be adversely affected by plugging problems. The
costs of scrubber corrosion have totalled $30,000 (1979 dollars)
for replacement of corroded stacks. Scrubber plugging has cost
a significant, but undetermined, amount in lost production and
labor.
SOURCE DESCRIPTION
Source 15 consists of three coal-fired rotary lime kilns
located at a limestone quarry. Each kiln has two wet-fan scrub-
bers, arranged in parallel, for removing particulate from kiln
exhausts (Figure A15-1). In the two larger kilns, the scrubber
pairs are used simultaneously, while in the smallest kiln, the
scrubbers are alternated.
A-91
-------
CLEAN GAS OUTLET
LIQUOR
INLET"
DIRTY GAS INLET
(TANGENTIAL)
DEMISTER CHAMBER
DEMISTER
CHAMBER
WET-FAN
CHAMBER
PRECLEANER
CHAMBER
I
LIQUOR OUTLET
Figure A15-1. Source 15 wet-fan scrubber.
Each scrubber contains three chambers: a tangential inlet,
cyclonic precleaner chamber at the bottom of the vessel; a
wet-fan chamber in the middle; and a centrifugal mist eliminator
at the top. All scrubber vessel parts are carbon steel. The
fan wheel is type 304 stainless steel and the fan housing is
carbon steel. The scrubber stacks were originally carbon steel,
but they are being replaced with type 304 stainless steel
stacks.
Scrubber inlet gas temperatures vary between 425° and
540°C. The 112-kW fans provide a gas flow of 33 m3/s (at 540°C)
A-92
-------
through each scrubber. Scrubber pumps can provide up to 4500
£/min of water at a pressure of 700 kPa. Actual water use is
not measured, but approximately 900 A/min of particulate-laden
slurry drains to the settling pond through 30 cm piping, but
they now flow through an open flume. Detention time in the
settling pond is 26 h.
MALFUNCTIONS DUE TO CORROSION
Corrosion first appeared in the carbon steel stacks immedi-
ately above the dampers. This corrosion progressed to the point
where the stacks became structurally unsound and had to be re-
moved. Portions of the scrubber stacks above the dampers were
replaced with type 304 stainless steel. Four of these stacks
have been removed because the portions below the dampers have
continued to corrode and cannot safely support them (Figure
A15-2).
Additional corrosion has occurred in the demister chamber
of the scrubber vessel. Several small holes in the chamber
walls have required patching and several hatches have required
replacement (Figures A15-2 and A15-3).
Neither the lower two chambers, the fan, nor the slurry
piping have suffered significant corrosion. The lower chamber,
however, has been subject to abrasion and plugging. Within 3
months of scrubber startup, the walls of the lower chambers had
been penetrated by particulate abrasion. Plant maintenance
personnel have welded thicker carbon steel plate over the af-
fected areas. Plugging has occurred in slurry outlet piping
necessitating replacement with an open channel to facilitate
frequent cleaning. Solids have also accumulated in the lower
chamber of the scrubber, requiring frequent manual cleaning.
DESIGN EVALUATION
The original intention of Source 15 was to use a single
scrubber for each kiln and to have the second scrubber as a
A-93
-------
Figure A15-2. Two wet scrubbers in operation at
Source 15. Stacks for both scrubbers have been
removed because of corrosion. Hatch for entry
into the demister section of the nearest scrubber
has recently been replaced because of corrosion.
Figure A15-3. Entry hatch of a scrubber vessel
showing leaks resulting from lime deposits and
corrosion at the seal.
-------
standby. This scheme would have permitted cleaning and other
maintenance of the off-line scrubbers without interrupting pro-
duction. After installation of the scrubbers, it became appar-
ent that a single scrubber and fan would not provide sufficient
gas flow through the larger kilns to maintain optimum produc-
tion. Because of this, the plant had to abandon the alternating
scrubber scheme, except on the smallest kiln. The plant now
must run the two larger kilns for shorter campaigns than other-
wise possible, because they must be shutdown frequently to allow
the removal of solids from the scrubber vessels.
Prior to selecting the scrubber system, Source 15 invested
little time investigating the materials of construction to be
used. Plant management contends that the compliance schedule
imposed by the State Agency left little time to adequately con-
sider all aspects of scrubber design. The plant, which had been
out of service briefly under previous ownership, was required by
the State agency to submit completed emissions control plans
prior to returning the kilns to production. The firm had sev-
eral stainless steel scrubbers at another plant that were pro-
viding good service. Plant management, however, incorrectly
concluded that these expensive materials of construction were
not actually necessary in lime kiln service.
Corrosion in the Source 15 scrubbers appears in the upper
chamber and in the stack because of acid mist (sulfuric acid and
possibly nitric acid) formed by the mixing of the flue gases
with the scrubbing water. Acids in the lower chambers and in
the scrubber piping are neutralized by the heavy burden of
lime-bearing particulate collected in the slurries. Liquors in
the upper portions of the scrubber do not collect as much neu-
tralizing materials. The sulfur and nitrogen oxides absorbed
into these liquors form acids that are not completely neutral-
ized.
The exterior surfaces of the carbon steel scrubber vessels
are coated with a red lead primer that has remained intact.
This coating, and the lime deposits that have formed from mist
A-95
-------
fallout from the plume and from leaks in the scrubber vessel and
at plumbing fittings have prevented corrosion on most outer
surfaces of the vessels. One noticeable exception is at areas
just above leaking hatch seals.
CORROSION-RELATED EMISSIONS AND COSTS
Scrubber efficiency could not be estimated during the in-
spection. The two scrubbers that were operating, however, ex-
hibited a light residual of particulate in the plume beyond the
point of steam dissipation. As noted earlier, there was also
evidence of poor mist elimination. Inspection of an out-of-
service scrubber revealed severe solids buildup in the lower
chamber and additional, less-severe buildup in other chambers.
It appears that solids buildup, rather than corrosion, is the
primary cause of any deteriorations in scrubber efficiencies.
Corrosion-related costs have been limited to the costs of
replacing the six carbon steel scrubber stacks with type 304
stainless steel stacks after less than 5 years of service.
(Designed life for scrubber components is 15 years.) The stain-
less steel stacks cost approximately $30,000 (1979 dollars).
This represents an additional 12 percent in capital expenditures
added to the original installed cost of the scrubber systems.
There have been lost production costs due to plugging but none
due to corrosion. The costs of repairing abrasion damage in the
scrubbers are also undetermined.
A-96
-------
SOURCE 16
FABRIC FILTER SERVING INDUSTRIAL BOILER
SUMMARY
This coal-fired boiler provides process steam to an indus-
trial facility. Boiler particulate emissions have been control-
led by a shaker-type fabric filter since 1974. After approxi-
mately 1 year of operation, acid-dewpoint corrosion, hopper
plugging, excessive bag wear, and fan abrasion had combined to
render the filter ineffective. Lack of thermal insulation and
poor fan design were the principal causes of the filter system
operation and maintenance problems.
Source 16 flue gases were bypassed around the disabled fil-
ter for most of a 4-year period beginning in 1975. Costs of
repairs and modifications to the filter and the fan since start-
up have totalled more than $300,000, which is more than the in-
stallation cost of the filter.
SOURCE DESCRIPTION
Source 16 is an industrial facility that uses several
coal-fired boilers to provide process steam. Particulate emis-
sions from one of the boilers are controlled by the fabric
filter described in this report, and emissions from the others
are controlled by venturi scrubbers. The boiler served by the
fabric filter produces steam at the rate of 6.9 kg/s and uses
coal that contains between 1 and 2H percent sulfur and between
10 and 15 percent ash.
The retrofitted fabric filter is a four-compartment shaker-
type, designed to handle 21.5 m3/s of flue gases at 175°C.
Cleaning cycles in the filter are timed by an electronic control
A-97
-------
system actuated by the filter differential manometers. At the
time of installation, the filter was not insulated. Since that
time, insulation has been added in stages to the shell and hop-
pers, and a wind screen has recently been added. Draft in the
filter system is generated by a single inlet, high-speed fan
located between the boiler and the filter. Because of its high
rotation speeds and its location on the dirty side of the fil-
ter, the fan has surface-hardened Nibraze alloy steel to reduce
abrasion wear.
MALFUNCTIONS DUE TO CORROSION
Corrosion became a problem in the filter within a year of
installation. By 1975 corrosion failures and other maintenance
problems had made the filter system so ineffective that it was
removed from service. The filter was not in use during most of
the period from 1975 through 1978.
Fan
A single-inlet, high-speed fan was selected for the filter
system because of limited space between the boiler and the
retrofitted filter. The fan's single-inlet design concentrates
larger fly ash particulates to one side of the fan wheel where
their high velocities cause severe abrasion (Figure A16-1).
Maintenance personnel report that the fan will operate only 3 to
6 months before a major overhaul is required. Use of Nibraze
alloy steel has not reduced wear to acceptable levels.
Dust Hoppers and Filter Shell
The mild steel dust hoppers and the filter shell exhibited
signs of acid dewpoint corrosion within the first year of filter
operation. In addition, the hoppers were subject to frequent
plugging because of the condensed moisture. Insulation added to
the hoppers and the filter shell was not sufficient to prevent
flue gas temperatures from cooling to below the dewpoint within
A-98
-------
Figure A16-la. Source 16 fan wheel (during
overhaul) showing abrasion on blades.
Figure A16-lb. Single blade of Source 16 fan
wheel showing abraded channel along
right side.
-------
the filter. Insulation on the hoppers eventually became dis-
bonded (Figure A16-2). Insulation on the filter shell was
installed in such a manner that filter structural supports
protruded through to ambient air (Figure A16-3). During the
inspection, corrosion was especially apparent on interior filter
walls along these cooler structural supports. These steel beams
conduct heat from localized sections of the filter wall creating
conditions conducive to acid condensation. By protruding
through the insulation blanket, these beams can also channel
rainwater to the filter walls.
Dampers
Dewpoint corrosion also occurred in the filter compartment
dampers. These damper failures prevented proper sealing of in-
dividual compartments, preventing access to the filter for
on-line maintenance.
Cleaning Mechanism Controls and Filter Instrumentation
The shaker mechanisms for cleaning the filter bags have
suffered repeated malfunction because the electrical controls
have corroded. Shaker-mechanism limit switches and fuses as
well as filter instrumentation were not properly protected from
the weather and therefore suffered atmospheric corrosion.
DESIGN EVALUATION
There are several design deficiencies in the Source 16 fab-
ric filter. One deficiency is the excessive loss of heat that
has led to the dewpoint corrosion. The original uninsulated de-
sign was grossly inadequate for maintaining temperatures above
the dewpoint in the filter. Subsequent addition of insulation
to the filter shell and the hoppers was only partially success-
ful in preventing heat losses. Insulation on the hoppers did
not remain intact and there were gaps in the shell insulation.
A second design deficiency is the location and size of the
fan. The space available for a dirty side (forced draft) fan
A-100
-------
o
Figure A16-2. Thermal insulation on hopper
of Source 16 filter showing large sections
where insulation has become disbonded.
X
Figure A16-3. Exterior of Source 16 fabric filter
showing structural steel protruding through the
insulation blanket. The exposed structural steel
promotes cold spots within the filter and may also
channel rainwater under the insulation blanket.
-------
after retrofitting the filter to the boiler was quite limited.
As a result, Source 16 had to install a small single-inlet fan
which must be operated at very high speeds to maintain draft on
the boiler. The combination of abrasive fly ash and high fan
rotation speeds has made fan abrasion inevitable. Source 16 has
also indicated that, despite high fan speeds, the filter system
has not provided sufficient draft to operate the boiler at maxi-
mum capacity. Source 16 should consider replacing the forced
draft fan with a larger, slower speed induced draft fan, if
economically feasible. Although this design would require a new
stack and additional ductwork and would reduce space in the
plant yard, it could eliminate fan abrasion and boiler draft
problems.
A third design deficiency in the filter is that the filter
bags suffer premature failure because of excessive bag abrasion
at the bag inlets. Possible solutions to this problem include
changes in inlet design to include a wear plate or a wear thim-
ble, or the use of cyclone precleaners to remove abrasive heavy
particulates. The latter change could also reduce fan abrasion,
but it is doubtful that space is available for precleaner de-
vices.
The boiler superintendent at Source 16 has recently begun a
step-by-step program to repair and modify the filter so that it
will operate for extended periods of time without major break-
downs. The first steps have been to repair corroded shaker
mechanism electrical components and to erect a windscreen to
further reduce filter heat losses. Other measures may include
improving existing insulation and adding heaters to the hoppers.
CORROSION-RELATED EMISSIONS AND COSTS
Based on reported emission factors,1 uncontrolled boiler
particulate emissions at Source 16 are likely to be in the range
of 90 to 230 kg/h depending on the boiler firing rate and the
ash content of the coal. During most of the 4-year period
A-102
-------
beginning in 1975, boiler exhausts bypassed the inoperative
fabric filter. The exact quantity of particulate emissions
resulting from the uncontrolled conditions is unknown because
the boiler operating schedule, firing rate, and coal ash content
are not readily available.
Installed costs for the filter were about $250,000. Re-
pairs since filter startup have amounted to nearly $250,000.
Source 16 has also spent an estimated $30,000 to $40,000 annu-
ally on fan and duct repair.
REFERENCE
1. U.S. Environmental Protection Agency. Compilation of Air
Pollutant Emission Factors, Part A. Third Edition. AP-42,
August 1977.
A-103
-------
SOURCE 17
VENTURI SCRUBBER SERVING COAL-FIRED INDUSTRIAL BOILERS
SUMMARY
This industrial facility operates three coal-fired boilers
to provide steam for various plant processes and for winter
heating. Exhausts from the boilers pass through a retrofitted
venturi scrubber for particulate removal. The scrubber was
first constructed of 304 stainless steel, which failed cata-
strophically within 2 months. The replacement scrubber was
constructed of fiberglass-reinforced plastic (FRP), which has
given satisfactory service for 6 years. However, several bolts
and a refractory-lined carbon steel quench chamber in the new
scrubber system have experienced corrosion problems; these items
were replaced with Carpenter 20 stainless steel.
The failure of the original scrubber caused a $140,000 fi-
nancial loss to the source and the manufacturer and caused the
plant to operate uncontrolled for a year longer than specified
in the compliance schedule. The responsibility for the incor-
rect choice of materials must be shared by the State Agency as
well as by the source and the scrubber manufacturer. The State
forced an unrealistic compliance schedule (5 months) on the
plant management only weeks after the facility had changed
ownership; thus, the plant had no time to perform a complete
engineering analysis of the problem and had little time to
purchase and install the control device.
SOURCE DESCRIPTION
This facility uses three 6.3-kg/s boilers to provide proc-
ess steam and plant heating. No more than two boilers are used
A-104
-------
at a time; the third is maintained as a standby. Typically, the
two operating boilers are fired with low-sulfur Appalachian coal
producing a combined steam rate of about 8.8 kg/s.
Prior to 1971, the plant operated the three boilers without
any control of particulate emissions. Because the plant owners
were planning to close the plant, the State Agency took little
interest in the facility. In 1971, the present owners acquired
the plant. Shortly thereafter, the State Agency imposed a
compliance schedule on the plant that required installation of
particulate controls within 5 months. Four scrubber manufac-
turers submitted bids; three of the four recommended 316L stain-
less steel as the main construction material, and one recom-
.mended 304 stainless steel. The latter, who had experience in
the wet scrubbing industry, claimed that 304 stainless steel
would perform just as well as 316L stainless steel and that it
would cost significantly less. Consequently, the plant in-
stalled the venturi scrubber made of 304 stainless steel.
The scrubber vessel was designed as a 3.6-m diameter cylin-
der. Dirty air that entered the side of the vessel was forced
through a 2.5-cm annular venturi opening formed by a conical
skirt inside the vessel. The boiler flue gases passed through
this venturi opening into a pool of scrubbing water where gas
turbulence caused impaction of particulates by the water.
Cleaned flue gases passed out of the water through radial vanes
in the top of the scrubber vessel which imparted cyclonic motion
to the gases and caused the droplets of scrubbing water to col-
lect on the vessel sides. These scrubbed gases exited from this
demister section through a stack on top of the scrubber, and the
collected scrubbing water drained back into the bottom of the
vessel. The scrubber used once-through municipal water as the
scrubbing medium. The plant engineer reports that the pH of the
water exiting the scrubber was approximately 2.0.
The design of the replacement scrubber was nearly identical
to that of the original scrubber, but the material of construc-
tion was FRP instead of stainless steel. Because FRP materials
A-105
-------
were used, the flue gases had to be cooled before they entered
into the scrubber. Consequently, a refractory-lined, mild-steel
quench chamber was installed upstream of the FRP scrubber.
The new scrubber uses 227 £/min of municipal water; of
this, 76 £/min evaporates and 201 £/min are discharged with the
plant wastewater. The water is not recycled, and there is no pH
control in the system.
The scrubber typically handles about 31 m3/s of flue gases.
The pressure drop across the scrubber averages about 2.7 kPa. A
bypass damper can direct the flue gases around the scrubber sys-
tem to the old plant stack. The bypass is designed to activate
whenever the instrumentation detects a temperature excursion in
the scrubber, a loss of water in the scrubber, a loss of water
flow to the system, or an unusual pressure differential across
the scrubber.
MALFUNCTIONS DUE TO CORROSION
Stainless Steel Scrubber
Within 2 months after startup, the 304 stainless steel
scrubber was removed from service because of severe corrosion.
All internal surfaces of the scrubber had been attacked by the
acidic scrubbing water (Figure A17-1). Corrosion was particu-
larly severe on the venturi cone. The bottom of the skirt that
formed the venturi cone had disintegrated (Figure A17-2) and
caused a major loss of scrubbing efficiency. The corrosion of
the venturi skirt was probably accelerated by the turbulent
conditions in the venturi zone, which rapidly removed the corro-
sion products and thereby exposed new metal to attack.
FRP Scrubber
The FRP scrubber, which provided better service, has also
experienced some corrosion problems. Originally, the scrubber
internals and the vessel were bolted together with 316L stain-
less steel bolts; these failed after 6 months of service and
were replaced with bolts made from Carpenter 20 stainless steel.
A-106
-------
Figure A17-1. Inside of Source 17 scrubber vessel showing (from top to
bottom) corrosion of 304 stainless steel rotating vanes,
vessel wall, and venturi skirt after only 2 months service.
A-107
-------
Figure A17-2. Inside of Source 17 scrubbing vessel showing corrosion of
304 stainless steel venturi skirt after only 2 months service.
The refractory-lined, mild-steel quench chamber that was
installed to lower the flue gas temperatures at the scrubber
inlet has required frequent maintenance of the refractory lin-
ing. The plant plans to replace the quench chamber with one
constructed from Carpenter 20 alloy steel to eliminate the need
for periodic maintenance; however, there is some question about
whether the Carpenter 20 can withstand fly ash erosion.
DESIGN EVALUATION
Process Characteristics
The stoker-type boilers produce flue gases containing nor-
mal coal combustion products such as carbon dioxide, sulfur
oxides, nitrogen oxides, and fly ash. When these gases contact
water, they form corrosive agents such as sulfurous acid. The
use of low-sulfur coal produces lower concentrations of the
A-108
-------
sulfur species than the use of high-sulfur coal, but the flue
gases still contain enough sulfur oxides to be quite corrosive
in a wet environment. The plant engineer reported that the pH of
scrubbing water is sometimes as low as 2.0 with low-sulfur coal.
Coal can also contain up to 0.5 percent chlorides.1'2 These
chlorides are volatilized during combustion,3 and can be detri-
mental to most stainless steels.4
In summary, the environment produced by boiler flue gases
in a scrubber is extremely corrosive to mild steel and can be
corrosive to 304 and 316L stainless steels. The addition of a
pH buffer system to the scrubber might have alleviated much of
the corrosion, but such a system would increase the operating
costs and would require frequent maintenance.
Materials of Construction
Selection of the original scrubber materials was based on
misleading advice from the scrubber manufacturer, who claimed to
have experience in manufacturing weft scrubbers. The manufac-
turer reasoned that the success of 304 stainless steel in other
applications assured success in the scrubber for Source 17;
unfortunately the manufacturer neglected to determine the spe-
cific chemical constituents of the fly ash and the flue gases to
determine their compatibilities with 304 stainless steel.
Several other manufacturers advised the source that type
316L stainless steel would be an appropriate material. As indi-
cated by subsequent corrosion of 316L bolts in the FRP scrubber,
it is doubtful that these recommendations were sound.
The use of FRP for the second scrubber has proven to be ap-
propriate. During the 6 years since installation, there have
been no failures of the venturi skirt or any other FRP compo-
nent. Compared with stainless steel, FRP has had one disadvan-
tage—a low tolerance of the polyester resin to heat; this
characteristic required the installation of the quench chamber,
which increased system maintenance requirements.
A-109
-------
CORROSION-RELATED EMISSIONS AND COSTS
Corrosion failures in the original scrubber resulted in
about 1 year of additional uncontrolled boiler emissions.
Corrosion failures in the replacement scrubber system (e.g.,
corroded bolts and eroded refractory lining) have not resulted
in increased emissions, and they were corrected during routine
maintenance.
The costs of the problems with the first scrubber were re-
ported by the plant engineer to be $140,000 (1972 dollars) in
materials and labor; these costs were shared by the source and
the equipment manufacturer. The costs of materials failures in
the replacement system cannot be accurately determined since
they were part of the overall maintenance costs. Source 17
incurred no loss-of-production costs due to scrubber corrosion,
because the enforcement agency permitted the source to operate
uncontrolled after major equipment failures.
REFERENCES
1. Magee, R. A., F. B. Meserole,- and R. G. Oldham. Coal-Fired
Power Plant Trace Element Study: A Three Station Compari-
son. EPA Region VII, September 1975.
2. lapalucci, T. L., R. J. Demski, and D. Bienstock. Chlorine
in Coal Combustion. Report 7260. U.S. Department of In-
terior, Bureau of Mines, 1969.
3. Cato, G. A. Field Testing: Trace Element and Organic
Emissions from Industrial Boilers. EPA-600/2-76-086b,
1976.
4. Harpel, W. I., D. T. Murray, A. J. Graf fee, and J. C.
Steelhammer. The Chemistry of Scrubbers. Combustion,
47(9):33.
A-110
-------
SOURCE 18
WIRE-BURNING INCINERATOR CONTROLLED BY WET SCRUBBER
SUMMARY
This facility is a metals recycling plant that used a
wire-burning incinerator to burn insulation from copper and
lead-sheathed wires. Emissions from the incinerator were con-
trolled by a wet scrubber that used sodium hydroxide for pH
control.
Severe corrosion of the stainless steel scrubber stack and
general operating problems with the scrubber resulted in the
shutdown of the system after a limited life of 1 year. The
principal reason for the corrosion of the stainless steel was
hydrochloric acid in the incinerator flue gases arising from the
combustion of polyvinyl chloride (PVC) and chlorinated rubbers.
Although there were no uncontrolled emissions resulted from the
corrosion of the scrubber system, the plant suffered a major
financial setback because of the damage to the stack and the
lost production resulting from shutdown of the system.
SOURCE DESCRIPTION
The wire-burning incinerator at this source used a wet
scrubber to control particulate emissions. The incinerator was
put into operation in 1976 to burn the coatings from copper
wires so that the scrap copper could be recovered. The incin-
erator was intended for use on cloth, polyethylene, and other
nonchloride-bearing insulation. However, all types of coated
wire were fed to the incinerator including PVC and chlorinated
rubber-coated wire and lead-sheathed cables. The incinerator
operated on a batch basis, as needed.
A-lll
-------
From the incinerator, the flue gases were drawn through a
dry fan to the scrubber. The scrubber, which was constructed
from 316L stainless steel, used 38 £/min of recirculated water.
Sodium hydroxide was used to control pH; additions were based on
hourly measurements using pH paper. The water was recirculated
in copper pipes.
Flue gases from the scrubber passed to the stack. The
lower 6.1 m of stack was designed with a double shell and an air
anulus; reportedly, stainless steel was the material of con-
struction. The upper portion of the stack (about 4.5 to 6.1 m)
was a single shell design also fabricated from stainless steel.
MALFUNCTIONS DUE TO CORROSION
Fan
The fan, which is forced draft with respect to the scrub-
ber, operated at a temperature above the dewpoint and did not
have any corrosion problems.
Scrubber Shell, Internals and Pipework
In the 1 year of operation, corrosion problems did not
occur in the scrubber or associated pipework.
Stack
The upper stack, which was fabricated from stainless steel
(304 or 316), was subject to severe corrosion that weakened its
structure and necessitated its removal after a service life of
only 1 year. Attack initiated at welds but eventually spread to
the plates. Inspection of the upper stack was not possible be-
cause the stack had been removed from the site.
The lower portion of the stack to just above the roof line
was still in place. Visual inspection found it to be of a
double-shell design. The inner surface that contacted the flue
gas was coated with a bitumastic lining. The lining appeared to
be in good condition. The outer surface of the stack and the
surfaces facing the air space of the anulus were uncoated and
A-112
-------
were badly corroded. Support brackets between the inner and
outer shell had failed because of corrosion (Figure A18-1). The
corrosion hart oanQcH --;cmificant thinning of the stack walls.
Figure A18-1. The double-wall structure of the lower portions
of the stack. Note failed support bracket and general
corrosion.
DESIGN EVALUATION
Materials of Construction
The combustion of PVC and chlorinated rubbers generates
hydrochloric acid that can be particularly deleterious to stain-
less steels. The use of sodium hydroxide to control the pH of
the recirculating water has protected the components of the
scrubber, but the stack has been subject to severe corrosion.
Presumably the scrubber is not totally effective in removing the
acidic gases from the flue gases.
A-113
-------
CORROSION-RELATED EMISSIONS AND COSTS
Corrosion and operating problems with the pollution control
system resulted in its shut down approximately 1 year after in-
stallation. Although no costs were reported by the plant, it
was clear that the costs of replacing the corroded stacks would
represent a significant percentage of the plant's total opera-
ting costs. The plant attempted to seek a financial settlement
with the scrubber manufacturer, but had little success. At the
time the plant was inspected, the incinerator and scrubber had
been shut down for almost 2 years awaiting resolution of the
scrubber problems. The plant has recently purchased (at a
significant cost) a machine designed to cut the insulation from
the copper and lead wire. This machine will perform the func-
tions of the incinerator system until that system is again
operative. At the time of the inspection, the plant had re-
ceived sections of a replacement stack that was awaiting instal-
lation. The replacement stack was fabricated from carbon steel
lined with about 5 cm of a gunite-type material.
A-114
-------
APPENDIX B: REVIEW OF LITERATURE
ON CORROSION IN AIR POLLUTION
CONTROL EQUIPMENT
Bibliography and Subject Index
B-l
-------
BIBLIOGRAPHY
Adams, A. B., Jr. Corrosion Problems with Wet Scrubbing
Equipment. Resolving Corrosion Problems in Air Pollution
Control Equipment, National Association of Corrosion Engi-
neers, Houston, Texas, 1976.
Keyword: fiberglass-reinforced plastic, scrubbers (partic-
ulate)
Adams, R. L. How to Design a Fabric Filter to Operate in a
Corrosive Atmosphere. Resolving Corrosion Problems in
Air Pollution Control Equipment, National Association of
Corrosion Engineers, Houston, Texas, 1976.
Keyword: dust hoppers, fabric filters, insulation '
Air Pollution Abstracts. U.S. Environmental Protection
Agency,OfficeofAir Programs, Research Triangle Park,
North Carolina, Published monthly.
Keyword: published abstracts
Air Pollution Aspects of Emission Sources: Municipal
Incineration. A bibliography with abstracts, Report AP-92,
U.S. EPA, May 1971.
Keyword: incinerators, published abstracts
Ashbaugh, W. G. Materials Selection for Chemical Process
Equipment. Resolving Corrosion Problems in Air Pollution
Control Equipment, National Association of Corrosion Engi-
neers, Houston, Texas, 1976.
Keyword: corrosion test data, costs, failure analysis,
protective coatings
Ashbaugh, W. G. Materials Selection for Chemical Process
Equipment. Presented at the Air Pollution Control Asso-
ciation Seminar on Corrosion Problems in Air Pollution
Control Equipment, Atlanta, Georgia., January 17-19, 1978.
Keyword: corrosion test data, costs, failure analysis,
protective coatings
B-2
-------
7. Ayer, F. A. (compiler). In: Proceedings of Symposium on
Flue Gas Desulfurization - Hollywood, Florida, November
1977, Vol. I. EPA-600/7-78-058a, March 1978.
Keyword: scrubbers (FGD)
8. Ayer, F. A. (compiler). In: Proceedings of Symposium on
Flue Gas Desulfurization - Hollywood, Florida, November
1977, Vol. II. EPA-600/7-78-058b, March 1978.
Keyword: scrubbers (FGD)
9. Balasic, P. J. Electrostatic Precipitator Corrosion Prob-
lems on Recovery Boiler Applications. Paper No. 185 pre-
sented at the National Association of Corrosion Engineers'
Corrosion/79, Atlanta, Georgia, March 12-16, 1979.
Keyword: acid dewpoint, chlorides, electrostatic precipi-
tators, paper mills, pH control
10. Banks, J. H. Corrosion Control with Fiberglass-Reinforced
Plastics in the Power Industry. Presented at the Air
Pollution Control Association Seminar on Corrosion Problems
in Air Pollution Control Equipment, Atlanta, Georgia,
January 17-19, 1978.
Keyword: boilers (utility), fiberglass-reinforced plastic,
scrubbers, stack liners
11. Baum, B., and C. H. Parker. Role of Plastics, Especially
PVC, in Incinerator Corrosion. Polymer News 2(5-6):37-47,
1975.
Keyword: hydrochloric acid, incinerators
12. Beggs, T. W., and U. M. Patankar. Accelerated Baghouse
Corrosion in a Waste Oil Burning Asphalt Concrete Plant.
Presented at the Air Pollution Control Association's 72nd
annual meeting, Cincinnati, June 24-29, 1979.
Keyword: asphalt plants, chlorides, fabric filters, oil
13. Bell, R. T. Protective Coatings and Positive Failure
Analysis. Resolving Corrosion Problems in Air Pollution
Control Equipment, National Association of r-nT-r-r.g-i rm Ev.a-i-
neers, Houston, Texas, 1976.
Keyword: protective coatings
14. Benzer, W. C. Steels. Chemical Engineering 77(22):
101-10, 1970. *
Keyword: cladding, costs, stainless steels
B-3
-------
15. Berger, D. M., R. J. Trewella, and C. J. Wymmer. Atlas
Test Cell Evaluation of Coating Material for SO2 Scrubber
Service. Paper No. 40 presented at the National Associ-
ation of Corrosion Engineers' Corrosion/79, Atlanta,
Georgia, March 12-16, 1979.
Keyword: corrosion test data, protective coatings, scrub-
bers (FGD)
16. Bibbo, P. P., and M. M. Peaces. Defining Preventive Main-
tenance Tasks for Electrostatic Precipitators. Power
119(8):56-58, 1975.
Keyword: electrostatic precipitators, maintenance
17. Booth, J. B. Control of Industrial Boiler Emissions.
Power 119(8):55-57, 1975.
Keyword: boilers (industrial), fiberglass-reinforced
plastic, scrubbers (particulate)
18. Borenstein, M. Special Construction Materials for Scrub-
bers. Air Pollution Control and Design Handbook, Ch. 40,
Marcel Dekker, New York, 1977.
Keyword: chlorides, fans, fiberglass-reinforced plastic,
linings, rubber, scrubbers
19. Boyd, W. K., and P. D. Miller. Materials Selection for
Design of Pollution Control Equipment. Paper 71-DE-12,
American Society of Mechanical Engineers, New York, 1971.
Keyword: ceramic and masonry materials, corrosion monitor-
ing, corrosion test data, incinerators, scrubbers
(particulate), stainless steels
20. Brasunas, A. deS., and N. E. Hamner (eds.). NACE Basic
Corrosion Course. National Association of Corrosion Engi-
neers, Houston, Texas, 1977.
Keyword: cathodic protection, corrosion inhibitors, corro-
sion monitoring, failure analysis, galvanic
corrosion and galvanic series, high-temperature
corrosion, special alloys, stainless steels,
textbooks (corrosion)
21. Bryers, R. W. (ed.), Ash Deposits and Corrosion due to
Impurities in Combustion Gases, Hemisphere Publishing
Corp., Washington, D.C., 1978.
Keyword: boilers, coal, fly ash, incinerators, oil
B-4
-------
22. Bump, R. L. Report Card on Electrostatic Precipitators.
Solid Wastes Management 20(3):16, 1977.
Keyword: electrostatic precipitators, incinerators
23. Buonicore, A. J., and E. S. Yankura. Cupola Emission
Control with Energy-Saving Centripetal Vortex Wet Scrubber:
A Pilot Investigation. Energy and the Environment - Pro-
ceedings of the Fourth National Conference, American Insti-
tute of Chemical Engineers, Dayton, Ohio, 1976.
Keyword: electrostatic precipitators, fabric filters, iron
and cceel industry, scrubbers (particulate)
24. Burda, P. A. Corrosion Protection of Wet Scrubbers.
Power Engineering 79(8):48-51, 1975.
Keyword: linings, scrubbers (FGD)
25. Burda, c>. A. Linear Polarization Method for Corrosion Rate
Measurements in Limestone Slurry Scrubber. Materials
Performance 14(6);27-31, 1975.
Keyword: corrosion monitoring, scrubbers (FGD)
26. Cato, G. A. Field Testing: Trace Element and Organic
Emissions from Industrial Boilers. EPA-600/2-76-086b, U.S.
EPA, Research Triangle Park, N.C., 1976.
Keyword: boilers (industrial), chlorides
27. Conybear, J. G. Corrosion Concerns in Waste Incinerators.
Resolving Corrosion Problems in Air Pollution Control
Equipment, National Association of Corrosion Engineers,
Houston, Texas, 1976.
Keyword: corrosion test data, incinerators
28. Corrosion (R. W. Staehle, ed.), National Association of
Corrosion Engineers, Houston, Texas, Published monthly.
Keyword: periodicals
29. Crow, G. L. Corrosion Tests Conducted in Prototype Scrub-
ber Systems. Presented at the Air Pollution Control Asso-
ciation Seminar on Corrosion Problems in Air Pollution
Control Equipment, Atlanta, Georgia, January 17-19, 1978.
Keyword: chlorides, corrosion test data, fiberglass-rein-
forced plastic, iron, plastics, rubber, scrub-
bers, steel
B-5
-------
30. Engdahl, R. B., and D. A. Vaughn. The Corrosion Research
Needs for Air Cleaning Equipment. Resolving Corrosion
Problems in Air Pollution Control Equipment, National
Association of Corrosion Engineers, Houston, Texas, 1976.
Keyword: incinerators, scrubbers (FGD), scrubbers (partic-
ulate)
31. Ewan, E. B. High Performance Functional Coatings. Paper
No. 259 presented at the National Association of Corrosion
Engineers' Corrosion/79, Atlanta, Georgia, March 12-16,
1979.
Keyword: protective coatings
32. Feige, N. D. Corrosion Service Experience and Economics of
Titanium's Usage in Gas Scrubbing Equipment for Refuse
Incinerators. Paper No. 138, presented at the National
Association of Corrosion Engineers International Corrosion
Forum, Chicago, March 4-8, 1974.
Keyword: costs, hydrochloric acid, incinerators, scrubbers
(particulate), titanium
33. Flanders, S. Miniguide to Elastomers. Materials Engineer-
ing 71(6):38, 1970.
Keyword: rubber
34. Fontana, M. G., and N. D. Greene, Corrosion Engineering,
2nd Ed., McGraw-Hill, New York, 1978.
Keyword: cathodic protection, corrosion monitoring, gal-
vanic corrosion and galvanic series, materials
testing, special alloys, stainless str '.Is, steel,
textbooks (corrosion)
35. Franconeri, P. Electrostatic Precipitatior Corrosion in
Incinerator Applications. Presented at the Air Pollution
Control Association's 68th annual meeting, Boston, June
15-20, 1975.
Keyword: electrostatic precipitators, incinerators
36. Frauenfelder, A. Overcoming Special Problems in Electrical
Precipitation. Filtration Separation 11(1) :52, 1974.
Keyword: aluminum plants, cement plants, electrostatic
precipitators
B-6
-------
37. Gilbert, W. Selecting Materials for Wet Scrubbing Systems.
Pollution Engineering 5(8):28-29, 1973.
Keyword: costs, linings, plastics, rubber, scrubbers,
stainless steels, steel
38. Gleason, T. G. How to Avoid Scrubber Corrosion. Air
Pollution Control and Design Handbook, Ch. 41, MarceT
Dekker, New York, 1977.
Keyword: chlorides, linings, rubber, scrubbers, special
alloys
39. Gleason, T. G. Halt Corrosion in Particulate Scrubbers.
Chemical Engineering 84(23):145-48. 1977.
Keyword: scrubbers (particulate)
40. Gleekm-n, L. W. Nonferrous Metals. Chemical Engineering
77(22):.111-18, 1970.
Keyword: aluminum alloys, copper alloys, nickel alloys,
special alloys, titanium
41. Gleekman, L. W. Nonferrous Metals. Chemical Engineering
79(27):47-49, 1972.
Keyword: cladding, special alloys, titanium
42. Graver, D. L. Forms of Corrosion. Resolving Corrosion
Problems in Air Pollution Control Equipment,National
Association of Corrosion Engineers, Houston, Texas, 1976.
Keyword: corrosion fundamentals, galvanic corrosion and
galvanic series
43. Haaland, H. H., and J. L. Ma. Corrosion Problems in Wet
Precipitator Design. Resolving Corrosion Problems in Air
Pollution Control Equipment, National Association of Corro-
sion Engineers, Houston, Texas, 1976.
Keyword: aluminum plants, electrostatic precipitators
(wet), lead, wood
44. Hall, H. J.; and J. Katz. Some Corrosion Problems and
Solutions in Utility, Cement, and Iron and Steel Elec-
trostatic Precipitators. Resolving Corrosion Problems
in Air Pollution Control Equipment, National Association of
Corrosion Engineers, Houston, Texas, 1976.
Keyword: boilers (utility), cement plants, electrostatic
precipitators, iron and steel industry
B-7
-------
45. Hamner, N. E. Corrosion Data Survey - Nonmetals, 5th Ed.
National Association of Corrosion Engineers, Houston,
Texas, 1975.
Keyword: concrete, corrosion test data, plastics, rubber,
wood
46. Hamner, N. E. Corrosion Data Survey - Metals, 5th Ed.
National Association of Corrosion Engineers, Houston,
Texas, 1975.
Keyword: corrosion test data, special alloys, stainless
steels, steel
47. Handwerk, R. J. Recycling Water Effectively. Foundry
100(7):40-43, 1972.
Keyword: iron and steel industry, scrubbers (particulate)
48. Hanf, E. W., and J. W. MacDonald. Economic Evaluation of
Wet Scrubbers. Chemical Engineering Progress -71(3):48,
1975.
Keyword: costs, scrubbers (particulate)
49. Harpel, W. L., D. T. Murray, A. J. Graffeo, and J. C.
Steelhammer. The Chemistry of Scrubbers. Combustion
47(9):33, 1976.
Keyword: fly ash, pH control, scrubbers (FGD)
50. Harpel, W. L., and J. P. Terry. Water Treatment Technique
for Preventing Scaling and Fouling in Gas Scrubbing Sys-
tems. Plant Engineering 32(2):155-57, 1978.
Keyword: aluminum plants, boilers, iron and £ -.eel indus-
try, scrubbers, water treatment
51. Harpel, W. L., and K. R. Lange. Water Treatment Chemical
Application to Gas Scrubber Systems. Air Pollution Control
and Industrial Energy Production, Ch. 13 (K. E. Noll, W. T.
Davis, and R. Duncan, eds.), Ann Arbor Science, Anr Arbor,
Michigan, 1974.
Keyword: aluminum plants, corrosion inhibitors, iron and
steel indsutry, scrubbers, water treatment
52. Hendricks, A. L., G. B. Wauters, and W. T. Singleton.
Vinyl Ester Coatings for Pollution Control Equipment.
Paper No. 41 presented at the National Association of
Corrosion Engineers' Corrosion/79, Atlanta, Georgia, March
12-16, 1979.
Keyword: protective coatings
B-8
-------
53. Holt, W. H.. Fiberglass-Reinforced Plastic Construction in
Baghouses. Resolving Corrosion Problems in Air Pollution
Control Equipment, National Association of Corrosion Engi-
neers, Houston, Texas, 1976.
Keyword: fabric filters, fiberglass-reinforced plastic
54. Hoxie, E. C. Discussion of Materials of Construction for
Wet Scrubbers for Incinerator Applications. Resolving
Corrosion Problems in Air Pollution Control Equipment,
National Association of Corrosion Engineers, Houston,
Texas, 1976.
Keyword: fans, incinerators
55. Hoxie, E. C., and G. W. Tuffnel. A Summary of INCO Corro-
sion TesJ-s in Power Plant Flue Gas Scrubbing Processes.
Resolvir Corrosion Problems in Air Pollution Control
Equipmf .^c~National Association of Corrosion Engineers,
Houstc ., Texas, 1976.
Keyword: corrosion test data, scrubbers (FGD)
56. Hughson, R. V. High-Nickel Alloys for Corrosion Resis-
tance. Chemical Engineering 83(24);125-36. 1976.
Keyword: nickel alloys
57. Humbert, C. 0. Factors Affecting the Failure of Discharge
and Collecting Electrodes in Electrostatic Precipitators.
Presented at the Air Pollution Control Association Con-
ference on the Operation and Maintenance of Electrostatic
Precipitators, Dearborn, Michigan, April 10-12, 1978.
Keyword: electrostatic precipitators
58. Hydrochloric Acid and Air Pollution: Annotated Bibliogra-
phy. Report No. AP-100, USEPA, Office of Air Programs,
Research Triangle Park, N.C., July 1971.
Keyword: hydrochloric acid, published abstracts
59. lapalucci, T. L., R. J. Demski, and D. Bienstock. Chlorine
in Coal Combustion. Report 7260, U.S. Department of the
Interior, Bureau of Mines, 1969.
Keyword: boilers, chlorides, coal
60. Jaffee, R. I., and R. H. Richman. A Program of Research on
Steel for Utility Applications: A Summary of a Workshop
Held at Palo Alto, California, December 1974. Report
EPRI/FP-274-SR, Electric Power Research Institute, Palo
Alto, California, November 1976.
B-9
-------
Keyword: boilers (utility), erosion, special alloys,
stainless steels, steel
61. Javetski, J. Solving Corrosion Problems in Air Pollution
Control Equipment - Part I. Power 122(5):72, 1978.
Keyword: chlorides, galvanic corrosion and galvanic se-
ries, scrubbers, special alloys
62. Javetski, J. Solving Corrosion Problems in Air Pollution
Control Equipment - Part II. Power 122(6):80, 1978.
Keyword: fiberglass-reinforced plastic, protective coat-
ings, vessel design, welding design
63. Johnson, G. H. Fan Applications on Corrosive Systems.
Resolving Corrosion Problems in Air Pollution Control
Equipment, National Association of Corrosion Engineers,
Houston, Texas, 1976.
Keyword: fans
64. Johnson, G. H. Fan Applications on Corrosive Systems.
Presented at the Air Pollution Control Association Seminar
on Corrosion Problems in Air Pollution Control Equipment,
Atlanta, Georgia, January 17-19, 1978.
Keyword: fans
65. Johnson, R. S. Materials Performance in a Flue Gas Partic-
ulate Removal and Desulfurization System. Presented at the
Air Pollution Control Association Seminar on Corrosion
Problems in Air Pollution Control Equipment, Atlanta,
Georgia, January 17-19, 1978.
Keyword: corrosion test data, iron, plasticr scrubbers,
stainless steels
66. Joneson, D. H. Hydrogen Analysis as a Method of Corrosion
Monitoring in Boilers. Combustion 51(2)-14-17, 1979.
Keyword: corrosion monitoring
67. Journal of the Air Pollution Control Association. (H. M.
Englund, ed.), Air Pollution Control Association, Pitts-
burgh, Pennsylvania, published monthly.
Keyword: periodicals
68. Kirchner, R. W. Corrosion of Pollution Control Equipment.
Chemical Engineering Progress 71(3):58-63, 1975.
B-10
-------
Keyword: corrosion test data, incinerators, paper mills
69. Klodt, D. T. Corrosion of Air Pollution Control Equipment
in the Mineral Industries. Mineral Industries Bulletin
16(1):1-14, 1973. '
Keyword: fans, iron and steel industry, sulfuric acid
plants
70. Kopecki, E. S. Corrosion Minimized/Efficiency Enhanced in
Wet Limestone Scrubbers. Power Engineering 80(40):86-89
]976. *
Keyword: chlorides, coal, scrubbers (FGD), stainless
steels
71. Lahr, P. T. Pumping Corrosive Scrubbing Liquids. Resolv-
ing Corrosion Problems in Air Pollution Control Equipment,
Natioi.il Association of Corrosion Engineers, Houston,
Texas, 1976.
Keyword: pumps, scrubbers
72. Landrum, R. J. Designing for Corrosion Resistance of Air
Pollution Control Equipment. Resolving Corrosion Problems
in Air Pollution Control Equipment, National Association of
Corrosion Engineers, Houston, Texas, 1976.
Keyword: scrubbers, vessel design, welding design
73. Landrum, R. J. Equipment. Chemical Engineering 77(22)-
75-82, 1970. a * l ;
Keyword: galvanic corrosion and galvanic series, vessel
design, welding design
74. Larsen, J. v., and J. H. Wilhelm. Corrosion Problems
Encountered in Flue-Gas Desulfurization Systems. Presented
at the Air Pollution Control Association Seminar on Corro-
sion Problems in Air Pollution Control Equipment, Atlanta,
Georgia, January 17-19, 1978.
Keyword: nickel alloys, scrubbers (FGD), special alloys,
stainless steels
75. Lewis, E. C., M. P. Stengel, and P. G. Marvin. Performance
of TP-316L SS and Other Materials in Electric Utility
Flue-Gas Scrubbers. Presented at the Air Pollution Control
Association Seminar on Corrosion Problems in Air Pollution
Control Equipment, Atlanta, Georgia, January 17-19, 1978.
Keyword: scrubbers (FGD), stainless steels
B-ll
-------
76. Lizlovs, E. A. Laboratory Corrosion Test for Stainless
Steels for S02 Scrubber Service. Materials Performance
17(3):36-37, 1978.
Keyword: corrosion test data, scrubbers (FGD), stainless
steels
77. Lomasney, H. L. Testing Wet Scrubber Lining Materials.
Paper No. 39 presented at the National Association of
Corrosion Engineers' Corrosion/79, Atlanta, Georgia, March
12-16, 1979.
Keyword: chlorides, linings, protective coatings, scrub-
bers (FGD)
78. Maier, J. H. Analysis of Wet, Dirty and Corrosive Combus-
tion Gases. Proceedings of the 20th Annual ISA Analysis
Instrument Symposium, Pittsburgh, Pennsylvania, May 12-15,
1974.
Keyword: acid dewpoint, combustion and combustion analysis
79. Materials Performance. (R. W. Staehle, ed.) National Asso-
ciationof Corrosion Engineers, Houston, Texas, Published
monthly.
Keyword: periodicals
80. McCarthy, J. E. Corrosion Considerations in an SO2 Partic-
ulate Scrubber. Proceedings of the 2nd Pacific Chemical
Engineers' Congress, Denver, Colorado, August 28-31, 1977,
Vol. II (pp. 1310-13), American Institute of Chemical
Engineers, New York, 1977.
Keyword: corrosion test data, stainless steels, steel
81. McDowell, D. W. Jr. Problems in Wet Gas Scrubbirq Systems.
Presented at the Air Pollution Control Association Seminar
on Corrosion Problems in Air Pollution Control Equipment,
Atlanta, Georgia, January 17-19, 1978.
Keyword: ceramic and masonry materials, chlorides, fiber-
glass-reinforced plastic, fluorides, livings,
mist eliminators, rubber, scrubbers, stainless
steels
82. Mellan, I., Corrosion Resistant Materials Handbook, Noyes
Data Corporation, Park Ridge, New Jersey, 1976.
Keyword: corrosion test data
B-12
-------
83. Michaels, H. T., and E. C. Hoxie. Some Insight into Corro-
sion in SO2 Exhaust Gas Scrubbers. Presented at the Air
Pollution Control Association Seminar on Corrosion Problems
in Air Pollution Control Equipment, Atlanta, Georgia,
January 17-19, 1978.
Keyword: chlorides, costs, nickel alloys, scrubbers,
stainless steels
84. Michaels, H. T., and H. C. Hoxie. How to Rate Alloys for
SO2 Scrubbers. Chemical Engineering 85(13): 161-64, 1978.
Keyword: boilers, chlorides, coal, scrubbers, stainless
steels
85. Miller, P. D. Corrosion Studies in Municipal Incinerators.
Battelle Memorial Laboratories, Columbus, Ohio, 1972.
Keyword; incinerators, scrubbers, special alloys, stain-
less steels
86. Mistry, N. T. Material Selection for Gas Scrubbers.
Materials Performance 15(4):27-33, 1976.
Keyword: fiberglass-reinforced plastic, scrubbers, stain-
less steels
87. Mockbridge, P. C., and A. Saleem. Some Corrosion Problems
and Solutions Encountered in Utility Boiler Flue-Gas Scrub-
bing Installations. Resolving Corrosion Problems in Air
Pollution Control Equipment, National Association of
Corrosion Engineers, Houston, Texas, 1976.
Keyword: scrubbers (FGD)
88. Mockbridge, P. C., and D. W. McDowell, Jr. Materials and
Corrosion Problems in a Fly Ash Scrubbing System.
Materials Performance 13(4):13-17, 1974.
Keyword: chlorides, corrosion monitoring, fly ash, scrub-
bers
89. Moreland, P. J., and J. G. Hines. The Concept and Develop-
ment of Corrosion Monitoring. Materials Performance 18(2):
65-70, 1979.
Keyword: corrosion monitoring
90. Nannen, L. W., and K. E. Yeager. Status of the EPRI Flue
Gas Desulfurization Development Program. EPA-600/2-76-
136a, Presented at the Flue Gas Desulfurization Symposium,
New Orleans, Louisiana, March 8-11, 1976.
B-13
-------
Keyword: boilers (utility), fly ash, mist eliminators,
scrubbers (FGD)
91. Nowak, F. Corrosion Problems in Incinerators. Combustion
40(5):32-40, 1968.
Keyword: incinerators
92. Olglesby, S., Jr., and G. B. Nichols. Maintenance of
Electrostatic Precipitators. Electrostatic Precipitation,
Ch. 13, Marcel Dekker, New York, 1978.
Keyword: dust hoppers, electrostatic precipitators, main-
tenance
93. Paul, G. Dealing with High Chloride Concentrations in
Closed-Loop Sulfur Dioxide Scrubbers. Industrial Water
Engineering 15(l):24-28, 1978.
Keyword: chlorides, protective coatings, scrubbers (FGD)
94. Pfoutz, P. D., and L. L. Stewart. Electrostatic Precip-
itators - Materials of Construction. Chemical Engineer-
ing Progress 71(3):53-57, 1975.
Keyword: electrostatic precipitators
95. Pierce, R. P. Estimating Acid Dewpoints in Stack Gases.
Chemical Engineering 84(4):125-28, 1977.
Keyword: acid dewpoint, coal, oil
96. Pitcher, J. H. Stainless Steels: CPI Workhorses. Chemi-
cal Engineering 83(24) :119-24, 1976.
Keyword: stainless steels
97. Ribald, E. Corrosion Guide, 2nd Ed., Else'^ .er Publishing
Co., New York, 1968.
Keyword: corrosion test data
98. Sakol, S. L., and R. A. Schwartz. Construction Materials
for Wet Scrubbers. Chemical Engineering Progress 70(8):
63, 1974.
Keyword: scrubbers (particulate), titanium
99. Semran, K. T. Emission of Fluoride from Industrial Pro-
cesses - A Review. Journal of Air Pollution Control Asso-
ciation 8(2):92-108, 1958.
Keyword: fluorides
B-14
-------
100. Sheppard, W. L. Uses of Chemically Resistant Masonry in
Air Pollution Control. Presented at the Air Pollution
Control Association Seminar on Corrosion in Air Pollution
Control Equipment, Atlanta, Georgia, January 17-19, 1978.
Keyword: boilers (utility), ceramic and masonry materials,
incinerators, smelters
101. Shinskey, F. G. pH Controls for S02 Scrubbers. Air Pollu-
tion Control and Design Handbook, Ch. 37, Marcel Dekker,
New York, 1977.
Keyword: pH control, scrubbers (FGD)
102. Sludge-Free Elimination of Sulfur Dioxide Stack Emissions.
Materials Performance 16(5):61-62, May 1977.
Keyword: scrubbers (FGD)
103. Smith, M. F. (ed.). Boiler Corrosion, NTIS/PS-78/0993,
National Technical InformationService, Springfield,
Virginia, September 1978.
Keyword: boilers, combustion and combustion analysis,
published abstracts
104. Smith, M. F. (ed.). Fly Ash, Vol. I, 1970-74, NTIS/PS-76/
0718, National Technical Information Service, Springfield,
Virginia, September 1976.
Keyword: combustion and combustion analysis, fly ash,
published abstracts
105. Smith, M. F. (ed.). Fly Ash, Vol. II, 1975 - July 1976,
NTIS/PS-76/0719, National Technical Information Service,
Springfield, Virginia, September 1976.
Keyword: combustion and combustion analysis, fly ash,
published abstracts
106. Sorell, G. Process Control. Chemical Engineering, 77(22):
83-88, 1970.
Keyword: costs, pH control, special alloys, stainless
steels, steel
107. Steel, C. J. Corrosion Protection Strategy for Pollution
Control Equipment. Pollution Engineering 10(3):49-50,
March 1978.
Keyword: cathodic protection, cladding, corrosion funda-
mentals, protective coatings
B-15
-------
108. Stern, R. D. (chairman), W. H. Ponder, and R. C. Christman
(vice-chairman), Symposium on Flue Gas Desulfurization, New
Orleans, Louisiana, March 1976, Vol. I, EPA 600/2-76-136a,
May 1976.
Keyword: scrubbers (FGD)
109. Stern, R. D. (chairman), W. H. Ponder, and R. C. Christman
(vice-chairman). Symposium on Flue Gas Desulfurization,
New Orleans, Louisiana, March 1976, Vol. II, EPA-600/2-76-
136b, May 1976.
Keyword: scrubbers (FGD)
110. Stewart, J. F. Materials Considerations for Utility S02
Scrubbing Systems. Resolving Corrosion Problems in Air
Pollution Control Equipment, National Association of Corro-
sion Engineers, Houston, Texas, 1976.
Keyword: scrubbers (FGD)
111. Thaxton, L. A., and A. G. Zourides. Corrosion Problems in
Specific Pollution Control Equipment. Paper No. 196 pre-
sented at the 5th International Pollution Engineering
Exposition and Congress, Anaheim, California, November
9-11, 1976.
Keyword: acid dewpoint, dust hoppers, electrostatic pre-
cipitators, fabric filters, protective coatings
112. Tice, E. A. Corrosion Tests in Flue Gas Desulfurization
Processes. Materials Performance 13(4):26-33, 1974.
Keyword: scrubbers (FGD), smelters, stainless steels
113. Townsend, R. Corrosion Protection in Effluer,'-. Treatment
Plants. Proceedings of Heavy Duty Coatings Corrosion
Protection Conference, Brintex Ltd, London, England, 1972.
Keyword: protective coatings, water treatment
114. Tracy, G. W. Corrosion Factors in the Operation and Main-
tenance of Precipitators. Presented at the Air Pollution
Control Association Conference on the Operation and Main-
tenance of Electrostatic Precipitators, Dearborn, Michigan,
April 10-12, 1978.
Keyword: aluminum alloys, cement plants, electrostatic
precipitators
115. Velzy, C. 0. Materials of Construction for Wet Scrubbers
for Incinerator Applications. Resolving Corrosion Problems
B-16
-------
in Air Pollution Control Equipment, National Association of
Corrosion Engineers, Houston, Texas, 1976.
Keyword: fans, incinerators, nickel alloys, scrubbers,
stainless steels
116. Williams, J. E. A Summary of EPA Experience Related to
FGD Corrosion Problems and Solutions. Resolving Corrosion
Problems in Air Pollution Control Equipment, National
Association of Corrosion Engineers, Houston, Texas, 1976.
Keyword: scrubbers (FGD)
117. Zarfoss, J. R. Clean Air from Paper Mill Recovery Boilers
Without Corrosion. Resolving Corrosion Problems in Air
Pollution Control Equipment, National Association of
Corrosior Engineers, Houston, Texas, 1976.
Keyworr- electrostatic precipitators, paper mills
B-17
-------
SUBJECT INDEX
Acid-resistant brick - (see ceramic and masonry materials)
Acid dewpoint - 9, 78, 95, 111
Aluminum alloys - 40, 114 (see also special alloys)
Aluminum plants - 36, 43, 50, 51
Asphalt plants - 12
B
Basic oxygen furnaces - (see iron and steel industry)
Blast furnaces - (see iron and steel industry)
Boilers - 21, 50, 59, 84, 103
Boilers (industrial) - 17, 26
Boilers (utility) - 10, 44, 60, 90, 100
Carbon steel - see steel
Cathodic protection - 20, 34, 107
Cement plants - 36, 44, 114
Ceramic and masonry materials - 19, 36, 45, 81, 100
Chlorides - 9, 12, 18, 26, 29, 38, 59, 61, 70, 77, 81, 83, 84,
88, 93
Cladding - 14, 41, 107
Coal - 21, 59, 70, 84, 95
Combustion and combustion analysis - 78, 103, 104, 105
Concrete - (see ceramic and masonry materials)
Copper alloys - 40
Corrosion fundamentals - 42, 107 (see also textbooks)
Corrosion inhibitors - 20, 51
Corrosion monitoring - 19, 20, 25, 34, 66, 88, 89
Corrosion test data - 5, 6, 15, 19, 27, 29, 45, 46, 55, 65, 68,
76, 80, 82, 97
Costs - 5, 6, 14, 32, 37, 48, 83, 106
Cupolas - (see iron and steel industry)
B-18
-------
D
Dust hoppers -2, 92, 111
Elastomers - see rubber
Electrostatic precipitators - 9, 16, 22, 23, 35, 36, 44, 57, 92,
94, 111, 114, 117
Electrostatic precipitators (wet) - 43
Erosion - 60
Fabric filters - 2, 12, 23, 53, 111
Failure analysis - 5, 6, 20
Fans - 18, 54, 63, 64, 69, 115
Fiberglass-reinforced plastic - 1, 10, 17, 18, 29, 53, 62,
81, 86
Fluorides - 81, 99
Fly ash - 21, 49, 88, 90, 104, 105
Fuel oil - (see oil)
Galvanic corrosion and galvanic series - 20, 34, 42, 61, 73
Galvanizing - see cladding
H
High-temperature corrosion - 20
Hoppers - see dust hoppers
Hydrochloric acid - 11, 32, 58 (see also chlorides)
Incinerators - 4, 11, 19, 21, 22, 27, 30, 32, 35, 54, 68, 85,
91, 100, 115
Inhibitors - see corrosion inhibitors
Insulation - 2
Iron - 29, 65
Iron and steel industry - 23, 44, 47, 50, 51, 69
B-19
-------
K
Kraft recovery boilers - (see paper mills)
Lead - 43
Linings - 18, 24, 37, 38, 77, 81
Literature searches - see published abstracts
M
Magnesium alloys - see special alloys
Maintenance - 16, 92
Masonry - see ceramic and masonry materials
Materials testing - 34
Mild steel - see steel
Mist eliminators - 81, 90
Molybdenum - (see special alloys, stainless steels)
N
Nickel alloys - 40, 56, 74, 83, 115 (see also special alloys
and stainless steels)
Oil - 12, 21, 95
Paper mills - 9, 68, 117
Periodicals - 28, 67, 79
pH control - 9, 49, 101, 106
Plastics - 29, 37, 45, 65
Protective coatings - 5, 6, 13, 15, 31, 52, 62, 77, 93, 107,
• 111, 113
Published abstracts - 3, 4, 58, 103, 104, 105
Pulp and paper mills - (see paper mills)
Pumps - 71
R
Refractory materials - (see ceramic and masonry materials)
Rubber - 18, 29, 33, 37, 38, 45, 81
Rubber linings - (see linings, rubber)
B-20
-------
Scrubbers - 10, 18, 29, 37, 38, 50, 51, 61, 65, 71, 72, 81, 83,
84, 85, 86, 88, 115
Scrubbers (FGD) - 7, 8, 15, 24, 25, 30, 49, 55, 70, 74, 75, 76,
77, 87, 90, 93, 101, 102, 108, 109, 110, 112,
116
Scrubbers (particulate) - 1, 17, 19, 23, 30, 32, 39, 47, 48, 98
Sewage sludge incinerators - see incinerators
Smelters - 100, 112
Special alloys - 20, 34, 38, 40, 41, 46, 60, 61, 74, 85, 106
Stack liners - 10 (see also ceramic and masonry materials)
Stainless steels - 14, 19, 20, 34, 37, 46, 60, 65, 70, 74, 75,
76, 80, 81, 83, 84, 85, 86, 96, 106, 112, 115
Steel - 29, 34, 37, 46, 60, 80, 106
Sulfuric acid plants - 69
Textbooks (corrosion) - 20, 34
Titanium - 32, 40, 41, 98 (see also special alloys)
V
Vessel design - 62, 72, 73
W
Waste oil fuel - (see oil)
Water treatment - 50, 51, 113
Welding design - 62, 72, 73
Wood - 43, 45
B-21
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NC.
EPA 340/1-81-002
2.
. RECIPIENT'S ACCESSIOI
4. TITLE AND SUBTITLE
An Investigation of Corrosion in Participate
Control Equipment
5 REPORT DATE
February 1981
. PERFORMING ORGANIZATION CODE
PN 3470-3-EE
AUTHOFUS)
T. E. Mappes and R. D. Tems
8. PEFIFORMING ORG>
PERFORMING ORGANIZATION NAME AND ADDRESS
PEDCo Environmental, Inc.
11499 Chester Road
Cincinnati, Ohio 45246
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-01-4147
2. SPONSORING AGENCY NAME AND ADDRESS
Environmental Protection Agency
Division of Stationary Source Enforcement
Washington, D. C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
Final
V4. SPONSORING AGENCY CODE
EPA 340/1
15'SUPfh|MbNSSERVprojeSct officer is Mr. K. E. Foster, Environmental Protection Agency
(MD-7), Research Triangle Park, N.C. 27711 (Tel. 919-541-4571; FTS 629-4571)
16.ABSj^|Tdocument presents the results of an investigation of corrosion problems in
particulate control equipment. During the investigation, corrosion problems in 38
fabric filters, wet scrubbers and electrostatic precipitators were observed and 7
control equipment manufacturers were interviewed. The document also contains
guidelines to assist control equipment users and enforcement personnel in reducing
particulate emissions resulting from corrosion-related malfunctions. Key
parameters discussed include scrubber liquor composition, materials selection, the
sulfuric acid dewpoint, and thermal insulation. Controlled particulate emissions
sources discussed include rotary dryers and kilns, cupolas, incinerators, and
steam boilers.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Corrosion, Fabric Filters, Scrubbers
Electrostatic Precipitators, Cupolas,
Kilns, Boilers, Municipal Incinerators,
Sludge Incinerators, Cement Plants, Lime
Plants, Sulfuric Acid Dewpoint,
Scrubbing Liquors
~3. DISTRIBUTION STATEMENT
Release Unlimited
• .IDENTIFIERS/OPEN ENDED TERMS
Particulate Control
Equipment, Maintenance
19. SECURITY CLASS (This Report)
Unclassified
20. SECURITY CLASS (Thispage)
Unclassified
COSATI Field/Group
21. NO. OF PAGES
196
22. PRICE
EPA Form 2220-1 (R.v. 4-77) PREVIOUS EDITION is OBSOLETE
-------
-------
United States
Environmental Protection
Agency
Office of General Enforcement
Division of Stationary Source Enforcement Series
Washington, DC 20460
Official Business
Penalty for Private Use
$300
Publication No. EPA-340/ 1-8 1-002
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
If your address is incorrect, please change on the above label,
tear off, and return to the above address
If you do not desire to continue receiving this technical report
series, CHECK HERE D ; tear off label, and return it to the
above address
------- |