United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S7-83-049 Dec. 1983
4>ERA Project Summary
Long Term Optimum
Performance/Corrosion
Tests of Combustion
Modifications for Utility Boilers-
Host Site: Columbus and
Southern Ohio Electric
Company Conesville No. 5
P.S. Natanson
The report gives results of part of a
study of corrosion in large utility boilers
burning high-sulfur coal, as possibly
affected by combustion modifications
for decreased nitrogen oxide (NO,)
emissions. During the first part of this
study, each boiler was characterized to
learn the short term effects of various
combustion modifications on boiler
operation and emissions. Later, a Level
1 environmental assessment (EA) of
boiler operation was performed, followed
by tests to measure corrosion rates in
the furnace. Abo, two 30-day continuous
emission monitoring (CEM) tests were
performed. This report discusses work
on Boiler No. 5 at Columbus and
Southern Ohio Electric Co.'s Conesville
(OH) generating station. During the
short-term characterization tests, full
load NO, emissions (as equivalent NO»)
were reduced from 263 to 152 ng/J
without adverse side effects. NO,
emissions during the 30-day CEM tests
were log-normally distributed with a
mean of 180-224 ng/J and a geometric
dispersion of 1.13 - 1.19. The Level 1
EA revealed no unusual environmental
hazards resulting froivi low-NO. opera-
tion. For the nearly 2-year study,
waterwall corrosion rates (measured
ultrasonically) were similar to rates
under normal operation at about 1-2
mils/yr.»
This Project Summary was developed
by EPA's Industrial Environmental
Research Laboratory. Research Triangle
Park, NC. to announce key findings of
the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction
For coal-fired utility boilers, NO,
emission regulations can often be met by
using such combustion modification
(CM) techniques as decreased total
excess air flow. However, this could lead
to a chemically reducing atmosphere in
the furnace and an increased corrosion
potential. Under this program, methods
of decreasing NO, emissions were
studied (characterization testing) from
environmental and corrosion points of
view. To this end, the program included
four parts: boiler characterization. Level
1 environmental assessment (EA), corro-
"Reeden mart femitter with metric unit* nay use
the conversion fectors ft the end of thi*
Summery.
-------
sion testing, and 30-day continuous
emission monitoring (CEM). Several
large utility boilers were studied during
this long-term performance/corrosion
study. This report deals with the No. 5
boiler at Columbus and Southern Ohio
Electric Co.'s Conesville (OH) generating
station.
Background
NO, has long been considered an
undesirable component of the earth's
atmosphere. In parts-per-million concen-
trations, NO, is considered to be a
precursor to acid precipitation, photochem-
ical smog, and irritation in human
respiratory systems. Since the early
1970s, after coal-fired utility boilers were
identified as a major NO, source, EPA has
supported research for NO, control at
these large boilers.
Under the sponsorship of the U.S.
Environmental Protection Agency (EPA),
Exxon Research and Engineering Co.
studied the effects of boiler operating
conditions on NO, emissions. These and
similar studies led to improved operating
procedures for NO, reduction. The new
procedures (known as combustion modifica-
tions (CMs) included decreased excess
air flow and staged combustion. However,
these CMs could increase the chemically
reducing nature of the gases in certain
regions of the furnace, increasing the
corrosion rate on the fireside of the
boiler's waterwall tubes, especially in
boilers burning high sulfur coal.
The program reported here was a
follow-on to the earlier work. During this
program, furnace wall corrosion was
studied at several coal-fired utility boilers
designed to use CM for reducing NO.
emissions. Some of the CMs included the
use of overfire air and low-NO, burners to
meet the applicable new source performance
standards (NSPS) for NO, emissions (301
ng/J as N02).
Work Plan
The test program contained four
phases: boiler characterization; corrosion
testing by probes, panels, and wall
measurements; Level 1 environmental
assessment; and 30-day continuous NO,
monitoring.
During boiler characterization, effects
of various boiler CMs were assessed and
their ability to reduce NO, emissions
without causing short-term, adverse side
effects (such as increased slagging) was
evaluated. This resulted in optimized
operating procedures for NO, control and
system flexibility. During this time, flue
gas emissions and fuel and ash composi-
tions were monitored. Also, furnace gas
composition was monitored at various
locations along the furnace wall (furnace
gas tap sampling) to identify locally
corrosive environments for further study.
Corrosion testing was initiated after
boiler characterization. Three methods
(corrosion probes, corrosion panels, and
wall thickness mapping) were used to
evaluate corrosion.
Corrosion probes were used to evaluate
short term (30 to 1000 hour) corrosion
effects. In this method, pieces of wall
tube material (called corrosion coupons
or test rings) were inserted at five
different locations in the furnace for
various times and then weighed to
determine the rate of metal loss. This
qualitative indirect method did not give a
true reading of actual wall losses, but it
was an economically attractive alternative
to methods which require entering the
furnace during an outage to measure the
tubes directly, as discussed below.
The corrosion panel method was
probably the most reliable method used
for measuring corrosion. In this method,
eight sections of the boiler wall were
removed, and replaced by new sections
(called corrosion panels) on which the
tube wall thickness had been carefully
measured. About 2 years later, the
furnace was re-entered (during an
outage) and the panels were remeasured.
The wall thickness measurements were
made ultrasonically, a nondestructive
test in which high frequency sound
•waves are bounced off the inner wall of
the tube and the thickness is determined
from the echo time delay.
Wall thickness mapping involved
ultrasonic thickness (UT) measurements
of walls throughout the furnace. The first
set of such wall measurements occurred
during the outage in which the corrosion
panels were installed, and the second set
was about 2 years later (coincident with
the final panel measurements).
During these long term corrosion tests,
an EPA Level 1 environmental assessment
(EA) test was performed. It included
physical, chemical, and biological (toxicity)
analysis of the major streams entering
and leaving the boiler. The Source
Assessment Sampling System (SASS)
train was used to sample the flue gas
stream, while EPA Method 17 was used
to measure paniculate removal efficiencies
(i.e., electrostatic precipitator [ESP]
collection efficiency). Other streams
sampled included sluice water, coal, and
ESP hopper ash.
Twice, during the 2-year corrosion
exposure period, NO, and other flue
gases were monitored continuously for
30 days using a continuous emission
monitoring (CEM) system. The 30-day
CEM tests quantified flue gas emissions
over normal boiler load cycles and for a
longer period than was possible during
the characterization tests. These tests
also indicated how CEMs may be expected
to perform (reliability, accuracy, mainte-
nance needs, etc.) in similar situations.
Boiler Description
Conesville No. 5 is the fifth boiler
constructed at Columbus and Southern
Ohio Electric Company's (C&SOE) Cones-
ville (OH) generating station. Burning
nearly 200 tons/hr of high-sulfur coal
(high corrosion potential) to generate
more than 400 MW of electric power, its
20 burners use tangential firing in a
single-compartment, balanced-draft,
waterwall boiler to produce about 3,000,
000 Ib/hr of super heated steam (1000°F,
2800 psi). Designed by Combustion Engi-
neering (CE) Inc., it uses CMs (e.g., over-
fire air) to guarantee compliance with the
applicable NSPS for NO, emissions (301
ng/J NO, as N02), and is one of the first
utility boilers to be governed by this
NSPS.
Results
Boiler Characterization
During the first part of this study, Exxon
evaluated several CMs which could be
used to reduce NO, emissions. Three
were selected for detailed study: decreased
total excess air flow, increased overfire
air flow, and upward overfire air nozzle
tilt.
By optimizing these parameters (Figure
1), full load NO, emissions (as equivalent
NO2) were decreased from 263 to 152
ng/J without adverse, short-term side
effects. (The applicable NSPS emission
rate is 301 ng/J as NO2.)
In addition, by using the data from the
boiler characterization tests, an improved
operating procedure was adopted for
medium and full load operation.
As expected, flue gas and fuel analysis
for sulfur confirmed that most fuel sulfur
is converted to S02 and then delivered to
the scrubber for subsequent cleanup.
Furnace gas sampled in the combustion
zone (near the furnace waterwalls)
showed that the lowest CO concentrations
were in the hopper zone, and that the
region where corrosion was most likely
(low Oz, high CO) was in the chemically
reducing environment around the burner
zone. After the characterization work,
corrosion in this and other regions was
studied in greater detail using corrosion
-------
I
d
600
500
446
400
300
259
200
700
! "As Found" Operating Point
93
Load Full
OF A Tilt 0
Burner Tilt 0
Normal Air (-4.3% Ot)
•-A- Low Air (-3.0% Ot)
Note: 12
By decreasing the total excess air
flow, and opening the overfire air
dampers to divert some of the air
to the top of the furnace, NO,
emissions were decreased from
446 to 259 pp_m. i
Improved
Operating
Point
25 50
Overfire Air (OFA) Dampers. % Open
75
too
Figure 1. NO* versus overfire air, Columbus and Southern Ohio Electric Co., Conesville No. 5.
probes and ultrasonic tubewall thickness
measurement methods.
In the "as-found" condition, NOX
emissions from the Conesville No. 5
boiler were well within the applicable
NSPS limit (301 ng/J). As a result of this
program, the full load NOX emissions, as
NOa were reduced 42% (from 263 to 152
ng/J). These data were used to set up
and "fine tune" the boiler, resulting in
improved operating procedures with no
adverse, short-term side effects. Long-
term effects (such as corrosion) were
studied in detail as the program continued.
Level 1 Environmental
Assessment (EA) Testing
The Level 1 EA test at Conesville No. 5
involved the major inlet and effluent
streams crossing the plant's process
boundaries. The streams sampled included
the inlet and outlet of the electrostatic
precipitator (ESP), the ESP hopper catch,
the fuel feed, and the sluice water flush-
ing the furnace hopper.
The Level 1 test was performed as a
screening study to determine potential
pollutants in the various streams and
included three types of analysis: (1)
chemical analysis was used to determine
the chemical composition of the various
streams; (2) physical analysis of particles
entrained in various streams provided
information on shape and size; and (3)
biological analysis provided information
on the toxic or mutagenic effects of the
various streams on living matter.
The test showed that, as expected,
most of the flue gases leaving the
combustor and going to cleanup devices
are fixed gases (02, CO& Nz, etc.). Other
components of this stream include sulfur
compounds, hydrocarbons, and entrained
paniculate matter.
Measurements taken at the ESP'
showed that the particulars matter (fly
ash) in the flue gas stream represents
more than 60% of the total ash leaving
the system (the rest, slag or bottom ash,
exits by another stream; e.g., the sluice
water stream to the settling pond). Also,
more than 98% of the fly ash (mostly the
larger particles) is removed by cleanup
devices, never reaching the stack.
The particulate matter and entraining
flue gases contained low levels of
organics (mostly alkyl esters): concentra-
tions of most metals were near or below
detectable levels. (Table 1 lists the metals
data for solid and liquid samples; EPA
priority metals for wastewater are shown
first.) Organic matter on particulate
favored the smaller particles. At the
precipitator inlet, the particles were
generally spherical, and composed
mostly of silicate material.
The fly ash particles sampled from the ESP
hopper were less spherical than the parti-
cles at the ESP inlet, ranged from 1 to 40
//m, and, like the ash entering the ESP,
contained mostly silicates.
While these tests were underway,
supplemental tests were performed by
Acurex Corporation under separate EPA
contract. The Acurex data shows that, as
the sluice water passes through the
system, there is an increase in acidity,
total suspended solids, chemical and
biological oxygen demand, and total
organic carbon. The effluent sluice water
samples contained very low levels of
metal anions.
The coal extract was separated into
seven fractions, but the spectra of each
fraction were generally too complex to be
analyzed by EPA Level 1 procedures.
Instead, the major peaks are listed in the
full report, and show that the extract was
highly aromatic, highly oxygenated
material containing phenolic material,
carboxylic acids, amides, and aliphatics.
Most aromatic rings had two or more
substrtuents. Mono-substituted aromatics
were rare.
Biological tests (using living matter)
were used to assess the environmental
effect of process effluents on life forms.
These tests indicated (Table 2) that the
coal, ESP hopper ash, and sluice water
are not mutagenic. The Rat Alveolar
Macrophage (RAM) assays were also
negative. The Chinese Hamster Ovary
(CHO) cell assay yielded toxic responses
with coal leachates and sluice water.
Moderate to high toxicity was also
recorded for some of the aquatic assays.
Sluice water caused no toxicity in any of
the aquatic species studied, but stimulated
the growth of algae.
Corrosion Tests
To more fully evaluate the longer term
side effects of low-NO« operation, corrosion
testing began after the characterization
period, and included corrosion probes,
corrosion panels, and wall thickness
mapping.
With corrosion probes, as exposure
time increased, the rate of metal loss
seemed to decrease from more than 70 to
less than 20 mils/yr; thus, the loss rate
for the probes approached the loss rate
for the furnace walls (which the probes
were designed to simulate). The corrosion
panels had an average loss rate of 1.04
mils/yr, which compares well with the
average wall loss rate (Table 3) of 1.39
mils/yr.
Table 3 seems to indicate that tubes in
the burner zone corrode faster than non-
burner-zone tubes. However, the difference
is small, and no strong correlation was
found between corrosion rate and location
in the furnace. Also, no eccentric tube
wear or loss of roundness was noticed,
and the tubes close to the burners did not
corrode measurably faster than those
farther from the flame. On the average,
the rate of corrosion in this boHer was
such that it would require more than 30
years for the tubes to lose half their
thickness. Therefore, corrosion rates,
while operating in compliance with NSPS
NOx regulations, appear to be acceptable.
-------
Table 1. Metal Concentrations and Emissions in Various Conesville No. 5 'Process Streams
(Results in ftg of metal/g of each stream.)
Flue Gas Stream
Element
Ag
As
Be
Cd
Cr
Cu
Hg
Ni
Pb
Sb
Se
Tl
Zn
Al
Ba
Ca
Co
Fe
K
Li
Mg
Mn
Mo
Na
P
Si
Sn
Sr
Ti
V
10 and 3 ion
Fly Ash
Particles*
<1.8
83."
-------
Table 2. Conesville No. 5 Boiler Bioassay Results (Summary) fcont.J
Bottom Precipitator ESP Outlet
Ash/Water Slurry (ESP) Hopper Ash Coal To Stack (XAD-2)
Magna (Water Flea)
Fresh Water Algae
£Csoc (Dose
that kills 5O%)
EC™ (Dose
that kills 95%)
Exxon Acurex Exxon
Data Data Data
U H
2.5 to
6.7
mg/L
4.4 to
16.2
mg/L
Acurex Exxon Acurex Exxon
Data Data Data Data
H
328 to
367
mg/L
830 to
963
mg/L
Acurex
Data
'Not mutagenic, but toxic to cells. (Up to the time of cell death, no mutations had been observed.)
"Dose = 1000 mg of sample/liter of culture.
^Bottom ash affected acidity in Acurex's algae test:
O ash, pH = 7. /, and
25 mg/L, pH = 3.6.
Key
H = High toxicity
M = Moderate toxicity
L = Low toxicity
U = Undetected
- = Not tested (usually due to insufficient sample)
Table 3. Conesville No. 5 Boiler
Long Term Corrosion Results
(On-line Time = 12.OOO hr)
Wall Loss
Location
mils mils/yr
Nose
Burner
Zone
Hopper
Averages:
Burner Zone
Non-Burner Zone
Grand Average
0.8
2.0
2.1
1.8
2.9
1.5
2.3
1.4
1.9
0.6
1.5
1.6
1.4
2.2
1.1
1.7
1.1
1.4
and 182 ng/J for the first and second
tests. The standard deviation was about
10 percent of the mean value. The
statistical parameters in Table 4 are in
good agreement with the probability plots
shown in the full report.
Similarly, for the second test, the NO«
emissions rate and excess oxygen levels
were fairly steady: both showed standard
deviations of about 9 % of the mean value.
The CEM system (utilizing an extractive
sampling system) passed the EPA perfor-
mance specifications for accuracy, drift,
response time, etc. and provided accurate
gaseous emissions data for the test.
However, both calibration of the instru-
Table 4. Statistical Parameters for Conesville Unit No. 5
30-Day Continuous Emission Monitoring (CEM) Tests
First 30-Day Test
Thirty-day Continuous
Emission Monitor (CEM) Tests
Twice during the long-term operating
period, flue gas emissions were measured
for 30 days using continuous emissions
monitors (CEMs). The data proved useful
for evaluating boiler emissions over
normal load cycles for longer periods than
during the earlier characterization tests.
The CEM data also helped to evaluate the
abilities, potential problems, and perfor-
mance limitations that might be associated
with CEMs in similar situations. Table 4
summarizes the data.
Although the boiler steam load varied
widely (e.g., from about 80 to 40% of
maximum capacity during the first CEM
test), the NO emission rate was fairly
steady with mean values (Table 4) of 226
Mean
Minimum
Maximum
Range
Standard Deviation
Percent Deviation
Variance
Count
Load
MWth
490.7
646.0
311.4
334.6
101.1
20.6
10.214
29
02
6.3
5.7
8.6
2.9
0.9
13.0
0.81
29
COi
11.5
10.2
12.5
2.3
0.7
6.1
0.49
29
CO
ppm'
320
0
914
914
280
87.5
78400
29
ng/J
114
0
329
329
100
87.5
WOOD
29
ppm'
387
303
447
144
40
10.3
1600
29
NO
ng/J
226
177
261
84
23
10.3
52.9
29
b/irfBtu*
0.526
0.412
0.607
0.195
0.053
10.3
1.230
29
Second 30-Day Test
Mean
Minimum
Maximum
Range
Standard Deviation
Percent Deviation
Variance
Count
Load
MWo,
647.2
445.1
801.9
356.8
87.7
13.6
7693.0
35
Ot
%
5.63
4.5
7.0
2.5
0.524
9.3
0.275
35
COt
%
11.94
10.5
13.2
2.7
0.558
4.7
0.312
35
CO
ppm'
49.2
14
135
121
25.07
52.0
628.6
35
ng/J
17.2
5.0
48.0
43.0
8.86
52.0
78.56
35
ppm'
310.1
233
381
148
28.46
9.0
96172
34
NO
ng/J
182.1
137.0
223.0
86.0
16.32
9.0
266.3
34
Ib/KfBtif
0.42
0.32
0.52
0.20
0.04
9.0
0.0016
34
'Corrected to 3% Oa dry.
"As N02.
-------
ments on a daily basis and maintenance
of the sampling system were necessary.
Conclusions
Combustion modification effectively
reduces NO, emissions from coal-fired
utility boilers, without adverse side
effects. On the average, for the Conesville
boiler, tube wall corrosion rates from low-
NOx operation will not decrease or limit
the useful life of the boiler. The Level 1
test confirms that no unusual environmen-
tal hazards result from Iow-N0« operation.
The 30-day CEM tests show that,
through a good maintenance and calibration
program, CEMs can be made to meet the
EPA performance specifications.
Recommendations
When possible, combustion modifications
(CMs) should be considered as an
effective approach to NO* control on coal-
fired utility boilers. However, slagging
and other possible side effects should be
monitored to ensure that satisfactory
operation continues.
Because CM for NOX control requires
close observation and tight control on
boiler operations, a well maintained contin-
uous emission monitor (CEM) for CO
should be used to help decrease excess
air to optimum levels.
Conversion Factors
Readers more familiar with the metric
system may use the following factors to
convert units to that system:
Non-metric Times Yields metric
°F
Ib
mil
psi
5/9(°F-32)
0.454
24.5
0.070
°C
kg
//m
kg/cm2
P. S. Natanson is with Exxon Research and Engineering Co.. Florham Park, NJ
07932.
David G. Lachapelle is the EPA Project Officer (see below).
The complete report, entitled "Long Term Optimum Performance/Corrosion
Tests of Combustion Modifications for Utility Boilers; Host Site: Columbus
and Southern Ohio Electric Company. Conesville No. 5," (Order No. PB 84-
102 698; Cost: $26.50, subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park. NC27711
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati-OH 45268
Official Business
Penalty for Private Use $300
U.S. GOVERNMENT PRINTING OFFICE: 1983-759
102/BCJ
------- |